1GCC(1)                                GNU                               GCC(1)
2
3
4

NAME

6       gcc - GNU project C and C++ compiler
7

SYNOPSIS

9       gcc [-c|-S|-E] [-std=standard]
10           [-g] [-pg] [-Olevel]
11           [-Wwarn...] [-Wpedantic]
12           [-Idir...] [-Ldir...]
13           [-Dmacro[=defn]...] [-Umacro]
14           [-foption...] [-mmachine-option...]
15           [-o outfile] [@file] infile...
16
17       Only the most useful options are listed here; see below for the
18       remainder.  g++ accepts mostly the same options as gcc.
19

DESCRIPTION

21       When you invoke GCC, it normally does preprocessing, compilation,
22       assembly and linking.  The "overall options" allow you to stop this
23       process at an intermediate stage.  For example, the -c option says not
24       to run the linker.  Then the output consists of object files output by
25       the assembler.
26
27       Other options are passed on to one or more stages of processing.  Some
28       options control the preprocessor and others the compiler itself.  Yet
29       other options control the assembler and linker; most of these are not
30       documented here, since you rarely need to use any of them.
31
32       Most of the command-line options that you can use with GCC are useful
33       for C programs; when an option is only useful with another language
34       (usually C++), the explanation says so explicitly.  If the description
35       for a particular option does not mention a source language, you can use
36       that option with all supported languages.
37
38       The usual way to run GCC is to run the executable called gcc, or
39       machine-gcc when cross-compiling, or machine-gcc-version to run a
40       specific version of GCC.  When you compile C++ programs, you should
41       invoke GCC as g++ instead.
42
43       The gcc program accepts options and file names as operands.  Many
44       options have multi-letter names; therefore multiple single-letter
45       options may not be grouped: -dv is very different from -d -v.
46
47       You can mix options and other arguments.  For the most part, the order
48       you use doesn't matter.  Order does matter when you use several options
49       of the same kind; for example, if you specify -L more than once, the
50       directories are searched in the order specified.  Also, the placement
51       of the -l option is significant.
52
53       Many options have long names starting with -f or with -W---for example,
54       -fmove-loop-invariants, -Wformat and so on.  Most of these have both
55       positive and negative forms; the negative form of -ffoo is -fno-foo.
56       This manual documents only one of these two forms, whichever one is not
57       the default.
58
59       Some options take one or more arguments typically separated either by a
60       space or by the equals sign (=) from the option name.  Unless
61       documented otherwise, an argument can be either numeric or a string.
62       Numeric arguments must typically be small unsigned decimal or
63       hexadecimal integers.  Hexadecimal arguments must begin with the 0x
64       prefix.  Arguments to options that specify a size threshold of some
65       sort may be arbitrarily large decimal or hexadecimal integers followed
66       by a byte size suffix designating a multiple of bytes such as "kB" and
67       "KiB" for kilobyte and kibibyte, respectively, "MB" and "MiB" for
68       megabyte and mebibyte, "GB" and "GiB" for gigabyte and gigibyte, and so
69       on.  Such arguments are designated by byte-size in the following text.
70       Refer to the NIST, IEC, and other relevant national and international
71       standards for the full listing and explanation of the binary and
72       decimal byte size prefixes.
73

OPTIONS

75   Option Summary
76       Here is a summary of all the options, grouped by type.  Explanations
77       are in the following sections.
78
79       Overall Options
80           -c  -S  -E  -o file  -x language -v  -###  --help[=class[,...]]
81           --target-help  --version -pass-exit-codes  -pipe  -specs=file
82           -wrapper @file  -ffile-prefix-map=old=new -fplugin=file
83           -fplugin-arg-name=arg -fdump-ada-spec[-slim]
84           -fada-spec-parent=unit  -fdump-go-spec=file
85
86       C Language Options
87           -ansi  -std=standard  -fgnu89-inline
88           -fpermitted-flt-eval-methods=standard -aux-info filename
89           -fallow-parameterless-variadic-functions -fno-asm  -fno-builtin
90           -fno-builtin-function  -fgimple -fhosted  -ffreestanding -fopenacc
91           -fopenacc-dim=geom -fopenmp  -fopenmp-simd -fms-extensions
92           -fplan9-extensions  -fsso-struct=endianness
93           -fallow-single-precision  -fcond-mismatch  -flax-vector-conversions
94           -fsigned-bitfields  -fsigned-char -funsigned-bitfields
95           -funsigned-char
96
97       C++ Language Options
98           -fabi-version=n  -fno-access-control -faligned-new=n
99           -fargs-in-order=n  -fchar8_t  -fcheck-new -fconstexpr-depth=n
100           -fconstexpr-cache-depth=n -fconstexpr-loop-limit=n
101           -fconstexpr-ops-limit=n -fno-elide-constructors
102           -fno-enforce-eh-specs -fno-gnu-keywords -fno-implicit-templates
103           -fno-implicit-inline-templates -fno-implement-inlines
104           -fms-extensions -fnew-inheriting-ctors -fnew-ttp-matching
105           -fno-nonansi-builtins  -fnothrow-opt  -fno-operator-names
106           -fno-optional-diags  -fpermissive -fno-pretty-templates -fno-rtti
107           -fsized-deallocation -ftemplate-backtrace-limit=n
108           -ftemplate-depth=n -fno-threadsafe-statics  -fuse-cxa-atexit
109           -fno-weak  -nostdinc++ -fvisibility-inlines-hidden
110           -fvisibility-ms-compat -fext-numeric-literals -Wabi-tag
111           -Wcatch-value  -Wcatch-value=n -Wno-class-conversion
112           -Wclass-memaccess -Wcomma-subscript  -Wconditionally-supported
113           -Wno-conversion-null  -Wctor-dtor-privacy  -Wno-delete-incomplete
114           -Wdelete-non-virtual-dtor  -Wdeprecated-copy
115           -Wdeprecated-copy-dtor -Weffc++  -Wextra-semi
116           -Wno-inaccessible-base -Wno-inherited-variadic-ctor
117           -Wno-init-list-lifetime -Wno-invalid-offsetof  -Wno-literal-suffix
118           -Wmismatched-tags -Wmultiple-inheritance  -Wnamespaces  -Wnarrowing
119           -Wnoexcept  -Wnoexcept-type  -Wnon-virtual-dtor -Wpessimizing-move
120           -Wno-placement-new  -Wplacement-new=n -Wredundant-move
121           -Wredundant-tags -Wreorder  -Wregister -Wstrict-null-sentinel
122           -Wno-subobject-linkage  -Wtemplates -Wno-non-template-friend
123           -Wold-style-cast -Woverloaded-virtual  -Wno-pmf-conversions
124           -Wsign-promo -Wsized-deallocation  -Wsuggest-final-methods
125           -Wsuggest-final-types  -Wsuggest-override -Wno-terminate
126           -Wuseless-cast  -Wvirtual-inheritance -Wno-virtual-move-assign
127           -Wvolatile  -Wzero-as-null-pointer-constant
128
129       Objective-C and Objective-C++ Language Options
130           -fconstant-string-class=class-name -fgnu-runtime  -fnext-runtime
131           -fno-nil-receivers -fobjc-abi-version=n -fobjc-call-cxx-cdtors
132           -fobjc-direct-dispatch -fobjc-exceptions -fobjc-gc -fobjc-nilcheck
133           -fobjc-std=objc1 -fno-local-ivars
134           -fivar-visibility=[public|protected|private|package]
135           -freplace-objc-classes -fzero-link -gen-decls -Wassign-intercept
136           -Wno-property-assign-default -Wno-protocol  -Wselector
137           -Wstrict-selector-match -Wundeclared-selector
138
139       Diagnostic Message Formatting Options
140           -fmessage-length=n -fdiagnostics-show-location=[once|every-line]
141           -fdiagnostics-color=[auto|never|always]
142           -fdiagnostics-urls=[auto|never|always]
143           -fdiagnostics-format=[text|json] -fno-diagnostics-show-option
144           -fno-diagnostics-show-caret -fno-diagnostics-show-labels
145           -fno-diagnostics-show-line-numbers -fno-diagnostics-show-cwe
146           -fdiagnostics-minimum-margin-width=width
147           -fdiagnostics-parseable-fixits  -fdiagnostics-generate-patch
148           -fdiagnostics-show-template-tree  -fno-elide-type
149           -fdiagnostics-path-format=[none|separate-events|inline-events]
150           -fdiagnostics-show-path-depths -fno-show-column
151
152       Warning Options
153           -fsyntax-only  -fmax-errors=n  -Wpedantic -pedantic-errors -w
154           -Wextra  -Wall  -Wabi=n -Waddress  -Wno-address-of-packed-member
155           -Waggregate-return -Walloc-size-larger-than=byte-size  -Walloc-zero
156           -Walloca  -Walloca-larger-than=byte-size
157           -Wno-aggressive-loop-optimizations -Warith-conversion
158           -Warray-bounds  -Warray-bounds=n -Wno-attributes
159           -Wattribute-alias=n -Wno-attribute-alias -Wno-attribute-warning
160           -Wbool-compare  -Wbool-operation -Wno-builtin-declaration-mismatch
161           -Wno-builtin-macro-redefined  -Wc90-c99-compat  -Wc99-c11-compat
162           -Wc11-c2x-compat -Wc++-compat  -Wc++11-compat  -Wc++14-compat
163           -Wc++17-compat -Wc++20-compat -Wcast-align  -Wcast-align=strict
164           -Wcast-function-type  -Wcast-qual -Wchar-subscripts -Wclobbered
165           -Wcomment -Wconversion  -Wno-coverage-mismatch  -Wno-cpp
166           -Wdangling-else  -Wdate-time -Wno-deprecated
167           -Wno-deprecated-declarations  -Wno-designated-init
168           -Wdisabled-optimization -Wno-discarded-array-qualifiers
169           -Wno-discarded-qualifiers -Wno-div-by-zero  -Wdouble-promotion
170           -Wduplicated-branches  -Wduplicated-cond -Wempty-body
171           -Wno-endif-labels  -Wenum-compare  -Wenum-conversion -Werror
172           -Werror=*  -Wexpansion-to-defined  -Wfatal-errors
173           -Wfloat-conversion  -Wfloat-equal  -Wformat  -Wformat=2
174           -Wno-format-contains-nul  -Wno-format-extra-args
175           -Wformat-nonliteral  -Wformat-overflow=n -Wformat-security
176           -Wformat-signedness  -Wformat-truncation=n -Wformat-y2k
177           -Wframe-address -Wframe-larger-than=byte-size
178           -Wno-free-nonheap-object -Wno-hsa  -Wno-if-not-aligned
179           -Wno-ignored-attributes -Wignored-qualifiers
180           -Wno-incompatible-pointer-types -Wimplicit  -Wimplicit-fallthrough
181           -Wimplicit-fallthrough=n -Wno-implicit-function-declaration
182           -Wno-implicit-int -Winit-self  -Winline  -Wno-int-conversion
183           -Wint-in-bool-context -Wno-int-to-pointer-cast
184           -Wno-invalid-memory-model -Winvalid-pch  -Wjump-misses-init
185           -Wlarger-than=byte-size -Wlogical-not-parentheses  -Wlogical-op
186           -Wlong-long -Wno-lto-type-mismatch -Wmain  -Wmaybe-uninitialized
187           -Wmemset-elt-size  -Wmemset-transposed-args
188           -Wmisleading-indentation  -Wmissing-attributes  -Wmissing-braces
189           -Wmissing-field-initializers  -Wmissing-format-attribute
190           -Wmissing-include-dirs  -Wmissing-noreturn  -Wno-missing-profile
191           -Wno-multichar  -Wmultistatement-macros  -Wnonnull
192           -Wnonnull-compare -Wnormalized=[none|id|nfc|nfkc]
193           -Wnull-dereference  -Wno-odr  -Wopenmp-simd -Wno-overflow
194           -Woverlength-strings  -Wno-override-init-side-effects -Wpacked
195           -Wno-packed-bitfield-compat  -Wpacked-not-aligned  -Wpadded
196           -Wparentheses  -Wno-pedantic-ms-format -Wpointer-arith
197           -Wno-pointer-compare  -Wno-pointer-to-int-cast -Wno-pragmas
198           -Wno-prio-ctor-dtor  -Wredundant-decls -Wrestrict
199           -Wno-return-local-addr  -Wreturn-type -Wno-scalar-storage-order
200           -Wsequence-point -Wshadow  -Wshadow=global  -Wshadow=local
201           -Wshadow=compatible-local -Wno-shadow-ivar
202           -Wno-shift-count-negative  -Wno-shift-count-overflow
203           -Wshift-negative-value -Wno-shift-overflow  -Wshift-overflow=n
204           -Wsign-compare  -Wsign-conversion -Wno-sizeof-array-argument
205           -Wsizeof-pointer-div  -Wsizeof-pointer-memaccess -Wstack-protector
206           -Wstack-usage=byte-size  -Wstrict-aliasing -Wstrict-aliasing=n
207           -Wstrict-overflow  -Wstrict-overflow=n -Wstring-compare
208           -Wstringop-overflow=n  -Wno-stringop-truncation
209           -Wsuggest-attribute=[pure|const|noreturn|format|malloc] -Wswitch
210           -Wno-switch-bool  -Wswitch-default  -Wswitch-enum
211           -Wno-switch-outside-range  -Wno-switch-unreachable  -Wsync-nand
212           -Wsystem-headers  -Wtautological-compare  -Wtrampolines
213           -Wtrigraphs -Wtype-limits  -Wundef -Wuninitialized
214           -Wunknown-pragmas -Wunsuffixed-float-constants  -Wunused
215           -Wunused-but-set-parameter  -Wunused-but-set-variable
216           -Wunused-const-variable  -Wunused-const-variable=n
217           -Wunused-function  -Wunused-label  -Wunused-local-typedefs
218           -Wunused-macros -Wunused-parameter  -Wno-unused-result
219           -Wunused-value  -Wunused-variable -Wno-varargs  -Wvariadic-macros
220           -Wvector-operation-performance -Wvla  -Wvla-larger-than=byte-size
221           -Wno-vla-larger-than -Wvolatile-register-var  -Wwrite-strings
222           -Wzero-length-bounds
223
224       Static Analyzer Options
225           -fanalyzer -fanalyzer-call-summaries -fanalyzer-checker=name
226           -fanalyzer-fine-grained -fanalyzer-state-merge
227           -fanalyzer-state-purge -fanalyzer-transitivity
228           -fanalyzer-verbose-edges -fanalyzer-verbose-state-changes
229           -fanalyzer-verbosity=level -fdump-analyzer -fdump-analyzer-stderr
230           -fdump-analyzer-callgraph -fdump-analyzer-exploded-graph
231           -fdump-analyzer-exploded-nodes -fdump-analyzer-exploded-nodes-2
232           -fdump-analyzer-exploded-nodes-3 -fdump-analyzer-state-purge
233           -fdump-analyzer-supergraph -Wno-analyzer-double-fclose
234           -Wno-analyzer-double-free
235           -Wno-analyzer-exposure-through-output-file -Wno-analyzer-file-leak
236           -Wno-analyzer-free-of-non-heap -Wno-analyzer-malloc-leak
237           -Wno-analyzer-null-argument -Wno-analyzer-null-dereference
238           -Wno-analyzer-possible-null-argument
239           -Wno-analyzer-possible-null-dereference
240           -Wno-analyzer-stale-setjmp-buffer -Wno-analyzer-tainted-array-index
241           -Wanalyzer-too-complex
242           -Wno-analyzer-unsafe-call-within-signal-handler
243           -Wno-analyzer-use-after-free
244           -Wno-analyzer-use-of-pointer-in-stale-stack-frame
245           -Wno-analyzer-use-of-uninitialized-value
246
247       C and Objective-C-only Warning Options
248           -Wbad-function-cast  -Wmissing-declarations
249           -Wmissing-parameter-type  -Wmissing-prototypes  -Wnested-externs
250           -Wold-style-declaration  -Wold-style-definition -Wstrict-prototypes
251           -Wtraditional  -Wtraditional-conversion
252           -Wdeclaration-after-statement  -Wpointer-sign
253
254       Debugging Options
255           -g  -glevel  -gdwarf  -gdwarf-version -ggdb  -grecord-gcc-switches
256           -gno-record-gcc-switches -gstabs  -gstabs+  -gstrict-dwarf
257           -gno-strict-dwarf -gas-loc-support  -gno-as-loc-support
258           -gas-locview-support  -gno-as-locview-support -gcolumn-info
259           -gno-column-info -gstatement-frontiers  -gno-statement-frontiers
260           -gvariable-location-views  -gno-variable-location-views
261           -ginternal-reset-location-views  -gno-internal-reset-location-views
262           -ginline-points  -gno-inline-points -gvms  -gxcoff  -gxcoff+
263           -gz[=type] -gsplit-dwarf  -gdescribe-dies  -gno-describe-dies
264           -fdebug-prefix-map=old=new  -fdebug-types-section
265           -fno-eliminate-unused-debug-types -femit-struct-debug-baseonly
266           -femit-struct-debug-reduced -femit-struct-debug-detailed[=spec-
267           list] -fno-eliminate-unused-debug-symbols
268           -femit-class-debug-always -fno-merge-debug-strings
269           -fno-dwarf2-cfi-asm -fvar-tracking  -fvar-tracking-assignments
270
271       Optimization Options
272           -faggressive-loop-optimizations -falign-functions[=n[:m:[n2[:m2]]]]
273           -falign-jumps[=n[:m:[n2[:m2]]]] -falign-labels[=n[:m:[n2[:m2]]]]
274           -falign-loops[=n[:m:[n2[:m2]]]] -fno-allocation-dce
275           -fallow-store-data-races -fassociative-math  -fauto-profile
276           -fauto-profile[=path] -fauto-inc-dec  -fbranch-probabilities
277           -fcaller-saves -fcombine-stack-adjustments  -fconserve-stack
278           -fcompare-elim  -fcprop-registers  -fcrossjumping
279           -fcse-follow-jumps  -fcse-skip-blocks  -fcx-fortran-rules
280           -fcx-limited-range -fdata-sections  -fdce  -fdelayed-branch
281           -fdelete-null-pointer-checks  -fdevirtualize
282           -fdevirtualize-speculatively -fdevirtualize-at-ltrans  -fdse
283           -fearly-inlining  -fipa-sra  -fexpensive-optimizations
284           -ffat-lto-objects -ffast-math  -ffinite-math-only  -ffloat-store
285           -fexcess-precision=style -ffinite-loops -fforward-propagate
286           -ffp-contract=style  -ffunction-sections -fgcse
287           -fgcse-after-reload  -fgcse-las  -fgcse-lm  -fgraphite-identity
288           -fgcse-sm  -fhoist-adjacent-loads  -fif-conversion -fif-conversion2
289           -findirect-inlining -finline-functions
290           -finline-functions-called-once  -finline-limit=n
291           -finline-small-functions  -fipa-cp  -fipa-cp-clone -fipa-bit-cp
292           -fipa-vrp  -fipa-pta  -fipa-profile  -fipa-pure-const
293           -fipa-reference  -fipa-reference-addressable -fipa-stack-alignment
294           -fipa-icf  -fira-algorithm=algorithm -flive-patching=level
295           -fira-region=region  -fira-hoist-pressure -fira-loop-pressure
296           -fno-ira-share-save-slots -fno-ira-share-spill-slots
297           -fisolate-erroneous-paths-dereference
298           -fisolate-erroneous-paths-attribute -fivopts
299           -fkeep-inline-functions  -fkeep-static-functions
300           -fkeep-static-consts  -flimit-function-alignment
301           -flive-range-shrinkage -floop-block  -floop-interchange
302           -floop-strip-mine -floop-unroll-and-jam  -floop-nest-optimize
303           -floop-parallelize-all  -flra-remat  -flto  -flto-compression-level
304           -flto-partition=alg  -fmerge-all-constants -fmerge-constants
305           -fmodulo-sched  -fmodulo-sched-allow-regmoves
306           -fmove-loop-invariants  -fno-branch-count-reg -fno-defer-pop
307           -fno-fp-int-builtin-inexact  -fno-function-cse
308           -fno-guess-branch-probability  -fno-inline  -fno-math-errno
309           -fno-peephole -fno-peephole2  -fno-printf-return-value
310           -fno-sched-interblock -fno-sched-spec  -fno-signed-zeros
311           -fno-toplevel-reorder  -fno-trapping-math
312           -fno-zero-initialized-in-bss -fomit-frame-pointer
313           -foptimize-sibling-calls -fpartial-inlining  -fpeel-loops
314           -fpredictive-commoning -fprefetch-loop-arrays -fprofile-correction
315           -fprofile-use  -fprofile-use=path -fprofile-partial-training
316           -fprofile-values -fprofile-reorder-functions -freciprocal-math
317           -free  -frename-registers  -freorder-blocks
318           -freorder-blocks-algorithm=algorithm -freorder-blocks-and-partition
319           -freorder-functions -frerun-cse-after-loop
320           -freschedule-modulo-scheduled-loops -frounding-math
321           -fsave-optimization-record -fsched2-use-superblocks
322           -fsched-pressure -fsched-spec-load  -fsched-spec-load-dangerous
323           -fsched-stalled-insns-dep[=n]  -fsched-stalled-insns[=n]
324           -fsched-group-heuristic  -fsched-critical-path-heuristic
325           -fsched-spec-insn-heuristic  -fsched-rank-heuristic
326           -fsched-last-insn-heuristic  -fsched-dep-count-heuristic
327           -fschedule-fusion -fschedule-insns  -fschedule-insns2
328           -fsection-anchors -fselective-scheduling  -fselective-scheduling2
329           -fsel-sched-pipelining  -fsel-sched-pipelining-outer-loops
330           -fsemantic-interposition  -fshrink-wrap  -fshrink-wrap-separate
331           -fsignaling-nans -fsingle-precision-constant
332           -fsplit-ivs-in-unroller  -fsplit-loops -fsplit-paths
333           -fsplit-wide-types  -fsplit-wide-types-early  -fssa-backprop
334           -fssa-phiopt -fstdarg-opt  -fstore-merging  -fstrict-aliasing
335           -fthread-jumps  -ftracer  -ftree-bit-ccp -ftree-builtin-call-dce
336           -ftree-ccp  -ftree-ch -ftree-coalesce-vars  -ftree-copy-prop
337           -ftree-dce  -ftree-dominator-opts -ftree-dse  -ftree-forwprop
338           -ftree-fre  -fcode-hoisting -ftree-loop-if-convert  -ftree-loop-im
339           -ftree-phiprop  -ftree-loop-distribution
340           -ftree-loop-distribute-patterns -ftree-loop-ivcanon
341           -ftree-loop-linear  -ftree-loop-optimize -ftree-loop-vectorize
342           -ftree-parallelize-loops=n  -ftree-pre  -ftree-partial-pre
343           -ftree-pta -ftree-reassoc  -ftree-scev-cprop  -ftree-sink
344           -ftree-slsr  -ftree-sra -ftree-switch-conversion  -ftree-tail-merge
345           -ftree-ter  -ftree-vectorize  -ftree-vrp  -funconstrained-commons
346           -funit-at-a-time  -funroll-all-loops  -funroll-loops
347           -funsafe-math-optimizations  -funswitch-loops -fipa-ra
348           -fvariable-expansion-in-unroller  -fvect-cost-model  -fvpt -fweb
349           -fwhole-program  -fwpa  -fuse-linker-plugin --param name=value -O
350           -O0  -O1  -O2  -O3  -Os  -Ofast  -Og
351
352       Program Instrumentation Options
353           -p  -pg  -fprofile-arcs  --coverage  -ftest-coverage
354           -fprofile-abs-path -fprofile-dir=path  -fprofile-generate
355           -fprofile-generate=path -fprofile-note=path
356           -fprofile-prefix-path=path -fprofile-update=method
357           -fprofile-filter-files=regex -fprofile-exclude-files=regex
358           -fprofile-reproducibility -fsanitize=style  -fsanitize-recover
359           -fsanitize-recover=style -fasan-shadow-offset=number
360           -fsanitize-sections=s1,s2,...  -fsanitize-undefined-trap-on-error
361           -fbounds-check -fcf-protection=[full|branch|return|none]
362           -fstack-protector  -fstack-protector-all  -fstack-protector-strong
363           -fstack-protector-explicit  -fstack-check
364           -fstack-limit-register=reg  -fstack-limit-symbol=sym
365           -fno-stack-limit  -fsplit-stack -fvtable-verify=[std|preinit|none]
366           -fvtv-counts  -fvtv-debug -finstrument-functions
367           -finstrument-functions-exclude-function-list=sym,sym,...
368           -finstrument-functions-exclude-file-list=file,file,...
369
370       Preprocessor Options
371           -Aquestion=answer -A-question[=answer] -C  -CC  -Dmacro[=defn] -dD
372           -dI  -dM  -dN  -dU -fdebug-cpp  -fdirectives-only
373           -fdollars-in-identifiers -fexec-charset=charset
374           -fextended-identifiers -finput-charset=charset
375           -fmacro-prefix-map=old=new -fmax-include-depth=depth
376           -fno-canonical-system-headers  -fpch-deps  -fpch-preprocess
377           -fpreprocessed  -ftabstop=width  -ftrack-macro-expansion
378           -fwide-exec-charset=charset  -fworking-directory -H  -imacros file
379           -include file -M  -MD  -MF  -MG  -MM  -MMD  -MP  -MQ  -MT
380           -no-integrated-cpp  -P  -pthread  -remap -traditional
381           -traditional-cpp  -trigraphs -Umacro  -undef -Wp,option
382           -Xpreprocessor option
383
384       Assembler Options
385           -Wa,option  -Xassembler option
386
387       Linker Options
388           object-file-name  -fuse-ld=linker  -llibrary -nostartfiles
389           -nodefaultlibs  -nolibc  -nostdlib -e entry  --entry=entry -pie
390           -pthread  -r  -rdynamic -s  -static  -static-pie  -static-libgcc
391           -static-libstdc++ -static-libasan  -static-libtsan  -static-liblsan
392           -static-libubsan -shared  -shared-libgcc  -symbolic -T script
393           -Wl,option  -Xlinker option -u symbol  -z keyword
394
395       Directory Options
396           -Bprefix  -Idir  -I- -idirafter dir -imacros file  -imultilib dir
397           -iplugindir=dir  -iprefix file -iquote dir  -isysroot dir  -isystem
398           dir -iwithprefix dir  -iwithprefixbefore dir -Ldir
399           -no-canonical-prefixes  --no-sysroot-suffix -nostdinc  -nostdinc++
400           --sysroot=dir
401
402       Code Generation Options
403           -fcall-saved-reg  -fcall-used-reg -ffixed-reg  -fexceptions
404           -fnon-call-exceptions  -fdelete-dead-exceptions  -funwind-tables
405           -fasynchronous-unwind-tables -fno-gnu-unique
406           -finhibit-size-directive  -fcommon  -fno-ident -fpcc-struct-return
407           -fpic  -fPIC  -fpie  -fPIE  -fno-plt -fno-jump-tables
408           -frecord-gcc-switches -freg-struct-return  -fshort-enums
409           -fshort-wchar -fverbose-asm  -fpack-struct[=n] -fleading-underscore
410           -ftls-model=model -fstack-reuse=reuse_level -ftrampolines  -ftrapv
411           -fwrapv -fvisibility=[default|internal|hidden|protected]
412           -fstrict-volatile-bitfields  -fsync-libcalls
413
414       Developer Options
415           -dletters  -dumpspecs  -dumpmachine  -dumpversion -dumpfullversion
416           -fcallgraph-info[=su,da] -fchecking  -fchecking=n -fdbg-cnt-list
417           -fdbg-cnt=counter-value-list -fdisable-ipa-pass_name
418           -fdisable-rtl-pass_name -fdisable-rtl-pass-name=range-list
419           -fdisable-tree-pass_name -fdisable-tree-pass-name=range-list
420           -fdump-debug  -fdump-earlydebug -fdump-noaddr  -fdump-unnumbered
421           -fdump-unnumbered-links -fdump-final-insns[=file] -fdump-ipa-all
422           -fdump-ipa-cgraph  -fdump-ipa-inline -fdump-lang-all
423           -fdump-lang-switch -fdump-lang-switch-options
424           -fdump-lang-switch-options=filename -fdump-passes -fdump-rtl-pass
425           -fdump-rtl-pass=filename -fdump-statistics -fdump-tree-all
426           -fdump-tree-switch -fdump-tree-switch-options
427           -fdump-tree-switch-options=filename -fcompare-debug[=opts]
428           -fcompare-debug-second -fenable-kind-pass -fenable-kind-pass=range-
429           list -fira-verbose=n -flto-report  -flto-report-wpa
430           -fmem-report-wpa -fmem-report  -fpre-ipa-mem-report
431           -fpost-ipa-mem-report -fopt-info  -fopt-info-options[=file]
432           -fprofile-report -frandom-seed=string  -fsched-verbose=n
433           -fsel-sched-verbose  -fsel-sched-dump-cfg
434           -fsel-sched-pipelining-verbose -fstats  -fstack-usage
435           -ftime-report  -ftime-report-details
436           -fvar-tracking-assignments-toggle  -gtoggle
437           -print-file-name=library  -print-libgcc-file-name
438           -print-multi-directory  -print-multi-lib  -print-multi-os-directory
439           -print-prog-name=program  -print-search-dirs  -Q -print-sysroot
440           -print-sysroot-headers-suffix -save-temps  -save-temps=cwd
441           -save-temps=obj  -time[=file]
442
443       Machine-Dependent Options
444           AArch64 Options -mabi=name  -mbig-endian  -mlittle-endian
445           -mgeneral-regs-only -mcmodel=tiny  -mcmodel=small  -mcmodel=large
446           -mstrict-align  -mno-strict-align -momit-leaf-frame-pointer
447           -mtls-dialect=desc  -mtls-dialect=traditional -mtls-size=size
448           -mfix-cortex-a53-835769  -mfix-cortex-a53-843419
449           -mlow-precision-recip-sqrt  -mlow-precision-sqrt
450           -mlow-precision-div -mpc-relative-literal-loads
451           -msign-return-address=scope -mbranch-protection=none|standard|pac-
452           ret[+leaf +b-key]|bti -march=name  -mcpu=name  -mtune=name
453           -moverride=string  -mverbose-cost-dump
454           -mstack-protector-guard=guard -mstack-protector-guard-reg=sysreg
455           -mstack-protector-guard-offset=offset -mtrack-speculation
456           -moutline-atomics
457
458           Adapteva Epiphany Options -mhalf-reg-file  -mprefer-short-insn-regs
459           -mbranch-cost=num  -mcmove  -mnops=num  -msoft-cmpsf -msplit-lohi
460           -mpost-inc  -mpost-modify  -mstack-offset=num -mround-nearest
461           -mlong-calls  -mshort-calls  -msmall16 -mfp-mode=mode
462           -mvect-double  -max-vect-align=num -msplit-vecmove-early
463           -m1reg-reg
464
465           AMD GCN Options -march=gpu -mtune=gpu -mstack-size=bytes
466
467           ARC Options -mbarrel-shifter  -mjli-always -mcpu=cpu  -mA6
468           -mARC600  -mA7  -mARC700 -mdpfp  -mdpfp-compact  -mdpfp-fast
469           -mno-dpfp-lrsr -mea  -mno-mpy  -mmul32x16  -mmul64  -matomic -mnorm
470           -mspfp  -mspfp-compact  -mspfp-fast  -msimd  -msoft-float  -mswap
471           -mcrc  -mdsp-packa  -mdvbf  -mlock  -mmac-d16  -mmac-24  -mrtsc
472           -mswape -mtelephony  -mxy  -misize  -mannotate-align  -marclinux
473           -marclinux_prof -mlong-calls  -mmedium-calls  -msdata
474           -mirq-ctrl-saved -mrgf-banked-regs  -mlpc-width=width  -G num
475           -mvolatile-cache  -mtp-regno=regno -malign-call  -mauto-modify-reg
476           -mbbit-peephole  -mno-brcc -mcase-vector-pcrel  -mcompact-casesi
477           -mno-cond-exec  -mearly-cbranchsi -mexpand-adddi  -mindexed-loads
478           -mlra  -mlra-priority-none -mlra-priority-compact mlra-priority-
479           noncompact  -mmillicode -mmixed-code  -mq-class  -mRcq  -mRcw
480           -msize-level=level -mtune=cpu  -mmultcost=num  -mcode-density-frame
481           -munalign-prob-threshold=probability  -mmpy-option=multo -mdiv-rem
482           -mcode-density  -mll64  -mfpu=fpu  -mrf16  -mbranch-index
483
484           ARM Options -mapcs-frame  -mno-apcs-frame -mabi=name
485           -mapcs-stack-check  -mno-apcs-stack-check -mapcs-reentrant
486           -mno-apcs-reentrant -mgeneral-regs-only -msched-prolog
487           -mno-sched-prolog -mlittle-endian  -mbig-endian -mbe8  -mbe32
488           -mfloat-abi=name -mfp16-format=name -mthumb-interwork
489           -mno-thumb-interwork -mcpu=name  -march=name  -mfpu=name
490           -mtune=name  -mprint-tune-info -mstructure-size-boundary=n
491           -mabort-on-noreturn -mlong-calls  -mno-long-calls -msingle-pic-base
492           -mno-single-pic-base -mpic-register=reg -mnop-fun-dllimport
493           -mpoke-function-name -mthumb  -marm  -mflip-thumb -mtpcs-frame
494           -mtpcs-leaf-frame -mcaller-super-interworking
495           -mcallee-super-interworking -mtp=name  -mtls-dialect=dialect
496           -mword-relocations -mfix-cortex-m3-ldrd -munaligned-access
497           -mneon-for-64bits -mslow-flash-data -masm-syntax-unified
498           -mrestrict-it -mverbose-cost-dump -mpure-code -mcmse -mfdpic
499
500           AVR Options -mmcu=mcu  -mabsdata  -maccumulate-args
501           -mbranch-cost=cost -mcall-prologues  -mgas-isr-prologues  -mint8
502           -mdouble=bits -mlong-double=bits -mn_flash=size  -mno-interrupts
503           -mmain-is-OS_task  -mrelax  -mrmw  -mstrict-X  -mtiny-stack
504           -mfract-convert-truncate -mshort-calls  -nodevicelib
505           -nodevicespecs -Waddr-space-convert  -Wmisspelled-isr
506
507           Blackfin Options -mcpu=cpu[-sirevision] -msim
508           -momit-leaf-frame-pointer  -mno-omit-leaf-frame-pointer
509           -mspecld-anomaly  -mno-specld-anomaly  -mcsync-anomaly
510           -mno-csync-anomaly -mlow-64k  -mno-low64k  -mstack-check-l1
511           -mid-shared-library -mno-id-shared-library  -mshared-library-id=n
512           -mleaf-id-shared-library  -mno-leaf-id-shared-library -msep-data
513           -mno-sep-data  -mlong-calls  -mno-long-calls -mfast-fp
514           -minline-plt  -mmulticore  -mcorea  -mcoreb  -msdram -micplb
515
516           C6X Options -mbig-endian  -mlittle-endian  -march=cpu -msim
517           -msdata=sdata-type
518
519           CRIS Options -mcpu=cpu  -march=cpu  -mtune=cpu -mmax-stack-frame=n
520           -melinux-stacksize=n -metrax4  -metrax100  -mpdebug  -mcc-init
521           -mno-side-effects -mstack-align  -mdata-align  -mconst-align
522           -m32-bit  -m16-bit  -m8-bit  -mno-prologue-epilogue  -mno-gotplt
523           -melf  -maout  -melinux  -mlinux  -sim  -sim2 -mmul-bug-workaround
524           -mno-mul-bug-workaround
525
526           CR16 Options -mmac -mcr16cplus  -mcr16c -msim  -mint32  -mbit-ops
527           -mdata-model=model
528
529           C-SKY Options -march=arch  -mcpu=cpu -mbig-endian  -EB
530           -mlittle-endian  -EL -mhard-float  -msoft-float  -mfpu=fpu
531           -mdouble-float  -mfdivdu -melrw  -mistack  -mmp  -mcp  -mcache
532           -msecurity  -mtrust -mdsp  -medsp  -mvdsp -mdiv  -msmart
533           -mhigh-registers  -manchor -mpushpop  -mmultiple-stld  -mconstpool
534           -mstack-size  -mccrt -mbranch-cost=n  -mcse-cc  -msched-prolog
535
536           Darwin Options -all_load  -allowable_client  -arch
537           -arch_errors_fatal -arch_only  -bind_at_load  -bundle
538           -bundle_loader -client_name  -compatibility_version
539           -current_version -dead_strip -dependency-file  -dylib_file
540           -dylinker_install_name -dynamic  -dynamiclib
541           -exported_symbols_list -filelist  -flat_namespace
542           -force_cpusubtype_ALL -force_flat_namespace
543           -headerpad_max_install_names -iframework -image_base  -init
544           -install_name  -keep_private_externs -multi_module
545           -multiply_defined  -multiply_defined_unused -noall_load
546           -no_dead_strip_inits_and_terms -nofixprebinding  -nomultidefs
547           -noprebind  -noseglinkedit -pagezero_size  -prebind
548           -prebind_all_twolevel_modules -private_bundle  -read_only_relocs
549           -sectalign -sectobjectsymbols  -whyload  -seg1addr -sectcreate
550           -sectobjectsymbols  -sectorder -segaddr  -segs_read_only_addr
551           -segs_read_write_addr -seg_addr_table  -seg_addr_table_filename
552           -seglinkedit -segprot  -segs_read_only_addr  -segs_read_write_addr
553           -single_module  -static  -sub_library  -sub_umbrella
554           -twolevel_namespace  -umbrella  -undefined -unexported_symbols_list
555           -weak_reference_mismatches -whatsloaded  -F  -gused  -gfull
556           -mmacosx-version-min=version -mkernel  -mone-byte-bool
557
558           DEC Alpha Options -mno-fp-regs  -msoft-float -mieee
559           -mieee-with-inexact  -mieee-conformant -mfp-trap-mode=mode
560           -mfp-rounding-mode=mode -mtrap-precision=mode  -mbuild-constants
561           -mcpu=cpu-type  -mtune=cpu-type -mbwx  -mmax  -mfix  -mcix
562           -mfloat-vax  -mfloat-ieee -mexplicit-relocs  -msmall-data
563           -mlarge-data -msmall-text  -mlarge-text -mmemory-latency=time
564
565           eBPF Options -mbig-endian -mlittle-endian -mkernel=version
566           -mframe-limit=bytes
567
568           FR30 Options -msmall-model  -mno-lsim
569
570           FT32 Options -msim  -mlra  -mnodiv  -mft32b  -mcompress  -mnopm
571
572           FRV Options -mgpr-32  -mgpr-64  -mfpr-32  -mfpr-64 -mhard-float
573           -msoft-float -malloc-cc  -mfixed-cc  -mdword  -mno-dword -mdouble
574           -mno-double -mmedia  -mno-media  -mmuladd  -mno-muladd -mfdpic
575           -minline-plt  -mgprel-ro  -multilib-library-pic -mlinked-fp
576           -mlong-calls  -malign-labels -mlibrary-pic  -macc-4  -macc-8 -mpack
577           -mno-pack  -mno-eflags  -mcond-move  -mno-cond-move
578           -moptimize-membar  -mno-optimize-membar -mscc  -mno-scc
579           -mcond-exec  -mno-cond-exec -mvliw-branch  -mno-vliw-branch
580           -mmulti-cond-exec  -mno-multi-cond-exec  -mnested-cond-exec
581           -mno-nested-cond-exec  -mtomcat-stats -mTLS  -mtls -mcpu=cpu
582
583           GNU/Linux Options -mglibc  -muclibc  -mmusl  -mbionic  -mandroid
584           -tno-android-cc  -tno-android-ld
585
586           H8/300 Options -mrelax  -mh  -ms  -mn  -mexr  -mno-exr  -mint32
587           -malign-300
588
589           HPPA Options -march=architecture-type -mcaller-copies
590           -mdisable-fpregs  -mdisable-indexing -mfast-indirect-calls  -mgas
591           -mgnu-ld   -mhp-ld -mfixed-range=register-range -mjump-in-delay
592           -mlinker-opt  -mlong-calls -mlong-load-store  -mno-disable-fpregs
593           -mno-disable-indexing  -mno-fast-indirect-calls  -mno-gas
594           -mno-jump-in-delay  -mno-long-load-store -mno-portable-runtime
595           -mno-soft-float -mno-space-regs  -msoft-float  -mpa-risc-1-0
596           -mpa-risc-1-1  -mpa-risc-2-0  -mportable-runtime -mschedule=cpu-
597           type  -mspace-regs  -msio  -mwsio -munix=unix-std  -nolibdld
598           -static  -threads
599
600           IA-64 Options -mbig-endian  -mlittle-endian  -mgnu-as  -mgnu-ld
601           -mno-pic -mvolatile-asm-stop  -mregister-names  -msdata  -mno-sdata
602           -mconstant-gp  -mauto-pic  -mfused-madd
603           -minline-float-divide-min-latency
604           -minline-float-divide-max-throughput -mno-inline-float-divide
605           -minline-int-divide-min-latency -minline-int-divide-max-throughput
606           -mno-inline-int-divide -minline-sqrt-min-latency
607           -minline-sqrt-max-throughput -mno-inline-sqrt -mdwarf2-asm
608           -mearly-stop-bits -mfixed-range=register-range  -mtls-size=tls-size
609           -mtune=cpu-type  -milp32  -mlp64 -msched-br-data-spec
610           -msched-ar-data-spec  -msched-control-spec -msched-br-in-data-spec
611           -msched-ar-in-data-spec  -msched-in-control-spec -msched-spec-ldc
612           -msched-spec-control-ldc -msched-prefer-non-data-spec-insns
613           -msched-prefer-non-control-spec-insns
614           -msched-stop-bits-after-every-cycle
615           -msched-count-spec-in-critical-path
616           -msel-sched-dont-check-control-spec  -msched-fp-mem-deps-zero-cost
617           -msched-max-memory-insns-hard-limit  -msched-max-memory-insns=max-
618           insns
619
620           LM32 Options -mbarrel-shift-enabled  -mdivide-enabled
621           -mmultiply-enabled -msign-extend-enabled  -muser-enabled
622
623           M32R/D Options -m32r2  -m32rx  -m32r -mdebug -malign-loops
624           -mno-align-loops -missue-rate=number -mbranch-cost=number
625           -mmodel=code-size-model-type -msdata=sdata-type -mno-flush-func
626           -mflush-func=name -mno-flush-trap  -mflush-trap=number -G num
627
628           M32C Options -mcpu=cpu  -msim  -memregs=number
629
630           M680x0 Options -march=arch  -mcpu=cpu  -mtune=tune -m68000  -m68020
631           -m68020-40  -m68020-60  -m68030  -m68040 -m68060  -mcpu32  -m5200
632           -m5206e  -m528x  -m5307  -m5407 -mcfv4e  -mbitfield  -mno-bitfield
633           -mc68000  -mc68020 -mnobitfield  -mrtd  -mno-rtd  -mdiv  -mno-div
634           -mshort -mno-short  -mhard-float  -m68881  -msoft-float  -mpcrel
635           -malign-int  -mstrict-align  -msep-data  -mno-sep-data
636           -mshared-library-id=n  -mid-shared-library  -mno-id-shared-library
637           -mxgot  -mno-xgot  -mlong-jump-table-offsets
638
639           MCore Options -mhardlit  -mno-hardlit  -mdiv  -mno-div
640           -mrelax-immediates -mno-relax-immediates  -mwide-bitfields
641           -mno-wide-bitfields -m4byte-functions  -mno-4byte-functions
642           -mcallgraph-data -mno-callgraph-data  -mslow-bytes  -mno-slow-bytes
643           -mno-lsim -mlittle-endian  -mbig-endian  -m210  -m340
644           -mstack-increment
645
646           MeP Options -mabsdiff  -mall-opts  -maverage  -mbased=n  -mbitops
647           -mc=n  -mclip  -mconfig=name  -mcop  -mcop32  -mcop64  -mivc2 -mdc
648           -mdiv  -meb  -mel  -mio-volatile  -ml  -mleadz  -mm  -mminmax
649           -mmult  -mno-opts  -mrepeat  -ms  -msatur  -msdram  -msim
650           -msimnovec  -mtf -mtiny=n
651
652           MicroBlaze Options -msoft-float  -mhard-float  -msmall-divides
653           -mcpu=cpu -mmemcpy  -mxl-soft-mul  -mxl-soft-div  -mxl-barrel-shift
654           -mxl-pattern-compare  -mxl-stack-check  -mxl-gp-opt  -mno-clearbss
655           -mxl-multiply-high  -mxl-float-convert  -mxl-float-sqrt
656           -mbig-endian  -mlittle-endian  -mxl-reorder  -mxl-mode-app-model
657           -mpic-data-is-text-relative
658
659           MIPS Options -EL  -EB  -march=arch  -mtune=arch -mips1  -mips2
660           -mips3  -mips4  -mips32  -mips32r2  -mips32r3  -mips32r5 -mips32r6
661           -mips64  -mips64r2  -mips64r3  -mips64r5  -mips64r6 -mips16
662           -mno-mips16  -mflip-mips16 -minterlink-compressed
663           -mno-interlink-compressed -minterlink-mips16  -mno-interlink-mips16
664           -mabi=abi  -mabicalls  -mno-abicalls -mshared  -mno-shared  -mplt
665           -mno-plt  -mxgot  -mno-xgot -mgp32  -mgp64  -mfp32  -mfpxx  -mfp64
666           -mhard-float  -msoft-float -mno-float  -msingle-float
667           -mdouble-float -modd-spreg  -mno-odd-spreg -mabs=mode
668           -mnan=encoding -mdsp  -mno-dsp  -mdspr2  -mno-dspr2 -mmcu
669           -mmno-mcu -meva  -mno-eva -mvirt  -mno-virt -mxpa  -mno-xpa -mcrc
670           -mno-crc -mginv  -mno-ginv -mmicromips  -mno-micromips -mmsa
671           -mno-msa -mloongson-mmi  -mno-loongson-mmi -mloongson-ext
672           -mno-loongson-ext -mloongson-ext2  -mno-loongson-ext2 -mfpu=fpu-
673           type -msmartmips  -mno-smartmips -mpaired-single
674           -mno-paired-single  -mdmx  -mno-mdmx -mips3d  -mno-mips3d  -mmt
675           -mno-mt  -mllsc  -mno-llsc -mlong64  -mlong32  -msym32  -mno-sym32
676           -Gnum  -mlocal-sdata  -mno-local-sdata -mextern-sdata
677           -mno-extern-sdata  -mgpopt  -mno-gopt -membedded-data
678           -mno-embedded-data -muninit-const-in-rodata
679           -mno-uninit-const-in-rodata -mcode-readable=setting
680           -msplit-addresses  -mno-split-addresses -mexplicit-relocs
681           -mno-explicit-relocs -mcheck-zero-division
682           -mno-check-zero-division -mdivide-traps  -mdivide-breaks
683           -mload-store-pairs  -mno-load-store-pairs -mmemcpy  -mno-memcpy
684           -mlong-calls  -mno-long-calls -mmad  -mno-mad  -mimadd  -mno-imadd
685           -mfused-madd  -mno-fused-madd  -nocpp -mfix-24k  -mno-fix-24k
686           -mfix-r4000  -mno-fix-r4000  -mfix-r4400  -mno-fix-r4400
687           -mfix-r5900  -mno-fix-r5900 -mfix-r10000  -mno-fix-r10000
688           -mfix-rm7000  -mno-fix-rm7000 -mfix-vr4120  -mno-fix-vr4120
689           -mfix-vr4130  -mno-fix-vr4130  -mfix-sb1  -mno-fix-sb1
690           -mflush-func=func  -mno-flush-func -mbranch-cost=num
691           -mbranch-likely  -mno-branch-likely -mcompact-branches=policy
692           -mfp-exceptions  -mno-fp-exceptions -mvr4130-align
693           -mno-vr4130-align  -msynci  -mno-synci -mlxc1-sxc1  -mno-lxc1-sxc1
694           -mmadd4  -mno-madd4 -mrelax-pic-calls  -mno-relax-pic-calls
695           -mmcount-ra-address -mframe-header-opt  -mno-frame-header-opt
696
697           MMIX Options -mlibfuncs  -mno-libfuncs  -mepsilon  -mno-epsilon
698           -mabi=gnu -mabi=mmixware  -mzero-extend  -mknuthdiv
699           -mtoplevel-symbols -melf  -mbranch-predict  -mno-branch-predict
700           -mbase-addresses -mno-base-addresses  -msingle-exit
701           -mno-single-exit
702
703           MN10300 Options -mmult-bug  -mno-mult-bug -mno-am33  -mam33
704           -mam33-2  -mam34 -mtune=cpu-type -mreturn-pointer-on-d0 -mno-crt0
705           -mrelax  -mliw  -msetlb
706
707           Moxie Options -meb  -mel  -mmul.x  -mno-crt0
708
709           MSP430 Options -msim  -masm-hex  -mmcu=  -mcpu=  -mlarge  -msmall
710           -mrelax -mwarn-mcu -mcode-region=  -mdata-region= -msilicon-errata=
711           -msilicon-errata-warn= -mhwmult=  -minrt  -mtiny-printf
712
713           NDS32 Options -mbig-endian  -mlittle-endian -mreduced-regs
714           -mfull-regs -mcmov  -mno-cmov -mext-perf  -mno-ext-perf -mext-perf2
715           -mno-ext-perf2 -mext-string  -mno-ext-string -mv3push  -mno-v3push
716           -m16bit  -mno-16bit -misr-vector-size=num -mcache-block-size=num
717           -march=arch -mcmodel=code-model -mctor-dtor  -mrelax
718
719           Nios II Options -G num  -mgpopt=option  -mgpopt  -mno-gpopt
720           -mgprel-sec=regexp  -mr0rel-sec=regexp -mel  -meb -mno-bypass-cache
721           -mbypass-cache -mno-cache-volatile  -mcache-volatile
722           -mno-fast-sw-div  -mfast-sw-div -mhw-mul  -mno-hw-mul  -mhw-mulx
723           -mno-hw-mulx  -mno-hw-div  -mhw-div -mcustom-insn=N
724           -mno-custom-insn -mcustom-fpu-cfg=name -mhal  -msmallc
725           -msys-crt0=name  -msys-lib=name -march=arch  -mbmx  -mno-bmx  -mcdx
726           -mno-cdx
727
728           Nvidia PTX Options -m32  -m64  -mmainkernel  -moptimize
729
730           OpenRISC Options -mboard=name  -mnewlib  -mhard-mul  -mhard-div
731           -msoft-mul  -msoft-div -msoft-float  -mhard-float  -mdouble-float
732           -munordered-float -mcmov  -mror  -mrori  -msext  -msfimm  -mshftimm
733
734           PDP-11 Options -mfpu  -msoft-float  -mac0  -mno-ac0  -m40  -m45
735           -m10 -mint32  -mno-int16  -mint16  -mno-int32 -msplit  -munix-asm
736           -mdec-asm  -mgnu-asm  -mlra
737
738           picoChip Options -mae=ae_type  -mvliw-lookahead=N
739           -msymbol-as-address  -mno-inefficient-warnings
740
741           PowerPC Options See RS/6000 and PowerPC Options.
742
743           PRU Options -mmcu=mcu  -minrt  -mno-relax  -mloop -mabi=variant
744
745           RISC-V Options -mbranch-cost=N-instruction -mplt  -mno-plt
746           -mabi=ABI-string -mfdiv  -mno-fdiv -mdiv  -mno-div -march=ISA-
747           string -mtune=processor-string -mpreferred-stack-boundary=num
748           -msmall-data-limit=N-bytes -msave-restore  -mno-save-restore
749           -mstrict-align  -mno-strict-align -mcmodel=medlow  -mcmodel=medany
750           -mexplicit-relocs  -mno-explicit-relocs -mrelax  -mno-relax
751           -mriscv-attribute  -mmo-riscv-attribute -malign-data=type
752
753           RL78 Options -msim  -mmul=none  -mmul=g13  -mmul=g14  -mallregs
754           -mcpu=g10  -mcpu=g13  -mcpu=g14  -mg10  -mg13  -mg14
755           -m64bit-doubles  -m32bit-doubles  -msave-mduc-in-interrupts
756
757           RS/6000 and PowerPC Options -mcpu=cpu-type -mtune=cpu-type
758           -mcmodel=code-model -mpowerpc64 -maltivec  -mno-altivec
759           -mpowerpc-gpopt  -mno-powerpc-gpopt -mpowerpc-gfxopt
760           -mno-powerpc-gfxopt -mmfcrf  -mno-mfcrf  -mpopcntb  -mno-popcntb
761           -mpopcntd  -mno-popcntd -mfprnd  -mno-fprnd -mcmpb  -mno-cmpb
762           -mhard-dfp  -mno-hard-dfp -mfull-toc   -mminimal-toc
763           -mno-fp-in-toc  -mno-sum-in-toc -m64  -m32  -mxl-compat
764           -mno-xl-compat  -mpe -malign-power  -malign-natural -msoft-float
765           -mhard-float  -mmultiple  -mno-multiple -mupdate  -mno-update
766           -mavoid-indexed-addresses  -mno-avoid-indexed-addresses
767           -mfused-madd  -mno-fused-madd  -mbit-align  -mno-bit-align
768           -mstrict-align  -mno-strict-align  -mrelocatable -mno-relocatable
769           -mrelocatable-lib  -mno-relocatable-lib -mtoc  -mno-toc  -mlittle
770           -mlittle-endian  -mbig  -mbig-endian -mdynamic-no-pic  -mswdiv
771           -msingle-pic-base -mprioritize-restricted-insns=priority
772           -msched-costly-dep=dependence_type -minsert-sched-nops=scheme
773           -mcall-aixdesc  -mcall-eabi  -mcall-freebsd -mcall-linux
774           -mcall-netbsd  -mcall-openbsd -mcall-sysv  -mcall-sysv-eabi
775           -mcall-sysv-noeabi -mtraceback=traceback_type -maix-struct-return
776           -msvr4-struct-return -mabi=abi-type  -msecure-plt  -mbss-plt
777           -mlongcall  -mno-longcall  -mpltseq  -mno-pltseq
778           -mblock-move-inline-limit=num -mblock-compare-inline-limit=num
779           -mblock-compare-inline-loop-limit=num
780           -mstring-compare-inline-limit=num -misel  -mno-isel -mvrsave
781           -mno-vrsave -mmulhw  -mno-mulhw -mdlmzb  -mno-dlmzb -mprototype
782           -mno-prototype -msim  -mmvme  -mads  -myellowknife  -memb  -msdata
783           -msdata=opt  -mreadonly-in-sdata  -mvxworks  -G num -mrecip
784           -mrecip=opt  -mno-recip  -mrecip-precision -mno-recip-precision
785           -mveclibabi=type  -mfriz  -mno-friz -mpointers-to-nested-functions
786           -mno-pointers-to-nested-functions -msave-toc-indirect
787           -mno-save-toc-indirect -mpower8-fusion  -mno-mpower8-fusion
788           -mpower8-vector  -mno-power8-vector -mcrypto  -mno-crypto  -mhtm
789           -mno-htm -mquad-memory  -mno-quad-memory -mquad-memory-atomic
790           -mno-quad-memory-atomic -mcompat-align-parm  -mno-compat-align-parm
791           -mfloat128  -mno-float128  -mfloat128-hardware
792           -mno-float128-hardware -mgnu-attribute  -mno-gnu-attribute
793           -mstack-protector-guard=guard -mstack-protector-guard-reg=reg
794           -mstack-protector-guard-offset=offset -mprefixed -mno-prefixed
795           -mpcrel -mno-pcrel -mmma -mno-mmma
796
797           RX Options -m64bit-doubles  -m32bit-doubles  -fpu  -nofpu -mcpu=
798           -mbig-endian-data  -mlittle-endian-data -msmall-data -msim
799           -mno-sim -mas100-syntax  -mno-as100-syntax -mrelax
800           -mmax-constant-size= -mint-register= -mpid -mallow-string-insns
801           -mno-allow-string-insns -mjsr -mno-warn-multiple-fast-interrupts
802           -msave-acc-in-interrupts
803
804           S/390 and zSeries Options -mtune=cpu-type  -march=cpu-type
805           -mhard-float  -msoft-float  -mhard-dfp  -mno-hard-dfp
806           -mlong-double-64  -mlong-double-128 -mbackchain  -mno-backchain
807           -mpacked-stack  -mno-packed-stack -msmall-exec  -mno-small-exec
808           -mmvcle  -mno-mvcle -m64  -m31  -mdebug  -mno-debug  -mesa  -mzarch
809           -mhtm  -mvx  -mzvector -mtpf-trace  -mno-tpf-trace
810           -mtpf-trace-skip  -mno-tpf-trace-skip -mfused-madd  -mno-fused-madd
811           -mwarn-framesize  -mwarn-dynamicstack  -mstack-size  -mstack-guard
812           -mhotpatch=halfwords,halfwords
813
814           Score Options -meb  -mel -mnhwloop -muls -mmac -mscore5  -mscore5u
815           -mscore7  -mscore7d
816
817           SH Options -m1  -m2  -m2e -m2a-nofpu  -m2a-single-only  -m2a-single
818           -m2a -m3  -m3e -m4-nofpu  -m4-single-only  -m4-single  -m4
819           -m4a-nofpu  -m4a-single-only  -m4a-single  -m4a  -m4al -mb  -ml
820           -mdalign  -mrelax -mbigtable  -mfmovd  -mrenesas  -mno-renesas
821           -mnomacsave -mieee  -mno-ieee  -mbitops  -misize
822           -minline-ic_invalidate  -mpadstruct -mprefergot  -musermode
823           -multcost=number  -mdiv=strategy -mdivsi3_libfunc=name
824           -mfixed-range=register-range -maccumulate-outgoing-args
825           -matomic-model=atomic-model -mbranch-cost=num  -mzdcbranch
826           -mno-zdcbranch -mcbranch-force-delay-slot -mfused-madd
827           -mno-fused-madd  -mfsca  -mno-fsca  -mfsrra  -mno-fsrra
828           -mpretend-cmove  -mtas
829
830           Solaris 2 Options -mclear-hwcap  -mno-clear-hwcap  -mimpure-text
831           -mno-impure-text -pthreads
832
833           SPARC Options -mcpu=cpu-type -mtune=cpu-type -mcmodel=code-model
834           -mmemory-model=mem-model -m32  -m64  -mapp-regs  -mno-app-regs
835           -mfaster-structs  -mno-faster-structs  -mflat  -mno-flat -mfpu
836           -mno-fpu  -mhard-float  -msoft-float -mhard-quad-float
837           -msoft-quad-float -mstack-bias  -mno-stack-bias -mstd-struct-return
838           -mno-std-struct-return -munaligned-doubles  -mno-unaligned-doubles
839           -muser-mode  -mno-user-mode -mv8plus  -mno-v8plus  -mvis  -mno-vis
840           -mvis2  -mno-vis2  -mvis3  -mno-vis3 -mvis4  -mno-vis4  -mvis4b
841           -mno-vis4b -mcbcond  -mno-cbcond  -mfmaf  -mno-fmaf  -mfsmuld
842           -mno-fsmuld -mpopc  -mno-popc  -msubxc  -mno-subxc -mfix-at697f
843           -mfix-ut699  -mfix-ut700  -mfix-gr712rc -mlra  -mno-lra
844
845           System V Options -Qy  -Qn  -YP,paths  -Ym,dir
846
847           TILE-Gx Options -mcpu=CPU  -m32  -m64  -mbig-endian
848           -mlittle-endian -mcmodel=code-model
849
850           TILEPro Options -mcpu=cpu  -m32
851
852           V850 Options -mlong-calls  -mno-long-calls  -mep  -mno-ep
853           -mprolog-function  -mno-prolog-function  -mspace -mtda=n  -msda=n
854           -mzda=n -mapp-regs  -mno-app-regs -mdisable-callt
855           -mno-disable-callt -mv850e2v3  -mv850e2  -mv850e1  -mv850es -mv850e
856           -mv850  -mv850e3v5 -mloop -mrelax -mlong-jumps -msoft-float
857           -mhard-float -mgcc-abi -mrh850-abi -mbig-switch
858
859           VAX Options -mg  -mgnu  -munix
860
861           Visium Options -mdebug  -msim  -mfpu  -mno-fpu  -mhard-float
862           -msoft-float -mcpu=cpu-type  -mtune=cpu-type  -msv-mode
863           -muser-mode
864
865           VMS Options -mvms-return-codes  -mdebug-main=prefix  -mmalloc64
866           -mpointer-size=size
867
868           VxWorks Options -mrtp  -non-static  -Bstatic  -Bdynamic -Xbind-lazy
869           -Xbind-now
870
871           x86 Options -mtune=cpu-type  -march=cpu-type -mtune-ctrl=feature-
872           list  -mdump-tune-features  -mno-default -mfpmath=unit
873           -masm=dialect  -mno-fancy-math-387 -mno-fp-ret-in-387  -m80387
874           -mhard-float  -msoft-float -mno-wide-multiply  -mrtd
875           -malign-double -mpreferred-stack-boundary=num
876           -mincoming-stack-boundary=num -mcld  -mcx16  -msahf  -mmovbe
877           -mcrc32 -mrecip  -mrecip=opt -mvzeroupper  -mprefer-avx128
878           -mprefer-vector-width=opt -mmmx  -msse  -msse2  -msse3  -mssse3
879           -msse4.1  -msse4.2  -msse4  -mavx -mavx2  -mavx512f  -mavx512pf
880           -mavx512er  -mavx512cd  -mavx512vl -mavx512bw  -mavx512dq
881           -mavx512ifma  -mavx512vbmi  -msha  -maes -mpclmul  -mfsgsbase
882           -mrdrnd  -mf16c  -mfma  -mpconfig  -mwbnoinvd -mptwrite
883           -mprefetchwt1  -mclflushopt  -mclwb  -mxsavec  -mxsaves -msse4a
884           -m3dnow  -m3dnowa  -mpopcnt  -mabm  -mbmi  -mtbm  -mfma4  -mxop
885           -madx  -mlzcnt  -mbmi2  -mfxsr  -mxsave  -mxsaveopt  -mrtm  -mhle
886           -mlwp -mmwaitx  -mclzero  -mpku  -mthreads  -mgfni  -mvaes
887           -mwaitpkg -mshstk -mmanual-endbr -mforce-indirect-call
888           -mavx512vbmi2 -mavx512bf16 -menqcmd -mvpclmulqdq  -mavx512bitalg
889           -mmovdiri  -mmovdir64b  -mavx512vpopcntdq -mavx5124fmaps
890           -mavx512vnni  -mavx5124vnniw  -mprfchw  -mrdpid -mrdseed  -msgx
891           -mavx512vp2intersect -mcldemote  -mms-bitfields
892           -mno-align-stringops  -minline-all-stringops
893           -minline-stringops-dynamically  -mstringop-strategy=alg
894           -mmemcpy-strategy=strategy  -mmemset-strategy=strategy -mpush-args
895           -maccumulate-outgoing-args  -m128bit-long-double
896           -m96bit-long-double  -mlong-double-64  -mlong-double-80
897           -mlong-double-128 -mregparm=num  -msseregparm -mveclibabi=type
898           -mvect8-ret-in-mem -mpc32  -mpc64  -mpc80  -mstackrealign
899           -momit-leaf-frame-pointer  -mno-red-zone  -mno-tls-direct-seg-refs
900           -mcmodel=code-model  -mabi=name  -maddress-mode=mode -m32  -m64
901           -mx32  -m16  -miamcu  -mlarge-data-threshold=num -msse2avx
902           -mfentry  -mrecord-mcount  -mnop-mcount  -m8bit-idiv
903           -minstrument-return=type -mfentry-name=name -mfentry-section=name
904           -mavx256-split-unaligned-load  -mavx256-split-unaligned-store
905           -malign-data=type  -mstack-protector-guard=guard
906           -mstack-protector-guard-reg=reg
907           -mstack-protector-guard-offset=offset
908           -mstack-protector-guard-symbol=symbol -mgeneral-regs-only
909           -mcall-ms2sysv-xlogues -mindirect-branch=choice
910           -mfunction-return=choice -mindirect-branch-register
911
912           x86 Windows Options -mconsole  -mcygwin  -mno-cygwin  -mdll
913           -mnop-fun-dllimport  -mthread -municode  -mwin32  -mwindows
914           -fno-set-stack-executable
915
916           Xstormy16 Options -msim
917
918           Xtensa Options -mconst16  -mno-const16 -mfused-madd
919           -mno-fused-madd -mforce-no-pic -mserialize-volatile
920           -mno-serialize-volatile -mtext-section-literals
921           -mno-text-section-literals -mauto-litpools  -mno-auto-litpools
922           -mtarget-align  -mno-target-align -mlongcalls  -mno-longcalls
923
924           zSeries Options See S/390 and zSeries Options.
925
926   Options Controlling the Kind of Output
927       Compilation can involve up to four stages: preprocessing, compilation
928       proper, assembly and linking, always in that order.  GCC is capable of
929       preprocessing and compiling several files either into several assembler
930       input files, or into one assembler input file; then each assembler
931       input file produces an object file, and linking combines all the object
932       files (those newly compiled, and those specified as input) into an
933       executable file.
934
935       For any given input file, the file name suffix determines what kind of
936       compilation is done:
937
938       file.c
939           C source code that must be preprocessed.
940
941       file.i
942           C source code that should not be preprocessed.
943
944       file.ii
945           C++ source code that should not be preprocessed.
946
947       file.m
948           Objective-C source code.  Note that you must link with the libobjc
949           library to make an Objective-C program work.
950
951       file.mi
952           Objective-C source code that should not be preprocessed.
953
954       file.mm
955       file.M
956           Objective-C++ source code.  Note that you must link with the
957           libobjc library to make an Objective-C++ program work.  Note that
958           .M refers to a literal capital M.
959
960       file.mii
961           Objective-C++ source code that should not be preprocessed.
962
963       file.h
964           C, C++, Objective-C or Objective-C++ header file to be turned into
965           a precompiled header (default), or C, C++ header file to be turned
966           into an Ada spec (via the -fdump-ada-spec switch).
967
968       file.cc
969       file.cp
970       file.cxx
971       file.cpp
972       file.CPP
973       file.c++
974       file.C
975           C++ source code that must be preprocessed.  Note that in .cxx, the
976           last two letters must both be literally x.  Likewise, .C refers to
977           a literal capital C.
978
979       file.mm
980       file.M
981           Objective-C++ source code that must be preprocessed.
982
983       file.mii
984           Objective-C++ source code that should not be preprocessed.
985
986       file.hh
987       file.H
988       file.hp
989       file.hxx
990       file.hpp
991       file.HPP
992       file.h++
993       file.tcc
994           C++ header file to be turned into a precompiled header or Ada spec.
995
996       file.f
997       file.for
998       file.ftn
999           Fixed form Fortran source code that should not be preprocessed.
1000
1001       file.F
1002       file.FOR
1003       file.fpp
1004       file.FPP
1005       file.FTN
1006           Fixed form Fortran source code that must be preprocessed (with the
1007           traditional preprocessor).
1008
1009       file.f90
1010       file.f95
1011       file.f03
1012       file.f08
1013           Free form Fortran source code that should not be preprocessed.
1014
1015       file.F90
1016       file.F95
1017       file.F03
1018       file.F08
1019           Free form Fortran source code that must be preprocessed (with the
1020           traditional preprocessor).
1021
1022       file.go
1023           Go source code.
1024
1025       file.brig
1026           BRIG files (binary representation of HSAIL).
1027
1028       file.d
1029           D source code.
1030
1031       file.di
1032           D interface file.
1033
1034       file.dd
1035           D documentation code (Ddoc).
1036
1037       file.ads
1038           Ada source code file that contains a library unit declaration (a
1039           declaration of a package, subprogram, or generic, or a generic
1040           instantiation), or a library unit renaming declaration (a package,
1041           generic, or subprogram renaming declaration).  Such files are also
1042           called specs.
1043
1044       file.adb
1045           Ada source code file containing a library unit body (a subprogram
1046           or package body).  Such files are also called bodies.
1047
1048       file.s
1049           Assembler code.
1050
1051       file.S
1052       file.sx
1053           Assembler code that must be preprocessed.
1054
1055       other
1056           An object file to be fed straight into linking.  Any file name with
1057           no recognized suffix is treated this way.
1058
1059       You can specify the input language explicitly with the -x option:
1060
1061       -x language
1062           Specify explicitly the language for the following input files
1063           (rather than letting the compiler choose a default based on the
1064           file name suffix).  This option applies to all following input
1065           files until the next -x option.  Possible values for language are:
1066
1067                   c  c-header  cpp-output
1068                   c++  c++-header  c++-cpp-output
1069                   objective-c  objective-c-header  objective-c-cpp-output
1070                   objective-c++ objective-c++-header objective-c++-cpp-output
1071                   assembler  assembler-with-cpp
1072                   ada
1073                   d
1074                   f77  f77-cpp-input f95  f95-cpp-input
1075                   go
1076                   brig
1077
1078       -x none
1079           Turn off any specification of a language, so that subsequent files
1080           are handled according to their file name suffixes (as they are if
1081           -x has not been used at all).
1082
1083       If you only want some of the stages of compilation, you can use -x (or
1084       filename suffixes) to tell gcc where to start, and one of the options
1085       -c, -S, or -E to say where gcc is to stop.  Note that some combinations
1086       (for example, -x cpp-output -E) instruct gcc to do nothing at all.
1087
1088       -c  Compile or assemble the source files, but do not link.  The linking
1089           stage simply is not done.  The ultimate output is in the form of an
1090           object file for each source file.
1091
1092           By default, the object file name for a source file is made by
1093           replacing the suffix .c, .i, .s, etc., with .o.
1094
1095           Unrecognized input files, not requiring compilation or assembly,
1096           are ignored.
1097
1098       -S  Stop after the stage of compilation proper; do not assemble.  The
1099           output is in the form of an assembler code file for each non-
1100           assembler input file specified.
1101
1102           By default, the assembler file name for a source file is made by
1103           replacing the suffix .c, .i, etc., with .s.
1104
1105           Input files that don't require compilation are ignored.
1106
1107       -E  Stop after the preprocessing stage; do not run the compiler proper.
1108           The output is in the form of preprocessed source code, which is
1109           sent to the standard output.
1110
1111           Input files that don't require preprocessing are ignored.
1112
1113       -o file
1114           Place output in file file.  This applies to whatever sort of output
1115           is being produced, whether it be an executable file, an object
1116           file, an assembler file or preprocessed C code.
1117
1118           If -o is not specified, the default is to put an executable file in
1119           a.out, the object file for source.suffix in source.o, its assembler
1120           file in source.s, a precompiled header file in source.suffix.gch,
1121           and all preprocessed C source on standard output.
1122
1123       -v  Print (on standard error output) the commands executed to run the
1124           stages of compilation.  Also print the version number of the
1125           compiler driver program and of the preprocessor and the compiler
1126           proper.
1127
1128       -###
1129           Like -v except the commands are not executed and arguments are
1130           quoted unless they contain only alphanumeric characters or "./-_".
1131           This is useful for shell scripts to capture the driver-generated
1132           command lines.
1133
1134       --help
1135           Print (on the standard output) a description of the command-line
1136           options understood by gcc.  If the -v option is also specified then
1137           --help is also passed on to the various processes invoked by gcc,
1138           so that they can display the command-line options they accept.  If
1139           the -Wextra option has also been specified (prior to the --help
1140           option), then command-line options that have no documentation
1141           associated with them are also displayed.
1142
1143       --target-help
1144           Print (on the standard output) a description of target-specific
1145           command-line options for each tool.  For some targets extra target-
1146           specific information may also be printed.
1147
1148       --help={class|[^]qualifier}[,...]
1149           Print (on the standard output) a description of the command-line
1150           options understood by the compiler that fit into all specified
1151           classes and qualifiers.  These are the supported classes:
1152
1153           optimizers
1154               Display all of the optimization options supported by the
1155               compiler.
1156
1157           warnings
1158               Display all of the options controlling warning messages
1159               produced by the compiler.
1160
1161           target
1162               Display target-specific options.  Unlike the --target-help
1163               option however, target-specific options of the linker and
1164               assembler are not displayed.  This is because those tools do
1165               not currently support the extended --help= syntax.
1166
1167           params
1168               Display the values recognized by the --param option.
1169
1170           language
1171               Display the options supported for language, where language is
1172               the name of one of the languages supported in this version of
1173               GCC.  If an option is supported by all languages, one needs to
1174               select common class.
1175
1176           common
1177               Display the options that are common to all languages.
1178
1179           These are the supported qualifiers:
1180
1181           undocumented
1182               Display only those options that are undocumented.
1183
1184           joined
1185               Display options taking an argument that appears after an equal
1186               sign in the same continuous piece of text, such as:
1187               --help=target.
1188
1189           separate
1190               Display options taking an argument that appears as a separate
1191               word following the original option, such as: -o output-file.
1192
1193           Thus for example to display all the undocumented target-specific
1194           switches supported by the compiler, use:
1195
1196                   --help=target,undocumented
1197
1198           The sense of a qualifier can be inverted by prefixing it with the ^
1199           character, so for example to display all binary warning options
1200           (i.e., ones that are either on or off and that do not take an
1201           argument) that have a description, use:
1202
1203                   --help=warnings,^joined,^undocumented
1204
1205           The argument to --help= should not consist solely of inverted
1206           qualifiers.
1207
1208           Combining several classes is possible, although this usually
1209           restricts the output so much that there is nothing to display.  One
1210           case where it does work, however, is when one of the classes is
1211           target.  For example, to display all the target-specific
1212           optimization options, use:
1213
1214                   --help=target,optimizers
1215
1216           The --help= option can be repeated on the command line.  Each
1217           successive use displays its requested class of options, skipping
1218           those that have already been displayed.  If --help is also
1219           specified anywhere on the command line then this takes precedence
1220           over any --help= option.
1221
1222           If the -Q option appears on the command line before the --help=
1223           option, then the descriptive text displayed by --help= is changed.
1224           Instead of describing the displayed options, an indication is given
1225           as to whether the option is enabled, disabled or set to a specific
1226           value (assuming that the compiler knows this at the point where the
1227           --help= option is used).
1228
1229           Here is a truncated example from the ARM port of gcc:
1230
1231                     % gcc -Q -mabi=2 --help=target -c
1232                     The following options are target specific:
1233                     -mabi=                                2
1234                     -mabort-on-noreturn                   [disabled]
1235                     -mapcs                                [disabled]
1236
1237           The output is sensitive to the effects of previous command-line
1238           options, so for example it is possible to find out which
1239           optimizations are enabled at -O2 by using:
1240
1241                   -Q -O2 --help=optimizers
1242
1243           Alternatively you can discover which binary optimizations are
1244           enabled by -O3 by using:
1245
1246                   gcc -c -Q -O3 --help=optimizers > /tmp/O3-opts
1247                   gcc -c -Q -O2 --help=optimizers > /tmp/O2-opts
1248                   diff /tmp/O2-opts /tmp/O3-opts | grep enabled
1249
1250       --version
1251           Display the version number and copyrights of the invoked GCC.
1252
1253       -pass-exit-codes
1254           Normally the gcc program exits with the code of 1 if any phase of
1255           the compiler returns a non-success return code.  If you specify
1256           -pass-exit-codes, the gcc program instead returns with the
1257           numerically highest error produced by any phase returning an error
1258           indication.  The C, C++, and Fortran front ends return 4 if an
1259           internal compiler error is encountered.
1260
1261       -pipe
1262           Use pipes rather than temporary files for communication between the
1263           various stages of compilation.  This fails to work on some systems
1264           where the assembler is unable to read from a pipe; but the GNU
1265           assembler has no trouble.
1266
1267       -specs=file
1268           Process file after the compiler reads in the standard specs file,
1269           in order to override the defaults which the gcc driver program uses
1270           when determining what switches to pass to cc1, cc1plus, as, ld,
1271           etc.  More than one -specs=file can be specified on the command
1272           line, and they are processed in order, from left to right.
1273
1274       -wrapper
1275           Invoke all subcommands under a wrapper program.  The name of the
1276           wrapper program and its parameters are passed as a comma separated
1277           list.
1278
1279                   gcc -c t.c -wrapper gdb,--args
1280
1281           This invokes all subprograms of gcc under gdb --args, thus the
1282           invocation of cc1 is gdb --args cc1 ....
1283
1284       -ffile-prefix-map=old=new
1285           When compiling files residing in directory old, record any
1286           references to them in the result of the compilation as if the files
1287           resided in directory new instead.  Specifying this option is
1288           equivalent to specifying all the individual -f*-prefix-map options.
1289           This can be used to make reproducible builds that are location
1290           independent.  See also -fmacro-prefix-map and -fdebug-prefix-map.
1291
1292       -fplugin=name.so
1293           Load the plugin code in file name.so, assumed to be a shared object
1294           to be dlopen'd by the compiler.  The base name of the shared object
1295           file is used to identify the plugin for the purposes of argument
1296           parsing (See -fplugin-arg-name-key=value below).  Each plugin
1297           should define the callback functions specified in the Plugins API.
1298
1299       -fplugin-arg-name-key=value
1300           Define an argument called key with a value of value for the plugin
1301           called name.
1302
1303       -fdump-ada-spec[-slim]
1304           For C and C++ source and include files, generate corresponding Ada
1305           specs.
1306
1307       -fada-spec-parent=unit
1308           In conjunction with -fdump-ada-spec[-slim] above, generate Ada
1309           specs as child units of parent unit.
1310
1311       -fdump-go-spec=file
1312           For input files in any language, generate corresponding Go
1313           declarations in file.  This generates Go "const", "type", "var",
1314           and "func" declarations which may be a useful way to start writing
1315           a Go interface to code written in some other language.
1316
1317       @file
1318           Read command-line options from file.  The options read are inserted
1319           in place of the original @file option.  If file does not exist, or
1320           cannot be read, then the option will be treated literally, and not
1321           removed.
1322
1323           Options in file are separated by whitespace.  A whitespace
1324           character may be included in an option by surrounding the entire
1325           option in either single or double quotes.  Any character (including
1326           a backslash) may be included by prefixing the character to be
1327           included with a backslash.  The file may itself contain additional
1328           @file options; any such options will be processed recursively.
1329
1330   Compiling C++ Programs
1331       C++ source files conventionally use one of the suffixes .C, .cc, .cpp,
1332       .CPP, .c++, .cp, or .cxx; C++ header files often use .hh, .hpp, .H, or
1333       (for shared template code) .tcc; and preprocessed C++ files use the
1334       suffix .ii.  GCC recognizes files with these names and compiles them as
1335       C++ programs even if you call the compiler the same way as for
1336       compiling C programs (usually with the name gcc).
1337
1338       However, the use of gcc does not add the C++ library.  g++ is a program
1339       that calls GCC and automatically specifies linking against the C++
1340       library.  It treats .c, .h and .i files as C++ source files instead of
1341       C source files unless -x is used.  This program is also useful when
1342       precompiling a C header file with a .h extension for use in C++
1343       compilations.  On many systems, g++ is also installed with the name
1344       c++.
1345
1346       When you compile C++ programs, you may specify many of the same
1347       command-line options that you use for compiling programs in any
1348       language; or command-line options meaningful for C and related
1349       languages; or options that are meaningful only for C++ programs.
1350
1351   Options Controlling C Dialect
1352       The following options control the dialect of C (or languages derived
1353       from C, such as C++, Objective-C and Objective-C++) that the compiler
1354       accepts:
1355
1356       -ansi
1357           In C mode, this is equivalent to -std=c90. In C++ mode, it is
1358           equivalent to -std=c++98.
1359
1360           This turns off certain features of GCC that are incompatible with
1361           ISO C90 (when compiling C code), or of standard C++ (when compiling
1362           C++ code), such as the "asm" and "typeof" keywords, and predefined
1363           macros such as "unix" and "vax" that identify the type of system
1364           you are using.  It also enables the undesirable and rarely used ISO
1365           trigraph feature.  For the C compiler, it disables recognition of
1366           C++ style // comments as well as the "inline" keyword.
1367
1368           The alternate keywords "__asm__", "__extension__", "__inline__" and
1369           "__typeof__" continue to work despite -ansi.  You would not want to
1370           use them in an ISO C program, of course, but it is useful to put
1371           them in header files that might be included in compilations done
1372           with -ansi.  Alternate predefined macros such as "__unix__" and
1373           "__vax__" are also available, with or without -ansi.
1374
1375           The -ansi option does not cause non-ISO programs to be rejected
1376           gratuitously.  For that, -Wpedantic is required in addition to
1377           -ansi.
1378
1379           The macro "__STRICT_ANSI__" is predefined when the -ansi option is
1380           used.  Some header files may notice this macro and refrain from
1381           declaring certain functions or defining certain macros that the ISO
1382           standard doesn't call for; this is to avoid interfering with any
1383           programs that might use these names for other things.
1384
1385           Functions that are normally built in but do not have semantics
1386           defined by ISO C (such as "alloca" and "ffs") are not built-in
1387           functions when -ansi is used.
1388
1389       -std=
1390           Determine the language standard.   This option is currently only
1391           supported when compiling C or C++.
1392
1393           The compiler can accept several base standards, such as c90 or
1394           c++98, and GNU dialects of those standards, such as gnu90 or
1395           gnu++98.  When a base standard is specified, the compiler accepts
1396           all programs following that standard plus those using GNU
1397           extensions that do not contradict it.  For example, -std=c90 turns
1398           off certain features of GCC that are incompatible with ISO C90,
1399           such as the "asm" and "typeof" keywords, but not other GNU
1400           extensions that do not have a meaning in ISO C90, such as omitting
1401           the middle term of a "?:" expression. On the other hand, when a GNU
1402           dialect of a standard is specified, all features supported by the
1403           compiler are enabled, even when those features change the meaning
1404           of the base standard.  As a result, some strict-conforming programs
1405           may be rejected.  The particular standard is used by -Wpedantic to
1406           identify which features are GNU extensions given that version of
1407           the standard. For example -std=gnu90 -Wpedantic warns about C++
1408           style // comments, while -std=gnu99 -Wpedantic does not.
1409
1410           A value for this option must be provided; possible values are
1411
1412           c90
1413           c89
1414           iso9899:1990
1415               Support all ISO C90 programs (certain GNU extensions that
1416               conflict with ISO C90 are disabled). Same as -ansi for C code.
1417
1418           iso9899:199409
1419               ISO C90 as modified in amendment 1.
1420
1421           c99
1422           c9x
1423           iso9899:1999
1424           iso9899:199x
1425               ISO C99.  This standard is substantially completely supported,
1426               modulo bugs and floating-point issues (mainly but not entirely
1427               relating to optional C99 features from Annexes F and G).  See
1428               <http://gcc.gnu.org/c99status.html> for more information.  The
1429               names c9x and iso9899:199x are deprecated.
1430
1431           c11
1432           c1x
1433           iso9899:2011
1434               ISO C11, the 2011 revision of the ISO C standard.  This
1435               standard is substantially completely supported, modulo bugs,
1436               floating-point issues (mainly but not entirely relating to
1437               optional C11 features from Annexes F and G) and the optional
1438               Annexes K (Bounds-checking interfaces) and L (Analyzability).
1439               The name c1x is deprecated.
1440
1441           c17
1442           c18
1443           iso9899:2017
1444           iso9899:2018
1445               ISO C17, the 2017 revision of the ISO C standard (published in
1446               2018).  This standard is same as C11 except for corrections of
1447               defects (all of which are also applied with -std=c11) and a new
1448               value of "__STDC_VERSION__", and so is supported to the same
1449               extent as C11.
1450
1451           c2x The next version of the ISO C standard, still under
1452               development.  The support for this version is experimental and
1453               incomplete.
1454
1455           gnu90
1456           gnu89
1457               GNU dialect of ISO C90 (including some C99 features).
1458
1459           gnu99
1460           gnu9x
1461               GNU dialect of ISO C99.  The name gnu9x is deprecated.
1462
1463           gnu11
1464           gnu1x
1465               GNU dialect of ISO C11.  The name gnu1x is deprecated.
1466
1467           gnu17
1468           gnu18
1469               GNU dialect of ISO C17.  This is the default for C code.
1470
1471           gnu2x
1472               The next version of the ISO C standard, still under
1473               development, plus GNU extensions.  The support for this version
1474               is experimental and incomplete.
1475
1476           c++98
1477           c++03
1478               The 1998 ISO C++ standard plus the 2003 technical corrigendum
1479               and some additional defect reports. Same as -ansi for C++ code.
1480
1481           gnu++98
1482           gnu++03
1483               GNU dialect of -std=c++98.
1484
1485           c++11
1486           c++0x
1487               The 2011 ISO C++ standard plus amendments.  The name c++0x is
1488               deprecated.
1489
1490           gnu++11
1491           gnu++0x
1492               GNU dialect of -std=c++11.  The name gnu++0x is deprecated.
1493
1494           c++14
1495           c++1y
1496               The 2014 ISO C++ standard plus amendments.  The name c++1y is
1497               deprecated.
1498
1499           gnu++14
1500           gnu++1y
1501               GNU dialect of -std=c++14.  This is the default for C++ code.
1502               The name gnu++1y is deprecated.
1503
1504           c++17
1505           c++1z
1506               The 2017 ISO C++ standard plus amendments.  The name c++1z is
1507               deprecated.
1508
1509           gnu++17
1510           gnu++1z
1511               GNU dialect of -std=c++17.  The name gnu++1z is deprecated.
1512
1513           c++20
1514           c++2a
1515               The next revision of the ISO C++ standard, planned for 2020.
1516               Support is highly experimental, and will almost certainly
1517               change in incompatible ways in future releases.
1518
1519           gnu++20
1520           gnu++2a
1521               GNU dialect of -std=c++20.  Support is highly experimental, and
1522               will almost certainly change in incompatible ways in future
1523               releases.
1524
1525       -fgnu89-inline
1526           The option -fgnu89-inline tells GCC to use the traditional GNU
1527           semantics for "inline" functions when in C99 mode.
1528
1529           Using this option is roughly equivalent to adding the "gnu_inline"
1530           function attribute to all inline functions.
1531
1532           The option -fno-gnu89-inline explicitly tells GCC to use the C99
1533           semantics for "inline" when in C99 or gnu99 mode (i.e., it
1534           specifies the default behavior).  This option is not supported in
1535           -std=c90 or -std=gnu90 mode.
1536
1537           The preprocessor macros "__GNUC_GNU_INLINE__" and
1538           "__GNUC_STDC_INLINE__" may be used to check which semantics are in
1539           effect for "inline" functions.
1540
1541       -fpermitted-flt-eval-methods=style
1542           ISO/IEC TS 18661-3 defines new permissible values for
1543           "FLT_EVAL_METHOD" that indicate that operations and constants with
1544           a semantic type that is an interchange or extended format should be
1545           evaluated to the precision and range of that type.  These new
1546           values are a superset of those permitted under C99/C11, which does
1547           not specify the meaning of other positive values of
1548           "FLT_EVAL_METHOD".  As such, code conforming to C11 may not have
1549           been written expecting the possibility of the new values.
1550
1551           -fpermitted-flt-eval-methods specifies whether the compiler should
1552           allow only the values of "FLT_EVAL_METHOD" specified in C99/C11, or
1553           the extended set of values specified in ISO/IEC TS 18661-3.
1554
1555           style is either "c11" or "ts-18661-3" as appropriate.
1556
1557           The default when in a standards compliant mode (-std=c11 or
1558           similar) is -fpermitted-flt-eval-methods=c11.  The default when in
1559           a GNU dialect (-std=gnu11 or similar) is
1560           -fpermitted-flt-eval-methods=ts-18661-3.
1561
1562       -aux-info filename
1563           Output to the given filename prototyped declarations for all
1564           functions declared and/or defined in a translation unit, including
1565           those in header files.  This option is silently ignored in any
1566           language other than C.
1567
1568           Besides declarations, the file indicates, in comments, the origin
1569           of each declaration (source file and line), whether the declaration
1570           was implicit, prototyped or unprototyped (I, N for new or O for
1571           old, respectively, in the first character after the line number and
1572           the colon), and whether it came from a declaration or a definition
1573           (C or F, respectively, in the following character).  In the case of
1574           function definitions, a K&R-style list of arguments followed by
1575           their declarations is also provided, inside comments, after the
1576           declaration.
1577
1578       -fallow-parameterless-variadic-functions
1579           Accept variadic functions without named parameters.
1580
1581           Although it is possible to define such a function, this is not very
1582           useful as it is not possible to read the arguments.  This is only
1583           supported for C as this construct is allowed by C++.
1584
1585       -fno-asm
1586           Do not recognize "asm", "inline" or "typeof" as a keyword, so that
1587           code can use these words as identifiers.  You can use the keywords
1588           "__asm__", "__inline__" and "__typeof__" instead.  -ansi implies
1589           -fno-asm.
1590
1591           In C++, this switch only affects the "typeof" keyword, since "asm"
1592           and "inline" are standard keywords.  You may want to use the
1593           -fno-gnu-keywords flag instead, which has the same effect.  In C99
1594           mode (-std=c99 or -std=gnu99), this switch only affects the "asm"
1595           and "typeof" keywords, since "inline" is a standard keyword in ISO
1596           C99.
1597
1598       -fno-builtin
1599       -fno-builtin-function
1600           Don't recognize built-in functions that do not begin with
1601           __builtin_ as prefix.
1602
1603           GCC normally generates special code to handle certain built-in
1604           functions more efficiently; for instance, calls to "alloca" may
1605           become single instructions which adjust the stack directly, and
1606           calls to "memcpy" may become inline copy loops.  The resulting code
1607           is often both smaller and faster, but since the function calls no
1608           longer appear as such, you cannot set a breakpoint on those calls,
1609           nor can you change the behavior of the functions by linking with a
1610           different library.  In addition, when a function is recognized as a
1611           built-in function, GCC may use information about that function to
1612           warn about problems with calls to that function, or to generate
1613           more efficient code, even if the resulting code still contains
1614           calls to that function.  For example, warnings are given with
1615           -Wformat for bad calls to "printf" when "printf" is built in and
1616           "strlen" is known not to modify global memory.
1617
1618           With the -fno-builtin-function option only the built-in function
1619           function is disabled.  function must not begin with __builtin_.  If
1620           a function is named that is not built-in in this version of GCC,
1621           this option is ignored.  There is no corresponding
1622           -fbuiltin-function option; if you wish to enable built-in functions
1623           selectively when using -fno-builtin or -ffreestanding, you may
1624           define macros such as:
1625
1626                   #define abs(n)          __builtin_abs ((n))
1627                   #define strcpy(d, s)    __builtin_strcpy ((d), (s))
1628
1629       -fgimple
1630           Enable parsing of function definitions marked with "__GIMPLE".
1631           This is an experimental feature that allows unit testing of GIMPLE
1632           passes.
1633
1634       -fhosted
1635           Assert that compilation targets a hosted environment.  This implies
1636           -fbuiltin.  A hosted environment is one in which the entire
1637           standard library is available, and in which "main" has a return
1638           type of "int".  Examples are nearly everything except a kernel.
1639           This is equivalent to -fno-freestanding.
1640
1641       -ffreestanding
1642           Assert that compilation targets a freestanding environment.  This
1643           implies -fno-builtin.  A freestanding environment is one in which
1644           the standard library may not exist, and program startup may not
1645           necessarily be at "main".  The most obvious example is an OS
1646           kernel.  This is equivalent to -fno-hosted.
1647
1648       -fopenacc
1649           Enable handling of OpenACC directives "#pragma acc" in C/C++ and
1650           "!$acc" in Fortran.  When -fopenacc is specified, the compiler
1651           generates accelerated code according to the OpenACC Application
1652           Programming Interface v2.6 <https://www.openacc.org>.  This option
1653           implies -pthread, and thus is only supported on targets that have
1654           support for -pthread.
1655
1656       -fopenacc-dim=geom
1657           Specify default compute dimensions for parallel offload regions
1658           that do not explicitly specify.  The geom value is a triple of
1659           ':'-separated sizes, in order 'gang', 'worker' and, 'vector'.  A
1660           size can be omitted, to use a target-specific default value.
1661
1662       -fopenmp
1663           Enable handling of OpenMP directives "#pragma omp" in C/C++ and
1664           "!$omp" in Fortran.  When -fopenmp is specified, the compiler
1665           generates parallel code according to the OpenMP Application Program
1666           Interface v4.5 <https://www.openmp.org>.  This option implies
1667           -pthread, and thus is only supported on targets that have support
1668           for -pthread. -fopenmp implies -fopenmp-simd.
1669
1670       -fopenmp-simd
1671           Enable handling of OpenMP's SIMD directives with "#pragma omp" in
1672           C/C++ and "!$omp" in Fortran. Other OpenMP directives are ignored.
1673
1674       -fgnu-tm
1675           When the option -fgnu-tm is specified, the compiler generates code
1676           for the Linux variant of Intel's current Transactional Memory ABI
1677           specification document (Revision 1.1, May 6 2009).  This is an
1678           experimental feature whose interface may change in future versions
1679           of GCC, as the official specification changes.  Please note that
1680           not all architectures are supported for this feature.
1681
1682           For more information on GCC's support for transactional memory,
1683
1684           Note that the transactional memory feature is not supported with
1685           non-call exceptions (-fnon-call-exceptions).
1686
1687       -fms-extensions
1688           Accept some non-standard constructs used in Microsoft header files.
1689
1690           In C++ code, this allows member names in structures to be similar
1691           to previous types declarations.
1692
1693                   typedef int UOW;
1694                   struct ABC {
1695                     UOW UOW;
1696                   };
1697
1698           Some cases of unnamed fields in structures and unions are only
1699           accepted with this option.
1700
1701           Note that this option is off for all targets except for x86 targets
1702           using ms-abi.
1703
1704       -fplan9-extensions
1705           Accept some non-standard constructs used in Plan 9 code.
1706
1707           This enables -fms-extensions, permits passing pointers to
1708           structures with anonymous fields to functions that expect pointers
1709           to elements of the type of the field, and permits referring to
1710           anonymous fields declared using a typedef.    This is only
1711           supported for C, not C++.
1712
1713       -fcond-mismatch
1714           Allow conditional expressions with mismatched types in the second
1715           and third arguments.  The value of such an expression is void.
1716           This option is not supported for C++.
1717
1718       -flax-vector-conversions
1719           Allow implicit conversions between vectors with differing numbers
1720           of elements and/or incompatible element types.  This option should
1721           not be used for new code.
1722
1723       -funsigned-char
1724           Let the type "char" be unsigned, like "unsigned char".
1725
1726           Each kind of machine has a default for what "char" should be.  It
1727           is either like "unsigned char" by default or like "signed char" by
1728           default.
1729
1730           Ideally, a portable program should always use "signed char" or
1731           "unsigned char" when it depends on the signedness of an object.
1732           But many programs have been written to use plain "char" and expect
1733           it to be signed, or expect it to be unsigned, depending on the
1734           machines they were written for.  This option, and its inverse, let
1735           you make such a program work with the opposite default.
1736
1737           The type "char" is always a distinct type from each of "signed
1738           char" or "unsigned char", even though its behavior is always just
1739           like one of those two.
1740
1741       -fsigned-char
1742           Let the type "char" be signed, like "signed char".
1743
1744           Note that this is equivalent to -fno-unsigned-char, which is the
1745           negative form of -funsigned-char.  Likewise, the option
1746           -fno-signed-char is equivalent to -funsigned-char.
1747
1748       -fsigned-bitfields
1749       -funsigned-bitfields
1750       -fno-signed-bitfields
1751       -fno-unsigned-bitfields
1752           These options control whether a bit-field is signed or unsigned,
1753           when the declaration does not use either "signed" or "unsigned".
1754           By default, such a bit-field is signed, because this is consistent:
1755           the basic integer types such as "int" are signed types.
1756
1757       -fsso-struct=endianness
1758           Set the default scalar storage order of structures and unions to
1759           the specified endianness.  The accepted values are big-endian,
1760           little-endian and native for the native endianness of the target
1761           (the default).  This option is not supported for C++.
1762
1763           Warning: the -fsso-struct switch causes GCC to generate code that
1764           is not binary compatible with code generated without it if the
1765           specified endianness is not the native endianness of the target.
1766
1767   Options Controlling C++ Dialect
1768       This section describes the command-line options that are only
1769       meaningful for C++ programs.  You can also use most of the GNU compiler
1770       options regardless of what language your program is in.  For example,
1771       you might compile a file firstClass.C like this:
1772
1773               g++ -g -fstrict-enums -O -c firstClass.C
1774
1775       In this example, only -fstrict-enums is an option meant only for C++
1776       programs; you can use the other options with any language supported by
1777       GCC.
1778
1779       Some options for compiling C programs, such as -std, are also relevant
1780       for C++ programs.
1781
1782       Here is a list of options that are only for compiling C++ programs:
1783
1784       -fabi-version=n
1785           Use version n of the C++ ABI.  The default is version 0.
1786
1787           Version 0 refers to the version conforming most closely to the C++
1788           ABI specification.  Therefore, the ABI obtained using version 0
1789           will change in different versions of G++ as ABI bugs are fixed.
1790
1791           Version 1 is the version of the C++ ABI that first appeared in G++
1792           3.2.
1793
1794           Version 2 is the version of the C++ ABI that first appeared in G++
1795           3.4, and was the default through G++ 4.9.
1796
1797           Version 3 corrects an error in mangling a constant address as a
1798           template argument.
1799
1800           Version 4, which first appeared in G++ 4.5, implements a standard
1801           mangling for vector types.
1802
1803           Version 5, which first appeared in G++ 4.6, corrects the mangling
1804           of attribute const/volatile on function pointer types, decltype of
1805           a plain decl, and use of a function parameter in the declaration of
1806           another parameter.
1807
1808           Version 6, which first appeared in G++ 4.7, corrects the promotion
1809           behavior of C++11 scoped enums and the mangling of template
1810           argument packs, const/static_cast, prefix ++ and --, and a class
1811           scope function used as a template argument.
1812
1813           Version 7, which first appeared in G++ 4.8, that treats nullptr_t
1814           as a builtin type and corrects the mangling of lambdas in default
1815           argument scope.
1816
1817           Version 8, which first appeared in G++ 4.9, corrects the
1818           substitution behavior of function types with function-cv-
1819           qualifiers.
1820
1821           Version 9, which first appeared in G++ 5.2, corrects the alignment
1822           of "nullptr_t".
1823
1824           Version 10, which first appeared in G++ 6.1, adds mangling of
1825           attributes that affect type identity, such as ia32 calling
1826           convention attributes (e.g. stdcall).
1827
1828           Version 11, which first appeared in G++ 7, corrects the mangling of
1829           sizeof... expressions and operator names.  For multiple entities
1830           with the same name within a function, that are declared in
1831           different scopes, the mangling now changes starting with the
1832           twelfth occurrence.  It also implies -fnew-inheriting-ctors.
1833
1834           Version 12, which first appeared in G++ 8, corrects the calling
1835           conventions for empty classes on the x86_64 target and for classes
1836           with only deleted copy/move constructors.  It accidentally changes
1837           the calling convention for classes with a deleted copy constructor
1838           and a trivial move constructor.
1839
1840           Version 13, which first appeared in G++ 8.2, fixes the accidental
1841           change in version 12.
1842
1843           Version 14, which first appeared in G++ 10, corrects the mangling
1844           of the nullptr expression.
1845
1846           See also -Wabi.
1847
1848       -fabi-compat-version=n
1849           On targets that support strong aliases, G++ works around mangling
1850           changes by creating an alias with the correct mangled name when
1851           defining a symbol with an incorrect mangled name.  This switch
1852           specifies which ABI version to use for the alias.
1853
1854           With -fabi-version=0 (the default), this defaults to 11 (GCC 7
1855           compatibility).  If another ABI version is explicitly selected,
1856           this defaults to 0.  For compatibility with GCC versions 3.2
1857           through 4.9, use -fabi-compat-version=2.
1858
1859           If this option is not provided but -Wabi=n is, that version is used
1860           for compatibility aliases.  If this option is provided along with
1861           -Wabi (without the version), the version from this option is used
1862           for the warning.
1863
1864       -fno-access-control
1865           Turn off all access checking.  This switch is mainly useful for
1866           working around bugs in the access control code.
1867
1868       -faligned-new
1869           Enable support for C++17 "new" of types that require more alignment
1870           than "void* ::operator new(std::size_t)" provides.  A numeric
1871           argument such as "-faligned-new=32" can be used to specify how much
1872           alignment (in bytes) is provided by that function, but few users
1873           will need to override the default of "alignof(std::max_align_t)".
1874
1875           This flag is enabled by default for -std=c++17.
1876
1877       -fchar8_t
1878       -fno-char8_t
1879           Enable support for "char8_t" as adopted for C++2a.  This includes
1880           the addition of a new "char8_t" fundamental type, changes to the
1881           types of UTF-8 string and character literals, new signatures for
1882           user-defined literals, associated standard library updates, and new
1883           "__cpp_char8_t" and "__cpp_lib_char8_t" feature test macros.
1884
1885           This option enables functions to be overloaded for ordinary and
1886           UTF-8 strings:
1887
1888                   int f(const char *);    // #1
1889                   int f(const char8_t *); // #2
1890                   int v1 = f("text");     // Calls #1
1891                   int v2 = f(u8"text");   // Calls #2
1892
1893           and introduces new signatures for user-defined literals:
1894
1895                   int operator""_udl1(char8_t);
1896                   int v3 = u8'x'_udl1;
1897                   int operator""_udl2(const char8_t*, std::size_t);
1898                   int v4 = u8"text"_udl2;
1899                   template<typename T, T...> int operator""_udl3();
1900                   int v5 = u8"text"_udl3;
1901
1902           The change to the types of UTF-8 string and character literals
1903           introduces incompatibilities with ISO C++11 and later standards.
1904           For example, the following code is well-formed under ISO C++11, but
1905           is ill-formed when -fchar8_t is specified.
1906
1907                   char ca[] = u8"xx";     // error: char-array initialized from wide
1908                                           //        string
1909                   const char *cp = u8"xx";// error: invalid conversion from
1910                                           //        `const char8_t*' to `const char*'
1911                   int f(const char*);
1912                   auto v = f(u8"xx");     // error: invalid conversion from
1913                                           //        `const char8_t*' to `const char*'
1914                   std::string s{u8"xx"};  // error: no matching function for call to
1915                                           //        `std::basic_string<char>::basic_string()'
1916                   using namespace std::literals;
1917                   s = u8"xx"s;            // error: conversion from
1918                                           //        `basic_string<char8_t>' to non-scalar
1919                                           //        type `basic_string<char>' requested
1920
1921       -fcheck-new
1922           Check that the pointer returned by "operator new" is non-null
1923           before attempting to modify the storage allocated.  This check is
1924           normally unnecessary because the C++ standard specifies that
1925           "operator new" only returns 0 if it is declared "throw()", in which
1926           case the compiler always checks the return value even without this
1927           option.  In all other cases, when "operator new" has a non-empty
1928           exception specification, memory exhaustion is signalled by throwing
1929           "std::bad_alloc".  See also new (nothrow).
1930
1931       -fconcepts
1932       -fconcepts-ts
1933           Below -std=c++2a, -fconcepts enables support for the C++ Extensions
1934           for Concepts Technical Specification, ISO 19217 (2015).
1935
1936           With -std=c++2a and above, Concepts are part of the language
1937           standard, so -fconcepts defaults to on.  But the standard
1938           specification of Concepts differs significantly from the TS, so
1939           some constructs that were allowed in the TS but didn't make it into
1940           the standard can still be enabled by -fconcepts-ts.
1941
1942       -fconstexpr-depth=n
1943           Set the maximum nested evaluation depth for C++11 constexpr
1944           functions to n.  A limit is needed to detect endless recursion
1945           during constant expression evaluation.  The minimum specified by
1946           the standard is 512.
1947
1948       -fconstexpr-cache-depth=n
1949           Set the maximum level of nested evaluation depth for C++11
1950           constexpr functions that will be cached to n.  This is a heuristic
1951           that trades off compilation speed (when the cache avoids repeated
1952           calculations) against memory consumption (when the cache grows very
1953           large from highly recursive evaluations).  The default is 8.  Very
1954           few users are likely to want to adjust it, but if your code does
1955           heavy constexpr calculations you might want to experiment to find
1956           which value works best for you.
1957
1958       -fconstexpr-loop-limit=n
1959           Set the maximum number of iterations for a loop in C++14 constexpr
1960           functions to n.  A limit is needed to detect infinite loops during
1961           constant expression evaluation.  The default is 262144 (1<<18).
1962
1963       -fconstexpr-ops-limit=n
1964           Set the maximum number of operations during a single constexpr
1965           evaluation.  Even when number of iterations of a single loop is
1966           limited with the above limit, if there are several nested loops and
1967           each of them has many iterations but still smaller than the above
1968           limit, or if in a body of some loop or even outside of a loop too
1969           many expressions need to be evaluated, the resulting constexpr
1970           evaluation might take too long.  The default is 33554432 (1<<25).
1971
1972       -fcoroutines
1973           Enable support for the C++ coroutines extension (experimental).
1974
1975       -fno-elide-constructors
1976           The C++ standard allows an implementation to omit creating a
1977           temporary that is only used to initialize another object of the
1978           same type.  Specifying this option disables that optimization, and
1979           forces G++ to call the copy constructor in all cases.  This option
1980           also causes G++ to call trivial member functions which otherwise
1981           would be expanded inline.
1982
1983           In C++17, the compiler is required to omit these temporaries, but
1984           this option still affects trivial member functions.
1985
1986       -fno-enforce-eh-specs
1987           Don't generate code to check for violation of exception
1988           specifications at run time.  This option violates the C++ standard,
1989           but may be useful for reducing code size in production builds, much
1990           like defining "NDEBUG".  This does not give user code permission to
1991           throw exceptions in violation of the exception specifications; the
1992           compiler still optimizes based on the specifications, so throwing
1993           an unexpected exception results in undefined behavior at run time.
1994
1995       -fextern-tls-init
1996       -fno-extern-tls-init
1997           The C++11 and OpenMP standards allow "thread_local" and
1998           "threadprivate" variables to have dynamic (runtime) initialization.
1999           To support this, any use of such a variable goes through a wrapper
2000           function that performs any necessary initialization.  When the use
2001           and definition of the variable are in the same translation unit,
2002           this overhead can be optimized away, but when the use is in a
2003           different translation unit there is significant overhead even if
2004           the variable doesn't actually need dynamic initialization.  If the
2005           programmer can be sure that no use of the variable in a non-
2006           defining TU needs to trigger dynamic initialization (either because
2007           the variable is statically initialized, or a use of the variable in
2008           the defining TU will be executed before any uses in another TU),
2009           they can avoid this overhead with the -fno-extern-tls-init option.
2010
2011           On targets that support symbol aliases, the default is
2012           -fextern-tls-init.  On targets that do not support symbol aliases,
2013           the default is -fno-extern-tls-init.
2014
2015       -fno-gnu-keywords
2016           Do not recognize "typeof" as a keyword, so that code can use this
2017           word as an identifier.  You can use the keyword "__typeof__"
2018           instead.  This option is implied by the strict ISO C++ dialects:
2019           -ansi, -std=c++98, -std=c++11, etc.
2020
2021       -fno-implicit-templates
2022           Never emit code for non-inline templates that are instantiated
2023           implicitly (i.e. by use); only emit code for explicit
2024           instantiations.  If you use this option, you must take care to
2025           structure your code to include all the necessary explicit
2026           instantiations to avoid getting undefined symbols at link time.
2027
2028       -fno-implicit-inline-templates
2029           Don't emit code for implicit instantiations of inline templates,
2030           either.  The default is to handle inlines differently so that
2031           compiles with and without optimization need the same set of
2032           explicit instantiations.
2033
2034       -fno-implement-inlines
2035           To save space, do not emit out-of-line copies of inline functions
2036           controlled by "#pragma implementation".  This causes linker errors
2037           if these functions are not inlined everywhere they are called.
2038
2039       -fms-extensions
2040           Disable Wpedantic warnings about constructs used in MFC, such as
2041           implicit int and getting a pointer to member function via non-
2042           standard syntax.
2043
2044       -fnew-inheriting-ctors
2045           Enable the P0136 adjustment to the semantics of C++11 constructor
2046           inheritance.  This is part of C++17 but also considered to be a
2047           Defect Report against C++11 and C++14.  This flag is enabled by
2048           default unless -fabi-version=10 or lower is specified.
2049
2050       -fnew-ttp-matching
2051           Enable the P0522 resolution to Core issue 150, template template
2052           parameters and default arguments: this allows a template with
2053           default template arguments as an argument for a template template
2054           parameter with fewer template parameters.  This flag is enabled by
2055           default for -std=c++17.
2056
2057       -fno-nonansi-builtins
2058           Disable built-in declarations of functions that are not mandated by
2059           ANSI/ISO C.  These include "ffs", "alloca", "_exit", "index",
2060           "bzero", "conjf", and other related functions.
2061
2062       -fnothrow-opt
2063           Treat a "throw()" exception specification as if it were a
2064           "noexcept" specification to reduce or eliminate the text size
2065           overhead relative to a function with no exception specification.
2066           If the function has local variables of types with non-trivial
2067           destructors, the exception specification actually makes the
2068           function smaller because the EH cleanups for those variables can be
2069           optimized away.  The semantic effect is that an exception thrown
2070           out of a function with such an exception specification results in a
2071           call to "terminate" rather than "unexpected".
2072
2073       -fno-operator-names
2074           Do not treat the operator name keywords "and", "bitand", "bitor",
2075           "compl", "not", "or" and "xor" as synonyms as keywords.
2076
2077       -fno-optional-diags
2078           Disable diagnostics that the standard says a compiler does not need
2079           to issue.  Currently, the only such diagnostic issued by G++ is the
2080           one for a name having multiple meanings within a class.
2081
2082       -fpermissive
2083           Downgrade some diagnostics about nonconformant code from errors to
2084           warnings.  Thus, using -fpermissive allows some nonconforming code
2085           to compile.
2086
2087       -fno-pretty-templates
2088           When an error message refers to a specialization of a function
2089           template, the compiler normally prints the signature of the
2090           template followed by the template arguments and any typedefs or
2091           typenames in the signature (e.g. "void f(T) [with T = int]" rather
2092           than "void f(int)") so that it's clear which template is involved.
2093           When an error message refers to a specialization of a class
2094           template, the compiler omits any template arguments that match the
2095           default template arguments for that template.  If either of these
2096           behaviors make it harder to understand the error message rather
2097           than easier, you can use -fno-pretty-templates to disable them.
2098
2099       -fno-rtti
2100           Disable generation of information about every class with virtual
2101           functions for use by the C++ run-time type identification features
2102           ("dynamic_cast" and "typeid").  If you don't use those parts of the
2103           language, you can save some space by using this flag.  Note that
2104           exception handling uses the same information, but G++ generates it
2105           as needed. The "dynamic_cast" operator can still be used for casts
2106           that do not require run-time type information, i.e. casts to "void
2107           *" or to unambiguous base classes.
2108
2109           Mixing code compiled with -frtti with that compiled with -fno-rtti
2110           may not work.  For example, programs may fail to link if a class
2111           compiled with -fno-rtti is used as a base for a class compiled with
2112           -frtti.
2113
2114       -fsized-deallocation
2115           Enable the built-in global declarations
2116
2117                   void operator delete (void *, std::size_t) noexcept;
2118                   void operator delete[] (void *, std::size_t) noexcept;
2119
2120           as introduced in C++14.  This is useful for user-defined
2121           replacement deallocation functions that, for example, use the size
2122           of the object to make deallocation faster.  Enabled by default
2123           under -std=c++14 and above.  The flag -Wsized-deallocation warns
2124           about places that might want to add a definition.
2125
2126       -fstrict-enums
2127           Allow the compiler to optimize using the assumption that a value of
2128           enumerated type can only be one of the values of the enumeration
2129           (as defined in the C++ standard; basically, a value that can be
2130           represented in the minimum number of bits needed to represent all
2131           the enumerators).  This assumption may not be valid if the program
2132           uses a cast to convert an arbitrary integer value to the enumerated
2133           type.
2134
2135       -fstrong-eval-order
2136           Evaluate member access, array subscripting, and shift expressions
2137           in left-to-right order, and evaluate assignment in right-to-left
2138           order, as adopted for C++17.  Enabled by default with -std=c++17.
2139           -fstrong-eval-order=some enables just the ordering of member access
2140           and shift expressions, and is the default without -std=c++17.
2141
2142       -ftemplate-backtrace-limit=n
2143           Set the maximum number of template instantiation notes for a single
2144           warning or error to n.  The default value is 10.
2145
2146       -ftemplate-depth=n
2147           Set the maximum instantiation depth for template classes to n.  A
2148           limit on the template instantiation depth is needed to detect
2149           endless recursions during template class instantiation.  ANSI/ISO
2150           C++ conforming programs must not rely on a maximum depth greater
2151           than 17 (changed to 1024 in C++11).  The default value is 900, as
2152           the compiler can run out of stack space before hitting 1024 in some
2153           situations.
2154
2155       -fno-threadsafe-statics
2156           Do not emit the extra code to use the routines specified in the C++
2157           ABI for thread-safe initialization of local statics.  You can use
2158           this option to reduce code size slightly in code that doesn't need
2159           to be thread-safe.
2160
2161       -fuse-cxa-atexit
2162           Register destructors for objects with static storage duration with
2163           the "__cxa_atexit" function rather than the "atexit" function.
2164           This option is required for fully standards-compliant handling of
2165           static destructors, but only works if your C library supports
2166           "__cxa_atexit".
2167
2168       -fno-use-cxa-get-exception-ptr
2169           Don't use the "__cxa_get_exception_ptr" runtime routine.  This
2170           causes "std::uncaught_exception" to be incorrect, but is necessary
2171           if the runtime routine is not available.
2172
2173       -fvisibility-inlines-hidden
2174           This switch declares that the user does not attempt to compare
2175           pointers to inline functions or methods where the addresses of the
2176           two functions are taken in different shared objects.
2177
2178           The effect of this is that GCC may, effectively, mark inline
2179           methods with "__attribute__ ((visibility ("hidden")))" so that they
2180           do not appear in the export table of a DSO and do not require a PLT
2181           indirection when used within the DSO.  Enabling this option can
2182           have a dramatic effect on load and link times of a DSO as it
2183           massively reduces the size of the dynamic export table when the
2184           library makes heavy use of templates.
2185
2186           The behavior of this switch is not quite the same as marking the
2187           methods as hidden directly, because it does not affect static
2188           variables local to the function or cause the compiler to deduce
2189           that the function is defined in only one shared object.
2190
2191           You may mark a method as having a visibility explicitly to negate
2192           the effect of the switch for that method.  For example, if you do
2193           want to compare pointers to a particular inline method, you might
2194           mark it as having default visibility.  Marking the enclosing class
2195           with explicit visibility has no effect.
2196
2197           Explicitly instantiated inline methods are unaffected by this
2198           option as their linkage might otherwise cross a shared library
2199           boundary.
2200
2201       -fvisibility-ms-compat
2202           This flag attempts to use visibility settings to make GCC's C++
2203           linkage model compatible with that of Microsoft Visual Studio.
2204
2205           The flag makes these changes to GCC's linkage model:
2206
2207           1.  It sets the default visibility to "hidden", like
2208               -fvisibility=hidden.
2209
2210           2.  Types, but not their members, are not hidden by default.
2211
2212           3.  The One Definition Rule is relaxed for types without explicit
2213               visibility specifications that are defined in more than one
2214               shared object: those declarations are permitted if they are
2215               permitted when this option is not used.
2216
2217           In new code it is better to use -fvisibility=hidden and export
2218           those classes that are intended to be externally visible.
2219           Unfortunately it is possible for code to rely, perhaps
2220           accidentally, on the Visual Studio behavior.
2221
2222           Among the consequences of these changes are that static data
2223           members of the same type with the same name but defined in
2224           different shared objects are different, so changing one does not
2225           change the other; and that pointers to function members defined in
2226           different shared objects may not compare equal.  When this flag is
2227           given, it is a violation of the ODR to define types with the same
2228           name differently.
2229
2230       -fno-weak
2231           Do not use weak symbol support, even if it is provided by the
2232           linker.  By default, G++ uses weak symbols if they are available.
2233           This option exists only for testing, and should not be used by end-
2234           users; it results in inferior code and has no benefits.  This
2235           option may be removed in a future release of G++.
2236
2237       -fext-numeric-literals (C++ and Objective-C++ only)
2238           Accept imaginary, fixed-point, or machine-defined literal number
2239           suffixes as GNU extensions.  When this option is turned off these
2240           suffixes are treated as C++11 user-defined literal numeric
2241           suffixes.  This is on by default for all pre-C++11 dialects and all
2242           GNU dialects: -std=c++98, -std=gnu++98, -std=gnu++11, -std=gnu++14.
2243           This option is off by default for ISO C++11 onwards (-std=c++11,
2244           ...).
2245
2246       -nostdinc++
2247           Do not search for header files in the standard directories specific
2248           to C++, but do still search the other standard directories.  (This
2249           option is used when building the C++ library.)
2250
2251       In addition, these warning options have meanings only for C++ programs:
2252
2253       -Wabi-tag (C++ and Objective-C++ only)
2254           Warn when a type with an ABI tag is used in a context that does not
2255           have that ABI tag.  See C++ Attributes for more information about
2256           ABI tags.
2257
2258       -Wcomma-subscript (C++ and Objective-C++ only)
2259           Warn about uses of a comma expression within a subscripting
2260           expression.  This usage was deprecated in C++2a.  However, a comma
2261           expression wrapped in "( )" is not deprecated.  Example:
2262
2263                   void f(int *a, int b, int c) {
2264                       a[b,c];     // deprecated
2265                       a[(b,c)];   // OK
2266                   }
2267
2268           Enabled by default with -std=c++2a.
2269
2270       -Wctor-dtor-privacy (C++ and Objective-C++ only)
2271           Warn when a class seems unusable because all the constructors or
2272           destructors in that class are private, and it has neither friends
2273           nor public static member functions.  Also warn if there are no non-
2274           private methods, and there's at least one private member function
2275           that isn't a constructor or destructor.
2276
2277       -Wdelete-non-virtual-dtor (C++ and Objective-C++ only)
2278           Warn when "delete" is used to destroy an instance of a class that
2279           has virtual functions and non-virtual destructor. It is unsafe to
2280           delete an instance of a derived class through a pointer to a base
2281           class if the base class does not have a virtual destructor.  This
2282           warning is enabled by -Wall.
2283
2284       -Wdeprecated-copy (C++ and Objective-C++ only)
2285           Warn that the implicit declaration of a copy constructor or copy
2286           assignment operator is deprecated if the class has a user-provided
2287           copy constructor or copy assignment operator, in C++11 and up.
2288           This warning is enabled by -Wextra.  With -Wdeprecated-copy-dtor,
2289           also deprecate if the class has a user-provided destructor.
2290
2291       -Wno-init-list-lifetime (C++ and Objective-C++ only)
2292           Do not warn about uses of "std::initializer_list" that are likely
2293           to result in dangling pointers.  Since the underlying array for an
2294           "initializer_list" is handled like a normal C++ temporary object,
2295           it is easy to inadvertently keep a pointer to the array past the
2296           end of the array's lifetime.  For example:
2297
2298           *   If a function returns a temporary "initializer_list", or a
2299               local "initializer_list" variable, the array's lifetime ends at
2300               the end of the return statement, so the value returned has a
2301               dangling pointer.
2302
2303           *   If a new-expression creates an "initializer_list", the array
2304               only lives until the end of the enclosing full-expression, so
2305               the "initializer_list" in the heap has a dangling pointer.
2306
2307           *   When an "initializer_list" variable is assigned from a brace-
2308               enclosed initializer list, the temporary array created for the
2309               right side of the assignment only lives until the end of the
2310               full-expression, so at the next statement the
2311               "initializer_list" variable has a dangling pointer.
2312
2313                       // li's initial underlying array lives as long as li
2314                       std::initializer_list<int> li = { 1,2,3 };
2315                       // assignment changes li to point to a temporary array
2316                       li = { 4, 5 };
2317                       // now the temporary is gone and li has a dangling pointer
2318                       int i = li.begin()[0] // undefined behavior
2319
2320           *   When a list constructor stores the "begin" pointer from the
2321               "initializer_list" argument, this doesn't extend the lifetime
2322               of the array, so if a class variable is constructed from a
2323               temporary "initializer_list", the pointer is left dangling by
2324               the end of the variable declaration statement.
2325
2326       -Wno-literal-suffix (C++ and Objective-C++ only)
2327           Do not warn when a string or character literal is followed by a ud-
2328           suffix which does not begin with an underscore.  As a conforming
2329           extension, GCC treats such suffixes as separate preprocessing
2330           tokens in order to maintain backwards compatibility with code that
2331           uses formatting macros from "<inttypes.h>".  For example:
2332
2333                   #define __STDC_FORMAT_MACROS
2334                   #include <inttypes.h>
2335                   #include <stdio.h>
2336
2337                   int main() {
2338                     int64_t i64 = 123;
2339                     printf("My int64: %" PRId64"\n", i64);
2340                   }
2341
2342           In this case, "PRId64" is treated as a separate preprocessing
2343           token.
2344
2345           This option also controls warnings when a user-defined literal
2346           operator is declared with a literal suffix identifier that doesn't
2347           begin with an underscore. Literal suffix identifiers that don't
2348           begin with an underscore are reserved for future standardization.
2349
2350           These warnings are enabled by default.
2351
2352       -Wno-narrowing (C++ and Objective-C++ only)
2353           For C++11 and later standards, narrowing conversions are diagnosed
2354           by default, as required by the standard.  A narrowing conversion
2355           from a constant produces an error, and a narrowing conversion from
2356           a non-constant produces a warning, but -Wno-narrowing suppresses
2357           the diagnostic.  Note that this does not affect the meaning of
2358           well-formed code; narrowing conversions are still considered ill-
2359           formed in SFINAE contexts.
2360
2361           With -Wnarrowing in C++98, warn when a narrowing conversion
2362           prohibited by C++11 occurs within { }, e.g.
2363
2364                   int i = { 2.2 }; // error: narrowing from double to int
2365
2366           This flag is included in -Wall and -Wc++11-compat.
2367
2368       -Wnoexcept (C++ and Objective-C++ only)
2369           Warn when a noexcept-expression evaluates to false because of a
2370           call to a function that does not have a non-throwing exception
2371           specification (i.e. "throw()" or "noexcept") but is known by the
2372           compiler to never throw an exception.
2373
2374       -Wnoexcept-type (C++ and Objective-C++ only)
2375           Warn if the C++17 feature making "noexcept" part of a function type
2376           changes the mangled name of a symbol relative to C++14.  Enabled by
2377           -Wabi and -Wc++17-compat.
2378
2379           As an example:
2380
2381                   template <class T> void f(T t) { t(); };
2382                   void g() noexcept;
2383                   void h() { f(g); }
2384
2385           In C++14, "f" calls "f<void(*)()>", but in C++17 it calls
2386           "f<void(*)()noexcept>".
2387
2388       -Wclass-memaccess (C++ and Objective-C++ only)
2389           Warn when the destination of a call to a raw memory function such
2390           as "memset" or "memcpy" is an object of class type, and when
2391           writing into such an object might bypass the class non-trivial or
2392           deleted constructor or copy assignment, violate const-correctness
2393           or encapsulation, or corrupt virtual table pointers.  Modifying the
2394           representation of such objects may violate invariants maintained by
2395           member functions of the class.  For example, the call to "memset"
2396           below is undefined because it modifies a non-trivial class object
2397           and is, therefore, diagnosed.  The safe way to either initialize or
2398           clear the storage of objects of such types is by using the
2399           appropriate constructor or assignment operator, if one is
2400           available.
2401
2402                   std::string str = "abc";
2403                   memset (&str, 0, sizeof str);
2404
2405           The -Wclass-memaccess option is enabled by -Wall.  Explicitly
2406           casting the pointer to the class object to "void *" or to a type
2407           that can be safely accessed by the raw memory function suppresses
2408           the warning.
2409
2410       -Wnon-virtual-dtor (C++ and Objective-C++ only)
2411           Warn when a class has virtual functions and an accessible non-
2412           virtual destructor itself or in an accessible polymorphic base
2413           class, in which case it is possible but unsafe to delete an
2414           instance of a derived class through a pointer to the class itself
2415           or base class.  This warning is automatically enabled if -Weffc++
2416           is specified.
2417
2418       -Wregister (C++ and Objective-C++ only)
2419           Warn on uses of the "register" storage class specifier, except when
2420           it is part of the GNU Explicit Register Variables extension.  The
2421           use of the "register" keyword as storage class specifier has been
2422           deprecated in C++11 and removed in C++17.  Enabled by default with
2423           -std=c++17.
2424
2425       -Wreorder (C++ and Objective-C++ only)
2426           Warn when the order of member initializers given in the code does
2427           not match the order in which they must be executed.  For instance:
2428
2429                   struct A {
2430                     int i;
2431                     int j;
2432                     A(): j (0), i (1) { }
2433                   };
2434
2435           The compiler rearranges the member initializers for "i" and "j" to
2436           match the declaration order of the members, emitting a warning to
2437           that effect.  This warning is enabled by -Wall.
2438
2439       -Wno-pessimizing-move (C++ and Objective-C++ only)
2440           This warning warns when a call to "std::move" prevents copy
2441           elision.  A typical scenario when copy elision can occur is when
2442           returning in a function with a class return type, when the
2443           expression being returned is the name of a non-volatile automatic
2444           object, and is not a function parameter, and has the same type as
2445           the function return type.
2446
2447                   struct T {
2448                   ...
2449                   };
2450                   T fn()
2451                   {
2452                     T t;
2453                     ...
2454                     return std::move (t);
2455                   }
2456
2457           But in this example, the "std::move" call prevents copy elision.
2458
2459           This warning is enabled by -Wall.
2460
2461       -Wno-redundant-move (C++ and Objective-C++ only)
2462           This warning warns about redundant calls to "std::move"; that is,
2463           when a move operation would have been performed even without the
2464           "std::move" call.  This happens because the compiler is forced to
2465           treat the object as if it were an rvalue in certain situations such
2466           as returning a local variable, where copy elision isn't applicable.
2467           Consider:
2468
2469                   struct T {
2470                   ...
2471                   };
2472                   T fn(T t)
2473                   {
2474                     ...
2475                     return std::move (t);
2476                   }
2477
2478           Here, the "std::move" call is redundant.  Because G++ implements
2479           Core Issue 1579, another example is:
2480
2481                   struct T { // convertible to U
2482                   ...
2483                   };
2484                   struct U {
2485                   ...
2486                   };
2487                   U fn()
2488                   {
2489                     T t;
2490                     ...
2491                     return std::move (t);
2492                   }
2493
2494           In this example, copy elision isn't applicable because the type of
2495           the expression being returned and the function return type differ,
2496           yet G++ treats the return value as if it were designated by an
2497           rvalue.
2498
2499           This warning is enabled by -Wextra.
2500
2501       -Wredundant-tags (C++ and Objective-C++ only)
2502           Warn about redundant class-key and enum-key in references to class
2503           types and enumerated types in contexts where the key can be
2504           eliminated without causing an ambiguity.  For example:
2505
2506                   struct foo;
2507                   struct foo *p;   // warn that keyword struct can be eliminated
2508
2509           On the other hand, in this example there is no warning:
2510
2511                   struct foo;
2512                   void foo ();   // "hides" struct foo
2513                   void bar (struct foo&);  // no warning, keyword struct is necessary
2514
2515       -Wno-subobject-linkage (C++ and Objective-C++ only)
2516           Do not warn if a class type has a base or a field whose type uses
2517           the anonymous namespace or depends on a type with no linkage.  If a
2518           type A depends on a type B with no or internal linkage, defining it
2519           in multiple translation units would be an ODR violation because the
2520           meaning of B is different in each translation unit.  If A only
2521           appears in a single translation unit, the best way to silence the
2522           warning is to give it internal linkage by putting it in an
2523           anonymous namespace as well.  The compiler doesn't give this
2524           warning for types defined in the main .C file, as those are
2525           unlikely to have multiple definitions.  -Wsubobject-linkage is
2526           enabled by default.
2527
2528       -Weffc++ (C++ and Objective-C++ only)
2529           Warn about violations of the following style guidelines from Scott
2530           Meyers' Effective C++ series of books:
2531
2532           *   Define a copy constructor and an assignment operator for
2533               classes with dynamically-allocated memory.
2534
2535           *   Prefer initialization to assignment in constructors.
2536
2537           *   Have "operator=" return a reference to *this.
2538
2539           *   Don't try to return a reference when you must return an object.
2540
2541           *   Distinguish between prefix and postfix forms of increment and
2542               decrement operators.
2543
2544           *   Never overload "&&", "||", or ",".
2545
2546           This option also enables -Wnon-virtual-dtor, which is also one of
2547           the effective C++ recommendations.  However, the check is extended
2548           to warn about the lack of virtual destructor in accessible non-
2549           polymorphic bases classes too.
2550
2551           When selecting this option, be aware that the standard library
2552           headers do not obey all of these guidelines; use grep -v to filter
2553           out those warnings.
2554
2555       -Wstrict-null-sentinel (C++ and Objective-C++ only)
2556           Warn about the use of an uncasted "NULL" as sentinel.  When
2557           compiling only with GCC this is a valid sentinel, as "NULL" is
2558           defined to "__null".  Although it is a null pointer constant rather
2559           than a null pointer, it is guaranteed to be of the same size as a
2560           pointer.  But this use is not portable across different compilers.
2561
2562       -Wno-non-template-friend (C++ and Objective-C++ only)
2563           Disable warnings when non-template friend functions are declared
2564           within a template.  In very old versions of GCC that predate
2565           implementation of the ISO standard, declarations such as friend int
2566           foo(int), where the name of the friend is an unqualified-id, could
2567           be interpreted as a particular specialization of a template
2568           function; the warning exists to diagnose compatibility problems,
2569           and is enabled by default.
2570
2571       -Wold-style-cast (C++ and Objective-C++ only)
2572           Warn if an old-style (C-style) cast to a non-void type is used
2573           within a C++ program.  The new-style casts ("dynamic_cast",
2574           "static_cast", "reinterpret_cast", and "const_cast") are less
2575           vulnerable to unintended effects and much easier to search for.
2576
2577       -Woverloaded-virtual (C++ and Objective-C++ only)
2578           Warn when a function declaration hides virtual functions from a
2579           base class.  For example, in:
2580
2581                   struct A {
2582                     virtual void f();
2583                   };
2584
2585                   struct B: public A {
2586                     void f(int);
2587                   };
2588
2589           the "A" class version of "f" is hidden in "B", and code like:
2590
2591                   B* b;
2592                   b->f();
2593
2594           fails to compile.
2595
2596       -Wno-pmf-conversions (C++ and Objective-C++ only)
2597           Disable the diagnostic for converting a bound pointer to member
2598           function to a plain pointer.
2599
2600       -Wsign-promo (C++ and Objective-C++ only)
2601           Warn when overload resolution chooses a promotion from unsigned or
2602           enumerated type to a signed type, over a conversion to an unsigned
2603           type of the same size.  Previous versions of G++ tried to preserve
2604           unsignedness, but the standard mandates the current behavior.
2605
2606       -Wtemplates (C++ and Objective-C++ only)
2607           Warn when a primary template declaration is encountered.  Some
2608           coding rules disallow templates, and this may be used to enforce
2609           that rule.  The warning is inactive inside a system header file,
2610           such as the STL, so one can still use the STL.  One may also
2611           instantiate or specialize templates.
2612
2613       -Wmismatched-tags (C++ and Objective-C++ only)
2614           Warn for declarations of structs, classes, and class templates and
2615           their specializations with a class-key that does not match either
2616           the definition or the first declaration if no definition is
2617           provided.
2618
2619           For example, the declaration of "struct Object" in the argument
2620           list of "draw" triggers the warning.  To avoid it, either remove
2621           the redundant class-key "struct" or replace it with "class" to
2622           match its definition.
2623
2624                   class Object {
2625                   public:
2626                     virtual ~Object () = 0;
2627                   };
2628                   void draw (struct Object*);
2629
2630           It is not wrong to declare a class with the class-key "struct" as
2631           the example above shows.  The -Wmismatched-tags option is intended
2632           to help achieve a consistent style of class declarations.  In code
2633           that is intended to be portable to Windows-based compilers the
2634           warning helps prevent unresolved references due to the difference
2635           in the mangling of symbols declared with different class-keys.  The
2636           option can be used either on its own or in conjunction with
2637           -Wredundant-tags.
2638
2639       -Wmultiple-inheritance (C++ and Objective-C++ only)
2640           Warn when a class is defined with multiple direct base classes.
2641           Some coding rules disallow multiple inheritance, and this may be
2642           used to enforce that rule.  The warning is inactive inside a system
2643           header file, such as the STL, so one can still use the STL.  One
2644           may also define classes that indirectly use multiple inheritance.
2645
2646       -Wvirtual-inheritance
2647           Warn when a class is defined with a virtual direct base class.
2648           Some coding rules disallow multiple inheritance, and this may be
2649           used to enforce that rule.  The warning is inactive inside a system
2650           header file, such as the STL, so one can still use the STL.  One
2651           may also define classes that indirectly use virtual inheritance.
2652
2653       -Wno-virtual-move-assign
2654           Suppress warnings about inheriting from a virtual base with a non-
2655           trivial C++11 move assignment operator.  This is dangerous because
2656           if the virtual base is reachable along more than one path, it is
2657           moved multiple times, which can mean both objects end up in the
2658           moved-from state.  If the move assignment operator is written to
2659           avoid moving from a moved-from object, this warning can be
2660           disabled.
2661
2662       -Wnamespaces
2663           Warn when a namespace definition is opened.  Some coding rules
2664           disallow namespaces, and this may be used to enforce that rule.
2665           The warning is inactive inside a system header file, such as the
2666           STL, so one can still use the STL.  One may also use using
2667           directives and qualified names.
2668
2669       -Wno-terminate (C++ and Objective-C++ only)
2670           Disable the warning about a throw-expression that will immediately
2671           result in a call to "terminate".
2672
2673       -Wno-class-conversion (C++ and Objective-C++ only)
2674           Do not warn when a conversion function converts an object to the
2675           same type, to a base class of that type, or to void; such a
2676           conversion function will never be called.
2677
2678       -Wvolatile (C++ and Objective-C++ only)
2679           Warn about deprecated uses of the "volatile" qualifier.  This
2680           includes postfix and prefix "++" and "--" expressions of
2681           "volatile"-qualified types, using simple assignments where the left
2682           operand is a "volatile"-qualified non-class type for their value,
2683           compound assignments where the left operand is a
2684           "volatile"-qualified non-class type, "volatile"-qualified function
2685           return type, "volatile"-qualified parameter type, and structured
2686           bindings of a "volatile"-qualified type.  This usage was deprecated
2687           in C++20.
2688
2689           Enabled by default with -std=c++2a.
2690
2691       -Wzero-as-null-pointer-constant (C++ and Objective-C++ only)
2692           Warn when a literal 0 is used as null pointer constant.  This can
2693           be useful to facilitate the conversion to "nullptr" in C++11.
2694
2695       -Waligned-new
2696           Warn about a new-expression of a type that requires greater
2697           alignment than the "alignof(std::max_align_t)" but uses an
2698           allocation function without an explicit alignment parameter. This
2699           option is enabled by -Wall.
2700
2701           Normally this only warns about global allocation functions, but
2702           -Waligned-new=all also warns about class member allocation
2703           functions.
2704
2705       -Wno-placement-new
2706       -Wplacement-new=n
2707           Warn about placement new expressions with undefined behavior, such
2708           as constructing an object in a buffer that is smaller than the type
2709           of the object.  For example, the placement new expression below is
2710           diagnosed because it attempts to construct an array of 64 integers
2711           in a buffer only 64 bytes large.
2712
2713                   char buf [64];
2714                   new (buf) int[64];
2715
2716           This warning is enabled by default.
2717
2718           -Wplacement-new=1
2719               This is the default warning level of -Wplacement-new.  At this
2720               level the warning is not issued for some strictly undefined
2721               constructs that GCC allows as extensions for compatibility with
2722               legacy code.  For example, the following "new" expression is
2723               not diagnosed at this level even though it has undefined
2724               behavior according to the C++ standard because it writes past
2725               the end of the one-element array.
2726
2727                       struct S { int n, a[1]; };
2728                       S *s = (S *)malloc (sizeof *s + 31 * sizeof s->a[0]);
2729                       new (s->a)int [32]();
2730
2731           -Wplacement-new=2
2732               At this level, in addition to diagnosing all the same
2733               constructs as at level 1, a diagnostic is also issued for
2734               placement new expressions that construct an object in the last
2735               member of structure whose type is an array of a single element
2736               and whose size is less than the size of the object being
2737               constructed.  While the previous example would be diagnosed,
2738               the following construct makes use of the flexible member array
2739               extension to avoid the warning at level 2.
2740
2741                       struct S { int n, a[]; };
2742                       S *s = (S *)malloc (sizeof *s + 32 * sizeof s->a[0]);
2743                       new (s->a)int [32]();
2744
2745       -Wcatch-value
2746       -Wcatch-value=n (C++ and Objective-C++ only)
2747           Warn about catch handlers that do not catch via reference.  With
2748           -Wcatch-value=1 (or -Wcatch-value for short) warn about polymorphic
2749           class types that are caught by value.  With -Wcatch-value=2 warn
2750           about all class types that are caught by value. With
2751           -Wcatch-value=3 warn about all types that are not caught by
2752           reference. -Wcatch-value is enabled by -Wall.
2753
2754       -Wconditionally-supported (C++ and Objective-C++ only)
2755           Warn for conditionally-supported (C++11 [intro.defs]) constructs.
2756
2757       -Wno-delete-incomplete (C++ and Objective-C++ only)
2758           Do not warn when deleting a pointer to incomplete type, which may
2759           cause undefined behavior at runtime.  This warning is enabled by
2760           default.
2761
2762       -Wextra-semi (C++, Objective-C++ only)
2763           Warn about redundant semicolons after in-class function
2764           definitions.
2765
2766       -Wno-inaccessible-base (C++, Objective-C++ only)
2767           This option controls warnings when a base class is inaccessible in
2768           a class derived from it due to ambiguity.  The warning is enabled
2769           by default.  Note that the warning for ambiguous virtual bases is
2770           enabled by the -Wextra option.
2771
2772                   struct A { int a; };
2773
2774                   struct B : A { };
2775
2776                   struct C : B, A { };
2777
2778       -Wno-inherited-variadic-ctor
2779           Suppress warnings about use of C++11 inheriting constructors when
2780           the base class inherited from has a C variadic constructor; the
2781           warning is on by default because the ellipsis is not inherited.
2782
2783       -Wno-invalid-offsetof (C++ and Objective-C++ only)
2784           Suppress warnings from applying the "offsetof" macro to a non-POD
2785           type.  According to the 2014 ISO C++ standard, applying "offsetof"
2786           to a non-standard-layout type is undefined.  In existing C++
2787           implementations, however, "offsetof" typically gives meaningful
2788           results.  This flag is for users who are aware that they are
2789           writing nonportable code and who have deliberately chosen to ignore
2790           the warning about it.
2791
2792           The restrictions on "offsetof" may be relaxed in a future version
2793           of the C++ standard.
2794
2795       -Wsized-deallocation (C++ and Objective-C++ only)
2796           Warn about a definition of an unsized deallocation function
2797
2798                   void operator delete (void *) noexcept;
2799                   void operator delete[] (void *) noexcept;
2800
2801           without a definition of the corresponding sized deallocation
2802           function
2803
2804                   void operator delete (void *, std::size_t) noexcept;
2805                   void operator delete[] (void *, std::size_t) noexcept;
2806
2807           or vice versa.  Enabled by -Wextra along with -fsized-deallocation.
2808
2809       -Wsuggest-final-types
2810           Warn about types with virtual methods where code quality would be
2811           improved if the type were declared with the C++11 "final"
2812           specifier, or, if possible, declared in an anonymous namespace.
2813           This allows GCC to more aggressively devirtualize the polymorphic
2814           calls. This warning is more effective with link-time optimization,
2815           where the information about the class hierarchy graph is more
2816           complete.
2817
2818       -Wsuggest-final-methods
2819           Warn about virtual methods where code quality would be improved if
2820           the method were declared with the C++11 "final" specifier, or, if
2821           possible, its type were declared in an anonymous namespace or with
2822           the "final" specifier.  This warning is more effective with link-
2823           time optimization, where the information about the class hierarchy
2824           graph is more complete. It is recommended to first consider
2825           suggestions of -Wsuggest-final-types and then rebuild with new
2826           annotations.
2827
2828       -Wsuggest-override
2829           Warn about overriding virtual functions that are not marked with
2830           the "override" keyword.
2831
2832       -Wuseless-cast (C++ and Objective-C++ only)
2833           Warn when an expression is casted to its own type.
2834
2835       -Wno-conversion-null (C++ and Objective-C++ only)
2836           Do not warn for conversions between "NULL" and non-pointer types.
2837           -Wconversion-null is enabled by default.
2838
2839   Options Controlling Objective-C and Objective-C++ Dialects
2840       (NOTE: This manual does not describe the Objective-C and Objective-C++
2841       languages themselves.
2842
2843       This section describes the command-line options that are only
2844       meaningful for Objective-C and Objective-C++ programs.  You can also
2845       use most of the language-independent GNU compiler options.  For
2846       example, you might compile a file some_class.m like this:
2847
2848               gcc -g -fgnu-runtime -O -c some_class.m
2849
2850       In this example, -fgnu-runtime is an option meant only for Objective-C
2851       and Objective-C++ programs; you can use the other options with any
2852       language supported by GCC.
2853
2854       Note that since Objective-C is an extension of the C language,
2855       Objective-C compilations may also use options specific to the C front-
2856       end (e.g., -Wtraditional).  Similarly, Objective-C++ compilations may
2857       use C++-specific options (e.g., -Wabi).
2858
2859       Here is a list of options that are only for compiling Objective-C and
2860       Objective-C++ programs:
2861
2862       -fconstant-string-class=class-name
2863           Use class-name as the name of the class to instantiate for each
2864           literal string specified with the syntax "@"..."".  The default
2865           class name is "NXConstantString" if the GNU runtime is being used,
2866           and "NSConstantString" if the NeXT runtime is being used (see
2867           below).  The -fconstant-cfstrings option, if also present,
2868           overrides the -fconstant-string-class setting and cause "@"...""
2869           literals to be laid out as constant CoreFoundation strings.
2870
2871       -fgnu-runtime
2872           Generate object code compatible with the standard GNU Objective-C
2873           runtime.  This is the default for most types of systems.
2874
2875       -fnext-runtime
2876           Generate output compatible with the NeXT runtime.  This is the
2877           default for NeXT-based systems, including Darwin and Mac OS X.  The
2878           macro "__NEXT_RUNTIME__" is predefined if (and only if) this option
2879           is used.
2880
2881       -fno-nil-receivers
2882           Assume that all Objective-C message dispatches ("[receiver
2883           message:arg]") in this translation unit ensure that the receiver is
2884           not "nil".  This allows for more efficient entry points in the
2885           runtime to be used.  This option is only available in conjunction
2886           with the NeXT runtime and ABI version 0 or 1.
2887
2888       -fobjc-abi-version=n
2889           Use version n of the Objective-C ABI for the selected runtime.
2890           This option is currently supported only for the NeXT runtime.  In
2891           that case, Version 0 is the traditional (32-bit) ABI without
2892           support for properties and other Objective-C 2.0 additions.
2893           Version 1 is the traditional (32-bit) ABI with support for
2894           properties and other Objective-C 2.0 additions.  Version 2 is the
2895           modern (64-bit) ABI.  If nothing is specified, the default is
2896           Version 0 on 32-bit target machines, and Version 2 on 64-bit target
2897           machines.
2898
2899       -fobjc-call-cxx-cdtors
2900           For each Objective-C class, check if any of its instance variables
2901           is a C++ object with a non-trivial default constructor.  If so,
2902           synthesize a special "- (id) .cxx_construct" instance method which
2903           runs non-trivial default constructors on any such instance
2904           variables, in order, and then return "self".  Similarly, check if
2905           any instance variable is a C++ object with a non-trivial
2906           destructor, and if so, synthesize a special "- (void)
2907           .cxx_destruct" method which runs all such default destructors, in
2908           reverse order.
2909
2910           The "- (id) .cxx_construct" and "- (void) .cxx_destruct" methods
2911           thusly generated only operate on instance variables declared in the
2912           current Objective-C class, and not those inherited from
2913           superclasses.  It is the responsibility of the Objective-C runtime
2914           to invoke all such methods in an object's inheritance hierarchy.
2915           The "- (id) .cxx_construct" methods are invoked by the runtime
2916           immediately after a new object instance is allocated; the "- (void)
2917           .cxx_destruct" methods are invoked immediately before the runtime
2918           deallocates an object instance.
2919
2920           As of this writing, only the NeXT runtime on Mac OS X 10.4 and
2921           later has support for invoking the "- (id) .cxx_construct" and "-
2922           (void) .cxx_destruct" methods.
2923
2924       -fobjc-direct-dispatch
2925           Allow fast jumps to the message dispatcher.  On Darwin this is
2926           accomplished via the comm page.
2927
2928       -fobjc-exceptions
2929           Enable syntactic support for structured exception handling in
2930           Objective-C, similar to what is offered by C++.  This option is
2931           required to use the Objective-C keywords @try, @throw, @catch,
2932           @finally and @synchronized.  This option is available with both the
2933           GNU runtime and the NeXT runtime (but not available in conjunction
2934           with the NeXT runtime on Mac OS X 10.2 and earlier).
2935
2936       -fobjc-gc
2937           Enable garbage collection (GC) in Objective-C and Objective-C++
2938           programs.  This option is only available with the NeXT runtime; the
2939           GNU runtime has a different garbage collection implementation that
2940           does not require special compiler flags.
2941
2942       -fobjc-nilcheck
2943           For the NeXT runtime with version 2 of the ABI, check for a nil
2944           receiver in method invocations before doing the actual method call.
2945           This is the default and can be disabled using -fno-objc-nilcheck.
2946           Class methods and super calls are never checked for nil in this way
2947           no matter what this flag is set to.  Currently this flag does
2948           nothing when the GNU runtime, or an older version of the NeXT
2949           runtime ABI, is used.
2950
2951       -fobjc-std=objc1
2952           Conform to the language syntax of Objective-C 1.0, the language
2953           recognized by GCC 4.0.  This only affects the Objective-C additions
2954           to the C/C++ language; it does not affect conformance to C/C++
2955           standards, which is controlled by the separate C/C++ dialect option
2956           flags.  When this option is used with the Objective-C or
2957           Objective-C++ compiler, any Objective-C syntax that is not
2958           recognized by GCC 4.0 is rejected.  This is useful if you need to
2959           make sure that your Objective-C code can be compiled with older
2960           versions of GCC.
2961
2962       -freplace-objc-classes
2963           Emit a special marker instructing ld(1) not to statically link in
2964           the resulting object file, and allow dyld(1) to load it in at run
2965           time instead.  This is used in conjunction with the Fix-and-
2966           Continue debugging mode, where the object file in question may be
2967           recompiled and dynamically reloaded in the course of program
2968           execution, without the need to restart the program itself.
2969           Currently, Fix-and-Continue functionality is only available in
2970           conjunction with the NeXT runtime on Mac OS X 10.3 and later.
2971
2972       -fzero-link
2973           When compiling for the NeXT runtime, the compiler ordinarily
2974           replaces calls to "objc_getClass("...")" (when the name of the
2975           class is known at compile time) with static class references that
2976           get initialized at load time, which improves run-time performance.
2977           Specifying the -fzero-link flag suppresses this behavior and causes
2978           calls to "objc_getClass("...")"  to be retained.  This is useful in
2979           Zero-Link debugging mode, since it allows for individual class
2980           implementations to be modified during program execution.  The GNU
2981           runtime currently always retains calls to "objc_get_class("...")"
2982           regardless of command-line options.
2983
2984       -fno-local-ivars
2985           By default instance variables in Objective-C can be accessed as if
2986           they were local variables from within the methods of the class
2987           they're declared in.  This can lead to shadowing between instance
2988           variables and other variables declared either locally inside a
2989           class method or globally with the same name.  Specifying the
2990           -fno-local-ivars flag disables this behavior thus avoiding variable
2991           shadowing issues.
2992
2993       -fivar-visibility=[public|protected|private|package]
2994           Set the default instance variable visibility to the specified
2995           option so that instance variables declared outside the scope of any
2996           access modifier directives default to the specified visibility.
2997
2998       -gen-decls
2999           Dump interface declarations for all classes seen in the source file
3000           to a file named sourcename.decl.
3001
3002       -Wassign-intercept (Objective-C and Objective-C++ only)
3003           Warn whenever an Objective-C assignment is being intercepted by the
3004           garbage collector.
3005
3006       -Wno-property-assign-default (Objective-C and Objective-C++ only)
3007           Do not warn if a property for an Objective-C object has no assign
3008           semantics specified.
3009
3010       -Wno-protocol (Objective-C and Objective-C++ only)
3011           If a class is declared to implement a protocol, a warning is issued
3012           for every method in the protocol that is not implemented by the
3013           class.  The default behavior is to issue a warning for every method
3014           not explicitly implemented in the class, even if a method
3015           implementation is inherited from the superclass.  If you use the
3016           -Wno-protocol option, then methods inherited from the superclass
3017           are considered to be implemented, and no warning is issued for
3018           them.
3019
3020       -Wselector (Objective-C and Objective-C++ only)
3021           Warn if multiple methods of different types for the same selector
3022           are found during compilation.  The check is performed on the list
3023           of methods in the final stage of compilation.  Additionally, a
3024           check is performed for each selector appearing in a
3025           "@selector(...)"  expression, and a corresponding method for that
3026           selector has been found during compilation.  Because these checks
3027           scan the method table only at the end of compilation, these
3028           warnings are not produced if the final stage of compilation is not
3029           reached, for example because an error is found during compilation,
3030           or because the -fsyntax-only option is being used.
3031
3032       -Wstrict-selector-match (Objective-C and Objective-C++ only)
3033           Warn if multiple methods with differing argument and/or return
3034           types are found for a given selector when attempting to send a
3035           message using this selector to a receiver of type "id" or "Class".
3036           When this flag is off (which is the default behavior), the compiler
3037           omits such warnings if any differences found are confined to types
3038           that share the same size and alignment.
3039
3040       -Wundeclared-selector (Objective-C and Objective-C++ only)
3041           Warn if a "@selector(...)" expression referring to an undeclared
3042           selector is found.  A selector is considered undeclared if no
3043           method with that name has been declared before the "@selector(...)"
3044           expression, either explicitly in an @interface or @protocol
3045           declaration, or implicitly in an @implementation section.  This
3046           option always performs its checks as soon as a "@selector(...)"
3047           expression is found, while -Wselector only performs its checks in
3048           the final stage of compilation.  This also enforces the coding
3049           style convention that methods and selectors must be declared before
3050           being used.
3051
3052       -print-objc-runtime-info
3053           Generate C header describing the largest structure that is passed
3054           by value, if any.
3055
3056   Options to Control Diagnostic Messages Formatting
3057       Traditionally, diagnostic messages have been formatted irrespective of
3058       the output device's aspect (e.g. its width, ...).  You can use the
3059       options described below to control the formatting algorithm for
3060       diagnostic messages, e.g. how many characters per line, how often
3061       source location information should be reported.  Note that some
3062       language front ends may not honor these options.
3063
3064       -fmessage-length=n
3065           Try to format error messages so that they fit on lines of about n
3066           characters.  If n is zero, then no line-wrapping is done; each
3067           error message appears on a single line.  This is the default for
3068           all front ends.
3069
3070           Note - this option also affects the display of the #error and
3071           #warning pre-processor directives, and the deprecated
3072           function/type/variable attribute.  It does not however affect the
3073           pragma GCC warning and pragma GCC error pragmas.
3074
3075       -fdiagnostics-show-location=once
3076           Only meaningful in line-wrapping mode.  Instructs the diagnostic
3077           messages reporter to emit source location information once; that
3078           is, in case the message is too long to fit on a single physical
3079           line and has to be wrapped, the source location won't be emitted
3080           (as prefix) again, over and over, in subsequent continuation lines.
3081           This is the default behavior.
3082
3083       -fdiagnostics-show-location=every-line
3084           Only meaningful in line-wrapping mode.  Instructs the diagnostic
3085           messages reporter to emit the same source location information (as
3086           prefix) for physical lines that result from the process of breaking
3087           a message which is too long to fit on a single line.
3088
3089       -fdiagnostics-color[=WHEN]
3090       -fno-diagnostics-color
3091           Use color in diagnostics.  WHEN is never, always, or auto.  The
3092           default depends on how the compiler has been configured, it can be
3093           any of the above WHEN options or also never if GCC_COLORS
3094           environment variable isn't present in the environment, and auto
3095           otherwise.  auto makes GCC use color only when the standard error
3096           is a terminal, and when not executing in an emacs shell.  The forms
3097           -fdiagnostics-color and -fno-diagnostics-color are aliases for
3098           -fdiagnostics-color=always and -fdiagnostics-color=never,
3099           respectively.
3100
3101           The colors are defined by the environment variable GCC_COLORS.  Its
3102           value is a colon-separated list of capabilities and Select Graphic
3103           Rendition (SGR) substrings. SGR commands are interpreted by the
3104           terminal or terminal emulator.  (See the section in the
3105           documentation of your text terminal for permitted values and their
3106           meanings as character attributes.)  These substring values are
3107           integers in decimal representation and can be concatenated with
3108           semicolons.  Common values to concatenate include 1 for bold, 4 for
3109           underline, 5 for blink, 7 for inverse, 39 for default foreground
3110           color, 30 to 37 for foreground colors, 90 to 97 for 16-color mode
3111           foreground colors, 38;5;0 to 38;5;255 for 88-color and 256-color
3112           modes foreground colors, 49 for default background color, 40 to 47
3113           for background colors, 100 to 107 for 16-color mode background
3114           colors, and 48;5;0 to 48;5;255 for 88-color and 256-color modes
3115           background colors.
3116
3117           The default GCC_COLORS is
3118
3119                   error=01;31:warning=01;35:note=01;36:range1=32:range2=34:locus=01:\
3120                   quote=01:path=01;36:fixit-insert=32:fixit-delete=31:\
3121                   diff-filename=01:diff-hunk=32:diff-delete=31:diff-insert=32:\
3122                   type-diff=01;32
3123
3124           where 01;31 is bold red, 01;35 is bold magenta, 01;36 is bold cyan,
3125           32 is green, 34 is blue, 01 is bold, and 31 is red.  Setting
3126           GCC_COLORS to the empty string disables colors.  Supported
3127           capabilities are as follows.
3128
3129           "error="
3130               SGR substring for error: markers.
3131
3132           "warning="
3133               SGR substring for warning: markers.
3134
3135           "note="
3136               SGR substring for note: markers.
3137
3138           "path="
3139               SGR substring for colorizing paths of control-flow events as
3140               printed via -fdiagnostics-path-format=, such as the identifiers
3141               of individual events and lines indicating interprocedural calls
3142               and returns.
3143
3144           "range1="
3145               SGR substring for first additional range.
3146
3147           "range2="
3148               SGR substring for second additional range.
3149
3150           "locus="
3151               SGR substring for location information, file:line or
3152               file:line:column etc.
3153
3154           "quote="
3155               SGR substring for information printed within quotes.
3156
3157           "fixit-insert="
3158               SGR substring for fix-it hints suggesting text to be inserted
3159               or replaced.
3160
3161           "fixit-delete="
3162               SGR substring for fix-it hints suggesting text to be deleted.
3163
3164           "diff-filename="
3165               SGR substring for filename headers within generated patches.
3166
3167           "diff-hunk="
3168               SGR substring for the starts of hunks within generated patches.
3169
3170           "diff-delete="
3171               SGR substring for deleted lines within generated patches.
3172
3173           "diff-insert="
3174               SGR substring for inserted lines within generated patches.
3175
3176           "type-diff="
3177               SGR substring for highlighting mismatching types within
3178               template arguments in the C++ frontend.
3179
3180       -fdiagnostics-urls[=WHEN]
3181           Use escape sequences to embed URLs in diagnostics.  For example,
3182           when -fdiagnostics-show-option emits text showing the command-line
3183           option controlling a diagnostic, embed a URL for documentation of
3184           that option.
3185
3186           WHEN is never, always, or auto.  auto makes GCC use URL escape
3187           sequences only when the standard error is a terminal, and when not
3188           executing in an emacs shell or any graphical terminal which is
3189           known to be incompatible with this feature, see below.
3190
3191           The default depends on how the compiler has been configured.  It
3192           can be any of the above WHEN options.
3193
3194           GCC can also be configured (via the
3195           --with-diagnostics-urls=auto-if-env configure-time option) so that
3196           the default is affected by environment variables.  Under such a
3197           configuration, GCC defaults to using auto if either GCC_URLS or
3198           TERM_URLS environment variables are present and non-empty in the
3199           environment of the compiler, or never if neither are.
3200
3201           However, even with -fdiagnostics-urls=always the behavior is
3202           dependent on those environment variables: If GCC_URLS is set to
3203           empty or no, do not embed URLs in diagnostics.  If set to st, URLs
3204           use ST escape sequences.  If set to bel, the default, URLs use BEL
3205           escape sequences.  Any other non-empty value enables the feature.
3206           If GCC_URLS is not set, use TERM_URLS as a fallback.  Note: ST is
3207           an ANSI escape sequence, string terminator ESC \, BEL is an ASCII
3208           character, CTRL-G that usually sounds like a beep.
3209
3210           At this time GCC tries to detect also a few terminals that are
3211           known to not implement the URL feature, and have bugs or at least
3212           had bugs in some versions that are still in use, where the URL
3213           escapes are likely to misbehave, i.e. print garbage on the screen.
3214           That list is currently xfce4-terminal, certain known to be buggy
3215           gnome-terminal versions, the linux console, and mingw.  This check
3216           can be skipped with the -fdiagnostics-urls=always.
3217
3218       -fno-diagnostics-show-option
3219           By default, each diagnostic emitted includes text indicating the
3220           command-line option that directly controls the diagnostic (if such
3221           an option is known to the diagnostic machinery).  Specifying the
3222           -fno-diagnostics-show-option flag suppresses that behavior.
3223
3224       -fno-diagnostics-show-caret
3225           By default, each diagnostic emitted includes the original source
3226           line and a caret ^ indicating the column.  This option suppresses
3227           this information.  The source line is truncated to n characters, if
3228           the -fmessage-length=n option is given.  When the output is done to
3229           the terminal, the width is limited to the width given by the
3230           COLUMNS environment variable or, if not set, to the terminal width.
3231
3232       -fno-diagnostics-show-labels
3233           By default, when printing source code (via
3234           -fdiagnostics-show-caret), diagnostics can label ranges of source
3235           code with pertinent information, such as the types of expressions:
3236
3237                       printf ("foo %s bar", long_i + long_j);
3238                                    ~^       ~~~~~~~~~~~~~~~
3239                                     |              |
3240                                     char *         long int
3241
3242           This option suppresses the printing of these labels (in the example
3243           above, the vertical bars and the "char *" and "long int" text).
3244
3245       -fno-diagnostics-show-cwe
3246           Diagnostic messages can optionally have an associated
3247           @url{https://cwe.mitre.org/index.html, CWE} identifier.  GCC itself
3248           only provides such metadata for some of the -fanalyzer diagnostics.
3249           GCC plugins may also provide diagnostics with such metadata.  By
3250           default, if this information is present, it will be printed with
3251           the diagnostic.  This option suppresses the printing of this
3252           metadata.
3253
3254       -fno-diagnostics-show-line-numbers
3255           By default, when printing source code (via
3256           -fdiagnostics-show-caret), a left margin is printed, showing line
3257           numbers.  This option suppresses this left margin.
3258
3259       -fdiagnostics-minimum-margin-width=width
3260           This option controls the minimum width of the left margin printed
3261           by -fdiagnostics-show-line-numbers.  It defaults to 6.
3262
3263       -fdiagnostics-parseable-fixits
3264           Emit fix-it hints in a machine-parseable format, suitable for
3265           consumption by IDEs.  For each fix-it, a line will be printed after
3266           the relevant diagnostic, starting with the string "fix-it:".  For
3267           example:
3268
3269                   fix-it:"test.c":{45:3-45:21}:"gtk_widget_show_all"
3270
3271           The location is expressed as a half-open range, expressed as a
3272           count of bytes, starting at byte 1 for the initial column.  In the
3273           above example, bytes 3 through 20 of line 45 of "test.c" are to be
3274           replaced with the given string:
3275
3276                   00000000011111111112222222222
3277                   12345678901234567890123456789
3278                     gtk_widget_showall (dlg);
3279                     ^^^^^^^^^^^^^^^^^^
3280                     gtk_widget_show_all
3281
3282           The filename and replacement string escape backslash as "\\", tab
3283           as "\t", newline as "\n", double quotes as "\"", non-printable
3284           characters as octal (e.g. vertical tab as "\013").
3285
3286           An empty replacement string indicates that the given range is to be
3287           removed.  An empty range (e.g. "45:3-45:3") indicates that the
3288           string is to be inserted at the given position.
3289
3290       -fdiagnostics-generate-patch
3291           Print fix-it hints to stderr in unified diff format, after any
3292           diagnostics are printed.  For example:
3293
3294                   --- test.c
3295                   +++ test.c
3296                   @ -42,5 +42,5 @
3297
3298                    void show_cb(GtkDialog *dlg)
3299                    {
3300                   -  gtk_widget_showall(dlg);
3301                   +  gtk_widget_show_all(dlg);
3302                    }
3303
3304           The diff may or may not be colorized, following the same rules as
3305           for diagnostics (see -fdiagnostics-color).
3306
3307       -fdiagnostics-show-template-tree
3308           In the C++ frontend, when printing diagnostics showing mismatching
3309           template types, such as:
3310
3311                     could not convert 'std::map<int, std::vector<double> >()'
3312                       from 'map<[...],vector<double>>' to 'map<[...],vector<float>>
3313
3314           the -fdiagnostics-show-template-tree flag enables printing a tree-
3315           like structure showing the common and differing parts of the types,
3316           such as:
3317
3318                     map<
3319                       [...],
3320                       vector<
3321                         [double != float]>>
3322
3323           The parts that differ are highlighted with color ("double" and
3324           "float" in this case).
3325
3326       -fno-elide-type
3327           By default when the C++ frontend prints diagnostics showing
3328           mismatching template types, common parts of the types are printed
3329           as "[...]" to simplify the error message.  For example:
3330
3331                     could not convert 'std::map<int, std::vector<double> >()'
3332                       from 'map<[...],vector<double>>' to 'map<[...],vector<float>>
3333
3334           Specifying the -fno-elide-type flag suppresses that behavior.  This
3335           flag also affects the output of the
3336           -fdiagnostics-show-template-tree flag.
3337
3338       -fdiagnostics-path-format=KIND
3339           Specify how to print paths of control-flow events for diagnostics
3340           that have such a path associated with them.
3341
3342           KIND is none, separate-events, or inline-events, the default.
3343
3344           none means to not print diagnostic paths.
3345
3346           separate-events means to print a separate "note" diagnostic for
3347           each event within the diagnostic.  For example:
3348
3349                   test.c:29:5: error: passing NULL as argument 1 to 'PyList_Append' which requires a non-NULL parameter
3350                   test.c:25:10: note: (1) when 'PyList_New' fails, returning NULL
3351                   test.c:27:3: note: (2) when 'i < count'
3352                   test.c:29:5: note: (3) when calling 'PyList_Append', passing NULL from (1) as argument 1
3353
3354           inline-events means to print the events "inline" within the source
3355           code.  This view attempts to consolidate the events into runs of
3356           sufficiently-close events, printing them as labelled ranges within
3357           the source.
3358
3359           For example, the same events as above might be printed as:
3360
3361                     'test': events 1-3
3362                       |
3363                       |   25 |   list = PyList_New(0);
3364                       |      |          ^~~~~~~~~~~~~
3365                       |      |          |
3366                       |      |          (1) when 'PyList_New' fails, returning NULL
3367                       |   26 |
3368                       |   27 |   for (i = 0; i < count; i++) {
3369                       |      |   ~~~
3370                       |      |   |
3371                       |      |   (2) when 'i < count'
3372                       |   28 |     item = PyLong_FromLong(random());
3373                       |   29 |     PyList_Append(list, item);
3374                       |      |     ~~~~~~~~~~~~~~~~~~~~~~~~~
3375                       |      |     |
3376                       |      |     (3) when calling 'PyList_Append', passing NULL from (1) as argument 1
3377                       |
3378
3379           Interprocedural control flow is shown by grouping the events by
3380           stack frame, and using indentation to show how stack frames are
3381           nested, pushed, and popped.
3382
3383           For example:
3384
3385                     'test': events 1-2
3386                       |
3387                       |  133 | {
3388                       |      | ^
3389                       |      | |
3390                       |      | (1) entering 'test'
3391                       |  134 |   boxed_int *obj = make_boxed_int (i);
3392                       |      |                    ~~~~~~~~~~~~~~~~~~
3393                       |      |                    |
3394                       |      |                    (2) calling 'make_boxed_int'
3395                       |
3396                       +--> 'make_boxed_int': events 3-4
3397                              |
3398                              |  120 | {
3399                              |      | ^
3400                              |      | |
3401                              |      | (3) entering 'make_boxed_int'
3402                              |  121 |   boxed_int *result = (boxed_int *)wrapped_malloc (sizeof (boxed_int));
3403                              |      |                                    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
3404                              |      |                                    |
3405                              |      |                                    (4) calling 'wrapped_malloc'
3406                              |
3407                              +--> 'wrapped_malloc': events 5-6
3408                                     |
3409                                     |    7 | {
3410                                     |      | ^
3411                                     |      | |
3412                                     |      | (5) entering 'wrapped_malloc'
3413                                     |    8 |   return malloc (size);
3414                                     |      |          ~~~~~~~~~~~~~
3415                                     |      |          |
3416                                     |      |          (6) calling 'malloc'
3417                                     |
3418                       <-------------+
3419                       |
3420                    'test': event 7
3421                       |
3422                       |  138 |   free_boxed_int (obj);
3423                       |      |   ^~~~~~~~~~~~~~~~~~~~
3424                       |      |   |
3425                       |      |   (7) calling 'free_boxed_int'
3426                       |
3427                   (etc)
3428
3429       -fdiagnostics-show-path-depths
3430           This option provides additional information when printing control-
3431           flow paths associated with a diagnostic.
3432
3433           If this is option is provided then the stack depth will be printed
3434           for each run of events within
3435           -fdiagnostics-path-format=separate-events.
3436
3437           This is intended for use by GCC developers and plugin developers
3438           when debugging diagnostics that report interprocedural control
3439           flow.
3440
3441       -fno-show-column
3442           Do not print column numbers in diagnostics.  This may be necessary
3443           if diagnostics are being scanned by a program that does not
3444           understand the column numbers, such as dejagnu.
3445
3446       -fdiagnostics-format=FORMAT
3447           Select a different format for printing diagnostics.  FORMAT is text
3448           or json.  The default is text.
3449
3450           The json format consists of a top-level JSON array containing JSON
3451           objects representing the diagnostics.
3452
3453           The JSON is emitted as one line, without formatting; the examples
3454           below have been formatted for clarity.
3455
3456           Diagnostics can have child diagnostics.  For example, this error
3457           and note:
3458
3459                   misleading-indentation.c:15:3: warning: this 'if' clause does not
3460                     guard... [-Wmisleading-indentation]
3461                      15 |   if (flag)
3462                         |   ^~
3463                   misleading-indentation.c:17:5: note: ...this statement, but the latter
3464                     is misleadingly indented as if it were guarded by the 'if'
3465                      17 |     y = 2;
3466                         |     ^
3467
3468           might be printed in JSON form (after formatting) like this:
3469
3470                   [
3471                       {
3472                           "kind": "warning",
3473                           "locations": [
3474                               {
3475                                   "caret": {
3476                                       "column": 3,
3477                                       "file": "misleading-indentation.c",
3478                                       "line": 15
3479                                   },
3480                                   "finish": {
3481                                       "column": 4,
3482                                       "file": "misleading-indentation.c",
3483                                       "line": 15
3484                                   }
3485                               }
3486                           ],
3487                           "message": "this \u2018if\u2019 clause does not guard...",
3488                           "option": "-Wmisleading-indentation",
3489                           "option_url": "https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html#index-Wmisleading-indentation",
3490                           "children": [
3491                               {
3492                                   "kind": "note",
3493                                   "locations": [
3494                                       {
3495                                           "caret": {
3496                                               "column": 5,
3497                                               "file": "misleading-indentation.c",
3498                                               "line": 17
3499                                           }
3500                                       }
3501                                   ],
3502                                   "message": "...this statement, but the latter is ..."
3503                               }
3504                           ]
3505                       },
3506                       ...
3507                   ]
3508
3509           where the "note" is a child of the "warning".
3510
3511           A diagnostic has a "kind".  If this is "warning", then there is an
3512           "option" key describing the command-line option controlling the
3513           warning.
3514
3515           A diagnostic can contain zero or more locations.  Each location has
3516           up to three positions within it: a "caret" position and optional
3517           "start" and "finish" positions.  A location can also have an
3518           optional "label" string.  For example, this error:
3519
3520                   bad-binary-ops.c:64:23: error: invalid operands to binary + (have 'S' {aka
3521                      'struct s'} and 'T' {aka 'struct t'})
3522                      64 |   return callee_4a () + callee_4b ();
3523                         |          ~~~~~~~~~~~~ ^ ~~~~~~~~~~~~
3524                         |          |              |
3525                         |          |              T {aka struct t}
3526                         |          S {aka struct s}
3527
3528           has three locations.  Its primary location is at the "+" token at
3529           column 23.  It has two secondary locations, describing the left and
3530           right-hand sides of the expression, which have labels.  It might be
3531           printed in JSON form as:
3532
3533                       {
3534                           "children": [],
3535                           "kind": "error",
3536                           "locations": [
3537                               {
3538                                   "caret": {
3539                                       "column": 23, "file": "bad-binary-ops.c", "line": 64
3540                                   }
3541                               },
3542                               {
3543                                   "caret": {
3544                                       "column": 10, "file": "bad-binary-ops.c", "line": 64
3545                                   },
3546                                   "finish": {
3547                                       "column": 21, "file": "bad-binary-ops.c", "line": 64
3548                                   },
3549                                   "label": "S {aka struct s}"
3550                               },
3551                               {
3552                                   "caret": {
3553                                       "column": 25, "file": "bad-binary-ops.c", "line": 64
3554                                   },
3555                                   "finish": {
3556                                       "column": 36, "file": "bad-binary-ops.c", "line": 64
3557                                   },
3558                                   "label": "T {aka struct t}"
3559                               }
3560                           ],
3561                           "message": "invalid operands to binary + ..."
3562                       }
3563
3564           If a diagnostic contains fix-it hints, it has a "fixits" array,
3565           consisting of half-open intervals, similar to the output of
3566           -fdiagnostics-parseable-fixits.  For example, this diagnostic with
3567           a replacement fix-it hint:
3568
3569                   demo.c:8:15: error: 'struct s' has no member named 'colour'; did you
3570                     mean 'color'?
3571                       8 |   return ptr->colour;
3572                         |               ^~~~~~
3573                         |               color
3574
3575           might be printed in JSON form as:
3576
3577                       {
3578                           "children": [],
3579                           "fixits": [
3580                               {
3581                                   "next": {
3582                                       "column": 21,
3583                                       "file": "demo.c",
3584                                       "line": 8
3585                                   },
3586                                   "start": {
3587                                       "column": 15,
3588                                       "file": "demo.c",
3589                                       "line": 8
3590                                   },
3591                                   "string": "color"
3592                               }
3593                           ],
3594                           "kind": "error",
3595                           "locations": [
3596                               {
3597                                   "caret": {
3598                                       "column": 15,
3599                                       "file": "demo.c",
3600                                       "line": 8
3601                                   },
3602                                   "finish": {
3603                                       "column": 20,
3604                                       "file": "demo.c",
3605                                       "line": 8
3606                                   }
3607                               }
3608                           ],
3609                           "message": "\u2018struct s\u2019 has no member named ..."
3610                       }
3611
3612           where the fix-it hint suggests replacing the text from "start" up
3613           to but not including "next" with "string"'s value.  Deletions are
3614           expressed via an empty value for "string", insertions by having
3615           "start" equal "next".
3616
3617           If the diagnostic has a path of control-flow events associated with
3618           it, it has a "path" array of objects representing the events.  Each
3619           event object has a "description" string, a "location" object, along
3620           with a "function" string and a "depth" number for representing
3621           interprocedural paths.  The "function" represents the current
3622           function at that event, and the "depth" represents the stack depth
3623           relative to some baseline: the higher, the more frames are within
3624           the stack.
3625
3626           For example, the intraprocedural example shown for
3627           -fdiagnostics-path-format= might have this JSON for its path:
3628
3629                       "path": [
3630                           {
3631                               "depth": 0,
3632                               "description": "when 'PyList_New' fails, returning NULL",
3633                               "function": "test",
3634                               "location": {
3635                                   "column": 10,
3636                                   "file": "test.c",
3637                                   "line": 25
3638                               }
3639                           },
3640                           {
3641                               "depth": 0,
3642                               "description": "when 'i < count'",
3643                               "function": "test",
3644                               "location": {
3645                                   "column": 3,
3646                                   "file": "test.c",
3647                                   "line": 27
3648                               }
3649                           },
3650                           {
3651                               "depth": 0,
3652                               "description": "when calling 'PyList_Append', passing NULL from (1) as argument 1",
3653                               "function": "test",
3654                               "location": {
3655                                   "column": 5,
3656                                   "file": "test.c",
3657                                   "line": 29
3658                               }
3659                           }
3660                       ]
3661
3662   Options to Request or Suppress Warnings
3663       Warnings are diagnostic messages that report constructions that are not
3664       inherently erroneous but that are risky or suggest there may have been
3665       an error.
3666
3667       The following language-independent options do not enable specific
3668       warnings but control the kinds of diagnostics produced by GCC.
3669
3670       -fsyntax-only
3671           Check the code for syntax errors, but don't do anything beyond
3672           that.
3673
3674       -fmax-errors=n
3675           Limits the maximum number of error messages to n, at which point
3676           GCC bails out rather than attempting to continue processing the
3677           source code.  If n is 0 (the default), there is no limit on the
3678           number of error messages produced.  If -Wfatal-errors is also
3679           specified, then -Wfatal-errors takes precedence over this option.
3680
3681       -w  Inhibit all warning messages.
3682
3683       -Werror
3684           Make all warnings into errors.
3685
3686       -Werror=
3687           Make the specified warning into an error.  The specifier for a
3688           warning is appended; for example -Werror=switch turns the warnings
3689           controlled by -Wswitch into errors.  This switch takes a negative
3690           form, to be used to negate -Werror for specific warnings; for
3691           example -Wno-error=switch makes -Wswitch warnings not be errors,
3692           even when -Werror is in effect.
3693
3694           The warning message for each controllable warning includes the
3695           option that controls the warning.  That option can then be used
3696           with -Werror= and -Wno-error= as described above.  (Printing of the
3697           option in the warning message can be disabled using the
3698           -fno-diagnostics-show-option flag.)
3699
3700           Note that specifying -Werror=foo automatically implies -Wfoo.
3701           However, -Wno-error=foo does not imply anything.
3702
3703       -Wfatal-errors
3704           This option causes the compiler to abort compilation on the first
3705           error occurred rather than trying to keep going and printing
3706           further error messages.
3707
3708       You can request many specific warnings with options beginning with -W,
3709       for example -Wimplicit to request warnings on implicit declarations.
3710       Each of these specific warning options also has a negative form
3711       beginning -Wno- to turn off warnings; for example, -Wno-implicit.  This
3712       manual lists only one of the two forms, whichever is not the default.
3713       For further language-specific options also refer to C++ Dialect Options
3714       and Objective-C and Objective-C++ Dialect Options.  Additional warnings
3715       can be produced by enabling the static analyzer;
3716
3717       Some options, such as -Wall and -Wextra, turn on other options, such as
3718       -Wunused, which may turn on further options, such as -Wunused-value.
3719       The combined effect of positive and negative forms is that more
3720       specific options have priority over less specific ones, independently
3721       of their position in the command-line. For options of the same
3722       specificity, the last one takes effect. Options enabled or disabled via
3723       pragmas take effect as if they appeared at the end of the command-line.
3724
3725       When an unrecognized warning option is requested (e.g.,
3726       -Wunknown-warning), GCC emits a diagnostic stating that the option is
3727       not recognized.  However, if the -Wno- form is used, the behavior is
3728       slightly different: no diagnostic is produced for -Wno-unknown-warning
3729       unless other diagnostics are being produced.  This allows the use of
3730       new -Wno- options with old compilers, but if something goes wrong, the
3731       compiler warns that an unrecognized option is present.
3732
3733       The effectiveness of some warnings depends on optimizations also being
3734       enabled. For example -Wsuggest-final-types is more effective with link-
3735       time optimization and -Wmaybe-uninitialized does not warn at all unless
3736       optimization is enabled.
3737
3738       -Wpedantic
3739       -pedantic
3740           Issue all the warnings demanded by strict ISO C and ISO C++; reject
3741           all programs that use forbidden extensions, and some other programs
3742           that do not follow ISO C and ISO C++.  For ISO C, follows the
3743           version of the ISO C standard specified by any -std option used.
3744
3745           Valid ISO C and ISO C++ programs should compile properly with or
3746           without this option (though a rare few require -ansi or a -std
3747           option specifying the required version of ISO C).  However, without
3748           this option, certain GNU extensions and traditional C and C++
3749           features are supported as well.  With this option, they are
3750           rejected.
3751
3752           -Wpedantic does not cause warning messages for use of the alternate
3753           keywords whose names begin and end with __.  This alternate format
3754           can also be used to disable warnings for non-ISO __intN types, i.e.
3755           __intN__.  Pedantic warnings are also disabled in the expression
3756           that follows "__extension__".  However, only system header files
3757           should use these escape routes; application programs should avoid
3758           them.
3759
3760           Some users try to use -Wpedantic to check programs for strict ISO C
3761           conformance.  They soon find that it does not do quite what they
3762           want: it finds some non-ISO practices, but not all---only those for
3763           which ISO C requires a diagnostic, and some others for which
3764           diagnostics have been added.
3765
3766           A feature to report any failure to conform to ISO C might be useful
3767           in some instances, but would require considerable additional work
3768           and would be quite different from -Wpedantic.  We don't have plans
3769           to support such a feature in the near future.
3770
3771           Where the standard specified with -std represents a GNU extended
3772           dialect of C, such as gnu90 or gnu99, there is a corresponding base
3773           standard, the version of ISO C on which the GNU extended dialect is
3774           based.  Warnings from -Wpedantic are given where they are required
3775           by the base standard.  (It does not make sense for such warnings to
3776           be given only for features not in the specified GNU C dialect,
3777           since by definition the GNU dialects of C include all features the
3778           compiler supports with the given option, and there would be nothing
3779           to warn about.)
3780
3781       -pedantic-errors
3782           Give an error whenever the base standard (see -Wpedantic) requires
3783           a diagnostic, in some cases where there is undefined behavior at
3784           compile-time and in some other cases that do not prevent
3785           compilation of programs that are valid according to the standard.
3786           This is not equivalent to -Werror=pedantic, since there are errors
3787           enabled by this option and not enabled by the latter and vice
3788           versa.
3789
3790       -Wall
3791           This enables all the warnings about constructions that some users
3792           consider questionable, and that are easy to avoid (or modify to
3793           prevent the warning), even in conjunction with macros.  This also
3794           enables some language-specific warnings described in C++ Dialect
3795           Options and Objective-C and Objective-C++ Dialect Options.
3796
3797           -Wall turns on the following warning flags:
3798
3799           -Waddress -Warray-bounds=1 (only with -O2) -Wbool-compare
3800           -Wbool-operation -Wc++11-compat  -Wc++14-compat -Wcatch-value (C++
3801           and Objective-C++ only) -Wchar-subscripts -Wcomment
3802           -Wduplicate-decl-specifier (C and Objective-C only) -Wenum-compare
3803           (in C/ObjC; this is on by default in C++) -Wenum-conversion in
3804           C/ObjC; -Wformat -Wformat-overflow -Wformat-truncation
3805           -Wint-in-bool-context -Wimplicit (C and Objective-C only)
3806           -Wimplicit-int (C and Objective-C only)
3807           -Wimplicit-function-declaration (C and Objective-C only)
3808           -Winit-self (only for C++) -Wlogical-not-parentheses -Wmain (only
3809           for C/ObjC and unless -ffreestanding) -Wmaybe-uninitialized
3810           -Wmemset-elt-size -Wmemset-transposed-args -Wmisleading-indentation
3811           (only for C/C++) -Wmissing-attributes -Wmissing-braces (only for
3812           C/ObjC) -Wmultistatement-macros -Wnarrowing (only for C++)
3813           -Wnonnull -Wnonnull-compare -Wopenmp-simd -Wparentheses
3814           -Wpessimizing-move (only for C++) -Wpointer-sign -Wreorder
3815           -Wrestrict -Wreturn-type -Wsequence-point -Wsign-compare (only in
3816           C++) -Wsizeof-pointer-div -Wsizeof-pointer-memaccess
3817           -Wstrict-aliasing -Wstrict-overflow=1 -Wswitch
3818           -Wtautological-compare -Wtrigraphs -Wuninitialized
3819           -Wunknown-pragmas -Wunused-function -Wunused-label -Wunused-value
3820           -Wunused-variable -Wvolatile-register-var -Wzero-length-bounds
3821
3822           Note that some warning flags are not implied by -Wall.  Some of
3823           them warn about constructions that users generally do not consider
3824           questionable, but which occasionally you might wish to check for;
3825           others warn about constructions that are necessary or hard to avoid
3826           in some cases, and there is no simple way to modify the code to
3827           suppress the warning. Some of them are enabled by -Wextra but many
3828           of them must be enabled individually.
3829
3830       -Wextra
3831           This enables some extra warning flags that are not enabled by
3832           -Wall. (This option used to be called -W.  The older name is still
3833           supported, but the newer name is more descriptive.)
3834
3835           -Wclobbered -Wcast-function-type -Wdeprecated-copy (C++ only)
3836           -Wempty-body -Wignored-qualifiers -Wimplicit-fallthrough=3
3837           -Wmissing-field-initializers -Wmissing-parameter-type (C only)
3838           -Wold-style-declaration (C only) -Woverride-init -Wsign-compare (C
3839           only) -Wstring-compare -Wredundant-move (only for C++)
3840           -Wtype-limits -Wuninitialized -Wshift-negative-value (in C++03 and
3841           in C99 and newer) -Wunused-parameter (only with -Wunused or -Wall)
3842           -Wunused-but-set-parameter (only with -Wunused or -Wall)
3843
3844           The option -Wextra also prints warning messages for the following
3845           cases:
3846
3847           *   A pointer is compared against integer zero with "<", "<=", ">",
3848               or ">=".
3849
3850           *   (C++ only) An enumerator and a non-enumerator both appear in a
3851               conditional expression.
3852
3853           *   (C++ only) Ambiguous virtual bases.
3854
3855           *   (C++ only) Subscripting an array that has been declared
3856               "register".
3857
3858           *   (C++ only) Taking the address of a variable that has been
3859               declared "register".
3860
3861           *   (C++ only) A base class is not initialized in the copy
3862               constructor of a derived class.
3863
3864       -Wabi (C, Objective-C, C++ and Objective-C++ only)
3865           Warn about code affected by ABI changes.  This includes code that
3866           may not be compatible with the vendor-neutral C++ ABI as well as
3867           the psABI for the particular target.
3868
3869           Since G++ now defaults to updating the ABI with each major release,
3870           normally -Wabi warns only about C++ ABI compatibility problems if
3871           there is a check added later in a release series for an ABI issue
3872           discovered since the initial release.  -Wabi warns about more
3873           things if an older ABI version is selected (with -fabi-version=n).
3874
3875           -Wabi can also be used with an explicit version number to warn
3876           about C++ ABI compatibility with a particular -fabi-version level,
3877           e.g. -Wabi=2 to warn about changes relative to -fabi-version=2.
3878
3879           If an explicit version number is provided and -fabi-compat-version
3880           is not specified, the version number from this option is used for
3881           compatibility aliases.  If no explicit version number is provided
3882           with this option, but -fabi-compat-version is specified, that
3883           version number is used for C++ ABI warnings.
3884
3885           Although an effort has been made to warn about all such cases,
3886           there are probably some cases that are not warned about, even
3887           though G++ is generating incompatible code.  There may also be
3888           cases where warnings are emitted even though the code that is
3889           generated is compatible.
3890
3891           You should rewrite your code to avoid these warnings if you are
3892           concerned about the fact that code generated by G++ may not be
3893           binary compatible with code generated by other compilers.
3894
3895           Known incompatibilities in -fabi-version=2 (which was the default
3896           from GCC 3.4 to 4.9) include:
3897
3898           *   A template with a non-type template parameter of reference type
3899               was mangled incorrectly:
3900
3901                       extern int N;
3902                       template <int &> struct S {};
3903                       void n (S<N>) {2}
3904
3905               This was fixed in -fabi-version=3.
3906
3907           *   SIMD vector types declared using "__attribute ((vector_size))"
3908               were mangled in a non-standard way that does not allow for
3909               overloading of functions taking vectors of different sizes.
3910
3911               The mangling was changed in -fabi-version=4.
3912
3913           *   "__attribute ((const))" and "noreturn" were mangled as type
3914               qualifiers, and "decltype" of a plain declaration was folded
3915               away.
3916
3917               These mangling issues were fixed in -fabi-version=5.
3918
3919           *   Scoped enumerators passed as arguments to a variadic function
3920               are promoted like unscoped enumerators, causing "va_arg" to
3921               complain.  On most targets this does not actually affect the
3922               parameter passing ABI, as there is no way to pass an argument
3923               smaller than "int".
3924
3925               Also, the ABI changed the mangling of template argument packs,
3926               "const_cast", "static_cast", prefix increment/decrement, and a
3927               class scope function used as a template argument.
3928
3929               These issues were corrected in -fabi-version=6.
3930
3931           *   Lambdas in default argument scope were mangled incorrectly, and
3932               the ABI changed the mangling of "nullptr_t".
3933
3934               These issues were corrected in -fabi-version=7.
3935
3936           *   When mangling a function type with function-cv-qualifiers, the
3937               un-qualified function type was incorrectly treated as a
3938               substitution candidate.
3939
3940               This was fixed in -fabi-version=8, the default for GCC 5.1.
3941
3942           *   "decltype(nullptr)" incorrectly had an alignment of 1, leading
3943               to unaligned accesses.  Note that this did not affect the ABI
3944               of a function with a "nullptr_t" parameter, as parameters have
3945               a minimum alignment.
3946
3947               This was fixed in -fabi-version=9, the default for GCC 5.2.
3948
3949           *   Target-specific attributes that affect the identity of a type,
3950               such as ia32 calling conventions on a function type (stdcall,
3951               regparm, etc.), did not affect the mangled name, leading to
3952               name collisions when function pointers were used as template
3953               arguments.
3954
3955               This was fixed in -fabi-version=10, the default for GCC 6.1.
3956
3957           This option also enables warnings about psABI-related changes.  The
3958           known psABI changes at this point include:
3959
3960           *   For SysV/x86-64, unions with "long double" members are passed
3961               in memory as specified in psABI.  Prior to GCC 4.4, this was
3962               not the case.  For example:
3963
3964                       union U {
3965                         long double ld;
3966                         int i;
3967                       };
3968
3969               "union U" is now always passed in memory.
3970
3971       -Wchar-subscripts
3972           Warn if an array subscript has type "char".  This is a common cause
3973           of error, as programmers often forget that this type is signed on
3974           some machines.  This warning is enabled by -Wall.
3975
3976       -Wno-coverage-mismatch
3977           Warn if feedback profiles do not match when using the -fprofile-use
3978           option.  If a source file is changed between compiling with
3979           -fprofile-generate and with -fprofile-use, the files with the
3980           profile feedback can fail to match the source file and GCC cannot
3981           use the profile feedback information.  By default, this warning is
3982           enabled and is treated as an error.  -Wno-coverage-mismatch can be
3983           used to disable the warning or -Wno-error=coverage-mismatch can be
3984           used to disable the error.  Disabling the error for this warning
3985           can result in poorly optimized code and is useful only in the case
3986           of very minor changes such as bug fixes to an existing code-base.
3987           Completely disabling the warning is not recommended.
3988
3989       -Wno-cpp
3990           (C, Objective-C, C++, Objective-C++ and Fortran only) Suppress
3991           warning messages emitted by "#warning" directives.
3992
3993       -Wdouble-promotion (C, C++, Objective-C and Objective-C++ only)
3994           Give a warning when a value of type "float" is implicitly promoted
3995           to "double".  CPUs with a 32-bit "single-precision" floating-point
3996           unit implement "float" in hardware, but emulate "double" in
3997           software.  On such a machine, doing computations using "double"
3998           values is much more expensive because of the overhead required for
3999           software emulation.
4000
4001           It is easy to accidentally do computations with "double" because
4002           floating-point literals are implicitly of type "double".  For
4003           example, in:
4004
4005                   float area(float radius)
4006                   {
4007                      return 3.14159 * radius * radius;
4008                   }
4009
4010           the compiler performs the entire computation with "double" because
4011           the floating-point literal is a "double".
4012
4013       -Wduplicate-decl-specifier (C and Objective-C only)
4014           Warn if a declaration has duplicate "const", "volatile", "restrict"
4015           or "_Atomic" specifier.  This warning is enabled by -Wall.
4016
4017       -Wformat
4018       -Wformat=n
4019           Check calls to "printf" and "scanf", etc., to make sure that the
4020           arguments supplied have types appropriate to the format string
4021           specified, and that the conversions specified in the format string
4022           make sense.  This includes standard functions, and others specified
4023           by format attributes, in the "printf", "scanf", "strftime" and
4024           "strfmon" (an X/Open extension, not in the C standard) families (or
4025           other target-specific families).  Which functions are checked
4026           without format attributes having been specified depends on the
4027           standard version selected, and such checks of functions without the
4028           attribute specified are disabled by -ffreestanding or -fno-builtin.
4029
4030           The formats are checked against the format features supported by
4031           GNU libc version 2.2.  These include all ISO C90 and C99 features,
4032           as well as features from the Single Unix Specification and some BSD
4033           and GNU extensions.  Other library implementations may not support
4034           all these features; GCC does not support warning about features
4035           that go beyond a particular library's limitations.  However, if
4036           -Wpedantic is used with -Wformat, warnings are given about format
4037           features not in the selected standard version (but not for
4038           "strfmon" formats, since those are not in any version of the C
4039           standard).
4040
4041           -Wformat=1
4042           -Wformat
4043               Option -Wformat is equivalent to -Wformat=1, and -Wno-format is
4044               equivalent to -Wformat=0.  Since -Wformat also checks for null
4045               format arguments for several functions, -Wformat also implies
4046               -Wnonnull.  Some aspects of this level of format checking can
4047               be disabled by the options: -Wno-format-contains-nul,
4048               -Wno-format-extra-args, and -Wno-format-zero-length.  -Wformat
4049               is enabled by -Wall.
4050
4051           -Wformat=2
4052               Enable -Wformat plus additional format checks.  Currently
4053               equivalent to -Wformat -Wformat-nonliteral -Wformat-security
4054               -Wformat-y2k.
4055
4056       -Wno-format-contains-nul
4057           If -Wformat is specified, do not warn about format strings that
4058           contain NUL bytes.
4059
4060       -Wno-format-extra-args
4061           If -Wformat is specified, do not warn about excess arguments to a
4062           "printf" or "scanf" format function.  The C standard specifies that
4063           such arguments are ignored.
4064
4065           Where the unused arguments lie between used arguments that are
4066           specified with $ operand number specifications, normally warnings
4067           are still given, since the implementation could not know what type
4068           to pass to "va_arg" to skip the unused arguments.  However, in the
4069           case of "scanf" formats, this option suppresses the warning if the
4070           unused arguments are all pointers, since the Single Unix
4071           Specification says that such unused arguments are allowed.
4072
4073       -Wformat-overflow
4074       -Wformat-overflow=level
4075           Warn about calls to formatted input/output functions such as
4076           "sprintf" and "vsprintf" that might overflow the destination
4077           buffer.  When the exact number of bytes written by a format
4078           directive cannot be determined at compile-time it is estimated
4079           based on heuristics that depend on the level argument and on
4080           optimization.  While enabling optimization will in most cases
4081           improve the accuracy of the warning, it may also result in false
4082           positives.
4083
4084           -Wformat-overflow
4085           -Wformat-overflow=1
4086               Level 1 of -Wformat-overflow enabled by -Wformat employs a
4087               conservative approach that warns only about calls that most
4088               likely overflow the buffer.  At this level, numeric arguments
4089               to format directives with unknown values are assumed to have
4090               the value of one, and strings of unknown length to be empty.
4091               Numeric arguments that are known to be bounded to a subrange of
4092               their type, or string arguments whose output is bounded either
4093               by their directive's precision or by a finite set of string
4094               literals, are assumed to take on the value within the range
4095               that results in the most bytes on output.  For example, the
4096               call to "sprintf" below is diagnosed because even with both a
4097               and b equal to zero, the terminating NUL character ('\0')
4098               appended by the function to the destination buffer will be
4099               written past its end.  Increasing the size of the buffer by a
4100               single byte is sufficient to avoid the warning, though it may
4101               not be sufficient to avoid the overflow.
4102
4103                       void f (int a, int b)
4104                       {
4105                         char buf [13];
4106                         sprintf (buf, "a = %i, b = %i\n", a, b);
4107                       }
4108
4109           -Wformat-overflow=2
4110               Level 2 warns also about calls that might overflow the
4111               destination buffer given an argument of sufficient length or
4112               magnitude.  At level 2, unknown numeric arguments are assumed
4113               to have the minimum representable value for signed types with a
4114               precision greater than 1, and the maximum representable value
4115               otherwise.  Unknown string arguments whose length cannot be
4116               assumed to be bounded either by the directive's precision, or
4117               by a finite set of string literals they may evaluate to, or the
4118               character array they may point to, are assumed to be 1
4119               character long.
4120
4121               At level 2, the call in the example above is again diagnosed,
4122               but this time because with a equal to a 32-bit "INT_MIN" the
4123               first %i directive will write some of its digits beyond the end
4124               of the destination buffer.  To make the call safe regardless of
4125               the values of the two variables, the size of the destination
4126               buffer must be increased to at least 34 bytes.  GCC includes
4127               the minimum size of the buffer in an informational note
4128               following the warning.
4129
4130               An alternative to increasing the size of the destination buffer
4131               is to constrain the range of formatted values.  The maximum
4132               length of string arguments can be bounded by specifying the
4133               precision in the format directive.  When numeric arguments of
4134               format directives can be assumed to be bounded by less than the
4135               precision of their type, choosing an appropriate length
4136               modifier to the format specifier will reduce the required
4137               buffer size.  For example, if a and b in the example above can
4138               be assumed to be within the precision of the "short int" type
4139               then using either the %hi format directive or casting the
4140               argument to "short" reduces the maximum required size of the
4141               buffer to 24 bytes.
4142
4143                       void f (int a, int b)
4144                       {
4145                         char buf [23];
4146                         sprintf (buf, "a = %hi, b = %i\n", a, (short)b);
4147                       }
4148
4149       -Wno-format-zero-length
4150           If -Wformat is specified, do not warn about zero-length formats.
4151           The C standard specifies that zero-length formats are allowed.
4152
4153       -Wformat-nonliteral
4154           If -Wformat is specified, also warn if the format string is not a
4155           string literal and so cannot be checked, unless the format function
4156           takes its format arguments as a "va_list".
4157
4158       -Wformat-security
4159           If -Wformat is specified, also warn about uses of format functions
4160           that represent possible security problems.  At present, this warns
4161           about calls to "printf" and "scanf" functions where the format
4162           string is not a string literal and there are no format arguments,
4163           as in "printf (foo);".  This may be a security hole if the format
4164           string came from untrusted input and contains %n.  (This is
4165           currently a subset of what -Wformat-nonliteral warns about, but in
4166           future warnings may be added to -Wformat-security that are not
4167           included in -Wformat-nonliteral.)
4168
4169       -Wformat-signedness
4170           If -Wformat is specified, also warn if the format string requires
4171           an unsigned argument and the argument is signed and vice versa.
4172
4173       -Wformat-truncation
4174       -Wformat-truncation=level
4175           Warn about calls to formatted input/output functions such as
4176           "snprintf" and "vsnprintf" that might result in output truncation.
4177           When the exact number of bytes written by a format directive cannot
4178           be determined at compile-time it is estimated based on heuristics
4179           that depend on the level argument and on optimization.  While
4180           enabling optimization will in most cases improve the accuracy of
4181           the warning, it may also result in false positives.  Except as
4182           noted otherwise, the option uses the same logic -Wformat-overflow.
4183
4184           -Wformat-truncation
4185           -Wformat-truncation=1
4186               Level 1 of -Wformat-truncation enabled by -Wformat employs a
4187               conservative approach that warns only about calls to bounded
4188               functions whose return value is unused and that will most
4189               likely result in output truncation.
4190
4191           -Wformat-truncation=2
4192               Level 2 warns also about calls to bounded functions whose
4193               return value is used and that might result in truncation given
4194               an argument of sufficient length or magnitude.
4195
4196       -Wformat-y2k
4197           If -Wformat is specified, also warn about "strftime" formats that
4198           may yield only a two-digit year.
4199
4200       -Wnonnull
4201           Warn about passing a null pointer for arguments marked as requiring
4202           a non-null value by the "nonnull" function attribute.
4203
4204           -Wnonnull is included in -Wall and -Wformat.  It can be disabled
4205           with the -Wno-nonnull option.
4206
4207       -Wnonnull-compare
4208           Warn when comparing an argument marked with the "nonnull" function
4209           attribute against null inside the function.
4210
4211           -Wnonnull-compare is included in -Wall.  It can be disabled with
4212           the -Wno-nonnull-compare option.
4213
4214       -Wnull-dereference
4215           Warn if the compiler detects paths that trigger erroneous or
4216           undefined behavior due to dereferencing a null pointer.  This
4217           option is only active when -fdelete-null-pointer-checks is active,
4218           which is enabled by optimizations in most targets.  The precision
4219           of the warnings depends on the optimization options used.
4220
4221       -Winit-self (C, C++, Objective-C and Objective-C++ only)
4222           Warn about uninitialized variables that are initialized with
4223           themselves.  Note this option can only be used with the
4224           -Wuninitialized option.
4225
4226           For example, GCC warns about "i" being uninitialized in the
4227           following snippet only when -Winit-self has been specified:
4228
4229                   int f()
4230                   {
4231                     int i = i;
4232                     return i;
4233                   }
4234
4235           This warning is enabled by -Wall in C++.
4236
4237       -Wno-implicit-int (C and Objective-C only)
4238           This option controls warnings when a declaration does not specify a
4239           type.  This warning is enabled by default in C99 and later dialects
4240           of C, and also by -Wall.
4241
4242       -Wno-implicit-function-declaration (C and Objective-C only)
4243           This option controls warnings when a function is used before being
4244           declared.  This warning is enabled by default in C99 and later
4245           dialects of C, and also by -Wall.  The warning is made into an
4246           error by -pedantic-errors.
4247
4248       -Wimplicit (C and Objective-C only)
4249           Same as -Wimplicit-int and -Wimplicit-function-declaration.  This
4250           warning is enabled by -Wall.
4251
4252       -Wimplicit-fallthrough
4253           -Wimplicit-fallthrough is the same as -Wimplicit-fallthrough=3 and
4254           -Wno-implicit-fallthrough is the same as -Wimplicit-fallthrough=0.
4255
4256       -Wimplicit-fallthrough=n
4257           Warn when a switch case falls through.  For example:
4258
4259                   switch (cond)
4260                     {
4261                     case 1:
4262                       a = 1;
4263                       break;
4264                     case 2:
4265                       a = 2;
4266                     case 3:
4267                       a = 3;
4268                       break;
4269                     }
4270
4271           This warning does not warn when the last statement of a case cannot
4272           fall through, e.g. when there is a return statement or a call to
4273           function declared with the noreturn attribute.
4274           -Wimplicit-fallthrough= also takes into account control flow
4275           statements, such as ifs, and only warns when appropriate.  E.g.
4276
4277                   switch (cond)
4278                     {
4279                     case 1:
4280                       if (i > 3) {
4281                         bar (5);
4282                         break;
4283                       } else if (i < 1) {
4284                         bar (0);
4285                       } else
4286                         return;
4287                     default:
4288                       ...
4289                     }
4290
4291           Since there are occasions where a switch case fall through is
4292           desirable, GCC provides an attribute, "__attribute__
4293           ((fallthrough))", that is to be used along with a null statement to
4294           suppress this warning that would normally occur:
4295
4296                   switch (cond)
4297                     {
4298                     case 1:
4299                       bar (0);
4300                       __attribute__ ((fallthrough));
4301                     default:
4302                       ...
4303                     }
4304
4305           C++17 provides a standard way to suppress the
4306           -Wimplicit-fallthrough warning using "[[fallthrough]];" instead of
4307           the GNU attribute.  In C++11 or C++14 users can use
4308           "[[gnu::fallthrough]];", which is a GNU extension.  Instead of
4309           these attributes, it is also possible to add a fallthrough comment
4310           to silence the warning.  The whole body of the C or C++ style
4311           comment should match the given regular expressions listed below.
4312           The option argument n specifies what kind of comments are accepted:
4313
4314           *<-Wimplicit-fallthrough=0 disables the warning altogether.>
4315           *<-Wimplicit-fallthrough=1 matches ".*" regular>
4316               expression, any comment is used as fallthrough comment.
4317
4318           *<-Wimplicit-fallthrough=2 case insensitively matches>
4319               ".*falls?[ \t-]*thr(ough|u).*" regular expression.
4320
4321           *<-Wimplicit-fallthrough=3 case sensitively matches one of the>
4322               following regular expressions:
4323
4324               *<"-fallthrough">
4325               *<"@fallthrough@">
4326               *<"lint -fallthrough[ \t]*">
4327               *<"[ \t.!]*(ELSE,? |INTENTIONAL(LY)? )?FALL(S |
4328               |-)?THR(OUGH|U)[ \t.!]*(-[^\n\r]*)?">
4329               *<"[ \t.!]*(Else,? |Intentional(ly)? )?Fall((s |
4330               |-)[Tt]|t)hr(ough|u)[ \t.!]*(-[^\n\r]*)?">
4331               *<"[ \t.!]*([Ee]lse,? |[Ii]ntentional(ly)? )?fall(s |
4332               |-)?thr(ough|u)[ \t.!]*(-[^\n\r]*)?">
4333           *<-Wimplicit-fallthrough=4 case sensitively matches one of the>
4334               following regular expressions:
4335
4336               *<"-fallthrough">
4337               *<"@fallthrough@">
4338               *<"lint -fallthrough[ \t]*">
4339               *<"[ \t]*FALLTHR(OUGH|U)[ \t]*">
4340           *<-Wimplicit-fallthrough=5 doesn't recognize any comments as>
4341               fallthrough comments, only attributes disable the warning.
4342
4343           The comment needs to be followed after optional whitespace and
4344           other comments by "case" or "default" keywords or by a user label
4345           that precedes some "case" or "default" label.
4346
4347                   switch (cond)
4348                     {
4349                     case 1:
4350                       bar (0);
4351                       /* FALLTHRU */
4352                     default:
4353                       ...
4354                     }
4355
4356           The -Wimplicit-fallthrough=3 warning is enabled by -Wextra.
4357
4358       -Wno-if-not-aligned (C, C++, Objective-C and Objective-C++ only)
4359           Control if warnings triggered by the "warn_if_not_aligned"
4360           attribute should be issued.  These warnings are enabled by default.
4361
4362       -Wignored-qualifiers (C and C++ only)
4363           Warn if the return type of a function has a type qualifier such as
4364           "const".  For ISO C such a type qualifier has no effect, since the
4365           value returned by a function is not an lvalue.  For C++, the
4366           warning is only emitted for scalar types or "void".  ISO C
4367           prohibits qualified "void" return types on function definitions, so
4368           such return types always receive a warning even without this
4369           option.
4370
4371           This warning is also enabled by -Wextra.
4372
4373       -Wno-ignored-attributes (C and C++ only)
4374           This option controls warnings when an attribute is ignored.  This
4375           is different from the -Wattributes option in that it warns whenever
4376           the compiler decides to drop an attribute, not that the attribute
4377           is either unknown, used in a wrong place, etc.  This warning is
4378           enabled by default.
4379
4380       -Wmain
4381           Warn if the type of "main" is suspicious.  "main" should be a
4382           function with external linkage, returning int, taking either zero
4383           arguments, two, or three arguments of appropriate types.  This
4384           warning is enabled by default in C++ and is enabled by either -Wall
4385           or -Wpedantic.
4386
4387       -Wmisleading-indentation (C and C++ only)
4388           Warn when the indentation of the code does not reflect the block
4389           structure.  Specifically, a warning is issued for "if", "else",
4390           "while", and "for" clauses with a guarded statement that does not
4391           use braces, followed by an unguarded statement with the same
4392           indentation.
4393
4394           In the following example, the call to "bar" is misleadingly
4395           indented as if it were guarded by the "if" conditional.
4396
4397                     if (some_condition ())
4398                       foo ();
4399                       bar ();  /* Gotcha: this is not guarded by the "if".  */
4400
4401           In the case of mixed tabs and spaces, the warning uses the
4402           -ftabstop= option to determine if the statements line up
4403           (defaulting to 8).
4404
4405           The warning is not issued for code involving multiline preprocessor
4406           logic such as the following example.
4407
4408                     if (flagA)
4409                       foo (0);
4410                   #if SOME_CONDITION_THAT_DOES_NOT_HOLD
4411                     if (flagB)
4412                   #endif
4413                       foo (1);
4414
4415           The warning is not issued after a "#line" directive, since this
4416           typically indicates autogenerated code, and no assumptions can be
4417           made about the layout of the file that the directive references.
4418
4419           This warning is enabled by -Wall in C and C++.
4420
4421       -Wmissing-attributes
4422           Warn when a declaration of a function is missing one or more
4423           attributes that a related function is declared with and whose
4424           absence may adversely affect the correctness or efficiency of
4425           generated code.  For example, the warning is issued for
4426           declarations of aliases that use attributes to specify less
4427           restrictive requirements than those of their targets.  This
4428           typically represents a potential optimization opportunity.  By
4429           contrast, the -Wattribute-alias=2 option controls warnings issued
4430           when the alias is more restrictive than the target, which could
4431           lead to incorrect code generation.  Attributes considered include
4432           "alloc_align", "alloc_size", "cold", "const", "hot", "leaf",
4433           "malloc", "nonnull", "noreturn", "nothrow", "pure",
4434           "returns_nonnull", and "returns_twice".
4435
4436           In C++, the warning is issued when an explicit specialization of a
4437           primary template declared with attribute "alloc_align",
4438           "alloc_size", "assume_aligned", "format", "format_arg", "malloc",
4439           or "nonnull" is declared without it.  Attributes "deprecated",
4440           "error", and "warning" suppress the warning..
4441
4442           You can use the "copy" attribute to apply the same set of
4443           attributes to a declaration as that on another declaration without
4444           explicitly enumerating the attributes. This attribute can be
4445           applied to declarations of functions, variables, or types.
4446
4447           -Wmissing-attributes is enabled by -Wall.
4448
4449           For example, since the declaration of the primary function template
4450           below makes use of both attribute "malloc" and "alloc_size" the
4451           declaration of the explicit specialization of the template is
4452           diagnosed because it is missing one of the attributes.
4453
4454                   template <class T>
4455                   T* __attribute__ ((malloc, alloc_size (1)))
4456                   allocate (size_t);
4457
4458                   template <>
4459                   void* __attribute__ ((malloc))   // missing alloc_size
4460                   allocate<void> (size_t);
4461
4462       -Wmissing-braces
4463           Warn if an aggregate or union initializer is not fully bracketed.
4464           In the following example, the initializer for "a" is not fully
4465           bracketed, but that for "b" is fully bracketed.
4466
4467                   int a[2][2] = { 0, 1, 2, 3 };
4468                   int b[2][2] = { { 0, 1 }, { 2, 3 } };
4469
4470           This warning is enabled by -Wall.
4471
4472       -Wmissing-include-dirs (C, C++, Objective-C and Objective-C++ only)
4473           Warn if a user-supplied include directory does not exist.
4474
4475       -Wno-missing-profile
4476           This option controls warnings if feedback profiles are missing when
4477           using the -fprofile-use option.  This option diagnoses those cases
4478           where a new function or a new file is added between compiling with
4479           -fprofile-generate and with -fprofile-use, without regenerating the
4480           profiles.  In these cases, the profile feedback data files do not
4481           contain any profile feedback information for the newly added
4482           function or file respectively.  Also, in the case when profile
4483           count data (.gcda) files are removed, GCC cannot use any profile
4484           feedback information.  In all these cases, warnings are issued to
4485           inform you that a profile generation step is due.  Ignoring the
4486           warning can result in poorly optimized code.  -Wno-missing-profile
4487           can be used to disable the warning, but this is not recommended and
4488           should be done only when non-existent profile data is justified.
4489
4490       -Wmultistatement-macros
4491           Warn about unsafe multiple statement macros that appear to be
4492           guarded by a clause such as "if", "else", "for", "switch", or
4493           "while", in which only the first statement is actually guarded
4494           after the macro is expanded.
4495
4496           For example:
4497
4498                   #define DOIT x++; y++
4499                   if (c)
4500                     DOIT;
4501
4502           will increment "y" unconditionally, not just when "c" holds.  The
4503           can usually be fixed by wrapping the macro in a do-while loop:
4504
4505                   #define DOIT do { x++; y++; } while (0)
4506                   if (c)
4507                     DOIT;
4508
4509           This warning is enabled by -Wall in C and C++.
4510
4511       -Wparentheses
4512           Warn if parentheses are omitted in certain contexts, such as when
4513           there is an assignment in a context where a truth value is
4514           expected, or when operators are nested whose precedence people
4515           often get confused about.
4516
4517           Also warn if a comparison like "x<=y<=z" appears; this is
4518           equivalent to "(x<=y ? 1 : 0) <= z", which is a different
4519           interpretation from that of ordinary mathematical notation.
4520
4521           Also warn for dangerous uses of the GNU extension to "?:" with
4522           omitted middle operand. When the condition in the "?": operator is
4523           a boolean expression, the omitted value is always 1.  Often
4524           programmers expect it to be a value computed inside the conditional
4525           expression instead.
4526
4527           For C++ this also warns for some cases of unnecessary parentheses
4528           in declarations, which can indicate an attempt at a function call
4529           instead of a declaration:
4530
4531                   {
4532                     // Declares a local variable called mymutex.
4533                     std::unique_lock<std::mutex> (mymutex);
4534                     // User meant std::unique_lock<std::mutex> lock (mymutex);
4535                   }
4536
4537           This warning is enabled by -Wall.
4538
4539       -Wsequence-point
4540           Warn about code that may have undefined semantics because of
4541           violations of sequence point rules in the C and C++ standards.
4542
4543           The C and C++ standards define the order in which expressions in a
4544           C/C++ program are evaluated in terms of sequence points, which
4545           represent a partial ordering between the execution of parts of the
4546           program: those executed before the sequence point, and those
4547           executed after it.  These occur after the evaluation of a full
4548           expression (one which is not part of a larger expression), after
4549           the evaluation of the first operand of a "&&", "||", "? :" or ","
4550           (comma) operator, before a function is called (but after the
4551           evaluation of its arguments and the expression denoting the called
4552           function), and in certain other places.  Other than as expressed by
4553           the sequence point rules, the order of evaluation of subexpressions
4554           of an expression is not specified.  All these rules describe only a
4555           partial order rather than a total order, since, for example, if two
4556           functions are called within one expression with no sequence point
4557           between them, the order in which the functions are called is not
4558           specified.  However, the standards committee have ruled that
4559           function calls do not overlap.
4560
4561           It is not specified when between sequence points modifications to
4562           the values of objects take effect.  Programs whose behavior depends
4563           on this have undefined behavior; the C and C++ standards specify
4564           that "Between the previous and next sequence point an object shall
4565           have its stored value modified at most once by the evaluation of an
4566           expression.  Furthermore, the prior value shall be read only to
4567           determine the value to be stored.".  If a program breaks these
4568           rules, the results on any particular implementation are entirely
4569           unpredictable.
4570
4571           Examples of code with undefined behavior are "a = a++;", "a[n] =
4572           b[n++]" and "a[i++] = i;".  Some more complicated cases are not
4573           diagnosed by this option, and it may give an occasional false
4574           positive result, but in general it has been found fairly effective
4575           at detecting this sort of problem in programs.
4576
4577           The C++17 standard will define the order of evaluation of operands
4578           in more cases: in particular it requires that the right-hand side
4579           of an assignment be evaluated before the left-hand side, so the
4580           above examples are no longer undefined.  But this option will still
4581           warn about them, to help people avoid writing code that is
4582           undefined in C and earlier revisions of C++.
4583
4584           The standard is worded confusingly, therefore there is some debate
4585           over the precise meaning of the sequence point rules in subtle
4586           cases.  Links to discussions of the problem, including proposed
4587           formal definitions, may be found on the GCC readings page, at
4588           <http://gcc.gnu.org/readings.html>.
4589
4590           This warning is enabled by -Wall for C and C++.
4591
4592       -Wno-return-local-addr
4593           Do not warn about returning a pointer (or in C++, a reference) to a
4594           variable that goes out of scope after the function returns.
4595
4596       -Wreturn-type
4597           Warn whenever a function is defined with a return type that
4598           defaults to "int".  Also warn about any "return" statement with no
4599           return value in a function whose return type is not "void" (falling
4600           off the end of the function body is considered returning without a
4601           value).
4602
4603           For C only, warn about a "return" statement with an expression in a
4604           function whose return type is "void", unless the expression type is
4605           also "void".  As a GNU extension, the latter case is accepted
4606           without a warning unless -Wpedantic is used.  Attempting to use the
4607           return value of a non-"void" function other than "main" that flows
4608           off the end by reaching the closing curly brace that terminates the
4609           function is undefined.
4610
4611           Unlike in C, in C++, flowing off the end of a non-"void" function
4612           other than "main" results in undefined behavior even when the value
4613           of the function is not used.
4614
4615           This warning is enabled by default in C++ and by -Wall otherwise.
4616
4617       -Wno-shift-count-negative
4618           Controls warnings if a shift count is negative.  This warning is
4619           enabled by default.
4620
4621       -Wno-shift-count-overflow
4622           Controls warnings if a shift count is greater than or equal to the
4623           bit width of the type.  This warning is enabled by default.
4624
4625       -Wshift-negative-value
4626           Warn if left shifting a negative value.  This warning is enabled by
4627           -Wextra in C99 and C++11 modes (and newer).
4628
4629       -Wno-shift-overflow
4630       -Wshift-overflow=n
4631           These options control warnings about left shift overflows.
4632
4633           -Wshift-overflow=1
4634               This is the warning level of -Wshift-overflow and is enabled by
4635               default in C99 and C++11 modes (and newer).  This warning level
4636               does not warn about left-shifting 1 into the sign bit.
4637               (However, in C, such an overflow is still rejected in contexts
4638               where an integer constant expression is required.)  No warning
4639               is emitted in C++2A mode (and newer), as signed left shifts
4640               always wrap.
4641
4642           -Wshift-overflow=2
4643               This warning level also warns about left-shifting 1 into the
4644               sign bit, unless C++14 mode (or newer) is active.
4645
4646       -Wswitch
4647           Warn whenever a "switch" statement has an index of enumerated type
4648           and lacks a "case" for one or more of the named codes of that
4649           enumeration.  (The presence of a "default" label prevents this
4650           warning.)  "case" labels outside the enumeration range also provoke
4651           warnings when this option is used (even if there is a "default"
4652           label).  This warning is enabled by -Wall.
4653
4654       -Wswitch-default
4655           Warn whenever a "switch" statement does not have a "default" case.
4656
4657       -Wswitch-enum
4658           Warn whenever a "switch" statement has an index of enumerated type
4659           and lacks a "case" for one or more of the named codes of that
4660           enumeration.  "case" labels outside the enumeration range also
4661           provoke warnings when this option is used.  The only difference
4662           between -Wswitch and this option is that this option gives a
4663           warning about an omitted enumeration code even if there is a
4664           "default" label.
4665
4666       -Wno-switch-bool
4667           Do not warn when a "switch" statement has an index of boolean type
4668           and the case values are outside the range of a boolean type.  It is
4669           possible to suppress this warning by casting the controlling
4670           expression to a type other than "bool".  For example:
4671
4672                   switch ((int) (a == 4))
4673                     {
4674                     ...
4675                     }
4676
4677           This warning is enabled by default for C and C++ programs.
4678
4679       -Wno-switch-outside-range
4680           This option controls warnings when a "switch" case has a value that
4681           is outside of its respective type range.  This warning is enabled
4682           by default for C and C++ programs.
4683
4684       -Wno-switch-unreachable
4685           Do not warn when a "switch" statement contains statements between
4686           the controlling expression and the first case label, which will
4687           never be executed.  For example:
4688
4689                   switch (cond)
4690                     {
4691                      i = 15;
4692                     ...
4693                      case 5:
4694                     ...
4695                     }
4696
4697           -Wswitch-unreachable does not warn if the statement between the
4698           controlling expression and the first case label is just a
4699           declaration:
4700
4701                   switch (cond)
4702                     {
4703                      int i;
4704                     ...
4705                      case 5:
4706                      i = 5;
4707                     ...
4708                     }
4709
4710           This warning is enabled by default for C and C++ programs.
4711
4712       -Wsync-nand (C and C++ only)
4713           Warn when "__sync_fetch_and_nand" and "__sync_nand_and_fetch"
4714           built-in functions are used.  These functions changed semantics in
4715           GCC 4.4.
4716
4717       -Wunused-but-set-parameter
4718           Warn whenever a function parameter is assigned to, but otherwise
4719           unused (aside from its declaration).
4720
4721           To suppress this warning use the "unused" attribute.
4722
4723           This warning is also enabled by -Wunused together with -Wextra.
4724
4725       -Wunused-but-set-variable
4726           Warn whenever a local variable is assigned to, but otherwise unused
4727           (aside from its declaration).  This warning is enabled by -Wall.
4728
4729           To suppress this warning use the "unused" attribute.
4730
4731           This warning is also enabled by -Wunused, which is enabled by
4732           -Wall.
4733
4734       -Wunused-function
4735           Warn whenever a static function is declared but not defined or a
4736           non-inline static function is unused.  This warning is enabled by
4737           -Wall.
4738
4739       -Wunused-label
4740           Warn whenever a label is declared but not used.  This warning is
4741           enabled by -Wall.
4742
4743           To suppress this warning use the "unused" attribute.
4744
4745       -Wunused-local-typedefs (C, Objective-C, C++ and Objective-C++ only)
4746           Warn when a typedef locally defined in a function is not used.
4747           This warning is enabled by -Wall.
4748
4749       -Wunused-parameter
4750           Warn whenever a function parameter is unused aside from its
4751           declaration.
4752
4753           To suppress this warning use the "unused" attribute.
4754
4755       -Wno-unused-result
4756           Do not warn if a caller of a function marked with attribute
4757           "warn_unused_result" does not use its return value. The default is
4758           -Wunused-result.
4759
4760       -Wunused-variable
4761           Warn whenever a local or static variable is unused aside from its
4762           declaration. This option implies -Wunused-const-variable=1 for C,
4763           but not for C++. This warning is enabled by -Wall.
4764
4765           To suppress this warning use the "unused" attribute.
4766
4767       -Wunused-const-variable
4768       -Wunused-const-variable=n
4769           Warn whenever a constant static variable is unused aside from its
4770           declaration.  -Wunused-const-variable=1 is enabled by
4771           -Wunused-variable for C, but not for C++. In C this declares
4772           variable storage, but in C++ this is not an error since const
4773           variables take the place of "#define"s.
4774
4775           To suppress this warning use the "unused" attribute.
4776
4777           -Wunused-const-variable=1
4778               This is the warning level that is enabled by -Wunused-variable
4779               for C.  It warns only about unused static const variables
4780               defined in the main compilation unit, but not about static
4781               const variables declared in any header included.
4782
4783           -Wunused-const-variable=2
4784               This warning level also warns for unused constant static
4785               variables in headers (excluding system headers).  This is the
4786               warning level of -Wunused-const-variable and must be explicitly
4787               requested since in C++ this isn't an error and in C it might be
4788               harder to clean up all headers included.
4789
4790       -Wunused-value
4791           Warn whenever a statement computes a result that is explicitly not
4792           used. To suppress this warning cast the unused expression to
4793           "void". This includes an expression-statement or the left-hand side
4794           of a comma expression that contains no side effects. For example,
4795           an expression such as "x[i,j]" causes a warning, while
4796           "x[(void)i,j]" does not.
4797
4798           This warning is enabled by -Wall.
4799
4800       -Wunused
4801           All the above -Wunused options combined.
4802
4803           In order to get a warning about an unused function parameter, you
4804           must either specify -Wextra -Wunused (note that -Wall implies
4805           -Wunused), or separately specify -Wunused-parameter.
4806
4807       -Wuninitialized
4808           Warn if an automatic variable is used without first being
4809           initialized.  In C++, warn if a non-static reference or non-static
4810           "const" member appears in a class without constructors.
4811
4812           If you want to warn about code that uses the uninitialized value of
4813           the variable in its own initializer, use the -Winit-self option.
4814
4815           These warnings occur for individual uninitialized elements of
4816           structure, union or array variables as well as for variables that
4817           are uninitialized as a whole.  They do not occur for variables or
4818           elements declared "volatile".  Because these warnings depend on
4819           optimization, the exact variables or elements for which there are
4820           warnings depend on the precise optimization options and version of
4821           GCC used.
4822
4823           Note that there may be no warning about a variable that is used
4824           only to compute a value that itself is never used, because such
4825           computations may be deleted by data flow analysis before the
4826           warnings are printed.
4827
4828       -Wno-invalid-memory-model
4829           This option controls warnings for invocations of __atomic Builtins,
4830           __sync Builtins, and the C11 atomic generic functions with a memory
4831           consistency argument that is either invalid for the operation or
4832           outside the range of values of the "memory_order" enumeration.  For
4833           example, since the "__atomic_store" and "__atomic_store_n" built-
4834           ins are only defined for the relaxed, release, and sequentially
4835           consistent memory orders the following code is diagnosed:
4836
4837                   void store (int *i)
4838                   {
4839                     __atomic_store_n (i, 0, memory_order_consume);
4840                   }
4841
4842           -Winvalid-memory-model is enabled by default.
4843
4844       -Wmaybe-uninitialized
4845           For an automatic (i.e. local) variable, if there exists a path from
4846           the function entry to a use of the variable that is initialized,
4847           but there exist some other paths for which the variable is not
4848           initialized, the compiler emits a warning if it cannot prove the
4849           uninitialized paths are not executed at run time.
4850
4851           These warnings are only possible in optimizing compilation, because
4852           otherwise GCC does not keep track of the state of variables.
4853
4854           These warnings are made optional because GCC may not be able to
4855           determine when the code is correct in spite of appearing to have an
4856           error.  Here is one example of how this can happen:
4857
4858                   {
4859                     int x;
4860                     switch (y)
4861                       {
4862                       case 1: x = 1;
4863                         break;
4864                       case 2: x = 4;
4865                         break;
4866                       case 3: x = 5;
4867                       }
4868                     foo (x);
4869                   }
4870
4871           If the value of "y" is always 1, 2 or 3, then "x" is always
4872           initialized, but GCC doesn't know this. To suppress the warning,
4873           you need to provide a default case with assert(0) or similar code.
4874
4875           This option also warns when a non-volatile automatic variable might
4876           be changed by a call to "longjmp".  The compiler sees only the
4877           calls to "setjmp".  It cannot know where "longjmp" will be called;
4878           in fact, a signal handler could call it at any point in the code.
4879           As a result, you may get a warning even when there is in fact no
4880           problem because "longjmp" cannot in fact be called at the place
4881           that would cause a problem.
4882
4883           Some spurious warnings can be avoided if you declare all the
4884           functions you use that never return as "noreturn".
4885
4886           This warning is enabled by -Wall or -Wextra.
4887
4888       -Wunknown-pragmas
4889           Warn when a "#pragma" directive is encountered that is not
4890           understood by GCC.  If this command-line option is used, warnings
4891           are even issued for unknown pragmas in system header files.  This
4892           is not the case if the warnings are only enabled by the -Wall
4893           command-line option.
4894
4895       -Wno-pragmas
4896           Do not warn about misuses of pragmas, such as incorrect parameters,
4897           invalid syntax, or conflicts between pragmas.  See also
4898           -Wunknown-pragmas.
4899
4900       -Wno-prio-ctor-dtor
4901           Do not warn if a priority from 0 to 100 is used for constructor or
4902           destructor.  The use of constructor and destructor attributes allow
4903           you to assign a priority to the constructor/destructor to control
4904           its order of execution before "main" is called or after it returns.
4905           The priority values must be greater than 100 as the compiler
4906           reserves priority values between 0--100 for the implementation.
4907
4908       -Wstrict-aliasing
4909           This option is only active when -fstrict-aliasing is active.  It
4910           warns about code that might break the strict aliasing rules that
4911           the compiler is using for optimization.  The warning does not catch
4912           all cases, but does attempt to catch the more common pitfalls.  It
4913           is included in -Wall.  It is equivalent to -Wstrict-aliasing=3
4914
4915       -Wstrict-aliasing=n
4916           This option is only active when -fstrict-aliasing is active.  It
4917           warns about code that might break the strict aliasing rules that
4918           the compiler is using for optimization.  Higher levels correspond
4919           to higher accuracy (fewer false positives).  Higher levels also
4920           correspond to more effort, similar to the way -O works.
4921           -Wstrict-aliasing is equivalent to -Wstrict-aliasing=3.
4922
4923           Level 1: Most aggressive, quick, least accurate.  Possibly useful
4924           when higher levels do not warn but -fstrict-aliasing still breaks
4925           the code, as it has very few false negatives.  However, it has many
4926           false positives.  Warns for all pointer conversions between
4927           possibly incompatible types, even if never dereferenced.  Runs in
4928           the front end only.
4929
4930           Level 2: Aggressive, quick, not too precise.  May still have many
4931           false positives (not as many as level 1 though), and few false
4932           negatives (but possibly more than level 1).  Unlike level 1, it
4933           only warns when an address is taken.  Warns about incomplete types.
4934           Runs in the front end only.
4935
4936           Level 3 (default for -Wstrict-aliasing): Should have very few false
4937           positives and few false negatives.  Slightly slower than levels 1
4938           or 2 when optimization is enabled.  Takes care of the common
4939           pun+dereference pattern in the front end: "*(int*)&some_float".  If
4940           optimization is enabled, it also runs in the back end, where it
4941           deals with multiple statement cases using flow-sensitive points-to
4942           information.  Only warns when the converted pointer is
4943           dereferenced.  Does not warn about incomplete types.
4944
4945       -Wstrict-overflow
4946       -Wstrict-overflow=n
4947           This option is only active when signed overflow is undefined.  It
4948           warns about cases where the compiler optimizes based on the
4949           assumption that signed overflow does not occur.  Note that it does
4950           not warn about all cases where the code might overflow: it only
4951           warns about cases where the compiler implements some optimization.
4952           Thus this warning depends on the optimization level.
4953
4954           An optimization that assumes that signed overflow does not occur is
4955           perfectly safe if the values of the variables involved are such
4956           that overflow never does, in fact, occur.  Therefore this warning
4957           can easily give a false positive: a warning about code that is not
4958           actually a problem.  To help focus on important issues, several
4959           warning levels are defined.  No warnings are issued for the use of
4960           undefined signed overflow when estimating how many iterations a
4961           loop requires, in particular when determining whether a loop will
4962           be executed at all.
4963
4964           -Wstrict-overflow=1
4965               Warn about cases that are both questionable and easy to avoid.
4966               For example the compiler simplifies "x + 1 > x" to 1.  This
4967               level of -Wstrict-overflow is enabled by -Wall; higher levels
4968               are not, and must be explicitly requested.
4969
4970           -Wstrict-overflow=2
4971               Also warn about other cases where a comparison is simplified to
4972               a constant.  For example: "abs (x) >= 0".  This can only be
4973               simplified when signed integer overflow is undefined, because
4974               "abs (INT_MIN)" overflows to "INT_MIN", which is less than
4975               zero.  -Wstrict-overflow (with no level) is the same as
4976               -Wstrict-overflow=2.
4977
4978           -Wstrict-overflow=3
4979               Also warn about other cases where a comparison is simplified.
4980               For example: "x + 1 > 1" is simplified to "x > 0".
4981
4982           -Wstrict-overflow=4
4983               Also warn about other simplifications not covered by the above
4984               cases.  For example: "(x * 10) / 5" is simplified to "x * 2".
4985
4986           -Wstrict-overflow=5
4987               Also warn about cases where the compiler reduces the magnitude
4988               of a constant involved in a comparison.  For example: "x + 2 >
4989               y" is simplified to "x + 1 >= y".  This is reported only at the
4990               highest warning level because this simplification applies to
4991               many comparisons, so this warning level gives a very large
4992               number of false positives.
4993
4994       -Wstring-compare
4995           Warn for calls to "strcmp" and "strncmp" whose result is determined
4996           to be either zero or non-zero in tests for such equality owing to
4997           the length of one argument being greater than the size of the array
4998           the other argument is stored in (or the bound in the case of
4999           "strncmp").  Such calls could be mistakes.  For example, the call
5000           to "strcmp" below is diagnosed because its result is necessarily
5001           non-zero irrespective of the contents of the array "a".
5002
5003                   extern char a[4];
5004                   void f (char *d)
5005                   {
5006                     strcpy (d, "string");
5007                     ...
5008                     if (0 == strcmp (a, d))   // cannot be true
5009                       puts ("a and d are the same");
5010                   }
5011
5012           -Wstring-compare is enabled by -Wextra.
5013
5014       -Wstringop-overflow
5015       -Wstringop-overflow=type
5016           Warn for calls to string manipulation functions such as "memcpy"
5017           and "strcpy" that are determined to overflow the destination
5018           buffer.  The optional argument is one greater than the type of
5019           Object Size Checking to perform to determine the size of the
5020           destination.  The argument is meaningful only for functions that
5021           operate on character arrays but not for raw memory functions like
5022           "memcpy" which always make use of Object Size type-0.  The option
5023           also warns for calls that specify a size in excess of the largest
5024           possible object or at most "SIZE_MAX / 2" bytes.  The option
5025           produces the best results with optimization enabled but can detect
5026           a small subset of simple buffer overflows even without optimization
5027           in calls to the GCC built-in functions like "__builtin_memcpy" that
5028           correspond to the standard functions.  In any case, the option
5029           warns about just a subset of buffer overflows detected by the
5030           corresponding overflow checking built-ins.  For example, the option
5031           issues a warning for the "strcpy" call below because it copies at
5032           least 5 characters (the string "blue" including the terminating
5033           NUL) into the buffer of size 4.
5034
5035                   enum Color { blue, purple, yellow };
5036                   const char* f (enum Color clr)
5037                   {
5038                     static char buf [4];
5039                     const char *str;
5040                     switch (clr)
5041                       {
5042                         case blue: str = "blue"; break;
5043                         case purple: str = "purple"; break;
5044                         case yellow: str = "yellow"; break;
5045                       }
5046
5047                     return strcpy (buf, str);   // warning here
5048                   }
5049
5050           Option -Wstringop-overflow=2 is enabled by default.
5051
5052           -Wstringop-overflow
5053           -Wstringop-overflow=1
5054               The -Wstringop-overflow=1 option uses type-zero Object Size
5055               Checking to determine the sizes of destination objects.  This
5056               is the default setting of the option.  At this setting the
5057               option does not warn for writes past the end of subobjects of
5058               larger objects accessed by pointers unless the size of the
5059               largest surrounding object is known.  When the destination may
5060               be one of several objects it is assumed to be the largest one
5061               of them.  On Linux systems, when optimization is enabled at
5062               this setting the option warns for the same code as when the
5063               "_FORTIFY_SOURCE" macro is defined to a non-zero value.
5064
5065           -Wstringop-overflow=2
5066               The -Wstringop-overflow=2 option uses type-one Object Size
5067               Checking to determine the sizes of destination objects.  At
5068               this setting the option warna about overflows when writing to
5069               members of the largest complete objects whose exact size is
5070               known.  However, it does not warn for excessive writes to the
5071               same members of unknown objects referenced by pointers since
5072               they may point to arrays containing unknown numbers of
5073               elements.
5074
5075           -Wstringop-overflow=3
5076               The -Wstringop-overflow=3 option uses type-two Object Size
5077               Checking to determine the sizes of destination objects.  At
5078               this setting the option warns about overflowing the smallest
5079               object or data member.  This is the most restrictive setting of
5080               the option that may result in warnings for safe code.
5081
5082           -Wstringop-overflow=4
5083               The -Wstringop-overflow=4 option uses type-three Object Size
5084               Checking to determine the sizes of destination objects.  At
5085               this setting the option warns about overflowing any data
5086               members, and when the destination is one of several objects it
5087               uses the size of the largest of them to decide whether to issue
5088               a warning.  Similarly to -Wstringop-overflow=3 this setting of
5089               the option may result in warnings for benign code.
5090
5091       -Wno-stringop-truncation
5092           Do not warn for calls to bounded string manipulation functions such
5093           as "strncat", "strncpy", and "stpncpy" that may either truncate the
5094           copied string or leave the destination unchanged.
5095
5096           In the following example, the call to "strncat" specifies a bound
5097           that is less than the length of the source string.  As a result,
5098           the copy of the source will be truncated and so the call is
5099           diagnosed.  To avoid the warning use "bufsize - strlen (buf) - 1)"
5100           as the bound.
5101
5102                   void append (char *buf, size_t bufsize)
5103                   {
5104                     strncat (buf, ".txt", 3);
5105                   }
5106
5107           As another example, the following call to "strncpy" results in
5108           copying to "d" just the characters preceding the terminating NUL,
5109           without appending the NUL to the end.  Assuming the result of
5110           "strncpy" is necessarily a NUL-terminated string is a common
5111           mistake, and so the call is diagnosed.  To avoid the warning when
5112           the result is not expected to be NUL-terminated, call "memcpy"
5113           instead.
5114
5115                   void copy (char *d, const char *s)
5116                   {
5117                     strncpy (d, s, strlen (s));
5118                   }
5119
5120           In the following example, the call to "strncpy" specifies the size
5121           of the destination buffer as the bound.  If the length of the
5122           source string is equal to or greater than this size the result of
5123           the copy will not be NUL-terminated.  Therefore, the call is also
5124           diagnosed.  To avoid the warning, specify "sizeof buf - 1" as the
5125           bound and set the last element of the buffer to "NUL".
5126
5127                   void copy (const char *s)
5128                   {
5129                     char buf[80];
5130                     strncpy (buf, s, sizeof buf);
5131                     ...
5132                   }
5133
5134           In situations where a character array is intended to store a
5135           sequence of bytes with no terminating "NUL" such an array may be
5136           annotated with attribute "nonstring" to avoid this warning.  Such
5137           arrays, however, are not suitable arguments to functions that
5138           expect "NUL"-terminated strings.  To help detect accidental misuses
5139           of such arrays GCC issues warnings unless it can prove that the use
5140           is safe.
5141
5142       -Wsuggest-attribute=[pure|const|noreturn|format|cold|malloc]
5143           Warn for cases where adding an attribute may be beneficial. The
5144           attributes currently supported are listed below.
5145
5146           -Wsuggest-attribute=pure
5147           -Wsuggest-attribute=const
5148           -Wsuggest-attribute=noreturn
5149           -Wmissing-noreturn
5150           -Wsuggest-attribute=malloc
5151               Warn about functions that might be candidates for attributes
5152               "pure", "const" or "noreturn" or "malloc". The compiler only
5153               warns for functions visible in other compilation units or (in
5154               the case of "pure" and "const") if it cannot prove that the
5155               function returns normally. A function returns normally if it
5156               doesn't contain an infinite loop or return abnormally by
5157               throwing, calling "abort" or trapping.  This analysis requires
5158               option -fipa-pure-const, which is enabled by default at -O and
5159               higher.  Higher optimization levels improve the accuracy of the
5160               analysis.
5161
5162           -Wsuggest-attribute=format
5163           -Wmissing-format-attribute
5164               Warn about function pointers that might be candidates for
5165               "format" attributes.  Note these are only possible candidates,
5166               not absolute ones.  GCC guesses that function pointers with
5167               "format" attributes that are used in assignment,
5168               initialization, parameter passing or return statements should
5169               have a corresponding "format" attribute in the resulting type.
5170               I.e. the left-hand side of the assignment or initialization,
5171               the type of the parameter variable, or the return type of the
5172               containing function respectively should also have a "format"
5173               attribute to avoid the warning.
5174
5175               GCC also warns about function definitions that might be
5176               candidates for "format" attributes.  Again, these are only
5177               possible candidates.  GCC guesses that "format" attributes
5178               might be appropriate for any function that calls a function
5179               like "vprintf" or "vscanf", but this might not always be the
5180               case, and some functions for which "format" attributes are
5181               appropriate may not be detected.
5182
5183           -Wsuggest-attribute=cold
5184               Warn about functions that might be candidates for "cold"
5185               attribute.  This is based on static detection and generally
5186               only warns about functions which always leads to a call to
5187               another "cold" function such as wrappers of C++ "throw" or
5188               fatal error reporting functions leading to "abort".
5189
5190       -Walloc-zero
5191           Warn about calls to allocation functions decorated with attribute
5192           "alloc_size" that specify zero bytes, including those to the built-
5193           in forms of the functions "aligned_alloc", "alloca", "calloc",
5194           "malloc", and "realloc".  Because the behavior of these functions
5195           when called with a zero size differs among implementations (and in
5196           the case of "realloc" has been deprecated) relying on it may result
5197           in subtle portability bugs and should be avoided.
5198
5199       -Walloc-size-larger-than=byte-size
5200           Warn about calls to functions decorated with attribute "alloc_size"
5201           that attempt to allocate objects larger than the specified number
5202           of bytes, or where the result of the size computation in an integer
5203           type with infinite precision would exceed the value of PTRDIFF_MAX
5204           on the target.  -Walloc-size-larger-than=PTRDIFF_MAX is enabled by
5205           default.  Warnings controlled by the option can be disabled either
5206           by specifying byte-size of SIZE_MAX or more or by
5207           -Wno-alloc-size-larger-than.
5208
5209       -Wno-alloc-size-larger-than
5210           Disable -Walloc-size-larger-than= warnings.  The option is
5211           equivalent to -Walloc-size-larger-than=SIZE_MAX or larger.
5212
5213       -Walloca
5214           This option warns on all uses of "alloca" in the source.
5215
5216       -Walloca-larger-than=byte-size
5217           This option warns on calls to "alloca" with an integer argument
5218           whose value is either zero, or that is not bounded by a controlling
5219           predicate that limits its value to at most byte-size.  It also
5220           warns for calls to "alloca" where the bound value is unknown.
5221           Arguments of non-integer types are considered unbounded even if
5222           they appear to be constrained to the expected range.
5223
5224           For example, a bounded case of "alloca" could be:
5225
5226                   void func (size_t n)
5227                   {
5228                     void *p;
5229                     if (n <= 1000)
5230                       p = alloca (n);
5231                     else
5232                       p = malloc (n);
5233                     f (p);
5234                   }
5235
5236           In the above example, passing "-Walloca-larger-than=1000" would not
5237           issue a warning because the call to "alloca" is known to be at most
5238           1000 bytes.  However, if "-Walloca-larger-than=500" were passed,
5239           the compiler would emit a warning.
5240
5241           Unbounded uses, on the other hand, are uses of "alloca" with no
5242           controlling predicate constraining its integer argument.  For
5243           example:
5244
5245                   void func ()
5246                   {
5247                     void *p = alloca (n);
5248                     f (p);
5249                   }
5250
5251           If "-Walloca-larger-than=500" were passed, the above would trigger
5252           a warning, but this time because of the lack of bounds checking.
5253
5254           Note, that even seemingly correct code involving signed integers
5255           could cause a warning:
5256
5257                   void func (signed int n)
5258                   {
5259                     if (n < 500)
5260                       {
5261                         p = alloca (n);
5262                         f (p);
5263                       }
5264                   }
5265
5266           In the above example, n could be negative, causing a larger than
5267           expected argument to be implicitly cast into the "alloca" call.
5268
5269           This option also warns when "alloca" is used in a loop.
5270
5271           -Walloca-larger-than=PTRDIFF_MAX is enabled by default but is
5272           usually only effective  when -ftree-vrp is active (default for -O2
5273           and above).
5274
5275           See also -Wvla-larger-than=byte-size.
5276
5277       -Wno-alloca-larger-than
5278           Disable -Walloca-larger-than= warnings.  The option is equivalent
5279           to -Walloca-larger-than=SIZE_MAX or larger.
5280
5281       -Warith-conversion
5282           Do warn about implicit conversions from arithmetic operations even
5283           when conversion of the operands to the same type cannot change
5284           their values.  This affects warnings from -Wconversion,
5285           -Wfloat-conversion, and -Wsign-conversion.
5286
5287                   void f (char c, int i)
5288                   {
5289                     c = c + i; // warns with B<-Wconversion>
5290                     c = c + 1; // only warns with B<-Warith-conversion>
5291                   }
5292
5293       -Warray-bounds
5294       -Warray-bounds=n
5295           This option is only active when -ftree-vrp is active (default for
5296           -O2 and above). It warns about subscripts to arrays that are always
5297           out of bounds. This warning is enabled by -Wall.
5298
5299           -Warray-bounds=1
5300               This is the warning level of -Warray-bounds and is enabled by
5301               -Wall; higher levels are not, and must be explicitly requested.
5302
5303           -Warray-bounds=2
5304               This warning level also warns about out of bounds access for
5305               arrays at the end of a struct and for arrays accessed through
5306               pointers. This warning level may give a larger number of false
5307               positives and is deactivated by default.
5308
5309       -Wattribute-alias=n
5310       -Wno-attribute-alias
5311           Warn about declarations using the "alias" and similar attributes
5312           whose target is incompatible with the type of the alias.
5313
5314           -Wattribute-alias=1
5315               The default warning level of the -Wattribute-alias option
5316               diagnoses incompatibilities between the type of the alias
5317               declaration and that of its target.  Such incompatibilities are
5318               typically indicative of bugs.
5319
5320           -Wattribute-alias=2
5321               At this level -Wattribute-alias also diagnoses cases where the
5322               attributes of the alias declaration are more restrictive than
5323               the attributes applied to its target.  These mismatches can
5324               potentially result in incorrect code generation.  In other
5325               cases they may be benign and could be resolved simply by adding
5326               the missing attribute to the target.  For comparison, see the
5327               -Wmissing-attributes option, which controls diagnostics when
5328               the alias declaration is less restrictive than the target,
5329               rather than more restrictive.
5330
5331               Attributes considered include "alloc_align", "alloc_size",
5332               "cold", "const", "hot", "leaf", "malloc", "nonnull",
5333               "noreturn", "nothrow", "pure", "returns_nonnull", and
5334               "returns_twice".
5335
5336           -Wattribute-alias is equivalent to -Wattribute-alias=1.  This is
5337           the default.  You can disable these warnings with either
5338           -Wno-attribute-alias or -Wattribute-alias=0.
5339
5340       -Wbool-compare
5341           Warn about boolean expression compared with an integer value
5342           different from "true"/"false".  For instance, the following
5343           comparison is always false:
5344
5345                   int n = 5;
5346                   ...
5347                   if ((n > 1) == 2) { ... }
5348
5349           This warning is enabled by -Wall.
5350
5351       -Wbool-operation
5352           Warn about suspicious operations on expressions of a boolean type.
5353           For instance, bitwise negation of a boolean is very likely a bug in
5354           the program.  For C, this warning also warns about incrementing or
5355           decrementing a boolean, which rarely makes sense.  (In C++,
5356           decrementing a boolean is always invalid.  Incrementing a boolean
5357           is invalid in C++17, and deprecated otherwise.)
5358
5359           This warning is enabled by -Wall.
5360
5361       -Wduplicated-branches
5362           Warn when an if-else has identical branches.  This warning detects
5363           cases like
5364
5365                   if (p != NULL)
5366                     return 0;
5367                   else
5368                     return 0;
5369
5370           It doesn't warn when both branches contain just a null statement.
5371           This warning also warn for conditional operators:
5372
5373                     int i = x ? *p : *p;
5374
5375       -Wduplicated-cond
5376           Warn about duplicated conditions in an if-else-if chain.  For
5377           instance, warn for the following code:
5378
5379                   if (p->q != NULL) { ... }
5380                   else if (p->q != NULL) { ... }
5381
5382       -Wframe-address
5383           Warn when the __builtin_frame_address or __builtin_return_address
5384           is called with an argument greater than 0.  Such calls may return
5385           indeterminate values or crash the program.  The warning is included
5386           in -Wall.
5387
5388       -Wno-discarded-qualifiers (C and Objective-C only)
5389           Do not warn if type qualifiers on pointers are being discarded.
5390           Typically, the compiler warns if a "const char *" variable is
5391           passed to a function that takes a "char *" parameter.  This option
5392           can be used to suppress such a warning.
5393
5394       -Wno-discarded-array-qualifiers (C and Objective-C only)
5395           Do not warn if type qualifiers on arrays which are pointer targets
5396           are being discarded.  Typically, the compiler warns if a "const int
5397           (*)[]" variable is passed to a function that takes a "int (*)[]"
5398           parameter.  This option can be used to suppress such a warning.
5399
5400       -Wno-incompatible-pointer-types (C and Objective-C only)
5401           Do not warn when there is a conversion between pointers that have
5402           incompatible types.  This warning is for cases not covered by
5403           -Wno-pointer-sign, which warns for pointer argument passing or
5404           assignment with different signedness.
5405
5406       -Wno-int-conversion (C and Objective-C only)
5407           Do not warn about incompatible integer to pointer and pointer to
5408           integer conversions.  This warning is about implicit conversions;
5409           for explicit conversions the warnings -Wno-int-to-pointer-cast and
5410           -Wno-pointer-to-int-cast may be used.
5411
5412       -Wzero-length-bounds
5413           Warn about accesses to elements of zero-length array members that
5414           might overlap other members of the same object.  Declaring interior
5415           zero-length arrays is discouraged because accesses to them are
5416           undefined.  See
5417
5418           For example, the first two stores in function "bad" are diagnosed
5419           because the array elements overlap the subsequent members "b" and
5420           "c".  The third store is diagnosed by -Warray-bounds because it is
5421           beyond the bounds of the enclosing object.
5422
5423                   struct X { int a[0]; int b, c; };
5424                   struct X x;
5425
5426                   void bad (void)
5427                   {
5428                     x.a[0] = 0;   // -Wzero-length-bounds
5429                     x.a[1] = 1;   // -Wzero-length-bounds
5430                     x.a[2] = 2;   // -Warray-bounds
5431                   }
5432
5433           Option -Wzero-length-bounds is enabled by -Warray-bounds.
5434
5435       -Wno-div-by-zero
5436           Do not warn about compile-time integer division by zero.  Floating-
5437           point division by zero is not warned about, as it can be a
5438           legitimate way of obtaining infinities and NaNs.
5439
5440       -Wsystem-headers
5441           Print warning messages for constructs found in system header files.
5442           Warnings from system headers are normally suppressed, on the
5443           assumption that they usually do not indicate real problems and
5444           would only make the compiler output harder to read.  Using this
5445           command-line option tells GCC to emit warnings from system headers
5446           as if they occurred in user code.  However, note that using -Wall
5447           in conjunction with this option does not warn about unknown pragmas
5448           in system headers---for that, -Wunknown-pragmas must also be used.
5449
5450       -Wtautological-compare
5451           Warn if a self-comparison always evaluates to true or false.  This
5452           warning detects various mistakes such as:
5453
5454                   int i = 1;
5455                   ...
5456                   if (i > i) { ... }
5457
5458           This warning also warns about bitwise comparisons that always
5459           evaluate to true or false, for instance:
5460
5461                   if ((a & 16) == 10) { ... }
5462
5463           will always be false.
5464
5465           This warning is enabled by -Wall.
5466
5467       -Wtrampolines
5468           Warn about trampolines generated for pointers to nested functions.
5469           A trampoline is a small piece of data or code that is created at
5470           run time on the stack when the address of a nested function is
5471           taken, and is used to call the nested function indirectly.  For
5472           some targets, it is made up of data only and thus requires no
5473           special treatment.  But, for most targets, it is made up of code
5474           and thus requires the stack to be made executable in order for the
5475           program to work properly.
5476
5477       -Wfloat-equal
5478           Warn if floating-point values are used in equality comparisons.
5479
5480           The idea behind this is that sometimes it is convenient (for the
5481           programmer) to consider floating-point values as approximations to
5482           infinitely precise real numbers.  If you are doing this, then you
5483           need to compute (by analyzing the code, or in some other way) the
5484           maximum or likely maximum error that the computation introduces,
5485           and allow for it when performing comparisons (and when producing
5486           output, but that's a different problem).  In particular, instead of
5487           testing for equality, you should check to see whether the two
5488           values have ranges that overlap; and this is done with the
5489           relational operators, so equality comparisons are probably
5490           mistaken.
5491
5492       -Wtraditional (C and Objective-C only)
5493           Warn about certain constructs that behave differently in
5494           traditional and ISO C.  Also warn about ISO C constructs that have
5495           no traditional C equivalent, and/or problematic constructs that
5496           should be avoided.
5497
5498           *   Macro parameters that appear within string literals in the
5499               macro body.  In traditional C macro replacement takes place
5500               within string literals, but in ISO C it does not.
5501
5502           *   In traditional C, some preprocessor directives did not exist.
5503               Traditional preprocessors only considered a line to be a
5504               directive if the # appeared in column 1 on the line.  Therefore
5505               -Wtraditional warns about directives that traditional C
5506               understands but ignores because the # does not appear as the
5507               first character on the line.  It also suggests you hide
5508               directives like "#pragma" not understood by traditional C by
5509               indenting them.  Some traditional implementations do not
5510               recognize "#elif", so this option suggests avoiding it
5511               altogether.
5512
5513           *   A function-like macro that appears without arguments.
5514
5515           *   The unary plus operator.
5516
5517           *   The U integer constant suffix, or the F or L floating-point
5518               constant suffixes.  (Traditional C does support the L suffix on
5519               integer constants.)  Note, these suffixes appear in macros
5520               defined in the system headers of most modern systems, e.g. the
5521               _MIN/_MAX macros in "<limits.h>".  Use of these macros in user
5522               code might normally lead to spurious warnings, however GCC's
5523               integrated preprocessor has enough context to avoid warning in
5524               these cases.
5525
5526           *   A function declared external in one block and then used after
5527               the end of the block.
5528
5529           *   A "switch" statement has an operand of type "long".
5530
5531           *   A non-"static" function declaration follows a "static" one.
5532               This construct is not accepted by some traditional C compilers.
5533
5534           *   The ISO type of an integer constant has a different width or
5535               signedness from its traditional type.  This warning is only
5536               issued if the base of the constant is ten.  I.e. hexadecimal or
5537               octal values, which typically represent bit patterns, are not
5538               warned about.
5539
5540           *   Usage of ISO string concatenation is detected.
5541
5542           *   Initialization of automatic aggregates.
5543
5544           *   Identifier conflicts with labels.  Traditional C lacks a
5545               separate namespace for labels.
5546
5547           *   Initialization of unions.  If the initializer is zero, the
5548               warning is omitted.  This is done under the assumption that the
5549               zero initializer in user code appears conditioned on e.g.
5550               "__STDC__" to avoid missing initializer warnings and relies on
5551               default initialization to zero in the traditional C case.
5552
5553           *   Conversions by prototypes between fixed/floating-point values
5554               and vice versa.  The absence of these prototypes when compiling
5555               with traditional C causes serious problems.  This is a subset
5556               of the possible conversion warnings; for the full set use
5557               -Wtraditional-conversion.
5558
5559           *   Use of ISO C style function definitions.  This warning
5560               intentionally is not issued for prototype declarations or
5561               variadic functions because these ISO C features appear in your
5562               code when using libiberty's traditional C compatibility macros,
5563               "PARAMS" and "VPARAMS".  This warning is also bypassed for
5564               nested functions because that feature is already a GCC
5565               extension and thus not relevant to traditional C compatibility.
5566
5567       -Wtraditional-conversion (C and Objective-C only)
5568           Warn if a prototype causes a type conversion that is different from
5569           what would happen to the same argument in the absence of a
5570           prototype.  This includes conversions of fixed point to floating
5571           and vice versa, and conversions changing the width or signedness of
5572           a fixed-point argument except when the same as the default
5573           promotion.
5574
5575       -Wdeclaration-after-statement (C and Objective-C only)
5576           Warn when a declaration is found after a statement in a block.
5577           This construct, known from C++, was introduced with ISO C99 and is
5578           by default allowed in GCC.  It is not supported by ISO C90.
5579
5580       -Wshadow
5581           Warn whenever a local variable or type declaration shadows another
5582           variable, parameter, type, class member (in C++), or instance
5583           variable (in Objective-C) or whenever a built-in function is
5584           shadowed.  Note that in C++, the compiler warns if a local variable
5585           shadows an explicit typedef, but not if it shadows a
5586           struct/class/enum.  If this warning is enabled, it includes also
5587           all instances of local shadowing.  This means that
5588           -Wno-shadow=local and -Wno-shadow=compatible-local are ignored when
5589           -Wshadow is used.  Same as -Wshadow=global.
5590
5591       -Wno-shadow-ivar (Objective-C only)
5592           Do not warn whenever a local variable shadows an instance variable
5593           in an Objective-C method.
5594
5595       -Wshadow=global
5596           Warn for any shadowing.  Same as -Wshadow.
5597
5598       -Wshadow=local
5599           Warn when a local variable shadows another local variable or
5600           parameter.
5601
5602       -Wshadow=compatible-local
5603           Warn when a local variable shadows another local variable or
5604           parameter whose type is compatible with that of the shadowing
5605           variable.  In C++, type compatibility here means the type of the
5606           shadowing variable can be converted to that of the shadowed
5607           variable.  The creation of this flag (in addition to
5608           -Wshadow=local) is based on the idea that when a local variable
5609           shadows another one of incompatible type, it is most likely
5610           intentional, not a bug or typo, as shown in the following example:
5611
5612                   for (SomeIterator i = SomeObj.begin(); i != SomeObj.end(); ++i)
5613                   {
5614                     for (int i = 0; i < N; ++i)
5615                     {
5616                       ...
5617                     }
5618                     ...
5619                   }
5620
5621           Since the two variable "i" in the example above have incompatible
5622           types, enabling only -Wshadow=compatible-local does not emit a
5623           warning.  Because their types are incompatible, if a programmer
5624           accidentally uses one in place of the other, type checking is
5625           expected to catch that and emit an error or warning.  Use of this
5626           flag instead of -Wshadow=local can possibly reduce the number of
5627           warnings triggered by intentional shadowing.  Note that this also
5628           means that shadowing "const char *i" by "char *i" does not emit a
5629           warning.
5630
5631           This warning is also enabled by -Wshadow=local.
5632
5633       -Wlarger-than=byte-size
5634           Warn whenever an object is defined whose size exceeds byte-size.
5635           -Wlarger-than=PTRDIFF_MAX is enabled by default.  Warnings
5636           controlled by the option can be disabled either by specifying byte-
5637           size of SIZE_MAX or more or by -Wno-larger-than.
5638
5639       -Wno-larger-than
5640           Disable -Wlarger-than= warnings.  The option is equivalent to
5641           -Wlarger-than=SIZE_MAX or larger.
5642
5643       -Wframe-larger-than=byte-size
5644           Warn if the size of a function frame exceeds byte-size.  The
5645           computation done to determine the stack frame size is approximate
5646           and not conservative.  The actual requirements may be somewhat
5647           greater than byte-size even if you do not get a warning.  In
5648           addition, any space allocated via "alloca", variable-length arrays,
5649           or related constructs is not included by the compiler when
5650           determining whether or not to issue a warning.
5651           -Wframe-larger-than=PTRDIFF_MAX is enabled by default.  Warnings
5652           controlled by the option can be disabled either by specifying byte-
5653           size of SIZE_MAX or more or by -Wno-frame-larger-than.
5654
5655       -Wno-frame-larger-than
5656           Disable -Wframe-larger-than= warnings.  The option is equivalent to
5657           -Wframe-larger-than=SIZE_MAX or larger.
5658
5659       -Wno-free-nonheap-object
5660           Do not warn when attempting to free an object that was not
5661           allocated on the heap.
5662
5663       -Wstack-usage=byte-size
5664           Warn if the stack usage of a function might exceed byte-size.  The
5665           computation done to determine the stack usage is conservative.  Any
5666           space allocated via "alloca", variable-length arrays, or related
5667           constructs is included by the compiler when determining whether or
5668           not to issue a warning.
5669
5670           The message is in keeping with the output of -fstack-usage.
5671
5672           *   If the stack usage is fully static but exceeds the specified
5673               amount, it's:
5674
5675                         warning: stack usage is 1120 bytes
5676
5677           *   If the stack usage is (partly) dynamic but bounded, it's:
5678
5679                         warning: stack usage might be 1648 bytes
5680
5681           *   If the stack usage is (partly) dynamic and not bounded, it's:
5682
5683                         warning: stack usage might be unbounded
5684
5685           -Wstack-usage=PTRDIFF_MAX is enabled by default.  Warnings
5686           controlled by the option can be disabled either by specifying byte-
5687           size of SIZE_MAX or more or by -Wno-stack-usage.
5688
5689       -Wno-stack-usage
5690           Disable -Wstack-usage= warnings.  The option is equivalent to
5691           -Wstack-usage=SIZE_MAX or larger.
5692
5693       -Wunsafe-loop-optimizations
5694           Warn if the loop cannot be optimized because the compiler cannot
5695           assume anything on the bounds of the loop indices.  With
5696           -funsafe-loop-optimizations warn if the compiler makes such
5697           assumptions.
5698
5699       -Wno-pedantic-ms-format (MinGW targets only)
5700           When used in combination with -Wformat and -pedantic without GNU
5701           extensions, this option disables the warnings about non-ISO
5702           "printf" / "scanf" format width specifiers "I32", "I64", and "I"
5703           used on Windows targets, which depend on the MS runtime.
5704
5705       -Wpointer-arith
5706           Warn about anything that depends on the "size of" a function type
5707           or of "void".  GNU C assigns these types a size of 1, for
5708           convenience in calculations with "void *" pointers and pointers to
5709           functions.  In C++, warn also when an arithmetic operation involves
5710           "NULL".  This warning is also enabled by -Wpedantic.
5711
5712       -Wno-pointer-compare
5713           Do not warn if a pointer is compared with a zero character
5714           constant.  This usually means that the pointer was meant to be
5715           dereferenced.  For example:
5716
5717                   const char *p = foo ();
5718                   if (p == '\0')
5719                     return 42;
5720
5721           Note that the code above is invalid in C++11.
5722
5723           This warning is enabled by default.
5724
5725       -Wtype-limits
5726           Warn if a comparison is always true or always false due to the
5727           limited range of the data type, but do not warn for constant
5728           expressions.  For example, warn if an unsigned variable is compared
5729           against zero with "<" or ">=".  This warning is also enabled by
5730           -Wextra.
5731
5732       -Wabsolute-value (C and Objective-C only)
5733           Warn for calls to standard functions that compute the absolute
5734           value of an argument when a more appropriate standard function is
5735           available.  For example, calling "abs(3.14)" triggers the warning
5736           because the appropriate function to call to compute the absolute
5737           value of a double argument is "fabs".  The option also triggers
5738           warnings when the argument in a call to such a function has an
5739           unsigned type.  This warning can be suppressed with an explicit
5740           type cast and it is also enabled by -Wextra.
5741
5742       -Wcomment
5743       -Wcomments
5744           Warn whenever a comment-start sequence /* appears in a /* comment,
5745           or whenever a backslash-newline appears in a // comment.  This
5746           warning is enabled by -Wall.
5747
5748       -Wtrigraphs
5749           Warn if any trigraphs are encountered that might change the meaning
5750           of the program.  Trigraphs within comments are not warned about,
5751           except those that would form escaped newlines.
5752
5753           This option is implied by -Wall.  If -Wall is not given, this
5754           option is still enabled unless trigraphs are enabled.  To get
5755           trigraph conversion without warnings, but get the other -Wall
5756           warnings, use -trigraphs -Wall -Wno-trigraphs.
5757
5758       -Wundef
5759           Warn if an undefined identifier is evaluated in an "#if" directive.
5760           Such identifiers are replaced with zero.
5761
5762       -Wexpansion-to-defined
5763           Warn whenever defined is encountered in the expansion of a macro
5764           (including the case where the macro is expanded by an #if
5765           directive).  Such usage is not portable.  This warning is also
5766           enabled by -Wpedantic and -Wextra.
5767
5768       -Wunused-macros
5769           Warn about macros defined in the main file that are unused.  A
5770           macro is used if it is expanded or tested for existence at least
5771           once.  The preprocessor also warns if the macro has not been used
5772           at the time it is redefined or undefined.
5773
5774           Built-in macros, macros defined on the command line, and macros
5775           defined in include files are not warned about.
5776
5777           Note: If a macro is actually used, but only used in skipped
5778           conditional blocks, then the preprocessor reports it as unused.  To
5779           avoid the warning in such a case, you might improve the scope of
5780           the macro's definition by, for example, moving it into the first
5781           skipped block.  Alternatively, you could provide a dummy use with
5782           something like:
5783
5784                   #if defined the_macro_causing_the_warning
5785                   #endif
5786
5787       -Wno-endif-labels
5788           Do not warn whenever an "#else" or an "#endif" are followed by
5789           text.  This sometimes happens in older programs with code of the
5790           form
5791
5792                   #if FOO
5793                   ...
5794                   #else FOO
5795                   ...
5796                   #endif FOO
5797
5798           The second and third "FOO" should be in comments.  This warning is
5799           on by default.
5800
5801       -Wbad-function-cast (C and Objective-C only)
5802           Warn when a function call is cast to a non-matching type.  For
5803           example, warn if a call to a function returning an integer type is
5804           cast to a pointer type.
5805
5806       -Wc90-c99-compat (C and Objective-C only)
5807           Warn about features not present in ISO C90, but present in ISO C99.
5808           For instance, warn about use of variable length arrays, "long long"
5809           type, "bool" type, compound literals, designated initializers, and
5810           so on.  This option is independent of the standards mode.  Warnings
5811           are disabled in the expression that follows "__extension__".
5812
5813       -Wc99-c11-compat (C and Objective-C only)
5814           Warn about features not present in ISO C99, but present in ISO C11.
5815           For instance, warn about use of anonymous structures and unions,
5816           "_Atomic" type qualifier, "_Thread_local" storage-class specifier,
5817           "_Alignas" specifier, "Alignof" operator, "_Generic" keyword, and
5818           so on.  This option is independent of the standards mode.  Warnings
5819           are disabled in the expression that follows "__extension__".
5820
5821       -Wc11-c2x-compat (C and Objective-C only)
5822           Warn about features not present in ISO C11, but present in ISO C2X.
5823           For instance, warn about omitting the string in "_Static_assert",
5824           use of [[]] syntax for attributes, use of decimal floating-point
5825           types, and so on.  This option is independent of the standards
5826           mode.  Warnings are disabled in the expression that follows
5827           "__extension__".
5828
5829       -Wc++-compat (C and Objective-C only)
5830           Warn about ISO C constructs that are outside of the common subset
5831           of ISO C and ISO C++, e.g. request for implicit conversion from
5832           "void *" to a pointer to non-"void" type.
5833
5834       -Wc++11-compat (C++ and Objective-C++ only)
5835           Warn about C++ constructs whose meaning differs between ISO C++
5836           1998 and ISO C++ 2011, e.g., identifiers in ISO C++ 1998 that are
5837           keywords in ISO C++ 2011.  This warning turns on -Wnarrowing and is
5838           enabled by -Wall.
5839
5840       -Wc++14-compat (C++ and Objective-C++ only)
5841           Warn about C++ constructs whose meaning differs between ISO C++
5842           2011 and ISO C++ 2014.  This warning is enabled by -Wall.
5843
5844       -Wc++17-compat (C++ and Objective-C++ only)
5845           Warn about C++ constructs whose meaning differs between ISO C++
5846           2014 and ISO C++ 2017.  This warning is enabled by -Wall.
5847
5848       -Wc++20-compat (C++ and Objective-C++ only)
5849           Warn about C++ constructs whose meaning differs between ISO C++
5850           2017 and ISO C++ 2020.  This warning is enabled by -Wall.
5851
5852       -Wcast-qual
5853           Warn whenever a pointer is cast so as to remove a type qualifier
5854           from the target type.  For example, warn if a "const char *" is
5855           cast to an ordinary "char *".
5856
5857           Also warn when making a cast that introduces a type qualifier in an
5858           unsafe way.  For example, casting "char **" to "const char **" is
5859           unsafe, as in this example:
5860
5861                     /* p is char ** value.  */
5862                     const char **q = (const char **) p;
5863                     /* Assignment of readonly string to const char * is OK.  */
5864                     *q = "string";
5865                     /* Now char** pointer points to read-only memory.  */
5866                     **p = 'b';
5867
5868       -Wcast-align
5869           Warn whenever a pointer is cast such that the required alignment of
5870           the target is increased.  For example, warn if a "char *" is cast
5871           to an "int *" on machines where integers can only be accessed at
5872           two- or four-byte boundaries.
5873
5874       -Wcast-align=strict
5875           Warn whenever a pointer is cast such that the required alignment of
5876           the target is increased.  For example, warn if a "char *" is cast
5877           to an "int *" regardless of the target machine.
5878
5879       -Wcast-function-type
5880           Warn when a function pointer is cast to an incompatible function
5881           pointer.  In a cast involving function types with a variable
5882           argument list only the types of initial arguments that are provided
5883           are considered.  Any parameter of pointer-type matches any other
5884           pointer-type.  Any benign differences in integral types are
5885           ignored, like "int" vs. "long" on ILP32 targets.  Likewise type
5886           qualifiers are ignored.  The function type "void (*) (void)" is
5887           special and matches everything, which can be used to suppress this
5888           warning.  In a cast involving pointer to member types this warning
5889           warns whenever the type cast is changing the pointer to member
5890           type.  This warning is enabled by -Wextra.
5891
5892       -Wwrite-strings
5893           When compiling C, give string constants the type "const
5894           char[length]" so that copying the address of one into a non-"const"
5895           "char *" pointer produces a warning.  These warnings help you find
5896           at compile time code that can try to write into a string constant,
5897           but only if you have been very careful about using "const" in
5898           declarations and prototypes.  Otherwise, it is just a nuisance.
5899           This is why we did not make -Wall request these warnings.
5900
5901           When compiling C++, warn about the deprecated conversion from
5902           string literals to "char *".  This warning is enabled by default
5903           for C++ programs.
5904
5905       -Wclobbered
5906           Warn for variables that might be changed by "longjmp" or "vfork".
5907           This warning is also enabled by -Wextra.
5908
5909       -Wconversion
5910           Warn for implicit conversions that may alter a value. This includes
5911           conversions between real and integer, like "abs (x)" when "x" is
5912           "double"; conversions between signed and unsigned, like "unsigned
5913           ui = -1"; and conversions to smaller types, like "sqrtf (M_PI)". Do
5914           not warn for explicit casts like "abs ((int) x)" and "ui =
5915           (unsigned) -1", or if the value is not changed by the conversion
5916           like in "abs (2.0)".  Warnings about conversions between signed and
5917           unsigned integers can be disabled by using -Wno-sign-conversion.
5918
5919           For C++, also warn for confusing overload resolution for user-
5920           defined conversions; and conversions that never use a type
5921           conversion operator: conversions to "void", the same type, a base
5922           class or a reference to them. Warnings about conversions between
5923           signed and unsigned integers are disabled by default in C++ unless
5924           -Wsign-conversion is explicitly enabled.
5925
5926           Warnings about conversion from arithmetic on a small type back to
5927           that type are only given with -Warith-conversion.
5928
5929       -Wdangling-else
5930           Warn about constructions where there may be confusion to which "if"
5931           statement an "else" branch belongs.  Here is an example of such a
5932           case:
5933
5934                   {
5935                     if (a)
5936                       if (b)
5937                         foo ();
5938                     else
5939                       bar ();
5940                   }
5941
5942           In C/C++, every "else" branch belongs to the innermost possible
5943           "if" statement, which in this example is "if (b)".  This is often
5944           not what the programmer expected, as illustrated in the above
5945           example by indentation the programmer chose.  When there is the
5946           potential for this confusion, GCC issues a warning when this flag
5947           is specified.  To eliminate the warning, add explicit braces around
5948           the innermost "if" statement so there is no way the "else" can
5949           belong to the enclosing "if".  The resulting code looks like this:
5950
5951                   {
5952                     if (a)
5953                       {
5954                         if (b)
5955                           foo ();
5956                         else
5957                           bar ();
5958                       }
5959                   }
5960
5961           This warning is enabled by -Wparentheses.
5962
5963       -Wdate-time
5964           Warn when macros "__TIME__", "__DATE__" or "__TIMESTAMP__" are
5965           encountered as they might prevent bit-wise-identical reproducible
5966           compilations.
5967
5968       -Wempty-body
5969           Warn if an empty body occurs in an "if", "else" or "do while"
5970           statement.  This warning is also enabled by -Wextra.
5971
5972       -Wno-endif-labels
5973           Do not warn about stray tokens after "#else" and "#endif".
5974
5975       -Wenum-compare
5976           Warn about a comparison between values of different enumerated
5977           types.  In C++ enumerated type mismatches in conditional
5978           expressions are also diagnosed and the warning is enabled by
5979           default.  In C this warning is enabled by -Wall.
5980
5981       -Wenum-conversion (C, Objective-C only)
5982           Warn when a value of enumerated type is implicitly converted to a
5983           different enumerated type.  This warning is enabled by -Wextra.
5984
5985       -Wjump-misses-init (C, Objective-C only)
5986           Warn if a "goto" statement or a "switch" statement jumps forward
5987           across the initialization of a variable, or jumps backward to a
5988           label after the variable has been initialized.  This only warns
5989           about variables that are initialized when they are declared.  This
5990           warning is only supported for C and Objective-C; in C++ this sort
5991           of branch is an error in any case.
5992
5993           -Wjump-misses-init is included in -Wc++-compat.  It can be disabled
5994           with the -Wno-jump-misses-init option.
5995
5996       -Wsign-compare
5997           Warn when a comparison between signed and unsigned values could
5998           produce an incorrect result when the signed value is converted to
5999           unsigned.  In C++, this warning is also enabled by -Wall.  In C, it
6000           is also enabled by -Wextra.
6001
6002       -Wsign-conversion
6003           Warn for implicit conversions that may change the sign of an
6004           integer value, like assigning a signed integer expression to an
6005           unsigned integer variable. An explicit cast silences the warning.
6006           In C, this option is enabled also by -Wconversion.
6007
6008       -Wfloat-conversion
6009           Warn for implicit conversions that reduce the precision of a real
6010           value.  This includes conversions from real to integer, and from
6011           higher precision real to lower precision real values.  This option
6012           is also enabled by -Wconversion.
6013
6014       -Wno-scalar-storage-order
6015           Do not warn on suspicious constructs involving reverse scalar
6016           storage order.
6017
6018       -Wsizeof-pointer-div
6019           Warn for suspicious divisions of two sizeof expressions that divide
6020           the pointer size by the element size, which is the usual way to
6021           compute the array size but won't work out correctly with pointers.
6022           This warning warns e.g. about "sizeof (ptr) / sizeof (ptr[0])" if
6023           "ptr" is not an array, but a pointer.  This warning is enabled by
6024           -Wall.
6025
6026       -Wsizeof-pointer-memaccess
6027           Warn for suspicious length parameters to certain string and memory
6028           built-in functions if the argument uses "sizeof".  This warning
6029           triggers for example for "memset (ptr, 0, sizeof (ptr));" if "ptr"
6030           is not an array, but a pointer, and suggests a possible fix, or
6031           about "memcpy (&foo, ptr, sizeof (&foo));".
6032           -Wsizeof-pointer-memaccess also warns about calls to bounded string
6033           copy functions like "strncat" or "strncpy" that specify as the
6034           bound a "sizeof" expression of the source array.  For example, in
6035           the following function the call to "strncat" specifies the size of
6036           the source string as the bound.  That is almost certainly a mistake
6037           and so the call is diagnosed.
6038
6039                   void make_file (const char *name)
6040                   {
6041                     char path[PATH_MAX];
6042                     strncpy (path, name, sizeof path - 1);
6043                     strncat (path, ".text", sizeof ".text");
6044                     ...
6045                   }
6046
6047           The -Wsizeof-pointer-memaccess option is enabled by -Wall.
6048
6049       -Wno-sizeof-array-argument
6050           Do not warn when the "sizeof" operator is applied to a parameter
6051           that is declared as an array in a function definition.  This
6052           warning is enabled by default for C and C++ programs.
6053
6054       -Wmemset-elt-size
6055           Warn for suspicious calls to the "memset" built-in function, if the
6056           first argument references an array, and the third argument is a
6057           number equal to the number of elements, but not equal to the size
6058           of the array in memory.  This indicates that the user has omitted a
6059           multiplication by the element size.  This warning is enabled by
6060           -Wall.
6061
6062       -Wmemset-transposed-args
6063           Warn for suspicious calls to the "memset" built-in function where
6064           the second argument is not zero and the third argument is zero.
6065           For example, the call "memset (buf, sizeof buf, 0)" is diagnosed
6066           because "memset (buf, 0, sizeof buf)" was meant instead.  The
6067           diagnostic is only emitted if the third argument is a literal zero.
6068           Otherwise, if it is an expression that is folded to zero, or a cast
6069           of zero to some type, it is far less likely that the arguments have
6070           been mistakenly transposed and no warning is emitted.  This warning
6071           is enabled by -Wall.
6072
6073       -Waddress
6074           Warn about suspicious uses of memory addresses. These include using
6075           the address of a function in a conditional expression, such as
6076           "void func(void); if (func)", and comparisons against the memory
6077           address of a string literal, such as "if (x == "abc")".  Such uses
6078           typically indicate a programmer error: the address of a function
6079           always evaluates to true, so their use in a conditional usually
6080           indicate that the programmer forgot the parentheses in a function
6081           call; and comparisons against string literals result in unspecified
6082           behavior and are not portable in C, so they usually indicate that
6083           the programmer intended to use "strcmp".  This warning is enabled
6084           by -Wall.
6085
6086       -Wno-address-of-packed-member
6087           Do not warn when the address of packed member of struct or union is
6088           taken, which usually results in an unaligned pointer value.  This
6089           is enabled by default.
6090
6091       -Wlogical-op
6092           Warn about suspicious uses of logical operators in expressions.
6093           This includes using logical operators in contexts where a bit-wise
6094           operator is likely to be expected.  Also warns when the operands of
6095           a logical operator are the same:
6096
6097                   extern int a;
6098                   if (a < 0 && a < 0) { ... }
6099
6100       -Wlogical-not-parentheses
6101           Warn about logical not used on the left hand side operand of a
6102           comparison.  This option does not warn if the right operand is
6103           considered to be a boolean expression.  Its purpose is to detect
6104           suspicious code like the following:
6105
6106                   int a;
6107                   ...
6108                   if (!a > 1) { ... }
6109
6110           It is possible to suppress the warning by wrapping the LHS into
6111           parentheses:
6112
6113                   if ((!a) > 1) { ... }
6114
6115           This warning is enabled by -Wall.
6116
6117       -Waggregate-return
6118           Warn if any functions that return structures or unions are defined
6119           or called.  (In languages where you can return an array, this also
6120           elicits a warning.)
6121
6122       -Wno-aggressive-loop-optimizations
6123           Warn if in a loop with constant number of iterations the compiler
6124           detects undefined behavior in some statement during one or more of
6125           the iterations.
6126
6127       -Wno-attributes
6128           Do not warn if an unexpected "__attribute__" is used, such as
6129           unrecognized attributes, function attributes applied to variables,
6130           etc.  This does not stop errors for incorrect use of supported
6131           attributes.
6132
6133       -Wno-builtin-declaration-mismatch
6134           Warn if a built-in function is declared with an incompatible
6135           signature or as a non-function, or when a built-in function
6136           declared with a type that does not include a prototype is called
6137           with arguments whose promoted types do not match those expected by
6138           the function.  When -Wextra is specified, also warn when a built-in
6139           function that takes arguments is declared without a prototype.  The
6140           -Wbuiltin-declaration-mismatch warning is enabled by default.  To
6141           avoid the warning include the appropriate header to bring the
6142           prototypes of built-in functions into scope.
6143
6144           For example, the call to "memset" below is diagnosed by the warning
6145           because the function expects a value of type "size_t" as its
6146           argument but the type of 32 is "int".  With -Wextra, the
6147           declaration of the function is diagnosed as well.
6148
6149                   extern void* memset ();
6150                   void f (void *d)
6151                   {
6152                     memset (d, '\0', 32);
6153                   }
6154
6155       -Wno-builtin-macro-redefined
6156           Do not warn if certain built-in macros are redefined.  This
6157           suppresses warnings for redefinition of "__TIMESTAMP__",
6158           "__TIME__", "__DATE__", "__FILE__", and "__BASE_FILE__".
6159
6160       -Wstrict-prototypes (C and Objective-C only)
6161           Warn if a function is declared or defined without specifying the
6162           argument types.  (An old-style function definition is permitted
6163           without a warning if preceded by a declaration that specifies the
6164           argument types.)
6165
6166       -Wold-style-declaration (C and Objective-C only)
6167           Warn for obsolescent usages, according to the C Standard, in a
6168           declaration. For example, warn if storage-class specifiers like
6169           "static" are not the first things in a declaration.  This warning
6170           is also enabled by -Wextra.
6171
6172       -Wold-style-definition (C and Objective-C only)
6173           Warn if an old-style function definition is used.  A warning is
6174           given even if there is a previous prototype.  A definition using ()
6175           is not considered an old-style definition in C2X mode, because it
6176           is equivalent to (void) in that case, but is considered an old-
6177           style definition for older standards.
6178
6179       -Wmissing-parameter-type (C and Objective-C only)
6180           A function parameter is declared without a type specifier in
6181           K&R-style functions:
6182
6183                   void foo(bar) { }
6184
6185           This warning is also enabled by -Wextra.
6186
6187       -Wmissing-prototypes (C and Objective-C only)
6188           Warn if a global function is defined without a previous prototype
6189           declaration.  This warning is issued even if the definition itself
6190           provides a prototype.  Use this option to detect global functions
6191           that do not have a matching prototype declaration in a header file.
6192           This option is not valid for C++ because all function declarations
6193           provide prototypes and a non-matching declaration declares an
6194           overload rather than conflict with an earlier declaration.  Use
6195           -Wmissing-declarations to detect missing declarations in C++.
6196
6197       -Wmissing-declarations
6198           Warn if a global function is defined without a previous
6199           declaration.  Do so even if the definition itself provides a
6200           prototype.  Use this option to detect global functions that are not
6201           declared in header files.  In C, no warnings are issued for
6202           functions with previous non-prototype declarations; use
6203           -Wmissing-prototypes to detect missing prototypes.  In C++, no
6204           warnings are issued for function templates, or for inline
6205           functions, or for functions in anonymous namespaces.
6206
6207       -Wmissing-field-initializers
6208           Warn if a structure's initializer has some fields missing.  For
6209           example, the following code causes such a warning, because "x.h" is
6210           implicitly zero:
6211
6212                   struct s { int f, g, h; };
6213                   struct s x = { 3, 4 };
6214
6215           This option does not warn about designated initializers, so the
6216           following modification does not trigger a warning:
6217
6218                   struct s { int f, g, h; };
6219                   struct s x = { .f = 3, .g = 4 };
6220
6221           In C this option does not warn about the universal zero initializer
6222           { 0 }:
6223
6224                   struct s { int f, g, h; };
6225                   struct s x = { 0 };
6226
6227           Likewise, in C++ this option does not warn about the empty { }
6228           initializer, for example:
6229
6230                   struct s { int f, g, h; };
6231                   s x = { };
6232
6233           This warning is included in -Wextra.  To get other -Wextra warnings
6234           without this one, use -Wextra -Wno-missing-field-initializers.
6235
6236       -Wno-multichar
6237           Do not warn if a multicharacter constant ('FOOF') is used.  Usually
6238           they indicate a typo in the user's code, as they have
6239           implementation-defined values, and should not be used in portable
6240           code.
6241
6242       -Wnormalized=[none|id|nfc|nfkc]
6243           In ISO C and ISO C++, two identifiers are different if they are
6244           different sequences of characters.  However, sometimes when
6245           characters outside the basic ASCII character set are used, you can
6246           have two different character sequences that look the same.  To
6247           avoid confusion, the ISO 10646 standard sets out some normalization
6248           rules which when applied ensure that two sequences that look the
6249           same are turned into the same sequence.  GCC can warn you if you
6250           are using identifiers that have not been normalized; this option
6251           controls that warning.
6252
6253           There are four levels of warning supported by GCC.  The default is
6254           -Wnormalized=nfc, which warns about any identifier that is not in
6255           the ISO 10646 "C" normalized form, NFC.  NFC is the recommended
6256           form for most uses.  It is equivalent to -Wnormalized.
6257
6258           Unfortunately, there are some characters allowed in identifiers by
6259           ISO C and ISO C++ that, when turned into NFC, are not allowed in
6260           identifiers.  That is, there's no way to use these symbols in
6261           portable ISO C or C++ and have all your identifiers in NFC.
6262           -Wnormalized=id suppresses the warning for these characters.  It is
6263           hoped that future versions of the standards involved will correct
6264           this, which is why this option is not the default.
6265
6266           You can switch the warning off for all characters by writing
6267           -Wnormalized=none or -Wno-normalized.  You should only do this if
6268           you are using some other normalization scheme (like "D"), because
6269           otherwise you can easily create bugs that are literally impossible
6270           to see.
6271
6272           Some characters in ISO 10646 have distinct meanings but look
6273           identical in some fonts or display methodologies, especially once
6274           formatting has been applied.  For instance "\u207F", "SUPERSCRIPT
6275           LATIN SMALL LETTER N", displays just like a regular "n" that has
6276           been placed in a superscript.  ISO 10646 defines the NFKC
6277           normalization scheme to convert all these into a standard form as
6278           well, and GCC warns if your code is not in NFKC if you use
6279           -Wnormalized=nfkc.  This warning is comparable to warning about
6280           every identifier that contains the letter O because it might be
6281           confused with the digit 0, and so is not the default, but may be
6282           useful as a local coding convention if the programming environment
6283           cannot be fixed to display these characters distinctly.
6284
6285       -Wno-attribute-warning
6286           Do not warn about usage of functions declared with "warning"
6287           attribute.  By default, this warning is enabled.
6288           -Wno-attribute-warning can be used to disable the warning or
6289           -Wno-error=attribute-warning can be used to disable the error when
6290           compiled with -Werror flag.
6291
6292       -Wno-deprecated
6293           Do not warn about usage of deprecated features.
6294
6295       -Wno-deprecated-declarations
6296           Do not warn about uses of functions, variables, and types marked as
6297           deprecated by using the "deprecated" attribute.
6298
6299       -Wno-overflow
6300           Do not warn about compile-time overflow in constant expressions.
6301
6302       -Wno-odr
6303           Warn about One Definition Rule violations during link-time
6304           optimization.  Enabled by default.
6305
6306       -Wopenmp-simd
6307           Warn if the vectorizer cost model overrides the OpenMP simd
6308           directive set by user.  The -fsimd-cost-model=unlimited option can
6309           be used to relax the cost model.
6310
6311       -Woverride-init (C and Objective-C only)
6312           Warn if an initialized field without side effects is overridden
6313           when using designated initializers.
6314
6315           This warning is included in -Wextra.  To get other -Wextra warnings
6316           without this one, use -Wextra -Wno-override-init.
6317
6318       -Wno-override-init-side-effects (C and Objective-C only)
6319           Do not warn if an initialized field with side effects is overridden
6320           when using designated initializers.  This warning is enabled by
6321           default.
6322
6323       -Wpacked
6324           Warn if a structure is given the packed attribute, but the packed
6325           attribute has no effect on the layout or size of the structure.
6326           Such structures may be mis-aligned for little benefit.  For
6327           instance, in this code, the variable "f.x" in "struct bar" is
6328           misaligned even though "struct bar" does not itself have the packed
6329           attribute:
6330
6331                   struct foo {
6332                     int x;
6333                     char a, b, c, d;
6334                   } __attribute__((packed));
6335                   struct bar {
6336                     char z;
6337                     struct foo f;
6338                   };
6339
6340       -Wnopacked-bitfield-compat
6341           The 4.1, 4.2 and 4.3 series of GCC ignore the "packed" attribute on
6342           bit-fields of type "char".  This was fixed in GCC 4.4 but the
6343           change can lead to differences in the structure layout.  GCC
6344           informs you when the offset of such a field has changed in GCC 4.4.
6345           For example there is no longer a 4-bit padding between field "a"
6346           and "b" in this structure:
6347
6348                   struct foo
6349                   {
6350                     char a:4;
6351                     char b:8;
6352                   } __attribute__ ((packed));
6353
6354           This warning is enabled by default.  Use
6355           -Wno-packed-bitfield-compat to disable this warning.
6356
6357       -Wpacked-not-aligned (C, C++, Objective-C and Objective-C++ only)
6358           Warn if a structure field with explicitly specified alignment in a
6359           packed struct or union is misaligned.  For example, a warning will
6360           be issued on "struct S", like, "warning: alignment 1 of 'struct S'
6361           is less than 8", in this code:
6362
6363                   struct __attribute__ ((aligned (8))) S8 { char a[8]; };
6364                   struct __attribute__ ((packed)) S {
6365                     struct S8 s8;
6366                   };
6367
6368           This warning is enabled by -Wall.
6369
6370       -Wpadded
6371           Warn if padding is included in a structure, either to align an
6372           element of the structure or to align the whole structure.
6373           Sometimes when this happens it is possible to rearrange the fields
6374           of the structure to reduce the padding and so make the structure
6375           smaller.
6376
6377       -Wredundant-decls
6378           Warn if anything is declared more than once in the same scope, even
6379           in cases where multiple declaration is valid and changes nothing.
6380
6381       -Wrestrict
6382           Warn when an object referenced by a "restrict"-qualified parameter
6383           (or, in C++, a "__restrict"-qualified parameter) is aliased by
6384           another argument, or when copies between such objects overlap.  For
6385           example, the call to the "strcpy" function below attempts to
6386           truncate the string by replacing its initial characters with the
6387           last four.  However, because the call writes the terminating NUL
6388           into "a[4]", the copies overlap and the call is diagnosed.
6389
6390                   void foo (void)
6391                   {
6392                     char a[] = "abcd1234";
6393                     strcpy (a, a + 4);
6394                     ...
6395                   }
6396
6397           The -Wrestrict option detects some instances of simple overlap even
6398           without optimization but works best at -O2 and above.  It is
6399           included in -Wall.
6400
6401       -Wnested-externs (C and Objective-C only)
6402           Warn if an "extern" declaration is encountered within a function.
6403
6404       -Winline
6405           Warn if a function that is declared as inline cannot be inlined.
6406           Even with this option, the compiler does not warn about failures to
6407           inline functions declared in system headers.
6408
6409           The compiler uses a variety of heuristics to determine whether or
6410           not to inline a function.  For example, the compiler takes into
6411           account the size of the function being inlined and the amount of
6412           inlining that has already been done in the current function.
6413           Therefore, seemingly insignificant changes in the source program
6414           can cause the warnings produced by -Winline to appear or disappear.
6415
6416       -Wint-in-bool-context
6417           Warn for suspicious use of integer values where boolean values are
6418           expected, such as conditional expressions (?:) using non-boolean
6419           integer constants in boolean context, like "if (a <= b ? 2 : 3)".
6420           Or left shifting of signed integers in boolean context, like "for
6421           (a = 0; 1 << a; a++);".  Likewise for all kinds of multiplications
6422           regardless of the data type.  This warning is enabled by -Wall.
6423
6424       -Wno-int-to-pointer-cast
6425           Suppress warnings from casts to pointer type of an integer of a
6426           different size. In C++, casting to a pointer type of smaller size
6427           is an error. Wint-to-pointer-cast is enabled by default.
6428
6429       -Wno-pointer-to-int-cast (C and Objective-C only)
6430           Suppress warnings from casts from a pointer to an integer type of a
6431           different size.
6432
6433       -Winvalid-pch
6434           Warn if a precompiled header is found in the search path but cannot
6435           be used.
6436
6437       -Wlong-long
6438           Warn if "long long" type is used.  This is enabled by either
6439           -Wpedantic or -Wtraditional in ISO C90 and C++98 modes.  To inhibit
6440           the warning messages, use -Wno-long-long.
6441
6442       -Wvariadic-macros
6443           Warn if variadic macros are used in ISO C90 mode, or if the GNU
6444           alternate syntax is used in ISO C99 mode.  This is enabled by
6445           either -Wpedantic or -Wtraditional.  To inhibit the warning
6446           messages, use -Wno-variadic-macros.
6447
6448       -Wno-varargs
6449           Do not warn upon questionable usage of the macros used to handle
6450           variable arguments like "va_start".  These warnings are enabled by
6451           default.
6452
6453       -Wvector-operation-performance
6454           Warn if vector operation is not implemented via SIMD capabilities
6455           of the architecture.  Mainly useful for the performance tuning.
6456           Vector operation can be implemented "piecewise", which means that
6457           the scalar operation is performed on every vector element; "in
6458           parallel", which means that the vector operation is implemented
6459           using scalars of wider type, which normally is more performance
6460           efficient; and "as a single scalar", which means that vector fits
6461           into a scalar type.
6462
6463       -Wvla
6464           Warn if a variable-length array is used in the code.  -Wno-vla
6465           prevents the -Wpedantic warning of the variable-length array.
6466
6467       -Wvla-larger-than=byte-size
6468           If this option is used, the compiler warns for declarations of
6469           variable-length arrays whose size is either unbounded, or bounded
6470           by an argument that allows the array size to exceed byte-size
6471           bytes.  This is similar to how -Walloca-larger-than=byte-size
6472           works, but with variable-length arrays.
6473
6474           Note that GCC may optimize small variable-length arrays of a known
6475           value into plain arrays, so this warning may not get triggered for
6476           such arrays.
6477
6478           -Wvla-larger-than=PTRDIFF_MAX is enabled by default but is
6479           typically only effective when -ftree-vrp is active (default for -O2
6480           and above).
6481
6482           See also -Walloca-larger-than=byte-size.
6483
6484       -Wno-vla-larger-than
6485           Disable -Wvla-larger-than= warnings.  The option is equivalent to
6486           -Wvla-larger-than=SIZE_MAX or larger.
6487
6488       -Wvolatile-register-var
6489           Warn if a register variable is declared volatile.  The volatile
6490           modifier does not inhibit all optimizations that may eliminate
6491           reads and/or writes to register variables.  This warning is enabled
6492           by -Wall.
6493
6494       -Wdisabled-optimization
6495           Warn if a requested optimization pass is disabled.  This warning
6496           does not generally indicate that there is anything wrong with your
6497           code; it merely indicates that GCC's optimizers are unable to
6498           handle the code effectively.  Often, the problem is that your code
6499           is too big or too complex; GCC refuses to optimize programs when
6500           the optimization itself is likely to take inordinate amounts of
6501           time.
6502
6503       -Wpointer-sign (C and Objective-C only)
6504           Warn for pointer argument passing or assignment with different
6505           signedness.  This option is only supported for C and Objective-C.
6506           It is implied by -Wall and by -Wpedantic, which can be disabled
6507           with -Wno-pointer-sign.
6508
6509       -Wstack-protector
6510           This option is only active when -fstack-protector is active.  It
6511           warns about functions that are not protected against stack
6512           smashing.
6513
6514       -Woverlength-strings
6515           Warn about string constants that are longer than the "minimum
6516           maximum" length specified in the C standard.  Modern compilers
6517           generally allow string constants that are much longer than the
6518           standard's minimum limit, but very portable programs should avoid
6519           using longer strings.
6520
6521           The limit applies after string constant concatenation, and does not
6522           count the trailing NUL.  In C90, the limit was 509 characters; in
6523           C99, it was raised to 4095.  C++98 does not specify a normative
6524           minimum maximum, so we do not diagnose overlength strings in C++.
6525
6526           This option is implied by -Wpedantic, and can be disabled with
6527           -Wno-overlength-strings.
6528
6529       -Wunsuffixed-float-constants (C and Objective-C only)
6530           Issue a warning for any floating constant that does not have a
6531           suffix.  When used together with -Wsystem-headers it warns about
6532           such constants in system header files.  This can be useful when
6533           preparing code to use with the "FLOAT_CONST_DECIMAL64" pragma from
6534           the decimal floating-point extension to C99.
6535
6536       -Wno-lto-type-mismatch
6537           During the link-time optimization, do not warn about type
6538           mismatches in global declarations from different compilation units.
6539           Requires -flto to be enabled.  Enabled by default.
6540
6541       -Wno-designated-init (C and Objective-C only)
6542           Suppress warnings when a positional initializer is used to
6543           initialize a structure that has been marked with the
6544           "designated_init" attribute.
6545
6546       -Wno-hsa
6547           Do not warn when HSAIL cannot be emitted for the compiled function
6548           or OpenMP construct.  These warnings are enabled by default.
6549
6550   Options That Control Static Analysis
6551       -fanalyzer
6552           This option enables an static analysis of program flow which looks
6553           for "interesting" interprocedural paths through the code, and
6554           issues warnings for problems found on them.
6555
6556           This analysis is much more expensive than other GCC warnings.
6557
6558           Enabling this option effectively enables the following warnings:
6559
6560           -Wanalyzer-double-fclose -Wanalyzer-double-free
6561           -Wanalyzer-exposure-through-output-file -Wanalyzer-file-leak
6562           -Wanalyzer-free-of-non-heap -Wanalyzer-malloc-leak
6563           -Wanalyzer-possible-null-argument
6564           -Wanalyzer-possible-null-dereference -Wanalyzer-null-argument
6565           -Wanalyzer-null-dereference -Wanalyzer-stale-setjmp-buffer
6566           -Wanalyzer-tainted-array-index
6567           -Wanalyzer-unsafe-call-within-signal-handler
6568           -Wanalyzer-use-after-free
6569           -Wanalyzer-use-of-pointer-in-stale-stack-frame
6570
6571           This option is only available if GCC was configured with analyzer
6572           support enabled.
6573
6574       -Wanalyzer-too-complex
6575           If -fanalyzer is enabled, the analyzer uses various heuristics to
6576           attempt to explore the control flow and data flow in the program,
6577           but these can be defeated by sufficiently complicated code.
6578
6579           By default, the analysis silently stops if the code is too
6580           complicated for the analyzer to fully explore and it reaches an
6581           internal limit.  The -Wanalyzer-too-complex option warns if this
6582           occurs.
6583
6584       -Wno-analyzer-double-fclose
6585           This warning requires -fanalyzer, which enables it; use
6586           -Wno-analyzer-double-fclose to disable it.
6587
6588           This diagnostic warns for paths through the code in which a "FILE
6589           *" can have "fclose" called on it more than once.
6590
6591       -Wno-analyzer-double-free
6592           This warning requires -fanalyzer, which enables it; use
6593           -Wno-analyzer-double-free to disable it.
6594
6595           This diagnostic warns for paths through the code in which a pointer
6596           can have "free" called on it more than once.
6597
6598       -Wno-analyzer-exposure-through-output-file
6599           This warning requires -fanalyzer, which enables it; use
6600           -Wno-analyzer-exposure-through-output-file to disable it.
6601
6602           This diagnostic warns for paths through the code in which a
6603           security-sensitive value is written to an output file (such as
6604           writing a password to a log file).
6605
6606       -Wno-analyzer-file-leak
6607           This warning requires -fanalyzer, which enables it; use
6608           -Wno-analyzer-file-leak to disable it.
6609
6610           This diagnostic warns for paths through the code in which a
6611           "<stdio.h>" "FILE *" stream object is leaked.
6612
6613       -Wno-analyzer-free-of-non-heap
6614           This warning requires -fanalyzer, which enables it; use
6615           -Wno-analyzer-free-of-non-heap to disable it.
6616
6617           This diagnostic warns for paths through the code in which "free" is
6618           called on a non-heap pointer (e.g. an on-stack buffer, or a
6619           global).
6620
6621       -Wno-analyzer-malloc-leak
6622           This warning requires -fanalyzer, which enables it; use
6623           -Wno-analyzer-malloc-leak to disable it.
6624
6625           This diagnostic warns for paths through the code in which a pointer
6626           allocated via "malloc" is leaked.
6627
6628       -Wno-analyzer-possible-null-argument
6629           This warning requires -fanalyzer, which enables it; use
6630           -Wno-analyzer-possible-null-argument to disable it.
6631
6632           This diagnostic warns for paths through the code in which a
6633           possibly-NULL value is passed to a function argument marked with
6634           "__attribute__((nonnull))" as requiring a non-NULL value.
6635
6636       -Wno-analyzer-possible-null-dereference
6637           This warning requires -fanalyzer, which enables it; use
6638           -Wno-analyzer-possible-null-dereference to disable it.
6639
6640           This diagnostic warns for paths through the code in which a
6641           possibly-NULL value is dereferenced.
6642
6643       -Wno-analyzer-null-argument
6644           This warning requires -fanalyzer, which enables it; use
6645           -Wno-analyzer-null-argument to disable it.
6646
6647           This diagnostic warns for paths through the code in which a value
6648           known to be NULL is passed to a function argument marked with
6649           "__attribute__((nonnull))" as requiring a non-NULL value.
6650
6651       -Wno-analyzer-null-dereference
6652           This warning requires -fanalyzer, which enables it; use
6653           -Wno-analyzer-null-dereference to disable it.
6654
6655           This diagnostic warns for paths through the code in which a value
6656           known to be NULL is dereferenced.
6657
6658       -Wno-analyzer-stale-setjmp-buffer
6659           This warning requires -fanalyzer, which enables it; use
6660           -Wno-analyzer-stale-setjmp-buffer to disable it.
6661
6662           This diagnostic warns for paths through the code in which "longjmp"
6663           is called to rewind to a "jmp_buf" relating to a "setjmp" call in a
6664           function that has returned.
6665
6666           When "setjmp" is called on a "jmp_buf" to record a rewind location,
6667           it records the stack frame.  The stack frame becomes invalid when
6668           the function containing the "setjmp" call returns.  Attempting to
6669           rewind to it via "longjmp" would reference a stack frame that no
6670           longer exists, and likely lead to a crash (or worse).
6671
6672       -Wno-analyzer-tainted-array-index
6673           This warning requires both -fanalyzer and -fanalyzer-checker=taint
6674           to enable it; use -Wno-analyzer-tainted-array-index to disable it.
6675
6676           This diagnostic warns for paths through the code in which a value
6677           that could be under an attacker's control is used as the index of
6678           an array access without being sanitized.
6679
6680       -Wno-analyzer-unsafe-call-within-signal-handler
6681           This warning requires -fanalyzer, which enables it; use
6682           -Wno-analyzer-unsafe-call-within-signal-handler to disable it.
6683
6684           This diagnostic warns for paths through the code in which a
6685           function known to be async-signal-unsafe (such as "fprintf") is
6686           called from a signal handler.
6687
6688       -Wno-analyzer-use-after-free
6689           This warning requires -fanalyzer, which enables it; use
6690           -Wno-analyzer-use-after-free to disable it.
6691
6692           This diagnostic warns for paths through the code in which a pointer
6693           is used after "free" is called on it.
6694
6695       -Wno-analyzer-use-of-pointer-in-stale-stack-frame
6696           This warning requires -fanalyzer, which enables it; use
6697           -Wno-analyzer-use-of-pointer-in-stale-stack-frame to disable it.
6698
6699           This diagnostic warns for paths through the code in which a pointer
6700           is dereferenced that points to a variable in a stale stack frame.
6701
6702       Pertinent parameters for controlling the exploration are: --param
6703       analyzer-bb-explosion-factor=value, --param
6704       analyzer-max-enodes-per-program-point=value, --param
6705       analyzer-max-recursion-depth=value, and --param
6706       analyzer-min-snodes-for-call-summary=value.
6707
6708       The following options control the analyzer.
6709
6710       -fanalyzer-call-summaries
6711           Simplify interprocedural analysis by computing the effect of
6712           certain calls, rather than exploring all paths through the function
6713           from callsite to each possible return.
6714
6715           If enabled, call summaries are only used for functions with more
6716           than one call site, and that are sufficiently complicated (as per
6717           --param analyzer-min-snodes-for-call-summary=value).
6718
6719       -fanalyzer-checker=name
6720           Restrict the analyzer to run just the named checker, and enable it.
6721
6722           Some checkers are disabled by default (even with -fanalyzer), such
6723           as the "taint" checker that implements
6724           -Wanalyzer-tainted-array-index, and this option is required to
6725           enable them.
6726
6727       -fanalyzer-fine-grained
6728           This option is intended for analyzer developers.
6729
6730           Internally the analyzer builds an "exploded graph" that combines
6731           control flow graphs with data flow information.
6732
6733           By default, an edge in this graph can contain the effects of a run
6734           of multiple statements within a basic block.  With
6735           -fanalyzer-fine-grained, each statement gets its own edge.
6736
6737       -fanalyzer-show-duplicate-count
6738           This option is intended for analyzer developers: if multiple
6739           diagnostics have been detected as being duplicates of each other,
6740           it emits a note when reporting the best diagnostic, giving the
6741           number of additional diagnostics that were suppressed by the
6742           deduplication logic.
6743
6744       -fno-analyzer-state-merge
6745           This option is intended for analyzer developers.
6746
6747           By default the analyzer attempts to simplify analysis by merging
6748           sufficiently similar states at each program point as it builds its
6749           "exploded graph".  With -fno-analyzer-state-merge this merging can
6750           be suppressed, for debugging state-handling issues.
6751
6752       -fno-analyzer-state-purge
6753           This option is intended for analyzer developers.
6754
6755           By default the analyzer attempts to simplify analysis by purging
6756           aspects of state at a program point that appear to no longer be
6757           relevant e.g. the values of locals that aren't accessed later in
6758           the function and which aren't relevant to leak analysis.
6759
6760           With -fno-analyzer-state-purge this purging of state can be
6761           suppressed, for debugging state-handling issues.
6762
6763       -fanalyzer-transitivity
6764           This option enables transitivity of constraints within the
6765           analyzer.
6766
6767       -fanalyzer-verbose-edges
6768           This option is intended for analyzer developers.  It enables more
6769           verbose, lower-level detail in the descriptions of control flow
6770           within diagnostic paths.
6771
6772       -fanalyzer-verbose-state-changes
6773           This option is intended for analyzer developers.  It enables more
6774           verbose, lower-level detail in the descriptions of events relating
6775           to state machines within diagnostic paths.
6776
6777       -fanalyzer-verbosity=level
6778           This option controls the complexity of the control flow paths that
6779           are emitted for analyzer diagnostics.
6780
6781           The level can be one of:
6782
6783           0   At this level, interprocedural call and return events are
6784               displayed, along with the most pertinent state-change events
6785               relating to a diagnostic.  For example, for a double-"free"
6786               diagnostic, both calls to "free" will be shown.
6787
6788           1   As per the previous level, but also show events for the entry
6789               to each function.
6790
6791           2   As per the previous level, but also show events relating to
6792               control flow that are significant to triggering the issue (e.g.
6793               "true path taken" at a conditional).
6794
6795               This level is the default.
6796
6797           3   As per the previous level, but show all control flow events,
6798               not just significant ones.
6799
6800           4   This level is intended for analyzer developers; it adds various
6801               other events intended for debugging the analyzer.
6802
6803       -fdump-analyzer
6804           Dump internal details about what the analyzer is doing to
6805           file.analyzer.txt.  This option is overridden by
6806           -fdump-analyzer-stderr.
6807
6808       -fdump-analyzer-stderr
6809           Dump internal details about what the analyzer is doing to stderr.
6810           This option overrides -fdump-analyzer.
6811
6812       -fdump-analyzer-callgraph
6813           Dump a representation of the call graph suitable for viewing with
6814           GraphViz to file.callgraph.dot.
6815
6816       -fdump-analyzer-exploded-graph
6817           Dump a representation of the "exploded graph" suitable for viewing
6818           with GraphViz to file.eg.dot.  Nodes are color-coded based on
6819           state-machine states to emphasize state changes.
6820
6821       -fdump-analyzer-exploded-nodes
6822           Emit diagnostics showing where nodes in the "exploded graph" are in
6823           relation to the program source.
6824
6825       -fdump-analyzer-exploded-nodes-2
6826           Dump a textual representation of the "exploded graph" to
6827           file.eg.txt.
6828
6829       -fdump-analyzer-exploded-nodes-3
6830           Dump a textual representation of the "exploded graph" to one dump
6831           file per node, to file.eg-id.txt.  This is typically a large number
6832           of dump files.
6833
6834       -fdump-analyzer-state-purge
6835           As per -fdump-analyzer-supergraph, dump a representation of the
6836           "supergraph" suitable for viewing with GraphViz, but annotate the
6837           graph with information on what state will be purged at each node.
6838           The graph is written to file.state-purge.dot.
6839
6840       -fdump-analyzer-supergraph
6841           Dump representations of the "supergraph" suitable for viewing with
6842           GraphViz to file.supergraph.dot and to file.supergraph-eg.dot.
6843           These show all of the control flow graphs in the program, with
6844           interprocedural edges for calls and returns.  The second dump
6845           contains annotations showing nodes in the "exploded graph" and
6846           diagnostics associated with them.
6847
6848   Options for Debugging Your Program
6849       To tell GCC to emit extra information for use by a debugger, in almost
6850       all cases you need only to add -g to your other options.
6851
6852       GCC allows you to use -g with -O.  The shortcuts taken by optimized
6853       code may occasionally be surprising: some variables you declared may
6854       not exist at all; flow of control may briefly move where you did not
6855       expect it; some statements may not be executed because they compute
6856       constant results or their values are already at hand; some statements
6857       may execute in different places because they have been moved out of
6858       loops.  Nevertheless it is possible to debug optimized output.  This
6859       makes it reasonable to use the optimizer for programs that might have
6860       bugs.
6861
6862       If you are not using some other optimization option, consider using -Og
6863       with -g.  With no -O option at all, some compiler passes that collect
6864       information useful for debugging do not run at all, so that -Og may
6865       result in a better debugging experience.
6866
6867       -g  Produce debugging information in the operating system's native
6868           format (stabs, COFF, XCOFF, or DWARF).  GDB can work with this
6869           debugging information.
6870
6871           On most systems that use stabs format, -g enables use of extra
6872           debugging information that only GDB can use; this extra information
6873           makes debugging work better in GDB but probably makes other
6874           debuggers crash or refuse to read the program.  If you want to
6875           control for certain whether to generate the extra information, use
6876           -gstabs+, -gstabs, -gxcoff+, -gxcoff, or -gvms (see below).
6877
6878       -ggdb
6879           Produce debugging information for use by GDB.  This means to use
6880           the most expressive format available (DWARF, stabs, or the native
6881           format if neither of those are supported), including GDB extensions
6882           if at all possible.
6883
6884       -gdwarf
6885       -gdwarf-version
6886           Produce debugging information in DWARF format (if that is
6887           supported).  The value of version may be either 2, 3, 4 or 5; the
6888           default version for most targets is 4.  DWARF Version 5 is only
6889           experimental.
6890
6891           Note that with DWARF Version 2, some ports require and always use
6892           some non-conflicting DWARF 3 extensions in the unwind tables.
6893
6894           Version 4 may require GDB 7.0 and -fvar-tracking-assignments for
6895           maximum benefit.
6896
6897           GCC no longer supports DWARF Version 1, which is substantially
6898           different than Version 2 and later.  For historical reasons, some
6899           other DWARF-related options such as -fno-dwarf2-cfi-asm) retain a
6900           reference to DWARF Version 2 in their names, but apply to all
6901           currently-supported versions of DWARF.
6902
6903       -gstabs
6904           Produce debugging information in stabs format (if that is
6905           supported), without GDB extensions.  This is the format used by DBX
6906           on most BSD systems.  On MIPS, Alpha and System V Release 4 systems
6907           this option produces stabs debugging output that is not understood
6908           by DBX.  On System V Release 4 systems this option requires the GNU
6909           assembler.
6910
6911       -gstabs+
6912           Produce debugging information in stabs format (if that is
6913           supported), using GNU extensions understood only by the GNU
6914           debugger (GDB).  The use of these extensions is likely to make
6915           other debuggers crash or refuse to read the program.
6916
6917       -gxcoff
6918           Produce debugging information in XCOFF format (if that is
6919           supported).  This is the format used by the DBX debugger on IBM
6920           RS/6000 systems.
6921
6922       -gxcoff+
6923           Produce debugging information in XCOFF format (if that is
6924           supported), using GNU extensions understood only by the GNU
6925           debugger (GDB).  The use of these extensions is likely to make
6926           other debuggers crash or refuse to read the program, and may cause
6927           assemblers other than the GNU assembler (GAS) to fail with an
6928           error.
6929
6930       -gvms
6931           Produce debugging information in Alpha/VMS debug format (if that is
6932           supported).  This is the format used by DEBUG on Alpha/VMS systems.
6933
6934       -glevel
6935       -ggdblevel
6936       -gstabslevel
6937       -gxcofflevel
6938       -gvmslevel
6939           Request debugging information and also use level to specify how
6940           much information.  The default level is 2.
6941
6942           Level 0 produces no debug information at all.  Thus, -g0 negates
6943           -g.
6944
6945           Level 1 produces minimal information, enough for making backtraces
6946           in parts of the program that you don't plan to debug.  This
6947           includes descriptions of functions and external variables, and line
6948           number tables, but no information about local variables.
6949
6950           Level 3 includes extra information, such as all the macro
6951           definitions present in the program.  Some debuggers support macro
6952           expansion when you use -g3.
6953
6954           If you use multiple -g options, with or without level numbers, the
6955           last such option is the one that is effective.
6956
6957           -gdwarf does not accept a concatenated debug level, to avoid
6958           confusion with -gdwarf-level.  Instead use an additional -glevel
6959           option to change the debug level for DWARF.
6960
6961       -fno-eliminate-unused-debug-symbols
6962           By default, no debug information is produced for symbols that are
6963           not actually used. Use this option if you want debug information
6964           for all symbols.
6965
6966       -femit-class-debug-always
6967           Instead of emitting debugging information for a C++ class in only
6968           one object file, emit it in all object files using the class.  This
6969           option should be used only with debuggers that are unable to handle
6970           the way GCC normally emits debugging information for classes
6971           because using this option increases the size of debugging
6972           information by as much as a factor of two.
6973
6974       -fno-merge-debug-strings
6975           Direct the linker to not merge together strings in the debugging
6976           information that are identical in different object files.  Merging
6977           is not supported by all assemblers or linkers.  Merging decreases
6978           the size of the debug information in the output file at the cost of
6979           increasing link processing time.  Merging is enabled by default.
6980
6981       -fdebug-prefix-map=old=new
6982           When compiling files residing in directory old, record debugging
6983           information describing them as if the files resided in directory
6984           new instead.  This can be used to replace a build-time path with an
6985           install-time path in the debug info.  It can also be used to change
6986           an absolute path to a relative path by using . for new.  This can
6987           give more reproducible builds, which are location independent, but
6988           may require an extra command to tell GDB where to find the source
6989           files. See also -ffile-prefix-map.
6990
6991       -fvar-tracking
6992           Run variable tracking pass.  It computes where variables are stored
6993           at each position in code.  Better debugging information is then
6994           generated (if the debugging information format supports this
6995           information).
6996
6997           It is enabled by default when compiling with optimization (-Os, -O,
6998           -O2, ...), debugging information (-g) and the debug info format
6999           supports it.
7000
7001       -fvar-tracking-assignments
7002           Annotate assignments to user variables early in the compilation and
7003           attempt to carry the annotations over throughout the compilation
7004           all the way to the end, in an attempt to improve debug information
7005           while optimizing.  Use of -gdwarf-4 is recommended along with it.
7006
7007           It can be enabled even if var-tracking is disabled, in which case
7008           annotations are created and maintained, but discarded at the end.
7009           By default, this flag is enabled together with -fvar-tracking,
7010           except when selective scheduling is enabled.
7011
7012       -gsplit-dwarf
7013           Separate as much DWARF debugging information as possible into a
7014           separate output file with the extension .dwo.  This option allows
7015           the build system to avoid linking files with debug information.  To
7016           be useful, this option requires a debugger capable of reading .dwo
7017           files.
7018
7019       -gdescribe-dies
7020           Add description attributes to some DWARF DIEs that have no name
7021           attribute, such as artificial variables, external references and
7022           call site parameter DIEs.
7023
7024       -gpubnames
7025           Generate DWARF ".debug_pubnames" and ".debug_pubtypes" sections.
7026
7027       -ggnu-pubnames
7028           Generate ".debug_pubnames" and ".debug_pubtypes" sections in a
7029           format suitable for conversion into a GDB index.  This option is
7030           only useful with a linker that can produce GDB index version 7.
7031
7032       -fdebug-types-section
7033           When using DWARF Version 4 or higher, type DIEs can be put into
7034           their own ".debug_types" section instead of making them part of the
7035           ".debug_info" section.  It is more efficient to put them in a
7036           separate comdat section since the linker can then remove
7037           duplicates.  But not all DWARF consumers support ".debug_types"
7038           sections yet and on some objects ".debug_types" produces larger
7039           instead of smaller debugging information.
7040
7041       -grecord-gcc-switches
7042       -gno-record-gcc-switches
7043           This switch causes the command-line options used to invoke the
7044           compiler that may affect code generation to be appended to the
7045           DW_AT_producer attribute in DWARF debugging information.  The
7046           options are concatenated with spaces separating them from each
7047           other and from the compiler version.  It is enabled by default.
7048           See also -frecord-gcc-switches for another way of storing compiler
7049           options into the object file.
7050
7051       -gstrict-dwarf
7052           Disallow using extensions of later DWARF standard version than
7053           selected with -gdwarf-version.  On most targets using non-
7054           conflicting DWARF extensions from later standard versions is
7055           allowed.
7056
7057       -gno-strict-dwarf
7058           Allow using extensions of later DWARF standard version than
7059           selected with -gdwarf-version.
7060
7061       -gas-loc-support
7062           Inform the compiler that the assembler supports ".loc" directives.
7063           It may then use them for the assembler to generate DWARF2+ line
7064           number tables.
7065
7066           This is generally desirable, because assembler-generated line-
7067           number tables are a lot more compact than those the compiler can
7068           generate itself.
7069
7070           This option will be enabled by default if, at GCC configure time,
7071           the assembler was found to support such directives.
7072
7073       -gno-as-loc-support
7074           Force GCC to generate DWARF2+ line number tables internally, if
7075           DWARF2+ line number tables are to be generated.
7076
7077       -gas-locview-support
7078           Inform the compiler that the assembler supports "view" assignment
7079           and reset assertion checking in ".loc" directives.
7080
7081           This option will be enabled by default if, at GCC configure time,
7082           the assembler was found to support them.
7083
7084       -gno-as-locview-support
7085           Force GCC to assign view numbers internally, if
7086           -gvariable-location-views are explicitly requested.
7087
7088       -gcolumn-info
7089       -gno-column-info
7090           Emit location column information into DWARF debugging information,
7091           rather than just file and line.  This option is enabled by default.
7092
7093       -gstatement-frontiers
7094       -gno-statement-frontiers
7095           This option causes GCC to create markers in the internal
7096           representation at the beginning of statements, and to keep them
7097           roughly in place throughout compilation, using them to guide the
7098           output of "is_stmt" markers in the line number table.  This is
7099           enabled by default when compiling with optimization (-Os, -O, -O2,
7100           ...), and outputting DWARF 2 debug information at the normal level.
7101
7102       -gvariable-location-views
7103       -gvariable-location-views=incompat5
7104       -gno-variable-location-views
7105           Augment variable location lists with progressive view numbers
7106           implied from the line number table.  This enables debug information
7107           consumers to inspect state at certain points of the program, even
7108           if no instructions associated with the corresponding source
7109           locations are present at that point.  If the assembler lacks
7110           support for view numbers in line number tables, this will cause the
7111           compiler to emit the line number table, which generally makes them
7112           somewhat less compact.  The augmented line number tables and
7113           location lists are fully backward-compatible, so they can be
7114           consumed by debug information consumers that are not aware of these
7115           augmentations, but they won't derive any benefit from them either.
7116
7117           This is enabled by default when outputting DWARF 2 debug
7118           information at the normal level, as long as there is assembler
7119           support, -fvar-tracking-assignments is enabled and -gstrict-dwarf
7120           is not.  When assembler support is not available, this may still be
7121           enabled, but it will force GCC to output internal line number
7122           tables, and if -ginternal-reset-location-views is not enabled, that
7123           will most certainly lead to silently mismatching location views.
7124
7125           There is a proposed representation for view numbers that is not
7126           backward compatible with the location list format introduced in
7127           DWARF 5, that can be enabled with
7128           -gvariable-location-views=incompat5.  This option may be removed in
7129           the future, is only provided as a reference implementation of the
7130           proposed representation.  Debug information consumers are not
7131           expected to support this extended format, and they would be
7132           rendered unable to decode location lists using it.
7133
7134       -ginternal-reset-location-views
7135       -gno-internal-reset-location-views
7136           Attempt to determine location views that can be omitted from
7137           location view lists.  This requires the compiler to have very
7138           accurate insn length estimates, which isn't always the case, and it
7139           may cause incorrect view lists to be generated silently when using
7140           an assembler that does not support location view lists.  The GNU
7141           assembler will flag any such error as a "view number mismatch".
7142           This is only enabled on ports that define a reliable estimation
7143           function.
7144
7145       -ginline-points
7146       -gno-inline-points
7147           Generate extended debug information for inlined functions.
7148           Location view tracking markers are inserted at inlined entry
7149           points, so that address and view numbers can be computed and output
7150           in debug information.  This can be enabled independently of
7151           location views, in which case the view numbers won't be output, but
7152           it can only be enabled along with statement frontiers, and it is
7153           only enabled by default if location views are enabled.
7154
7155       -gz[=type]
7156           Produce compressed debug sections in DWARF format, if that is
7157           supported.  If type is not given, the default type depends on the
7158           capabilities of the assembler and linker used.  type may be one of
7159           none (don't compress debug sections), zlib (use zlib compression in
7160           ELF gABI format), or zlib-gnu (use zlib compression in traditional
7161           GNU format).  If the linker doesn't support writing compressed
7162           debug sections, the option is rejected.  Otherwise, if the
7163           assembler does not support them, -gz is silently ignored when
7164           producing object files.
7165
7166       -femit-struct-debug-baseonly
7167           Emit debug information for struct-like types only when the base
7168           name of the compilation source file matches the base name of file
7169           in which the struct is defined.
7170
7171           This option substantially reduces the size of debugging
7172           information, but at significant potential loss in type information
7173           to the debugger.  See -femit-struct-debug-reduced for a less
7174           aggressive option.  See -femit-struct-debug-detailed for more
7175           detailed control.
7176
7177           This option works only with DWARF debug output.
7178
7179       -femit-struct-debug-reduced
7180           Emit debug information for struct-like types only when the base
7181           name of the compilation source file matches the base name of file
7182           in which the type is defined, unless the struct is a template or
7183           defined in a system header.
7184
7185           This option significantly reduces the size of debugging
7186           information, with some potential loss in type information to the
7187           debugger.  See -femit-struct-debug-baseonly for a more aggressive
7188           option.  See -femit-struct-debug-detailed for more detailed
7189           control.
7190
7191           This option works only with DWARF debug output.
7192
7193       -femit-struct-debug-detailed[=spec-list]
7194           Specify the struct-like types for which the compiler generates
7195           debug information.  The intent is to reduce duplicate struct debug
7196           information between different object files within the same program.
7197
7198           This option is a detailed version of -femit-struct-debug-reduced
7199           and -femit-struct-debug-baseonly, which serves for most needs.
7200
7201           A specification has the
7202           syntax[dir:|ind:][ord:|gen:](any|sys|base|none)
7203
7204           The optional first word limits the specification to structs that
7205           are used directly (dir:) or used indirectly (ind:).  A struct type
7206           is used directly when it is the type of a variable, member.
7207           Indirect uses arise through pointers to structs.  That is, when use
7208           of an incomplete struct is valid, the use is indirect.  An example
7209           is struct one direct; struct two * indirect;.
7210
7211           The optional second word limits the specification to ordinary
7212           structs (ord:) or generic structs (gen:).  Generic structs are a
7213           bit complicated to explain.  For C++, these are non-explicit
7214           specializations of template classes, or non-template classes within
7215           the above.  Other programming languages have generics, but
7216           -femit-struct-debug-detailed does not yet implement them.
7217
7218           The third word specifies the source files for those structs for
7219           which the compiler should emit debug information.  The values none
7220           and any have the normal meaning.  The value base means that the
7221           base of name of the file in which the type declaration appears must
7222           match the base of the name of the main compilation file.  In
7223           practice, this means that when compiling foo.c, debug information
7224           is generated for types declared in that file and foo.h, but not
7225           other header files.  The value sys means those types satisfying
7226           base or declared in system or compiler headers.
7227
7228           You may need to experiment to determine the best settings for your
7229           application.
7230
7231           The default is -femit-struct-debug-detailed=all.
7232
7233           This option works only with DWARF debug output.
7234
7235       -fno-dwarf2-cfi-asm
7236           Emit DWARF unwind info as compiler generated ".eh_frame" section
7237           instead of using GAS ".cfi_*" directives.
7238
7239       -fno-eliminate-unused-debug-types
7240           Normally, when producing DWARF output, GCC avoids producing debug
7241           symbol output for types that are nowhere used in the source file
7242           being compiled.  Sometimes it is useful to have GCC emit debugging
7243           information for all types declared in a compilation unit,
7244           regardless of whether or not they are actually used in that
7245           compilation unit, for example if, in the debugger, you want to cast
7246           a value to a type that is not actually used in your program (but is
7247           declared).  More often, however, this results in a significant
7248           amount of wasted space.
7249
7250   Options That Control Optimization
7251       These options control various sorts of optimizations.
7252
7253       Without any optimization option, the compiler's goal is to reduce the
7254       cost of compilation and to make debugging produce the expected results.
7255       Statements are independent: if you stop the program with a breakpoint
7256       between statements, you can then assign a new value to any variable or
7257       change the program counter to any other statement in the function and
7258       get exactly the results you expect from the source code.
7259
7260       Turning on optimization flags makes the compiler attempt to improve the
7261       performance and/or code size at the expense of compilation time and
7262       possibly the ability to debug the program.
7263
7264       The compiler performs optimization based on the knowledge it has of the
7265       program.  Compiling multiple files at once to a single output file mode
7266       allows the compiler to use information gained from all of the files
7267       when compiling each of them.
7268
7269       Not all optimizations are controlled directly by a flag.  Only
7270       optimizations that have a flag are listed in this section.
7271
7272       Most optimizations are completely disabled at -O0 or if an -O level is
7273       not set on the command line, even if individual optimization flags are
7274       specified.  Similarly, -Og suppresses many optimization passes.
7275
7276       Depending on the target and how GCC was configured, a slightly
7277       different set of optimizations may be enabled at each -O level than
7278       those listed here.  You can invoke GCC with -Q --help=optimizers to
7279       find out the exact set of optimizations that are enabled at each level.
7280
7281       -O
7282       -O1 Optimize.  Optimizing compilation takes somewhat more time, and a
7283           lot more memory for a large function.
7284
7285           With -O, the compiler tries to reduce code size and execution time,
7286           without performing any optimizations that take a great deal of
7287           compilation time.
7288
7289           -O turns on the following optimization flags:
7290
7291           -fauto-inc-dec -fbranch-count-reg -fcombine-stack-adjustments
7292           -fcompare-elim -fcprop-registers -fdce -fdefer-pop -fdelayed-branch
7293           -fdse -fforward-propagate -fguess-branch-probability
7294           -fif-conversion -fif-conversion2 -finline-functions-called-once
7295           -fipa-profile -fipa-pure-const -fipa-reference
7296           -fipa-reference-addressable -fmerge-constants
7297           -fmove-loop-invariants -fomit-frame-pointer -freorder-blocks
7298           -fshrink-wrap -fshrink-wrap-separate -fsplit-wide-types
7299           -fssa-backprop -fssa-phiopt -ftree-bit-ccp -ftree-ccp -ftree-ch
7300           -ftree-coalesce-vars -ftree-copy-prop -ftree-dce
7301           -ftree-dominator-opts -ftree-dse -ftree-forwprop -ftree-fre
7302           -ftree-phiprop -ftree-pta -ftree-scev-cprop -ftree-sink -ftree-slsr
7303           -ftree-sra -ftree-ter -funit-at-a-time
7304
7305       -O2 Optimize even more.  GCC performs nearly all supported
7306           optimizations that do not involve a space-speed tradeoff.  As
7307           compared to -O, this option increases both compilation time and the
7308           performance of the generated code.
7309
7310           -O2 turns on all optimization flags specified by -O.  It also turns
7311           on the following optimization flags:
7312
7313           -falign-functions  -falign-jumps -falign-labels  -falign-loops
7314           -fcaller-saves -fcode-hoisting -fcrossjumping -fcse-follow-jumps
7315           -fcse-skip-blocks -fdelete-null-pointer-checks -fdevirtualize
7316           -fdevirtualize-speculatively -fexpensive-optimizations
7317           -ffinite-loops -fgcse  -fgcse-lm -fhoist-adjacent-loads
7318           -finline-functions -finline-small-functions -findirect-inlining
7319           -fipa-bit-cp  -fipa-cp  -fipa-icf -fipa-ra  -fipa-sra  -fipa-vrp
7320           -fisolate-erroneous-paths-dereference -flra-remat
7321           -foptimize-sibling-calls -foptimize-strlen -fpartial-inlining
7322           -fpeephole2 -freorder-blocks-algorithm=stc
7323           -freorder-blocks-and-partition  -freorder-functions
7324           -frerun-cse-after-loop -fschedule-insns  -fschedule-insns2
7325           -fsched-interblock  -fsched-spec -fstore-merging -fstrict-aliasing
7326           -fthread-jumps -ftree-builtin-call-dce -ftree-pre
7327           -ftree-switch-conversion  -ftree-tail-merge -ftree-vrp
7328
7329           Please note the warning under -fgcse about invoking -O2 on programs
7330           that use computed gotos.
7331
7332       -O3 Optimize yet more.  -O3 turns on all optimizations specified by -O2
7333           and also turns on the following optimization flags:
7334
7335           -fgcse-after-reload -fipa-cp-clone -floop-interchange
7336           -floop-unroll-and-jam -fpeel-loops -fpredictive-commoning
7337           -fsplit-loops -fsplit-paths -ftree-loop-distribution
7338           -ftree-loop-vectorize -ftree-partial-pre -ftree-slp-vectorize
7339           -funswitch-loops -fvect-cost-model -fvect-cost-model=dynamic
7340           -fversion-loops-for-strides
7341
7342       -O0 Reduce compilation time and make debugging produce the expected
7343           results.  This is the default.
7344
7345       -Os Optimize for size.  -Os enables all -O2 optimizations except those
7346           that often increase code size:
7347
7348           -falign-functions  -falign-jumps -falign-labels  -falign-loops
7349           -fprefetch-loop-arrays  -freorder-blocks-algorithm=stc
7350
7351           It also enables -finline-functions, causes the compiler to tune for
7352           code size rather than execution speed, and performs further
7353           optimizations designed to reduce code size.
7354
7355       -Ofast
7356           Disregard strict standards compliance.  -Ofast enables all -O3
7357           optimizations.  It also enables optimizations that are not valid
7358           for all standard-compliant programs.  It turns on -ffast-math,
7359           -fallow-store-data-races and the Fortran-specific -fstack-arrays,
7360           unless -fmax-stack-var-size is specified, and -fno-protect-parens.
7361
7362       -Og Optimize debugging experience.  -Og should be the optimization
7363           level of choice for the standard edit-compile-debug cycle, offering
7364           a reasonable level of optimization while maintaining fast
7365           compilation and a good debugging experience.  It is a better choice
7366           than -O0 for producing debuggable code because some compiler passes
7367           that collect debug information are disabled at -O0.
7368
7369           Like -O0, -Og completely disables a number of optimization passes
7370           so that individual options controlling them have no effect.
7371           Otherwise -Og enables all -O1 optimization flags except for those
7372           that may interfere with debugging:
7373
7374           -fbranch-count-reg  -fdelayed-branch -fdse  -fif-conversion
7375           -fif-conversion2 -finline-functions-called-once
7376           -fmove-loop-invariants  -fssa-phiopt -ftree-bit-ccp  -ftree-dse
7377           -ftree-pta  -ftree-sra
7378
7379       If you use multiple -O options, with or without level numbers, the last
7380       such option is the one that is effective.
7381
7382       Options of the form -fflag specify machine-independent flags.  Most
7383       flags have both positive and negative forms; the negative form of -ffoo
7384       is -fno-foo.  In the table below, only one of the forms is listed---the
7385       one you typically use.  You can figure out the other form by either
7386       removing no- or adding it.
7387
7388       The following options control specific optimizations.  They are either
7389       activated by -O options or are related to ones that are.  You can use
7390       the following flags in the rare cases when "fine-tuning" of
7391       optimizations to be performed is desired.
7392
7393       -fno-defer-pop
7394           For machines that must pop arguments after a function call, always
7395           pop the arguments as soon as each function returns.  At levels -O1
7396           and higher, -fdefer-pop is the default; this allows the compiler to
7397           let arguments accumulate on the stack for several function calls
7398           and pop them all at once.
7399
7400       -fforward-propagate
7401           Perform a forward propagation pass on RTL.  The pass tries to
7402           combine two instructions and checks if the result can be
7403           simplified.  If loop unrolling is active, two passes are performed
7404           and the second is scheduled after loop unrolling.
7405
7406           This option is enabled by default at optimization levels -O, -O2,
7407           -O3, -Os.
7408
7409       -ffp-contract=style
7410           -ffp-contract=off disables floating-point expression contraction.
7411           -ffp-contract=fast enables floating-point expression contraction
7412           such as forming of fused multiply-add operations if the target has
7413           native support for them.  -ffp-contract=on enables floating-point
7414           expression contraction if allowed by the language standard.  This
7415           is currently not implemented and treated equal to
7416           -ffp-contract=off.
7417
7418           The default is -ffp-contract=fast.
7419
7420       -fomit-frame-pointer
7421           Omit the frame pointer in functions that don't need one.  This
7422           avoids the instructions to save, set up and restore the frame
7423           pointer; on many targets it also makes an extra register available.
7424
7425           On some targets this flag has no effect because the standard
7426           calling sequence always uses a frame pointer, so it cannot be
7427           omitted.
7428
7429           Note that -fno-omit-frame-pointer doesn't guarantee the frame
7430           pointer is used in all functions.  Several targets always omit the
7431           frame pointer in leaf functions.
7432
7433           Enabled by default at -O and higher.
7434
7435       -foptimize-sibling-calls
7436           Optimize sibling and tail recursive calls.
7437
7438           Enabled at levels -O2, -O3, -Os.
7439
7440       -foptimize-strlen
7441           Optimize various standard C string functions (e.g. "strlen",
7442           "strchr" or "strcpy") and their "_FORTIFY_SOURCE" counterparts into
7443           faster alternatives.
7444
7445           Enabled at levels -O2, -O3.
7446
7447       -fno-inline
7448           Do not expand any functions inline apart from those marked with the
7449           "always_inline" attribute.  This is the default when not
7450           optimizing.
7451
7452           Single functions can be exempted from inlining by marking them with
7453           the "noinline" attribute.
7454
7455       -finline-small-functions
7456           Integrate functions into their callers when their body is smaller
7457           than expected function call code (so overall size of program gets
7458           smaller).  The compiler heuristically decides which functions are
7459           simple enough to be worth integrating in this way.  This inlining
7460           applies to all functions, even those not declared inline.
7461
7462           Enabled at levels -O2, -O3, -Os.
7463
7464       -findirect-inlining
7465           Inline also indirect calls that are discovered to be known at
7466           compile time thanks to previous inlining.  This option has any
7467           effect only when inlining itself is turned on by the
7468           -finline-functions or -finline-small-functions options.
7469
7470           Enabled at levels -O2, -O3, -Os.
7471
7472       -finline-functions
7473           Consider all functions for inlining, even if they are not declared
7474           inline.  The compiler heuristically decides which functions are
7475           worth integrating in this way.
7476
7477           If all calls to a given function are integrated, and the function
7478           is declared "static", then the function is normally not output as
7479           assembler code in its own right.
7480
7481           Enabled at levels -O2, -O3, -Os.  Also enabled by -fprofile-use and
7482           -fauto-profile.
7483
7484       -finline-functions-called-once
7485           Consider all "static" functions called once for inlining into their
7486           caller even if they are not marked "inline".  If a call to a given
7487           function is integrated, then the function is not output as
7488           assembler code in its own right.
7489
7490           Enabled at levels -O1, -O2, -O3 and -Os, but not -Og.
7491
7492       -fearly-inlining
7493           Inline functions marked by "always_inline" and functions whose body
7494           seems smaller than the function call overhead early before doing
7495           -fprofile-generate instrumentation and real inlining pass.  Doing
7496           so makes profiling significantly cheaper and usually inlining
7497           faster on programs having large chains of nested wrapper functions.
7498
7499           Enabled by default.
7500
7501       -fipa-sra
7502           Perform interprocedural scalar replacement of aggregates, removal
7503           of unused parameters and replacement of parameters passed by
7504           reference by parameters passed by value.
7505
7506           Enabled at levels -O2, -O3 and -Os.
7507
7508       -finline-limit=n
7509           By default, GCC limits the size of functions that can be inlined.
7510           This flag allows coarse control of this limit.  n is the size of
7511           functions that can be inlined in number of pseudo instructions.
7512
7513           Inlining is actually controlled by a number of parameters, which
7514           may be specified individually by using --param name=value.  The
7515           -finline-limit=n option sets some of these parameters as follows:
7516
7517           max-inline-insns-single
7518               is set to n/2.
7519
7520           max-inline-insns-auto
7521               is set to n/2.
7522
7523           See below for a documentation of the individual parameters
7524           controlling inlining and for the defaults of these parameters.
7525
7526           Note: there may be no value to -finline-limit that results in
7527           default behavior.
7528
7529           Note: pseudo instruction represents, in this particular context, an
7530           abstract measurement of function's size.  In no way does it
7531           represent a count of assembly instructions and as such its exact
7532           meaning might change from one release to an another.
7533
7534       -fno-keep-inline-dllexport
7535           This is a more fine-grained version of -fkeep-inline-functions,
7536           which applies only to functions that are declared using the
7537           "dllexport" attribute or declspec.
7538
7539       -fkeep-inline-functions
7540           In C, emit "static" functions that are declared "inline" into the
7541           object file, even if the function has been inlined into all of its
7542           callers.  This switch does not affect functions using the "extern
7543           inline" extension in GNU C90.  In C++, emit any and all inline
7544           functions into the object file.
7545
7546       -fkeep-static-functions
7547           Emit "static" functions into the object file, even if the function
7548           is never used.
7549
7550       -fkeep-static-consts
7551           Emit variables declared "static const" when optimization isn't
7552           turned on, even if the variables aren't referenced.
7553
7554           GCC enables this option by default.  If you want to force the
7555           compiler to check if a variable is referenced, regardless of
7556           whether or not optimization is turned on, use the
7557           -fno-keep-static-consts option.
7558
7559       -fmerge-constants
7560           Attempt to merge identical constants (string constants and
7561           floating-point constants) across compilation units.
7562
7563           This option is the default for optimized compilation if the
7564           assembler and linker support it.  Use -fno-merge-constants to
7565           inhibit this behavior.
7566
7567           Enabled at levels -O, -O2, -O3, -Os.
7568
7569       -fmerge-all-constants
7570           Attempt to merge identical constants and identical variables.
7571
7572           This option implies -fmerge-constants.  In addition to
7573           -fmerge-constants this considers e.g. even constant initialized
7574           arrays or initialized constant variables with integral or floating-
7575           point types.  Languages like C or C++ require each variable,
7576           including multiple instances of the same variable in recursive
7577           calls, to have distinct locations, so using this option results in
7578           non-conforming behavior.
7579
7580       -fmodulo-sched
7581           Perform swing modulo scheduling immediately before the first
7582           scheduling pass.  This pass looks at innermost loops and reorders
7583           their instructions by overlapping different iterations.
7584
7585       -fmodulo-sched-allow-regmoves
7586           Perform more aggressive SMS-based modulo scheduling with register
7587           moves allowed.  By setting this flag certain anti-dependences edges
7588           are deleted, which triggers the generation of reg-moves based on
7589           the life-range analysis.  This option is effective only with
7590           -fmodulo-sched enabled.
7591
7592       -fno-branch-count-reg
7593           Disable the optimization pass that scans for opportunities to use
7594           "decrement and branch" instructions on a count register instead of
7595           instruction sequences that decrement a register, compare it against
7596           zero, and then branch based upon the result.  This option is only
7597           meaningful on architectures that support such instructions, which
7598           include x86, PowerPC, IA-64 and S/390.  Note that the
7599           -fno-branch-count-reg option doesn't remove the decrement and
7600           branch instructions from the generated instruction stream
7601           introduced by other optimization passes.
7602
7603           The default is -fbranch-count-reg at -O1 and higher, except for
7604           -Og.
7605
7606       -fno-function-cse
7607           Do not put function addresses in registers; make each instruction
7608           that calls a constant function contain the function's address
7609           explicitly.
7610
7611           This option results in less efficient code, but some strange hacks
7612           that alter the assembler output may be confused by the
7613           optimizations performed when this option is not used.
7614
7615           The default is -ffunction-cse
7616
7617       -fno-zero-initialized-in-bss
7618           If the target supports a BSS section, GCC by default puts variables
7619           that are initialized to zero into BSS.  This can save space in the
7620           resulting code.
7621
7622           This option turns off this behavior because some programs
7623           explicitly rely on variables going to the data section---e.g., so
7624           that the resulting executable can find the beginning of that
7625           section and/or make assumptions based on that.
7626
7627           The default is -fzero-initialized-in-bss.
7628
7629       -fthread-jumps
7630           Perform optimizations that check to see if a jump branches to a
7631           location where another comparison subsumed by the first is found.
7632           If so, the first branch is redirected to either the destination of
7633           the second branch or a point immediately following it, depending on
7634           whether the condition is known to be true or false.
7635
7636           Enabled at levels -O2, -O3, -Os.
7637
7638       -fsplit-wide-types
7639           When using a type that occupies multiple registers, such as "long
7640           long" on a 32-bit system, split the registers apart and allocate
7641           them independently.  This normally generates better code for those
7642           types, but may make debugging more difficult.
7643
7644           Enabled at levels -O, -O2, -O3, -Os.
7645
7646       -fsplit-wide-types-early
7647           Fully split wide types early, instead of very late.  This option
7648           has no effect unless -fsplit-wide-types is turned on.
7649
7650           This is the default on some targets.
7651
7652       -fcse-follow-jumps
7653           In common subexpression elimination (CSE), scan through jump
7654           instructions when the target of the jump is not reached by any
7655           other path.  For example, when CSE encounters an "if" statement
7656           with an "else" clause, CSE follows the jump when the condition
7657           tested is false.
7658
7659           Enabled at levels -O2, -O3, -Os.
7660
7661       -fcse-skip-blocks
7662           This is similar to -fcse-follow-jumps, but causes CSE to follow
7663           jumps that conditionally skip over blocks.  When CSE encounters a
7664           simple "if" statement with no else clause, -fcse-skip-blocks causes
7665           CSE to follow the jump around the body of the "if".
7666
7667           Enabled at levels -O2, -O3, -Os.
7668
7669       -frerun-cse-after-loop
7670           Re-run common subexpression elimination after loop optimizations
7671           are performed.
7672
7673           Enabled at levels -O2, -O3, -Os.
7674
7675       -fgcse
7676           Perform a global common subexpression elimination pass.  This pass
7677           also performs global constant and copy propagation.
7678
7679           Note: When compiling a program using computed gotos, a GCC
7680           extension, you may get better run-time performance if you disable
7681           the global common subexpression elimination pass by adding
7682           -fno-gcse to the command line.
7683
7684           Enabled at levels -O2, -O3, -Os.
7685
7686       -fgcse-lm
7687           When -fgcse-lm is enabled, global common subexpression elimination
7688           attempts to move loads that are only killed by stores into
7689           themselves.  This allows a loop containing a load/store sequence to
7690           be changed to a load outside the loop, and a copy/store within the
7691           loop.
7692
7693           Enabled by default when -fgcse is enabled.
7694
7695       -fgcse-sm
7696           When -fgcse-sm is enabled, a store motion pass is run after global
7697           common subexpression elimination.  This pass attempts to move
7698           stores out of loops.  When used in conjunction with -fgcse-lm,
7699           loops containing a load/store sequence can be changed to a load
7700           before the loop and a store after the loop.
7701
7702           Not enabled at any optimization level.
7703
7704       -fgcse-las
7705           When -fgcse-las is enabled, the global common subexpression
7706           elimination pass eliminates redundant loads that come after stores
7707           to the same memory location (both partial and full redundancies).
7708
7709           Not enabled at any optimization level.
7710
7711       -fgcse-after-reload
7712           When -fgcse-after-reload is enabled, a redundant load elimination
7713           pass is performed after reload.  The purpose of this pass is to
7714           clean up redundant spilling.
7715
7716           Enabled by -fprofile-use and -fauto-profile.
7717
7718       -faggressive-loop-optimizations
7719           This option tells the loop optimizer to use language constraints to
7720           derive bounds for the number of iterations of a loop.  This assumes
7721           that loop code does not invoke undefined behavior by for example
7722           causing signed integer overflows or out-of-bound array accesses.
7723           The bounds for the number of iterations of a loop are used to guide
7724           loop unrolling and peeling and loop exit test optimizations.  This
7725           option is enabled by default.
7726
7727       -funconstrained-commons
7728           This option tells the compiler that variables declared in common
7729           blocks (e.g. Fortran) may later be overridden with longer trailing
7730           arrays. This prevents certain optimizations that depend on knowing
7731           the array bounds.
7732
7733       -fcrossjumping
7734           Perform cross-jumping transformation.  This transformation unifies
7735           equivalent code and saves code size.  The resulting code may or may
7736           not perform better than without cross-jumping.
7737
7738           Enabled at levels -O2, -O3, -Os.
7739
7740       -fauto-inc-dec
7741           Combine increments or decrements of addresses with memory accesses.
7742           This pass is always skipped on architectures that do not have
7743           instructions to support this.  Enabled by default at -O and higher
7744           on architectures that support this.
7745
7746       -fdce
7747           Perform dead code elimination (DCE) on RTL.  Enabled by default at
7748           -O and higher.
7749
7750       -fdse
7751           Perform dead store elimination (DSE) on RTL.  Enabled by default at
7752           -O and higher.
7753
7754       -fif-conversion
7755           Attempt to transform conditional jumps into branch-less
7756           equivalents.  This includes use of conditional moves, min, max, set
7757           flags and abs instructions, and some tricks doable by standard
7758           arithmetics.  The use of conditional execution on chips where it is
7759           available is controlled by -fif-conversion2.
7760
7761           Enabled at levels -O, -O2, -O3, -Os, but not with -Og.
7762
7763       -fif-conversion2
7764           Use conditional execution (where available) to transform
7765           conditional jumps into branch-less equivalents.
7766
7767           Enabled at levels -O, -O2, -O3, -Os, but not with -Og.
7768
7769       -fdeclone-ctor-dtor
7770           The C++ ABI requires multiple entry points for constructors and
7771           destructors: one for a base subobject, one for a complete object,
7772           and one for a virtual destructor that calls operator delete
7773           afterwards.  For a hierarchy with virtual bases, the base and
7774           complete variants are clones, which means two copies of the
7775           function.  With this option, the base and complete variants are
7776           changed to be thunks that call a common implementation.
7777
7778           Enabled by -Os.
7779
7780       -fdelete-null-pointer-checks
7781           Assume that programs cannot safely dereference null pointers, and
7782           that no code or data element resides at address zero.  This option
7783           enables simple constant folding optimizations at all optimization
7784           levels.  In addition, other optimization passes in GCC use this
7785           flag to control global dataflow analyses that eliminate useless
7786           checks for null pointers; these assume that a memory access to
7787           address zero always results in a trap, so that if a pointer is
7788           checked after it has already been dereferenced, it cannot be null.
7789
7790           Note however that in some environments this assumption is not true.
7791           Use -fno-delete-null-pointer-checks to disable this optimization
7792           for programs that depend on that behavior.
7793
7794           This option is enabled by default on most targets.  On Nios II ELF,
7795           it defaults to off.  On AVR, CR16, and MSP430, this option is
7796           completely disabled.
7797
7798           Passes that use the dataflow information are enabled independently
7799           at different optimization levels.
7800
7801       -fdevirtualize
7802           Attempt to convert calls to virtual functions to direct calls.
7803           This is done both within a procedure and interprocedurally as part
7804           of indirect inlining (-findirect-inlining) and interprocedural
7805           constant propagation (-fipa-cp).  Enabled at levels -O2, -O3, -Os.
7806
7807       -fdevirtualize-speculatively
7808           Attempt to convert calls to virtual functions to speculative direct
7809           calls.  Based on the analysis of the type inheritance graph,
7810           determine for a given call the set of likely targets. If the set is
7811           small, preferably of size 1, change the call into a conditional
7812           deciding between direct and indirect calls.  The speculative calls
7813           enable more optimizations, such as inlining.  When they seem
7814           useless after further optimization, they are converted back into
7815           original form.
7816
7817       -fdevirtualize-at-ltrans
7818           Stream extra information needed for aggressive devirtualization
7819           when running the link-time optimizer in local transformation mode.
7820           This option enables more devirtualization but significantly
7821           increases the size of streamed data. For this reason it is disabled
7822           by default.
7823
7824       -fexpensive-optimizations
7825           Perform a number of minor optimizations that are relatively
7826           expensive.
7827
7828           Enabled at levels -O2, -O3, -Os.
7829
7830       -free
7831           Attempt to remove redundant extension instructions.  This is
7832           especially helpful for the x86-64 architecture, which implicitly
7833           zero-extends in 64-bit registers after writing to their lower
7834           32-bit half.
7835
7836           Enabled for Alpha, AArch64 and x86 at levels -O2, -O3, -Os.
7837
7838       -fno-lifetime-dse
7839           In C++ the value of an object is only affected by changes within
7840           its lifetime: when the constructor begins, the object has an
7841           indeterminate value, and any changes during the lifetime of the
7842           object are dead when the object is destroyed.  Normally dead store
7843           elimination will take advantage of this; if your code relies on the
7844           value of the object storage persisting beyond the lifetime of the
7845           object, you can use this flag to disable this optimization.  To
7846           preserve stores before the constructor starts (e.g. because your
7847           operator new clears the object storage) but still treat the object
7848           as dead after the destructor, you can use -flifetime-dse=1.  The
7849           default behavior can be explicitly selected with -flifetime-dse=2.
7850           -flifetime-dse=0 is equivalent to -fno-lifetime-dse.
7851
7852       -flive-range-shrinkage
7853           Attempt to decrease register pressure through register live range
7854           shrinkage.  This is helpful for fast processors with small or
7855           moderate size register sets.
7856
7857       -fira-algorithm=algorithm
7858           Use the specified coloring algorithm for the integrated register
7859           allocator.  The algorithm argument can be priority, which specifies
7860           Chow's priority coloring, or CB, which specifies Chaitin-Briggs
7861           coloring.  Chaitin-Briggs coloring is not implemented for all
7862           architectures, but for those targets that do support it, it is the
7863           default because it generates better code.
7864
7865       -fira-region=region
7866           Use specified regions for the integrated register allocator.  The
7867           region argument should be one of the following:
7868
7869           all Use all loops as register allocation regions.  This can give
7870               the best results for machines with a small and/or irregular
7871               register set.
7872
7873           mixed
7874               Use all loops except for loops with small register pressure as
7875               the regions.  This value usually gives the best results in most
7876               cases and for most architectures, and is enabled by default
7877               when compiling with optimization for speed (-O, -O2, ...).
7878
7879           one Use all functions as a single region.  This typically results
7880               in the smallest code size, and is enabled by default for -Os or
7881               -O0.
7882
7883       -fira-hoist-pressure
7884           Use IRA to evaluate register pressure in the code hoisting pass for
7885           decisions to hoist expressions.  This option usually results in
7886           smaller code, but it can slow the compiler down.
7887
7888           This option is enabled at level -Os for all targets.
7889
7890       -fira-loop-pressure
7891           Use IRA to evaluate register pressure in loops for decisions to
7892           move loop invariants.  This option usually results in generation of
7893           faster and smaller code on machines with large register files (>=
7894           32 registers), but it can slow the compiler down.
7895
7896           This option is enabled at level -O3 for some targets.
7897
7898       -fno-ira-share-save-slots
7899           Disable sharing of stack slots used for saving call-used hard
7900           registers living through a call.  Each hard register gets a
7901           separate stack slot, and as a result function stack frames are
7902           larger.
7903
7904       -fno-ira-share-spill-slots
7905           Disable sharing of stack slots allocated for pseudo-registers.
7906           Each pseudo-register that does not get a hard register gets a
7907           separate stack slot, and as a result function stack frames are
7908           larger.
7909
7910       -flra-remat
7911           Enable CFG-sensitive rematerialization in LRA.  Instead of loading
7912           values of spilled pseudos, LRA tries to rematerialize (recalculate)
7913           values if it is profitable.
7914
7915           Enabled at levels -O2, -O3, -Os.
7916
7917       -fdelayed-branch
7918           If supported for the target machine, attempt to reorder
7919           instructions to exploit instruction slots available after delayed
7920           branch instructions.
7921
7922           Enabled at levels -O, -O2, -O3, -Os, but not at -Og.
7923
7924       -fschedule-insns
7925           If supported for the target machine, attempt to reorder
7926           instructions to eliminate execution stalls due to required data
7927           being unavailable.  This helps machines that have slow floating
7928           point or memory load instructions by allowing other instructions to
7929           be issued until the result of the load or floating-point
7930           instruction is required.
7931
7932           Enabled at levels -O2, -O3.
7933
7934       -fschedule-insns2
7935           Similar to -fschedule-insns, but requests an additional pass of
7936           instruction scheduling after register allocation has been done.
7937           This is especially useful on machines with a relatively small
7938           number of registers and where memory load instructions take more
7939           than one cycle.
7940
7941           Enabled at levels -O2, -O3, -Os.
7942
7943       -fno-sched-interblock
7944           Disable instruction scheduling across basic blocks, which is
7945           normally enabled when scheduling before register allocation, i.e.
7946           with -fschedule-insns or at -O2 or higher.
7947
7948       -fno-sched-spec
7949           Disable speculative motion of non-load instructions, which is
7950           normally enabled when scheduling before register allocation, i.e.
7951           with -fschedule-insns or at -O2 or higher.
7952
7953       -fsched-pressure
7954           Enable register pressure sensitive insn scheduling before register
7955           allocation.  This only makes sense when scheduling before register
7956           allocation is enabled, i.e. with -fschedule-insns or at -O2 or
7957           higher.  Usage of this option can improve the generated code and
7958           decrease its size by preventing register pressure increase above
7959           the number of available hard registers and subsequent spills in
7960           register allocation.
7961
7962       -fsched-spec-load
7963           Allow speculative motion of some load instructions.  This only
7964           makes sense when scheduling before register allocation, i.e. with
7965           -fschedule-insns or at -O2 or higher.
7966
7967       -fsched-spec-load-dangerous
7968           Allow speculative motion of more load instructions.  This only
7969           makes sense when scheduling before register allocation, i.e. with
7970           -fschedule-insns or at -O2 or higher.
7971
7972       -fsched-stalled-insns
7973       -fsched-stalled-insns=n
7974           Define how many insns (if any) can be moved prematurely from the
7975           queue of stalled insns into the ready list during the second
7976           scheduling pass.  -fno-sched-stalled-insns means that no insns are
7977           moved prematurely, -fsched-stalled-insns=0 means there is no limit
7978           on how many queued insns can be moved prematurely.
7979           -fsched-stalled-insns without a value is equivalent to
7980           -fsched-stalled-insns=1.
7981
7982       -fsched-stalled-insns-dep
7983       -fsched-stalled-insns-dep=n
7984           Define how many insn groups (cycles) are examined for a dependency
7985           on a stalled insn that is a candidate for premature removal from
7986           the queue of stalled insns.  This has an effect only during the
7987           second scheduling pass, and only if -fsched-stalled-insns is used.
7988           -fno-sched-stalled-insns-dep is equivalent to
7989           -fsched-stalled-insns-dep=0.  -fsched-stalled-insns-dep without a
7990           value is equivalent to -fsched-stalled-insns-dep=1.
7991
7992       -fsched2-use-superblocks
7993           When scheduling after register allocation, use superblock
7994           scheduling.  This allows motion across basic block boundaries,
7995           resulting in faster schedules.  This option is experimental, as not
7996           all machine descriptions used by GCC model the CPU closely enough
7997           to avoid unreliable results from the algorithm.
7998
7999           This only makes sense when scheduling after register allocation,
8000           i.e. with -fschedule-insns2 or at -O2 or higher.
8001
8002       -fsched-group-heuristic
8003           Enable the group heuristic in the scheduler.  This heuristic favors
8004           the instruction that belongs to a schedule group.  This is enabled
8005           by default when scheduling is enabled, i.e. with -fschedule-insns
8006           or -fschedule-insns2 or at -O2 or higher.
8007
8008       -fsched-critical-path-heuristic
8009           Enable the critical-path heuristic in the scheduler.  This
8010           heuristic favors instructions on the critical path.  This is
8011           enabled by default when scheduling is enabled, i.e. with
8012           -fschedule-insns or -fschedule-insns2 or at -O2 or higher.
8013
8014       -fsched-spec-insn-heuristic
8015           Enable the speculative instruction heuristic in the scheduler.
8016           This heuristic favors speculative instructions with greater
8017           dependency weakness.  This is enabled by default when scheduling is
8018           enabled, i.e.  with -fschedule-insns or -fschedule-insns2 or at -O2
8019           or higher.
8020
8021       -fsched-rank-heuristic
8022           Enable the rank heuristic in the scheduler.  This heuristic favors
8023           the instruction belonging to a basic block with greater size or
8024           frequency.  This is enabled by default when scheduling is enabled,
8025           i.e.  with -fschedule-insns or -fschedule-insns2 or at -O2 or
8026           higher.
8027
8028       -fsched-last-insn-heuristic
8029           Enable the last-instruction heuristic in the scheduler.  This
8030           heuristic favors the instruction that is less dependent on the last
8031           instruction scheduled.  This is enabled by default when scheduling
8032           is enabled, i.e. with -fschedule-insns or -fschedule-insns2 or at
8033           -O2 or higher.
8034
8035       -fsched-dep-count-heuristic
8036           Enable the dependent-count heuristic in the scheduler.  This
8037           heuristic favors the instruction that has more instructions
8038           depending on it.  This is enabled by default when scheduling is
8039           enabled, i.e.  with -fschedule-insns or -fschedule-insns2 or at -O2
8040           or higher.
8041
8042       -freschedule-modulo-scheduled-loops
8043           Modulo scheduling is performed before traditional scheduling.  If a
8044           loop is modulo scheduled, later scheduling passes may change its
8045           schedule.  Use this option to control that behavior.
8046
8047       -fselective-scheduling
8048           Schedule instructions using selective scheduling algorithm.
8049           Selective scheduling runs instead of the first scheduler pass.
8050
8051       -fselective-scheduling2
8052           Schedule instructions using selective scheduling algorithm.
8053           Selective scheduling runs instead of the second scheduler pass.
8054
8055       -fsel-sched-pipelining
8056           Enable software pipelining of innermost loops during selective
8057           scheduling.  This option has no effect unless one of
8058           -fselective-scheduling or -fselective-scheduling2 is turned on.
8059
8060       -fsel-sched-pipelining-outer-loops
8061           When pipelining loops during selective scheduling, also pipeline
8062           outer loops.  This option has no effect unless
8063           -fsel-sched-pipelining is turned on.
8064
8065       -fsemantic-interposition
8066           Some object formats, like ELF, allow interposing of symbols by the
8067           dynamic linker.  This means that for symbols exported from the DSO,
8068           the compiler cannot perform interprocedural propagation, inlining
8069           and other optimizations in anticipation that the function or
8070           variable in question may change. While this feature is useful, for
8071           example, to rewrite memory allocation functions by a debugging
8072           implementation, it is expensive in the terms of code quality.  With
8073           -fno-semantic-interposition the compiler assumes that if
8074           interposition happens for functions the overwriting function will
8075           have precisely the same semantics (and side effects).  Similarly if
8076           interposition happens for variables, the constructor of the
8077           variable will be the same. The flag has no effect for functions
8078           explicitly declared inline (where it is never allowed for
8079           interposition to change semantics) and for symbols explicitly
8080           declared weak.
8081
8082       -fshrink-wrap
8083           Emit function prologues only before parts of the function that need
8084           it, rather than at the top of the function.  This flag is enabled
8085           by default at -O and higher.
8086
8087       -fshrink-wrap-separate
8088           Shrink-wrap separate parts of the prologue and epilogue separately,
8089           so that those parts are only executed when needed.  This option is
8090           on by default, but has no effect unless -fshrink-wrap is also
8091           turned on and the target supports this.
8092
8093       -fcaller-saves
8094           Enable allocation of values to registers that are clobbered by
8095           function calls, by emitting extra instructions to save and restore
8096           the registers around such calls.  Such allocation is done only when
8097           it seems to result in better code.
8098
8099           This option is always enabled by default on certain machines,
8100           usually those which have no call-preserved registers to use
8101           instead.
8102
8103           Enabled at levels -O2, -O3, -Os.
8104
8105       -fcombine-stack-adjustments
8106           Tracks stack adjustments (pushes and pops) and stack memory
8107           references and then tries to find ways to combine them.
8108
8109           Enabled by default at -O1 and higher.
8110
8111       -fipa-ra
8112           Use caller save registers for allocation if those registers are not
8113           used by any called function.  In that case it is not necessary to
8114           save and restore them around calls.  This is only possible if
8115           called functions are part of same compilation unit as current
8116           function and they are compiled before it.
8117
8118           Enabled at levels -O2, -O3, -Os, however the option is disabled if
8119           generated code will be instrumented for profiling (-p, or -pg) or
8120           if callee's register usage cannot be known exactly (this happens on
8121           targets that do not expose prologues and epilogues in RTL).
8122
8123       -fconserve-stack
8124           Attempt to minimize stack usage.  The compiler attempts to use less
8125           stack space, even if that makes the program slower.  This option
8126           implies setting the large-stack-frame parameter to 100 and the
8127           large-stack-frame-growth parameter to 400.
8128
8129       -ftree-reassoc
8130           Perform reassociation on trees.  This flag is enabled by default at
8131           -O and higher.
8132
8133       -fcode-hoisting
8134           Perform code hoisting.  Code hoisting tries to move the evaluation
8135           of expressions executed on all paths to the function exit as early
8136           as possible.  This is especially useful as a code size
8137           optimization, but it often helps for code speed as well.  This flag
8138           is enabled by default at -O2 and higher.
8139
8140       -ftree-pre
8141           Perform partial redundancy elimination (PRE) on trees.  This flag
8142           is enabled by default at -O2 and -O3.
8143
8144       -ftree-partial-pre
8145           Make partial redundancy elimination (PRE) more aggressive.  This
8146           flag is enabled by default at -O3.
8147
8148       -ftree-forwprop
8149           Perform forward propagation on trees.  This flag is enabled by
8150           default at -O and higher.
8151
8152       -ftree-fre
8153           Perform full redundancy elimination (FRE) on trees.  The difference
8154           between FRE and PRE is that FRE only considers expressions that are
8155           computed on all paths leading to the redundant computation.  This
8156           analysis is faster than PRE, though it exposes fewer redundancies.
8157           This flag is enabled by default at -O and higher.
8158
8159       -ftree-phiprop
8160           Perform hoisting of loads from conditional pointers on trees.  This
8161           pass is enabled by default at -O and higher.
8162
8163       -fhoist-adjacent-loads
8164           Speculatively hoist loads from both branches of an if-then-else if
8165           the loads are from adjacent locations in the same structure and the
8166           target architecture has a conditional move instruction.  This flag
8167           is enabled by default at -O2 and higher.
8168
8169       -ftree-copy-prop
8170           Perform copy propagation on trees.  This pass eliminates
8171           unnecessary copy operations.  This flag is enabled by default at -O
8172           and higher.
8173
8174       -fipa-pure-const
8175           Discover which functions are pure or constant.  Enabled by default
8176           at -O and higher.
8177
8178       -fipa-reference
8179           Discover which static variables do not escape the compilation unit.
8180           Enabled by default at -O and higher.
8181
8182       -fipa-reference-addressable
8183           Discover read-only, write-only and non-addressable static
8184           variables.  Enabled by default at -O and higher.
8185
8186       -fipa-stack-alignment
8187           Reduce stack alignment on call sites if possible.  Enabled by
8188           default.
8189
8190       -fipa-pta
8191           Perform interprocedural pointer analysis and interprocedural
8192           modification and reference analysis.  This option can cause
8193           excessive memory and compile-time usage on large compilation units.
8194           It is not enabled by default at any optimization level.
8195
8196       -fipa-profile
8197           Perform interprocedural profile propagation.  The functions called
8198           only from cold functions are marked as cold. Also functions
8199           executed once (such as "cold", "noreturn", static constructors or
8200           destructors) are identified. Cold functions and loop less parts of
8201           functions executed once are then optimized for size.  Enabled by
8202           default at -O and higher.
8203
8204       -fipa-cp
8205           Perform interprocedural constant propagation.  This optimization
8206           analyzes the program to determine when values passed to functions
8207           are constants and then optimizes accordingly.  This optimization
8208           can substantially increase performance if the application has
8209           constants passed to functions.  This flag is enabled by default at
8210           -O2, -Os and -O3.  It is also enabled by -fprofile-use and
8211           -fauto-profile.
8212
8213       -fipa-cp-clone
8214           Perform function cloning to make interprocedural constant
8215           propagation stronger.  When enabled, interprocedural constant
8216           propagation performs function cloning when externally visible
8217           function can be called with constant arguments.  Because this
8218           optimization can create multiple copies of functions, it may
8219           significantly increase code size (see --param
8220           ipa-cp-unit-growth=value).  This flag is enabled by default at -O3.
8221           It is also enabled by -fprofile-use and -fauto-profile.
8222
8223       -fipa-bit-cp
8224           When enabled, perform interprocedural bitwise constant propagation.
8225           This flag is enabled by default at -O2 and by -fprofile-use and
8226           -fauto-profile.  It requires that -fipa-cp is enabled.
8227
8228       -fipa-vrp
8229           When enabled, perform interprocedural propagation of value ranges.
8230           This flag is enabled by default at -O2. It requires that -fipa-cp
8231           is enabled.
8232
8233       -fipa-icf
8234           Perform Identical Code Folding for functions and read-only
8235           variables.  The optimization reduces code size and may disturb
8236           unwind stacks by replacing a function by equivalent one with a
8237           different name. The optimization works more effectively with link-
8238           time optimization enabled.
8239
8240           Although the behavior is similar to the Gold Linker's ICF
8241           optimization, GCC ICF works on different levels and thus the
8242           optimizations are not same - there are equivalences that are found
8243           only by GCC and equivalences found only by Gold.
8244
8245           This flag is enabled by default at -O2 and -Os.
8246
8247       -flive-patching=level
8248           Control GCC's optimizations to produce output suitable for live-
8249           patching.
8250
8251           If the compiler's optimization uses a function's body or
8252           information extracted from its body to optimize/change another
8253           function, the latter is called an impacted function of the former.
8254           If a function is patched, its impacted functions should be patched
8255           too.
8256
8257           The impacted functions are determined by the compiler's
8258           interprocedural optimizations.  For example, a caller is impacted
8259           when inlining a function into its caller, cloning a function and
8260           changing its caller to call this new clone, or extracting a
8261           function's pureness/constness information to optimize its direct or
8262           indirect callers, etc.
8263
8264           Usually, the more IPA optimizations enabled, the larger the number
8265           of impacted functions for each function.  In order to control the
8266           number of impacted functions and more easily compute the list of
8267           impacted function, IPA optimizations can be partially enabled at
8268           two different levels.
8269
8270           The level argument should be one of the following:
8271
8272           inline-clone
8273               Only enable inlining and cloning optimizations, which includes
8274               inlining, cloning, interprocedural scalar replacement of
8275               aggregates and partial inlining.  As a result, when patching a
8276               function, all its callers and its clones' callers are impacted,
8277               therefore need to be patched as well.
8278
8279               -flive-patching=inline-clone disables the following
8280               optimization flags: -fwhole-program  -fipa-pta  -fipa-reference
8281               -fipa-ra -fipa-icf  -fipa-icf-functions  -fipa-icf-variables
8282               -fipa-bit-cp  -fipa-vrp  -fipa-pure-const
8283               -fipa-reference-addressable -fipa-stack-alignment
8284
8285           inline-only-static
8286               Only enable inlining of static functions.  As a result, when
8287               patching a static function, all its callers are impacted and so
8288               need to be patched as well.
8289
8290               In addition to all the flags that -flive-patching=inline-clone
8291               disables, -flive-patching=inline-only-static disables the
8292               following additional optimization flags: -fipa-cp-clone
8293               -fipa-sra  -fpartial-inlining  -fipa-cp
8294
8295           When -flive-patching is specified without any value, the default
8296           value is inline-clone.
8297
8298           This flag is disabled by default.
8299
8300           Note that -flive-patching is not supported with link-time
8301           optimization (-flto).
8302
8303       -fisolate-erroneous-paths-dereference
8304           Detect paths that trigger erroneous or undefined behavior due to
8305           dereferencing a null pointer.  Isolate those paths from the main
8306           control flow and turn the statement with erroneous or undefined
8307           behavior into a trap.  This flag is enabled by default at -O2 and
8308           higher and depends on -fdelete-null-pointer-checks also being
8309           enabled.
8310
8311       -fisolate-erroneous-paths-attribute
8312           Detect paths that trigger erroneous or undefined behavior due to a
8313           null value being used in a way forbidden by a "returns_nonnull" or
8314           "nonnull" attribute.  Isolate those paths from the main control
8315           flow and turn the statement with erroneous or undefined behavior
8316           into a trap.  This is not currently enabled, but may be enabled by
8317           -O2 in the future.
8318
8319       -ftree-sink
8320           Perform forward store motion on trees.  This flag is enabled by
8321           default at -O and higher.
8322
8323       -ftree-bit-ccp
8324           Perform sparse conditional bit constant propagation on trees and
8325           propagate pointer alignment information.  This pass only operates
8326           on local scalar variables and is enabled by default at -O1 and
8327           higher, except for -Og.  It requires that -ftree-ccp is enabled.
8328
8329       -ftree-ccp
8330           Perform sparse conditional constant propagation (CCP) on trees.
8331           This pass only operates on local scalar variables and is enabled by
8332           default at -O and higher.
8333
8334       -fssa-backprop
8335           Propagate information about uses of a value up the definition chain
8336           in order to simplify the definitions.  For example, this pass
8337           strips sign operations if the sign of a value never matters.  The
8338           flag is enabled by default at -O and higher.
8339
8340       -fssa-phiopt
8341           Perform pattern matching on SSA PHI nodes to optimize conditional
8342           code.  This pass is enabled by default at -O1 and higher, except
8343           for -Og.
8344
8345       -ftree-switch-conversion
8346           Perform conversion of simple initializations in a switch to
8347           initializations from a scalar array.  This flag is enabled by
8348           default at -O2 and higher.
8349
8350       -ftree-tail-merge
8351           Look for identical code sequences.  When found, replace one with a
8352           jump to the other.  This optimization is known as tail merging or
8353           cross jumping.  This flag is enabled by default at -O2 and higher.
8354           The compilation time in this pass can be limited using max-tail-
8355           merge-comparisons parameter and max-tail-merge-iterations
8356           parameter.
8357
8358       -ftree-dce
8359           Perform dead code elimination (DCE) on trees.  This flag is enabled
8360           by default at -O and higher.
8361
8362       -ftree-builtin-call-dce
8363           Perform conditional dead code elimination (DCE) for calls to built-
8364           in functions that may set "errno" but are otherwise free of side
8365           effects.  This flag is enabled by default at -O2 and higher if -Os
8366           is not also specified.
8367
8368       -ffinite-loops
8369           Assume that a loop with an exit will eventually take the exit and
8370           not loop indefinitely.  This allows the compiler to remove loops
8371           that otherwise have no side-effects, not considering eventual
8372           endless looping as such.
8373
8374           This option is enabled by default at -O2 for C++ with -std=c++11 or
8375           higher.
8376
8377       -ftree-dominator-opts
8378           Perform a variety of simple scalar cleanups (constant/copy
8379           propagation, redundancy elimination, range propagation and
8380           expression simplification) based on a dominator tree traversal.
8381           This also performs jump threading (to reduce jumps to jumps). This
8382           flag is enabled by default at -O and higher.
8383
8384       -ftree-dse
8385           Perform dead store elimination (DSE) on trees.  A dead store is a
8386           store into a memory location that is later overwritten by another
8387           store without any intervening loads.  In this case the earlier
8388           store can be deleted.  This flag is enabled by default at -O and
8389           higher.
8390
8391       -ftree-ch
8392           Perform loop header copying on trees.  This is beneficial since it
8393           increases effectiveness of code motion optimizations.  It also
8394           saves one jump.  This flag is enabled by default at -O and higher.
8395           It is not enabled for -Os, since it usually increases code size.
8396
8397       -ftree-loop-optimize
8398           Perform loop optimizations on trees.  This flag is enabled by
8399           default at -O and higher.
8400
8401       -ftree-loop-linear
8402       -floop-strip-mine
8403       -floop-block
8404           Perform loop nest optimizations.  Same as -floop-nest-optimize.  To
8405           use this code transformation, GCC has to be configured with
8406           --with-isl to enable the Graphite loop transformation
8407           infrastructure.
8408
8409       -fgraphite-identity
8410           Enable the identity transformation for graphite.  For every SCoP we
8411           generate the polyhedral representation and transform it back to
8412           gimple.  Using -fgraphite-identity we can check the costs or
8413           benefits of the GIMPLE -> GRAPHITE -> GIMPLE transformation.  Some
8414           minimal optimizations are also performed by the code generator isl,
8415           like index splitting and dead code elimination in loops.
8416
8417       -floop-nest-optimize
8418           Enable the isl based loop nest optimizer.  This is a generic loop
8419           nest optimizer based on the Pluto optimization algorithms.  It
8420           calculates a loop structure optimized for data-locality and
8421           parallelism.  This option is experimental.
8422
8423       -floop-parallelize-all
8424           Use the Graphite data dependence analysis to identify loops that
8425           can be parallelized.  Parallelize all the loops that can be
8426           analyzed to not contain loop carried dependences without checking
8427           that it is profitable to parallelize the loops.
8428
8429       -ftree-coalesce-vars
8430           While transforming the program out of the SSA representation,
8431           attempt to reduce copying by coalescing versions of different user-
8432           defined variables, instead of just compiler temporaries.  This may
8433           severely limit the ability to debug an optimized program compiled
8434           with -fno-var-tracking-assignments.  In the negated form, this flag
8435           prevents SSA coalescing of user variables.  This option is enabled
8436           by default if optimization is enabled, and it does very little
8437           otherwise.
8438
8439       -ftree-loop-if-convert
8440           Attempt to transform conditional jumps in the innermost loops to
8441           branch-less equivalents.  The intent is to remove control-flow from
8442           the innermost loops in order to improve the ability of the
8443           vectorization pass to handle these loops.  This is enabled by
8444           default if vectorization is enabled.
8445
8446       -ftree-loop-distribution
8447           Perform loop distribution.  This flag can improve cache performance
8448           on big loop bodies and allow further loop optimizations, like
8449           parallelization or vectorization, to take place.  For example, the
8450           loop
8451
8452                   DO I = 1, N
8453                     A(I) = B(I) + C
8454                     D(I) = E(I) * F
8455                   ENDDO
8456
8457           is transformed to
8458
8459                   DO I = 1, N
8460                      A(I) = B(I) + C
8461                   ENDDO
8462                   DO I = 1, N
8463                      D(I) = E(I) * F
8464                   ENDDO
8465
8466           This flag is enabled by default at -O3.  It is also enabled by
8467           -fprofile-use and -fauto-profile.
8468
8469       -ftree-loop-distribute-patterns
8470           Perform loop distribution of patterns that can be code generated
8471           with calls to a library.  This flag is enabled by default at -O2
8472           and higher, and by -fprofile-use and -fauto-profile.
8473
8474           This pass distributes the initialization loops and generates a call
8475           to memset zero.  For example, the loop
8476
8477                   DO I = 1, N
8478                     A(I) = 0
8479                     B(I) = A(I) + I
8480                   ENDDO
8481
8482           is transformed to
8483
8484                   DO I = 1, N
8485                      A(I) = 0
8486                   ENDDO
8487                   DO I = 1, N
8488                      B(I) = A(I) + I
8489                   ENDDO
8490
8491           and the initialization loop is transformed into a call to memset
8492           zero.  This flag is enabled by default at -O3.  It is also enabled
8493           by -fprofile-use and -fauto-profile.
8494
8495       -floop-interchange
8496           Perform loop interchange outside of graphite.  This flag can
8497           improve cache performance on loop nest and allow further loop
8498           optimizations, like vectorization, to take place.  For example, the
8499           loop
8500
8501                   for (int i = 0; i < N; i++)
8502                     for (int j = 0; j < N; j++)
8503                       for (int k = 0; k < N; k++)
8504                         c[i][j] = c[i][j] + a[i][k]*b[k][j];
8505
8506           is transformed to
8507
8508                   for (int i = 0; i < N; i++)
8509                     for (int k = 0; k < N; k++)
8510                       for (int j = 0; j < N; j++)
8511                         c[i][j] = c[i][j] + a[i][k]*b[k][j];
8512
8513           This flag is enabled by default at -O3.  It is also enabled by
8514           -fprofile-use and -fauto-profile.
8515
8516       -floop-unroll-and-jam
8517           Apply unroll and jam transformations on feasible loops.  In a loop
8518           nest this unrolls the outer loop by some factor and fuses the
8519           resulting multiple inner loops.  This flag is enabled by default at
8520           -O3.  It is also enabled by -fprofile-use and -fauto-profile.
8521
8522       -ftree-loop-im
8523           Perform loop invariant motion on trees.  This pass moves only
8524           invariants that are hard to handle at RTL level (function calls,
8525           operations that expand to nontrivial sequences of insns).  With
8526           -funswitch-loops it also moves operands of conditions that are
8527           invariant out of the loop, so that we can use just trivial
8528           invariantness analysis in loop unswitching.  The pass also includes
8529           store motion.
8530
8531       -ftree-loop-ivcanon
8532           Create a canonical counter for number of iterations in loops for
8533           which determining number of iterations requires complicated
8534           analysis.  Later optimizations then may determine the number
8535           easily.  Useful especially in connection with unrolling.
8536
8537       -ftree-scev-cprop
8538           Perform final value replacement.  If a variable is modified in a
8539           loop in such a way that its value when exiting the loop can be
8540           determined using only its initial value and the number of loop
8541           iterations, replace uses of the final value by such a computation,
8542           provided it is sufficiently cheap.  This reduces data dependencies
8543           and may allow further simplifications.  Enabled by default at -O
8544           and higher.
8545
8546       -fivopts
8547           Perform induction variable optimizations (strength reduction,
8548           induction variable merging and induction variable elimination) on
8549           trees.
8550
8551       -ftree-parallelize-loops=n
8552           Parallelize loops, i.e., split their iteration space to run in n
8553           threads.  This is only possible for loops whose iterations are
8554           independent and can be arbitrarily reordered.  The optimization is
8555           only profitable on multiprocessor machines, for loops that are CPU-
8556           intensive, rather than constrained e.g. by memory bandwidth.  This
8557           option implies -pthread, and thus is only supported on targets that
8558           have support for -pthread.
8559
8560       -ftree-pta
8561           Perform function-local points-to analysis on trees.  This flag is
8562           enabled by default at -O1 and higher, except for -Og.
8563
8564       -ftree-sra
8565           Perform scalar replacement of aggregates.  This pass replaces
8566           structure references with scalars to prevent committing structures
8567           to memory too early.  This flag is enabled by default at -O1 and
8568           higher, except for -Og.
8569
8570       -fstore-merging
8571           Perform merging of narrow stores to consecutive memory addresses.
8572           This pass merges contiguous stores of immediate values narrower
8573           than a word into fewer wider stores to reduce the number of
8574           instructions.  This is enabled by default at -O2 and higher as well
8575           as -Os.
8576
8577       -ftree-ter
8578           Perform temporary expression replacement during the SSA->normal
8579           phase.  Single use/single def temporaries are replaced at their use
8580           location with their defining expression.  This results in non-
8581           GIMPLE code, but gives the expanders much more complex trees to
8582           work on resulting in better RTL generation.  This is enabled by
8583           default at -O and higher.
8584
8585       -ftree-slsr
8586           Perform straight-line strength reduction on trees.  This recognizes
8587           related expressions involving multiplications and replaces them by
8588           less expensive calculations when possible.  This is enabled by
8589           default at -O and higher.
8590
8591       -ftree-vectorize
8592           Perform vectorization on trees. This flag enables
8593           -ftree-loop-vectorize and -ftree-slp-vectorize if not explicitly
8594           specified.
8595
8596       -ftree-loop-vectorize
8597           Perform loop vectorization on trees. This flag is enabled by
8598           default at -O3 and by -ftree-vectorize, -fprofile-use, and
8599           -fauto-profile.
8600
8601       -ftree-slp-vectorize
8602           Perform basic block vectorization on trees. This flag is enabled by
8603           default at -O3 and by -ftree-vectorize, -fprofile-use, and
8604           -fauto-profile.
8605
8606       -fvect-cost-model=model
8607           Alter the cost model used for vectorization.  The model argument
8608           should be one of unlimited, dynamic or cheap.  With the unlimited
8609           model the vectorized code-path is assumed to be profitable while
8610           with the dynamic model a runtime check guards the vectorized code-
8611           path to enable it only for iteration counts that will likely
8612           execute faster than when executing the original scalar loop.  The
8613           cheap model disables vectorization of loops where doing so would be
8614           cost prohibitive for example due to required runtime checks for
8615           data dependence or alignment but otherwise is equal to the dynamic
8616           model.  The default cost model depends on other optimization flags
8617           and is either dynamic or cheap.
8618
8619       -fsimd-cost-model=model
8620           Alter the cost model used for vectorization of loops marked with
8621           the OpenMP simd directive.  The model argument should be one of
8622           unlimited, dynamic, cheap.  All values of model have the same
8623           meaning as described in -fvect-cost-model and by default a cost
8624           model defined with -fvect-cost-model is used.
8625
8626       -ftree-vrp
8627           Perform Value Range Propagation on trees.  This is similar to the
8628           constant propagation pass, but instead of values, ranges of values
8629           are propagated.  This allows the optimizers to remove unnecessary
8630           range checks like array bound checks and null pointer checks.  This
8631           is enabled by default at -O2 and higher.  Null pointer check
8632           elimination is only done if -fdelete-null-pointer-checks is
8633           enabled.
8634
8635       -fsplit-paths
8636           Split paths leading to loop backedges.  This can improve dead code
8637           elimination and common subexpression elimination.  This is enabled
8638           by default at -O3 and above.
8639
8640       -fsplit-ivs-in-unroller
8641           Enables expression of values of induction variables in later
8642           iterations of the unrolled loop using the value in the first
8643           iteration.  This breaks long dependency chains, thus improving
8644           efficiency of the scheduling passes.
8645
8646           A combination of -fweb and CSE is often sufficient to obtain the
8647           same effect.  However, that is not reliable in cases where the loop
8648           body is more complicated than a single basic block.  It also does
8649           not work at all on some architectures due to restrictions in the
8650           CSE pass.
8651
8652           This optimization is enabled by default.
8653
8654       -fvariable-expansion-in-unroller
8655           With this option, the compiler creates multiple copies of some
8656           local variables when unrolling a loop, which can result in superior
8657           code.
8658
8659           This optimization is enabled by default for PowerPC targets, but
8660           disabled by default otherwise.
8661
8662       -fpartial-inlining
8663           Inline parts of functions.  This option has any effect only when
8664           inlining itself is turned on by the -finline-functions or
8665           -finline-small-functions options.
8666
8667           Enabled at levels -O2, -O3, -Os.
8668
8669       -fpredictive-commoning
8670           Perform predictive commoning optimization, i.e., reusing
8671           computations (especially memory loads and stores) performed in
8672           previous iterations of loops.
8673
8674           This option is enabled at level -O3.  It is also enabled by
8675           -fprofile-use and -fauto-profile.
8676
8677       -fprefetch-loop-arrays
8678           If supported by the target machine, generate instructions to
8679           prefetch memory to improve the performance of loops that access
8680           large arrays.
8681
8682           This option may generate better or worse code; results are highly
8683           dependent on the structure of loops within the source code.
8684
8685           Disabled at level -Os.
8686
8687       -fno-printf-return-value
8688           Do not substitute constants for known return value of formatted
8689           output functions such as "sprintf", "snprintf", "vsprintf", and
8690           "vsnprintf" (but not "printf" of "fprintf").  This transformation
8691           allows GCC to optimize or even eliminate branches based on the
8692           known return value of these functions called with arguments that
8693           are either constant, or whose values are known to be in a range
8694           that makes determining the exact return value possible.  For
8695           example, when -fprintf-return-value is in effect, both the branch
8696           and the body of the "if" statement (but not the call to "snprint")
8697           can be optimized away when "i" is a 32-bit or smaller integer
8698           because the return value is guaranteed to be at most 8.
8699
8700                   char buf[9];
8701                   if (snprintf (buf, "%08x", i) >= sizeof buf)
8702                     ...
8703
8704           The -fprintf-return-value option relies on other optimizations and
8705           yields best results with -O2 and above.  It works in tandem with
8706           the -Wformat-overflow and -Wformat-truncation options.  The
8707           -fprintf-return-value option is enabled by default.
8708
8709       -fno-peephole
8710       -fno-peephole2
8711           Disable any machine-specific peephole optimizations.  The
8712           difference between -fno-peephole and -fno-peephole2 is in how they
8713           are implemented in the compiler; some targets use one, some use the
8714           other, a few use both.
8715
8716           -fpeephole is enabled by default.  -fpeephole2 enabled at levels
8717           -O2, -O3, -Os.
8718
8719       -fno-guess-branch-probability
8720           Do not guess branch probabilities using heuristics.
8721
8722           GCC uses heuristics to guess branch probabilities if they are not
8723           provided by profiling feedback (-fprofile-arcs).  These heuristics
8724           are based on the control flow graph.  If some branch probabilities
8725           are specified by "__builtin_expect", then the heuristics are used
8726           to guess branch probabilities for the rest of the control flow
8727           graph, taking the "__builtin_expect" info into account.  The
8728           interactions between the heuristics and "__builtin_expect" can be
8729           complex, and in some cases, it may be useful to disable the
8730           heuristics so that the effects of "__builtin_expect" are easier to
8731           understand.
8732
8733           It is also possible to specify expected probability of the
8734           expression with "__builtin_expect_with_probability" built-in
8735           function.
8736
8737           The default is -fguess-branch-probability at levels -O, -O2, -O3,
8738           -Os.
8739
8740       -freorder-blocks
8741           Reorder basic blocks in the compiled function in order to reduce
8742           number of taken branches and improve code locality.
8743
8744           Enabled at levels -O, -O2, -O3, -Os.
8745
8746       -freorder-blocks-algorithm=algorithm
8747           Use the specified algorithm for basic block reordering.  The
8748           algorithm argument can be simple, which does not increase code size
8749           (except sometimes due to secondary effects like alignment), or stc,
8750           the "software trace cache" algorithm, which tries to put all often
8751           executed code together, minimizing the number of branches executed
8752           by making extra copies of code.
8753
8754           The default is simple at levels -O, -Os, and stc at levels -O2,
8755           -O3.
8756
8757       -freorder-blocks-and-partition
8758           In addition to reordering basic blocks in the compiled function, in
8759           order to reduce number of taken branches, partitions hot and cold
8760           basic blocks into separate sections of the assembly and .o files,
8761           to improve paging and cache locality performance.
8762
8763           This optimization is automatically turned off in the presence of
8764           exception handling or unwind tables (on targets using
8765           setjump/longjump or target specific scheme), for linkonce sections,
8766           for functions with a user-defined section attribute and on any
8767           architecture that does not support named sections.  When
8768           -fsplit-stack is used this option is not enabled by default (to
8769           avoid linker errors), but may be enabled explicitly (if using a
8770           working linker).
8771
8772           Enabled for x86 at levels -O2, -O3, -Os.
8773
8774       -freorder-functions
8775           Reorder functions in the object file in order to improve code
8776           locality.  This is implemented by using special subsections
8777           ".text.hot" for most frequently executed functions and
8778           ".text.unlikely" for unlikely executed functions.  Reordering is
8779           done by the linker so object file format must support named
8780           sections and linker must place them in a reasonable way.
8781
8782           This option isn't effective unless you either provide profile
8783           feedback (see -fprofile-arcs for details) or manually annotate
8784           functions with "hot" or "cold" attributes.
8785
8786           Enabled at levels -O2, -O3, -Os.
8787
8788       -fstrict-aliasing
8789           Allow the compiler to assume the strictest aliasing rules
8790           applicable to the language being compiled.  For C (and C++), this
8791           activates optimizations based on the type of expressions.  In
8792           particular, an object of one type is assumed never to reside at the
8793           same address as an object of a different type, unless the types are
8794           almost the same.  For example, an "unsigned int" can alias an
8795           "int", but not a "void*" or a "double".  A character type may alias
8796           any other type.
8797
8798           Pay special attention to code like this:
8799
8800                   union a_union {
8801                     int i;
8802                     double d;
8803                   };
8804
8805                   int f() {
8806                     union a_union t;
8807                     t.d = 3.0;
8808                     return t.i;
8809                   }
8810
8811           The practice of reading from a different union member than the one
8812           most recently written to (called "type-punning") is common.  Even
8813           with -fstrict-aliasing, type-punning is allowed, provided the
8814           memory is accessed through the union type.  So, the code above
8815           works as expected.    However, this code might not:
8816
8817                   int f() {
8818                     union a_union t;
8819                     int* ip;
8820                     t.d = 3.0;
8821                     ip = &t.i;
8822                     return *ip;
8823                   }
8824
8825           Similarly, access by taking the address, casting the resulting
8826           pointer and dereferencing the result has undefined behavior, even
8827           if the cast uses a union type, e.g.:
8828
8829                   int f() {
8830                     double d = 3.0;
8831                     return ((union a_union *) &d)->i;
8832                   }
8833
8834           The -fstrict-aliasing option is enabled at levels -O2, -O3, -Os.
8835
8836       -falign-functions
8837       -falign-functions=n
8838       -falign-functions=n:m
8839       -falign-functions=n:m:n2
8840       -falign-functions=n:m:n2:m2
8841           Align the start of functions to the next power-of-two greater than
8842           or equal to n, skipping up to m-1 bytes.  This ensures that at
8843           least the first m bytes of the function can be fetched by the CPU
8844           without crossing an n-byte alignment boundary.
8845
8846           If m is not specified, it defaults to n.
8847
8848           Examples: -falign-functions=32 aligns functions to the next 32-byte
8849           boundary, -falign-functions=24 aligns to the next 32-byte boundary
8850           only if this can be done by skipping 23 bytes or less,
8851           -falign-functions=32:7 aligns to the next 32-byte boundary only if
8852           this can be done by skipping 6 bytes or less.
8853
8854           The second pair of n2:m2 values allows you to specify a secondary
8855           alignment: -falign-functions=64:7:32:3 aligns to the next 64-byte
8856           boundary if this can be done by skipping 6 bytes or less, otherwise
8857           aligns to the next 32-byte boundary if this can be done by skipping
8858           2 bytes or less.  If m2 is not specified, it defaults to n2.
8859
8860           Some assemblers only support this flag when n is a power of two; in
8861           that case, it is rounded up.
8862
8863           -fno-align-functions and -falign-functions=1 are equivalent and
8864           mean that functions are not aligned.
8865
8866           If n is not specified or is zero, use a machine-dependent default.
8867           The maximum allowed n option value is 65536.
8868
8869           Enabled at levels -O2, -O3.
8870
8871       -flimit-function-alignment
8872           If this option is enabled, the compiler tries to avoid
8873           unnecessarily overaligning functions. It attempts to instruct the
8874           assembler to align by the amount specified by -falign-functions,
8875           but not to skip more bytes than the size of the function.
8876
8877       -falign-labels
8878       -falign-labels=n
8879       -falign-labels=n:m
8880       -falign-labels=n:m:n2
8881       -falign-labels=n:m:n2:m2
8882           Align all branch targets to a power-of-two boundary.
8883
8884           Parameters of this option are analogous to the -falign-functions
8885           option.  -fno-align-labels and -falign-labels=1 are equivalent and
8886           mean that labels are not aligned.
8887
8888           If -falign-loops or -falign-jumps are applicable and are greater
8889           than this value, then their values are used instead.
8890
8891           If n is not specified or is zero, use a machine-dependent default
8892           which is very likely to be 1, meaning no alignment.  The maximum
8893           allowed n option value is 65536.
8894
8895           Enabled at levels -O2, -O3.
8896
8897       -falign-loops
8898       -falign-loops=n
8899       -falign-loops=n:m
8900       -falign-loops=n:m:n2
8901       -falign-loops=n:m:n2:m2
8902           Align loops to a power-of-two boundary.  If the loops are executed
8903           many times, this makes up for any execution of the dummy padding
8904           instructions.
8905
8906           If -falign-labels is greater than this value, then its value is
8907           used instead.
8908
8909           Parameters of this option are analogous to the -falign-functions
8910           option.  -fno-align-loops and -falign-loops=1 are equivalent and
8911           mean that loops are not aligned.  The maximum allowed n option
8912           value is 65536.
8913
8914           If n is not specified or is zero, use a machine-dependent default.
8915
8916           Enabled at levels -O2, -O3.
8917
8918       -falign-jumps
8919       -falign-jumps=n
8920       -falign-jumps=n:m
8921       -falign-jumps=n:m:n2
8922       -falign-jumps=n:m:n2:m2
8923           Align branch targets to a power-of-two boundary, for branch targets
8924           where the targets can only be reached by jumping.  In this case, no
8925           dummy operations need be executed.
8926
8927           If -falign-labels is greater than this value, then its value is
8928           used instead.
8929
8930           Parameters of this option are analogous to the -falign-functions
8931           option.  -fno-align-jumps and -falign-jumps=1 are equivalent and
8932           mean that loops are not aligned.
8933
8934           If n is not specified or is zero, use a machine-dependent default.
8935           The maximum allowed n option value is 65536.
8936
8937           Enabled at levels -O2, -O3.
8938
8939       -fno-allocation-dce
8940           Do not remove unused C++ allocations in dead code elimination.
8941
8942       -fallow-store-data-races
8943           Allow the compiler to introduce new data races on stores.
8944
8945           Enabled at level -Ofast.
8946
8947       -funit-at-a-time
8948           This option is left for compatibility reasons. -funit-at-a-time has
8949           no effect, while -fno-unit-at-a-time implies -fno-toplevel-reorder
8950           and -fno-section-anchors.
8951
8952           Enabled by default.
8953
8954       -fno-toplevel-reorder
8955           Do not reorder top-level functions, variables, and "asm"
8956           statements.  Output them in the same order that they appear in the
8957           input file.  When this option is used, unreferenced static
8958           variables are not removed.  This option is intended to support
8959           existing code that relies on a particular ordering.  For new code,
8960           it is better to use attributes when possible.
8961
8962           -ftoplevel-reorder is the default at -O1 and higher, and also at
8963           -O0 if -fsection-anchors is explicitly requested.  Additionally
8964           -fno-toplevel-reorder implies -fno-section-anchors.
8965
8966       -fweb
8967           Constructs webs as commonly used for register allocation purposes
8968           and assign each web individual pseudo register.  This allows the
8969           register allocation pass to operate on pseudos directly, but also
8970           strengthens several other optimization passes, such as CSE, loop
8971           optimizer and trivial dead code remover.  It can, however, make
8972           debugging impossible, since variables no longer stay in a "home
8973           register".
8974
8975           Enabled by default with -funroll-loops.
8976
8977       -fwhole-program
8978           Assume that the current compilation unit represents the whole
8979           program being compiled.  All public functions and variables with
8980           the exception of "main" and those merged by attribute
8981           "externally_visible" become static functions and in effect are
8982           optimized more aggressively by interprocedural optimizers.
8983
8984           This option should not be used in combination with -flto.  Instead
8985           relying on a linker plugin should provide safer and more precise
8986           information.
8987
8988       -flto[=n]
8989           This option runs the standard link-time optimizer.  When invoked
8990           with source code, it generates GIMPLE (one of GCC's internal
8991           representations) and writes it to special ELF sections in the
8992           object file.  When the object files are linked together, all the
8993           function bodies are read from these ELF sections and instantiated
8994           as if they had been part of the same translation unit.
8995
8996           To use the link-time optimizer, -flto and optimization options
8997           should be specified at compile time and during the final link.  It
8998           is recommended that you compile all the files participating in the
8999           same link with the same options and also specify those options at
9000           link time.  For example:
9001
9002                   gcc -c -O2 -flto foo.c
9003                   gcc -c -O2 -flto bar.c
9004                   gcc -o myprog -flto -O2 foo.o bar.o
9005
9006           The first two invocations to GCC save a bytecode representation of
9007           GIMPLE into special ELF sections inside foo.o and bar.o.  The final
9008           invocation reads the GIMPLE bytecode from foo.o and bar.o, merges
9009           the two files into a single internal image, and compiles the result
9010           as usual.  Since both foo.o and bar.o are merged into a single
9011           image, this causes all the interprocedural analyses and
9012           optimizations in GCC to work across the two files as if they were a
9013           single one.  This means, for example, that the inliner is able to
9014           inline functions in bar.o into functions in foo.o and vice-versa.
9015
9016           Another (simpler) way to enable link-time optimization is:
9017
9018                   gcc -o myprog -flto -O2 foo.c bar.c
9019
9020           The above generates bytecode for foo.c and bar.c, merges them
9021           together into a single GIMPLE representation and optimizes them as
9022           usual to produce myprog.
9023
9024           The important thing to keep in mind is that to enable link-time
9025           optimizations you need to use the GCC driver to perform the link
9026           step.  GCC automatically performs link-time optimization if any of
9027           the objects involved were compiled with the -flto command-line
9028           option.  You can always override the automatic decision to do link-
9029           time optimization by passing -fno-lto to the link command.
9030
9031           To make whole program optimization effective, it is necessary to
9032           make certain whole program assumptions.  The compiler needs to know
9033           what functions and variables can be accessed by libraries and
9034           runtime outside of the link-time optimized unit.  When supported by
9035           the linker, the linker plugin (see -fuse-linker-plugin) passes
9036           information to the compiler about used and externally visible
9037           symbols.  When the linker plugin is not available, -fwhole-program
9038           should be used to allow the compiler to make these assumptions,
9039           which leads to more aggressive optimization decisions.
9040
9041           When a file is compiled with -flto without -fuse-linker-plugin, the
9042           generated object file is larger than a regular object file because
9043           it contains GIMPLE bytecodes and the usual final code (see
9044           -ffat-lto-objects.  This means that object files with LTO
9045           information can be linked as normal object files; if -fno-lto is
9046           passed to the linker, no interprocedural optimizations are applied.
9047           Note that when -fno-fat-lto-objects is enabled the compile stage is
9048           faster but you cannot perform a regular, non-LTO link on them.
9049
9050           When producing the final binary, GCC only applies link-time
9051           optimizations to those files that contain bytecode.  Therefore, you
9052           can mix and match object files and libraries with GIMPLE bytecodes
9053           and final object code.  GCC automatically selects which files to
9054           optimize in LTO mode and which files to link without further
9055           processing.
9056
9057           Generally, options specified at link time override those specified
9058           at compile time, although in some cases GCC attempts to infer link-
9059           time options from the settings used to compile the input files.
9060
9061           If you do not specify an optimization level option -O at link time,
9062           then GCC uses the highest optimization level used when compiling
9063           the object files.  Note that it is generally ineffective to specify
9064           an optimization level option only at link time and not at compile
9065           time, for two reasons.  First, compiling without optimization
9066           suppresses compiler passes that gather information needed for
9067           effective optimization at link time.  Second, some early
9068           optimization passes can be performed only at compile time and not
9069           at link time.
9070
9071           There are some code generation flags preserved by GCC when
9072           generating bytecodes, as they need to be used during the final
9073           link.  Currently, the following options and their settings are
9074           taken from the first object file that explicitly specifies them:
9075           -fPIC, -fpic, -fpie, -fcommon, -fexceptions, -fnon-call-exceptions,
9076           -fgnu-tm and all the -m target flags.
9077
9078           Certain ABI-changing flags are required to match in all compilation
9079           units, and trying to override this at link time with a conflicting
9080           value is ignored.  This includes options such as
9081           -freg-struct-return and -fpcc-struct-return.
9082
9083           Other options such as -ffp-contract, -fno-strict-overflow, -fwrapv,
9084           -fno-trapv or -fno-strict-aliasing are passed through to the link
9085           stage and merged conservatively for conflicting translation units.
9086           Specifically -fno-strict-overflow, -fwrapv and -fno-trapv take
9087           precedence; and for example -ffp-contract=off takes precedence over
9088           -ffp-contract=fast.  You can override them at link time.
9089
9090           Diagnostic options such as -Wstringop-overflow are passed through
9091           to the link stage and their setting matches that of the compile-
9092           step at function granularity.  Note that this matters only for
9093           diagnostics emitted during optimization.  Note that code transforms
9094           such as inlining can lead to warnings being enabled or disabled for
9095           regions if code not consistent with the setting at compile time.
9096
9097           When you need to pass options to the assembler via -Wa or
9098           -Xassembler make sure to either compile such translation units with
9099           -fno-lto or consistently use the same assembler options on all
9100           translation units.  You can alternatively also specify assembler
9101           options at LTO link time.
9102
9103           To enable debug info generation you need to supply -g at compile
9104           time.  If any of the input files at link time were built with debug
9105           info generation enabled the link will enable debug info generation
9106           as well.  Any elaborate debug info settings like the dwarf level
9107           -gdwarf-5 need to be explicitly repeated at the linker command line
9108           and mixing different settings in different translation units is
9109           discouraged.
9110
9111           If LTO encounters objects with C linkage declared with incompatible
9112           types in separate translation units to be linked together
9113           (undefined behavior according to ISO C99 6.2.7), a non-fatal
9114           diagnostic may be issued.  The behavior is still undefined at run
9115           time.  Similar diagnostics may be raised for other languages.
9116
9117           Another feature of LTO is that it is possible to apply
9118           interprocedural optimizations on files written in different
9119           languages:
9120
9121                   gcc -c -flto foo.c
9122                   g++ -c -flto bar.cc
9123                   gfortran -c -flto baz.f90
9124                   g++ -o myprog -flto -O3 foo.o bar.o baz.o -lgfortran
9125
9126           Notice that the final link is done with g++ to get the C++ runtime
9127           libraries and -lgfortran is added to get the Fortran runtime
9128           libraries.  In general, when mixing languages in LTO mode, you
9129           should use the same link command options as when mixing languages
9130           in a regular (non-LTO) compilation.
9131
9132           If object files containing GIMPLE bytecode are stored in a library
9133           archive, say libfoo.a, it is possible to extract and use them in an
9134           LTO link if you are using a linker with plugin support.  To create
9135           static libraries suitable for LTO, use gcc-ar and gcc-ranlib
9136           instead of ar and ranlib; to show the symbols of object files with
9137           GIMPLE bytecode, use gcc-nm.  Those commands require that ar,
9138           ranlib and nm have been compiled with plugin support.  At link
9139           time, use the flag -fuse-linker-plugin to ensure that the library
9140           participates in the LTO optimization process:
9141
9142                   gcc -o myprog -O2 -flto -fuse-linker-plugin a.o b.o -lfoo
9143
9144           With the linker plugin enabled, the linker extracts the needed
9145           GIMPLE files from libfoo.a and passes them on to the running GCC to
9146           make them part of the aggregated GIMPLE image to be optimized.
9147
9148           If you are not using a linker with plugin support and/or do not
9149           enable the linker plugin, then the objects inside libfoo.a are
9150           extracted and linked as usual, but they do not participate in the
9151           LTO optimization process.  In order to make a static library
9152           suitable for both LTO optimization and usual linkage, compile its
9153           object files with -flto -ffat-lto-objects.
9154
9155           Link-time optimizations do not require the presence of the whole
9156           program to operate.  If the program does not require any symbols to
9157           be exported, it is possible to combine -flto and -fwhole-program to
9158           allow the interprocedural optimizers to use more aggressive
9159           assumptions which may lead to improved optimization opportunities.
9160           Use of -fwhole-program is not needed when linker plugin is active
9161           (see -fuse-linker-plugin).
9162
9163           The current implementation of LTO makes no attempt to generate
9164           bytecode that is portable between different types of hosts.  The
9165           bytecode files are versioned and there is a strict version check,
9166           so bytecode files generated in one version of GCC do not work with
9167           an older or newer version of GCC.
9168
9169           Link-time optimization does not work well with generation of
9170           debugging information on systems other than those using a
9171           combination of ELF and DWARF.
9172
9173           If you specify the optional n, the optimization and code generation
9174           done at link time is executed in parallel using n parallel jobs by
9175           utilizing an installed make program.  The environment variable MAKE
9176           may be used to override the program used.
9177
9178           You can also specify -flto=jobserver to use GNU make's job server
9179           mode to determine the number of parallel jobs. This is useful when
9180           the Makefile calling GCC is already executing in parallel.  You
9181           must prepend a + to the command recipe in the parent Makefile for
9182           this to work.  This option likely only works if MAKE is GNU make.
9183           Even without the option value, GCC tries to automatically detect a
9184           running GNU make's job server.
9185
9186           Use -flto=auto to use GNU make's job server, if available, or
9187           otherwise fall back to autodetection of the number of CPU threads
9188           present in your system.
9189
9190       -flto-partition=alg
9191           Specify the partitioning algorithm used by the link-time optimizer.
9192           The value is either 1to1 to specify a partitioning mirroring the
9193           original source files or balanced to specify partitioning into
9194           equally sized chunks (whenever possible) or max to create new
9195           partition for every symbol where possible.  Specifying none as an
9196           algorithm disables partitioning and streaming completely.  The
9197           default value is balanced. While 1to1 can be used as an workaround
9198           for various code ordering issues, the max partitioning is intended
9199           for internal testing only.  The value one specifies that exactly
9200           one partition should be used while the value none bypasses
9201           partitioning and executes the link-time optimization step directly
9202           from the WPA phase.
9203
9204       -flto-compression-level=n
9205           This option specifies the level of compression used for
9206           intermediate language written to LTO object files, and is only
9207           meaningful in conjunction with LTO mode (-flto).  Valid values are
9208           0 (no compression) to 9 (maximum compression).  Values outside this
9209           range are clamped to either 0 or 9.  If the option is not given, a
9210           default balanced compression setting is used.
9211
9212       -fuse-linker-plugin
9213           Enables the use of a linker plugin during link-time optimization.
9214           This option relies on plugin support in the linker, which is
9215           available in gold or in GNU ld 2.21 or newer.
9216
9217           This option enables the extraction of object files with GIMPLE
9218           bytecode out of library archives. This improves the quality of
9219           optimization by exposing more code to the link-time optimizer.
9220           This information specifies what symbols can be accessed externally
9221           (by non-LTO object or during dynamic linking).  Resulting code
9222           quality improvements on binaries (and shared libraries that use
9223           hidden visibility) are similar to -fwhole-program.  See -flto for a
9224           description of the effect of this flag and how to use it.
9225
9226           This option is enabled by default when LTO support in GCC is
9227           enabled and GCC was configured for use with a linker supporting
9228           plugins (GNU ld 2.21 or newer or gold).
9229
9230       -ffat-lto-objects
9231           Fat LTO objects are object files that contain both the intermediate
9232           language and the object code. This makes them usable for both LTO
9233           linking and normal linking. This option is effective only when
9234           compiling with -flto and is ignored at link time.
9235
9236           -fno-fat-lto-objects improves compilation time over plain LTO, but
9237           requires the complete toolchain to be aware of LTO. It requires a
9238           linker with linker plugin support for basic functionality.
9239           Additionally, nm, ar and ranlib need to support linker plugins to
9240           allow a full-featured build environment (capable of building static
9241           libraries etc).  GCC provides the gcc-ar, gcc-nm, gcc-ranlib
9242           wrappers to pass the right options to these tools. With non fat LTO
9243           makefiles need to be modified to use them.
9244
9245           Note that modern binutils provide plugin auto-load mechanism.
9246           Installing the linker plugin into $libdir/bfd-plugins has the same
9247           effect as usage of the command wrappers (gcc-ar, gcc-nm and gcc-
9248           ranlib).
9249
9250           The default is -fno-fat-lto-objects on targets with linker plugin
9251           support.
9252
9253       -fcompare-elim
9254           After register allocation and post-register allocation instruction
9255           splitting, identify arithmetic instructions that compute processor
9256           flags similar to a comparison operation based on that arithmetic.
9257           If possible, eliminate the explicit comparison operation.
9258
9259           This pass only applies to certain targets that cannot explicitly
9260           represent the comparison operation before register allocation is
9261           complete.
9262
9263           Enabled at levels -O, -O2, -O3, -Os.
9264
9265       -fcprop-registers
9266           After register allocation and post-register allocation instruction
9267           splitting, perform a copy-propagation pass to try to reduce
9268           scheduling dependencies and occasionally eliminate the copy.
9269
9270           Enabled at levels -O, -O2, -O3, -Os.
9271
9272       -fprofile-correction
9273           Profiles collected using an instrumented binary for multi-threaded
9274           programs may be inconsistent due to missed counter updates. When
9275           this option is specified, GCC uses heuristics to correct or smooth
9276           out such inconsistencies. By default, GCC emits an error message
9277           when an inconsistent profile is detected.
9278
9279           This option is enabled by -fauto-profile.
9280
9281       -fprofile-partial-training
9282           With "-fprofile-use" all portions of programs not executed during
9283           train run are optimized agressively for size rather than speed.  In
9284           some cases it is not practical to train all possible hot paths in
9285           the program. (For example, program may contain functions specific
9286           for a given hardware and trianing may not cover all hardware
9287           configurations program is run on.)  With
9288           "-fprofile-partial-training" profile feedback will be ignored for
9289           all functions not executed during the train run leading them to be
9290           optimized as if they were compiled without profile feedback. This
9291           leads to better performance when train run is not representative
9292           but also leads to significantly bigger code.
9293
9294       -fprofile-use
9295       -fprofile-use=path
9296           Enable profile feedback-directed optimizations, and the following
9297           optimizations, many of which are generally profitable only with
9298           profile feedback available:
9299
9300           -fbranch-probabilities  -fprofile-values -funroll-loops
9301           -fpeel-loops  -ftracer  -fvpt -finline-functions  -fipa-cp
9302           -fipa-cp-clone  -fipa-bit-cp -fpredictive-commoning  -fsplit-loops
9303           -funswitch-loops -fgcse-after-reload  -ftree-loop-vectorize
9304           -ftree-slp-vectorize -fvect-cost-model=dynamic
9305           -ftree-loop-distribute-patterns -fprofile-reorder-functions
9306
9307           Before you can use this option, you must first generate profiling
9308           information.
9309
9310           By default, GCC emits an error message if the feedback profiles do
9311           not match the source code.  This error can be turned into a warning
9312           by using -Wno-error=coverage-mismatch.  Note this may result in
9313           poorly optimized code.  Additionally, by default, GCC also emits a
9314           warning message if the feedback profiles do not exist (see
9315           -Wmissing-profile).
9316
9317           If path is specified, GCC looks at the path to find the profile
9318           feedback data files. See -fprofile-dir.
9319
9320       -fauto-profile
9321       -fauto-profile=path
9322           Enable sampling-based feedback-directed optimizations, and the
9323           following optimizations, many of which are generally profitable
9324           only with profile feedback available:
9325
9326           -fbranch-probabilities  -fprofile-values -funroll-loops
9327           -fpeel-loops  -ftracer  -fvpt -finline-functions  -fipa-cp
9328           -fipa-cp-clone  -fipa-bit-cp -fpredictive-commoning  -fsplit-loops
9329           -funswitch-loops -fgcse-after-reload  -ftree-loop-vectorize
9330           -ftree-slp-vectorize -fvect-cost-model=dynamic
9331           -ftree-loop-distribute-patterns -fprofile-correction
9332
9333           path is the name of a file containing AutoFDO profile information.
9334           If omitted, it defaults to fbdata.afdo in the current directory.
9335
9336           Producing an AutoFDO profile data file requires running your
9337           program with the perf utility on a supported GNU/Linux target
9338           system.  For more information, see <https://perf.wiki.kernel.org/>.
9339
9340           E.g.
9341
9342                   perf record -e br_inst_retired:near_taken -b -o perf.data \
9343                       -- your_program
9344
9345           Then use the create_gcov tool to convert the raw profile data to a
9346           format that can be used by GCC.  You must also supply the
9347           unstripped binary for your program to this tool.  See
9348           <https://github.com/google/autofdo>.
9349
9350           E.g.
9351
9352                   create_gcov --binary=your_program.unstripped --profile=perf.data \
9353                       --gcov=profile.afdo
9354
9355       The following options control compiler behavior regarding floating-
9356       point arithmetic.  These options trade off between speed and
9357       correctness.  All must be specifically enabled.
9358
9359       -ffloat-store
9360           Do not store floating-point variables in registers, and inhibit
9361           other options that might change whether a floating-point value is
9362           taken from a register or memory.
9363
9364           This option prevents undesirable excess precision on machines such
9365           as the 68000 where the floating registers (of the 68881) keep more
9366           precision than a "double" is supposed to have.  Similarly for the
9367           x86 architecture.  For most programs, the excess precision does
9368           only good, but a few programs rely on the precise definition of
9369           IEEE floating point.  Use -ffloat-store for such programs, after
9370           modifying them to store all pertinent intermediate computations
9371           into variables.
9372
9373       -fexcess-precision=style
9374           This option allows further control over excess precision on
9375           machines where floating-point operations occur in a format with
9376           more precision or range than the IEEE standard and interchange
9377           floating-point types.  By default, -fexcess-precision=fast is in
9378           effect; this means that operations may be carried out in a wider
9379           precision than the types specified in the source if that would
9380           result in faster code, and it is unpredictable when rounding to the
9381           types specified in the source code takes place.  When compiling C,
9382           if -fexcess-precision=standard is specified then excess precision
9383           follows the rules specified in ISO C99; in particular, both casts
9384           and assignments cause values to be rounded to their semantic types
9385           (whereas -ffloat-store only affects assignments).  This option is
9386           enabled by default for C if a strict conformance option such as
9387           -std=c99 is used.  -ffast-math enables -fexcess-precision=fast by
9388           default regardless of whether a strict conformance option is used.
9389
9390           -fexcess-precision=standard is not implemented for languages other
9391           than C.  On the x86, it has no effect if -mfpmath=sse or
9392           -mfpmath=sse+387 is specified; in the former case, IEEE semantics
9393           apply without excess precision, and in the latter, rounding is
9394           unpredictable.
9395
9396       -ffast-math
9397           Sets the options -fno-math-errno, -funsafe-math-optimizations,
9398           -ffinite-math-only, -fno-rounding-math, -fno-signaling-nans,
9399           -fcx-limited-range and -fexcess-precision=fast.
9400
9401           This option causes the preprocessor macro "__FAST_MATH__" to be
9402           defined.
9403
9404           This option is not turned on by any -O option besides -Ofast since
9405           it can result in incorrect output for programs that depend on an
9406           exact implementation of IEEE or ISO rules/specifications for math
9407           functions. It may, however, yield faster code for programs that do
9408           not require the guarantees of these specifications.
9409
9410       -fno-math-errno
9411           Do not set "errno" after calling math functions that are executed
9412           with a single instruction, e.g., "sqrt".  A program that relies on
9413           IEEE exceptions for math error handling may want to use this flag
9414           for speed while maintaining IEEE arithmetic compatibility.
9415
9416           This option is not turned on by any -O option since it can result
9417           in incorrect output for programs that depend on an exact
9418           implementation of IEEE or ISO rules/specifications for math
9419           functions. It may, however, yield faster code for programs that do
9420           not require the guarantees of these specifications.
9421
9422           The default is -fmath-errno.
9423
9424           On Darwin systems, the math library never sets "errno".  There is
9425           therefore no reason for the compiler to consider the possibility
9426           that it might, and -fno-math-errno is the default.
9427
9428       -funsafe-math-optimizations
9429           Allow optimizations for floating-point arithmetic that (a) assume
9430           that arguments and results are valid and (b) may violate IEEE or
9431           ANSI standards.  When used at link time, it may include libraries
9432           or startup files that change the default FPU control word or other
9433           similar optimizations.
9434
9435           This option is not turned on by any -O option since it can result
9436           in incorrect output for programs that depend on an exact
9437           implementation of IEEE or ISO rules/specifications for math
9438           functions. It may, however, yield faster code for programs that do
9439           not require the guarantees of these specifications.  Enables
9440           -fno-signed-zeros, -fno-trapping-math, -fassociative-math and
9441           -freciprocal-math.
9442
9443           The default is -fno-unsafe-math-optimizations.
9444
9445       -fassociative-math
9446           Allow re-association of operands in series of floating-point
9447           operations.  This violates the ISO C and C++ language standard by
9448           possibly changing computation result.  NOTE: re-ordering may change
9449           the sign of zero as well as ignore NaNs and inhibit or create
9450           underflow or overflow (and thus cannot be used on code that relies
9451           on rounding behavior like "(x + 2**52) - 2**52".  May also reorder
9452           floating-point comparisons and thus may not be used when ordered
9453           comparisons are required.  This option requires that both
9454           -fno-signed-zeros and -fno-trapping-math be in effect.  Moreover,
9455           it doesn't make much sense with -frounding-math. For Fortran the
9456           option is automatically enabled when both -fno-signed-zeros and
9457           -fno-trapping-math are in effect.
9458
9459           The default is -fno-associative-math.
9460
9461       -freciprocal-math
9462           Allow the reciprocal of a value to be used instead of dividing by
9463           the value if this enables optimizations.  For example "x / y" can
9464           be replaced with "x * (1/y)", which is useful if "(1/y)" is subject
9465           to common subexpression elimination.  Note that this loses
9466           precision and increases the number of flops operating on the value.
9467
9468           The default is -fno-reciprocal-math.
9469
9470       -ffinite-math-only
9471           Allow optimizations for floating-point arithmetic that assume that
9472           arguments and results are not NaNs or +-Infs.
9473
9474           This option is not turned on by any -O option since it can result
9475           in incorrect output for programs that depend on an exact
9476           implementation of IEEE or ISO rules/specifications for math
9477           functions. It may, however, yield faster code for programs that do
9478           not require the guarantees of these specifications.
9479
9480           The default is -fno-finite-math-only.
9481
9482       -fno-signed-zeros
9483           Allow optimizations for floating-point arithmetic that ignore the
9484           signedness of zero.  IEEE arithmetic specifies the behavior of
9485           distinct +0.0 and -0.0 values, which then prohibits simplification
9486           of expressions such as x+0.0 or 0.0*x (even with
9487           -ffinite-math-only).  This option implies that the sign of a zero
9488           result isn't significant.
9489
9490           The default is -fsigned-zeros.
9491
9492       -fno-trapping-math
9493           Compile code assuming that floating-point operations cannot
9494           generate user-visible traps.  These traps include division by zero,
9495           overflow, underflow, inexact result and invalid operation.  This
9496           option requires that -fno-signaling-nans be in effect.  Setting
9497           this option may allow faster code if one relies on "non-stop" IEEE
9498           arithmetic, for example.
9499
9500           This option should never be turned on by any -O option since it can
9501           result in incorrect output for programs that depend on an exact
9502           implementation of IEEE or ISO rules/specifications for math
9503           functions.
9504
9505           The default is -ftrapping-math.
9506
9507       -frounding-math
9508           Disable transformations and optimizations that assume default
9509           floating-point rounding behavior.  This is round-to-zero for all
9510           floating point to integer conversions, and round-to-nearest for all
9511           other arithmetic truncations.  This option should be specified for
9512           programs that change the FP rounding mode dynamically, or that may
9513           be executed with a non-default rounding mode.  This option disables
9514           constant folding of floating-point expressions at compile time
9515           (which may be affected by rounding mode) and arithmetic
9516           transformations that are unsafe in the presence of sign-dependent
9517           rounding modes.
9518
9519           The default is -fno-rounding-math.
9520
9521           This option is experimental and does not currently guarantee to
9522           disable all GCC optimizations that are affected by rounding mode.
9523           Future versions of GCC may provide finer control of this setting
9524           using C99's "FENV_ACCESS" pragma.  This command-line option will be
9525           used to specify the default state for "FENV_ACCESS".
9526
9527       -fsignaling-nans
9528           Compile code assuming that IEEE signaling NaNs may generate user-
9529           visible traps during floating-point operations.  Setting this
9530           option disables optimizations that may change the number of
9531           exceptions visible with signaling NaNs.  This option implies
9532           -ftrapping-math.
9533
9534           This option causes the preprocessor macro "__SUPPORT_SNAN__" to be
9535           defined.
9536
9537           The default is -fno-signaling-nans.
9538
9539           This option is experimental and does not currently guarantee to
9540           disable all GCC optimizations that affect signaling NaN behavior.
9541
9542       -fno-fp-int-builtin-inexact
9543           Do not allow the built-in functions "ceil", "floor", "round" and
9544           "trunc", and their "float" and "long double" variants, to generate
9545           code that raises the "inexact" floating-point exception for
9546           noninteger arguments.  ISO C99 and C11 allow these functions to
9547           raise the "inexact" exception, but ISO/IEC TS 18661-1:2014, the C
9548           bindings to IEEE 754-2008, as integrated into ISO C2X, does not
9549           allow these functions to do so.
9550
9551           The default is -ffp-int-builtin-inexact, allowing the exception to
9552           be raised, unless C2X or a later C standard is selected.  This
9553           option does nothing unless -ftrapping-math is in effect.
9554
9555           Even if -fno-fp-int-builtin-inexact is used, if the functions
9556           generate a call to a library function then the "inexact" exception
9557           may be raised if the library implementation does not follow TS
9558           18661.
9559
9560       -fsingle-precision-constant
9561           Treat floating-point constants as single precision instead of
9562           implicitly converting them to double-precision constants.
9563
9564       -fcx-limited-range
9565           When enabled, this option states that a range reduction step is not
9566           needed when performing complex division.  Also, there is no
9567           checking whether the result of a complex multiplication or division
9568           is "NaN + I*NaN", with an attempt to rescue the situation in that
9569           case.  The default is -fno-cx-limited-range, but is enabled by
9570           -ffast-math.
9571
9572           This option controls the default setting of the ISO C99
9573           "CX_LIMITED_RANGE" pragma.  Nevertheless, the option applies to all
9574           languages.
9575
9576       -fcx-fortran-rules
9577           Complex multiplication and division follow Fortran rules.  Range
9578           reduction is done as part of complex division, but there is no
9579           checking whether the result of a complex multiplication or division
9580           is "NaN + I*NaN", with an attempt to rescue the situation in that
9581           case.
9582
9583           The default is -fno-cx-fortran-rules.
9584
9585       The following options control optimizations that may improve
9586       performance, but are not enabled by any -O options.  This section
9587       includes experimental options that may produce broken code.
9588
9589       -fbranch-probabilities
9590           After running a program compiled with -fprofile-arcs, you can
9591           compile it a second time using -fbranch-probabilities, to improve
9592           optimizations based on the number of times each branch was taken.
9593           When a program compiled with -fprofile-arcs exits, it saves arc
9594           execution counts to a file called sourcename.gcda for each source
9595           file.  The information in this data file is very dependent on the
9596           structure of the generated code, so you must use the same source
9597           code and the same optimization options for both compilations.
9598
9599           With -fbranch-probabilities, GCC puts a REG_BR_PROB note on each
9600           JUMP_INSN and CALL_INSN.  These can be used to improve
9601           optimization.  Currently, they are only used in one place: in
9602           reorg.c, instead of guessing which path a branch is most likely to
9603           take, the REG_BR_PROB values are used to exactly determine which
9604           path is taken more often.
9605
9606           Enabled by -fprofile-use and -fauto-profile.
9607
9608       -fprofile-values
9609           If combined with -fprofile-arcs, it adds code so that some data
9610           about values of expressions in the program is gathered.
9611
9612           With -fbranch-probabilities, it reads back the data gathered from
9613           profiling values of expressions for usage in optimizations.
9614
9615           Enabled by -fprofile-generate, -fprofile-use, and -fauto-profile.
9616
9617       -fprofile-reorder-functions
9618           Function reordering based on profile instrumentation collects first
9619           time of execution of a function and orders these functions in
9620           ascending order.
9621
9622           Enabled with -fprofile-use.
9623
9624       -fvpt
9625           If combined with -fprofile-arcs, this option instructs the compiler
9626           to add code to gather information about values of expressions.
9627
9628           With -fbranch-probabilities, it reads back the data gathered and
9629           actually performs the optimizations based on them.  Currently the
9630           optimizations include specialization of division operations using
9631           the knowledge about the value of the denominator.
9632
9633           Enabled with -fprofile-use and -fauto-profile.
9634
9635       -frename-registers
9636           Attempt to avoid false dependencies in scheduled code by making use
9637           of registers left over after register allocation.  This
9638           optimization most benefits processors with lots of registers.
9639           Depending on the debug information format adopted by the target,
9640           however, it can make debugging impossible, since variables no
9641           longer stay in a "home register".
9642
9643           Enabled by default with -funroll-loops.
9644
9645       -fschedule-fusion
9646           Performs a target dependent pass over the instruction stream to
9647           schedule instructions of same type together because target machine
9648           can execute them more efficiently if they are adjacent to each
9649           other in the instruction flow.
9650
9651           Enabled at levels -O2, -O3, -Os.
9652
9653       -ftracer
9654           Perform tail duplication to enlarge superblock size.  This
9655           transformation simplifies the control flow of the function allowing
9656           other optimizations to do a better job.
9657
9658           Enabled by -fprofile-use and -fauto-profile.
9659
9660       -funroll-loops
9661           Unroll loops whose number of iterations can be determined at
9662           compile time or upon entry to the loop.  -funroll-loops implies
9663           -frerun-cse-after-loop, -fweb and -frename-registers.  It also
9664           turns on complete loop peeling (i.e. complete removal of loops with
9665           a small constant number of iterations).  This option makes code
9666           larger, and may or may not make it run faster.
9667
9668           Enabled by -fprofile-use and -fauto-profile.
9669
9670       -funroll-all-loops
9671           Unroll all loops, even if their number of iterations is uncertain
9672           when the loop is entered.  This usually makes programs run more
9673           slowly.  -funroll-all-loops implies the same options as
9674           -funroll-loops.
9675
9676       -fpeel-loops
9677           Peels loops for which there is enough information that they do not
9678           roll much (from profile feedback or static analysis).  It also
9679           turns on complete loop peeling (i.e. complete removal of loops with
9680           small constant number of iterations).
9681
9682           Enabled by -O3, -fprofile-use, and -fauto-profile.
9683
9684       -fmove-loop-invariants
9685           Enables the loop invariant motion pass in the RTL loop optimizer.
9686           Enabled at level -O1 and higher, except for -Og.
9687
9688       -fsplit-loops
9689           Split a loop into two if it contains a condition that's always true
9690           for one side of the iteration space and false for the other.
9691
9692           Enabled by -fprofile-use and -fauto-profile.
9693
9694       -funswitch-loops
9695           Move branches with loop invariant conditions out of the loop, with
9696           duplicates of the loop on both branches (modified according to
9697           result of the condition).
9698
9699           Enabled by -fprofile-use and -fauto-profile.
9700
9701       -fversion-loops-for-strides
9702           If a loop iterates over an array with a variable stride, create
9703           another version of the loop that assumes the stride is always one.
9704           For example:
9705
9706                   for (int i = 0; i < n; ++i)
9707                     x[i * stride] = ...;
9708
9709           becomes:
9710
9711                   if (stride == 1)
9712                     for (int i = 0; i < n; ++i)
9713                       x[i] = ...;
9714                   else
9715                     for (int i = 0; i < n; ++i)
9716                       x[i * stride] = ...;
9717
9718           This is particularly useful for assumed-shape arrays in Fortran
9719           where (for example) it allows better vectorization assuming
9720           contiguous accesses.  This flag is enabled by default at -O3.  It
9721           is also enabled by -fprofile-use and -fauto-profile.
9722
9723       -ffunction-sections
9724       -fdata-sections
9725           Place each function or data item into its own section in the output
9726           file if the target supports arbitrary sections.  The name of the
9727           function or the name of the data item determines the section's name
9728           in the output file.
9729
9730           Use these options on systems where the linker can perform
9731           optimizations to improve locality of reference in the instruction
9732           space.  Most systems using the ELF object format have linkers with
9733           such optimizations.  On AIX, the linker rearranges sections
9734           (CSECTs) based on the call graph.  The performance impact varies.
9735
9736           Together with a linker garbage collection (linker --gc-sections
9737           option) these options may lead to smaller statically-linked
9738           executables (after stripping).
9739
9740           On ELF/DWARF systems these options do not degenerate the quality of
9741           the debug information.  There could be issues with other object
9742           files/debug info formats.
9743
9744           Only use these options when there are significant benefits from
9745           doing so.  When you specify these options, the assembler and linker
9746           create larger object and executable files and are also slower.
9747           These options affect code generation.  They prevent optimizations
9748           by the compiler and assembler using relative locations inside a
9749           translation unit since the locations are unknown until link time.
9750           An example of such an optimization is relaxing calls to short call
9751           instructions.
9752
9753       -fstdarg-opt
9754           Optimize the prologue of variadic argument functions with respect
9755           to usage of those arguments.
9756
9757       -fsection-anchors
9758           Try to reduce the number of symbolic address calculations by using
9759           shared "anchor" symbols to address nearby objects.  This
9760           transformation can help to reduce the number of GOT entries and GOT
9761           accesses on some targets.
9762
9763           For example, the implementation of the following function "foo":
9764
9765                   static int a, b, c;
9766                   int foo (void) { return a + b + c; }
9767
9768           usually calculates the addresses of all three variables, but if you
9769           compile it with -fsection-anchors, it accesses the variables from a
9770           common anchor point instead.  The effect is similar to the
9771           following pseudocode (which isn't valid C):
9772
9773                   int foo (void)
9774                   {
9775                     register int *xr = &x;
9776                     return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
9777                   }
9778
9779           Not all targets support this option.
9780
9781       --param name=value
9782           In some places, GCC uses various constants to control the amount of
9783           optimization that is done.  For example, GCC does not inline
9784           functions that contain more than a certain number of instructions.
9785           You can control some of these constants on the command line using
9786           the --param option.
9787
9788           The names of specific parameters, and the meaning of the values,
9789           are tied to the internals of the compiler, and are subject to
9790           change without notice in future releases.
9791
9792           In order to get minimal, maximal and default value of a parameter,
9793           one can use --help=param -Q options.
9794
9795           In each case, the value is an integer.  The following choices of
9796           name are recognized for all targets:
9797
9798           predictable-branch-outcome
9799               When branch is predicted to be taken with probability lower
9800               than this threshold (in percent), then it is considered well
9801               predictable.
9802
9803           max-rtl-if-conversion-insns
9804               RTL if-conversion tries to remove conditional branches around a
9805               block and replace them with conditionally executed
9806               instructions.  This parameter gives the maximum number of
9807               instructions in a block which should be considered for if-
9808               conversion.  The compiler will also use other heuristics to
9809               decide whether if-conversion is likely to be profitable.
9810
9811           max-rtl-if-conversion-predictable-cost
9812           max-rtl-if-conversion-unpredictable-cost
9813               RTL if-conversion will try to remove conditional branches
9814               around a block and replace them with conditionally executed
9815               instructions.  These parameters give the maximum permissible
9816               cost for the sequence that would be generated by if-conversion
9817               depending on whether the branch is statically determined to be
9818               predictable or not.  The units for this parameter are the same
9819               as those for the GCC internal seq_cost metric.  The compiler
9820               will try to provide a reasonable default for this parameter
9821               using the BRANCH_COST target macro.
9822
9823           max-crossjump-edges
9824               The maximum number of incoming edges to consider for cross-
9825               jumping.  The algorithm used by -fcrossjumping is O(N^2) in the
9826               number of edges incoming to each block.  Increasing values mean
9827               more aggressive optimization, making the compilation time
9828               increase with probably small improvement in executable size.
9829
9830           min-crossjump-insns
9831               The minimum number of instructions that must be matched at the
9832               end of two blocks before cross-jumping is performed on them.
9833               This value is ignored in the case where all instructions in the
9834               block being cross-jumped from are matched.
9835
9836           max-grow-copy-bb-insns
9837               The maximum code size expansion factor when copying basic
9838               blocks instead of jumping.  The expansion is relative to a jump
9839               instruction.
9840
9841           max-goto-duplication-insns
9842               The maximum number of instructions to duplicate to a block that
9843               jumps to a computed goto.  To avoid O(N^2) behavior in a number
9844               of passes, GCC factors computed gotos early in the compilation
9845               process, and unfactors them as late as possible.  Only computed
9846               jumps at the end of a basic blocks with no more than max-goto-
9847               duplication-insns are unfactored.
9848
9849           max-delay-slot-insn-search
9850               The maximum number of instructions to consider when looking for
9851               an instruction to fill a delay slot.  If more than this
9852               arbitrary number of instructions are searched, the time savings
9853               from filling the delay slot are minimal, so stop searching.
9854               Increasing values mean more aggressive optimization, making the
9855               compilation time increase with probably small improvement in
9856               execution time.
9857
9858           max-delay-slot-live-search
9859               When trying to fill delay slots, the maximum number of
9860               instructions to consider when searching for a block with valid
9861               live register information.  Increasing this arbitrarily chosen
9862               value means more aggressive optimization, increasing the
9863               compilation time.  This parameter should be removed when the
9864               delay slot code is rewritten to maintain the control-flow
9865               graph.
9866
9867           max-gcse-memory
9868               The approximate maximum amount of memory that can be allocated
9869               in order to perform the global common subexpression elimination
9870               optimization.  If more memory than specified is required, the
9871               optimization is not done.
9872
9873           max-gcse-insertion-ratio
9874               If the ratio of expression insertions to deletions is larger
9875               than this value for any expression, then RTL PRE inserts or
9876               removes the expression and thus leaves partially redundant
9877               computations in the instruction stream.
9878
9879           max-pending-list-length
9880               The maximum number of pending dependencies scheduling allows
9881               before flushing the current state and starting over.  Large
9882               functions with few branches or calls can create excessively
9883               large lists which needlessly consume memory and resources.
9884
9885           max-modulo-backtrack-attempts
9886               The maximum number of backtrack attempts the scheduler should
9887               make when modulo scheduling a loop.  Larger values can
9888               exponentially increase compilation time.
9889
9890           max-inline-insns-single
9891               Several parameters control the tree inliner used in GCC.  This
9892               number sets the maximum number of instructions (counted in
9893               GCC's internal representation) in a single function that the
9894               tree inliner considers for inlining.  This only affects
9895               functions declared inline and methods implemented in a class
9896               declaration (C++).
9897
9898           max-inline-insns-auto
9899               When you use -finline-functions (included in -O3), a lot of
9900               functions that would otherwise not be considered for inlining
9901               by the compiler are investigated.  To those functions, a
9902               different (more restrictive) limit compared to functions
9903               declared inline can be applied (--param max-inline-insns-auto).
9904
9905           max-inline-insns-small
9906               This is bound applied to calls which are considered relevant
9907               with -finline-small-functions.
9908
9909           max-inline-insns-size
9910               This is bound applied to calls which are optimized for size.
9911               Small growth may be desirable to anticipate optimization
9912               oppurtunities exposed by inlining.
9913
9914           uninlined-function-insns
9915               Number of instructions accounted by inliner for function
9916               overhead such as function prologue and epilogue.
9917
9918           uninlined-function-time
9919               Extra time accounted by inliner for function overhead such as
9920               time needed to execute function prologue and epilogue
9921
9922           inline-heuristics-hint-percent
9923               The scale (in percents) applied to inline-insns-single,
9924               inline-insns-single-O2, inline-insns-auto when inline
9925               heuristics hints that inlining is very profitable (will enable
9926               later optimizations).
9927
9928           uninlined-thunk-insns
9929           uninlined-thunk-time
9930               Same as --param uninlined-function-insns and --param uninlined-
9931               function-time but applied to function thunks
9932
9933           inline-min-speedup
9934               When estimated performance improvement of caller + callee
9935               runtime exceeds this threshold (in percent), the function can
9936               be inlined regardless of the limit on --param max-inline-insns-
9937               single and --param max-inline-insns-auto.
9938
9939           large-function-insns
9940               The limit specifying really large functions.  For functions
9941               larger than this limit after inlining, inlining is constrained
9942               by --param large-function-growth.  This parameter is useful
9943               primarily to avoid extreme compilation time caused by non-
9944               linear algorithms used by the back end.
9945
9946           large-function-growth
9947               Specifies maximal growth of large function caused by inlining
9948               in percents.  For example, parameter value 100 limits large
9949               function growth to 2.0 times the original size.
9950
9951           large-unit-insns
9952               The limit specifying large translation unit.  Growth caused by
9953               inlining of units larger than this limit is limited by --param
9954               inline-unit-growth.  For small units this might be too tight.
9955               For example, consider a unit consisting of function A that is
9956               inline and B that just calls A three times.  If B is small
9957               relative to A, the growth of unit is 300\% and yet such
9958               inlining is very sane.  For very large units consisting of
9959               small inlineable functions, however, the overall unit growth
9960               limit is needed to avoid exponential explosion of code size.
9961               Thus for smaller units, the size is increased to --param large-
9962               unit-insns before applying --param inline-unit-growth.
9963
9964           inline-unit-growth
9965               Specifies maximal overall growth of the compilation unit caused
9966               by inlining.  For example, parameter value 20 limits unit
9967               growth to 1.2 times the original size. Cold functions (either
9968               marked cold via an attribute or by profile feedback) are not
9969               accounted into the unit size.
9970
9971           ipa-cp-unit-growth
9972               Specifies maximal overall growth of the compilation unit caused
9973               by interprocedural constant propagation.  For example,
9974               parameter value 10 limits unit growth to 1.1 times the original
9975               size.
9976
9977           large-stack-frame
9978               The limit specifying large stack frames.  While inlining the
9979               algorithm is trying to not grow past this limit too much.
9980
9981           large-stack-frame-growth
9982               Specifies maximal growth of large stack frames caused by
9983               inlining in percents.  For example, parameter value 1000 limits
9984               large stack frame growth to 11 times the original size.
9985
9986           max-inline-insns-recursive
9987           max-inline-insns-recursive-auto
9988               Specifies the maximum number of instructions an out-of-line
9989               copy of a self-recursive inline function can grow into by
9990               performing recursive inlining.
9991
9992               --param max-inline-insns-recursive applies to functions
9993               declared inline.  For functions not declared inline, recursive
9994               inlining happens only when -finline-functions (included in -O3)
9995               is enabled; --param max-inline-insns-recursive-auto applies
9996               instead.
9997
9998           max-inline-recursive-depth
9999           max-inline-recursive-depth-auto
10000               Specifies the maximum recursion depth used for recursive
10001               inlining.
10002
10003               --param max-inline-recursive-depth applies to functions
10004               declared inline.  For functions not declared inline, recursive
10005               inlining happens only when -finline-functions (included in -O3)
10006               is enabled; --param max-inline-recursive-depth-auto applies
10007               instead.
10008
10009           min-inline-recursive-probability
10010               Recursive inlining is profitable only for function having deep
10011               recursion in average and can hurt for function having little
10012               recursion depth by increasing the prologue size or complexity
10013               of function body to other optimizers.
10014
10015               When profile feedback is available (see -fprofile-generate) the
10016               actual recursion depth can be guessed from the probability that
10017               function recurses via a given call expression.  This parameter
10018               limits inlining only to call expressions whose probability
10019               exceeds the given threshold (in percents).
10020
10021           early-inlining-insns
10022               Specify growth that the early inliner can make.  In effect it
10023               increases the amount of inlining for code having a large
10024               abstraction penalty.
10025
10026           max-early-inliner-iterations
10027               Limit of iterations of the early inliner.  This basically
10028               bounds the number of nested indirect calls the early inliner
10029               can resolve.  Deeper chains are still handled by late inlining.
10030
10031           comdat-sharing-probability
10032               Probability (in percent) that C++ inline function with comdat
10033               visibility are shared across multiple compilation units.
10034
10035           profile-func-internal-id
10036               A parameter to control whether to use function internal id in
10037               profile database lookup. If the value is 0, the compiler uses
10038               an id that is based on function assembler name and filename,
10039               which makes old profile data more tolerant to source changes
10040               such as function reordering etc.
10041
10042           min-vect-loop-bound
10043               The minimum number of iterations under which loops are not
10044               vectorized when -ftree-vectorize is used.  The number of
10045               iterations after vectorization needs to be greater than the
10046               value specified by this option to allow vectorization.
10047
10048           gcse-cost-distance-ratio
10049               Scaling factor in calculation of maximum distance an expression
10050               can be moved by GCSE optimizations.  This is currently
10051               supported only in the code hoisting pass.  The bigger the
10052               ratio, the more aggressive code hoisting is with simple
10053               expressions, i.e., the expressions that have cost less than
10054               gcse-unrestricted-cost.  Specifying 0 disables hoisting of
10055               simple expressions.
10056
10057           gcse-unrestricted-cost
10058               Cost, roughly measured as the cost of a single typical machine
10059               instruction, at which GCSE optimizations do not constrain the
10060               distance an expression can travel.  This is currently supported
10061               only in the code hoisting pass.  The lesser the cost, the more
10062               aggressive code hoisting is.  Specifying 0 allows all
10063               expressions to travel unrestricted distances.
10064
10065           max-hoist-depth
10066               The depth of search in the dominator tree for expressions to
10067               hoist.  This is used to avoid quadratic behavior in hoisting
10068               algorithm.  The value of 0 does not limit on the search, but
10069               may slow down compilation of huge functions.
10070
10071           max-tail-merge-comparisons
10072               The maximum amount of similar bbs to compare a bb with.  This
10073               is used to avoid quadratic behavior in tree tail merging.
10074
10075           max-tail-merge-iterations
10076               The maximum amount of iterations of the pass over the function.
10077               This is used to limit compilation time in tree tail merging.
10078
10079           store-merging-allow-unaligned
10080               Allow the store merging pass to introduce unaligned stores if
10081               it is legal to do so.
10082
10083           max-stores-to-merge
10084               The maximum number of stores to attempt to merge into wider
10085               stores in the store merging pass.
10086
10087           max-unrolled-insns
10088               The maximum number of instructions that a loop may have to be
10089               unrolled.  If a loop is unrolled, this parameter also
10090               determines how many times the loop code is unrolled.
10091
10092           max-average-unrolled-insns
10093               The maximum number of instructions biased by probabilities of
10094               their execution that a loop may have to be unrolled.  If a loop
10095               is unrolled, this parameter also determines how many times the
10096               loop code is unrolled.
10097
10098           max-unroll-times
10099               The maximum number of unrollings of a single loop.
10100
10101           max-peeled-insns
10102               The maximum number of instructions that a loop may have to be
10103               peeled.  If a loop is peeled, this parameter also determines
10104               how many times the loop code is peeled.
10105
10106           max-peel-times
10107               The maximum number of peelings of a single loop.
10108
10109           max-peel-branches
10110               The maximum number of branches on the hot path through the
10111               peeled sequence.
10112
10113           max-completely-peeled-insns
10114               The maximum number of insns of a completely peeled loop.
10115
10116           max-completely-peel-times
10117               The maximum number of iterations of a loop to be suitable for
10118               complete peeling.
10119
10120           max-completely-peel-loop-nest-depth
10121               The maximum depth of a loop nest suitable for complete peeling.
10122
10123           max-unswitch-insns
10124               The maximum number of insns of an unswitched loop.
10125
10126           max-unswitch-level
10127               The maximum number of branches unswitched in a single loop.
10128
10129           lim-expensive
10130               The minimum cost of an expensive expression in the loop
10131               invariant motion.
10132
10133           min-loop-cond-split-prob
10134               When FDO profile information is available, min-loop-cond-split-
10135               prob specifies minimum threshold for probability of semi-
10136               invariant condition statement to trigger loop split.
10137
10138           iv-consider-all-candidates-bound
10139               Bound on number of candidates for induction variables, below
10140               which all candidates are considered for each use in induction
10141               variable optimizations.  If there are more candidates than
10142               this, only the most relevant ones are considered to avoid
10143               quadratic time complexity.
10144
10145           iv-max-considered-uses
10146               The induction variable optimizations give up on loops that
10147               contain more induction variable uses.
10148
10149           iv-always-prune-cand-set-bound
10150               If the number of candidates in the set is smaller than this
10151               value, always try to remove unnecessary ivs from the set when
10152               adding a new one.
10153
10154           avg-loop-niter
10155               Average number of iterations of a loop.
10156
10157           dse-max-object-size
10158               Maximum size (in bytes) of objects tracked bytewise by dead
10159               store elimination.  Larger values may result in larger
10160               compilation times.
10161
10162           dse-max-alias-queries-per-store
10163               Maximum number of queries into the alias oracle per store.
10164               Larger values result in larger compilation times and may result
10165               in more removed dead stores.
10166
10167           scev-max-expr-size
10168               Bound on size of expressions used in the scalar evolutions
10169               analyzer.  Large expressions slow the analyzer.
10170
10171           scev-max-expr-complexity
10172               Bound on the complexity of the expressions in the scalar
10173               evolutions analyzer.  Complex expressions slow the analyzer.
10174
10175           max-tree-if-conversion-phi-args
10176               Maximum number of arguments in a PHI supported by TREE if
10177               conversion unless the loop is marked with simd pragma.
10178
10179           vect-max-version-for-alignment-checks
10180               The maximum number of run-time checks that can be performed
10181               when doing loop versioning for alignment in the vectorizer.
10182
10183           vect-max-version-for-alias-checks
10184               The maximum number of run-time checks that can be performed
10185               when doing loop versioning for alias in the vectorizer.
10186
10187           vect-max-peeling-for-alignment
10188               The maximum number of loop peels to enhance access alignment
10189               for vectorizer. Value -1 means no limit.
10190
10191           max-iterations-to-track
10192               The maximum number of iterations of a loop the brute-force
10193               algorithm for analysis of the number of iterations of the loop
10194               tries to evaluate.
10195
10196           hot-bb-count-fraction
10197               The denominator n of fraction 1/n of the maximal execution
10198               count of a basic block in the entire program that a basic block
10199               needs to at least have in order to be considered hot.  The
10200               default is 10000, which means that a basic block is considered
10201               hot if its execution count is greater than 1/10000 of the
10202               maximal execution count.  0 means that it is never considered
10203               hot.  Used in non-LTO mode.
10204
10205           hot-bb-count-ws-permille
10206               The number of most executed permilles, ranging from 0 to 1000,
10207               of the profiled execution of the entire program to which the
10208               execution count of a basic block must be part of in order to be
10209               considered hot.  The default is 990, which means that a basic
10210               block is considered hot if its execution count contributes to
10211               the upper 990 permilles, or 99.0%, of the profiled execution of
10212               the entire program.  0 means that it is never considered hot.
10213               Used in LTO mode.
10214
10215           hot-bb-frequency-fraction
10216               The denominator n of fraction 1/n of the execution frequency of
10217               the entry block of a function that a basic block of this
10218               function needs to at least have in order to be considered hot.
10219               The default is 1000, which means that a basic block is
10220               considered hot in a function if it is executed more frequently
10221               than 1/1000 of the frequency of the entry block of the
10222               function.  0 means that it is never considered hot.
10223
10224           unlikely-bb-count-fraction
10225               The denominator n of fraction 1/n of the number of profiled
10226               runs of the entire program below which the execution count of a
10227               basic block must be in order for the basic block to be
10228               considered unlikely executed.  The default is 20, which means
10229               that a basic block is considered unlikely executed if it is
10230               executed in fewer than 1/20, or 5%, of the runs of the program.
10231               0 means that it is always considered unlikely executed.
10232
10233           max-predicted-iterations
10234               The maximum number of loop iterations we predict statically.
10235               This is useful in cases where a function contains a single loop
10236               with known bound and another loop with unknown bound.  The
10237               known number of iterations is predicted correctly, while the
10238               unknown number of iterations average to roughly 10.  This means
10239               that the loop without bounds appears artificially cold relative
10240               to the other one.
10241
10242           builtin-expect-probability
10243               Control the probability of the expression having the specified
10244               value. This parameter takes a percentage (i.e. 0 ... 100) as
10245               input.
10246
10247           builtin-string-cmp-inline-length
10248               The maximum length of a constant string for a builtin string
10249               cmp call eligible for inlining.
10250
10251           align-threshold
10252               Select fraction of the maximal frequency of executions of a
10253               basic block in a function to align the basic block.
10254
10255           align-loop-iterations
10256               A loop expected to iterate at least the selected number of
10257               iterations is aligned.
10258
10259           tracer-dynamic-coverage
10260           tracer-dynamic-coverage-feedback
10261               This value is used to limit superblock formation once the given
10262               percentage of executed instructions is covered.  This limits
10263               unnecessary code size expansion.
10264
10265               The tracer-dynamic-coverage-feedback parameter is used only
10266               when profile feedback is available.  The real profiles (as
10267               opposed to statically estimated ones) are much less balanced
10268               allowing the threshold to be larger value.
10269
10270           tracer-max-code-growth
10271               Stop tail duplication once code growth has reached given
10272               percentage.  This is a rather artificial limit, as most of the
10273               duplicates are eliminated later in cross jumping, so it may be
10274               set to much higher values than is the desired code growth.
10275
10276           tracer-min-branch-ratio
10277               Stop reverse growth when the reverse probability of best edge
10278               is less than this threshold (in percent).
10279
10280           tracer-min-branch-probability
10281           tracer-min-branch-probability-feedback
10282               Stop forward growth if the best edge has probability lower than
10283               this threshold.
10284
10285               Similarly to tracer-dynamic-coverage two parameters are
10286               provided.  tracer-min-branch-probability-feedback is used for
10287               compilation with profile feedback and tracer-min-branch-
10288               probability compilation without.  The value for compilation
10289               with profile feedback needs to be more conservative (higher) in
10290               order to make tracer effective.
10291
10292           stack-clash-protection-guard-size
10293               Specify the size of the operating system provided stack guard
10294               as 2 raised to num bytes.  Higher values may reduce the number
10295               of explicit probes, but a value larger than the operating
10296               system provided guard will leave code vulnerable to stack clash
10297               style attacks.
10298
10299           stack-clash-protection-probe-interval
10300               Stack clash protection involves probing stack space as it is
10301               allocated.  This param controls the maximum distance between
10302               probes into the stack as 2 raised to num bytes.  Higher values
10303               may reduce the number of explicit probes, but a value larger
10304               than the operating system provided guard will leave code
10305               vulnerable to stack clash style attacks.
10306
10307           max-cse-path-length
10308               The maximum number of basic blocks on path that CSE considers.
10309
10310           max-cse-insns
10311               The maximum number of instructions CSE processes before
10312               flushing.
10313
10314           ggc-min-expand
10315               GCC uses a garbage collector to manage its own memory
10316               allocation.  This parameter specifies the minimum percentage by
10317               which the garbage collector's heap should be allowed to expand
10318               between collections.  Tuning this may improve compilation
10319               speed; it has no effect on code generation.
10320
10321               The default is 30% + 70% * (RAM/1GB) with an upper bound of
10322               100% when RAM >= 1GB.  If "getrlimit" is available, the notion
10323               of "RAM" is the smallest of actual RAM and "RLIMIT_DATA" or
10324               "RLIMIT_AS".  If GCC is not able to calculate RAM on a
10325               particular platform, the lower bound of 30% is used.  Setting
10326               this parameter and ggc-min-heapsize to zero causes a full
10327               collection to occur at every opportunity.  This is extremely
10328               slow, but can be useful for debugging.
10329
10330           ggc-min-heapsize
10331               Minimum size of the garbage collector's heap before it begins
10332               bothering to collect garbage.  The first collection occurs
10333               after the heap expands by ggc-min-expand% beyond ggc-min-
10334               heapsize.  Again, tuning this may improve compilation speed,
10335               and has no effect on code generation.
10336
10337               The default is the smaller of RAM/8, RLIMIT_RSS, or a limit
10338               that tries to ensure that RLIMIT_DATA or RLIMIT_AS are not
10339               exceeded, but with a lower bound of 4096 (four megabytes) and
10340               an upper bound of 131072 (128 megabytes).  If GCC is not able
10341               to calculate RAM on a particular platform, the lower bound is
10342               used.  Setting this parameter very large effectively disables
10343               garbage collection.  Setting this parameter and ggc-min-expand
10344               to zero causes a full collection to occur at every opportunity.
10345
10346           max-reload-search-insns
10347               The maximum number of instruction reload should look backward
10348               for equivalent register.  Increasing values mean more
10349               aggressive optimization, making the compilation time increase
10350               with probably slightly better performance.
10351
10352           max-cselib-memory-locations
10353               The maximum number of memory locations cselib should take into
10354               account.  Increasing values mean more aggressive optimization,
10355               making the compilation time increase with probably slightly
10356               better performance.
10357
10358           max-sched-ready-insns
10359               The maximum number of instructions ready to be issued the
10360               scheduler should consider at any given time during the first
10361               scheduling pass.  Increasing values mean more thorough
10362               searches, making the compilation time increase with probably
10363               little benefit.
10364
10365           max-sched-region-blocks
10366               The maximum number of blocks in a region to be considered for
10367               interblock scheduling.
10368
10369           max-pipeline-region-blocks
10370               The maximum number of blocks in a region to be considered for
10371               pipelining in the selective scheduler.
10372
10373           max-sched-region-insns
10374               The maximum number of insns in a region to be considered for
10375               interblock scheduling.
10376
10377           max-pipeline-region-insns
10378               The maximum number of insns in a region to be considered for
10379               pipelining in the selective scheduler.
10380
10381           min-spec-prob
10382               The minimum probability (in percents) of reaching a source
10383               block for interblock speculative scheduling.
10384
10385           max-sched-extend-regions-iters
10386               The maximum number of iterations through CFG to extend regions.
10387               A value of 0 disables region extensions.
10388
10389           max-sched-insn-conflict-delay
10390               The maximum conflict delay for an insn to be considered for
10391               speculative motion.
10392
10393           sched-spec-prob-cutoff
10394               The minimal probability of speculation success (in percents),
10395               so that speculative insns are scheduled.
10396
10397           sched-state-edge-prob-cutoff
10398               The minimum probability an edge must have for the scheduler to
10399               save its state across it.
10400
10401           sched-mem-true-dep-cost
10402               Minimal distance (in CPU cycles) between store and load
10403               targeting same memory locations.
10404
10405           selsched-max-lookahead
10406               The maximum size of the lookahead window of selective
10407               scheduling.  It is a depth of search for available
10408               instructions.
10409
10410           selsched-max-sched-times
10411               The maximum number of times that an instruction is scheduled
10412               during selective scheduling.  This is the limit on the number
10413               of iterations through which the instruction may be pipelined.
10414
10415           selsched-insns-to-rename
10416               The maximum number of best instructions in the ready list that
10417               are considered for renaming in the selective scheduler.
10418
10419           sms-min-sc
10420               The minimum value of stage count that swing modulo scheduler
10421               generates.
10422
10423           max-last-value-rtl
10424               The maximum size measured as number of RTLs that can be
10425               recorded in an expression in combiner for a pseudo register as
10426               last known value of that register.
10427
10428           max-combine-insns
10429               The maximum number of instructions the RTL combiner tries to
10430               combine.
10431
10432           integer-share-limit
10433               Small integer constants can use a shared data structure,
10434               reducing the compiler's memory usage and increasing its speed.
10435               This sets the maximum value of a shared integer constant.
10436
10437           ssp-buffer-size
10438               The minimum size of buffers (i.e. arrays) that receive stack
10439               smashing protection when -fstack-protection is used.
10440
10441           min-size-for-stack-sharing
10442               The minimum size of variables taking part in stack slot sharing
10443               when not optimizing.
10444
10445           max-jump-thread-duplication-stmts
10446               Maximum number of statements allowed in a block that needs to
10447               be duplicated when threading jumps.
10448
10449           max-fields-for-field-sensitive
10450               Maximum number of fields in a structure treated in a field
10451               sensitive manner during pointer analysis.
10452
10453           prefetch-latency
10454               Estimate on average number of instructions that are executed
10455               before prefetch finishes.  The distance prefetched ahead is
10456               proportional to this constant.  Increasing this number may also
10457               lead to less streams being prefetched (see simultaneous-
10458               prefetches).
10459
10460           simultaneous-prefetches
10461               Maximum number of prefetches that can run at the same time.
10462
10463           l1-cache-line-size
10464               The size of cache line in L1 data cache, in bytes.
10465
10466           l1-cache-size
10467               The size of L1 data cache, in kilobytes.
10468
10469           l2-cache-size
10470               The size of L2 data cache, in kilobytes.
10471
10472           prefetch-dynamic-strides
10473               Whether the loop array prefetch pass should issue software
10474               prefetch hints for strides that are non-constant.  In some
10475               cases this may be beneficial, though the fact the stride is
10476               non-constant may make it hard to predict when there is clear
10477               benefit to issuing these hints.
10478
10479               Set to 1 if the prefetch hints should be issued for non-
10480               constant strides.  Set to 0 if prefetch hints should be issued
10481               only for strides that are known to be constant and below
10482               prefetch-minimum-stride.
10483
10484           prefetch-minimum-stride
10485               Minimum constant stride, in bytes, to start using prefetch
10486               hints for.  If the stride is less than this threshold, prefetch
10487               hints will not be issued.
10488
10489               This setting is useful for processors that have hardware
10490               prefetchers, in which case there may be conflicts between the
10491               hardware prefetchers and the software prefetchers.  If the
10492               hardware prefetchers have a maximum stride they can handle, it
10493               should be used here to improve the use of software prefetchers.
10494
10495               A value of -1 means we don't have a threshold and therefore
10496               prefetch hints can be issued for any constant stride.
10497
10498               This setting is only useful for strides that are known and
10499               constant.
10500
10501           loop-interchange-max-num-stmts
10502               The maximum number of stmts in a loop to be interchanged.
10503
10504           loop-interchange-stride-ratio
10505               The minimum ratio between stride of two loops for interchange
10506               to be profitable.
10507
10508           min-insn-to-prefetch-ratio
10509               The minimum ratio between the number of instructions and the
10510               number of prefetches to enable prefetching in a loop.
10511
10512           prefetch-min-insn-to-mem-ratio
10513               The minimum ratio between the number of instructions and the
10514               number of memory references to enable prefetching in a loop.
10515
10516           use-canonical-types
10517               Whether the compiler should use the "canonical" type system.
10518               Should always be 1, which uses a more efficient internal
10519               mechanism for comparing types in C++ and Objective-C++.
10520               However, if bugs in the canonical type system are causing
10521               compilation failures, set this value to 0 to disable canonical
10522               types.
10523
10524           switch-conversion-max-branch-ratio
10525               Switch initialization conversion refuses to create arrays that
10526               are bigger than switch-conversion-max-branch-ratio times the
10527               number of branches in the switch.
10528
10529           max-partial-antic-length
10530               Maximum length of the partial antic set computed during the
10531               tree partial redundancy elimination optimization (-ftree-pre)
10532               when optimizing at -O3 and above.  For some sorts of source
10533               code the enhanced partial redundancy elimination optimization
10534               can run away, consuming all of the memory available on the host
10535               machine.  This parameter sets a limit on the length of the sets
10536               that are computed, which prevents the runaway behavior.
10537               Setting a value of 0 for this parameter allows an unlimited set
10538               length.
10539
10540           rpo-vn-max-loop-depth
10541               Maximum loop depth that is value-numbered optimistically.  When
10542               the limit hits the innermost rpo-vn-max-loop-depth loops and
10543               the outermost loop in the loop nest are value-numbered
10544               optimistically and the remaining ones not.
10545
10546           sccvn-max-alias-queries-per-access
10547               Maximum number of alias-oracle queries we perform when looking
10548               for redundancies for loads and stores.  If this limit is hit
10549               the search is aborted and the load or store is not considered
10550               redundant.  The number of queries is algorithmically limited to
10551               the number of stores on all paths from the load to the function
10552               entry.
10553
10554           ira-max-loops-num
10555               IRA uses regional register allocation by default.  If a
10556               function contains more loops than the number given by this
10557               parameter, only at most the given number of the most
10558               frequently-executed loops form regions for regional register
10559               allocation.
10560
10561           ira-max-conflict-table-size
10562               Although IRA uses a sophisticated algorithm to compress the
10563               conflict table, the table can still require excessive amounts
10564               of memory for huge functions.  If the conflict table for a
10565               function could be more than the size in MB given by this
10566               parameter, the register allocator instead uses a faster,
10567               simpler, and lower-quality algorithm that does not require
10568               building a pseudo-register conflict table.
10569
10570           ira-loop-reserved-regs
10571               IRA can be used to evaluate more accurate register pressure in
10572               loops for decisions to move loop invariants (see -O3).  The
10573               number of available registers reserved for some other purposes
10574               is given by this parameter.  Default of the parameter is the
10575               best found from numerous experiments.
10576
10577           lra-inheritance-ebb-probability-cutoff
10578               LRA tries to reuse values reloaded in registers in subsequent
10579               insns.  This optimization is called inheritance.  EBB is used
10580               as a region to do this optimization.  The parameter defines a
10581               minimal fall-through edge probability in percentage used to add
10582               BB to inheritance EBB in LRA.  The default value was chosen
10583               from numerous runs of SPEC2000 on x86-64.
10584
10585           loop-invariant-max-bbs-in-loop
10586               Loop invariant motion can be very expensive, both in
10587               compilation time and in amount of needed compile-time memory,
10588               with very large loops.  Loops with more basic blocks than this
10589               parameter won't have loop invariant motion optimization
10590               performed on them.
10591
10592           loop-max-datarefs-for-datadeps
10593               Building data dependencies is expensive for very large loops.
10594               This parameter limits the number of data references in loops
10595               that are considered for data dependence analysis.  These large
10596               loops are no handled by the optimizations using loop data
10597               dependencies.
10598
10599           max-vartrack-size
10600               Sets a maximum number of hash table slots to use during
10601               variable tracking dataflow analysis of any function.  If this
10602               limit is exceeded with variable tracking at assignments
10603               enabled, analysis for that function is retried without it,
10604               after removing all debug insns from the function.  If the limit
10605               is exceeded even without debug insns, var tracking analysis is
10606               completely disabled for the function.  Setting the parameter to
10607               zero makes it unlimited.
10608
10609           max-vartrack-expr-depth
10610               Sets a maximum number of recursion levels when attempting to
10611               map variable names or debug temporaries to value expressions.
10612               This trades compilation time for more complete debug
10613               information.  If this is set too low, value expressions that
10614               are available and could be represented in debug information may
10615               end up not being used; setting this higher may enable the
10616               compiler to find more complex debug expressions, but compile
10617               time and memory use may grow.
10618
10619           max-debug-marker-count
10620               Sets a threshold on the number of debug markers (e.g. begin
10621               stmt markers) to avoid complexity explosion at inlining or
10622               expanding to RTL.  If a function has more such gimple stmts
10623               than the set limit, such stmts will be dropped from the inlined
10624               copy of a function, and from its RTL expansion.
10625
10626           min-nondebug-insn-uid
10627               Use uids starting at this parameter for nondebug insns.  The
10628               range below the parameter is reserved exclusively for debug
10629               insns created by -fvar-tracking-assignments, but debug insns
10630               may get (non-overlapping) uids above it if the reserved range
10631               is exhausted.
10632
10633           ipa-sra-ptr-growth-factor
10634               IPA-SRA replaces a pointer to an aggregate with one or more new
10635               parameters only when their cumulative size is less or equal to
10636               ipa-sra-ptr-growth-factor times the size of the original
10637               pointer parameter.
10638
10639           ipa-sra-max-replacements
10640               Maximum pieces of an aggregate that IPA-SRA tracks.  As a
10641               consequence, it is also the maximum number of replacements of a
10642               formal parameter.
10643
10644           sra-max-scalarization-size-Ospeed
10645           sra-max-scalarization-size-Osize
10646               The two Scalar Reduction of Aggregates passes (SRA and IPA-SRA)
10647               aim to replace scalar parts of aggregates with uses of
10648               independent scalar variables.  These parameters control the
10649               maximum size, in storage units, of aggregate which is
10650               considered for replacement when compiling for speed (sra-max-
10651               scalarization-size-Ospeed) or size (sra-max-scalarization-size-
10652               Osize) respectively.
10653
10654           sra-max-propagations
10655               The maximum number of artificial accesses that Scalar
10656               Replacement of Aggregates (SRA) will track, per one local
10657               variable, in order to facilitate copy propagation.
10658
10659           tm-max-aggregate-size
10660               When making copies of thread-local variables in a transaction,
10661               this parameter specifies the size in bytes after which
10662               variables are saved with the logging functions as opposed to
10663               save/restore code sequence pairs.  This option only applies
10664               when using -fgnu-tm.
10665
10666           graphite-max-nb-scop-params
10667               To avoid exponential effects in the Graphite loop transforms,
10668               the number of parameters in a Static Control Part (SCoP) is
10669               bounded.  A value of zero can be used to lift the bound.  A
10670               variable whose value is unknown at compilation time and defined
10671               outside a SCoP is a parameter of the SCoP.
10672
10673           loop-block-tile-size
10674               Loop blocking or strip mining transforms, enabled with
10675               -floop-block or -floop-strip-mine, strip mine each loop in the
10676               loop nest by a given number of iterations.  The strip length
10677               can be changed using the loop-block-tile-size parameter.
10678
10679           ipa-cp-value-list-size
10680               IPA-CP attempts to track all possible values and types passed
10681               to a function's parameter in order to propagate them and
10682               perform devirtualization.  ipa-cp-value-list-size is the
10683               maximum number of values and types it stores per one formal
10684               parameter of a function.
10685
10686           ipa-cp-eval-threshold
10687               IPA-CP calculates its own score of cloning profitability
10688               heuristics and performs those cloning opportunities with scores
10689               that exceed ipa-cp-eval-threshold.
10690
10691           ipa-cp-max-recursive-depth
10692               Maximum depth of recursive cloning for self-recursive function.
10693
10694           ipa-cp-min-recursive-probability
10695               Recursive cloning only when the probability of call being
10696               executed exceeds the parameter.
10697
10698           ipa-cp-recursion-penalty
10699               Percentage penalty the recursive functions will receive when
10700               they are evaluated for cloning.
10701
10702           ipa-cp-single-call-penalty
10703               Percentage penalty functions containing a single call to
10704               another function will receive when they are evaluated for
10705               cloning.
10706
10707           ipa-max-agg-items
10708               IPA-CP is also capable to propagate a number of scalar values
10709               passed in an aggregate. ipa-max-agg-items controls the maximum
10710               number of such values per one parameter.
10711
10712           ipa-cp-loop-hint-bonus
10713               When IPA-CP determines that a cloning candidate would make the
10714               number of iterations of a loop known, it adds a bonus of ipa-
10715               cp-loop-hint-bonus to the profitability score of the candidate.
10716
10717           ipa-max-aa-steps
10718               During its analysis of function bodies, IPA-CP employs alias
10719               analysis in order to track values pointed to by function
10720               parameters.  In order not spend too much time analyzing huge
10721               functions, it gives up and consider all memory clobbered after
10722               examining ipa-max-aa-steps statements modifying memory.
10723
10724           ipa-max-switch-predicate-bounds
10725               Maximal number of boundary endpoints of case ranges of switch
10726               statement.  For switch exceeding this limit, IPA-CP will not
10727               construct cloning cost predicate, which is used to estimate
10728               cloning benefit, for default case of the switch statement.
10729
10730           ipa-max-param-expr-ops
10731               IPA-CP will analyze conditional statement that references some
10732               function parameter to estimate benefit for cloning upon certain
10733               constant value.  But if number of operations in a parameter
10734               expression exceeds ipa-max-param-expr-ops, the expression is
10735               treated as complicated one, and is not handled by IPA analysis.
10736
10737           lto-partitions
10738               Specify desired number of partitions produced during WHOPR
10739               compilation.  The number of partitions should exceed the number
10740               of CPUs used for compilation.
10741
10742           lto-min-partition
10743               Size of minimal partition for WHOPR (in estimated
10744               instructions).  This prevents expenses of splitting very small
10745               programs into too many partitions.
10746
10747           lto-max-partition
10748               Size of max partition for WHOPR (in estimated instructions).
10749               to provide an upper bound for individual size of partition.
10750               Meant to be used only with balanced partitioning.
10751
10752           lto-max-streaming-parallelism
10753               Maximal number of parallel processes used for LTO streaming.
10754
10755           cxx-max-namespaces-for-diagnostic-help
10756               The maximum number of namespaces to consult for suggestions
10757               when C++ name lookup fails for an identifier.
10758
10759           sink-frequency-threshold
10760               The maximum relative execution frequency (in percents) of the
10761               target block relative to a statement's original block to allow
10762               statement sinking of a statement.  Larger numbers result in
10763               more aggressive statement sinking.  A small positive adjustment
10764               is applied for statements with memory operands as those are
10765               even more profitable so sink.
10766
10767           max-stores-to-sink
10768               The maximum number of conditional store pairs that can be sunk.
10769               Set to 0 if either vectorization (-ftree-vectorize) or if-
10770               conversion (-ftree-loop-if-convert) is disabled.
10771
10772           case-values-threshold
10773               The smallest number of different values for which it is best to
10774               use a jump-table instead of a tree of conditional branches.  If
10775               the value is 0, use the default for the machine.
10776
10777           jump-table-max-growth-ratio-for-size
10778               The maximum code size growth ratio when expanding into a jump
10779               table (in percent).  The parameter is used when optimizing for
10780               size.
10781
10782           jump-table-max-growth-ratio-for-speed
10783               The maximum code size growth ratio when expanding into a jump
10784               table (in percent).  The parameter is used when optimizing for
10785               speed.
10786
10787           tree-reassoc-width
10788               Set the maximum number of instructions executed in parallel in
10789               reassociated tree. This parameter overrides target dependent
10790               heuristics used by default if has non zero value.
10791
10792           sched-pressure-algorithm
10793               Choose between the two available implementations of
10794               -fsched-pressure.  Algorithm 1 is the original implementation
10795               and is the more likely to prevent instructions from being
10796               reordered.  Algorithm 2 was designed to be a compromise between
10797               the relatively conservative approach taken by algorithm 1 and
10798               the rather aggressive approach taken by the default scheduler.
10799               It relies more heavily on having a regular register file and
10800               accurate register pressure classes.  See haifa-sched.c in the
10801               GCC sources for more details.
10802
10803               The default choice depends on the target.
10804
10805           max-slsr-cand-scan
10806               Set the maximum number of existing candidates that are
10807               considered when seeking a basis for a new straight-line
10808               strength reduction candidate.
10809
10810           asan-globals
10811               Enable buffer overflow detection for global objects.  This kind
10812               of protection is enabled by default if you are using
10813               -fsanitize=address option.  To disable global objects
10814               protection use --param asan-globals=0.
10815
10816           asan-stack
10817               Enable buffer overflow detection for stack objects.  This kind
10818               of protection is enabled by default when using
10819               -fsanitize=address.  To disable stack protection use --param
10820               asan-stack=0 option.
10821
10822           asan-instrument-reads
10823               Enable buffer overflow detection for memory reads.  This kind
10824               of protection is enabled by default when using
10825               -fsanitize=address.  To disable memory reads protection use
10826               --param asan-instrument-reads=0.
10827
10828           asan-instrument-writes
10829               Enable buffer overflow detection for memory writes.  This kind
10830               of protection is enabled by default when using
10831               -fsanitize=address.  To disable memory writes protection use
10832               --param asan-instrument-writes=0 option.
10833
10834           asan-memintrin
10835               Enable detection for built-in functions.  This kind of
10836               protection is enabled by default when using -fsanitize=address.
10837               To disable built-in functions protection use --param
10838               asan-memintrin=0.
10839
10840           asan-use-after-return
10841               Enable detection of use-after-return.  This kind of protection
10842               is enabled by default when using the -fsanitize=address option.
10843               To disable it use --param asan-use-after-return=0.
10844
10845               Note: By default the check is disabled at run time.  To enable
10846               it, add "detect_stack_use_after_return=1" to the environment
10847               variable ASAN_OPTIONS.
10848
10849           asan-instrumentation-with-call-threshold
10850               If number of memory accesses in function being instrumented is
10851               greater or equal to this number, use callbacks instead of
10852               inline checks.  E.g. to disable inline code use --param
10853               asan-instrumentation-with-call-threshold=0.
10854
10855           use-after-scope-direct-emission-threshold
10856               If the size of a local variable in bytes is smaller or equal to
10857               this number, directly poison (or unpoison) shadow memory
10858               instead of using run-time callbacks.
10859
10860           max-fsm-thread-path-insns
10861               Maximum number of instructions to copy when duplicating blocks
10862               on a finite state automaton jump thread path.
10863
10864           max-fsm-thread-length
10865               Maximum number of basic blocks on a finite state automaton jump
10866               thread path.
10867
10868           max-fsm-thread-paths
10869               Maximum number of new jump thread paths to create for a finite
10870               state automaton.
10871
10872           parloops-chunk-size
10873               Chunk size of omp schedule for loops parallelized by parloops.
10874
10875           parloops-schedule
10876               Schedule type of omp schedule for loops parallelized by
10877               parloops (static, dynamic, guided, auto, runtime).
10878
10879           parloops-min-per-thread
10880               The minimum number of iterations per thread of an innermost
10881               parallelized loop for which the parallelized variant is
10882               preferred over the single threaded one.  Note that for a
10883               parallelized loop nest the minimum number of iterations of the
10884               outermost loop per thread is two.
10885
10886           max-ssa-name-query-depth
10887               Maximum depth of recursion when querying properties of SSA
10888               names in things like fold routines.  One level of recursion
10889               corresponds to following a use-def chain.
10890
10891           hsa-gen-debug-stores
10892               Enable emission of special debug stores within HSA kernels
10893               which are then read and reported by libgomp plugin.  Generation
10894               of these stores is disabled by default, use --param
10895               hsa-gen-debug-stores=1 to enable it.
10896
10897           max-speculative-devirt-maydefs
10898               The maximum number of may-defs we analyze when looking for a
10899               must-def specifying the dynamic type of an object that invokes
10900               a virtual call we may be able to devirtualize speculatively.
10901
10902           max-vrp-switch-assertions
10903               The maximum number of assertions to add along the default edge
10904               of a switch statement during VRP.
10905
10906           unroll-jam-min-percent
10907               The minimum percentage of memory references that must be
10908               optimized away for the unroll-and-jam transformation to be
10909               considered profitable.
10910
10911           unroll-jam-max-unroll
10912               The maximum number of times the outer loop should be unrolled
10913               by the unroll-and-jam transformation.
10914
10915           max-rtl-if-conversion-unpredictable-cost
10916               Maximum permissible cost for the sequence that would be
10917               generated by the RTL if-conversion pass for a branch that is
10918               considered unpredictable.
10919
10920           max-variable-expansions-in-unroller
10921               If -fvariable-expansion-in-unroller is used, the maximum number
10922               of times that an individual variable will be expanded during
10923               loop unrolling.
10924
10925           tracer-min-branch-probability-feedback
10926               Stop forward growth if the probability of best edge is less
10927               than this threshold (in percent). Used when profile feedback is
10928               available.
10929
10930           partial-inlining-entry-probability
10931               Maximum probability of the entry BB of split region (in percent
10932               relative to entry BB of the function) to make partial inlining
10933               happen.
10934
10935           max-tracked-strlens
10936               Maximum number of strings for which strlen optimization pass
10937               will track string lengths.
10938
10939           gcse-after-reload-partial-fraction
10940               The threshold ratio for performing partial redundancy
10941               elimination after reload.
10942
10943           gcse-after-reload-critical-fraction
10944               The threshold ratio of critical edges execution count that
10945               permit performing redundancy elimination after reload.
10946
10947           max-loop-header-insns
10948               The maximum number of insns in loop header duplicated by the
10949               copy loop headers pass.
10950
10951           vect-epilogues-nomask
10952               Enable loop epilogue vectorization using smaller vector size.
10953
10954           slp-max-insns-in-bb
10955               Maximum number of instructions in basic block to be considered
10956               for SLP vectorization.
10957
10958           avoid-fma-max-bits
10959               Maximum number of bits for which we avoid creating FMAs.
10960
10961           sms-loop-average-count-threshold
10962               A threshold on the average loop count considered by the swing
10963               modulo scheduler.
10964
10965           sms-dfa-history
10966               The number of cycles the swing modulo scheduler considers when
10967               checking conflicts using DFA.
10968
10969           max-inline-insns-recursive-auto
10970               The maximum number of instructions non-inline function can grow
10971               to via recursive inlining.
10972
10973           graphite-allow-codegen-errors
10974               Whether codegen errors should be ICEs when -fchecking.
10975
10976           sms-max-ii-factor
10977               A factor for tuning the upper bound that swing modulo scheduler
10978               uses for scheduling a loop.
10979
10980           lra-max-considered-reload-pseudos
10981               The max number of reload pseudos which are considered during
10982               spilling a non-reload pseudo.
10983
10984           max-pow-sqrt-depth
10985               Maximum depth of sqrt chains to use when synthesizing
10986               exponentiation by a real constant.
10987
10988           max-dse-active-local-stores
10989               Maximum number of active local stores in RTL dead store
10990               elimination.
10991
10992           asan-instrument-allocas
10993               Enable asan allocas/VLAs protection.
10994
10995           max-iterations-computation-cost
10996               Bound on the cost of an expression to compute the number of
10997               iterations.
10998
10999           max-isl-operations
11000               Maximum number of isl operations, 0 means unlimited.
11001
11002           graphite-max-arrays-per-scop
11003               Maximum number of arrays per scop.
11004
11005           max-vartrack-reverse-op-size
11006               Max. size of loc list for which reverse ops should be added.
11007
11008           tracer-dynamic-coverage-feedback
11009               The percentage of function, weighted by execution frequency,
11010               that must be covered by trace formation.  Used when profile
11011               feedback is available.
11012
11013           max-inline-recursive-depth-auto
11014               The maximum depth of recursive inlining for non-inline
11015               functions.
11016
11017           fsm-scale-path-stmts
11018               Scale factor to apply to the number of statements in a
11019               threading path when comparing to the number of (scaled) blocks.
11020
11021           fsm-maximum-phi-arguments
11022               Maximum number of arguments a PHI may have before the FSM
11023               threader will not try to thread through its block.
11024
11025           uninit-control-dep-attempts
11026               Maximum number of nested calls to search for control
11027               dependencies during uninitialized variable analysis.
11028
11029           sra-max-scalarization-size-Osize
11030               Maximum size, in storage units, of an aggregate which should be
11031               considered for scalarization when compiling for size.
11032
11033           fsm-scale-path-blocks
11034               Scale factor to apply to the number of blocks in a threading
11035               path when comparing to the number of (scaled) statements.
11036
11037           sched-autopref-queue-depth
11038               Hardware autoprefetcher scheduler model control flag.  Number
11039               of lookahead cycles the model looks into; at ' ' only enable
11040               instruction sorting heuristic.
11041
11042           loop-versioning-max-inner-insns
11043               The maximum number of instructions that an inner loop can have
11044               before the loop versioning pass considers it too big to copy.
11045
11046           loop-versioning-max-outer-insns
11047               The maximum number of instructions that an outer loop can have
11048               before the loop versioning pass considers it too big to copy,
11049               discounting any instructions in inner loops that directly
11050               benefit from versioning.
11051
11052           ssa-name-def-chain-limit
11053               The maximum number of SSA_NAME assignments to follow in
11054               determining a property of a variable such as its value.  This
11055               limits the number of iterations or recursive calls GCC performs
11056               when optimizing certain statements or when determining their
11057               validity prior to issuing diagnostics.
11058
11059           store-merging-max-size
11060               Maximum size of a single store merging region in bytes.
11061
11062           hash-table-verification-limit
11063               The number of elements for which hash table verification is
11064               done for each searched element.
11065
11066           max-find-base-term-values
11067               Maximum number of VALUEs handled during a single find_base_term
11068               call.
11069
11070           analyzer-max-enodes-per-program-point
11071               The maximum number of exploded nodes per program point within
11072               the analyzer, before terminating analysis of that point.
11073
11074           analyzer-min-snodes-for-call-summary
11075               The minimum number of supernodes within a function for the
11076               analyzer to consider summarizing its effects at call sites.
11077
11078           analyzer-max-recursion-depth
11079               The maximum number of times a callsite can appear in a call
11080               stack within the analyzer, before terminating analysis of a
11081               call that would recurse deeper.
11082
11083           gimple-fe-computed-hot-bb-threshold
11084               The number of executions of a basic block which is considered
11085               hot.  The parameter is used only in GIMPLE FE.
11086
11087           analyzer-bb-explosion-factor
11088               The maximum number of 'after supernode' exploded nodes within
11089               the analyzer per supernode, before terminating analysis.
11090
11091           The following choices of name are available on AArch64 targets:
11092
11093           aarch64-sve-compare-costs
11094               When vectorizing for SVE, consider using "unpacked" vectors for
11095               smaller elements and use the cost model to pick the cheapest
11096               approach.  Also use the cost model to choose between SVE and
11097               Advanced SIMD vectorization.
11098
11099               Using unpacked vectors includes storing smaller elements in
11100               larger containers and accessing elements with extending loads
11101               and truncating stores.
11102
11103           aarch64-float-recp-precision
11104               The number of Newton iterations for calculating the reciprocal
11105               for float type.  The precision of division is proportional to
11106               this param when division approximation is enabled.  The default
11107               value is 1.
11108
11109           aarch64-double-recp-precision
11110               The number of Newton iterations for calculating the reciprocal
11111               for double type.  The precision of division is propotional to
11112               this param when division approximation is enabled.  The default
11113               value is 2.
11114
11115   Program Instrumentation Options
11116       GCC supports a number of command-line options that control adding run-
11117       time instrumentation to the code it normally generates.  For example,
11118       one purpose of instrumentation is collect profiling statistics for use
11119       in finding program hot spots, code coverage analysis, or profile-guided
11120       optimizations.  Another class of program instrumentation is adding run-
11121       time checking to detect programming errors like invalid pointer
11122       dereferences or out-of-bounds array accesses, as well as deliberately
11123       hostile attacks such as stack smashing or C++ vtable hijacking.  There
11124       is also a general hook which can be used to implement other forms of
11125       tracing or function-level instrumentation for debug or program analysis
11126       purposes.
11127
11128       -p
11129       -pg Generate extra code to write profile information suitable for the
11130           analysis program prof (for -p) or gprof (for -pg).  You must use
11131           this option when compiling the source files you want data about,
11132           and you must also use it when linking.
11133
11134           You can use the function attribute "no_instrument_function" to
11135           suppress profiling of individual functions when compiling with
11136           these options.
11137
11138       -fprofile-arcs
11139           Add code so that program flow arcs are instrumented.  During
11140           execution the program records how many times each branch and call
11141           is executed and how many times it is taken or returns.  On targets
11142           that support constructors with priority support, profiling properly
11143           handles constructors, destructors and C++ constructors (and
11144           destructors) of classes which are used as a type of a global
11145           variable.
11146
11147           When the compiled program exits it saves this data to a file called
11148           auxname.gcda for each source file.  The data may be used for
11149           profile-directed optimizations (-fbranch-probabilities), or for
11150           test coverage analysis (-ftest-coverage).  Each object file's
11151           auxname is generated from the name of the output file, if
11152           explicitly specified and it is not the final executable, otherwise
11153           it is the basename of the source file.  In both cases any suffix is
11154           removed (e.g. foo.gcda for input file dir/foo.c, or dir/foo.gcda
11155           for output file specified as -o dir/foo.o).
11156
11157       --coverage
11158           This option is used to compile and link code instrumented for
11159           coverage analysis.  The option is a synonym for -fprofile-arcs
11160           -ftest-coverage (when compiling) and -lgcov (when linking).  See
11161           the documentation for those options for more details.
11162
11163           *   Compile the source files with -fprofile-arcs plus optimization
11164               and code generation options.  For test coverage analysis, use
11165               the additional -ftest-coverage option.  You do not need to
11166               profile every source file in a program.
11167
11168           *   Compile the source files additionally with -fprofile-abs-path
11169               to create absolute path names in the .gcno files.  This allows
11170               gcov to find the correct sources in projects where compilations
11171               occur with different working directories.
11172
11173           *   Link your object files with -lgcov or -fprofile-arcs (the
11174               latter implies the former).
11175
11176           *   Run the program on a representative workload to generate the
11177               arc profile information.  This may be repeated any number of
11178               times.  You can run concurrent instances of your program, and
11179               provided that the file system supports locking, the data files
11180               will be correctly updated.  Unless a strict ISO C dialect
11181               option is in effect, "fork" calls are detected and correctly
11182               handled without double counting.
11183
11184           *   For profile-directed optimizations, compile the source files
11185               again with the same optimization and code generation options
11186               plus -fbranch-probabilities.
11187
11188           *   For test coverage analysis, use gcov to produce human readable
11189               information from the .gcno and .gcda files.  Refer to the gcov
11190               documentation for further information.
11191
11192           With -fprofile-arcs, for each function of your program GCC creates
11193           a program flow graph, then finds a spanning tree for the graph.
11194           Only arcs that are not on the spanning tree have to be
11195           instrumented: the compiler adds code to count the number of times
11196           that these arcs are executed.  When an arc is the only exit or only
11197           entrance to a block, the instrumentation code can be added to the
11198           block; otherwise, a new basic block must be created to hold the
11199           instrumentation code.
11200
11201       -ftest-coverage
11202           Produce a notes file that the gcov code-coverage utility can use to
11203           show program coverage.  Each source file's note file is called
11204           auxname.gcno.  Refer to the -fprofile-arcs option above for a
11205           description of auxname and instructions on how to generate test
11206           coverage data.  Coverage data matches the source files more closely
11207           if you do not optimize.
11208
11209       -fprofile-abs-path
11210           Automatically convert relative source file names to absolute path
11211           names in the .gcno files.  This allows gcov to find the correct
11212           sources in projects where compilations occur with different working
11213           directories.
11214
11215       -fprofile-dir=path
11216           Set the directory to search for the profile data files in to path.
11217           This option affects only the profile data generated by
11218           -fprofile-generate, -ftest-coverage, -fprofile-arcs and used by
11219           -fprofile-use and -fbranch-probabilities and its related options.
11220           Both absolute and relative paths can be used.  By default, GCC uses
11221           the current directory as path, thus the profile data file appears
11222           in the same directory as the object file.  In order to prevent the
11223           file name clashing, if the object file name is not an absolute
11224           path, we mangle the absolute path of the sourcename.gcda file and
11225           use it as the file name of a .gcda file.  See similar option
11226           -fprofile-note.
11227
11228           When an executable is run in a massive parallel environment, it is
11229           recommended to save profile to different folders.  That can be done
11230           with variables in path that are exported during run-time:
11231
11232           %p  process ID.
11233
11234           %q{VAR}
11235               value of environment variable VAR
11236
11237       -fprofile-generate
11238       -fprofile-generate=path
11239           Enable options usually used for instrumenting application to
11240           produce profile useful for later recompilation with profile
11241           feedback based optimization.  You must use -fprofile-generate both
11242           when compiling and when linking your program.
11243
11244           The following options are enabled: -fprofile-arcs,
11245           -fprofile-values, -finline-functions, and -fipa-bit-cp.
11246
11247           If path is specified, GCC looks at the path to find the profile
11248           feedback data files. See -fprofile-dir.
11249
11250           To optimize the program based on the collected profile information,
11251           use -fprofile-use.
11252
11253       -fprofile-note=path
11254           If path is specified, GCC saves .gcno file into path location.  If
11255           you combine the option with multiple source files, the .gcno file
11256           will be overwritten.
11257
11258       -fprofile-prefix-path=path
11259           This option can be used in combination with
11260           profile-generate=profile_dir and profile-use=profile_dir to inform
11261           GCC where is the base directory of built source tree.  By default
11262           profile_dir will contain files with mangled absolute paths of all
11263           object files in the built project.  This is not desirable when
11264           directory used to build the instrumented binary differs from the
11265           directory used to build the binary optimized with profile feedback
11266           because the profile data will not be found during the optimized
11267           build.  In such setups -fprofile-prefix-path=path with path
11268           pointing to the base directory of the build can be used to strip
11269           the irrelevant part of the path and keep all file names relative to
11270           the main build directory.
11271
11272       -fprofile-update=method
11273           Alter the update method for an application instrumented for profile
11274           feedback based optimization.  The method argument should be one of
11275           single, atomic or prefer-atomic.  The first one is useful for
11276           single-threaded applications, while the second one prevents profile
11277           corruption by emitting thread-safe code.
11278
11279           Warning: When an application does not properly join all threads (or
11280           creates an detached thread), a profile file can be still corrupted.
11281
11282           Using prefer-atomic would be transformed either to atomic, when
11283           supported by a target, or to single otherwise.  The GCC driver
11284           automatically selects prefer-atomic when -pthread is present in the
11285           command line.
11286
11287       -fprofile-filter-files=regex
11288           Instrument only functions from files where names match any regular
11289           expression (separated by a semi-colon).
11290
11291           For example, -fprofile-filter-files=main.c;module.*.c will
11292           instrument only main.c and all C files starting with 'module'.
11293
11294       -fprofile-exclude-files=regex
11295           Instrument only functions from files where names do not match all
11296           the regular expressions (separated by a semi-colon).
11297
11298           For example, -fprofile-exclude-files=/usr/* will prevent
11299           instrumentation of all files that are located in /usr/ folder.
11300
11301       -fprofile-reproducible
11302           Control level of reproducibility of profile gathered by
11303           "-fprofile-generate".  This makes it possible to rebuild program
11304           with same outcome which is useful, for example, for distribution
11305           packages.
11306
11307           With -fprofile-reproducibility=serial the profile gathered by
11308           -fprofile-generate is reproducible provided the trained program
11309           behaves the same at each invocation of the train run, it is not
11310           multi-threaded and profile data streaming is always done in the
11311           same order.  Note that profile streaming happens at the end of
11312           program run but also before "fork" function is invoked.
11313
11314           Note that it is quite common that execution counts of some part of
11315           programs depends, for example, on length of temporary file names or
11316           memory space randomization (that may affect hash-table collision
11317           rate).  Such non-reproducible part of programs may be annotated by
11318           "no_instrument_function" function attribute. "gcov-dump" with -l
11319           can be used to dump gathered data and verify that they are indeed
11320           reproducible.
11321
11322           With -fprofile-reproducibility=parallel-runs collected profile
11323           stays reproducible regardless the order of streaming of the data
11324           into gcda files.  This setting makes it possible to run multiple
11325           instances of instrumented program in parallel (such as with "make
11326           -j"). This reduces quality of gathered data, in particular of
11327           indirect call profiling.
11328
11329       -fsanitize=address
11330           Enable AddressSanitizer, a fast memory error detector.  Memory
11331           access instructions are instrumented to detect out-of-bounds and
11332           use-after-free bugs.  The option enables
11333           -fsanitize-address-use-after-scope.  See
11334           <https://github.com/google/sanitizers/wiki/AddressSanitizer> for
11335           more details.  The run-time behavior can be influenced using the
11336           ASAN_OPTIONS environment variable.  When set to "help=1", the
11337           available options are shown at startup of the instrumented program.
11338           See
11339           <https://github.com/google/sanitizers/wiki/AddressSanitizerFlags#run-time-flags>
11340           for a list of supported options.  The option cannot be combined
11341           with -fsanitize=thread.
11342
11343       -fsanitize=kernel-address
11344           Enable AddressSanitizer for Linux kernel.  See
11345           <https://github.com/google/kasan/wiki> for more details.
11346
11347       -fsanitize=pointer-compare
11348           Instrument comparison operation (<, <=, >, >=) with pointer
11349           operands.  The option must be combined with either
11350           -fsanitize=kernel-address or -fsanitize=address The option cannot
11351           be combined with -fsanitize=thread.  Note: By default the check is
11352           disabled at run time.  To enable it, add
11353           "detect_invalid_pointer_pairs=2" to the environment variable
11354           ASAN_OPTIONS. Using "detect_invalid_pointer_pairs=1" detects
11355           invalid operation only when both pointers are non-null.
11356
11357       -fsanitize=pointer-subtract
11358           Instrument subtraction with pointer operands.  The option must be
11359           combined with either -fsanitize=kernel-address or
11360           -fsanitize=address The option cannot be combined with
11361           -fsanitize=thread.  Note: By default the check is disabled at run
11362           time.  To enable it, add "detect_invalid_pointer_pairs=2" to the
11363           environment variable ASAN_OPTIONS. Using
11364           "detect_invalid_pointer_pairs=1" detects invalid operation only
11365           when both pointers are non-null.
11366
11367       -fsanitize=thread
11368           Enable ThreadSanitizer, a fast data race detector.  Memory access
11369           instructions are instrumented to detect data race bugs.  See
11370           <https://github.com/google/sanitizers/wiki#threadsanitizer> for
11371           more details. The run-time behavior can be influenced using the
11372           TSAN_OPTIONS environment variable; see
11373           <https://github.com/google/sanitizers/wiki/ThreadSanitizerFlags>
11374           for a list of supported options.  The option cannot be combined
11375           with -fsanitize=address, -fsanitize=leak.
11376
11377           Note that sanitized atomic builtins cannot throw exceptions when
11378           operating on invalid memory addresses with non-call exceptions
11379           (-fnon-call-exceptions).
11380
11381       -fsanitize=leak
11382           Enable LeakSanitizer, a memory leak detector.  This option only
11383           matters for linking of executables and the executable is linked
11384           against a library that overrides "malloc" and other allocator
11385           functions.  See
11386           <https://github.com/google/sanitizers/wiki/AddressSanitizerLeakSanitizer>
11387           for more details.  The run-time behavior can be influenced using
11388           the LSAN_OPTIONS environment variable.  The option cannot be
11389           combined with -fsanitize=thread.
11390
11391       -fsanitize=undefined
11392           Enable UndefinedBehaviorSanitizer, a fast undefined behavior
11393           detector.  Various computations are instrumented to detect
11394           undefined behavior at runtime.  Current suboptions are:
11395
11396           -fsanitize=shift
11397               This option enables checking that the result of a shift
11398               operation is not undefined.  Note that what exactly is
11399               considered undefined differs slightly between C and C++, as
11400               well as between ISO C90 and C99, etc.  This option has two
11401               suboptions, -fsanitize=shift-base and
11402               -fsanitize=shift-exponent.
11403
11404           -fsanitize=shift-exponent
11405               This option enables checking that the second argument of a
11406               shift operation is not negative and is smaller than the
11407               precision of the promoted first argument.
11408
11409           -fsanitize=shift-base
11410               If the second argument of a shift operation is within range,
11411               check that the result of a shift operation is not undefined.
11412               Note that what exactly is considered undefined differs slightly
11413               between C and C++, as well as between ISO C90 and C99, etc.
11414
11415           -fsanitize=integer-divide-by-zero
11416               Detect integer division by zero as well as "INT_MIN / -1"
11417               division.
11418
11419           -fsanitize=unreachable
11420               With this option, the compiler turns the
11421               "__builtin_unreachable" call into a diagnostics message call
11422               instead.  When reaching the "__builtin_unreachable" call, the
11423               behavior is undefined.
11424
11425           -fsanitize=vla-bound
11426               This option instructs the compiler to check that the size of a
11427               variable length array is positive.
11428
11429           -fsanitize=null
11430               This option enables pointer checking.  Particularly, the
11431               application built with this option turned on will issue an
11432               error message when it tries to dereference a NULL pointer, or
11433               if a reference (possibly an rvalue reference) is bound to a
11434               NULL pointer, or if a method is invoked on an object pointed by
11435               a NULL pointer.
11436
11437           -fsanitize=return
11438               This option enables return statement checking.  Programs built
11439               with this option turned on will issue an error message when the
11440               end of a non-void function is reached without actually
11441               returning a value.  This option works in C++ only.
11442
11443           -fsanitize=signed-integer-overflow
11444               This option enables signed integer overflow checking.  We check
11445               that the result of "+", "*", and both unary and binary "-" does
11446               not overflow in the signed arithmetics.  Note, integer
11447               promotion rules must be taken into account.  That is, the
11448               following is not an overflow:
11449
11450                       signed char a = SCHAR_MAX;
11451                       a++;
11452
11453           -fsanitize=bounds
11454               This option enables instrumentation of array bounds.  Various
11455               out of bounds accesses are detected.  Flexible array members,
11456               flexible array member-like arrays, and initializers of
11457               variables with static storage are not instrumented.
11458
11459           -fsanitize=bounds-strict
11460               This option enables strict instrumentation of array bounds.
11461               Most out of bounds accesses are detected, including flexible
11462               array members and flexible array member-like arrays.
11463               Initializers of variables with static storage are not
11464               instrumented.
11465
11466           -fsanitize=alignment
11467               This option enables checking of alignment of pointers when they
11468               are dereferenced, or when a reference is bound to
11469               insufficiently aligned target, or when a method or constructor
11470               is invoked on insufficiently aligned object.
11471
11472           -fsanitize=object-size
11473               This option enables instrumentation of memory references using
11474               the "__builtin_object_size" function.  Various out of bounds
11475               pointer accesses are detected.
11476
11477           -fsanitize=float-divide-by-zero
11478               Detect floating-point division by zero.  Unlike other similar
11479               options, -fsanitize=float-divide-by-zero is not enabled by
11480               -fsanitize=undefined, since floating-point division by zero can
11481               be a legitimate way of obtaining infinities and NaNs.
11482
11483           -fsanitize=float-cast-overflow
11484               This option enables floating-point type to integer conversion
11485               checking.  We check that the result of the conversion does not
11486               overflow.  Unlike other similar options,
11487               -fsanitize=float-cast-overflow is not enabled by
11488               -fsanitize=undefined.  This option does not work well with
11489               "FE_INVALID" exceptions enabled.
11490
11491           -fsanitize=nonnull-attribute
11492               This option enables instrumentation of calls, checking whether
11493               null values are not passed to arguments marked as requiring a
11494               non-null value by the "nonnull" function attribute.
11495
11496           -fsanitize=returns-nonnull-attribute
11497               This option enables instrumentation of return statements in
11498               functions marked with "returns_nonnull" function attribute, to
11499               detect returning of null values from such functions.
11500
11501           -fsanitize=bool
11502               This option enables instrumentation of loads from bool.  If a
11503               value other than 0/1 is loaded, a run-time error is issued.
11504
11505           -fsanitize=enum
11506               This option enables instrumentation of loads from an enum type.
11507               If a value outside the range of values for the enum type is
11508               loaded, a run-time error is issued.
11509
11510           -fsanitize=vptr
11511               This option enables instrumentation of C++ member function
11512               calls, member accesses and some conversions between pointers to
11513               base and derived classes, to verify the referenced object has
11514               the correct dynamic type.
11515
11516           -fsanitize=pointer-overflow
11517               This option enables instrumentation of pointer arithmetics.  If
11518               the pointer arithmetics overflows, a run-time error is issued.
11519
11520           -fsanitize=builtin
11521               This option enables instrumentation of arguments to selected
11522               builtin functions.  If an invalid value is passed to such
11523               arguments, a run-time error is issued.  E.g. passing 0 as the
11524               argument to "__builtin_ctz" or "__builtin_clz" invokes
11525               undefined behavior and is diagnosed by this option.
11526
11527           While -ftrapv causes traps for signed overflows to be emitted,
11528           -fsanitize=undefined gives a diagnostic message.  This currently
11529           works only for the C family of languages.
11530
11531       -fno-sanitize=all
11532           This option disables all previously enabled sanitizers.
11533           -fsanitize=all is not allowed, as some sanitizers cannot be used
11534           together.
11535
11536       -fasan-shadow-offset=number
11537           This option forces GCC to use custom shadow offset in
11538           AddressSanitizer checks.  It is useful for experimenting with
11539           different shadow memory layouts in Kernel AddressSanitizer.
11540
11541       -fsanitize-sections=s1,s2,...
11542           Sanitize global variables in selected user-defined sections.  si
11543           may contain wildcards.
11544
11545       -fsanitize-recover[=opts]
11546           -fsanitize-recover= controls error recovery mode for sanitizers
11547           mentioned in comma-separated list of opts.  Enabling this option
11548           for a sanitizer component causes it to attempt to continue running
11549           the program as if no error happened.  This means multiple runtime
11550           errors can be reported in a single program run, and the exit code
11551           of the program may indicate success even when errors have been
11552           reported.  The -fno-sanitize-recover= option can be used to alter
11553           this behavior: only the first detected error is reported and
11554           program then exits with a non-zero exit code.
11555
11556           Currently this feature only works for -fsanitize=undefined (and its
11557           suboptions except for -fsanitize=unreachable and
11558           -fsanitize=return), -fsanitize=float-cast-overflow,
11559           -fsanitize=float-divide-by-zero, -fsanitize=bounds-strict,
11560           -fsanitize=kernel-address and -fsanitize=address.  For these
11561           sanitizers error recovery is turned on by default, except
11562           -fsanitize=address, for which this feature is experimental.
11563           -fsanitize-recover=all and -fno-sanitize-recover=all is also
11564           accepted, the former enables recovery for all sanitizers that
11565           support it, the latter disables recovery for all sanitizers that
11566           support it.
11567
11568           Even if a recovery mode is turned on the compiler side, it needs to
11569           be also enabled on the runtime library side, otherwise the failures
11570           are still fatal.  The runtime library defaults to "halt_on_error=0"
11571           for ThreadSanitizer and UndefinedBehaviorSanitizer, while default
11572           value for AddressSanitizer is "halt_on_error=1". This can be
11573           overridden through setting the "halt_on_error" flag in the
11574           corresponding environment variable.
11575
11576           Syntax without an explicit opts parameter is deprecated.  It is
11577           equivalent to specifying an opts list of:
11578
11579                   undefined,float-cast-overflow,float-divide-by-zero,bounds-strict
11580
11581       -fsanitize-address-use-after-scope
11582           Enable sanitization of local variables to detect use-after-scope
11583           bugs.  The option sets -fstack-reuse to none.
11584
11585       -fsanitize-undefined-trap-on-error
11586           The -fsanitize-undefined-trap-on-error option instructs the
11587           compiler to report undefined behavior using "__builtin_trap" rather
11588           than a "libubsan" library routine.  The advantage of this is that
11589           the "libubsan" library is not needed and is not linked in, so this
11590           is usable even in freestanding environments.
11591
11592       -fsanitize-coverage=trace-pc
11593           Enable coverage-guided fuzzing code instrumentation.  Inserts a
11594           call to "__sanitizer_cov_trace_pc" into every basic block.
11595
11596       -fsanitize-coverage=trace-cmp
11597           Enable dataflow guided fuzzing code instrumentation.  Inserts a
11598           call to "__sanitizer_cov_trace_cmp1", "__sanitizer_cov_trace_cmp2",
11599           "__sanitizer_cov_trace_cmp4" or "__sanitizer_cov_trace_cmp8" for
11600           integral comparison with both operands variable or
11601           "__sanitizer_cov_trace_const_cmp1",
11602           "__sanitizer_cov_trace_const_cmp2",
11603           "__sanitizer_cov_trace_const_cmp4" or
11604           "__sanitizer_cov_trace_const_cmp8" for integral comparison with one
11605           operand constant, "__sanitizer_cov_trace_cmpf" or
11606           "__sanitizer_cov_trace_cmpd" for float or double comparisons and
11607           "__sanitizer_cov_trace_switch" for switch statements.
11608
11609       -fcf-protection=[full|branch|return|none]
11610           Enable code instrumentation of control-flow transfers to increase
11611           program security by checking that target addresses of control-flow
11612           transfer instructions (such as indirect function call, function
11613           return, indirect jump) are valid.  This prevents diverting the flow
11614           of control to an unexpected target.  This is intended to protect
11615           against such threats as Return-oriented Programming (ROP), and
11616           similarly call/jmp-oriented programming (COP/JOP).
11617
11618           The value "branch" tells the compiler to implement checking of
11619           validity of control-flow transfer at the point of indirect branch
11620           instructions, i.e. call/jmp instructions.  The value "return"
11621           implements checking of validity at the point of returning from a
11622           function.  The value "full" is an alias for specifying both
11623           "branch" and "return". The value "none" turns off instrumentation.
11624
11625           The macro "__CET__" is defined when -fcf-protection is used.  The
11626           first bit of "__CET__" is set to 1 for the value "branch" and the
11627           second bit of "__CET__" is set to 1 for the "return".
11628
11629           You can also use the "nocf_check" attribute to identify which
11630           functions and calls should be skipped from instrumentation.
11631
11632           Currently the x86 GNU/Linux target provides an implementation based
11633           on Intel Control-flow Enforcement Technology (CET).
11634
11635       -fstack-protector
11636           Emit extra code to check for buffer overflows, such as stack
11637           smashing attacks.  This is done by adding a guard variable to
11638           functions with vulnerable objects.  This includes functions that
11639           call "alloca", and functions with buffers larger than or equal to 8
11640           bytes.  The guards are initialized when a function is entered and
11641           then checked when the function exits.  If a guard check fails, an
11642           error message is printed and the program exits.  Only variables
11643           that are actually allocated on the stack are considered, optimized
11644           away variables or variables allocated in registers don't count.
11645
11646       -fstack-protector-all
11647           Like -fstack-protector except that all functions are protected.
11648
11649       -fstack-protector-strong
11650           Like -fstack-protector but includes additional functions to be
11651           protected --- those that have local array definitions, or have
11652           references to local frame addresses.  Only variables that are
11653           actually allocated on the stack are considered, optimized away
11654           variables or variables allocated in registers don't count.
11655
11656       -fstack-protector-explicit
11657           Like -fstack-protector but only protects those functions which have
11658           the "stack_protect" attribute.
11659
11660       -fstack-check
11661           Generate code to verify that you do not go beyond the boundary of
11662           the stack.  You should specify this flag if you are running in an
11663           environment with multiple threads, but you only rarely need to
11664           specify it in a single-threaded environment since stack overflow is
11665           automatically detected on nearly all systems if there is only one
11666           stack.
11667
11668           Note that this switch does not actually cause checking to be done;
11669           the operating system or the language runtime must do that.  The
11670           switch causes generation of code to ensure that they see the stack
11671           being extended.
11672
11673           You can additionally specify a string parameter: no means no
11674           checking, generic means force the use of old-style checking,
11675           specific means use the best checking method and is equivalent to
11676           bare -fstack-check.
11677
11678           Old-style checking is a generic mechanism that requires no specific
11679           target support in the compiler but comes with the following
11680           drawbacks:
11681
11682           1.  Modified allocation strategy for large objects: they are always
11683               allocated dynamically if their size exceeds a fixed threshold.
11684               Note this may change the semantics of some code.
11685
11686           2.  Fixed limit on the size of the static frame of functions: when
11687               it is topped by a particular function, stack checking is not
11688               reliable and a warning is issued by the compiler.
11689
11690           3.  Inefficiency: because of both the modified allocation strategy
11691               and the generic implementation, code performance is hampered.
11692
11693           Note that old-style stack checking is also the fallback method for
11694           specific if no target support has been added in the compiler.
11695
11696           -fstack-check= is designed for Ada's needs to detect infinite
11697           recursion and stack overflows.  specific is an excellent choice
11698           when compiling Ada code.  It is not generally sufficient to protect
11699           against stack-clash attacks.  To protect against those you want
11700           -fstack-clash-protection.
11701
11702       -fstack-clash-protection
11703           Generate code to prevent stack clash style attacks.  When this
11704           option is enabled, the compiler will only allocate one page of
11705           stack space at a time and each page is accessed immediately after
11706           allocation.  Thus, it prevents allocations from jumping over any
11707           stack guard page provided by the operating system.
11708
11709           Most targets do not fully support stack clash protection.  However,
11710           on those targets -fstack-clash-protection will protect dynamic
11711           stack allocations.  -fstack-clash-protection may also provide
11712           limited protection for static stack allocations if the target
11713           supports -fstack-check=specific.
11714
11715       -fstack-limit-register=reg
11716       -fstack-limit-symbol=sym
11717       -fno-stack-limit
11718           Generate code to ensure that the stack does not grow beyond a
11719           certain value, either the value of a register or the address of a
11720           symbol.  If a larger stack is required, a signal is raised at run
11721           time.  For most targets, the signal is raised before the stack
11722           overruns the boundary, so it is possible to catch the signal
11723           without taking special precautions.
11724
11725           For instance, if the stack starts at absolute address 0x80000000
11726           and grows downwards, you can use the flags
11727           -fstack-limit-symbol=__stack_limit and
11728           -Wl,--defsym,__stack_limit=0x7ffe0000 to enforce a stack limit of
11729           128KB.  Note that this may only work with the GNU linker.
11730
11731           You can locally override stack limit checking by using the
11732           "no_stack_limit" function attribute.
11733
11734       -fsplit-stack
11735           Generate code to automatically split the stack before it overflows.
11736           The resulting program has a discontiguous stack which can only
11737           overflow if the program is unable to allocate any more memory.
11738           This is most useful when running threaded programs, as it is no
11739           longer necessary to calculate a good stack size to use for each
11740           thread.  This is currently only implemented for the x86 targets
11741           running GNU/Linux.
11742
11743           When code compiled with -fsplit-stack calls code compiled without
11744           -fsplit-stack, there may not be much stack space available for the
11745           latter code to run.  If compiling all code, including library code,
11746           with -fsplit-stack is not an option, then the linker can fix up
11747           these calls so that the code compiled without -fsplit-stack always
11748           has a large stack.  Support for this is implemented in the gold
11749           linker in GNU binutils release 2.21 and later.
11750
11751       -fvtable-verify=[std|preinit|none]
11752           This option is only available when compiling C++ code.  It turns on
11753           (or off, if using -fvtable-verify=none) the security feature that
11754           verifies at run time, for every virtual call, that the vtable
11755           pointer through which the call is made is valid for the type of the
11756           object, and has not been corrupted or overwritten.  If an invalid
11757           vtable pointer is detected at run time, an error is reported and
11758           execution of the program is immediately halted.
11759
11760           This option causes run-time data structures to be built at program
11761           startup, which are used for verifying the vtable pointers.  The
11762           options std and preinit control the timing of when these data
11763           structures are built.  In both cases the data structures are built
11764           before execution reaches "main".  Using -fvtable-verify=std causes
11765           the data structures to be built after shared libraries have been
11766           loaded and initialized.  -fvtable-verify=preinit causes them to be
11767           built before shared libraries have been loaded and initialized.
11768
11769           If this option appears multiple times in the command line with
11770           different values specified, none takes highest priority over both
11771           std and preinit; preinit takes priority over std.
11772
11773       -fvtv-debug
11774           When used in conjunction with -fvtable-verify=std or
11775           -fvtable-verify=preinit, causes debug versions of the runtime
11776           functions for the vtable verification feature to be called.  This
11777           flag also causes the compiler to log information about which vtable
11778           pointers it finds for each class.  This information is written to a
11779           file named vtv_set_ptr_data.log in the directory named by the
11780           environment variable VTV_LOGS_DIR if that is defined or the current
11781           working directory otherwise.
11782
11783           Note:  This feature appends data to the log file. If you want a
11784           fresh log file, be sure to delete any existing one.
11785
11786       -fvtv-counts
11787           This is a debugging flag.  When used in conjunction with
11788           -fvtable-verify=std or -fvtable-verify=preinit, this causes the
11789           compiler to keep track of the total number of virtual calls it
11790           encounters and the number of verifications it inserts.  It also
11791           counts the number of calls to certain run-time library functions
11792           that it inserts and logs this information for each compilation
11793           unit.  The compiler writes this information to a file named
11794           vtv_count_data.log in the directory named by the environment
11795           variable VTV_LOGS_DIR if that is defined or the current working
11796           directory otherwise.  It also counts the size of the vtable pointer
11797           sets for each class, and writes this information to
11798           vtv_class_set_sizes.log in the same directory.
11799
11800           Note:  This feature appends data to the log files.  To get fresh
11801           log files, be sure to delete any existing ones.
11802
11803       -finstrument-functions
11804           Generate instrumentation calls for entry and exit to functions.
11805           Just after function entry and just before function exit, the
11806           following profiling functions are called with the address of the
11807           current function and its call site.  (On some platforms,
11808           "__builtin_return_address" does not work beyond the current
11809           function, so the call site information may not be available to the
11810           profiling functions otherwise.)
11811
11812                   void __cyg_profile_func_enter (void *this_fn,
11813                                                  void *call_site);
11814                   void __cyg_profile_func_exit  (void *this_fn,
11815                                                  void *call_site);
11816
11817           The first argument is the address of the start of the current
11818           function, which may be looked up exactly in the symbol table.
11819
11820           This instrumentation is also done for functions expanded inline in
11821           other functions.  The profiling calls indicate where, conceptually,
11822           the inline function is entered and exited.  This means that
11823           addressable versions of such functions must be available.  If all
11824           your uses of a function are expanded inline, this may mean an
11825           additional expansion of code size.  If you use "extern inline" in
11826           your C code, an addressable version of such functions must be
11827           provided.  (This is normally the case anyway, but if you get lucky
11828           and the optimizer always expands the functions inline, you might
11829           have gotten away without providing static copies.)
11830
11831           A function may be given the attribute "no_instrument_function", in
11832           which case this instrumentation is not done.  This can be used, for
11833           example, for the profiling functions listed above, high-priority
11834           interrupt routines, and any functions from which the profiling
11835           functions cannot safely be called (perhaps signal handlers, if the
11836           profiling routines generate output or allocate memory).
11837
11838       -finstrument-functions-exclude-file-list=file,file,...
11839           Set the list of functions that are excluded from instrumentation
11840           (see the description of -finstrument-functions).  If the file that
11841           contains a function definition matches with one of file, then that
11842           function is not instrumented.  The match is done on substrings: if
11843           the file parameter is a substring of the file name, it is
11844           considered to be a match.
11845
11846           For example:
11847
11848                   -finstrument-functions-exclude-file-list=/bits/stl,include/sys
11849
11850           excludes any inline function defined in files whose pathnames
11851           contain /bits/stl or include/sys.
11852
11853           If, for some reason, you want to include letter , in one of sym,
11854           write ,. For example,
11855           -finstrument-functions-exclude-file-list=',,tmp' (note the single
11856           quote surrounding the option).
11857
11858       -finstrument-functions-exclude-function-list=sym,sym,...
11859           This is similar to -finstrument-functions-exclude-file-list, but
11860           this option sets the list of function names to be excluded from
11861           instrumentation.  The function name to be matched is its user-
11862           visible name, such as "vector<int> blah(const vector<int> &)", not
11863           the internal mangled name (e.g., "_Z4blahRSt6vectorIiSaIiEE").  The
11864           match is done on substrings: if the sym parameter is a substring of
11865           the function name, it is considered to be a match.  For C99 and C++
11866           extended identifiers, the function name must be given in UTF-8, not
11867           using universal character names.
11868
11869       -fpatchable-function-entry=N[,M]
11870           Generate N NOPs right at the beginning of each function, with the
11871           function entry point before the Mth NOP.  If M is omitted, it
11872           defaults to 0 so the function entry points to the address just at
11873           the first NOP.  The NOP instructions reserve extra space which can
11874           be used to patch in any desired instrumentation at run time,
11875           provided that the code segment is writable.  The amount of space is
11876           controllable indirectly via the number of NOPs; the NOP instruction
11877           used corresponds to the instruction emitted by the internal GCC
11878           back-end interface "gen_nop".  This behavior is target-specific and
11879           may also depend on the architecture variant and/or other
11880           compilation options.
11881
11882           For run-time identification, the starting addresses of these areas,
11883           which correspond to their respective function entries minus M, are
11884           additionally collected in the "__patchable_function_entries"
11885           section of the resulting binary.
11886
11887           Note that the value of "__attribute__ ((patchable_function_entry
11888           (N,M)))" takes precedence over command-line option
11889           -fpatchable-function-entry=N,M.  This can be used to increase the
11890           area size or to remove it completely on a single function.  If
11891           "N=0", no pad location is recorded.
11892
11893           The NOP instructions are inserted at---and maybe before, depending
11894           on M---the function entry address, even before the prologue.
11895
11896   Options Controlling the Preprocessor
11897       These options control the C preprocessor, which is run on each C source
11898       file before actual compilation.
11899
11900       If you use the -E option, nothing is done except preprocessing.  Some
11901       of these options make sense only together with -E because they cause
11902       the preprocessor output to be unsuitable for actual compilation.
11903
11904       In addition to the options listed here, there are a number of options
11905       to control search paths for include files documented in Directory
11906       Options.  Options to control preprocessor diagnostics are listed in
11907       Warning Options.
11908
11909       -D name
11910           Predefine name as a macro, with definition 1.
11911
11912       -D name=definition
11913           The contents of definition are tokenized and processed as if they
11914           appeared during translation phase three in a #define directive.  In
11915           particular, the definition is truncated by embedded newline
11916           characters.
11917
11918           If you are invoking the preprocessor from a shell or shell-like
11919           program you may need to use the shell's quoting syntax to protect
11920           characters such as spaces that have a meaning in the shell syntax.
11921
11922           If you wish to define a function-like macro on the command line,
11923           write its argument list with surrounding parentheses before the
11924           equals sign (if any).  Parentheses are meaningful to most shells,
11925           so you should quote the option.  With sh and csh,
11926           -D'name(args...)=definition' works.
11927
11928           -D and -U options are processed in the order they are given on the
11929           command line.  All -imacros file and -include file options are
11930           processed after all -D and -U options.
11931
11932       -U name
11933           Cancel any previous definition of name, either built in or provided
11934           with a -D option.
11935
11936       -include file
11937           Process file as if "#include "file"" appeared as the first line of
11938           the primary source file.  However, the first directory searched for
11939           file is the preprocessor's working directory instead of the
11940           directory containing the main source file.  If not found there, it
11941           is searched for in the remainder of the "#include "..."" search
11942           chain as normal.
11943
11944           If multiple -include options are given, the files are included in
11945           the order they appear on the command line.
11946
11947       -imacros file
11948           Exactly like -include, except that any output produced by scanning
11949           file is thrown away.  Macros it defines remain defined.  This
11950           allows you to acquire all the macros from a header without also
11951           processing its declarations.
11952
11953           All files specified by -imacros are processed before all files
11954           specified by -include.
11955
11956       -undef
11957           Do not predefine any system-specific or GCC-specific macros.  The
11958           standard predefined macros remain defined.
11959
11960       -pthread
11961           Define additional macros required for using the POSIX threads
11962           library.  You should use this option consistently for both
11963           compilation and linking.  This option is supported on GNU/Linux
11964           targets, most other Unix derivatives, and also on x86 Cygwin and
11965           MinGW targets.
11966
11967       -M  Instead of outputting the result of preprocessing, output a rule
11968           suitable for make describing the dependencies of the main source
11969           file.  The preprocessor outputs one make rule containing the object
11970           file name for that source file, a colon, and the names of all the
11971           included files, including those coming from -include or -imacros
11972           command-line options.
11973
11974           Unless specified explicitly (with -MT or -MQ), the object file name
11975           consists of the name of the source file with any suffix replaced
11976           with object file suffix and with any leading directory parts
11977           removed.  If there are many included files then the rule is split
11978           into several lines using \-newline.  The rule has no commands.
11979
11980           This option does not suppress the preprocessor's debug output, such
11981           as -dM.  To avoid mixing such debug output with the dependency
11982           rules you should explicitly specify the dependency output file with
11983           -MF, or use an environment variable like DEPENDENCIES_OUTPUT.
11984           Debug output is still sent to the regular output stream as normal.
11985
11986           Passing -M to the driver implies -E, and suppresses warnings with
11987           an implicit -w.
11988
11989       -MM Like -M but do not mention header files that are found in system
11990           header directories, nor header files that are included, directly or
11991           indirectly, from such a header.
11992
11993           This implies that the choice of angle brackets or double quotes in
11994           an #include directive does not in itself determine whether that
11995           header appears in -MM dependency output.
11996
11997       -MF file
11998           When used with -M or -MM, specifies a file to write the
11999           dependencies to.  If no -MF switch is given the preprocessor sends
12000           the rules to the same place it would send preprocessed output.
12001
12002           When used with the driver options -MD or -MMD, -MF overrides the
12003           default dependency output file.
12004
12005           If file is -, then the dependencies are written to stdout.
12006
12007       -MG In conjunction with an option such as -M requesting dependency
12008           generation, -MG assumes missing header files are generated files
12009           and adds them to the dependency list without raising an error.  The
12010           dependency filename is taken directly from the "#include" directive
12011           without prepending any path.  -MG also suppresses preprocessed
12012           output, as a missing header file renders this useless.
12013
12014           This feature is used in automatic updating of makefiles.
12015
12016       -MP This option instructs CPP to add a phony target for each dependency
12017           other than the main file, causing each to depend on nothing.  These
12018           dummy rules work around errors make gives if you remove header
12019           files without updating the Makefile to match.
12020
12021           This is typical output:
12022
12023                   test.o: test.c test.h
12024
12025                   test.h:
12026
12027       -MT target
12028           Change the target of the rule emitted by dependency generation.  By
12029           default CPP takes the name of the main input file, deletes any
12030           directory components and any file suffix such as .c, and appends
12031           the platform's usual object suffix.  The result is the target.
12032
12033           An -MT option sets the target to be exactly the string you specify.
12034           If you want multiple targets, you can specify them as a single
12035           argument to -MT, or use multiple -MT options.
12036
12037           For example, -MT '$(objpfx)foo.o' might give
12038
12039                   $(objpfx)foo.o: foo.c
12040
12041       -MQ target
12042           Same as -MT, but it quotes any characters which are special to
12043           Make.  -MQ '$(objpfx)foo.o' gives
12044
12045                   $$(objpfx)foo.o: foo.c
12046
12047           The default target is automatically quoted, as if it were given
12048           with -MQ.
12049
12050       -MD -MD is equivalent to -M -MF file, except that -E is not implied.
12051           The driver determines file based on whether an -o option is given.
12052           If it is, the driver uses its argument but with a suffix of .d,
12053           otherwise it takes the name of the input file, removes any
12054           directory components and suffix, and applies a .d suffix.
12055
12056           If -MD is used in conjunction with -E, any -o switch is understood
12057           to specify the dependency output file, but if used without -E, each
12058           -o is understood to specify a target object file.
12059
12060           Since -E is not implied, -MD can be used to generate a dependency
12061           output file as a side effect of the compilation process.
12062
12063       -MMD
12064           Like -MD except mention only user header files, not system header
12065           files.
12066
12067       -fpreprocessed
12068           Indicate to the preprocessor that the input file has already been
12069           preprocessed.  This suppresses things like macro expansion,
12070           trigraph conversion, escaped newline splicing, and processing of
12071           most directives.  The preprocessor still recognizes and removes
12072           comments, so that you can pass a file preprocessed with -C to the
12073           compiler without problems.  In this mode the integrated
12074           preprocessor is little more than a tokenizer for the front ends.
12075
12076           -fpreprocessed is implicit if the input file has one of the
12077           extensions .i, .ii or .mi.  These are the extensions that GCC uses
12078           for preprocessed files created by -save-temps.
12079
12080       -fdirectives-only
12081           When preprocessing, handle directives, but do not expand macros.
12082
12083           The option's behavior depends on the -E and -fpreprocessed options.
12084
12085           With -E, preprocessing is limited to the handling of directives
12086           such as "#define", "#ifdef", and "#error".  Other preprocessor
12087           operations, such as macro expansion and trigraph conversion are not
12088           performed.  In addition, the -dD option is implicitly enabled.
12089
12090           With -fpreprocessed, predefinition of command line and most builtin
12091           macros is disabled.  Macros such as "__LINE__", which are
12092           contextually dependent, are handled normally.  This enables
12093           compilation of files previously preprocessed with "-E
12094           -fdirectives-only".
12095
12096           With both -E and -fpreprocessed, the rules for -fpreprocessed take
12097           precedence.  This enables full preprocessing of files previously
12098           preprocessed with "-E -fdirectives-only".
12099
12100       -fdollars-in-identifiers
12101           Accept $ in identifiers.
12102
12103       -fextended-identifiers
12104           Accept universal character names and extended characters in
12105           identifiers.  This option is enabled by default for C99 (and later
12106           C standard versions) and C++.
12107
12108       -fno-canonical-system-headers
12109           When preprocessing, do not shorten system header paths with
12110           canonicalization.
12111
12112       -fmax-include-depth=depth
12113           Set the maximum depth of the nested #include. The default is 200.
12114
12115       -ftabstop=width
12116           Set the distance between tab stops.  This helps the preprocessor
12117           report correct column numbers in warnings or errors, even if tabs
12118           appear on the line.  If the value is less than 1 or greater than
12119           100, the option is ignored.  The default is 8.
12120
12121       -ftrack-macro-expansion[=level]
12122           Track locations of tokens across macro expansions. This allows the
12123           compiler to emit diagnostic about the current macro expansion stack
12124           when a compilation error occurs in a macro expansion. Using this
12125           option makes the preprocessor and the compiler consume more memory.
12126           The level parameter can be used to choose the level of precision of
12127           token location tracking thus decreasing the memory consumption if
12128           necessary. Value 0 of level de-activates this option. Value 1
12129           tracks tokens locations in a degraded mode for the sake of minimal
12130           memory overhead. In this mode all tokens resulting from the
12131           expansion of an argument of a function-like macro have the same
12132           location. Value 2 tracks tokens locations completely. This value is
12133           the most memory hungry.  When this option is given no argument, the
12134           default parameter value is 2.
12135
12136           Note that "-ftrack-macro-expansion=2" is activated by default.
12137
12138       -fmacro-prefix-map=old=new
12139           When preprocessing files residing in directory old, expand the
12140           "__FILE__" and "__BASE_FILE__" macros as if the files resided in
12141           directory new instead.  This can be used to change an absolute path
12142           to a relative path by using . for new which can result in more
12143           reproducible builds that are location independent.  This option
12144           also affects "__builtin_FILE()" during compilation.  See also
12145           -ffile-prefix-map.
12146
12147       -fexec-charset=charset
12148           Set the execution character set, used for string and character
12149           constants.  The default is UTF-8.  charset can be any encoding
12150           supported by the system's "iconv" library routine.
12151
12152       -fwide-exec-charset=charset
12153           Set the wide execution character set, used for wide string and
12154           character constants.  The default is UTF-32 or UTF-16, whichever
12155           corresponds to the width of "wchar_t".  As with -fexec-charset,
12156           charset can be any encoding supported by the system's "iconv"
12157           library routine; however, you will have problems with encodings
12158           that do not fit exactly in "wchar_t".
12159
12160       -finput-charset=charset
12161           Set the input character set, used for translation from the
12162           character set of the input file to the source character set used by
12163           GCC.  If the locale does not specify, or GCC cannot get this
12164           information from the locale, the default is UTF-8.  This can be
12165           overridden by either the locale or this command-line option.
12166           Currently the command-line option takes precedence if there's a
12167           conflict.  charset can be any encoding supported by the system's
12168           "iconv" library routine.
12169
12170       -fpch-deps
12171           When using precompiled headers, this flag causes the dependency-
12172           output flags to also list the files from the precompiled header's
12173           dependencies.  If not specified, only the precompiled header are
12174           listed and not the files that were used to create it, because those
12175           files are not consulted when a precompiled header is used.
12176
12177       -fpch-preprocess
12178           This option allows use of a precompiled header together with -E.
12179           It inserts a special "#pragma", "#pragma GCC pch_preprocess
12180           "filename"" in the output to mark the place where the precompiled
12181           header was found, and its filename.  When -fpreprocessed is in use,
12182           GCC recognizes this "#pragma" and loads the PCH.
12183
12184           This option is off by default, because the resulting preprocessed
12185           output is only really suitable as input to GCC.  It is switched on
12186           by -save-temps.
12187
12188           You should not write this "#pragma" in your own code, but it is
12189           safe to edit the filename if the PCH file is available in a
12190           different location.  The filename may be absolute or it may be
12191           relative to GCC's current directory.
12192
12193       -fworking-directory
12194           Enable generation of linemarkers in the preprocessor output that
12195           let the compiler know the current working directory at the time of
12196           preprocessing.  When this option is enabled, the preprocessor
12197           emits, after the initial linemarker, a second linemarker with the
12198           current working directory followed by two slashes.  GCC uses this
12199           directory, when it's present in the preprocessed input, as the
12200           directory emitted as the current working directory in some
12201           debugging information formats.  This option is implicitly enabled
12202           if debugging information is enabled, but this can be inhibited with
12203           the negated form -fno-working-directory.  If the -P flag is present
12204           in the command line, this option has no effect, since no "#line"
12205           directives are emitted whatsoever.
12206
12207       -A predicate=answer
12208           Make an assertion with the predicate predicate and answer answer.
12209           This form is preferred to the older form -A predicate(answer),
12210           which is still supported, because it does not use shell special
12211           characters.
12212
12213       -A -predicate=answer
12214           Cancel an assertion with the predicate predicate and answer answer.
12215
12216       -C  Do not discard comments.  All comments are passed through to the
12217           output file, except for comments in processed directives, which are
12218           deleted along with the directive.
12219
12220           You should be prepared for side effects when using -C; it causes
12221           the preprocessor to treat comments as tokens in their own right.
12222           For example, comments appearing at the start of what would be a
12223           directive line have the effect of turning that line into an
12224           ordinary source line, since the first token on the line is no
12225           longer a #.
12226
12227       -CC Do not discard comments, including during macro expansion.  This is
12228           like -C, except that comments contained within macros are also
12229           passed through to the output file where the macro is expanded.
12230
12231           In addition to the side effects of the -C option, the -CC option
12232           causes all C++-style comments inside a macro to be converted to
12233           C-style comments.  This is to prevent later use of that macro from
12234           inadvertently commenting out the remainder of the source line.
12235
12236           The -CC option is generally used to support lint comments.
12237
12238       -P  Inhibit generation of linemarkers in the output from the
12239           preprocessor.  This might be useful when running the preprocessor
12240           on something that is not C code, and will be sent to a program
12241           which might be confused by the linemarkers.
12242
12243       -traditional
12244       -traditional-cpp
12245           Try to imitate the behavior of pre-standard C preprocessors, as
12246           opposed to ISO C preprocessors.  See the GNU CPP manual for
12247           details.
12248
12249           Note that GCC does not otherwise attempt to emulate a pre-standard
12250           C compiler, and these options are only supported with the -E
12251           switch, or when invoking CPP explicitly.
12252
12253       -trigraphs
12254           Support ISO C trigraphs.  These are three-character sequences, all
12255           starting with ??, that are defined by ISO C to stand for single
12256           characters.  For example, ??/ stands for \, so '??/n' is a
12257           character constant for a newline.
12258
12259           The nine trigraphs and their replacements are
12260
12261                   Trigraph:       ??(  ??)  ??<  ??>  ??=  ??/  ??'  ??!  ??-
12262                   Replacement:      [    ]    {    }    #    \    ^    |    ~
12263
12264           By default, GCC ignores trigraphs, but in standard-conforming modes
12265           it converts them.  See the -std and -ansi options.
12266
12267       -remap
12268           Enable special code to work around file systems which only permit
12269           very short file names, such as MS-DOS.
12270
12271       -H  Print the name of each header file used, in addition to other
12272           normal activities.  Each name is indented to show how deep in the
12273           #include stack it is.  Precompiled header files are also printed,
12274           even if they are found to be invalid; an invalid precompiled header
12275           file is printed with ...x and a valid one with ...! .
12276
12277       -dletters
12278           Says to make debugging dumps during compilation as specified by
12279           letters.  The flags documented here are those relevant to the
12280           preprocessor.  Other letters are interpreted by the compiler
12281           proper, or reserved for future versions of GCC, and so are silently
12282           ignored.  If you specify letters whose behavior conflicts, the
12283           result is undefined.
12284
12285           -dM Instead of the normal output, generate a list of #define
12286               directives for all the macros defined during the execution of
12287               the preprocessor, including predefined macros.  This gives you
12288               a way of finding out what is predefined in your version of the
12289               preprocessor.  Assuming you have no file foo.h, the command
12290
12291                       touch foo.h; cpp -dM foo.h
12292
12293               shows all the predefined macros.
12294
12295               If you use -dM without the -E option, -dM is interpreted as a
12296               synonym for -fdump-rtl-mach.
12297
12298           -dD Like -dM except in two respects: it does not include the
12299               predefined macros, and it outputs both the #define directives
12300               and the result of preprocessing.  Both kinds of output go to
12301               the standard output file.
12302
12303           -dN Like -dD, but emit only the macro names, not their expansions.
12304
12305           -dI Output #include directives in addition to the result of
12306               preprocessing.
12307
12308           -dU Like -dD except that only macros that are expanded, or whose
12309               definedness is tested in preprocessor directives, are output;
12310               the output is delayed until the use or test of the macro; and
12311               #undef directives are also output for macros tested but
12312               undefined at the time.
12313
12314       -fdebug-cpp
12315           This option is only useful for debugging GCC.  When used from CPP
12316           or with -E, it dumps debugging information about location maps.
12317           Every token in the output is preceded by the dump of the map its
12318           location belongs to.
12319
12320           When used from GCC without -E, this option has no effect.
12321
12322       -Wp,option
12323           You can use -Wp,option to bypass the compiler driver and pass
12324           option directly through to the preprocessor.  If option contains
12325           commas, it is split into multiple options at the commas.  However,
12326           many options are modified, translated or interpreted by the
12327           compiler driver before being passed to the preprocessor, and -Wp
12328           forcibly bypasses this phase.  The preprocessor's direct interface
12329           is undocumented and subject to change, so whenever possible you
12330           should avoid using -Wp and let the driver handle the options
12331           instead.
12332
12333       -Xpreprocessor option
12334           Pass option as an option to the preprocessor.  You can use this to
12335           supply system-specific preprocessor options that GCC does not
12336           recognize.
12337
12338           If you want to pass an option that takes an argument, you must use
12339           -Xpreprocessor twice, once for the option and once for the
12340           argument.
12341
12342       -no-integrated-cpp
12343           Perform preprocessing as a separate pass before compilation.  By
12344           default, GCC performs preprocessing as an integrated part of input
12345           tokenization and parsing.  If this option is provided, the
12346           appropriate language front end (cc1, cc1plus, or cc1obj for C, C++,
12347           and Objective-C, respectively) is instead invoked twice, once for
12348           preprocessing only and once for actual compilation of the
12349           preprocessed input.  This option may be useful in conjunction with
12350           the -B or -wrapper options to specify an alternate preprocessor or
12351           perform additional processing of the program source between normal
12352           preprocessing and compilation.
12353
12354   Passing Options to the Assembler
12355       You can pass options to the assembler.
12356
12357       -Wa,option
12358           Pass option as an option to the assembler.  If option contains
12359           commas, it is split into multiple options at the commas.
12360
12361       -Xassembler option
12362           Pass option as an option to the assembler.  You can use this to
12363           supply system-specific assembler options that GCC does not
12364           recognize.
12365
12366           If you want to pass an option that takes an argument, you must use
12367           -Xassembler twice, once for the option and once for the argument.
12368
12369   Options for Linking
12370       These options come into play when the compiler links object files into
12371       an executable output file.  They are meaningless if the compiler is not
12372       doing a link step.
12373
12374       object-file-name
12375           A file name that does not end in a special recognized suffix is
12376           considered to name an object file or library.  (Object files are
12377           distinguished from libraries by the linker according to the file
12378           contents.)  If linking is done, these object files are used as
12379           input to the linker.
12380
12381       -c
12382       -S
12383       -E  If any of these options is used, then the linker is not run, and
12384           object file names should not be used as arguments.
12385
12386       -flinker-output=type
12387           This option controls code generation of the link-time optimizer.
12388           By default the linker output is automatically determined by the
12389           linker plugin.  For debugging the compiler and if incremental
12390           linking with a non-LTO object file is desired, it may be useful to
12391           control the type manually.
12392
12393           If type is exec, code generation produces a static binary. In this
12394           case -fpic and -fpie are both disabled.
12395
12396           If type is dyn, code generation produces a shared library.  In this
12397           case -fpic or -fPIC is preserved, but not enabled automatically.
12398           This allows to build shared libraries without position-independent
12399           code on architectures where this is possible, i.e. on x86.
12400
12401           If type is pie, code generation produces an -fpie executable. This
12402           results in similar optimizations as exec except that -fpie is not
12403           disabled if specified at compilation time.
12404
12405           If type is rel, the compiler assumes that incremental linking is
12406           done.  The sections containing intermediate code for link-time
12407           optimization are merged, pre-optimized, and output to the resulting
12408           object file. In addition, if -ffat-lto-objects is specified, binary
12409           code is produced for future non-LTO linking. The object file
12410           produced by incremental linking is smaller than a static library
12411           produced from the same object files.  At link time the result of
12412           incremental linking also loads faster than a static library
12413           assuming that the majority of objects in the library are used.
12414
12415           Finally nolto-rel configures the compiler for incremental linking
12416           where code generation is forced, a final binary is produced, and
12417           the intermediate code for later link-time optimization is stripped.
12418           When multiple object files are linked together the resulting code
12419           is better optimized than with link-time optimizations disabled (for
12420           example, cross-module inlining happens), but most of benefits of
12421           whole program optimizations are lost.
12422
12423           During the incremental link (by -r) the linker plugin defaults to
12424           rel. With current interfaces to GNU Binutils it is however not
12425           possible to incrementally link LTO objects and non-LTO objects into
12426           a single mixed object file.  If any of object files in incremental
12427           link cannot be used for link-time optimization, the linker plugin
12428           issues a warning and uses nolto-rel. To maintain whole program
12429           optimization, it is recommended to link such objects into static
12430           library instead. Alternatively it is possible to use H.J. Lu's
12431           binutils with support for mixed objects.
12432
12433       -fuse-ld=bfd
12434           Use the bfd linker instead of the default linker.
12435
12436       -fuse-ld=gold
12437           Use the gold linker instead of the default linker.
12438
12439       -fuse-ld=lld
12440           Use the LLVM lld linker instead of the default linker.
12441
12442       -llibrary
12443       -l library
12444           Search the library named library when linking.  (The second
12445           alternative with the library as a separate argument is only for
12446           POSIX compliance and is not recommended.)
12447
12448           The -l option is passed directly to the linker by GCC.  Refer to
12449           your linker documentation for exact details.  The general
12450           description below applies to the GNU linker.
12451
12452           The linker searches a standard list of directories for the library.
12453           The directories searched include several standard system
12454           directories plus any that you specify with -L.
12455
12456           Static libraries are archives of object files, and have file names
12457           like liblibrary.a.  Some targets also support shared libraries,
12458           which typically have names like liblibrary.so.  If both static and
12459           shared libraries are found, the linker gives preference to linking
12460           with the shared library unless the -static option is used.
12461
12462           It makes a difference where in the command you write this option;
12463           the linker searches and processes libraries and object files in the
12464           order they are specified.  Thus, foo.o -lz bar.o searches library z
12465           after file foo.o but before bar.o.  If bar.o refers to functions in
12466           z, those functions may not be loaded.
12467
12468       -lobjc
12469           You need this special case of the -l option in order to link an
12470           Objective-C or Objective-C++ program.
12471
12472       -nostartfiles
12473           Do not use the standard system startup files when linking.  The
12474           standard system libraries are used normally, unless -nostdlib,
12475           -nolibc, or -nodefaultlibs is used.
12476
12477       -nodefaultlibs
12478           Do not use the standard system libraries when linking.  Only the
12479           libraries you specify are passed to the linker, and options
12480           specifying linkage of the system libraries, such as -static-libgcc
12481           or -shared-libgcc, are ignored.  The standard startup files are
12482           used normally, unless -nostartfiles is used.
12483
12484           The compiler may generate calls to "memcmp", "memset", "memcpy" and
12485           "memmove".  These entries are usually resolved by entries in libc.
12486           These entry points should be supplied through some other mechanism
12487           when this option is specified.
12488
12489       -nolibc
12490           Do not use the C library or system libraries tightly coupled with
12491           it when linking.  Still link with the startup files, libgcc or
12492           toolchain provided language support libraries such as libgnat,
12493           libgfortran or libstdc++ unless options preventing their inclusion
12494           are used as well.  This typically removes -lc from the link command
12495           line, as well as system libraries that normally go with it and
12496           become meaningless when absence of a C library is assumed, for
12497           example -lpthread or -lm in some configurations.  This is intended
12498           for bare-board targets when there is indeed no C library available.
12499
12500       -nostdlib
12501           Do not use the standard system startup files or libraries when
12502           linking.  No startup files and only the libraries you specify are
12503           passed to the linker, and options specifying linkage of the system
12504           libraries, such as -static-libgcc or -shared-libgcc, are ignored.
12505
12506           The compiler may generate calls to "memcmp", "memset", "memcpy" and
12507           "memmove".  These entries are usually resolved by entries in libc.
12508           These entry points should be supplied through some other mechanism
12509           when this option is specified.
12510
12511           One of the standard libraries bypassed by -nostdlib and
12512           -nodefaultlibs is libgcc.a, a library of internal subroutines which
12513           GCC uses to overcome shortcomings of particular machines, or
12514           special needs for some languages.
12515
12516           In most cases, you need libgcc.a even when you want to avoid other
12517           standard libraries.  In other words, when you specify -nostdlib or
12518           -nodefaultlibs you should usually specify -lgcc as well.  This
12519           ensures that you have no unresolved references to internal GCC
12520           library subroutines.  (An example of such an internal subroutine is
12521           "__main", used to ensure C++ constructors are called.)
12522
12523       -e entry
12524       --entry=entry
12525           Specify that the program entry point is entry.  The argument is
12526           interpreted by the linker; the GNU linker accepts either a symbol
12527           name or an address.
12528
12529       -pie
12530           Produce a dynamically linked position independent executable on
12531           targets that support it.  For predictable results, you must also
12532           specify the same set of options used for compilation (-fpie, -fPIE,
12533           or model suboptions) when you specify this linker option.
12534
12535       -no-pie
12536           Don't produce a dynamically linked position independent executable.
12537
12538       -static-pie
12539           Produce a static position independent executable on targets that
12540           support it.  A static position independent executable is similar to
12541           a static executable, but can be loaded at any address without a
12542           dynamic linker.  For predictable results, you must also specify the
12543           same set of options used for compilation (-fpie, -fPIE, or model
12544           suboptions) when you specify this linker option.
12545
12546       -pthread
12547           Link with the POSIX threads library.  This option is supported on
12548           GNU/Linux targets, most other Unix derivatives, and also on x86
12549           Cygwin and MinGW targets.  On some targets this option also sets
12550           flags for the preprocessor, so it should be used consistently for
12551           both compilation and linking.
12552
12553       -r  Produce a relocatable object as output.  This is also known as
12554           partial linking.
12555
12556       -rdynamic
12557           Pass the flag -export-dynamic to the ELF linker, on targets that
12558           support it. This instructs the linker to add all symbols, not only
12559           used ones, to the dynamic symbol table. This option is needed for
12560           some uses of "dlopen" or to allow obtaining backtraces from within
12561           a program.
12562
12563       -s  Remove all symbol table and relocation information from the
12564           executable.
12565
12566       -static
12567           On systems that support dynamic linking, this overrides -pie and
12568           prevents linking with the shared libraries.  On other systems, this
12569           option has no effect.
12570
12571       -shared
12572           Produce a shared object which can then be linked with other objects
12573           to form an executable.  Not all systems support this option.  For
12574           predictable results, you must also specify the same set of options
12575           used for compilation (-fpic, -fPIC, or model suboptions) when you
12576           specify this linker option.[1]
12577
12578       -shared-libgcc
12579       -static-libgcc
12580           On systems that provide libgcc as a shared library, these options
12581           force the use of either the shared or static version, respectively.
12582           If no shared version of libgcc was built when the compiler was
12583           configured, these options have no effect.
12584
12585           There are several situations in which an application should use the
12586           shared libgcc instead of the static version.  The most common of
12587           these is when the application wishes to throw and catch exceptions
12588           across different shared libraries.  In that case, each of the
12589           libraries as well as the application itself should use the shared
12590           libgcc.
12591
12592           Therefore, the G++ driver automatically adds -shared-libgcc
12593           whenever you build a shared library or a main executable, because
12594           C++ programs typically use exceptions, so this is the right thing
12595           to do.
12596
12597           If, instead, you use the GCC driver to create shared libraries, you
12598           may find that they are not always linked with the shared libgcc.
12599           If GCC finds, at its configuration time, that you have a non-GNU
12600           linker or a GNU linker that does not support option --eh-frame-hdr,
12601           it links the shared version of libgcc into shared libraries by
12602           default.  Otherwise, it takes advantage of the linker and optimizes
12603           away the linking with the shared version of libgcc, linking with
12604           the static version of libgcc by default.  This allows exceptions to
12605           propagate through such shared libraries, without incurring
12606           relocation costs at library load time.
12607
12608           However, if a library or main executable is supposed to throw or
12609           catch exceptions, you must link it using the G++ driver, or using
12610           the option -shared-libgcc, such that it is linked with the shared
12611           libgcc.
12612
12613       -static-libasan
12614           When the -fsanitize=address option is used to link a program, the
12615           GCC driver automatically links against libasan.  If libasan is
12616           available as a shared library, and the -static option is not used,
12617           then this links against the shared version of libasan.  The
12618           -static-libasan option directs the GCC driver to link libasan
12619           statically, without necessarily linking other libraries statically.
12620
12621       -static-libtsan
12622           When the -fsanitize=thread option is used to link a program, the
12623           GCC driver automatically links against libtsan.  If libtsan is
12624           available as a shared library, and the -static option is not used,
12625           then this links against the shared version of libtsan.  The
12626           -static-libtsan option directs the GCC driver to link libtsan
12627           statically, without necessarily linking other libraries statically.
12628
12629       -static-liblsan
12630           When the -fsanitize=leak option is used to link a program, the GCC
12631           driver automatically links against liblsan.  If liblsan is
12632           available as a shared library, and the -static option is not used,
12633           then this links against the shared version of liblsan.  The
12634           -static-liblsan option directs the GCC driver to link liblsan
12635           statically, without necessarily linking other libraries statically.
12636
12637       -static-libubsan
12638           When the -fsanitize=undefined option is used to link a program, the
12639           GCC driver automatically links against libubsan.  If libubsan is
12640           available as a shared library, and the -static option is not used,
12641           then this links against the shared version of libubsan.  The
12642           -static-libubsan option directs the GCC driver to link libubsan
12643           statically, without necessarily linking other libraries statically.
12644
12645       -static-libstdc++
12646           When the g++ program is used to link a C++ program, it normally
12647           automatically links against libstdc++.  If libstdc++ is available
12648           as a shared library, and the -static option is not used, then this
12649           links against the shared version of libstdc++.  That is normally
12650           fine.  However, it is sometimes useful to freeze the version of
12651           libstdc++ used by the program without going all the way to a fully
12652           static link.  The -static-libstdc++ option directs the g++ driver
12653           to link libstdc++ statically, without necessarily linking other
12654           libraries statically.
12655
12656       -symbolic
12657           Bind references to global symbols when building a shared object.
12658           Warn about any unresolved references (unless overridden by the link
12659           editor option -Xlinker -z -Xlinker defs).  Only a few systems
12660           support this option.
12661
12662       -T script
12663           Use script as the linker script.  This option is supported by most
12664           systems using the GNU linker.  On some targets, such as bare-board
12665           targets without an operating system, the -T option may be required
12666           when linking to avoid references to undefined symbols.
12667
12668       -Xlinker option
12669           Pass option as an option to the linker.  You can use this to supply
12670           system-specific linker options that GCC does not recognize.
12671
12672           If you want to pass an option that takes a separate argument, you
12673           must use -Xlinker twice, once for the option and once for the
12674           argument.  For example, to pass -assert definitions, you must write
12675           -Xlinker -assert -Xlinker definitions.  It does not work to write
12676           -Xlinker "-assert definitions", because this passes the entire
12677           string as a single argument, which is not what the linker expects.
12678
12679           When using the GNU linker, it is usually more convenient to pass
12680           arguments to linker options using the option=value syntax than as
12681           separate arguments.  For example, you can specify -Xlinker
12682           -Map=output.map rather than -Xlinker -Map -Xlinker output.map.
12683           Other linkers may not support this syntax for command-line options.
12684
12685       -Wl,option
12686           Pass option as an option to the linker.  If option contains commas,
12687           it is split into multiple options at the commas.  You can use this
12688           syntax to pass an argument to the option.  For example,
12689           -Wl,-Map,output.map passes -Map output.map to the linker.  When
12690           using the GNU linker, you can also get the same effect with
12691           -Wl,-Map=output.map.
12692
12693       -u symbol
12694           Pretend the symbol symbol is undefined, to force linking of library
12695           modules to define it.  You can use -u multiple times with different
12696           symbols to force loading of additional library modules.
12697
12698       -z keyword
12699           -z is passed directly on to the linker along with the keyword
12700           keyword. See the section in the documentation of your linker for
12701           permitted values and their meanings.
12702
12703   Options for Directory Search
12704       These options specify directories to search for header files, for
12705       libraries and for parts of the compiler:
12706
12707       -I dir
12708       -iquote dir
12709       -isystem dir
12710       -idirafter dir
12711           Add the directory dir to the list of directories to be searched for
12712           header files during preprocessing.  If dir begins with = or
12713           $SYSROOT, then the = or $SYSROOT is replaced by the sysroot prefix;
12714           see --sysroot and -isysroot.
12715
12716           Directories specified with -iquote apply only to the quote form of
12717           the directive, "#include "file"".  Directories specified with -I,
12718           -isystem, or -idirafter apply to lookup for both the
12719           "#include "file"" and "#include <file>" directives.
12720
12721           You can specify any number or combination of these options on the
12722           command line to search for header files in several directories.
12723           The lookup order is as follows:
12724
12725           1.  For the quote form of the include directive, the directory of
12726               the current file is searched first.
12727
12728           2.  For the quote form of the include directive, the directories
12729               specified by -iquote options are searched in left-to-right
12730               order, as they appear on the command line.
12731
12732           3.  Directories specified with -I options are scanned in left-to-
12733               right order.
12734
12735           4.  Directories specified with -isystem options are scanned in
12736               left-to-right order.
12737
12738           5.  Standard system directories are scanned.
12739
12740           6.  Directories specified with -idirafter options are scanned in
12741               left-to-right order.
12742
12743           You can use -I to override a system header file, substituting your
12744           own version, since these directories are searched before the
12745           standard system header file directories.  However, you should not
12746           use this option to add directories that contain vendor-supplied
12747           system header files; use -isystem for that.
12748
12749           The -isystem and -idirafter options also mark the directory as a
12750           system directory, so that it gets the same special treatment that
12751           is applied to the standard system directories.
12752
12753           If a standard system include directory, or a directory specified
12754           with -isystem, is also specified with -I, the -I option is ignored.
12755           The directory is still searched but as a system directory at its
12756           normal position in the system include chain.  This is to ensure
12757           that GCC's procedure to fix buggy system headers and the ordering
12758           for the "#include_next" directive are not inadvertently changed.
12759           If you really need to change the search order for system
12760           directories, use the -nostdinc and/or -isystem options.
12761
12762       -I- Split the include path.  This option has been deprecated.  Please
12763           use -iquote instead for -I directories before the -I- and remove
12764           the -I- option.
12765
12766           Any directories specified with -I options before -I- are searched
12767           only for headers requested with "#include "file""; they are not
12768           searched for "#include <file>".  If additional directories are
12769           specified with -I options after the -I-, those directories are
12770           searched for all #include directives.
12771
12772           In addition, -I- inhibits the use of the directory of the current
12773           file directory as the first search directory for "#include "file"".
12774           There is no way to override this effect of -I-.
12775
12776       -iprefix prefix
12777           Specify prefix as the prefix for subsequent -iwithprefix options.
12778           If the prefix represents a directory, you should include the final
12779           /.
12780
12781       -iwithprefix dir
12782       -iwithprefixbefore dir
12783           Append dir to the prefix specified previously with -iprefix, and
12784           add the resulting directory to the include search path.
12785           -iwithprefixbefore puts it in the same place -I would; -iwithprefix
12786           puts it where -idirafter would.
12787
12788       -isysroot dir
12789           This option is like the --sysroot option, but applies only to
12790           header files (except for Darwin targets, where it applies to both
12791           header files and libraries).  See the --sysroot option for more
12792           information.
12793
12794       -imultilib dir
12795           Use dir as a subdirectory of the directory containing target-
12796           specific C++ headers.
12797
12798       -nostdinc
12799           Do not search the standard system directories for header files.
12800           Only the directories explicitly specified with -I, -iquote,
12801           -isystem, and/or -idirafter options (and the directory of the
12802           current file, if appropriate) are searched.
12803
12804       -nostdinc++
12805           Do not search for header files in the C++-specific standard
12806           directories, but do still search the other standard directories.
12807           (This option is used when building the C++ library.)
12808
12809       -iplugindir=dir
12810           Set the directory to search for plugins that are passed by
12811           -fplugin=name instead of -fplugin=path/name.so.  This option is not
12812           meant to be used by the user, but only passed by the driver.
12813
12814       -Ldir
12815           Add directory dir to the list of directories to be searched for -l.
12816
12817       -Bprefix
12818           This option specifies where to find the executables, libraries,
12819           include files, and data files of the compiler itself.
12820
12821           The compiler driver program runs one or more of the subprograms
12822           cpp, cc1, as and ld.  It tries prefix as a prefix for each program
12823           it tries to run, both with and without machine/version/ for the
12824           corresponding target machine and compiler version.
12825
12826           For each subprogram to be run, the compiler driver first tries the
12827           -B prefix, if any.  If that name is not found, or if -B is not
12828           specified, the driver tries two standard prefixes, /usr/lib/gcc/
12829           and /usr/local/lib/gcc/.  If neither of those results in a file
12830           name that is found, the unmodified program name is searched for
12831           using the directories specified in your PATH environment variable.
12832
12833           The compiler checks to see if the path provided by -B refers to a
12834           directory, and if necessary it adds a directory separator character
12835           at the end of the path.
12836
12837           -B prefixes that effectively specify directory names also apply to
12838           libraries in the linker, because the compiler translates these
12839           options into -L options for the linker.  They also apply to include
12840           files in the preprocessor, because the compiler translates these
12841           options into -isystem options for the preprocessor.  In this case,
12842           the compiler appends include to the prefix.
12843
12844           The runtime support file libgcc.a can also be searched for using
12845           the -B prefix, if needed.  If it is not found there, the two
12846           standard prefixes above are tried, and that is all.  The file is
12847           left out of the link if it is not found by those means.
12848
12849           Another way to specify a prefix much like the -B prefix is to use
12850           the environment variable GCC_EXEC_PREFIX.
12851
12852           As a special kludge, if the path provided by -B is [dir/]stageN/,
12853           where N is a number in the range 0 to 9, then it is replaced by
12854           [dir/]include.  This is to help with boot-strapping the compiler.
12855
12856       -no-canonical-prefixes
12857           Do not expand any symbolic links, resolve references to /../ or
12858           /./, or make the path absolute when generating a relative prefix.
12859
12860       --sysroot=dir
12861           Use dir as the logical root directory for headers and libraries.
12862           For example, if the compiler normally searches for headers in
12863           /usr/include and libraries in /usr/lib, it instead searches
12864           dir/usr/include and dir/usr/lib.
12865
12866           If you use both this option and the -isysroot option, then the
12867           --sysroot option applies to libraries, but the -isysroot option
12868           applies to header files.
12869
12870           The GNU linker (beginning with version 2.16) has the necessary
12871           support for this option.  If your linker does not support this
12872           option, the header file aspect of --sysroot still works, but the
12873           library aspect does not.
12874
12875       --no-sysroot-suffix
12876           For some targets, a suffix is added to the root directory specified
12877           with --sysroot, depending on the other options used, so that
12878           headers may for example be found in dir/suffix/usr/include instead
12879           of dir/usr/include.  This option disables the addition of such a
12880           suffix.
12881
12882   Options for Code Generation Conventions
12883       These machine-independent options control the interface conventions
12884       used in code generation.
12885
12886       Most of them have both positive and negative forms; the negative form
12887       of -ffoo is -fno-foo.  In the table below, only one of the forms is
12888       listed---the one that is not the default.  You can figure out the other
12889       form by either removing no- or adding it.
12890
12891       -fstack-reuse=reuse-level
12892           This option controls stack space reuse for user declared local/auto
12893           variables and compiler generated temporaries.  reuse_level can be
12894           all, named_vars, or none. all enables stack reuse for all local
12895           variables and temporaries, named_vars enables the reuse only for
12896           user defined local variables with names, and none disables stack
12897           reuse completely. The default value is all. The option is needed
12898           when the program extends the lifetime of a scoped local variable or
12899           a compiler generated temporary beyond the end point defined by the
12900           language.  When a lifetime of a variable ends, and if the variable
12901           lives in memory, the optimizing compiler has the freedom to reuse
12902           its stack space with other temporaries or scoped local variables
12903           whose live range does not overlap with it. Legacy code extending
12904           local lifetime is likely to break with the stack reuse
12905           optimization.
12906
12907           For example,
12908
12909                      int *p;
12910                      {
12911                        int local1;
12912
12913                        p = &local1;
12914                        local1 = 10;
12915                        ....
12916                      }
12917                      {
12918                         int local2;
12919                         local2 = 20;
12920                         ...
12921                      }
12922
12923                      if (*p == 10)  // out of scope use of local1
12924                        {
12925
12926                        }
12927
12928           Another example:
12929
12930                      struct A
12931                      {
12932                          A(int k) : i(k), j(k) { }
12933                          int i;
12934                          int j;
12935                      };
12936
12937                      A *ap;
12938
12939                      void foo(const A& ar)
12940                      {
12941                         ap = &ar;
12942                      }
12943
12944                      void bar()
12945                      {
12946                         foo(A(10)); // temp object's lifetime ends when foo returns
12947
12948                         {
12949                           A a(20);
12950                           ....
12951                         }
12952                         ap->i+= 10;  // ap references out of scope temp whose space
12953                                      // is reused with a. What is the value of ap->i?
12954                      }
12955
12956           The lifetime of a compiler generated temporary is well defined by
12957           the C++ standard. When a lifetime of a temporary ends, and if the
12958           temporary lives in memory, the optimizing compiler has the freedom
12959           to reuse its stack space with other temporaries or scoped local
12960           variables whose live range does not overlap with it. However some
12961           of the legacy code relies on the behavior of older compilers in
12962           which temporaries' stack space is not reused, the aggressive stack
12963           reuse can lead to runtime errors. This option is used to control
12964           the temporary stack reuse optimization.
12965
12966       -ftrapv
12967           This option generates traps for signed overflow on addition,
12968           subtraction, multiplication operations.  The options -ftrapv and
12969           -fwrapv override each other, so using -ftrapv -fwrapv on the
12970           command-line results in -fwrapv being effective.  Note that only
12971           active options override, so using -ftrapv -fwrapv -fno-wrapv on the
12972           command-line results in -ftrapv being effective.
12973
12974       -fwrapv
12975           This option instructs the compiler to assume that signed arithmetic
12976           overflow of addition, subtraction and multiplication wraps around
12977           using twos-complement representation.  This flag enables some
12978           optimizations and disables others.  The options -ftrapv and -fwrapv
12979           override each other, so using -ftrapv -fwrapv on the command-line
12980           results in -fwrapv being effective.  Note that only active options
12981           override, so using -ftrapv -fwrapv -fno-wrapv on the command-line
12982           results in -ftrapv being effective.
12983
12984       -fwrapv-pointer
12985           This option instructs the compiler to assume that pointer
12986           arithmetic overflow on addition and subtraction wraps around using
12987           twos-complement representation.  This flag disables some
12988           optimizations which assume pointer overflow is invalid.
12989
12990       -fstrict-overflow
12991           This option implies -fno-wrapv -fno-wrapv-pointer and when negated
12992           implies -fwrapv -fwrapv-pointer.
12993
12994       -fexceptions
12995           Enable exception handling.  Generates extra code needed to
12996           propagate exceptions.  For some targets, this implies GCC generates
12997           frame unwind information for all functions, which can produce
12998           significant data size overhead, although it does not affect
12999           execution.  If you do not specify this option, GCC enables it by
13000           default for languages like C++ that normally require exception
13001           handling, and disables it for languages like C that do not normally
13002           require it.  However, you may need to enable this option when
13003           compiling C code that needs to interoperate properly with exception
13004           handlers written in C++.  You may also wish to disable this option
13005           if you are compiling older C++ programs that don't use exception
13006           handling.
13007
13008       -fnon-call-exceptions
13009           Generate code that allows trapping instructions to throw
13010           exceptions.  Note that this requires platform-specific runtime
13011           support that does not exist everywhere.  Moreover, it only allows
13012           trapping instructions to throw exceptions, i.e. memory references
13013           or floating-point instructions.  It does not allow exceptions to be
13014           thrown from arbitrary signal handlers such as "SIGALRM".
13015
13016       -fdelete-dead-exceptions
13017           Consider that instructions that may throw exceptions but don't
13018           otherwise contribute to the execution of the program can be
13019           optimized away.  This option is enabled by default for the Ada
13020           front end, as permitted by the Ada language specification.
13021           Optimization passes that cause dead exceptions to be removed are
13022           enabled independently at different optimization levels.
13023
13024       -funwind-tables
13025           Similar to -fexceptions, except that it just generates any needed
13026           static data, but does not affect the generated code in any other
13027           way.  You normally do not need to enable this option; instead, a
13028           language processor that needs this handling enables it on your
13029           behalf.
13030
13031       -fasynchronous-unwind-tables
13032           Generate unwind table in DWARF format, if supported by target
13033           machine.  The table is exact at each instruction boundary, so it
13034           can be used for stack unwinding from asynchronous events (such as
13035           debugger or garbage collector).
13036
13037       -fno-gnu-unique
13038           On systems with recent GNU assembler and C library, the C++
13039           compiler uses the "STB_GNU_UNIQUE" binding to make sure that
13040           definitions of template static data members and static local
13041           variables in inline functions are unique even in the presence of
13042           "RTLD_LOCAL"; this is necessary to avoid problems with a library
13043           used by two different "RTLD_LOCAL" plugins depending on a
13044           definition in one of them and therefore disagreeing with the other
13045           one about the binding of the symbol.  But this causes "dlclose" to
13046           be ignored for affected DSOs; if your program relies on
13047           reinitialization of a DSO via "dlclose" and "dlopen", you can use
13048           -fno-gnu-unique.
13049
13050       -fpcc-struct-return
13051           Return "short" "struct" and "union" values in memory like longer
13052           ones, rather than in registers.  This convention is less efficient,
13053           but it has the advantage of allowing intercallability between GCC-
13054           compiled files and files compiled with other compilers,
13055           particularly the Portable C Compiler (pcc).
13056
13057           The precise convention for returning structures in memory depends
13058           on the target configuration macros.
13059
13060           Short structures and unions are those whose size and alignment
13061           match that of some integer type.
13062
13063           Warning: code compiled with the -fpcc-struct-return switch is not
13064           binary compatible with code compiled with the -freg-struct-return
13065           switch.  Use it to conform to a non-default application binary
13066           interface.
13067
13068       -freg-struct-return
13069           Return "struct" and "union" values in registers when possible.
13070           This is more efficient for small structures than
13071           -fpcc-struct-return.
13072
13073           If you specify neither -fpcc-struct-return nor -freg-struct-return,
13074           GCC defaults to whichever convention is standard for the target.
13075           If there is no standard convention, GCC defaults to
13076           -fpcc-struct-return, except on targets where GCC is the principal
13077           compiler.  In those cases, we can choose the standard, and we chose
13078           the more efficient register return alternative.
13079
13080           Warning: code compiled with the -freg-struct-return switch is not
13081           binary compatible with code compiled with the -fpcc-struct-return
13082           switch.  Use it to conform to a non-default application binary
13083           interface.
13084
13085       -fshort-enums
13086           Allocate to an "enum" type only as many bytes as it needs for the
13087           declared range of possible values.  Specifically, the "enum" type
13088           is equivalent to the smallest integer type that has enough room.
13089
13090           Warning: the -fshort-enums switch causes GCC to generate code that
13091           is not binary compatible with code generated without that switch.
13092           Use it to conform to a non-default application binary interface.
13093
13094       -fshort-wchar
13095           Override the underlying type for "wchar_t" to be "short unsigned
13096           int" instead of the default for the target.  This option is useful
13097           for building programs to run under WINE.
13098
13099           Warning: the -fshort-wchar switch causes GCC to generate code that
13100           is not binary compatible with code generated without that switch.
13101           Use it to conform to a non-default application binary interface.
13102
13103       -fcommon
13104           In C code, this option controls the placement of global variables
13105           defined without an initializer, known as tentative definitions in
13106           the C standard.  Tentative definitions are distinct from
13107           declarations of a variable with the "extern" keyword, which do not
13108           allocate storage.
13109
13110           The default is -fno-common, which specifies that the compiler
13111           places uninitialized global variables in the BSS section of the
13112           object file.  This inhibits the merging of tentative definitions by
13113           the linker so you get a multiple-definition error if the same
13114           variable is accidentally defined in more than one compilation unit.
13115
13116           The -fcommon places uninitialized global variables in a common
13117           block.  This allows the linker to resolve all tentative definitions
13118           of the same variable in different compilation units to the same
13119           object, or to a non-tentative definition.  This behavior is
13120           inconsistent with C++, and on many targets implies a speed and code
13121           size penalty on global variable references.  It is mainly useful to
13122           enable legacy code to link without errors.
13123
13124       -fno-ident
13125           Ignore the "#ident" directive.
13126
13127       -finhibit-size-directive
13128           Don't output a ".size" assembler directive, or anything else that
13129           would cause trouble if the function is split in the middle, and the
13130           two halves are placed at locations far apart in memory.  This
13131           option is used when compiling crtstuff.c; you should not need to
13132           use it for anything else.
13133
13134       -fverbose-asm
13135           Put extra commentary information in the generated assembly code to
13136           make it more readable.  This option is generally only of use to
13137           those who actually need to read the generated assembly code
13138           (perhaps while debugging the compiler itself).
13139
13140           -fno-verbose-asm, the default, causes the extra information to be
13141           omitted and is useful when comparing two assembler files.
13142
13143           The added comments include:
13144
13145           *   information on the compiler version and command-line options,
13146
13147           *   the source code lines associated with the assembly
13148               instructions, in the form FILENAME:LINENUMBER:CONTENT OF LINE,
13149
13150           *   hints on which high-level expressions correspond to the various
13151               assembly instruction operands.
13152
13153           For example, given this C source file:
13154
13155                   int test (int n)
13156                   {
13157                     int i;
13158                     int total = 0;
13159
13160                     for (i = 0; i < n; i++)
13161                       total += i * i;
13162
13163                     return total;
13164                   }
13165
13166           compiling to (x86_64) assembly via -S and emitting the result
13167           direct to stdout via -o -
13168
13169                   gcc -S test.c -fverbose-asm -Os -o -
13170
13171           gives output similar to this:
13172
13173                           .file   "test.c"
13174                   # GNU C11 (GCC) version 7.0.0 20160809 (experimental) (x86_64-pc-linux-gnu)
13175                     [...snip...]
13176                   # options passed:
13177                     [...snip...]
13178
13179                           .text
13180                           .globl  test
13181                           .type   test, @function
13182                   test:
13183                   .LFB0:
13184                           .cfi_startproc
13185                   # test.c:4:   int total = 0;
13186                           xorl    %eax, %eax      # <retval>
13187                   # test.c:6:   for (i = 0; i < n; i++)
13188                           xorl    %edx, %edx      # i
13189                   .L2:
13190                   # test.c:6:   for (i = 0; i < n; i++)
13191                           cmpl    %edi, %edx      # n, i
13192                           jge     .L5     #,
13193                   # test.c:7:     total += i * i;
13194                           movl    %edx, %ecx      # i, tmp92
13195                           imull   %edx, %ecx      # i, tmp92
13196                   # test.c:6:   for (i = 0; i < n; i++)
13197                           incl    %edx    # i
13198                   # test.c:7:     total += i * i;
13199                           addl    %ecx, %eax      # tmp92, <retval>
13200                           jmp     .L2     #
13201                   .L5:
13202                   # test.c:10: }
13203                           ret
13204                           .cfi_endproc
13205                   .LFE0:
13206                           .size   test, .-test
13207                           .ident  "GCC: (GNU) 7.0.0 20160809 (experimental)"
13208                           .section        .note.GNU-stack,"",@progbits
13209
13210           The comments are intended for humans rather than machines and hence
13211           the precise format of the comments is subject to change.
13212
13213       -frecord-gcc-switches
13214           This switch causes the command line used to invoke the compiler to
13215           be recorded into the object file that is being created.  This
13216           switch is only implemented on some targets and the exact format of
13217           the recording is target and binary file format dependent, but it
13218           usually takes the form of a section containing ASCII text.  This
13219           switch is related to the -fverbose-asm switch, but that switch only
13220           records information in the assembler output file as comments, so it
13221           never reaches the object file.  See also -grecord-gcc-switches for
13222           another way of storing compiler options into the object file.
13223
13224       -fpic
13225           Generate position-independent code (PIC) suitable for use in a
13226           shared library, if supported for the target machine.  Such code
13227           accesses all constant addresses through a global offset table
13228           (GOT).  The dynamic loader resolves the GOT entries when the
13229           program starts (the dynamic loader is not part of GCC; it is part
13230           of the operating system).  If the GOT size for the linked
13231           executable exceeds a machine-specific maximum size, you get an
13232           error message from the linker indicating that -fpic does not work;
13233           in that case, recompile with -fPIC instead.  (These maximums are 8k
13234           on the SPARC, 28k on AArch64 and 32k on the m68k and RS/6000.  The
13235           x86 has no such limit.)
13236
13237           Position-independent code requires special support, and therefore
13238           works only on certain machines.  For the x86, GCC supports PIC for
13239           System V but not for the Sun 386i.  Code generated for the IBM
13240           RS/6000 is always position-independent.
13241
13242           When this flag is set, the macros "__pic__" and "__PIC__" are
13243           defined to 1.
13244
13245       -fPIC
13246           If supported for the target machine, emit position-independent
13247           code, suitable for dynamic linking and avoiding any limit on the
13248           size of the global offset table.  This option makes a difference on
13249           AArch64, m68k, PowerPC and SPARC.
13250
13251           Position-independent code requires special support, and therefore
13252           works only on certain machines.
13253
13254           When this flag is set, the macros "__pic__" and "__PIC__" are
13255           defined to 2.
13256
13257       -fpie
13258       -fPIE
13259           These options are similar to -fpic and -fPIC, but the generated
13260           position-independent code can be only linked into executables.
13261           Usually these options are used to compile code that will be linked
13262           using the -pie GCC option.
13263
13264           -fpie and -fPIE both define the macros "__pie__" and "__PIE__".
13265           The macros have the value 1 for -fpie and 2 for -fPIE.
13266
13267       -fno-plt
13268           Do not use the PLT for external function calls in position-
13269           independent code.  Instead, load the callee address at call sites
13270           from the GOT and branch to it.  This leads to more efficient code
13271           by eliminating PLT stubs and exposing GOT loads to optimizations.
13272           On architectures such as 32-bit x86 where PLT stubs expect the GOT
13273           pointer in a specific register, this gives more register allocation
13274           freedom to the compiler.  Lazy binding requires use of the PLT;
13275           with -fno-plt all external symbols are resolved at load time.
13276
13277           Alternatively, the function attribute "noplt" can be used to avoid
13278           calls through the PLT for specific external functions.
13279
13280           In position-dependent code, a few targets also convert calls to
13281           functions that are marked to not use the PLT to use the GOT
13282           instead.
13283
13284       -fno-jump-tables
13285           Do not use jump tables for switch statements even where it would be
13286           more efficient than other code generation strategies.  This option
13287           is of use in conjunction with -fpic or -fPIC for building code that
13288           forms part of a dynamic linker and cannot reference the address of
13289           a jump table.  On some targets, jump tables do not require a GOT
13290           and this option is not needed.
13291
13292       -ffixed-reg
13293           Treat the register named reg as a fixed register; generated code
13294           should never refer to it (except perhaps as a stack pointer, frame
13295           pointer or in some other fixed role).
13296
13297           reg must be the name of a register.  The register names accepted
13298           are machine-specific and are defined in the "REGISTER_NAMES" macro
13299           in the machine description macro file.
13300
13301           This flag does not have a negative form, because it specifies a
13302           three-way choice.
13303
13304       -fcall-used-reg
13305           Treat the register named reg as an allocable register that is
13306           clobbered by function calls.  It may be allocated for temporaries
13307           or variables that do not live across a call.  Functions compiled
13308           this way do not save and restore the register reg.
13309
13310           It is an error to use this flag with the frame pointer or stack
13311           pointer.  Use of this flag for other registers that have fixed
13312           pervasive roles in the machine's execution model produces
13313           disastrous results.
13314
13315           This flag does not have a negative form, because it specifies a
13316           three-way choice.
13317
13318       -fcall-saved-reg
13319           Treat the register named reg as an allocable register saved by
13320           functions.  It may be allocated even for temporaries or variables
13321           that live across a call.  Functions compiled this way save and
13322           restore the register reg if they use it.
13323
13324           It is an error to use this flag with the frame pointer or stack
13325           pointer.  Use of this flag for other registers that have fixed
13326           pervasive roles in the machine's execution model produces
13327           disastrous results.
13328
13329           A different sort of disaster results from the use of this flag for
13330           a register in which function values may be returned.
13331
13332           This flag does not have a negative form, because it specifies a
13333           three-way choice.
13334
13335       -fpack-struct[=n]
13336           Without a value specified, pack all structure members together
13337           without holes.  When a value is specified (which must be a small
13338           power of two), pack structure members according to this value,
13339           representing the maximum alignment (that is, objects with default
13340           alignment requirements larger than this are output potentially
13341           unaligned at the next fitting location.
13342
13343           Warning: the -fpack-struct switch causes GCC to generate code that
13344           is not binary compatible with code generated without that switch.
13345           Additionally, it makes the code suboptimal.  Use it to conform to a
13346           non-default application binary interface.
13347
13348       -fleading-underscore
13349           This option and its counterpart, -fno-leading-underscore, forcibly
13350           change the way C symbols are represented in the object file.  One
13351           use is to help link with legacy assembly code.
13352
13353           Warning: the -fleading-underscore switch causes GCC to generate
13354           code that is not binary compatible with code generated without that
13355           switch.  Use it to conform to a non-default application binary
13356           interface.  Not all targets provide complete support for this
13357           switch.
13358
13359       -ftls-model=model
13360           Alter the thread-local storage model to be used.  The model
13361           argument should be one of global-dynamic, local-dynamic, initial-
13362           exec or local-exec.  Note that the choice is subject to
13363           optimization: the compiler may use a more efficient model for
13364           symbols not visible outside of the translation unit, or if -fpic is
13365           not given on the command line.
13366
13367           The default without -fpic is initial-exec; with -fpic the default
13368           is global-dynamic.
13369
13370       -ftrampolines
13371           For targets that normally need trampolines for nested functions,
13372           always generate them instead of using descriptors.  Otherwise, for
13373           targets that do not need them, like for example HP-PA or IA-64, do
13374           nothing.
13375
13376           A trampoline is a small piece of code that is created at run time
13377           on the stack when the address of a nested function is taken, and is
13378           used to call the nested function indirectly.  Therefore, it
13379           requires the stack to be made executable in order for the program
13380           to work properly.
13381
13382           -fno-trampolines is enabled by default on a language by language
13383           basis to let the compiler avoid generating them, if it computes
13384           that this is safe, and replace them with descriptors.  Descriptors
13385           are made up of data only, but the generated code must be prepared
13386           to deal with them.  As of this writing, -fno-trampolines is enabled
13387           by default only for Ada.
13388
13389           Moreover, code compiled with -ftrampolines and code compiled with
13390           -fno-trampolines are not binary compatible if nested functions are
13391           present.  This option must therefore be used on a program-wide
13392           basis and be manipulated with extreme care.
13393
13394       -fvisibility=[default|internal|hidden|protected]
13395           Set the default ELF image symbol visibility to the specified
13396           option---all symbols are marked with this unless overridden within
13397           the code.  Using this feature can very substantially improve
13398           linking and load times of shared object libraries, produce more
13399           optimized code, provide near-perfect API export and prevent symbol
13400           clashes.  It is strongly recommended that you use this in any
13401           shared objects you distribute.
13402
13403           Despite the nomenclature, default always means public; i.e.,
13404           available to be linked against from outside the shared object.
13405           protected and internal are pretty useless in real-world usage so
13406           the only other commonly used option is hidden.  The default if
13407           -fvisibility isn't specified is default, i.e., make every symbol
13408           public.
13409
13410           A good explanation of the benefits offered by ensuring ELF symbols
13411           have the correct visibility is given by "How To Write Shared
13412           Libraries" by Ulrich Drepper (which can be found at
13413           <https://www.akkadia.org/drepper/>)---however a superior solution
13414           made possible by this option to marking things hidden when the
13415           default is public is to make the default hidden and mark things
13416           public.  This is the norm with DLLs on Windows and with
13417           -fvisibility=hidden and "__attribute__ ((visibility("default")))"
13418           instead of "__declspec(dllexport)" you get almost identical
13419           semantics with identical syntax.  This is a great boon to those
13420           working with cross-platform projects.
13421
13422           For those adding visibility support to existing code, you may find
13423           "#pragma GCC visibility" of use.  This works by you enclosing the
13424           declarations you wish to set visibility for with (for example)
13425           "#pragma GCC visibility push(hidden)" and "#pragma GCC visibility
13426           pop".  Bear in mind that symbol visibility should be viewed as part
13427           of the API interface contract and thus all new code should always
13428           specify visibility when it is not the default; i.e., declarations
13429           only for use within the local DSO should always be marked
13430           explicitly as hidden as so to avoid PLT indirection
13431           overheads---making this abundantly clear also aids readability and
13432           self-documentation of the code.  Note that due to ISO C++
13433           specification requirements, "operator new" and "operator delete"
13434           must always be of default visibility.
13435
13436           Be aware that headers from outside your project, in particular
13437           system headers and headers from any other library you use, may not
13438           be expecting to be compiled with visibility other than the default.
13439           You may need to explicitly say "#pragma GCC visibility
13440           push(default)" before including any such headers.
13441
13442           "extern" declarations are not affected by -fvisibility, so a lot of
13443           code can be recompiled with -fvisibility=hidden with no
13444           modifications.  However, this means that calls to "extern"
13445           functions with no explicit visibility use the PLT, so it is more
13446           effective to use "__attribute ((visibility))" and/or "#pragma GCC
13447           visibility" to tell the compiler which "extern" declarations should
13448           be treated as hidden.
13449
13450           Note that -fvisibility does affect C++ vague linkage entities. This
13451           means that, for instance, an exception class that is be thrown
13452           between DSOs must be explicitly marked with default visibility so
13453           that the type_info nodes are unified between the DSOs.
13454
13455           An overview of these techniques, their benefits and how to use them
13456           is at <http://gcc.gnu.org/wiki/Visibility>.
13457
13458       -fstrict-volatile-bitfields
13459           This option should be used if accesses to volatile bit-fields (or
13460           other structure fields, although the compiler usually honors those
13461           types anyway) should use a single access of the width of the
13462           field's type, aligned to a natural alignment if possible.  For
13463           example, targets with memory-mapped peripheral registers might
13464           require all such accesses to be 16 bits wide; with this flag you
13465           can declare all peripheral bit-fields as "unsigned short" (assuming
13466           short is 16 bits on these targets) to force GCC to use 16-bit
13467           accesses instead of, perhaps, a more efficient 32-bit access.
13468
13469           If this option is disabled, the compiler uses the most efficient
13470           instruction.  In the previous example, that might be a 32-bit load
13471           instruction, even though that accesses bytes that do not contain
13472           any portion of the bit-field, or memory-mapped registers unrelated
13473           to the one being updated.
13474
13475           In some cases, such as when the "packed" attribute is applied to a
13476           structure field, it may not be possible to access the field with a
13477           single read or write that is correctly aligned for the target
13478           machine.  In this case GCC falls back to generating multiple
13479           accesses rather than code that will fault or truncate the result at
13480           run time.
13481
13482           Note:  Due to restrictions of the C/C++11 memory model, write
13483           accesses are not allowed to touch non bit-field members.  It is
13484           therefore recommended to define all bits of the field's type as
13485           bit-field members.
13486
13487           The default value of this option is determined by the application
13488           binary interface for the target processor.
13489
13490       -fsync-libcalls
13491           This option controls whether any out-of-line instance of the
13492           "__sync" family of functions may be used to implement the C++11
13493           "__atomic" family of functions.
13494
13495           The default value of this option is enabled, thus the only useful
13496           form of the option is -fno-sync-libcalls.  This option is used in
13497           the implementation of the libatomic runtime library.
13498
13499   GCC Developer Options
13500       This section describes command-line options that are primarily of
13501       interest to GCC developers, including options to support compiler
13502       testing and investigation of compiler bugs and compile-time performance
13503       problems.  This includes options that produce debug dumps at various
13504       points in the compilation; that print statistics such as memory use and
13505       execution time; and that print information about GCC's configuration,
13506       such as where it searches for libraries.  You should rarely need to use
13507       any of these options for ordinary compilation and linking tasks.
13508
13509       Many developer options that cause GCC to dump output to a file take an
13510       optional =filename suffix. You can specify stdout or - to dump to
13511       standard output, and stderr for standard error.
13512
13513       If =filename is omitted, a default dump file name is constructed by
13514       concatenating the base dump file name, a pass number, phase letter, and
13515       pass name.  The base dump file name is the name of output file produced
13516       by the compiler if explicitly specified and not an executable;
13517       otherwise it is the source file name.  The pass number is determined by
13518       the order passes are registered with the compiler's pass manager.  This
13519       is generally the same as the order of execution, but passes registered
13520       by plugins, target-specific passes, or passes that are otherwise
13521       registered late are numbered higher than the pass named final, even if
13522       they are executed earlier.  The phase letter is one of i (inter-
13523       procedural analysis), l (language-specific), r (RTL), or t (tree).  The
13524       files are created in the directory of the output file.
13525
13526       -fcallgraph-info
13527       -fcallgraph-info=MARKERS
13528           Makes the compiler output callgraph information for the program, on
13529           a per-object-file basis.  The information is generated in the
13530           common VCG format.  It can be decorated with additional, per-node
13531           and/or per-edge information, if a list of comma-separated markers
13532           is additionally specified.  When the "su" marker is specified, the
13533           callgraph is decorated with stack usage information; it is
13534           equivalent to -fstack-usage.  When the "da" marker is specified,
13535           the callgraph is decorated with information about dynamically
13536           allocated objects.
13537
13538           When compiling with -flto, no callgraph information is output along
13539           with the object file.  At LTO link time, -fcallgraph-info may
13540           generate multiple callgraph information files next to intermediate
13541           LTO output files.
13542
13543       -dletters
13544       -fdump-rtl-pass
13545       -fdump-rtl-pass=filename
13546           Says to make debugging dumps during compilation at times specified
13547           by letters.  This is used for debugging the RTL-based passes of the
13548           compiler.
13549
13550           Some -dletters switches have different meaning when -E is used for
13551           preprocessing.
13552
13553           Debug dumps can be enabled with a -fdump-rtl switch or some -d
13554           option letters.  Here are the possible letters for use in pass and
13555           letters, and their meanings:
13556
13557           -fdump-rtl-alignments
13558               Dump after branch alignments have been computed.
13559
13560           -fdump-rtl-asmcons
13561               Dump after fixing rtl statements that have unsatisfied in/out
13562               constraints.
13563
13564           -fdump-rtl-auto_inc_dec
13565               Dump after auto-inc-dec discovery.  This pass is only run on
13566               architectures that have auto inc or auto dec instructions.
13567
13568           -fdump-rtl-barriers
13569               Dump after cleaning up the barrier instructions.
13570
13571           -fdump-rtl-bbpart
13572               Dump after partitioning hot and cold basic blocks.
13573
13574           -fdump-rtl-bbro
13575               Dump after block reordering.
13576
13577           -fdump-rtl-btl1
13578           -fdump-rtl-btl2
13579               -fdump-rtl-btl1 and -fdump-rtl-btl2 enable dumping after the
13580               two branch target load optimization passes.
13581
13582           -fdump-rtl-bypass
13583               Dump after jump bypassing and control flow optimizations.
13584
13585           -fdump-rtl-combine
13586               Dump after the RTL instruction combination pass.
13587
13588           -fdump-rtl-compgotos
13589               Dump after duplicating the computed gotos.
13590
13591           -fdump-rtl-ce1
13592           -fdump-rtl-ce2
13593           -fdump-rtl-ce3
13594               -fdump-rtl-ce1, -fdump-rtl-ce2, and -fdump-rtl-ce3 enable
13595               dumping after the three if conversion passes.
13596
13597           -fdump-rtl-cprop_hardreg
13598               Dump after hard register copy propagation.
13599
13600           -fdump-rtl-csa
13601               Dump after combining stack adjustments.
13602
13603           -fdump-rtl-cse1
13604           -fdump-rtl-cse2
13605               -fdump-rtl-cse1 and -fdump-rtl-cse2 enable dumping after the
13606               two common subexpression elimination passes.
13607
13608           -fdump-rtl-dce
13609               Dump after the standalone dead code elimination passes.
13610
13611           -fdump-rtl-dbr
13612               Dump after delayed branch scheduling.
13613
13614           -fdump-rtl-dce1
13615           -fdump-rtl-dce2
13616               -fdump-rtl-dce1 and -fdump-rtl-dce2 enable dumping after the
13617               two dead store elimination passes.
13618
13619           -fdump-rtl-eh
13620               Dump after finalization of EH handling code.
13621
13622           -fdump-rtl-eh_ranges
13623               Dump after conversion of EH handling range regions.
13624
13625           -fdump-rtl-expand
13626               Dump after RTL generation.
13627
13628           -fdump-rtl-fwprop1
13629           -fdump-rtl-fwprop2
13630               -fdump-rtl-fwprop1 and -fdump-rtl-fwprop2 enable dumping after
13631               the two forward propagation passes.
13632
13633           -fdump-rtl-gcse1
13634           -fdump-rtl-gcse2
13635               -fdump-rtl-gcse1 and -fdump-rtl-gcse2 enable dumping after
13636               global common subexpression elimination.
13637
13638           -fdump-rtl-init-regs
13639               Dump after the initialization of the registers.
13640
13641           -fdump-rtl-initvals
13642               Dump after the computation of the initial value sets.
13643
13644           -fdump-rtl-into_cfglayout
13645               Dump after converting to cfglayout mode.
13646
13647           -fdump-rtl-ira
13648               Dump after iterated register allocation.
13649
13650           -fdump-rtl-jump
13651               Dump after the second jump optimization.
13652
13653           -fdump-rtl-loop2
13654               -fdump-rtl-loop2 enables dumping after the rtl loop
13655               optimization passes.
13656
13657           -fdump-rtl-mach
13658               Dump after performing the machine dependent reorganization
13659               pass, if that pass exists.
13660
13661           -fdump-rtl-mode_sw
13662               Dump after removing redundant mode switches.
13663
13664           -fdump-rtl-rnreg
13665               Dump after register renumbering.
13666
13667           -fdump-rtl-outof_cfglayout
13668               Dump after converting from cfglayout mode.
13669
13670           -fdump-rtl-peephole2
13671               Dump after the peephole pass.
13672
13673           -fdump-rtl-postreload
13674               Dump after post-reload optimizations.
13675
13676           -fdump-rtl-pro_and_epilogue
13677               Dump after generating the function prologues and epilogues.
13678
13679           -fdump-rtl-sched1
13680           -fdump-rtl-sched2
13681               -fdump-rtl-sched1 and -fdump-rtl-sched2 enable dumping after
13682               the basic block scheduling passes.
13683
13684           -fdump-rtl-ree
13685               Dump after sign/zero extension elimination.
13686
13687           -fdump-rtl-seqabstr
13688               Dump after common sequence discovery.
13689
13690           -fdump-rtl-shorten
13691               Dump after shortening branches.
13692
13693           -fdump-rtl-sibling
13694               Dump after sibling call optimizations.
13695
13696           -fdump-rtl-split1
13697           -fdump-rtl-split2
13698           -fdump-rtl-split3
13699           -fdump-rtl-split4
13700           -fdump-rtl-split5
13701               These options enable dumping after five rounds of instruction
13702               splitting.
13703
13704           -fdump-rtl-sms
13705               Dump after modulo scheduling.  This pass is only run on some
13706               architectures.
13707
13708           -fdump-rtl-stack
13709               Dump after conversion from GCC's "flat register file" registers
13710               to the x87's stack-like registers.  This pass is only run on
13711               x86 variants.
13712
13713           -fdump-rtl-subreg1
13714           -fdump-rtl-subreg2
13715               -fdump-rtl-subreg1 and -fdump-rtl-subreg2 enable dumping after
13716               the two subreg expansion passes.
13717
13718           -fdump-rtl-unshare
13719               Dump after all rtl has been unshared.
13720
13721           -fdump-rtl-vartrack
13722               Dump after variable tracking.
13723
13724           -fdump-rtl-vregs
13725               Dump after converting virtual registers to hard registers.
13726
13727           -fdump-rtl-web
13728               Dump after live range splitting.
13729
13730           -fdump-rtl-regclass
13731           -fdump-rtl-subregs_of_mode_init
13732           -fdump-rtl-subregs_of_mode_finish
13733           -fdump-rtl-dfinit
13734           -fdump-rtl-dfinish
13735               These dumps are defined but always produce empty files.
13736
13737           -da
13738           -fdump-rtl-all
13739               Produce all the dumps listed above.
13740
13741           -dA Annotate the assembler output with miscellaneous debugging
13742               information.
13743
13744           -dD Dump all macro definitions, at the end of preprocessing, in
13745               addition to normal output.
13746
13747           -dH Produce a core dump whenever an error occurs.
13748
13749           -dp Annotate the assembler output with a comment indicating which
13750               pattern and alternative is used.  The length and cost of each
13751               instruction are also printed.
13752
13753           -dP Dump the RTL in the assembler output as a comment before each
13754               instruction.  Also turns on -dp annotation.
13755
13756           -dx Just generate RTL for a function instead of compiling it.
13757               Usually used with -fdump-rtl-expand.
13758
13759       -fdump-debug
13760           Dump debugging information generated during the debug generation
13761           phase.
13762
13763       -fdump-earlydebug
13764           Dump debugging information generated during the early debug
13765           generation phase.
13766
13767       -fdump-noaddr
13768           When doing debugging dumps, suppress address output.  This makes it
13769           more feasible to use diff on debugging dumps for compiler
13770           invocations with different compiler binaries and/or different text
13771           / bss / data / heap / stack / dso start locations.
13772
13773       -freport-bug
13774           Collect and dump debug information into a temporary file if an
13775           internal compiler error (ICE) occurs.
13776
13777       -fdump-unnumbered
13778           When doing debugging dumps, suppress instruction numbers and
13779           address output.  This makes it more feasible to use diff on
13780           debugging dumps for compiler invocations with different options, in
13781           particular with and without -g.
13782
13783       -fdump-unnumbered-links
13784           When doing debugging dumps (see -d option above), suppress
13785           instruction numbers for the links to the previous and next
13786           instructions in a sequence.
13787
13788       -fdump-ipa-switch
13789       -fdump-ipa-switch-options
13790           Control the dumping at various stages of inter-procedural analysis
13791           language tree to a file.  The file name is generated by appending a
13792           switch specific suffix to the source file name, and the file is
13793           created in the same directory as the output file.  The following
13794           dumps are possible:
13795
13796           all Enables all inter-procedural analysis dumps.
13797
13798           cgraph
13799               Dumps information about call-graph optimization, unused
13800               function removal, and inlining decisions.
13801
13802           inline
13803               Dump after function inlining.
13804
13805           Additionally, the options -optimized, -missed, -note, and -all can
13806           be provided, with the same meaning as for -fopt-info, defaulting to
13807           -optimized.
13808
13809           For example, -fdump-ipa-inline-optimized-missed will emit
13810           information on callsites that were inlined, along with callsites
13811           that were not inlined.
13812
13813           By default, the dump will contain messages about successful
13814           optimizations (equivalent to -optimized) together with low-level
13815           details about the analysis.
13816
13817       -fdump-lang-all
13818       -fdump-lang-switch
13819       -fdump-lang-switch-options
13820       -fdump-lang-switch-options=filename
13821           Control the dumping of language-specific information.  The options
13822           and filename portions behave as described in the -fdump-tree
13823           option.  The following switch values are accepted:
13824
13825           all Enable all language-specific dumps.
13826
13827           class
13828               Dump class hierarchy information.  Virtual table information is
13829               emitted unless 'slim' is specified.  This option is applicable
13830               to C++ only.
13831
13832           raw Dump the raw internal tree data.  This option is applicable to
13833               C++ only.
13834
13835       -fdump-passes
13836           Print on stderr the list of optimization passes that are turned on
13837           and off by the current command-line options.
13838
13839       -fdump-statistics-option
13840           Enable and control dumping of pass statistics in a separate file.
13841           The file name is generated by appending a suffix ending in
13842           .statistics to the source file name, and the file is created in the
13843           same directory as the output file.  If the -option form is used,
13844           -stats causes counters to be summed over the whole compilation unit
13845           while -details dumps every event as the passes generate them.  The
13846           default with no option is to sum counters for each function
13847           compiled.
13848
13849       -fdump-tree-all
13850       -fdump-tree-switch
13851       -fdump-tree-switch-options
13852       -fdump-tree-switch-options=filename
13853           Control the dumping at various stages of processing the
13854           intermediate language tree to a file.  If the -options form is
13855           used, options is a list of - separated options which control the
13856           details of the dump.  Not all options are applicable to all dumps;
13857           those that are not meaningful are ignored.  The following options
13858           are available
13859
13860           address
13861               Print the address of each node.  Usually this is not meaningful
13862               as it changes according to the environment and source file.
13863               Its primary use is for tying up a dump file with a debug
13864               environment.
13865
13866           asmname
13867               If "DECL_ASSEMBLER_NAME" has been set for a given decl, use
13868               that in the dump instead of "DECL_NAME".  Its primary use is
13869               ease of use working backward from mangled names in the assembly
13870               file.
13871
13872           slim
13873               When dumping front-end intermediate representations, inhibit
13874               dumping of members of a scope or body of a function merely
13875               because that scope has been reached.  Only dump such items when
13876               they are directly reachable by some other path.
13877
13878               When dumping pretty-printed trees, this option inhibits dumping
13879               the bodies of control structures.
13880
13881               When dumping RTL, print the RTL in slim (condensed) form
13882               instead of the default LISP-like representation.
13883
13884           raw Print a raw representation of the tree.  By default, trees are
13885               pretty-printed into a C-like representation.
13886
13887           details
13888               Enable more detailed dumps (not honored by every dump option).
13889               Also include information from the optimization passes.
13890
13891           stats
13892               Enable dumping various statistics about the pass (not honored
13893               by every dump option).
13894
13895           blocks
13896               Enable showing basic block boundaries (disabled in raw dumps).
13897
13898           graph
13899               For each of the other indicated dump files (-fdump-rtl-pass),
13900               dump a representation of the control flow graph suitable for
13901               viewing with GraphViz to file.passid.pass.dot.  Each function
13902               in the file is pretty-printed as a subgraph, so that GraphViz
13903               can render them all in a single plot.
13904
13905               This option currently only works for RTL dumps, and the RTL is
13906               always dumped in slim form.
13907
13908           vops
13909               Enable showing virtual operands for every statement.
13910
13911           lineno
13912               Enable showing line numbers for statements.
13913
13914           uid Enable showing the unique ID ("DECL_UID") for each variable.
13915
13916           verbose
13917               Enable showing the tree dump for each statement.
13918
13919           eh  Enable showing the EH region number holding each statement.
13920
13921           scev
13922               Enable showing scalar evolution analysis details.
13923
13924           optimized
13925               Enable showing optimization information (only available in
13926               certain passes).
13927
13928           missed
13929               Enable showing missed optimization information (only available
13930               in certain passes).
13931
13932           note
13933               Enable other detailed optimization information (only available
13934               in certain passes).
13935
13936           all Turn on all options, except raw, slim, verbose and lineno.
13937
13938           optall
13939               Turn on all optimization options, i.e., optimized, missed, and
13940               note.
13941
13942           To determine what tree dumps are available or find the dump for a
13943           pass of interest follow the steps below.
13944
13945           1.  Invoke GCC with -fdump-passes and in the stderr output look for
13946               a code that corresponds to the pass you are interested in.  For
13947               example, the codes "tree-evrp", "tree-vrp1", and "tree-vrp2"
13948               correspond to the three Value Range Propagation passes.  The
13949               number at the end distinguishes distinct invocations of the
13950               same pass.
13951
13952           2.  To enable the creation of the dump file, append the pass code
13953               to the -fdump- option prefix and invoke GCC with it.  For
13954               example, to enable the dump from the Early Value Range
13955               Propagation pass, invoke GCC with the -fdump-tree-evrp option.
13956               Optionally, you may specify the name of the dump file.  If you
13957               don't specify one, GCC creates as described below.
13958
13959           3.  Find the pass dump in a file whose name is composed of three
13960               components separated by a period: the name of the source file
13961               GCC was invoked to compile, a numeric suffix indicating the
13962               pass number followed by the letter t for tree passes (and the
13963               letter r for RTL passes), and finally the pass code.  For
13964               example, the Early VRP pass dump might be in a file named
13965               myfile.c.038t.evrp in the current working directory.  Note that
13966               the numeric codes are not stable and may change from one
13967               version of GCC to another.
13968
13969       -fopt-info
13970       -fopt-info-options
13971       -fopt-info-options=filename
13972           Controls optimization dumps from various optimization passes. If
13973           the -options form is used, options is a list of - separated option
13974           keywords to select the dump details and optimizations.
13975
13976           The options can be divided into three groups:
13977
13978           1.  options describing what kinds of messages should be emitted,
13979
13980           2.  options describing the verbosity of the dump, and
13981
13982           3.  options describing which optimizations should be included.
13983
13984           The options from each group can be freely mixed as they are non-
13985           overlapping. However, in case of any conflicts, the later options
13986           override the earlier options on the command line.
13987
13988           The following options control which kinds of messages should be
13989           emitted:
13990
13991           optimized
13992               Print information when an optimization is successfully applied.
13993               It is up to a pass to decide which information is relevant. For
13994               example, the vectorizer passes print the source location of
13995               loops which are successfully vectorized.
13996
13997           missed
13998               Print information about missed optimizations. Individual passes
13999               control which information to include in the output.
14000
14001           note
14002               Print verbose information about optimizations, such as certain
14003               transformations, more detailed messages about decisions etc.
14004
14005           all Print detailed optimization information. This includes
14006               optimized, missed, and note.
14007
14008           The following option controls the dump verbosity:
14009
14010           internals
14011               By default, only "high-level" messages are emitted. This option
14012               enables additional, more detailed, messages, which are likely
14013               to only be of interest to GCC developers.
14014
14015           One or more of the following option keywords can be used to
14016           describe a group of optimizations:
14017
14018           ipa Enable dumps from all interprocedural optimizations.
14019
14020           loop
14021               Enable dumps from all loop optimizations.
14022
14023           inline
14024               Enable dumps from all inlining optimizations.
14025
14026           omp Enable dumps from all OMP (Offloading and Multi Processing)
14027               optimizations.
14028
14029           vec Enable dumps from all vectorization optimizations.
14030
14031           optall
14032               Enable dumps from all optimizations. This is a superset of the
14033               optimization groups listed above.
14034
14035           If options is omitted, it defaults to optimized-optall, which means
14036           to dump messages about successful optimizations from all the
14037           passes, omitting messages that are treated as "internals".
14038
14039           If the filename is provided, then the dumps from all the applicable
14040           optimizations are concatenated into the filename.  Otherwise the
14041           dump is output onto stderr. Though multiple -fopt-info options are
14042           accepted, only one of them can include a filename. If other
14043           filenames are provided then all but the first such option are
14044           ignored.
14045
14046           Note that the output filename is overwritten in case of multiple
14047           translation units. If a combined output from multiple translation
14048           units is desired, stderr should be used instead.
14049
14050           In the following example, the optimization info is output to
14051           stderr:
14052
14053                   gcc -O3 -fopt-info
14054
14055           This example:
14056
14057                   gcc -O3 -fopt-info-missed=missed.all
14058
14059           outputs missed optimization report from all the passes into
14060           missed.all, and this one:
14061
14062                   gcc -O2 -ftree-vectorize -fopt-info-vec-missed
14063
14064           prints information about missed optimization opportunities from
14065           vectorization passes on stderr.  Note that -fopt-info-vec-missed is
14066           equivalent to -fopt-info-missed-vec.  The order of the optimization
14067           group names and message types listed after -fopt-info does not
14068           matter.
14069
14070           As another example,
14071
14072                   gcc -O3 -fopt-info-inline-optimized-missed=inline.txt
14073
14074           outputs information about missed optimizations as well as optimized
14075           locations from all the inlining passes into inline.txt.
14076
14077           Finally, consider:
14078
14079                   gcc -fopt-info-vec-missed=vec.miss -fopt-info-loop-optimized=loop.opt
14080
14081           Here the two output filenames vec.miss and loop.opt are in conflict
14082           since only one output file is allowed. In this case, only the first
14083           option takes effect and the subsequent options are ignored. Thus
14084           only vec.miss is produced which contains dumps from the vectorizer
14085           about missed opportunities.
14086
14087       -fsave-optimization-record
14088           Write a SRCFILE.opt-record.json.gz file detailing what
14089           optimizations were performed, for those optimizations that support
14090           -fopt-info.
14091
14092           This option is experimental and the format of the data within the
14093           compressed JSON file is subject to change.
14094
14095           It is roughly equivalent to a machine-readable version of
14096           -fopt-info-all, as a collection of messages with source file, line
14097           number and column number, with the following additional data for
14098           each message:
14099
14100           *   the execution count of the code being optimized, along with
14101               metadata about whether this was from actual profile data, or
14102               just an estimate, allowing consumers to prioritize messages by
14103               code hotness,
14104
14105           *   the function name of the code being optimized, where
14106               applicable,
14107
14108           *   the "inlining chain" for the code being optimized, so that when
14109               a function is inlined into several different places (which
14110               might themselves be inlined), the reader can distinguish
14111               between the copies,
14112
14113           *   objects identifying those parts of the message that refer to
14114               expressions, statements or symbol-table nodes, which of these
14115               categories they are, and, when available, their source code
14116               location,
14117
14118           *   the GCC pass that emitted the message, and
14119
14120           *   the location in GCC's own code from which the message was
14121               emitted
14122
14123           Additionally, some messages are logically nested within other
14124           messages, reflecting implementation details of the optimization
14125           passes.
14126
14127       -fsched-verbose=n
14128           On targets that use instruction scheduling, this option controls
14129           the amount of debugging output the scheduler prints to the dump
14130           files.
14131
14132           For n greater than zero, -fsched-verbose outputs the same
14133           information as -fdump-rtl-sched1 and -fdump-rtl-sched2.  For n
14134           greater than one, it also output basic block probabilities,
14135           detailed ready list information and unit/insn info.  For n greater
14136           than two, it includes RTL at abort point, control-flow and regions
14137           info.  And for n over four, -fsched-verbose also includes
14138           dependence info.
14139
14140       -fenable-kind-pass
14141       -fdisable-kind-pass=range-list
14142           This is a set of options that are used to explicitly disable/enable
14143           optimization passes.  These options are intended for use for
14144           debugging GCC.  Compiler users should use regular options for
14145           enabling/disabling passes instead.
14146
14147           -fdisable-ipa-pass
14148               Disable IPA pass pass. pass is the pass name.  If the same pass
14149               is statically invoked in the compiler multiple times, the pass
14150               name should be appended with a sequential number starting from
14151               1.
14152
14153           -fdisable-rtl-pass
14154           -fdisable-rtl-pass=range-list
14155               Disable RTL pass pass.  pass is the pass name.  If the same
14156               pass is statically invoked in the compiler multiple times, the
14157               pass name should be appended with a sequential number starting
14158               from 1.  range-list is a comma-separated list of function
14159               ranges or assembler names.  Each range is a number pair
14160               separated by a colon.  The range is inclusive in both ends.  If
14161               the range is trivial, the number pair can be simplified as a
14162               single number.  If the function's call graph node's uid falls
14163               within one of the specified ranges, the pass is disabled for
14164               that function.  The uid is shown in the function header of a
14165               dump file, and the pass names can be dumped by using option
14166               -fdump-passes.
14167
14168           -fdisable-tree-pass
14169           -fdisable-tree-pass=range-list
14170               Disable tree pass pass.  See -fdisable-rtl for the description
14171               of option arguments.
14172
14173           -fenable-ipa-pass
14174               Enable IPA pass pass.  pass is the pass name.  If the same pass
14175               is statically invoked in the compiler multiple times, the pass
14176               name should be appended with a sequential number starting from
14177               1.
14178
14179           -fenable-rtl-pass
14180           -fenable-rtl-pass=range-list
14181               Enable RTL pass pass.  See -fdisable-rtl for option argument
14182               description and examples.
14183
14184           -fenable-tree-pass
14185           -fenable-tree-pass=range-list
14186               Enable tree pass pass.  See -fdisable-rtl for the description
14187               of option arguments.
14188
14189           Here are some examples showing uses of these options.
14190
14191                   # disable ccp1 for all functions
14192                      -fdisable-tree-ccp1
14193                   # disable complete unroll for function whose cgraph node uid is 1
14194                      -fenable-tree-cunroll=1
14195                   # disable gcse2 for functions at the following ranges [1,1],
14196                   # [300,400], and [400,1000]
14197                   # disable gcse2 for functions foo and foo2
14198                      -fdisable-rtl-gcse2=foo,foo2
14199                   # disable early inlining
14200                      -fdisable-tree-einline
14201                   # disable ipa inlining
14202                      -fdisable-ipa-inline
14203                   # enable tree full unroll
14204                      -fenable-tree-unroll
14205
14206       -fchecking
14207       -fchecking=n
14208           Enable internal consistency checking.  The default depends on the
14209           compiler configuration.  -fchecking=2 enables further internal
14210           consistency checking that might affect code generation.
14211
14212       -frandom-seed=string
14213           This option provides a seed that GCC uses in place of random
14214           numbers in generating certain symbol names that have to be
14215           different in every compiled file.  It is also used to place unique
14216           stamps in coverage data files and the object files that produce
14217           them.  You can use the -frandom-seed option to produce reproducibly
14218           identical object files.
14219
14220           The string can either be a number (decimal, octal or hex) or an
14221           arbitrary string (in which case it's converted to a number by
14222           computing CRC32).
14223
14224           The string should be different for every file you compile.
14225
14226       -save-temps
14227       -save-temps=cwd
14228           Store the usual "temporary" intermediate files permanently; place
14229           them in the current directory and name them based on the source
14230           file.  Thus, compiling foo.c with -c -save-temps produces files
14231           foo.i and foo.s, as well as foo.o.  This creates a preprocessed
14232           foo.i output file even though the compiler now normally uses an
14233           integrated preprocessor.
14234
14235           When used in combination with the -x command-line option,
14236           -save-temps is sensible enough to avoid over writing an input
14237           source file with the same extension as an intermediate file.  The
14238           corresponding intermediate file may be obtained by renaming the
14239           source file before using -save-temps.
14240
14241           If you invoke GCC in parallel, compiling several different source
14242           files that share a common base name in different subdirectories or
14243           the same source file compiled for multiple output destinations, it
14244           is likely that the different parallel compilers will interfere with
14245           each other, and overwrite the temporary files.  For instance:
14246
14247                   gcc -save-temps -o outdir1/foo.o indir1/foo.c&
14248                   gcc -save-temps -o outdir2/foo.o indir2/foo.c&
14249
14250           may result in foo.i and foo.o being written to simultaneously by
14251           both compilers.
14252
14253       -save-temps=obj
14254           Store the usual "temporary" intermediate files permanently.  If the
14255           -o option is used, the temporary files are based on the object
14256           file.  If the -o option is not used, the -save-temps=obj switch
14257           behaves like -save-temps.
14258
14259           For example:
14260
14261                   gcc -save-temps=obj -c foo.c
14262                   gcc -save-temps=obj -c bar.c -o dir/xbar.o
14263                   gcc -save-temps=obj foobar.c -o dir2/yfoobar
14264
14265           creates foo.i, foo.s, dir/xbar.i, dir/xbar.s, dir2/yfoobar.i,
14266           dir2/yfoobar.s, and dir2/yfoobar.o.
14267
14268       -time[=file]
14269           Report the CPU time taken by each subprocess in the compilation
14270           sequence.  For C source files, this is the compiler proper and
14271           assembler (plus the linker if linking is done).
14272
14273           Without the specification of an output file, the output looks like
14274           this:
14275
14276                   # cc1 0.12 0.01
14277                   # as 0.00 0.01
14278
14279           The first number on each line is the "user time", that is time
14280           spent executing the program itself.  The second number is "system
14281           time", time spent executing operating system routines on behalf of
14282           the program.  Both numbers are in seconds.
14283
14284           With the specification of an output file, the output is appended to
14285           the named file, and it looks like this:
14286
14287                   0.12 0.01 cc1 <options>
14288                   0.00 0.01 as <options>
14289
14290           The "user time" and the "system time" are moved before the program
14291           name, and the options passed to the program are displayed, so that
14292           one can later tell what file was being compiled, and with which
14293           options.
14294
14295       -fdump-final-insns[=file]
14296           Dump the final internal representation (RTL) to file.  If the
14297           optional argument is omitted (or if file is "."), the name of the
14298           dump file is determined by appending ".gkd" to the compilation
14299           output file name.
14300
14301       -fcompare-debug[=opts]
14302           If no error occurs during compilation, run the compiler a second
14303           time, adding opts and -fcompare-debug-second to the arguments
14304           passed to the second compilation.  Dump the final internal
14305           representation in both compilations, and print an error if they
14306           differ.
14307
14308           If the equal sign is omitted, the default -gtoggle is used.
14309
14310           The environment variable GCC_COMPARE_DEBUG, if defined, non-empty
14311           and nonzero, implicitly enables -fcompare-debug.  If
14312           GCC_COMPARE_DEBUG is defined to a string starting with a dash, then
14313           it is used for opts, otherwise the default -gtoggle is used.
14314
14315           -fcompare-debug=, with the equal sign but without opts, is
14316           equivalent to -fno-compare-debug, which disables the dumping of the
14317           final representation and the second compilation, preventing even
14318           GCC_COMPARE_DEBUG from taking effect.
14319
14320           To verify full coverage during -fcompare-debug testing, set
14321           GCC_COMPARE_DEBUG to say -fcompare-debug-not-overridden, which GCC
14322           rejects as an invalid option in any actual compilation (rather than
14323           preprocessing, assembly or linking).  To get just a warning,
14324           setting GCC_COMPARE_DEBUG to -w%n-fcompare-debug not overridden
14325           will do.
14326
14327       -fcompare-debug-second
14328           This option is implicitly passed to the compiler for the second
14329           compilation requested by -fcompare-debug, along with options to
14330           silence warnings, and omitting other options that would cause the
14331           compiler to produce output to files or to standard output as a side
14332           effect.  Dump files and preserved temporary files are renamed so as
14333           to contain the ".gk" additional extension during the second
14334           compilation, to avoid overwriting those generated by the first.
14335
14336           When this option is passed to the compiler driver, it causes the
14337           first compilation to be skipped, which makes it useful for little
14338           other than debugging the compiler proper.
14339
14340       -gtoggle
14341           Turn off generation of debug info, if leaving out this option
14342           generates it, or turn it on at level 2 otherwise.  The position of
14343           this argument in the command line does not matter; it takes effect
14344           after all other options are processed, and it does so only once, no
14345           matter how many times it is given.  This is mainly intended to be
14346           used with -fcompare-debug.
14347
14348       -fvar-tracking-assignments-toggle
14349           Toggle -fvar-tracking-assignments, in the same way that -gtoggle
14350           toggles -g.
14351
14352       -Q  Makes the compiler print out each function name as it is compiled,
14353           and print some statistics about each pass when it finishes.
14354
14355       -ftime-report
14356           Makes the compiler print some statistics about the time consumed by
14357           each pass when it finishes.
14358
14359       -ftime-report-details
14360           Record the time consumed by infrastructure parts separately for
14361           each pass.
14362
14363       -fira-verbose=n
14364           Control the verbosity of the dump file for the integrated register
14365           allocator.  The default value is 5.  If the value n is greater or
14366           equal to 10, the dump output is sent to stderr using the same
14367           format as n minus 10.
14368
14369       -flto-report
14370           Prints a report with internal details on the workings of the link-
14371           time optimizer.  The contents of this report vary from version to
14372           version.  It is meant to be useful to GCC developers when
14373           processing object files in LTO mode (via -flto).
14374
14375           Disabled by default.
14376
14377       -flto-report-wpa
14378           Like -flto-report, but only print for the WPA phase of link-time
14379           optimization.
14380
14381       -fmem-report
14382           Makes the compiler print some statistics about permanent memory
14383           allocation when it finishes.
14384
14385       -fmem-report-wpa
14386           Makes the compiler print some statistics about permanent memory
14387           allocation for the WPA phase only.
14388
14389       -fpre-ipa-mem-report
14390       -fpost-ipa-mem-report
14391           Makes the compiler print some statistics about permanent memory
14392           allocation before or after interprocedural optimization.
14393
14394       -fprofile-report
14395           Makes the compiler print some statistics about consistency of the
14396           (estimated) profile and effect of individual passes.
14397
14398       -fstack-usage
14399           Makes the compiler output stack usage information for the program,
14400           on a per-function basis.  The filename for the dump is made by
14401           appending .su to the auxname.  auxname is generated from the name
14402           of the output file, if explicitly specified and it is not an
14403           executable, otherwise it is the basename of the source file.  An
14404           entry is made up of three fields:
14405
14406           *   The name of the function.
14407
14408           *   A number of bytes.
14409
14410           *   One or more qualifiers: "static", "dynamic", "bounded".
14411
14412           The qualifier "static" means that the function manipulates the
14413           stack statically: a fixed number of bytes are allocated for the
14414           frame on function entry and released on function exit; no stack
14415           adjustments are otherwise made in the function.  The second field
14416           is this fixed number of bytes.
14417
14418           The qualifier "dynamic" means that the function manipulates the
14419           stack dynamically: in addition to the static allocation described
14420           above, stack adjustments are made in the body of the function, for
14421           example to push/pop arguments around function calls.  If the
14422           qualifier "bounded" is also present, the amount of these
14423           adjustments is bounded at compile time and the second field is an
14424           upper bound of the total amount of stack used by the function.  If
14425           it is not present, the amount of these adjustments is not bounded
14426           at compile time and the second field only represents the bounded
14427           part.
14428
14429       -fstats
14430           Emit statistics about front-end processing at the end of the
14431           compilation.  This option is supported only by the C++ front end,
14432           and the information is generally only useful to the G++ development
14433           team.
14434
14435       -fdbg-cnt-list
14436           Print the name and the counter upper bound for all debug counters.
14437
14438       -fdbg-cnt=counter-value-list
14439           Set the internal debug counter lower and upper bound.  counter-
14440           value-list is a comma-separated list of
14441           name:lower_bound1-upper_bound1 [:lower_bound2-upper_bound2...]
14442           tuples which sets the name of the counter and list of closed
14443           intervals.  The lower_bound is optional and is zero initialized if
14444           not set.  For example, with -fdbg-cnt=dce:2-4:10-11,tail_call:10,
14445           "dbg_cnt(dce)" returns true only for second, third, fourth, tenth
14446           and eleventh invocation.  For "dbg_cnt(tail_call)" true is returned
14447           for first 10 invocations.
14448
14449       -print-file-name=library
14450           Print the full absolute name of the library file library that would
14451           be used when linking---and don't do anything else.  With this
14452           option, GCC does not compile or link anything; it just prints the
14453           file name.
14454
14455       -print-multi-directory
14456           Print the directory name corresponding to the multilib selected by
14457           any other switches present in the command line.  This directory is
14458           supposed to exist in GCC_EXEC_PREFIX.
14459
14460       -print-multi-lib
14461           Print the mapping from multilib directory names to compiler
14462           switches that enable them.  The directory name is separated from
14463           the switches by ;, and each switch starts with an @ instead of the
14464           -, without spaces between multiple switches.  This is supposed to
14465           ease shell processing.
14466
14467       -print-multi-os-directory
14468           Print the path to OS libraries for the selected multilib, relative
14469           to some lib subdirectory.  If OS libraries are present in the lib
14470           subdirectory and no multilibs are used, this is usually just ., if
14471           OS libraries are present in libsuffix sibling directories this
14472           prints e.g. ../lib64, ../lib or ../lib32, or if OS libraries are
14473           present in lib/subdir subdirectories it prints e.g. amd64, sparcv9
14474           or ev6.
14475
14476       -print-multiarch
14477           Print the path to OS libraries for the selected multiarch, relative
14478           to some lib subdirectory.
14479
14480       -print-prog-name=program
14481           Like -print-file-name, but searches for a program such as cpp.
14482
14483       -print-libgcc-file-name
14484           Same as -print-file-name=libgcc.a.
14485
14486           This is useful when you use -nostdlib or -nodefaultlibs but you do
14487           want to link with libgcc.a.  You can do:
14488
14489                   gcc -nostdlib <files>... `gcc -print-libgcc-file-name`
14490
14491       -print-search-dirs
14492           Print the name of the configured installation directory and a list
14493           of program and library directories gcc searches---and don't do
14494           anything else.
14495
14496           This is useful when gcc prints the error message installation
14497           problem, cannot exec cpp0: No such file or directory.  To resolve
14498           this you either need to put cpp0 and the other compiler components
14499           where gcc expects to find them, or you can set the environment
14500           variable GCC_EXEC_PREFIX to the directory where you installed them.
14501           Don't forget the trailing /.
14502
14503       -print-sysroot
14504           Print the target sysroot directory that is used during compilation.
14505           This is the target sysroot specified either at configure time or
14506           using the --sysroot option, possibly with an extra suffix that
14507           depends on compilation options.  If no target sysroot is specified,
14508           the option prints nothing.
14509
14510       -print-sysroot-headers-suffix
14511           Print the suffix added to the target sysroot when searching for
14512           headers, or give an error if the compiler is not configured with
14513           such a suffix---and don't do anything else.
14514
14515       -dumpmachine
14516           Print the compiler's target machine (for example,
14517           i686-pc-linux-gnu)---and don't do anything else.
14518
14519       -dumpversion
14520           Print the compiler version (for example, 3.0, 6.3.0 or 7)---and
14521           don't do anything else.  This is the compiler version used in
14522           filesystem paths and specs. Depending on how the compiler has been
14523           configured it can be just a single number (major version), two
14524           numbers separated by a dot (major and minor version) or three
14525           numbers separated by dots (major, minor and patchlevel version).
14526
14527       -dumpfullversion
14528           Print the full compiler version---and don't do anything else. The
14529           output is always three numbers separated by dots, major, minor and
14530           patchlevel version.
14531
14532       -dumpspecs
14533           Print the compiler's built-in specs---and don't do anything else.
14534           (This is used when GCC itself is being built.)
14535
14536   Machine-Dependent Options
14537       Each target machine supported by GCC can have its own options---for
14538       example, to allow you to compile for a particular processor variant or
14539       ABI, or to control optimizations specific to that machine.  By
14540       convention, the names of machine-specific options start with -m.
14541
14542       Some configurations of the compiler also support additional target-
14543       specific options, usually for compatibility with other compilers on the
14544       same platform.
14545
14546       AArch64 Options
14547
14548       These options are defined for AArch64 implementations:
14549
14550       -mabi=name
14551           Generate code for the specified data model.  Permissible values are
14552           ilp32 for SysV-like data model where int, long int and pointers are
14553           32 bits, and lp64 for SysV-like data model where int is 32 bits,
14554           but long int and pointers are 64 bits.
14555
14556           The default depends on the specific target configuration.  Note
14557           that the LP64 and ILP32 ABIs are not link-compatible; you must
14558           compile your entire program with the same ABI, and link with a
14559           compatible set of libraries.
14560
14561       -mbig-endian
14562           Generate big-endian code.  This is the default when GCC is
14563           configured for an aarch64_be-*-* target.
14564
14565       -mgeneral-regs-only
14566           Generate code which uses only the general-purpose registers.  This
14567           will prevent the compiler from using floating-point and Advanced
14568           SIMD registers but will not impose any restrictions on the
14569           assembler.
14570
14571       -mlittle-endian
14572           Generate little-endian code.  This is the default when GCC is
14573           configured for an aarch64-*-* but not an aarch64_be-*-* target.
14574
14575       -mcmodel=tiny
14576           Generate code for the tiny code model.  The program and its
14577           statically defined symbols must be within 1MB of each other.
14578           Programs can be statically or dynamically linked.
14579
14580       -mcmodel=small
14581           Generate code for the small code model.  The program and its
14582           statically defined symbols must be within 4GB of each other.
14583           Programs can be statically or dynamically linked.  This is the
14584           default code model.
14585
14586       -mcmodel=large
14587           Generate code for the large code model.  This makes no assumptions
14588           about addresses and sizes of sections.  Programs can be statically
14589           linked only.  The -mcmodel=large option is incompatible with
14590           -mabi=ilp32, -fpic and -fPIC.
14591
14592       -mstrict-align
14593       -mno-strict-align
14594           Avoid or allow generating memory accesses that may not be aligned
14595           on a natural object boundary as described in the architecture
14596           specification.
14597
14598       -momit-leaf-frame-pointer
14599       -mno-omit-leaf-frame-pointer
14600           Omit or keep the frame pointer in leaf functions.  The former
14601           behavior is the default.
14602
14603       -mstack-protector-guard=guard
14604       -mstack-protector-guard-reg=reg
14605       -mstack-protector-guard-offset=offset
14606           Generate stack protection code using canary at guard.  Supported
14607           locations are global for a global canary or sysreg for a canary in
14608           an appropriate system register.
14609
14610           With the latter choice the options -mstack-protector-guard-reg=reg
14611           and -mstack-protector-guard-offset=offset furthermore specify which
14612           system register to use as base register for reading the canary, and
14613           from what offset from that base register. There is no default
14614           register or offset as this is entirely for use within the Linux
14615           kernel.
14616
14617       -mstack-protector-guard=guard
14618       -mstack-protector-guard-reg=reg
14619       -mstack-protector-guard-offset=offset
14620           Generate stack protection code using canary at guard.  Supported
14621           locations are global for a global canary or sysreg for a canary in
14622           an appropriate system register.
14623
14624           With the latter choice the options -mstack-protector-guard-reg=reg
14625           and -mstack-protector-guard-offset=offset furthermore specify which
14626           system register to use as base register for reading the canary, and
14627           from what offset from that base register. There is no default
14628           register or offset as this is entirely for use within the Linux
14629           kernel.
14630
14631       -mtls-dialect=desc
14632           Use TLS descriptors as the thread-local storage mechanism for
14633           dynamic accesses of TLS variables.  This is the default.
14634
14635       -mtls-dialect=traditional
14636           Use traditional TLS as the thread-local storage mechanism for
14637           dynamic accesses of TLS variables.
14638
14639       -mtls-size=size
14640           Specify bit size of immediate TLS offsets.  Valid values are 12,
14641           24, 32, 48.  This option requires binutils 2.26 or newer.
14642
14643       -mfix-cortex-a53-835769
14644       -mno-fix-cortex-a53-835769
14645           Enable or disable the workaround for the ARM Cortex-A53 erratum
14646           number 835769.  This involves inserting a NOP instruction between
14647           memory instructions and 64-bit integer multiply-accumulate
14648           instructions.
14649
14650       -mfix-cortex-a53-843419
14651       -mno-fix-cortex-a53-843419
14652           Enable or disable the workaround for the ARM Cortex-A53 erratum
14653           number 843419.  This erratum workaround is made at link time and
14654           this will only pass the corresponding flag to the linker.
14655
14656       -mlow-precision-recip-sqrt
14657       -mno-low-precision-recip-sqrt
14658           Enable or disable the reciprocal square root approximation.  This
14659           option only has an effect if -ffast-math or
14660           -funsafe-math-optimizations is used as well.  Enabling this reduces
14661           precision of reciprocal square root results to about 16 bits for
14662           single precision and to 32 bits for double precision.
14663
14664       -mlow-precision-sqrt
14665       -mno-low-precision-sqrt
14666           Enable or disable the square root approximation.  This option only
14667           has an effect if -ffast-math or -funsafe-math-optimizations is used
14668           as well.  Enabling this reduces precision of square root results to
14669           about 16 bits for single precision and to 32 bits for double
14670           precision.  If enabled, it implies -mlow-precision-recip-sqrt.
14671
14672       -mlow-precision-div
14673       -mno-low-precision-div
14674           Enable or disable the division approximation.  This option only has
14675           an effect if -ffast-math or -funsafe-math-optimizations is used as
14676           well.  Enabling this reduces precision of division results to about
14677           16 bits for single precision and to 32 bits for double precision.
14678
14679       -mtrack-speculation
14680       -mno-track-speculation
14681           Enable or disable generation of additional code to track
14682           speculative execution through conditional branches.  The tracking
14683           state can then be used by the compiler when expanding calls to
14684           "__builtin_speculation_safe_copy" to permit a more efficient code
14685           sequence to be generated.
14686
14687       -moutline-atomics
14688       -mno-outline-atomics
14689           Enable or disable calls to out-of-line helpers to implement atomic
14690           operations.  These helpers will, at runtime, determine if the LSE
14691           instructions from ARMv8.1-A can be used; if not, they will use the
14692           load/store-exclusive instructions that are present in the base
14693           ARMv8.0 ISA.
14694
14695           This option is only applicable when compiling for the base ARMv8.0
14696           instruction set.  If using a later revision, e.g. -march=armv8.1-a
14697           or -march=armv8-a+lse, the ARMv8.1-Atomics instructions will be
14698           used directly.  The same applies when using -mcpu= when the
14699           selected cpu supports the lse feature.  This option is on by
14700           default.
14701
14702       -march=name
14703           Specify the name of the target architecture and, optionally, one or
14704           more feature modifiers.  This option has the form
14705           -march=arch{+[no]feature}*.
14706
14707           The table below summarizes the permissible values for arch and the
14708           features that they enable by default:
14709
14710           arch value : Architecture : Includes by default
14711           armv8-a : Armv8-A : +fp, +simd
14712           armv8.1-a : Armv8.1-A : armv8-a, +crc, +lse, +rdma
14713           armv8.2-a : Armv8.2-A : armv8.1-a
14714           armv8.3-a : Armv8.3-A : armv8.2-a
14715           armv8.4-a : Armv8.4-A : armv8.3-a, +fp16fml, +dotprod
14716           armv8.5-a : Armv8.5-A : armv8.4-a, +sb, +ssbs, +predres
14717           armv8.6-a : Armv8.6-A : armv8.5-a, +bf16, +i8mm
14718
14719           The value native is available on native AArch64 GNU/Linux and
14720           causes the compiler to pick the architecture of the host system.
14721           This option has no effect if the compiler is unable to recognize
14722           the architecture of the host system,
14723
14724           The permissible values for feature are listed in the sub-section on
14725           aarch64-feature-modifiers,,-march and -mcpu Feature Modifiers.
14726           Where conflicting feature modifiers are specified, the right-most
14727           feature is used.
14728
14729           GCC uses name to determine what kind of instructions it can emit
14730           when generating assembly code.  If -march is specified without
14731           either of -mtune or -mcpu also being specified, the code is tuned
14732           to perform well across a range of target processors implementing
14733           the target architecture.
14734
14735       -mtune=name
14736           Specify the name of the target processor for which GCC should tune
14737           the performance of the code.  Permissible values for this option
14738           are: generic, cortex-a35, cortex-a53, cortex-a55, cortex-a57,
14739           cortex-a72, cortex-a73, cortex-a75, cortex-a76, cortex-a76ae,
14740           cortex-a77, cortex-a65, cortex-a65ae, cortex-a34, ares, exynos-m1,
14741           emag, falkor, neoverse-e1,neoverse-n1,qdf24xx, saphira, phecda,
14742           xgene1, vulcan, octeontx, octeontx81,  octeontx83, octeontx2,
14743           octeontx2t98, octeontx2t96 octeontx2t93, octeontx2f95,
14744           octeontx2f95n, octeontx2f95mm thunderx, thunderxt88, thunderxt88p1,
14745           thunderxt81, tsv110, thunderxt83, thunderx2t99, thunderx3t110,
14746           zeus, cortex-a57.cortex-a53, cortex-a72.cortex-a53,
14747           cortex-a73.cortex-a35, cortex-a73.cortex-a53,
14748           cortex-a75.cortex-a55, cortex-a76.cortex-a55 native.
14749
14750           The values cortex-a57.cortex-a53, cortex-a72.cortex-a53,
14751           cortex-a73.cortex-a35, cortex-a73.cortex-a53,
14752           cortex-a75.cortex-a55, cortex-a76.cortex-a55 specify that GCC
14753           should tune for a big.LITTLE system.
14754
14755           Additionally on native AArch64 GNU/Linux systems the value native
14756           tunes performance to the host system.  This option has no effect if
14757           the compiler is unable to recognize the processor of the host
14758           system.
14759
14760           Where none of -mtune=, -mcpu= or -march= are specified, the code is
14761           tuned to perform well across a range of target processors.
14762
14763           This option cannot be suffixed by feature modifiers.
14764
14765       -mcpu=name
14766           Specify the name of the target processor, optionally suffixed by
14767           one or more feature modifiers.  This option has the form
14768           -mcpu=cpu{+[no]feature}*, where the permissible values for cpu are
14769           the same as those available for -mtune.  The permissible values for
14770           feature are documented in the sub-section on
14771           aarch64-feature-modifiers,,-march and -mcpu Feature Modifiers.
14772           Where conflicting feature modifiers are specified, the right-most
14773           feature is used.
14774
14775           GCC uses name to determine what kind of instructions it can emit
14776           when generating assembly code (as if by -march) and to determine
14777           the target processor for which to tune for performance (as if by
14778           -mtune).  Where this option is used in conjunction with -march or
14779           -mtune, those options take precedence over the appropriate part of
14780           this option.
14781
14782       -moverride=string
14783           Override tuning decisions made by the back-end in response to a
14784           -mtune= switch.  The syntax, semantics, and accepted values for
14785           string in this option are not guaranteed to be consistent across
14786           releases.
14787
14788           This option is only intended to be useful when developing GCC.
14789
14790       -mverbose-cost-dump
14791           Enable verbose cost model dumping in the debug dump files.  This
14792           option is provided for use in debugging the compiler.
14793
14794       -mpc-relative-literal-loads
14795       -mno-pc-relative-literal-loads
14796           Enable or disable PC-relative literal loads.  With this option
14797           literal pools are accessed using a single instruction and emitted
14798           after each function.  This limits the maximum size of functions to
14799           1MB.  This is enabled by default for -mcmodel=tiny.
14800
14801       -msign-return-address=scope
14802           Select the function scope on which return address signing will be
14803           applied.  Permissible values are none, which disables return
14804           address signing, non-leaf, which enables pointer signing for
14805           functions which are not leaf functions, and all, which enables
14806           pointer signing for all functions.  The default value is none. This
14807           option has been deprecated by -mbranch-protection.
14808
14809       -mbranch-protection=none|standard|pac-ret[+leaf+b-key]|bti
14810           Select the branch protection features to use.  none is the default
14811           and turns off all types of branch protection.  standard turns on
14812           all types of branch protection features.  If a feature has
14813           additional tuning options, then standard sets it to its standard
14814           level.  pac-ret[+leaf] turns on return address signing to its
14815           standard level: signing functions that save the return address to
14816           memory (non-leaf functions will practically always do this) using
14817           the a-key.  The optional argument leaf can be used to extend the
14818           signing to include leaf functions.  The optional argument b-key can
14819           be used to sign the functions with the B-key instead of the A-key.
14820           bti turns on branch target identification mechanism.
14821
14822       -msve-vector-bits=bits
14823           Specify the number of bits in an SVE vector register.  This option
14824           only has an effect when SVE is enabled.
14825
14826           GCC supports two forms of SVE code generation: "vector-length
14827           agnostic" output that works with any size of vector register and
14828           "vector-length specific" output that allows GCC to make assumptions
14829           about the vector length when it is useful for optimization reasons.
14830           The possible values of bits are: scalable, 128, 256, 512, 1024 and
14831           2048.  Specifying scalable selects vector-length agnostic output.
14832           At present -msve-vector-bits=128 also generates vector-length
14833           agnostic output for big-endian targets.  All other values generate
14834           vector-length specific code.  The behavior of these values may
14835           change in future releases and no value except scalable should be
14836           relied on for producing code that is portable across different
14837           hardware SVE vector lengths.
14838
14839           The default is -msve-vector-bits=scalable, which produces vector-
14840           length agnostic code.
14841
14842       -march and -mcpu Feature Modifiers
14843
14844       Feature modifiers used with -march and -mcpu can be any of the
14845       following and their inverses nofeature:
14846
14847       crc Enable CRC extension.  This is on by default for -march=armv8.1-a.
14848
14849       crypto
14850           Enable Crypto extension.  This also enables Advanced SIMD and
14851           floating-point instructions.
14852
14853       fp  Enable floating-point instructions.  This is on by default for all
14854           possible values for options -march and -mcpu.
14855
14856       simd
14857           Enable Advanced SIMD instructions.  This also enables floating-
14858           point instructions.  This is on by default for all possible values
14859           for options -march and -mcpu.
14860
14861       sve Enable Scalable Vector Extension instructions.  This also enables
14862           Advanced SIMD and floating-point instructions.
14863
14864       lse Enable Large System Extension instructions.  This is on by default
14865           for -march=armv8.1-a.
14866
14867       rdma
14868           Enable Round Double Multiply Accumulate instructions.  This is on
14869           by default for -march=armv8.1-a.
14870
14871       fp16
14872           Enable FP16 extension.  This also enables floating-point
14873           instructions.
14874
14875       fp16fml
14876           Enable FP16 fmla extension.  This also enables FP16 extensions and
14877           floating-point instructions. This option is enabled by default for
14878           -march=armv8.4-a. Use of this option with architectures prior to
14879           Armv8.2-A is not supported.
14880
14881       rcpc
14882           Enable the RcPc extension.  This does not change code generation
14883           from GCC, but is passed on to the assembler, enabling inline asm
14884           statements to use instructions from the RcPc extension.
14885
14886       dotprod
14887           Enable the Dot Product extension.  This also enables Advanced SIMD
14888           instructions.
14889
14890       aes Enable the Armv8-a aes and pmull crypto extension.  This also
14891           enables Advanced SIMD instructions.
14892
14893       sha2
14894           Enable the Armv8-a sha2 crypto extension.  This also enables
14895           Advanced SIMD instructions.
14896
14897       sha3
14898           Enable the sha512 and sha3 crypto extension.  This also enables
14899           Advanced SIMD instructions. Use of this option with architectures
14900           prior to Armv8.2-A is not supported.
14901
14902       sm4 Enable the sm3 and sm4 crypto extension.  This also enables
14903           Advanced SIMD instructions.  Use of this option with architectures
14904           prior to Armv8.2-A is not supported.
14905
14906       profile
14907           Enable the Statistical Profiling extension.  This option is only to
14908           enable the extension at the assembler level and does not affect
14909           code generation.
14910
14911       rng Enable the Armv8.5-a Random Number instructions.  This option is
14912           only to enable the extension at the assembler level and does not
14913           affect code generation.
14914
14915       memtag
14916           Enable the Armv8.5-a Memory Tagging Extensions.  Use of this option
14917           with architectures prior to Armv8.5-A is not supported.
14918
14919       sb  Enable the Armv8-a Speculation Barrier instruction.  This option is
14920           only to enable the extension at the assembler level and does not
14921           affect code generation.  This option is enabled by default for
14922           -march=armv8.5-a.
14923
14924       ssbs
14925           Enable the Armv8-a Speculative Store Bypass Safe instruction.  This
14926           option is only to enable the extension at the assembler level and
14927           does not affect code generation.  This option is enabled by default
14928           for -march=armv8.5-a.
14929
14930       predres
14931           Enable the Armv8-a Execution and Data Prediction Restriction
14932           instructions.  This option is only to enable the extension at the
14933           assembler level and does not affect code generation.  This option
14934           is enabled by default for -march=armv8.5-a.
14935
14936       sve2
14937           Enable the Armv8-a Scalable Vector Extension 2.  This also enables
14938           SVE instructions.
14939
14940       sve2-bitperm
14941           Enable SVE2 bitperm instructions.  This also enables SVE2
14942           instructions.
14943
14944       sve2-sm4
14945           Enable SVE2 sm4 instructions.  This also enables SVE2 instructions.
14946
14947       sve2-aes
14948           Enable SVE2 aes instructions.  This also enables SVE2 instructions.
14949
14950       sve2-sha3
14951           Enable SVE2 sha3 instructions.  This also enables SVE2
14952           instructions.
14953
14954       tme Enable the Transactional Memory Extension.
14955
14956       i8mm
14957           Enable 8-bit Integer Matrix Multiply instructions.  This also
14958           enables Advanced SIMD and floating-point instructions.  This option
14959           is enabled by default for -march=armv8.6-a.  Use of this option
14960           with architectures prior to Armv8.2-A is not supported.
14961
14962       f32mm
14963           Enable 32-bit Floating point Matrix Multiply instructions.  This
14964           also enables SVE instructions.  Use of this option with
14965           architectures prior to Armv8.2-A is not supported.
14966
14967       f64mm
14968           Enable 64-bit Floating point Matrix Multiply instructions.  This
14969           also enables SVE instructions.  Use of this option with
14970           architectures prior to Armv8.2-A is not supported.
14971
14972       bf16
14973           Enable brain half-precision floating-point instructions.  This also
14974           enables Advanced SIMD and floating-point instructions.  This option
14975           is enabled by default for -march=armv8.6-a.  Use of this option
14976           with architectures prior to Armv8.2-A is not supported.
14977
14978       Feature crypto implies aes, sha2, and simd, which implies fp.
14979       Conversely, nofp implies nosimd, which implies nocrypto, noaes and
14980       nosha2.
14981
14982       Adapteva Epiphany Options
14983
14984       These -m options are defined for Adapteva Epiphany:
14985
14986       -mhalf-reg-file
14987           Don't allocate any register in the range "r32"..."r63".  That
14988           allows code to run on hardware variants that lack these registers.
14989
14990       -mprefer-short-insn-regs
14991           Preferentially allocate registers that allow short instruction
14992           generation.  This can result in increased instruction count, so
14993           this may either reduce or increase overall code size.
14994
14995       -mbranch-cost=num
14996           Set the cost of branches to roughly num "simple" instructions.
14997           This cost is only a heuristic and is not guaranteed to produce
14998           consistent results across releases.
14999
15000       -mcmove
15001           Enable the generation of conditional moves.
15002
15003       -mnops=num
15004           Emit num NOPs before every other generated instruction.
15005
15006       -mno-soft-cmpsf
15007           For single-precision floating-point comparisons, emit an "fsub"
15008           instruction and test the flags.  This is faster than a software
15009           comparison, but can get incorrect results in the presence of NaNs,
15010           or when two different small numbers are compared such that their
15011           difference is calculated as zero.  The default is -msoft-cmpsf,
15012           which uses slower, but IEEE-compliant, software comparisons.
15013
15014       -mstack-offset=num
15015           Set the offset between the top of the stack and the stack pointer.
15016           E.g., a value of 8 means that the eight bytes in the range
15017           "sp+0...sp+7" can be used by leaf functions without stack
15018           allocation.  Values other than 8 or 16 are untested and unlikely to
15019           work.  Note also that this option changes the ABI; compiling a
15020           program with a different stack offset than the libraries have been
15021           compiled with generally does not work.  This option can be useful
15022           if you want to evaluate if a different stack offset would give you
15023           better code, but to actually use a different stack offset to build
15024           working programs, it is recommended to configure the toolchain with
15025           the appropriate --with-stack-offset=num option.
15026
15027       -mno-round-nearest
15028           Make the scheduler assume that the rounding mode has been set to
15029           truncating.  The default is -mround-nearest.
15030
15031       -mlong-calls
15032           If not otherwise specified by an attribute, assume all calls might
15033           be beyond the offset range of the "b" / "bl" instructions, and
15034           therefore load the function address into a register before
15035           performing a (otherwise direct) call.  This is the default.
15036
15037       -mshort-calls
15038           If not otherwise specified by an attribute, assume all direct calls
15039           are in the range of the "b" / "bl" instructions, so use these
15040           instructions for direct calls.  The default is -mlong-calls.
15041
15042       -msmall16
15043           Assume addresses can be loaded as 16-bit unsigned values.  This
15044           does not apply to function addresses for which -mlong-calls
15045           semantics are in effect.
15046
15047       -mfp-mode=mode
15048           Set the prevailing mode of the floating-point unit.  This
15049           determines the floating-point mode that is provided and expected at
15050           function call and return time.  Making this mode match the mode you
15051           predominantly need at function start can make your programs smaller
15052           and faster by avoiding unnecessary mode switches.
15053
15054           mode can be set to one the following values:
15055
15056           caller
15057               Any mode at function entry is valid, and retained or restored
15058               when the function returns, and when it calls other functions.
15059               This mode is useful for compiling libraries or other
15060               compilation units you might want to incorporate into different
15061               programs with different prevailing FPU modes, and the
15062               convenience of being able to use a single object file outweighs
15063               the size and speed overhead for any extra mode switching that
15064               might be needed, compared with what would be needed with a more
15065               specific choice of prevailing FPU mode.
15066
15067           truncate
15068               This is the mode used for floating-point calculations with
15069               truncating (i.e. round towards zero) rounding mode.  That
15070               includes conversion from floating point to integer.
15071
15072           round-nearest
15073               This is the mode used for floating-point calculations with
15074               round-to-nearest-or-even rounding mode.
15075
15076           int This is the mode used to perform integer calculations in the
15077               FPU, e.g.  integer multiply, or integer multiply-and-
15078               accumulate.
15079
15080           The default is -mfp-mode=caller
15081
15082       -mno-split-lohi
15083       -mno-postinc
15084       -mno-postmodify
15085           Code generation tweaks that disable, respectively, splitting of
15086           32-bit loads, generation of post-increment addresses, and
15087           generation of post-modify addresses.  The defaults are msplit-lohi,
15088           -mpost-inc, and -mpost-modify.
15089
15090       -mnovect-double
15091           Change the preferred SIMD mode to SImode.  The default is
15092           -mvect-double, which uses DImode as preferred SIMD mode.
15093
15094       -max-vect-align=num
15095           The maximum alignment for SIMD vector mode types.  num may be 4 or
15096           8.  The default is 8.  Note that this is an ABI change, even though
15097           many library function interfaces are unaffected if they don't use
15098           SIMD vector modes in places that affect size and/or alignment of
15099           relevant types.
15100
15101       -msplit-vecmove-early
15102           Split vector moves into single word moves before reload.  In theory
15103           this can give better register allocation, but so far the reverse
15104           seems to be generally the case.
15105
15106       -m1reg-reg
15107           Specify a register to hold the constant -1, which makes loading
15108           small negative constants and certain bitmasks faster.  Allowable
15109           values for reg are r43 and r63, which specify use of that register
15110           as a fixed register, and none, which means that no register is used
15111           for this purpose.  The default is -m1reg-none.
15112
15113       AMD GCN Options
15114
15115       These options are defined specifically for the AMD GCN port.
15116
15117       -march=gpu
15118       -mtune=gpu
15119           Set architecture type or tuning for gpu. Supported values for gpu
15120           are
15121
15122           fiji
15123               Compile for GCN3 Fiji devices (gfx803).
15124
15125           gfx900
15126               Compile for GCN5 Vega 10 devices (gfx900).
15127
15128           gfx906
15129               Compile for GCN5 Vega 20 devices (gfx906).
15130
15131       -mstack-size=bytes
15132           Specify how many bytes of stack space will be requested for each
15133           GPU thread (wave-front).  Beware that there may be many threads and
15134           limited memory available.  The size of the stack allocation may
15135           also have an impact on run-time performance.  The default is 32KB
15136           when using OpenACC or OpenMP, and 1MB otherwise.
15137
15138       ARC Options
15139
15140       The following options control the architecture variant for which code
15141       is being compiled:
15142
15143       -mbarrel-shifter
15144           Generate instructions supported by barrel shifter.  This is the
15145           default unless -mcpu=ARC601 or -mcpu=ARCEM is in effect.
15146
15147       -mjli-always
15148           Force to call a function using jli_s instruction.  This option is
15149           valid only for ARCv2 architecture.
15150
15151       -mcpu=cpu
15152           Set architecture type, register usage, and instruction scheduling
15153           parameters for cpu.  There are also shortcut alias options
15154           available for backward compatibility and convenience.  Supported
15155           values for cpu are
15156
15157           arc600
15158               Compile for ARC600.  Aliases: -mA6, -mARC600.
15159
15160           arc601
15161               Compile for ARC601.  Alias: -mARC601.
15162
15163           arc700
15164               Compile for ARC700.  Aliases: -mA7, -mARC700.  This is the
15165               default when configured with --with-cpu=arc700.
15166
15167           arcem
15168               Compile for ARC EM.
15169
15170           archs
15171               Compile for ARC HS.
15172
15173           em  Compile for ARC EM CPU with no hardware extensions.
15174
15175           em4 Compile for ARC EM4 CPU.
15176
15177           em4_dmips
15178               Compile for ARC EM4 DMIPS CPU.
15179
15180           em4_fpus
15181               Compile for ARC EM4 DMIPS CPU with the single-precision
15182               floating-point extension.
15183
15184           em4_fpuda
15185               Compile for ARC EM4 DMIPS CPU with single-precision floating-
15186               point and double assist instructions.
15187
15188           hs  Compile for ARC HS CPU with no hardware extensions except the
15189               atomic instructions.
15190
15191           hs34
15192               Compile for ARC HS34 CPU.
15193
15194           hs38
15195               Compile for ARC HS38 CPU.
15196
15197           hs38_linux
15198               Compile for ARC HS38 CPU with all hardware extensions on.
15199
15200           arc600_norm
15201               Compile for ARC 600 CPU with "norm" instructions enabled.
15202
15203           arc600_mul32x16
15204               Compile for ARC 600 CPU with "norm" and 32x16-bit multiply
15205               instructions enabled.
15206
15207           arc600_mul64
15208               Compile for ARC 600 CPU with "norm" and "mul64"-family
15209               instructions enabled.
15210
15211           arc601_norm
15212               Compile for ARC 601 CPU with "norm" instructions enabled.
15213
15214           arc601_mul32x16
15215               Compile for ARC 601 CPU with "norm" and 32x16-bit multiply
15216               instructions enabled.
15217
15218           arc601_mul64
15219               Compile for ARC 601 CPU with "norm" and "mul64"-family
15220               instructions enabled.
15221
15222           nps400
15223               Compile for ARC 700 on NPS400 chip.
15224
15225           em_mini
15226               Compile for ARC EM minimalist configuration featuring reduced
15227               register set.
15228
15229       -mdpfp
15230       -mdpfp-compact
15231           Generate double-precision FPX instructions, tuned for the compact
15232           implementation.
15233
15234       -mdpfp-fast
15235           Generate double-precision FPX instructions, tuned for the fast
15236           implementation.
15237
15238       -mno-dpfp-lrsr
15239           Disable "lr" and "sr" instructions from using FPX extension aux
15240           registers.
15241
15242       -mea
15243           Generate extended arithmetic instructions.  Currently only "divaw",
15244           "adds", "subs", and "sat16" are supported.  Only valid for
15245           -mcpu=ARC700.
15246
15247       -mno-mpy
15248           Do not generate "mpy"-family instructions for ARC700.  This option
15249           is deprecated.
15250
15251       -mmul32x16
15252           Generate 32x16-bit multiply and multiply-accumulate instructions.
15253
15254       -mmul64
15255           Generate "mul64" and "mulu64" instructions.  Only valid for
15256           -mcpu=ARC600.
15257
15258       -mnorm
15259           Generate "norm" instructions.  This is the default if -mcpu=ARC700
15260           is in effect.
15261
15262       -mspfp
15263       -mspfp-compact
15264           Generate single-precision FPX instructions, tuned for the compact
15265           implementation.
15266
15267       -mspfp-fast
15268           Generate single-precision FPX instructions, tuned for the fast
15269           implementation.
15270
15271       -msimd
15272           Enable generation of ARC SIMD instructions via target-specific
15273           builtins.  Only valid for -mcpu=ARC700.
15274
15275       -msoft-float
15276           This option ignored; it is provided for compatibility purposes
15277           only.  Software floating-point code is emitted by default, and this
15278           default can overridden by FPX options; -mspfp, -mspfp-compact, or
15279           -mspfp-fast for single precision, and -mdpfp, -mdpfp-compact, or
15280           -mdpfp-fast for double precision.
15281
15282       -mswap
15283           Generate "swap" instructions.
15284
15285       -matomic
15286           This enables use of the locked load/store conditional extension to
15287           implement atomic memory built-in functions.  Not available for ARC
15288           6xx or ARC EM cores.
15289
15290       -mdiv-rem
15291           Enable "div" and "rem" instructions for ARCv2 cores.
15292
15293       -mcode-density
15294           Enable code density instructions for ARC EM.  This option is on by
15295           default for ARC HS.
15296
15297       -mll64
15298           Enable double load/store operations for ARC HS cores.
15299
15300       -mtp-regno=regno
15301           Specify thread pointer register number.
15302
15303       -mmpy-option=multo
15304           Compile ARCv2 code with a multiplier design option.  You can
15305           specify the option using either a string or numeric value for
15306           multo.  wlh1 is the default value.  The recognized values are:
15307
15308           0
15309           none
15310               No multiplier available.
15311
15312           1
15313           w   16x16 multiplier, fully pipelined.  The following instructions
15314               are enabled: "mpyw" and "mpyuw".
15315
15316           2
15317           wlh1
15318               32x32 multiplier, fully pipelined (1 stage).  The following
15319               instructions are additionally enabled: "mpy", "mpyu", "mpym",
15320               "mpymu", and "mpy_s".
15321
15322           3
15323           wlh2
15324               32x32 multiplier, fully pipelined (2 stages).  The following
15325               instructions are additionally enabled: "mpy", "mpyu", "mpym",
15326               "mpymu", and "mpy_s".
15327
15328           4
15329           wlh3
15330               Two 16x16 multipliers, blocking, sequential.  The following
15331               instructions are additionally enabled: "mpy", "mpyu", "mpym",
15332               "mpymu", and "mpy_s".
15333
15334           5
15335           wlh4
15336               One 16x16 multiplier, blocking, sequential.  The following
15337               instructions are additionally enabled: "mpy", "mpyu", "mpym",
15338               "mpymu", and "mpy_s".
15339
15340           6
15341           wlh5
15342               One 32x4 multiplier, blocking, sequential.  The following
15343               instructions are additionally enabled: "mpy", "mpyu", "mpym",
15344               "mpymu", and "mpy_s".
15345
15346           7
15347           plus_dmpy
15348               ARC HS SIMD support.
15349
15350           8
15351           plus_macd
15352               ARC HS SIMD support.
15353
15354           9
15355           plus_qmacw
15356               ARC HS SIMD support.
15357
15358           This option is only available for ARCv2 cores.
15359
15360       -mfpu=fpu
15361           Enables support for specific floating-point hardware extensions for
15362           ARCv2 cores.  Supported values for fpu are:
15363
15364           fpus
15365               Enables support for single-precision floating-point hardware
15366               extensions.
15367
15368           fpud
15369               Enables support for double-precision floating-point hardware
15370               extensions.  The single-precision floating-point extension is
15371               also enabled.  Not available for ARC EM.
15372
15373           fpuda
15374               Enables support for double-precision floating-point hardware
15375               extensions using double-precision assist instructions.  The
15376               single-precision floating-point extension is also enabled.
15377               This option is only available for ARC EM.
15378
15379           fpuda_div
15380               Enables support for double-precision floating-point hardware
15381               extensions using double-precision assist instructions.  The
15382               single-precision floating-point, square-root, and divide
15383               extensions are also enabled.  This option is only available for
15384               ARC EM.
15385
15386           fpuda_fma
15387               Enables support for double-precision floating-point hardware
15388               extensions using double-precision assist instructions.  The
15389               single-precision floating-point and fused multiply and add
15390               hardware extensions are also enabled.  This option is only
15391               available for ARC EM.
15392
15393           fpuda_all
15394               Enables support for double-precision floating-point hardware
15395               extensions using double-precision assist instructions.  All
15396               single-precision floating-point hardware extensions are also
15397               enabled.  This option is only available for ARC EM.
15398
15399           fpus_div
15400               Enables support for single-precision floating-point, square-
15401               root and divide hardware extensions.
15402
15403           fpud_div
15404               Enables support for double-precision floating-point, square-
15405               root and divide hardware extensions.  This option includes
15406               option fpus_div. Not available for ARC EM.
15407
15408           fpus_fma
15409               Enables support for single-precision floating-point and fused
15410               multiply and add hardware extensions.
15411
15412           fpud_fma
15413               Enables support for double-precision floating-point and fused
15414               multiply and add hardware extensions.  This option includes
15415               option fpus_fma.  Not available for ARC EM.
15416
15417           fpus_all
15418               Enables support for all single-precision floating-point
15419               hardware extensions.
15420
15421           fpud_all
15422               Enables support for all single- and double-precision floating-
15423               point hardware extensions.  Not available for ARC EM.
15424
15425       -mirq-ctrl-saved=register-range, blink, lp_count
15426           Specifies general-purposes registers that the processor
15427           automatically saves/restores on interrupt entry and exit.
15428           register-range is specified as two registers separated by a dash.
15429           The register range always starts with "r0", the upper limit is "fp"
15430           register.  blink and lp_count are optional.  This option is only
15431           valid for ARC EM and ARC HS cores.
15432
15433       -mrgf-banked-regs=number
15434           Specifies the number of registers replicated in second register
15435           bank on entry to fast interrupt.  Fast interrupts are interrupts
15436           with the highest priority level P0.  These interrupts save only PC
15437           and STATUS32 registers to avoid memory transactions during
15438           interrupt entry and exit sequences.  Use this option when you are
15439           using fast interrupts in an ARC V2 family processor.  Permitted
15440           values are 4, 8, 16, and 32.
15441
15442       -mlpc-width=width
15443           Specify the width of the "lp_count" register.  Valid values for
15444           width are 8, 16, 20, 24, 28 and 32 bits.  The default width is
15445           fixed to 32 bits.  If the width is less than 32, the compiler does
15446           not attempt to transform loops in your program to use the zero-
15447           delay loop mechanism unless it is known that the "lp_count"
15448           register can hold the required loop-counter value.  Depending on
15449           the width specified, the compiler and run-time library might
15450           continue to use the loop mechanism for various needs.  This option
15451           defines macro "__ARC_LPC_WIDTH__" with the value of width.
15452
15453       -mrf16
15454           This option instructs the compiler to generate code for a 16-entry
15455           register file.  This option defines the "__ARC_RF16__" preprocessor
15456           macro.
15457
15458       -mbranch-index
15459           Enable use of "bi" or "bih" instructions to implement jump tables.
15460
15461       The following options are passed through to the assembler, and also
15462       define preprocessor macro symbols.
15463
15464       -mdsp-packa
15465           Passed down to the assembler to enable the DSP Pack A extensions.
15466           Also sets the preprocessor symbol "__Xdsp_packa".  This option is
15467           deprecated.
15468
15469       -mdvbf
15470           Passed down to the assembler to enable the dual Viterbi butterfly
15471           extension.  Also sets the preprocessor symbol "__Xdvbf".  This
15472           option is deprecated.
15473
15474       -mlock
15475           Passed down to the assembler to enable the locked load/store
15476           conditional extension.  Also sets the preprocessor symbol
15477           "__Xlock".
15478
15479       -mmac-d16
15480           Passed down to the assembler.  Also sets the preprocessor symbol
15481           "__Xxmac_d16".  This option is deprecated.
15482
15483       -mmac-24
15484           Passed down to the assembler.  Also sets the preprocessor symbol
15485           "__Xxmac_24".  This option is deprecated.
15486
15487       -mrtsc
15488           Passed down to the assembler to enable the 64-bit time-stamp
15489           counter extension instruction.  Also sets the preprocessor symbol
15490           "__Xrtsc".  This option is deprecated.
15491
15492       -mswape
15493           Passed down to the assembler to enable the swap byte ordering
15494           extension instruction.  Also sets the preprocessor symbol
15495           "__Xswape".
15496
15497       -mtelephony
15498           Passed down to the assembler to enable dual- and single-operand
15499           instructions for telephony.  Also sets the preprocessor symbol
15500           "__Xtelephony".  This option is deprecated.
15501
15502       -mxy
15503           Passed down to the assembler to enable the XY memory extension.
15504           Also sets the preprocessor symbol "__Xxy".
15505
15506       The following options control how the assembly code is annotated:
15507
15508       -misize
15509           Annotate assembler instructions with estimated addresses.
15510
15511       -mannotate-align
15512           Explain what alignment considerations lead to the decision to make
15513           an instruction short or long.
15514
15515       The following options are passed through to the linker:
15516
15517       -marclinux
15518           Passed through to the linker, to specify use of the "arclinux"
15519           emulation.  This option is enabled by default in tool chains built
15520           for "arc-linux-uclibc" and "arceb-linux-uclibc" targets when
15521           profiling is not requested.
15522
15523       -marclinux_prof
15524           Passed through to the linker, to specify use of the "arclinux_prof"
15525           emulation.  This option is enabled by default in tool chains built
15526           for "arc-linux-uclibc" and "arceb-linux-uclibc" targets when
15527           profiling is requested.
15528
15529       The following options control the semantics of generated code:
15530
15531       -mlong-calls
15532           Generate calls as register indirect calls, thus providing access to
15533           the full 32-bit address range.
15534
15535       -mmedium-calls
15536           Don't use less than 25-bit addressing range for calls, which is the
15537           offset available for an unconditional branch-and-link instruction.
15538           Conditional execution of function calls is suppressed, to allow use
15539           of the 25-bit range, rather than the 21-bit range with conditional
15540           branch-and-link.  This is the default for tool chains built for
15541           "arc-linux-uclibc" and "arceb-linux-uclibc" targets.
15542
15543       -G num
15544           Put definitions of externally-visible data in a small data section
15545           if that data is no bigger than num bytes.  The default value of num
15546           is 4 for any ARC configuration, or 8 when we have double load/store
15547           operations.
15548
15549       -mno-sdata
15550           Do not generate sdata references.  This is the default for tool
15551           chains built for "arc-linux-uclibc" and "arceb-linux-uclibc"
15552           targets.
15553
15554       -mvolatile-cache
15555           Use ordinarily cached memory accesses for volatile references.
15556           This is the default.
15557
15558       -mno-volatile-cache
15559           Enable cache bypass for volatile references.
15560
15561       The following options fine tune code generation:
15562
15563       -malign-call
15564           Do alignment optimizations for call instructions.
15565
15566       -mauto-modify-reg
15567           Enable the use of pre/post modify with register displacement.
15568
15569       -mbbit-peephole
15570           Enable bbit peephole2.
15571
15572       -mno-brcc
15573           This option disables a target-specific pass in arc_reorg to
15574           generate compare-and-branch ("brcc") instructions.  It has no
15575           effect on generation of these instructions driven by the combiner
15576           pass.
15577
15578       -mcase-vector-pcrel
15579           Use PC-relative switch case tables to enable case table shortening.
15580           This is the default for -Os.
15581
15582       -mcompact-casesi
15583           Enable compact "casesi" pattern.  This is the default for -Os, and
15584           only available for ARCv1 cores.  This option is deprecated.
15585
15586       -mno-cond-exec
15587           Disable the ARCompact-specific pass to generate conditional
15588           execution instructions.
15589
15590           Due to delay slot scheduling and interactions between operand
15591           numbers, literal sizes, instruction lengths, and the support for
15592           conditional execution, the target-independent pass to generate
15593           conditional execution is often lacking, so the ARC port has kept a
15594           special pass around that tries to find more conditional execution
15595           generation opportunities after register allocation, branch
15596           shortening, and delay slot scheduling have been done.  This pass
15597           generally, but not always, improves performance and code size, at
15598           the cost of extra compilation time, which is why there is an option
15599           to switch it off.  If you have a problem with call instructions
15600           exceeding their allowable offset range because they are
15601           conditionalized, you should consider using -mmedium-calls instead.
15602
15603       -mearly-cbranchsi
15604           Enable pre-reload use of the "cbranchsi" pattern.
15605
15606       -mexpand-adddi
15607           Expand "adddi3" and "subdi3" at RTL generation time into "add.f",
15608           "adc" etc.  This option is deprecated.
15609
15610       -mindexed-loads
15611           Enable the use of indexed loads.  This can be problematic because
15612           some optimizers then assume that indexed stores exist, which is not
15613           the case.
15614
15615       -mlra
15616           Enable Local Register Allocation.  This is still experimental for
15617           ARC, so by default the compiler uses standard reload (i.e.
15618           -mno-lra).
15619
15620       -mlra-priority-none
15621           Don't indicate any priority for target registers.
15622
15623       -mlra-priority-compact
15624           Indicate target register priority for r0..r3 / r12..r15.
15625
15626       -mlra-priority-noncompact
15627           Reduce target register priority for r0..r3 / r12..r15.
15628
15629       -mmillicode
15630           When optimizing for size (using -Os), prologues and epilogues that
15631           have to save or restore a large number of registers are often
15632           shortened by using call to a special function in libgcc; this is
15633           referred to as a millicode call.  As these calls can pose
15634           performance issues, and/or cause linking issues when linking in a
15635           nonstandard way, this option is provided to turn on or off
15636           millicode call generation.
15637
15638       -mcode-density-frame
15639           This option enable the compiler to emit "enter" and "leave"
15640           instructions.  These instructions are only valid for CPUs with
15641           code-density feature.
15642
15643       -mmixed-code
15644           Tweak register allocation to help 16-bit instruction generation.
15645           This generally has the effect of decreasing the average instruction
15646           size while increasing the instruction count.
15647
15648       -mq-class
15649           Ths option is deprecated.  Enable q instruction alternatives.  This
15650           is the default for -Os.
15651
15652       -mRcq
15653           Enable Rcq constraint handling.  Most short code generation depends
15654           on this.  This is the default.
15655
15656       -mRcw
15657           Enable Rcw constraint handling.  Most ccfsm condexec mostly depends
15658           on this.  This is the default.
15659
15660       -msize-level=level
15661           Fine-tune size optimization with regards to instruction lengths and
15662           alignment.  The recognized values for level are:
15663
15664           0   No size optimization.  This level is deprecated and treated
15665               like 1.
15666
15667           1   Short instructions are used opportunistically.
15668
15669           2   In addition, alignment of loops and of code after barriers are
15670               dropped.
15671
15672           3   In addition, optional data alignment is dropped, and the option
15673               Os is enabled.
15674
15675           This defaults to 3 when -Os is in effect.  Otherwise, the behavior
15676           when this is not set is equivalent to level 1.
15677
15678       -mtune=cpu
15679           Set instruction scheduling parameters for cpu, overriding any
15680           implied by -mcpu=.
15681
15682           Supported values for cpu are
15683
15684           ARC600
15685               Tune for ARC600 CPU.
15686
15687           ARC601
15688               Tune for ARC601 CPU.
15689
15690           ARC700
15691               Tune for ARC700 CPU with standard multiplier block.
15692
15693           ARC700-xmac
15694               Tune for ARC700 CPU with XMAC block.
15695
15696           ARC725D
15697               Tune for ARC725D CPU.
15698
15699           ARC750D
15700               Tune for ARC750D CPU.
15701
15702       -mmultcost=num
15703           Cost to assume for a multiply instruction, with 4 being equal to a
15704           normal instruction.
15705
15706       -munalign-prob-threshold=probability
15707           Set probability threshold for unaligning branches.  When tuning for
15708           ARC700 and optimizing for speed, branches without filled delay slot
15709           are preferably emitted unaligned and long, unless profiling
15710           indicates that the probability for the branch to be taken is below
15711           probability.  The default is (REG_BR_PROB_BASE/2), i.e. 5000.
15712
15713       The following options are maintained for backward compatibility, but
15714       are now deprecated and will be removed in a future release:
15715
15716       -margonaut
15717           Obsolete FPX.
15718
15719       -mbig-endian
15720       -EB Compile code for big-endian targets.  Use of these options is now
15721           deprecated.  Big-endian code is supported by configuring GCC to
15722           build "arceb-elf32" and "arceb-linux-uclibc" targets, for which big
15723           endian is the default.
15724
15725       -mlittle-endian
15726       -EL Compile code for little-endian targets.  Use of these options is
15727           now deprecated.  Little-endian code is supported by configuring GCC
15728           to build "arc-elf32" and "arc-linux-uclibc" targets, for which
15729           little endian is the default.
15730
15731       -mbarrel_shifter
15732           Replaced by -mbarrel-shifter.
15733
15734       -mdpfp_compact
15735           Replaced by -mdpfp-compact.
15736
15737       -mdpfp_fast
15738           Replaced by -mdpfp-fast.
15739
15740       -mdsp_packa
15741           Replaced by -mdsp-packa.
15742
15743       -mEA
15744           Replaced by -mea.
15745
15746       -mmac_24
15747           Replaced by -mmac-24.
15748
15749       -mmac_d16
15750           Replaced by -mmac-d16.
15751
15752       -mspfp_compact
15753           Replaced by -mspfp-compact.
15754
15755       -mspfp_fast
15756           Replaced by -mspfp-fast.
15757
15758       -mtune=cpu
15759           Values arc600, arc601, arc700 and arc700-xmac for cpu are replaced
15760           by ARC600, ARC601, ARC700 and ARC700-xmac respectively.
15761
15762       -multcost=num
15763           Replaced by -mmultcost.
15764
15765       ARM Options
15766
15767       These -m options are defined for the ARM port:
15768
15769       -mabi=name
15770           Generate code for the specified ABI.  Permissible values are: apcs-
15771           gnu, atpcs, aapcs, aapcs-linux and iwmmxt.
15772
15773       -mapcs-frame
15774           Generate a stack frame that is compliant with the ARM Procedure
15775           Call Standard for all functions, even if this is not strictly
15776           necessary for correct execution of the code.  Specifying
15777           -fomit-frame-pointer with this option causes the stack frames not
15778           to be generated for leaf functions.  The default is
15779           -mno-apcs-frame.  This option is deprecated.
15780
15781       -mapcs
15782           This is a synonym for -mapcs-frame and is deprecated.
15783
15784       -mthumb-interwork
15785           Generate code that supports calling between the ARM and Thumb
15786           instruction sets.  Without this option, on pre-v5 architectures,
15787           the two instruction sets cannot be reliably used inside one
15788           program.  The default is -mno-thumb-interwork, since slightly
15789           larger code is generated when -mthumb-interwork is specified.  In
15790           AAPCS configurations this option is meaningless.
15791
15792       -mno-sched-prolog
15793           Prevent the reordering of instructions in the function prologue, or
15794           the merging of those instruction with the instructions in the
15795           function's body.  This means that all functions start with a
15796           recognizable set of instructions (or in fact one of a choice from a
15797           small set of different function prologues), and this information
15798           can be used to locate the start of functions inside an executable
15799           piece of code.  The default is -msched-prolog.
15800
15801       -mfloat-abi=name
15802           Specifies which floating-point ABI to use.  Permissible values are:
15803           soft, softfp and hard.
15804
15805           Specifying soft causes GCC to generate output containing library
15806           calls for floating-point operations.  softfp allows the generation
15807           of code using hardware floating-point instructions, but still uses
15808           the soft-float calling conventions.  hard allows generation of
15809           floating-point instructions and uses FPU-specific calling
15810           conventions.
15811
15812           The default depends on the specific target configuration.  Note
15813           that the hard-float and soft-float ABIs are not link-compatible;
15814           you must compile your entire program with the same ABI, and link
15815           with a compatible set of libraries.
15816
15817       -mgeneral-regs-only
15818           Generate code which uses only the general-purpose registers.  This
15819           will prevent the compiler from using floating-point and Advanced
15820           SIMD registers but will not impose any restrictions on the
15821           assembler.
15822
15823       -mlittle-endian
15824           Generate code for a processor running in little-endian mode.  This
15825           is the default for all standard configurations.
15826
15827       -mbig-endian
15828           Generate code for a processor running in big-endian mode; the
15829           default is to compile code for a little-endian processor.
15830
15831       -mbe8
15832       -mbe32
15833           When linking a big-endian image select between BE8 and BE32
15834           formats.  The option has no effect for little-endian images and is
15835           ignored.  The default is dependent on the selected target
15836           architecture.  For ARMv6 and later architectures the default is
15837           BE8, for older architectures the default is BE32.  BE32 format has
15838           been deprecated by ARM.
15839
15840       -march=name[+extension...]
15841           This specifies the name of the target ARM architecture.  GCC uses
15842           this name to determine what kind of instructions it can emit when
15843           generating assembly code.  This option can be used in conjunction
15844           with or instead of the -mcpu= option.
15845
15846           Permissible names are: armv4t, armv5t, armv5te, armv6, armv6j,
15847           armv6k, armv6kz, armv6t2, armv6z, armv6zk, armv7, armv7-a, armv7ve,
15848           armv8-a, armv8.1-a, armv8.2-a, armv8.3-a, armv8.4-a, armv8.5-a,
15849           armv8.6-a, armv7-r, armv8-r, armv6-m, armv6s-m, armv7-m, armv7e-m,
15850           armv8-m.base, armv8-m.main, armv8.1-m.main, iwmmxt and iwmmxt2.
15851
15852           Additionally, the following architectures, which lack support for
15853           the Thumb execution state, are recognized but support is
15854           deprecated: armv4.
15855
15856           Many of the architectures support extensions.  These can be added
15857           by appending +extension to the architecture name.  Extension
15858           options are processed in order and capabilities accumulate.  An
15859           extension will also enable any necessary base extensions upon which
15860           it depends.  For example, the +crypto extension will always enable
15861           the +simd extension.  The exception to the additive construction is
15862           for extensions that are prefixed with +no...: these extensions
15863           disable the specified option and any other extensions that may
15864           depend on the presence of that extension.
15865
15866           For example, -march=armv7-a+simd+nofp+vfpv4 is equivalent to
15867           writing -march=armv7-a+vfpv4 since the +simd option is entirely
15868           disabled by the +nofp option that follows it.
15869
15870           Most extension names are generically named, but have an effect that
15871           is dependent upon the architecture to which it is applied.  For
15872           example, the +simd option can be applied to both armv7-a and
15873           armv8-a architectures, but will enable the original ARMv7-A
15874           Advanced SIMD (Neon) extensions for armv7-a and the ARMv8-A variant
15875           for armv8-a.
15876
15877           The table below lists the supported extensions for each
15878           architecture.  Architectures not mentioned do not support any
15879           extensions.
15880
15881           armv5te
15882           armv6
15883           armv6j
15884           armv6k
15885           armv6kz
15886           armv6t2
15887           armv6z
15888           armv6zk
15889               +fp The VFPv2 floating-point instructions.  The extension
15890                   +vfpv2 can be used as an alias for this extension.
15891
15892               +nofp
15893                   Disable the floating-point instructions.
15894
15895           armv7
15896               The common subset of the ARMv7-A, ARMv7-R and ARMv7-M
15897               architectures.
15898
15899               +fp The VFPv3 floating-point instructions, with 16 double-
15900                   precision registers.  The extension +vfpv3-d16 can be used
15901                   as an alias for this extension.  Note that floating-point
15902                   is not supported by the base ARMv7-M architecture, but is
15903                   compatible with both the ARMv7-A and ARMv7-R architectures.
15904
15905               +nofp
15906                   Disable the floating-point instructions.
15907
15908           armv7-a
15909               +mp The multiprocessing extension.
15910
15911               +sec
15912                   The security extension.
15913
15914               +fp The VFPv3 floating-point instructions, with 16 double-
15915                   precision registers.  The extension +vfpv3-d16 can be used
15916                   as an alias for this extension.
15917
15918               +simd
15919                   The Advanced SIMD (Neon) v1 and the VFPv3 floating-point
15920                   instructions.  The extensions +neon and +neon-vfpv3 can be
15921                   used as aliases for this extension.
15922
15923               +vfpv3
15924                   The VFPv3 floating-point instructions, with 32 double-
15925                   precision registers.
15926
15927               +vfpv3-d16-fp16
15928                   The VFPv3 floating-point instructions, with 16 double-
15929                   precision registers and the half-precision floating-point
15930                   conversion operations.
15931
15932               +vfpv3-fp16
15933                   The VFPv3 floating-point instructions, with 32 double-
15934                   precision registers and the half-precision floating-point
15935                   conversion operations.
15936
15937               +vfpv4-d16
15938                   The VFPv4 floating-point instructions, with 16 double-
15939                   precision registers.
15940
15941               +vfpv4
15942                   The VFPv4 floating-point instructions, with 32 double-
15943                   precision registers.
15944
15945               +neon-fp16
15946                   The Advanced SIMD (Neon) v1 and the VFPv3 floating-point
15947                   instructions, with the half-precision floating-point
15948                   conversion operations.
15949
15950               +neon-vfpv4
15951                   The Advanced SIMD (Neon) v2 and the VFPv4 floating-point
15952                   instructions.
15953
15954               +nosimd
15955                   Disable the Advanced SIMD instructions (does not disable
15956                   floating point).
15957
15958               +nofp
15959                   Disable the floating-point and Advanced SIMD instructions.
15960
15961           armv7ve
15962               The extended version of the ARMv7-A architecture with support
15963               for virtualization.
15964
15965               +fp The VFPv4 floating-point instructions, with 16 double-
15966                   precision registers.  The extension +vfpv4-d16 can be used
15967                   as an alias for this extension.
15968
15969               +simd
15970                   The Advanced SIMD (Neon) v2 and the VFPv4 floating-point
15971                   instructions.  The extension +neon-vfpv4 can be used as an
15972                   alias for this extension.
15973
15974               +vfpv3-d16
15975                   The VFPv3 floating-point instructions, with 16 double-
15976                   precision registers.
15977
15978               +vfpv3
15979                   The VFPv3 floating-point instructions, with 32 double-
15980                   precision registers.
15981
15982               +vfpv3-d16-fp16
15983                   The VFPv3 floating-point instructions, with 16 double-
15984                   precision registers and the half-precision floating-point
15985                   conversion operations.
15986
15987               +vfpv3-fp16
15988                   The VFPv3 floating-point instructions, with 32 double-
15989                   precision registers and the half-precision floating-point
15990                   conversion operations.
15991
15992               +vfpv4-d16
15993                   The VFPv4 floating-point instructions, with 16 double-
15994                   precision registers.
15995
15996               +vfpv4
15997                   The VFPv4 floating-point instructions, with 32 double-
15998                   precision registers.
15999
16000               +neon
16001                   The Advanced SIMD (Neon) v1 and the VFPv3 floating-point
16002                   instructions.  The extension +neon-vfpv3 can be used as an
16003                   alias for this extension.
16004
16005               +neon-fp16
16006                   The Advanced SIMD (Neon) v1 and the VFPv3 floating-point
16007                   instructions, with the half-precision floating-point
16008                   conversion operations.
16009
16010               +nosimd
16011                   Disable the Advanced SIMD instructions (does not disable
16012                   floating point).
16013
16014               +nofp
16015                   Disable the floating-point and Advanced SIMD instructions.
16016
16017           armv8-a
16018               +crc
16019                   The Cyclic Redundancy Check (CRC) instructions.
16020
16021               +simd
16022                   The ARMv8-A Advanced SIMD and floating-point instructions.
16023
16024               +crypto
16025                   The cryptographic instructions.
16026
16027               +nocrypto
16028                   Disable the cryptographic instructions.
16029
16030               +nofp
16031                   Disable the floating-point, Advanced SIMD and cryptographic
16032                   instructions.
16033
16034               +sb Speculation Barrier Instruction.
16035
16036               +predres
16037                   Execution and Data Prediction Restriction Instructions.
16038
16039           armv8.1-a
16040               +simd
16041                   The ARMv8.1-A Advanced SIMD and floating-point
16042                   instructions.
16043
16044               +crypto
16045                   The cryptographic instructions.  This also enables the
16046                   Advanced SIMD and floating-point instructions.
16047
16048               +nocrypto
16049                   Disable the cryptographic instructions.
16050
16051               +nofp
16052                   Disable the floating-point, Advanced SIMD and cryptographic
16053                   instructions.
16054
16055               +sb Speculation Barrier Instruction.
16056
16057               +predres
16058                   Execution and Data Prediction Restriction Instructions.
16059
16060           armv8.2-a
16061           armv8.3-a
16062               +fp16
16063                   The half-precision floating-point data processing
16064                   instructions.  This also enables the Advanced SIMD and
16065                   floating-point instructions.
16066
16067               +fp16fml
16068                   The half-precision floating-point fmla extension.  This
16069                   also enables the half-precision floating-point extension
16070                   and Advanced SIMD and floating-point instructions.
16071
16072               +simd
16073                   The ARMv8.1-A Advanced SIMD and floating-point
16074                   instructions.
16075
16076               +crypto
16077                   The cryptographic instructions.  This also enables the
16078                   Advanced SIMD and floating-point instructions.
16079
16080               +dotprod
16081                   Enable the Dot Product extension.  This also enables
16082                   Advanced SIMD instructions.
16083
16084               +nocrypto
16085                   Disable the cryptographic extension.
16086
16087               +nofp
16088                   Disable the floating-point, Advanced SIMD and cryptographic
16089                   instructions.
16090
16091               +sb Speculation Barrier Instruction.
16092
16093               +predres
16094                   Execution and Data Prediction Restriction Instructions.
16095
16096               +i8mm
16097                   8-bit Integer Matrix Multiply instructions.  This also
16098                   enables Advanced SIMD and floating-point instructions.
16099
16100               +bf16
16101                   Brain half-precision floating-point instructions.  This
16102                   also enables Advanced SIMD and floating-point instructions.
16103
16104           armv8.4-a
16105               +fp16
16106                   The half-precision floating-point data processing
16107                   instructions.  This also enables the Advanced SIMD and
16108                   floating-point instructions as well as the Dot Product
16109                   extension and the half-precision floating-point fmla
16110                   extension.
16111
16112               +simd
16113                   The ARMv8.3-A Advanced SIMD and floating-point instructions
16114                   as well as the Dot Product extension.
16115
16116               +crypto
16117                   The cryptographic instructions.  This also enables the
16118                   Advanced SIMD and floating-point instructions as well as
16119                   the Dot Product extension.
16120
16121               +nocrypto
16122                   Disable the cryptographic extension.
16123
16124               +nofp
16125                   Disable the floating-point, Advanced SIMD and cryptographic
16126                   instructions.
16127
16128               +sb Speculation Barrier Instruction.
16129
16130               +predres
16131                   Execution and Data Prediction Restriction Instructions.
16132
16133               +i8mm
16134                   8-bit Integer Matrix Multiply instructions.  This also
16135                   enables Advanced SIMD and floating-point instructions.
16136
16137               +bf16
16138                   Brain half-precision floating-point instructions.  This
16139                   also enables Advanced SIMD and floating-point instructions.
16140
16141           armv8.5-a
16142               +fp16
16143                   The half-precision floating-point data processing
16144                   instructions.  This also enables the Advanced SIMD and
16145                   floating-point instructions as well as the Dot Product
16146                   extension and the half-precision floating-point fmla
16147                   extension.
16148
16149               +simd
16150                   The ARMv8.3-A Advanced SIMD and floating-point instructions
16151                   as well as the Dot Product extension.
16152
16153               +crypto
16154                   The cryptographic instructions.  This also enables the
16155                   Advanced SIMD and floating-point instructions as well as
16156                   the Dot Product extension.
16157
16158               +nocrypto
16159                   Disable the cryptographic extension.
16160
16161               +nofp
16162                   Disable the floating-point, Advanced SIMD and cryptographic
16163                   instructions.
16164
16165               +i8mm
16166                   8-bit Integer Matrix Multiply instructions.  This also
16167                   enables Advanced SIMD and floating-point instructions.
16168
16169               +bf16
16170                   Brain half-precision floating-point instructions.  This
16171                   also enables Advanced SIMD and floating-point instructions.
16172
16173           armv8.6-a
16174               +fp16
16175                   The half-precision floating-point data processing
16176                   instructions.  This also enables the Advanced SIMD and
16177                   floating-point instructions as well as the Dot Product
16178                   extension and the half-precision floating-point fmla
16179                   extension.
16180
16181               +simd
16182                   The ARMv8.3-A Advanced SIMD and floating-point instructions
16183                   as well as the Dot Product extension.
16184
16185               +crypto
16186                   The cryptographic instructions.  This also enables the
16187                   Advanced SIMD and floating-point instructions as well as
16188                   the Dot Product extension.
16189
16190               +nocrypto
16191                   Disable the cryptographic extension.
16192
16193               +nofp
16194                   Disable the floating-point, Advanced SIMD and cryptographic
16195                   instructions.
16196
16197               +i8mm
16198                   8-bit Integer Matrix Multiply instructions.  This also
16199                   enables Advanced SIMD and floating-point instructions.
16200
16201               +bf16
16202                   Brain half-precision floating-point instructions.  This
16203                   also enables Advanced SIMD and floating-point instructions.
16204
16205           armv7-r
16206               +fp.sp
16207                   The single-precision VFPv3 floating-point instructions.
16208                   The extension +vfpv3xd can be used as an alias for this
16209                   extension.
16210
16211               +fp The VFPv3 floating-point instructions with 16 double-
16212                   precision registers.  The extension +vfpv3-d16 can be used
16213                   as an alias for this extension.
16214
16215               +vfpv3xd-d16-fp16
16216                   The single-precision VFPv3 floating-point instructions with
16217                   16 double-precision registers and the half-precision
16218                   floating-point conversion operations.
16219
16220               +vfpv3-d16-fp16
16221                   The VFPv3 floating-point instructions with 16 double-
16222                   precision registers and the half-precision floating-point
16223                   conversion operations.
16224
16225               +nofp
16226                   Disable the floating-point extension.
16227
16228               +idiv
16229                   The ARM-state integer division instructions.
16230
16231               +noidiv
16232                   Disable the ARM-state integer division extension.
16233
16234           armv7e-m
16235               +fp The single-precision VFPv4 floating-point instructions.
16236
16237               +fpv5
16238                   The single-precision FPv5 floating-point instructions.
16239
16240               +fp.dp
16241                   The single- and double-precision FPv5 floating-point
16242                   instructions.
16243
16244               +nofp
16245                   Disable the floating-point extensions.
16246
16247           armv8.1-m.main
16248               +dsp
16249                   The DSP instructions.
16250
16251               +mve
16252                   The M-Profile Vector Extension (MVE) integer instructions.
16253
16254               +mve.fp
16255                   The M-Profile Vector Extension (MVE) integer and single
16256                   precision floating-point instructions.
16257
16258               +fp The single-precision floating-point instructions.
16259
16260               +fp.dp
16261                   The single- and double-precision floating-point
16262                   instructions.
16263
16264               +nofp
16265                   Disable the floating-point extension.
16266
16267               +cdecp0, +cdecp1, ... , +cdecp7
16268                   Enable the Custom Datapath Extension (CDE) on selected
16269                   coprocessors according to the numbers given in the options
16270                   in the range 0 to 7.
16271
16272           armv8-m.main
16273               +dsp
16274                   The DSP instructions.
16275
16276               +nodsp
16277                   Disable the DSP extension.
16278
16279               +fp The single-precision floating-point instructions.
16280
16281               +fp.dp
16282                   The single- and double-precision floating-point
16283                   instructions.
16284
16285               +nofp
16286                   Disable the floating-point extension.
16287
16288               +cdecp0, +cdecp1, ... , +cdecp7
16289                   Enable the Custom Datapath Extension (CDE) on selected
16290                   coprocessors according to the numbers given in the options
16291                   in the range 0 to 7.
16292
16293           armv8-r
16294               +crc
16295                   The Cyclic Redundancy Check (CRC) instructions.
16296
16297               +fp.sp
16298                   The single-precision FPv5 floating-point instructions.
16299
16300               +simd
16301                   The ARMv8-A Advanced SIMD and floating-point instructions.
16302
16303               +crypto
16304                   The cryptographic instructions.
16305
16306               +nocrypto
16307                   Disable the cryptographic instructions.
16308
16309               +nofp
16310                   Disable the floating-point, Advanced SIMD and cryptographic
16311                   instructions.
16312
16313           -march=native causes the compiler to auto-detect the architecture
16314           of the build computer.  At present, this feature is only supported
16315           on GNU/Linux, and not all architectures are recognized.  If the
16316           auto-detect is unsuccessful the option has no effect.
16317
16318       -mtune=name
16319           This option specifies the name of the target ARM processor for
16320           which GCC should tune the performance of the code.  For some ARM
16321           implementations better performance can be obtained by using this
16322           option.  Permissible names are: arm7tdmi, arm7tdmi-s, arm710t,
16323           arm720t, arm740t, strongarm, strongarm110, strongarm1100,
16324           0strongarm1110, arm8, arm810, arm9, arm9e, arm920, arm920t,
16325           arm922t, arm946e-s, arm966e-s, arm968e-s, arm926ej-s, arm940t,
16326           arm9tdmi, arm10tdmi, arm1020t, arm1026ej-s, arm10e, arm1020e,
16327           arm1022e, arm1136j-s, arm1136jf-s, mpcore, mpcorenovfp,
16328           arm1156t2-s, arm1156t2f-s, arm1176jz-s, arm1176jzf-s,
16329           generic-armv7-a, cortex-a5, cortex-a7, cortex-a8, cortex-a9,
16330           cortex-a12, cortex-a15, cortex-a17, cortex-a32, cortex-a35,
16331           cortex-a53, cortex-a55, cortex-a57, cortex-a72, cortex-a73,
16332           cortex-a75, cortex-a76, cortex-a76ae, cortex-a77, ares, cortex-r4,
16333           cortex-r4f, cortex-r5, cortex-r7, cortex-r8, cortex-r52, cortex-m0,
16334           cortex-m0plus, cortex-m1, cortex-m3, cortex-m4, cortex-m7,
16335           cortex-m23, cortex-m33, cortex-m35p, cortex-m55,
16336           cortex-m1.small-multiply, cortex-m0.small-multiply,
16337           cortex-m0plus.small-multiply, exynos-m1, marvell-pj4, neoverse-n1,
16338           xscale, iwmmxt, iwmmxt2, ep9312, fa526, fa626, fa606te, fa626te,
16339           fmp626, fa726te, xgene1.
16340
16341           Additionally, this option can specify that GCC should tune the
16342           performance of the code for a big.LITTLE system.  Permissible names
16343           are: cortex-a15.cortex-a7, cortex-a17.cortex-a7,
16344           cortex-a57.cortex-a53, cortex-a72.cortex-a53,
16345           cortex-a72.cortex-a35, cortex-a73.cortex-a53,
16346           cortex-a75.cortex-a55, cortex-a76.cortex-a55.
16347
16348           -mtune=generic-arch specifies that GCC should tune the performance
16349           for a blend of processors within architecture arch.  The aim is to
16350           generate code that run well on the current most popular processors,
16351           balancing between optimizations that benefit some CPUs in the
16352           range, and avoiding performance pitfalls of other CPUs.  The
16353           effects of this option may change in future GCC versions as CPU
16354           models come and go.
16355
16356           -mtune permits the same extension options as -mcpu, but the
16357           extension options do not affect the tuning of the generated code.
16358
16359           -mtune=native causes the compiler to auto-detect the CPU of the
16360           build computer.  At present, this feature is only supported on
16361           GNU/Linux, and not all architectures are recognized.  If the auto-
16362           detect is unsuccessful the option has no effect.
16363
16364       -mcpu=name[+extension...]
16365           This specifies the name of the target ARM processor.  GCC uses this
16366           name to derive the name of the target ARM architecture (as if
16367           specified by -march) and the ARM processor type for which to tune
16368           for performance (as if specified by -mtune).  Where this option is
16369           used in conjunction with -march or -mtune, those options take
16370           precedence over the appropriate part of this option.
16371
16372           Many of the supported CPUs implement optional architectural
16373           extensions.  Where this is so the architectural extensions are
16374           normally enabled by default.  If implementations that lack the
16375           extension exist, then the extension syntax can be used to disable
16376           those extensions that have been omitted.  For floating-point and
16377           Advanced SIMD (Neon) instructions, the settings of the options
16378           -mfloat-abi and -mfpu must also be considered: floating-point and
16379           Advanced SIMD instructions will only be used if -mfloat-abi is not
16380           set to soft; and any setting of -mfpu other than auto will override
16381           the available floating-point and SIMD extension instructions.
16382
16383           For example, cortex-a9 can be found in three major configurations:
16384           integer only, with just a floating-point unit or with floating-
16385           point and Advanced SIMD.  The default is to enable all the
16386           instructions, but the extensions +nosimd and +nofp can be used to
16387           disable just the SIMD or both the SIMD and floating-point
16388           instructions respectively.
16389
16390           Permissible names for this option are the same as those for -mtune.
16391
16392           The following extension options are common to the listed CPUs:
16393
16394           +nodsp
16395               Disable the DSP instructions on cortex-m33, cortex-m35p.
16396
16397           +nofp
16398               Disables the floating-point instructions on arm9e, arm946e-s,
16399               arm966e-s, arm968e-s, arm10e, arm1020e, arm1022e, arm926ej-s,
16400               arm1026ej-s, cortex-r5, cortex-r7, cortex-r8, cortex-m4,
16401               cortex-m7, cortex-m33 and cortex-m35p.  Disables the floating-
16402               point and SIMD instructions on generic-armv7-a, cortex-a5,
16403               cortex-a7, cortex-a8, cortex-a9, cortex-a12, cortex-a15,
16404               cortex-a17, cortex-a15.cortex-a7, cortex-a17.cortex-a7,
16405               cortex-a32, cortex-a35, cortex-a53 and cortex-a55.
16406
16407           +nofp.dp
16408               Disables the double-precision component of the floating-point
16409               instructions on cortex-r5, cortex-r7, cortex-r8, cortex-r52 and
16410               cortex-m7.
16411
16412           +nosimd
16413               Disables the SIMD (but not floating-point) instructions on
16414               generic-armv7-a, cortex-a5, cortex-a7 and cortex-a9.
16415
16416           +crypto
16417               Enables the cryptographic instructions on cortex-a32,
16418               cortex-a35, cortex-a53, cortex-a55, cortex-a57, cortex-a72,
16419               cortex-a73, cortex-a75, exynos-m1, xgene1,
16420               cortex-a57.cortex-a53, cortex-a72.cortex-a53,
16421               cortex-a73.cortex-a35, cortex-a73.cortex-a53 and
16422               cortex-a75.cortex-a55.
16423
16424           Additionally the generic-armv7-a pseudo target defaults to VFPv3
16425           with 16 double-precision registers.  It supports the following
16426           extension options: mp, sec, vfpv3-d16, vfpv3, vfpv3-d16-fp16,
16427           vfpv3-fp16, vfpv4-d16, vfpv4, neon, neon-vfpv3, neon-fp16,
16428           neon-vfpv4.  The meanings are the same as for the extensions to
16429           -march=armv7-a.
16430
16431           -mcpu=generic-arch is also permissible, and is equivalent to
16432           -march=arch -mtune=generic-arch.  See -mtune for more information.
16433
16434           -mcpu=native causes the compiler to auto-detect the CPU of the
16435           build computer.  At present, this feature is only supported on
16436           GNU/Linux, and not all architectures are recognized.  If the auto-
16437           detect is unsuccessful the option has no effect.
16438
16439       -mfpu=name
16440           This specifies what floating-point hardware (or hardware emulation)
16441           is available on the target.  Permissible names are: auto, vfpv2,
16442           vfpv3, vfpv3-fp16, vfpv3-d16, vfpv3-d16-fp16, vfpv3xd,
16443           vfpv3xd-fp16, neon-vfpv3, neon-fp16, vfpv4, vfpv4-d16, fpv4-sp-d16,
16444           neon-vfpv4, fpv5-d16, fpv5-sp-d16, fp-armv8, neon-fp-armv8 and
16445           crypto-neon-fp-armv8.  Note that neon is an alias for neon-vfpv3
16446           and vfp is an alias for vfpv2.
16447
16448           The setting auto is the default and is special.  It causes the
16449           compiler to select the floating-point and Advanced SIMD
16450           instructions based on the settings of -mcpu and -march.
16451
16452           If the selected floating-point hardware includes the NEON extension
16453           (e.g. -mfpu=neon), note that floating-point operations are not
16454           generated by GCC's auto-vectorization pass unless
16455           -funsafe-math-optimizations is also specified.  This is because
16456           NEON hardware does not fully implement the IEEE 754 standard for
16457           floating-point arithmetic (in particular denormal values are
16458           treated as zero), so the use of NEON instructions may lead to a
16459           loss of precision.
16460
16461           You can also set the fpu name at function level by using the
16462           "target("fpu=")" function attributes or pragmas.
16463
16464       -mfp16-format=name
16465           Specify the format of the "__fp16" half-precision floating-point
16466           type.  Permissible names are none, ieee, and alternative; the
16467           default is none, in which case the "__fp16" type is not defined.
16468
16469       -mstructure-size-boundary=n
16470           The sizes of all structures and unions are rounded up to a multiple
16471           of the number of bits set by this option.  Permissible values are
16472           8, 32 and 64.  The default value varies for different toolchains.
16473           For the COFF targeted toolchain the default value is 8.  A value of
16474           64 is only allowed if the underlying ABI supports it.
16475
16476           Specifying a larger number can produce faster, more efficient code,
16477           but can also increase the size of the program.  Different values
16478           are potentially incompatible.  Code compiled with one value cannot
16479           necessarily expect to work with code or libraries compiled with
16480           another value, if they exchange information using structures or
16481           unions.
16482
16483           This option is deprecated.
16484
16485       -mabort-on-noreturn
16486           Generate a call to the function "abort" at the end of a "noreturn"
16487           function.  It is executed if the function tries to return.
16488
16489       -mlong-calls
16490       -mno-long-calls
16491           Tells the compiler to perform function calls by first loading the
16492           address of the function into a register and then performing a
16493           subroutine call on this register.  This switch is needed if the
16494           target function lies outside of the 64-megabyte addressing range of
16495           the offset-based version of subroutine call instruction.
16496
16497           Even if this switch is enabled, not all function calls are turned
16498           into long calls.  The heuristic is that static functions, functions
16499           that have the "short_call" attribute, functions that are inside the
16500           scope of a "#pragma no_long_calls" directive, and functions whose
16501           definitions have already been compiled within the current
16502           compilation unit are not turned into long calls.  The exceptions to
16503           this rule are that weak function definitions, functions with the
16504           "long_call" attribute or the "section" attribute, and functions
16505           that are within the scope of a "#pragma long_calls" directive are
16506           always turned into long calls.
16507
16508           This feature is not enabled by default.  Specifying -mno-long-calls
16509           restores the default behavior, as does placing the function calls
16510           within the scope of a "#pragma long_calls_off" directive.  Note
16511           these switches have no effect on how the compiler generates code to
16512           handle function calls via function pointers.
16513
16514       -msingle-pic-base
16515           Treat the register used for PIC addressing as read-only, rather
16516           than loading it in the prologue for each function.  The runtime
16517           system is responsible for initializing this register with an
16518           appropriate value before execution begins.
16519
16520       -mpic-register=reg
16521           Specify the register to be used for PIC addressing.  For standard
16522           PIC base case, the default is any suitable register determined by
16523           compiler.  For single PIC base case, the default is R9 if target is
16524           EABI based or stack-checking is enabled, otherwise the default is
16525           R10.
16526
16527       -mpic-data-is-text-relative
16528           Assume that the displacement between the text and data segments is
16529           fixed at static link time.  This permits using PC-relative
16530           addressing operations to access data known to be in the data
16531           segment.  For non-VxWorks RTP targets, this option is enabled by
16532           default.  When disabled on such targets, it will enable
16533           -msingle-pic-base by default.
16534
16535       -mpoke-function-name
16536           Write the name of each function into the text section, directly
16537           preceding the function prologue.  The generated code is similar to
16538           this:
16539
16540                        t0
16541                            .ascii "arm_poke_function_name", 0
16542                            .align
16543                        t1
16544                            .word 0xff000000 + (t1 - t0)
16545                        arm_poke_function_name
16546                            mov     ip, sp
16547                            stmfd   sp!, {fp, ip, lr, pc}
16548                            sub     fp, ip, #4
16549
16550           When performing a stack backtrace, code can inspect the value of
16551           "pc" stored at "fp + 0".  If the trace function then looks at
16552           location "pc - 12" and the top 8 bits are set, then we know that
16553           there is a function name embedded immediately preceding this
16554           location and has length "((pc[-3]) & 0xff000000)".
16555
16556       -mthumb
16557       -marm
16558           Select between generating code that executes in ARM and Thumb
16559           states.  The default for most configurations is to generate code
16560           that executes in ARM state, but the default can be changed by
16561           configuring GCC with the --with-mode=state configure option.
16562
16563           You can also override the ARM and Thumb mode for each function by
16564           using the "target("thumb")" and "target("arm")" function attributes
16565           or pragmas.
16566
16567       -mflip-thumb
16568           Switch ARM/Thumb modes on alternating functions.  This option is
16569           provided for regression testing of mixed Thumb/ARM code generation,
16570           and is not intended for ordinary use in compiling code.
16571
16572       -mtpcs-frame
16573           Generate a stack frame that is compliant with the Thumb Procedure
16574           Call Standard for all non-leaf functions.  (A leaf function is one
16575           that does not call any other functions.)  The default is
16576           -mno-tpcs-frame.
16577
16578       -mtpcs-leaf-frame
16579           Generate a stack frame that is compliant with the Thumb Procedure
16580           Call Standard for all leaf functions.  (A leaf function is one that
16581           does not call any other functions.)  The default is
16582           -mno-apcs-leaf-frame.
16583
16584       -mcallee-super-interworking
16585           Gives all externally visible functions in the file being compiled
16586           an ARM instruction set header which switches to Thumb mode before
16587           executing the rest of the function.  This allows these functions to
16588           be called from non-interworking code.  This option is not valid in
16589           AAPCS configurations because interworking is enabled by default.
16590
16591       -mcaller-super-interworking
16592           Allows calls via function pointers (including virtual functions) to
16593           execute correctly regardless of whether the target code has been
16594           compiled for interworking or not.  There is a small overhead in the
16595           cost of executing a function pointer if this option is enabled.
16596           This option is not valid in AAPCS configurations because
16597           interworking is enabled by default.
16598
16599       -mtp=name
16600           Specify the access model for the thread local storage pointer.  The
16601           valid models are soft, which generates calls to "__aeabi_read_tp",
16602           cp15, which fetches the thread pointer from "cp15" directly
16603           (supported in the arm6k architecture), and auto, which uses the
16604           best available method for the selected processor.  The default
16605           setting is auto.
16606
16607       -mtls-dialect=dialect
16608           Specify the dialect to use for accessing thread local storage.  Two
16609           dialects are supported---gnu and gnu2.  The gnu dialect selects the
16610           original GNU scheme for supporting local and global dynamic TLS
16611           models.  The gnu2 dialect selects the GNU descriptor scheme, which
16612           provides better performance for shared libraries.  The GNU
16613           descriptor scheme is compatible with the original scheme, but does
16614           require new assembler, linker and library support.  Initial and
16615           local exec TLS models are unaffected by this option and always use
16616           the original scheme.
16617
16618       -mword-relocations
16619           Only generate absolute relocations on word-sized values (i.e.
16620           R_ARM_ABS32).  This is enabled by default on targets (uClinux,
16621           SymbianOS) where the runtime loader imposes this restriction, and
16622           when -fpic or -fPIC is specified. This option conflicts with
16623           -mslow-flash-data.
16624
16625       -mfix-cortex-m3-ldrd
16626           Some Cortex-M3 cores can cause data corruption when "ldrd"
16627           instructions with overlapping destination and base registers are
16628           used.  This option avoids generating these instructions.  This
16629           option is enabled by default when -mcpu=cortex-m3 is specified.
16630
16631       -munaligned-access
16632       -mno-unaligned-access
16633           Enables (or disables) reading and writing of 16- and 32- bit values
16634           from addresses that are not 16- or 32- bit aligned.  By default
16635           unaligned access is disabled for all pre-ARMv6, all ARMv6-M and for
16636           ARMv8-M Baseline architectures, and enabled for all other
16637           architectures.  If unaligned access is not enabled then words in
16638           packed data structures are accessed a byte at a time.
16639
16640           The ARM attribute "Tag_CPU_unaligned_access" is set in the
16641           generated object file to either true or false, depending upon the
16642           setting of this option.  If unaligned access is enabled then the
16643           preprocessor symbol "__ARM_FEATURE_UNALIGNED" is also defined.
16644
16645       -mneon-for-64bits
16646           This option is deprecated and has no effect.
16647
16648       -mslow-flash-data
16649           Assume loading data from flash is slower than fetching instruction.
16650           Therefore literal load is minimized for better performance.  This
16651           option is only supported when compiling for ARMv7 M-profile and off
16652           by default. It conflicts with -mword-relocations.
16653
16654       -masm-syntax-unified
16655           Assume inline assembler is using unified asm syntax.  The default
16656           is currently off which implies divided syntax.  This option has no
16657           impact on Thumb2. However, this may change in future releases of
16658           GCC.  Divided syntax should be considered deprecated.
16659
16660       -mrestrict-it
16661           Restricts generation of IT blocks to conform to the rules of
16662           ARMv8-A.  IT blocks can only contain a single 16-bit instruction
16663           from a select set of instructions. This option is on by default for
16664           ARMv8-A Thumb mode.
16665
16666       -mprint-tune-info
16667           Print CPU tuning information as comment in assembler file.  This is
16668           an option used only for regression testing of the compiler and not
16669           intended for ordinary use in compiling code.  This option is
16670           disabled by default.
16671
16672       -mverbose-cost-dump
16673           Enable verbose cost model dumping in the debug dump files.  This
16674           option is provided for use in debugging the compiler.
16675
16676       -mpure-code
16677           Do not allow constant data to be placed in code sections.
16678           Additionally, when compiling for ELF object format give all text
16679           sections the ELF processor-specific section attribute
16680           "SHF_ARM_PURECODE".  This option is only available when generating
16681           non-pic code for M-profile targets.
16682
16683       -mcmse
16684           Generate secure code as per the "ARMv8-M Security Extensions:
16685           Requirements on Development Tools Engineering Specification", which
16686           can be found on
16687           <http://infocenter.arm.com/help/topic/com.arm.doc.ecm0359818/ECM0359818_armv8m_security_extensions_reqs_on_dev_tools_1_0.pdf>.
16688
16689       -mfdpic
16690       -mno-fdpic
16691           Select the FDPIC ABI, which uses 64-bit function descriptors to
16692           represent pointers to functions.  When the compiler is configured
16693           for "arm-*-uclinuxfdpiceabi" targets, this option is on by default
16694           and implies -fPIE if none of the PIC/PIE-related options is
16695           provided.  On other targets, it only enables the FDPIC-specific
16696           code generation features, and the user should explicitly provide
16697           the PIC/PIE-related options as needed.
16698
16699           Note that static linking is not supported because it would still
16700           involve the dynamic linker when the program self-relocates.  If
16701           such behavior is acceptable, use -static and -Wl,-dynamic-linker
16702           options.
16703
16704           The opposite -mno-fdpic option is useful (and required) to build
16705           the Linux kernel using the same ("arm-*-uclinuxfdpiceabi")
16706           toolchain as the one used to build the userland programs.
16707
16708       AVR Options
16709
16710       These options are defined for AVR implementations:
16711
16712       -mmcu=mcu
16713           Specify Atmel AVR instruction set architectures (ISA) or MCU type.
16714
16715           The default for this option is@tie{}avr2.
16716
16717           GCC supports the following AVR devices and ISAs:
16718
16719           "avr2"
16720               "Classic" devices with up to 8@tie{}KiB of program memory.
16721               mcu@tie{}= "attiny22", "attiny26", "at90s2313", "at90s2323",
16722               "at90s2333", "at90s2343", "at90s4414", "at90s4433",
16723               "at90s4434", "at90c8534", "at90s8515", "at90s8535".
16724
16725           "avr25"
16726               "Classic" devices with up to 8@tie{}KiB of program memory and
16727               with the "MOVW" instruction.  mcu@tie{}= "attiny13",
16728               "attiny13a", "attiny24", "attiny24a", "attiny25", "attiny261",
16729               "attiny261a", "attiny2313", "attiny2313a", "attiny43u",
16730               "attiny44", "attiny44a", "attiny45", "attiny48", "attiny441",
16731               "attiny461", "attiny461a", "attiny4313", "attiny84",
16732               "attiny84a", "attiny85", "attiny87", "attiny88", "attiny828",
16733               "attiny841", "attiny861", "attiny861a", "ata5272", "ata6616c",
16734               "at86rf401".
16735
16736           "avr3"
16737               "Classic" devices with 16@tie{}KiB up to 64@tie{}KiB of program
16738               memory.  mcu@tie{}= "at76c711", "at43usb355".
16739
16740           "avr31"
16741               "Classic" devices with 128@tie{}KiB of program memory.
16742               mcu@tie{}= "atmega103", "at43usb320".
16743
16744           "avr35"
16745               "Classic" devices with 16@tie{}KiB up to 64@tie{}KiB of program
16746               memory and with the "MOVW" instruction.  mcu@tie{}=
16747               "attiny167", "attiny1634", "atmega8u2", "atmega16u2",
16748               "atmega32u2", "ata5505", "ata6617c", "ata664251", "at90usb82",
16749               "at90usb162".
16750
16751           "avr4"
16752               "Enhanced" devices with up to 8@tie{}KiB of program memory.
16753               mcu@tie{}= "atmega48", "atmega48a", "atmega48p", "atmega48pa",
16754               "atmega48pb", "atmega8", "atmega8a", "atmega8hva", "atmega88",
16755               "atmega88a", "atmega88p", "atmega88pa", "atmega88pb",
16756               "atmega8515", "atmega8535", "ata6285", "ata6286", "ata6289",
16757               "ata6612c", "at90pwm1", "at90pwm2", "at90pwm2b", "at90pwm3",
16758               "at90pwm3b", "at90pwm81".
16759
16760           "avr5"
16761               "Enhanced" devices with 16@tie{}KiB up to 64@tie{}KiB of
16762               program memory.  mcu@tie{}= "atmega16", "atmega16a",
16763               "atmega16hva", "atmega16hva2", "atmega16hvb",
16764               "atmega16hvbrevb", "atmega16m1", "atmega16u4", "atmega161",
16765               "atmega162", "atmega163", "atmega164a", "atmega164p",
16766               "atmega164pa", "atmega165", "atmega165a", "atmega165p",
16767               "atmega165pa", "atmega168", "atmega168a", "atmega168p",
16768               "atmega168pa", "atmega168pb", "atmega169", "atmega169a",
16769               "atmega169p", "atmega169pa", "atmega32", "atmega32a",
16770               "atmega32c1", "atmega32hvb", "atmega32hvbrevb", "atmega32m1",
16771               "atmega32u4", "atmega32u6", "atmega323", "atmega324a",
16772               "atmega324p", "atmega324pa", "atmega325", "atmega325a",
16773               "atmega325p", "atmega325pa", "atmega328", "atmega328p",
16774               "atmega328pb", "atmega329", "atmega329a", "atmega329p",
16775               "atmega329pa", "atmega3250", "atmega3250a", "atmega3250p",
16776               "atmega3250pa", "atmega3290", "atmega3290a", "atmega3290p",
16777               "atmega3290pa", "atmega406", "atmega64", "atmega64a",
16778               "atmega64c1", "atmega64hve", "atmega64hve2", "atmega64m1",
16779               "atmega64rfr2", "atmega640", "atmega644", "atmega644a",
16780               "atmega644p", "atmega644pa", "atmega644rfr2", "atmega645",
16781               "atmega645a", "atmega645p", "atmega649", "atmega649a",
16782               "atmega649p", "atmega6450", "atmega6450a", "atmega6450p",
16783               "atmega6490", "atmega6490a", "atmega6490p", "ata5795",
16784               "ata5790", "ata5790n", "ata5791", "ata6613c", "ata6614q",
16785               "ata5782", "ata5831", "ata8210", "ata8510", "ata5702m322",
16786               "at90pwm161", "at90pwm216", "at90pwm316", "at90can32",
16787               "at90can64", "at90scr100", "at90usb646", "at90usb647", "at94k",
16788               "m3000".
16789
16790           "avr51"
16791               "Enhanced" devices with 128@tie{}KiB of program memory.
16792               mcu@tie{}= "atmega128", "atmega128a", "atmega128rfa1",
16793               "atmega128rfr2", "atmega1280", "atmega1281", "atmega1284",
16794               "atmega1284p", "atmega1284rfr2", "at90can128", "at90usb1286",
16795               "at90usb1287".
16796
16797           "avr6"
16798               "Enhanced" devices with 3-byte PC, i.e. with more than
16799               128@tie{}KiB of program memory.  mcu@tie{}= "atmega256rfr2",
16800               "atmega2560", "atmega2561", "atmega2564rfr2".
16801
16802           "avrxmega2"
16803               "XMEGA" devices with more than 8@tie{}KiB and up to 64@tie{}KiB
16804               of program memory.  mcu@tie{}= "atxmega8e5", "atxmega16a4",
16805               "atxmega16a4u", "atxmega16c4", "atxmega16d4", "atxmega16e5",
16806               "atxmega32a4", "atxmega32a4u", "atxmega32c3", "atxmega32c4",
16807               "atxmega32d3", "atxmega32d4", "atxmega32e5".
16808
16809           "avrxmega3"
16810               "XMEGA" devices with up to 64@tie{}KiB of combined program
16811               memory and RAM, and with program memory visible in the RAM
16812               address space.  mcu@tie{}= "attiny202", "attiny204",
16813               "attiny212", "attiny214", "attiny402", "attiny404",
16814               "attiny406", "attiny412", "attiny414", "attiny416",
16815               "attiny417", "attiny804", "attiny806", "attiny807",
16816               "attiny814", "attiny816", "attiny817", "attiny1604",
16817               "attiny1606", "attiny1607", "attiny1614", "attiny1616",
16818               "attiny1617", "attiny3214", "attiny3216", "attiny3217",
16819               "atmega808", "atmega809", "atmega1608", "atmega1609",
16820               "atmega3208", "atmega3209", "atmega4808", "atmega4809".
16821
16822           "avrxmega4"
16823               "XMEGA" devices with more than 64@tie{}KiB and up to
16824               128@tie{}KiB of program memory.  mcu@tie{}= "atxmega64a3",
16825               "atxmega64a3u", "atxmega64a4u", "atxmega64b1", "atxmega64b3",
16826               "atxmega64c3", "atxmega64d3", "atxmega64d4".
16827
16828           "avrxmega5"
16829               "XMEGA" devices with more than 64@tie{}KiB and up to
16830               128@tie{}KiB of program memory and more than 64@tie{}KiB of
16831               RAM.  mcu@tie{}= "atxmega64a1", "atxmega64a1u".
16832
16833           "avrxmega6"
16834               "XMEGA" devices with more than 128@tie{}KiB of program memory.
16835               mcu@tie{}= "atxmega128a3", "atxmega128a3u", "atxmega128b1",
16836               "atxmega128b3", "atxmega128c3", "atxmega128d3", "atxmega128d4",
16837               "atxmega192a3", "atxmega192a3u", "atxmega192c3",
16838               "atxmega192d3", "atxmega256a3", "atxmega256a3b",
16839               "atxmega256a3bu", "atxmega256a3u", "atxmega256c3",
16840               "atxmega256d3", "atxmega384c3", "atxmega384d3".
16841
16842           "avrxmega7"
16843               "XMEGA" devices with more than 128@tie{}KiB of program memory
16844               and more than 64@tie{}KiB of RAM.  mcu@tie{}= "atxmega128a1",
16845               "atxmega128a1u", "atxmega128a4u".
16846
16847           "avrtiny"
16848               "TINY" Tiny core devices with 512@tie{}B up to 4@tie{}KiB of
16849               program memory.  mcu@tie{}= "attiny4", "attiny5", "attiny9",
16850               "attiny10", "attiny20", "attiny40".
16851
16852           "avr1"
16853               This ISA is implemented by the minimal AVR core and supported
16854               for assembler only.  mcu@tie{}= "attiny11", "attiny12",
16855               "attiny15", "attiny28", "at90s1200".
16856
16857       -mabsdata
16858           Assume that all data in static storage can be accessed by LDS / STS
16859           instructions.  This option has only an effect on reduced Tiny
16860           devices like ATtiny40.  See also the "absdata" AVR Variable
16861           Attributes,variable attribute.
16862
16863       -maccumulate-args
16864           Accumulate outgoing function arguments and acquire/release the
16865           needed stack space for outgoing function arguments once in function
16866           prologue/epilogue.  Without this option, outgoing arguments are
16867           pushed before calling a function and popped afterwards.
16868
16869           Popping the arguments after the function call can be expensive on
16870           AVR so that accumulating the stack space might lead to smaller
16871           executables because arguments need not be removed from the stack
16872           after such a function call.
16873
16874           This option can lead to reduced code size for functions that
16875           perform several calls to functions that get their arguments on the
16876           stack like calls to printf-like functions.
16877
16878       -mbranch-cost=cost
16879           Set the branch costs for conditional branch instructions to cost.
16880           Reasonable values for cost are small, non-negative integers. The
16881           default branch cost is 0.
16882
16883       -mcall-prologues
16884           Functions prologues/epilogues are expanded as calls to appropriate
16885           subroutines.  Code size is smaller.
16886
16887       -mdouble=bits
16888       -mlong-double=bits
16889           Set the size (in bits) of the "double" or "long double" type,
16890           respectively.  Possible values for bits are 32 and 64.  Whether or
16891           not a specific value for bits is allowed depends on the
16892           "--with-double=" and "--with-long-double=" configure options
16893           ("https://gcc.gnu.org/install/configure.html#avr"), and the same
16894           applies for the default values of the options.
16895
16896       -mgas-isr-prologues
16897           Interrupt service routines (ISRs) may use the "__gcc_isr" pseudo
16898           instruction supported by GNU Binutils.  If this option is on, the
16899           feature can still be disabled for individual ISRs by means of the
16900           AVR Function Attributes,,"no_gccisr" function attribute.  This
16901           feature is activated per default if optimization is on (but not
16902           with -Og, @pxref{Optimize Options}), and if GNU Binutils support
16903           PR21683 ("https://sourceware.org/PR21683").
16904
16905       -mint8
16906           Assume "int" to be 8-bit integer.  This affects the sizes of all
16907           types: a "char" is 1 byte, an "int" is 1 byte, a "long" is 2 bytes,
16908           and "long long" is 4 bytes.  Please note that this option does not
16909           conform to the C standards, but it results in smaller code size.
16910
16911       -mmain-is-OS_task
16912           Do not save registers in "main".  The effect is the same like
16913           attaching attribute AVR Function Attributes,,"OS_task" to "main".
16914           It is activated per default if optimization is on.
16915
16916       -mn-flash=num
16917           Assume that the flash memory has a size of num times 64@tie{}KiB.
16918
16919       -mno-interrupts
16920           Generated code is not compatible with hardware interrupts.  Code
16921           size is smaller.
16922
16923       -mrelax
16924           Try to replace "CALL" resp. "JMP" instruction by the shorter
16925           "RCALL" resp. "RJMP" instruction if applicable.  Setting -mrelax
16926           just adds the --mlink-relax option to the assembler's command line
16927           and the --relax option to the linker's command line.
16928
16929           Jump relaxing is performed by the linker because jump offsets are
16930           not known before code is located. Therefore, the assembler code
16931           generated by the compiler is the same, but the instructions in the
16932           executable may differ from instructions in the assembler code.
16933
16934           Relaxing must be turned on if linker stubs are needed, see the
16935           section on "EIND" and linker stubs below.
16936
16937       -mrmw
16938           Assume that the device supports the Read-Modify-Write instructions
16939           "XCH", "LAC", "LAS" and "LAT".
16940
16941       -mshort-calls
16942           Assume that "RJMP" and "RCALL" can target the whole program memory.
16943
16944           This option is used internally for multilib selection.  It is not
16945           an optimization option, and you don't need to set it by hand.
16946
16947       -msp8
16948           Treat the stack pointer register as an 8-bit register, i.e. assume
16949           the high byte of the stack pointer is zero.  In general, you don't
16950           need to set this option by hand.
16951
16952           This option is used internally by the compiler to select and build
16953           multilibs for architectures "avr2" and "avr25".  These
16954           architectures mix devices with and without "SPH".  For any setting
16955           other than -mmcu=avr2 or -mmcu=avr25 the compiler driver adds or
16956           removes this option from the compiler proper's command line,
16957           because the compiler then knows if the device or architecture has
16958           an 8-bit stack pointer and thus no "SPH" register or not.
16959
16960       -mstrict-X
16961           Use address register "X" in a way proposed by the hardware.  This
16962           means that "X" is only used in indirect, post-increment or pre-
16963           decrement addressing.
16964
16965           Without this option, the "X" register may be used in the same way
16966           as "Y" or "Z" which then is emulated by additional instructions.
16967           For example, loading a value with "X+const" addressing with a small
16968           non-negative "const < 64" to a register Rn is performed as
16969
16970                   adiw r26, const   ; X += const
16971                   ld   <Rn>, X        ; <Rn> = *X
16972                   sbiw r26, const   ; X -= const
16973
16974       -mtiny-stack
16975           Only change the lower 8@tie{}bits of the stack pointer.
16976
16977       -mfract-convert-truncate
16978           Allow to use truncation instead of rounding towards zero for
16979           fractional fixed-point types.
16980
16981       -nodevicelib
16982           Don't link against AVR-LibC's device specific library "lib<mcu>.a".
16983
16984       -nodevicespecs
16985           Don't add -specs=device-specs/specs-mcu to the compiler driver's
16986           command line.  The user takes responsibility for supplying the sub-
16987           processes like compiler proper, assembler and linker with
16988           appropriate command line options.  This means that the user has to
16989           supply her private device specs file by means of -specs=path-to-
16990           specs-file.  There is no more need for option -mmcu=mcu.
16991
16992           This option can also serve as a replacement for the older way of
16993           specifying custom device-specs files that needed -B some-path to
16994           point to a directory which contains a folder named "device-specs"
16995           which contains a specs file named "specs-mcu", where mcu was
16996           specified by -mmcu=mcu.
16997
16998       -Waddr-space-convert
16999           Warn about conversions between address spaces in the case where the
17000           resulting address space is not contained in the incoming address
17001           space.
17002
17003       -Wmisspelled-isr
17004           Warn if the ISR is misspelled, i.e. without __vector prefix.
17005           Enabled by default.
17006
17007       "EIND" and Devices with More Than 128 Ki Bytes of Flash
17008
17009       Pointers in the implementation are 16@tie{}bits wide.  The address of a
17010       function or label is represented as word address so that indirect jumps
17011       and calls can target any code address in the range of 64@tie{}Ki words.
17012
17013       In order to facilitate indirect jump on devices with more than
17014       128@tie{}Ki bytes of program memory space, there is a special function
17015       register called "EIND" that serves as most significant part of the
17016       target address when "EICALL" or "EIJMP" instructions are used.
17017
17018       Indirect jumps and calls on these devices are handled as follows by the
17019       compiler and are subject to some limitations:
17020
17021       *   The compiler never sets "EIND".
17022
17023       *   The compiler uses "EIND" implicitly in "EICALL"/"EIJMP"
17024           instructions or might read "EIND" directly in order to emulate an
17025           indirect call/jump by means of a "RET" instruction.
17026
17027       *   The compiler assumes that "EIND" never changes during the startup
17028           code or during the application. In particular, "EIND" is not
17029           saved/restored in function or interrupt service routine
17030           prologue/epilogue.
17031
17032       *   For indirect calls to functions and computed goto, the linker
17033           generates stubs. Stubs are jump pads sometimes also called
17034           trampolines. Thus, the indirect call/jump jumps to such a stub.
17035           The stub contains a direct jump to the desired address.
17036
17037       *   Linker relaxation must be turned on so that the linker generates
17038           the stubs correctly in all situations. See the compiler option
17039           -mrelax and the linker option --relax.  There are corner cases
17040           where the linker is supposed to generate stubs but aborts without
17041           relaxation and without a helpful error message.
17042
17043       *   The default linker script is arranged for code with "EIND = 0".  If
17044           code is supposed to work for a setup with "EIND != 0", a custom
17045           linker script has to be used in order to place the sections whose
17046           name start with ".trampolines" into the segment where "EIND" points
17047           to.
17048
17049       *   The startup code from libgcc never sets "EIND".  Notice that
17050           startup code is a blend of code from libgcc and AVR-LibC.  For the
17051           impact of AVR-LibC on "EIND", see the AVR-LibC user manual
17052           ("http://nongnu.org/avr-libc/user-manual/").
17053
17054       *   It is legitimate for user-specific startup code to set up "EIND"
17055           early, for example by means of initialization code located in
17056           section ".init3". Such code runs prior to general startup code that
17057           initializes RAM and calls constructors, but after the bit of
17058           startup code from AVR-LibC that sets "EIND" to the segment where
17059           the vector table is located.
17060
17061                   #include <avr/io.h>
17062
17063                   static void
17064                   __attribute__((section(".init3"),naked,used,no_instrument_function))
17065                   init3_set_eind (void)
17066                   {
17067                     __asm volatile ("ldi r24,pm_hh8(__trampolines_start)\n\t"
17068                                     "out %i0,r24" :: "n" (&EIND) : "r24","memory");
17069                   }
17070
17071           The "__trampolines_start" symbol is defined in the linker script.
17072
17073       *   Stubs are generated automatically by the linker if the following
17074           two conditions are met:
17075
17076           -<The address of a label is taken by means of the "gs" modifier>
17077               (short for generate stubs) like so:
17078
17079                       LDI r24, lo8(gs(<func>))
17080                       LDI r25, hi8(gs(<func>))
17081
17082           -<The final location of that label is in a code segment>
17083               outside the segment where the stubs are located.
17084
17085       *   The compiler emits such "gs" modifiers for code labels in the
17086           following situations:
17087
17088           -<Taking address of a function or code label.>
17089           -<Computed goto.>
17090           -<If prologue-save function is used, see -mcall-prologues>
17091               command-line option.
17092
17093           -<Switch/case dispatch tables. If you do not want such dispatch>
17094               tables you can specify the -fno-jump-tables command-line
17095               option.
17096
17097           -<C and C++ constructors/destructors called during
17098           startup/shutdown.>
17099           -<If the tools hit a "gs()" modifier explained above.>
17100       *   Jumping to non-symbolic addresses like so is not supported:
17101
17102                   int main (void)
17103                   {
17104                       /* Call function at word address 0x2 */
17105                       return ((int(*)(void)) 0x2)();
17106                   }
17107
17108           Instead, a stub has to be set up, i.e. the function has to be
17109           called through a symbol ("func_4" in the example):
17110
17111                   int main (void)
17112                   {
17113                       extern int func_4 (void);
17114
17115                       /* Call function at byte address 0x4 */
17116                       return func_4();
17117                   }
17118
17119           and the application be linked with -Wl,--defsym,func_4=0x4.
17120           Alternatively, "func_4" can be defined in the linker script.
17121
17122       Handling of the "RAMPD", "RAMPX", "RAMPY" and "RAMPZ" Special Function
17123       Registers
17124
17125       Some AVR devices support memories larger than the 64@tie{}KiB range
17126       that can be accessed with 16-bit pointers.  To access memory locations
17127       outside this 64@tie{}KiB range, the content of a "RAMP" register is
17128       used as high part of the address: The "X", "Y", "Z" address register is
17129       concatenated with the "RAMPX", "RAMPY", "RAMPZ" special function
17130       register, respectively, to get a wide address. Similarly, "RAMPD" is
17131       used together with direct addressing.
17132
17133       *   The startup code initializes the "RAMP" special function registers
17134           with zero.
17135
17136       *   If a AVR Named Address Spaces,named address space other than
17137           generic or "__flash" is used, then "RAMPZ" is set as needed before
17138           the operation.
17139
17140       *   If the device supports RAM larger than 64@tie{}KiB and the compiler
17141           needs to change "RAMPZ" to accomplish an operation, "RAMPZ" is
17142           reset to zero after the operation.
17143
17144       *   If the device comes with a specific "RAMP" register, the ISR
17145           prologue/epilogue saves/restores that SFR and initializes it with
17146           zero in case the ISR code might (implicitly) use it.
17147
17148       *   RAM larger than 64@tie{}KiB is not supported by GCC for AVR
17149           targets.  If you use inline assembler to read from locations
17150           outside the 16-bit address range and change one of the "RAMP"
17151           registers, you must reset it to zero after the access.
17152
17153       AVR Built-in Macros
17154
17155       GCC defines several built-in macros so that the user code can test for
17156       the presence or absence of features.  Almost any of the following
17157       built-in macros are deduced from device capabilities and thus triggered
17158       by the -mmcu= command-line option.
17159
17160       For even more AVR-specific built-in macros see AVR Named Address Spaces
17161       and AVR Built-in Functions.
17162
17163       "__AVR_ARCH__"
17164           Build-in macro that resolves to a decimal number that identifies
17165           the architecture and depends on the -mmcu=mcu option.  Possible
17166           values are:
17167
17168           2, 25, 3, 31, 35, 4, 5, 51, 6
17169
17170           for mcu="avr2", "avr25", "avr3", "avr31", "avr35", "avr4", "avr5",
17171           "avr51", "avr6",
17172
17173           respectively and
17174
17175           100, 102, 103, 104, 105, 106, 107
17176
17177           for mcu="avrtiny", "avrxmega2", "avrxmega3", "avrxmega4",
17178           "avrxmega5", "avrxmega6", "avrxmega7", respectively.  If mcu
17179           specifies a device, this built-in macro is set accordingly. For
17180           example, with -mmcu=atmega8 the macro is defined to 4.
17181
17182       "__AVR_Device__"
17183           Setting -mmcu=device defines this built-in macro which reflects the
17184           device's name. For example, -mmcu=atmega8 defines the built-in
17185           macro "__AVR_ATmega8__", -mmcu=attiny261a defines
17186           "__AVR_ATtiny261A__", etc.
17187
17188           The built-in macros' names follow the scheme "__AVR_Device__" where
17189           Device is the device name as from the AVR user manual. The
17190           difference between Device in the built-in macro and device in
17191           -mmcu=device is that the latter is always lowercase.
17192
17193           If device is not a device but only a core architecture like avr51,
17194           this macro is not defined.
17195
17196       "__AVR_DEVICE_NAME__"
17197           Setting -mmcu=device defines this built-in macro to the device's
17198           name. For example, with -mmcu=atmega8 the macro is defined to
17199           "atmega8".
17200
17201           If device is not a device but only a core architecture like avr51,
17202           this macro is not defined.
17203
17204       "__AVR_XMEGA__"
17205           The device / architecture belongs to the XMEGA family of devices.
17206
17207       "__AVR_HAVE_ELPM__"
17208           The device has the "ELPM" instruction.
17209
17210       "__AVR_HAVE_ELPMX__"
17211           The device has the "ELPM Rn,Z" and "ELPM Rn,Z+" instructions.
17212
17213       "__AVR_HAVE_MOVW__"
17214           The device has the "MOVW" instruction to perform 16-bit register-
17215           register moves.
17216
17217       "__AVR_HAVE_LPMX__"
17218           The device has the "LPM Rn,Z" and "LPM Rn,Z+" instructions.
17219
17220       "__AVR_HAVE_MUL__"
17221           The device has a hardware multiplier.
17222
17223       "__AVR_HAVE_JMP_CALL__"
17224           The device has the "JMP" and "CALL" instructions.  This is the case
17225           for devices with more than 8@tie{}KiB of program memory.
17226
17227       "__AVR_HAVE_EIJMP_EICALL__"
17228       "__AVR_3_BYTE_PC__"
17229           The device has the "EIJMP" and "EICALL" instructions.  This is the
17230           case for devices with more than 128@tie{}KiB of program memory.
17231           This also means that the program counter (PC) is 3@tie{}bytes wide.
17232
17233       "__AVR_2_BYTE_PC__"
17234           The program counter (PC) is 2@tie{}bytes wide. This is the case for
17235           devices with up to 128@tie{}KiB of program memory.
17236
17237       "__AVR_HAVE_8BIT_SP__"
17238       "__AVR_HAVE_16BIT_SP__"
17239           The stack pointer (SP) register is treated as 8-bit respectively
17240           16-bit register by the compiler.  The definition of these macros is
17241           affected by -mtiny-stack.
17242
17243       "__AVR_HAVE_SPH__"
17244       "__AVR_SP8__"
17245           The device has the SPH (high part of stack pointer) special
17246           function register or has an 8-bit stack pointer, respectively.  The
17247           definition of these macros is affected by -mmcu= and in the cases
17248           of -mmcu=avr2 and -mmcu=avr25 also by -msp8.
17249
17250       "__AVR_HAVE_RAMPD__"
17251       "__AVR_HAVE_RAMPX__"
17252       "__AVR_HAVE_RAMPY__"
17253       "__AVR_HAVE_RAMPZ__"
17254           The device has the "RAMPD", "RAMPX", "RAMPY", "RAMPZ" special
17255           function register, respectively.
17256
17257       "__NO_INTERRUPTS__"
17258           This macro reflects the -mno-interrupts command-line option.
17259
17260       "__AVR_ERRATA_SKIP__"
17261       "__AVR_ERRATA_SKIP_JMP_CALL__"
17262           Some AVR devices (AT90S8515, ATmega103) must not skip 32-bit
17263           instructions because of a hardware erratum.  Skip instructions are
17264           "SBRS", "SBRC", "SBIS", "SBIC" and "CPSE".  The second macro is
17265           only defined if "__AVR_HAVE_JMP_CALL__" is also set.
17266
17267       "__AVR_ISA_RMW__"
17268           The device has Read-Modify-Write instructions (XCH, LAC, LAS and
17269           LAT).
17270
17271       "__AVR_SFR_OFFSET__=offset"
17272           Instructions that can address I/O special function registers
17273           directly like "IN", "OUT", "SBI", etc. may use a different address
17274           as if addressed by an instruction to access RAM like "LD" or "STS".
17275           This offset depends on the device architecture and has to be
17276           subtracted from the RAM address in order to get the respective
17277           I/O@tie{}address.
17278
17279       "__AVR_SHORT_CALLS__"
17280           The -mshort-calls command line option is set.
17281
17282       "__AVR_PM_BASE_ADDRESS__=addr"
17283           Some devices support reading from flash memory by means of "LD*"
17284           instructions.  The flash memory is seen in the data address space
17285           at an offset of "__AVR_PM_BASE_ADDRESS__".  If this macro is not
17286           defined, this feature is not available.  If defined, the address
17287           space is linear and there is no need to put ".rodata" into RAM.
17288           This is handled by the default linker description file, and is
17289           currently available for "avrtiny" and "avrxmega3".  Even more
17290           convenient, there is no need to use address spaces like "__flash"
17291           or features like attribute "progmem" and "pgm_read_*".
17292
17293       "__WITH_AVRLIBC__"
17294           The compiler is configured to be used together with AVR-Libc.  See
17295           the --with-avrlibc configure option.
17296
17297       "__HAVE_DOUBLE_MULTILIB__"
17298           Defined if -mdouble= acts as a multilib option.
17299
17300       "__HAVE_DOUBLE32__"
17301       "__HAVE_DOUBLE64__"
17302           Defined if the compiler supports 32-bit double resp. 64-bit double.
17303           The actual layout is specified by option -mdouble=.
17304
17305       "__DEFAULT_DOUBLE__"
17306           The size in bits of "double" if -mdouble= is not set.  To test the
17307           layout of "double" in a program, use the built-in macro
17308           "__SIZEOF_DOUBLE__".
17309
17310       "__HAVE_LONG_DOUBLE32__"
17311       "__HAVE_LONG_DOUBLE64__"
17312       "__HAVE_LONG_DOUBLE_MULTILIB__"
17313       "__DEFAULT_LONG_DOUBLE__"
17314           Same as above, but for "long double" instead of "double".
17315
17316       "__WITH_DOUBLE_COMPARISON__"
17317           Reflects the "--with-double-comparison={tristate|bool|libf7}"
17318           configure option ("https://gcc.gnu.org/install/configure.html#avr")
17319           and is defined to 2 or 3.
17320
17321       "__WITH_LIBF7_LIBGCC__"
17322       "__WITH_LIBF7_MATH__"
17323       "__WITH_LIBF7_MATH_SYMBOLS__"
17324           Reflects the "--with-libf7={libgcc|math|math-symbols}"
17325           configure option
17326           ("https://gcc.gnu.org/install/configure.html#avr").
17327
17328       Blackfin Options
17329
17330       -mcpu=cpu[-sirevision]
17331           Specifies the name of the target Blackfin processor.  Currently,
17332           cpu can be one of bf512, bf514, bf516, bf518, bf522, bf523, bf524,
17333           bf525, bf526, bf527, bf531, bf532, bf533, bf534, bf536, bf537,
17334           bf538, bf539, bf542, bf544, bf547, bf548, bf549, bf542m, bf544m,
17335           bf547m, bf548m, bf549m, bf561, bf592.
17336
17337           The optional sirevision specifies the silicon revision of the
17338           target Blackfin processor.  Any workarounds available for the
17339           targeted silicon revision are enabled.  If sirevision is none, no
17340           workarounds are enabled.  If sirevision is any, all workarounds for
17341           the targeted processor are enabled.  The "__SILICON_REVISION__"
17342           macro is defined to two hexadecimal digits representing the major
17343           and minor numbers in the silicon revision.  If sirevision is none,
17344           the "__SILICON_REVISION__" is not defined.  If sirevision is any,
17345           the "__SILICON_REVISION__" is defined to be 0xffff.  If this
17346           optional sirevision is not used, GCC assumes the latest known
17347           silicon revision of the targeted Blackfin processor.
17348
17349           GCC defines a preprocessor macro for the specified cpu.  For the
17350           bfin-elf toolchain, this option causes the hardware BSP provided by
17351           libgloss to be linked in if -msim is not given.
17352
17353           Without this option, bf532 is used as the processor by default.
17354
17355           Note that support for bf561 is incomplete.  For bf561, only the
17356           preprocessor macro is defined.
17357
17358       -msim
17359           Specifies that the program will be run on the simulator.  This
17360           causes the simulator BSP provided by libgloss to be linked in.
17361           This option has effect only for bfin-elf toolchain.  Certain other
17362           options, such as -mid-shared-library and -mfdpic, imply -msim.
17363
17364       -momit-leaf-frame-pointer
17365           Don't keep the frame pointer in a register for leaf functions.
17366           This avoids the instructions to save, set up and restore frame
17367           pointers and makes an extra register available in leaf functions.
17368
17369       -mspecld-anomaly
17370           When enabled, the compiler ensures that the generated code does not
17371           contain speculative loads after jump instructions. If this option
17372           is used, "__WORKAROUND_SPECULATIVE_LOADS" is defined.
17373
17374       -mno-specld-anomaly
17375           Don't generate extra code to prevent speculative loads from
17376           occurring.
17377
17378       -mcsync-anomaly
17379           When enabled, the compiler ensures that the generated code does not
17380           contain CSYNC or SSYNC instructions too soon after conditional
17381           branches.  If this option is used, "__WORKAROUND_SPECULATIVE_SYNCS"
17382           is defined.
17383
17384       -mno-csync-anomaly
17385           Don't generate extra code to prevent CSYNC or SSYNC instructions
17386           from occurring too soon after a conditional branch.
17387
17388       -mlow64k
17389           When enabled, the compiler is free to take advantage of the
17390           knowledge that the entire program fits into the low 64k of memory.
17391
17392       -mno-low64k
17393           Assume that the program is arbitrarily large.  This is the default.
17394
17395       -mstack-check-l1
17396           Do stack checking using information placed into L1 scratchpad
17397           memory by the uClinux kernel.
17398
17399       -mid-shared-library
17400           Generate code that supports shared libraries via the library ID
17401           method.  This allows for execute in place and shared libraries in
17402           an environment without virtual memory management.  This option
17403           implies -fPIC.  With a bfin-elf target, this option implies -msim.
17404
17405       -mno-id-shared-library
17406           Generate code that doesn't assume ID-based shared libraries are
17407           being used.  This is the default.
17408
17409       -mleaf-id-shared-library
17410           Generate code that supports shared libraries via the library ID
17411           method, but assumes that this library or executable won't link
17412           against any other ID shared libraries.  That allows the compiler to
17413           use faster code for jumps and calls.
17414
17415       -mno-leaf-id-shared-library
17416           Do not assume that the code being compiled won't link against any
17417           ID shared libraries.  Slower code is generated for jump and call
17418           insns.
17419
17420       -mshared-library-id=n
17421           Specifies the identification number of the ID-based shared library
17422           being compiled.  Specifying a value of 0 generates more compact
17423           code; specifying other values forces the allocation of that number
17424           to the current library but is no more space- or time-efficient than
17425           omitting this option.
17426
17427       -msep-data
17428           Generate code that allows the data segment to be located in a
17429           different area of memory from the text segment.  This allows for
17430           execute in place in an environment without virtual memory
17431           management by eliminating relocations against the text section.
17432
17433       -mno-sep-data
17434           Generate code that assumes that the data segment follows the text
17435           segment.  This is the default.
17436
17437       -mlong-calls
17438       -mno-long-calls
17439           Tells the compiler to perform function calls by first loading the
17440           address of the function into a register and then performing a
17441           subroutine call on this register.  This switch is needed if the
17442           target function lies outside of the 24-bit addressing range of the
17443           offset-based version of subroutine call instruction.
17444
17445           This feature is not enabled by default.  Specifying -mno-long-calls
17446           restores the default behavior.  Note these switches have no effect
17447           on how the compiler generates code to handle function calls via
17448           function pointers.
17449
17450       -mfast-fp
17451           Link with the fast floating-point library. This library relaxes
17452           some of the IEEE floating-point standard's rules for checking
17453           inputs against Not-a-Number (NAN), in the interest of performance.
17454
17455       -minline-plt
17456           Enable inlining of PLT entries in function calls to functions that
17457           are not known to bind locally.  It has no effect without -mfdpic.
17458
17459       -mmulticore
17460           Build a standalone application for multicore Blackfin processors.
17461           This option causes proper start files and link scripts supporting
17462           multicore to be used, and defines the macro "__BFIN_MULTICORE".  It
17463           can only be used with -mcpu=bf561[-sirevision].
17464
17465           This option can be used with -mcorea or -mcoreb, which selects the
17466           one-application-per-core programming model.  Without -mcorea or
17467           -mcoreb, the single-application/dual-core programming model is
17468           used. In this model, the main function of Core B should be named as
17469           "coreb_main".
17470
17471           If this option is not used, the single-core application programming
17472           model is used.
17473
17474       -mcorea
17475           Build a standalone application for Core A of BF561 when using the
17476           one-application-per-core programming model. Proper start files and
17477           link scripts are used to support Core A, and the macro
17478           "__BFIN_COREA" is defined.  This option can only be used in
17479           conjunction with -mmulticore.
17480
17481       -mcoreb
17482           Build a standalone application for Core B of BF561 when using the
17483           one-application-per-core programming model. Proper start files and
17484           link scripts are used to support Core B, and the macro
17485           "__BFIN_COREB" is defined. When this option is used, "coreb_main"
17486           should be used instead of "main".  This option can only be used in
17487           conjunction with -mmulticore.
17488
17489       -msdram
17490           Build a standalone application for SDRAM. Proper start files and
17491           link scripts are used to put the application into SDRAM, and the
17492           macro "__BFIN_SDRAM" is defined.  The loader should initialize
17493           SDRAM before loading the application.
17494
17495       -micplb
17496           Assume that ICPLBs are enabled at run time.  This has an effect on
17497           certain anomaly workarounds.  For Linux targets, the default is to
17498           assume ICPLBs are enabled; for standalone applications the default
17499           is off.
17500
17501       C6X Options
17502
17503       -march=name
17504           This specifies the name of the target architecture.  GCC uses this
17505           name to determine what kind of instructions it can emit when
17506           generating assembly code.  Permissible names are: c62x, c64x,
17507           c64x+, c67x, c67x+, c674x.
17508
17509       -mbig-endian
17510           Generate code for a big-endian target.
17511
17512       -mlittle-endian
17513           Generate code for a little-endian target.  This is the default.
17514
17515       -msim
17516           Choose startup files and linker script suitable for the simulator.
17517
17518       -msdata=default
17519           Put small global and static data in the ".neardata" section, which
17520           is pointed to by register "B14".  Put small uninitialized global
17521           and static data in the ".bss" section, which is adjacent to the
17522           ".neardata" section.  Put small read-only data into the ".rodata"
17523           section.  The corresponding sections used for large pieces of data
17524           are ".fardata", ".far" and ".const".
17525
17526       -msdata=all
17527           Put all data, not just small objects, into the sections reserved
17528           for small data, and use addressing relative to the "B14" register
17529           to access them.
17530
17531       -msdata=none
17532           Make no use of the sections reserved for small data, and use
17533           absolute addresses to access all data.  Put all initialized global
17534           and static data in the ".fardata" section, and all uninitialized
17535           data in the ".far" section.  Put all constant data into the
17536           ".const" section.
17537
17538       CRIS Options
17539
17540       These options are defined specifically for the CRIS ports.
17541
17542       -march=architecture-type
17543       -mcpu=architecture-type
17544           Generate code for the specified architecture.  The choices for
17545           architecture-type are v3, v8 and v10 for respectively ETRAX 4,
17546           ETRAX 100, and ETRAX 100 LX.  Default is v0 except for cris-axis-
17547           linux-gnu, where the default is v10.
17548
17549       -mtune=architecture-type
17550           Tune to architecture-type everything applicable about the generated
17551           code, except for the ABI and the set of available instructions.
17552           The choices for architecture-type are the same as for
17553           -march=architecture-type.
17554
17555       -mmax-stack-frame=n
17556           Warn when the stack frame of a function exceeds n bytes.
17557
17558       -metrax4
17559       -metrax100
17560           The options -metrax4 and -metrax100 are synonyms for -march=v3 and
17561           -march=v8 respectively.
17562
17563       -mmul-bug-workaround
17564       -mno-mul-bug-workaround
17565           Work around a bug in the "muls" and "mulu" instructions for CPU
17566           models where it applies.  This option is active by default.
17567
17568       -mpdebug
17569           Enable CRIS-specific verbose debug-related information in the
17570           assembly code.  This option also has the effect of turning off the
17571           #NO_APP formatted-code indicator to the assembler at the beginning
17572           of the assembly file.
17573
17574       -mcc-init
17575           Do not use condition-code results from previous instruction; always
17576           emit compare and test instructions before use of condition codes.
17577
17578       -mno-side-effects
17579           Do not emit instructions with side effects in addressing modes
17580           other than post-increment.
17581
17582       -mstack-align
17583       -mno-stack-align
17584       -mdata-align
17585       -mno-data-align
17586       -mconst-align
17587       -mno-const-align
17588           These options (no- options) arrange (eliminate arrangements) for
17589           the stack frame, individual data and constants to be aligned for
17590           the maximum single data access size for the chosen CPU model.  The
17591           default is to arrange for 32-bit alignment.  ABI details such as
17592           structure layout are not affected by these options.
17593
17594       -m32-bit
17595       -m16-bit
17596       -m8-bit
17597           Similar to the stack- data- and const-align options above, these
17598           options arrange for stack frame, writable data and constants to all
17599           be 32-bit, 16-bit or 8-bit aligned.  The default is 32-bit
17600           alignment.
17601
17602       -mno-prologue-epilogue
17603       -mprologue-epilogue
17604           With -mno-prologue-epilogue, the normal function prologue and
17605           epilogue which set up the stack frame are omitted and no return
17606           instructions or return sequences are generated in the code.  Use
17607           this option only together with visual inspection of the compiled
17608           code: no warnings or errors are generated when call-saved registers
17609           must be saved, or storage for local variables needs to be
17610           allocated.
17611
17612       -mno-gotplt
17613       -mgotplt
17614           With -fpic and -fPIC, don't generate (do generate) instruction
17615           sequences that load addresses for functions from the PLT part of
17616           the GOT rather than (traditional on other architectures) calls to
17617           the PLT.  The default is -mgotplt.
17618
17619       -melf
17620           Legacy no-op option only recognized with the cris-axis-elf and
17621           cris-axis-linux-gnu targets.
17622
17623       -mlinux
17624           Legacy no-op option only recognized with the cris-axis-linux-gnu
17625           target.
17626
17627       -sim
17628           This option, recognized for the cris-axis-elf, arranges to link
17629           with input-output functions from a simulator library.  Code,
17630           initialized data and zero-initialized data are allocated
17631           consecutively.
17632
17633       -sim2
17634           Like -sim, but pass linker options to locate initialized data at
17635           0x40000000 and zero-initialized data at 0x80000000.
17636
17637       CR16 Options
17638
17639       These options are defined specifically for the CR16 ports.
17640
17641       -mmac
17642           Enable the use of multiply-accumulate instructions. Disabled by
17643           default.
17644
17645       -mcr16cplus
17646       -mcr16c
17647           Generate code for CR16C or CR16C+ architecture. CR16C+ architecture
17648           is default.
17649
17650       -msim
17651           Links the library libsim.a which is in compatible with simulator.
17652           Applicable to ELF compiler only.
17653
17654       -mint32
17655           Choose integer type as 32-bit wide.
17656
17657       -mbit-ops
17658           Generates "sbit"/"cbit" instructions for bit manipulations.
17659
17660       -mdata-model=model
17661           Choose a data model. The choices for model are near, far or medium.
17662           medium is default.  However, far is not valid with -mcr16c, as the
17663           CR16C architecture does not support the far data model.
17664
17665       C-SKY Options
17666
17667       GCC supports these options when compiling for C-SKY V2 processors.
17668
17669       -march=arch
17670           Specify the C-SKY target architecture.  Valid values for arch are:
17671           ck801, ck802, ck803, ck807, and ck810.  The default is ck810.
17672
17673       -mcpu=cpu
17674           Specify the C-SKY target processor.  Valid values for cpu are:
17675           ck801, ck801t, ck802, ck802t, ck802j, ck803, ck803h, ck803t,
17676           ck803ht, ck803f, ck803fh, ck803e, ck803eh, ck803et, ck803eht,
17677           ck803ef, ck803efh, ck803ft, ck803eft, ck803efht, ck803r1, ck803hr1,
17678           ck803tr1, ck803htr1, ck803fr1, ck803fhr1, ck803er1, ck803ehr1,
17679           ck803etr1, ck803ehtr1, ck803efr1, ck803efhr1, ck803ftr1,
17680           ck803eftr1, ck803efhtr1, ck803s, ck803st, ck803se, ck803sf,
17681           ck803sef, ck803seft, ck807e, ck807ef, ck807, ck807f, ck810e,
17682           ck810et, ck810ef, ck810eft, ck810, ck810v, ck810f, ck810t, ck810fv,
17683           ck810tv, ck810ft, and ck810ftv.
17684
17685       -mbig-endian
17686       -EB
17687       -mlittle-endian
17688       -EL Select big- or little-endian code.  The default is little-endian.
17689
17690       -mhard-float
17691       -msoft-float
17692           Select hardware or software floating-point implementations.  The
17693           default is soft float.
17694
17695       -mdouble-float
17696       -mno-double-float
17697           When -mhard-float is in effect, enable generation of double-
17698           precision float instructions.  This is the default except when
17699           compiling for CK803.
17700
17701       -mfdivdu
17702       -mno-fdivdu
17703           When -mhard-float is in effect, enable generation of "frecipd",
17704           "fsqrtd", and "fdivd" instructions.  This is the default except
17705           when compiling for CK803.
17706
17707       -mfpu=fpu
17708           Select the floating-point processor.  This option can only be used
17709           with -mhard-float.  Values for fpu are fpv2_sf (equivalent to
17710           -mno-double-float -mno-fdivdu), fpv2 (-mdouble-float -mno-divdu),
17711           and fpv2_divd (-mdouble-float -mdivdu).
17712
17713       -melrw
17714       -mno-elrw
17715           Enable the extended "lrw" instruction.  This option defaults to on
17716           for CK801 and off otherwise.
17717
17718       -mistack
17719       -mno-istack
17720           Enable interrupt stack instructions; the default is off.
17721
17722           The -mistack option is required to handle the "interrupt" and "isr"
17723           function attributes.
17724
17725       -mmp
17726           Enable multiprocessor instructions; the default is off.
17727
17728       -mcp
17729           Enable coprocessor instructions; the default is off.
17730
17731       -mcache
17732           Enable coprocessor instructions; the default is off.
17733
17734       -msecurity
17735           Enable C-SKY security instructions; the default is off.
17736
17737       -mtrust
17738           Enable C-SKY trust instructions; the default is off.
17739
17740       -mdsp
17741       -medsp
17742       -mvdsp
17743           Enable C-SKY DSP, Enhanced DSP, or Vector DSP instructions,
17744           respectively.  All of these options default to off.
17745
17746       -mdiv
17747       -mno-div
17748           Generate divide instructions.  Default is off.
17749
17750       -msmart
17751       -mno-smart
17752           Generate code for Smart Mode, using only registers numbered 0-7 to
17753           allow use of 16-bit instructions.  This option is ignored for CK801
17754           where this is the required behavior, and it defaults to on for
17755           CK802.  For other targets, the default is off.
17756
17757       -mhigh-registers
17758       -mno-high-registers
17759           Generate code using the high registers numbered 16-31.  This option
17760           is not supported on CK801, CK802, or CK803, and is enabled by
17761           default for other processors.
17762
17763       -manchor
17764       -mno-anchor
17765           Generate code using global anchor symbol addresses.
17766
17767       -mpushpop
17768       -mno-pushpop
17769           Generate code using "push" and "pop" instructions.  This option
17770           defaults to on.
17771
17772       -mmultiple-stld
17773       -mstm
17774       -mno-multiple-stld
17775       -mno-stm
17776           Generate code using "stm" and "ldm" instructions.  This option
17777           isn't supported on CK801 but is enabled by default on other
17778           processors.
17779
17780       -mconstpool
17781       -mno-constpool
17782           Create constant pools in the compiler instead of deferring it to
17783           the assembler.  This option is the default and required for correct
17784           code generation on CK801 and CK802, and is optional on other
17785           processors.
17786
17787       -mstack-size
17788       -mno-stack-size
17789           Emit ".stack_size" directives for each function in the assembly
17790           output.  This option defaults to off.
17791
17792       -mccrt
17793       -mno-ccrt
17794           Generate code for the C-SKY compiler runtime instead of libgcc.
17795           This option defaults to off.
17796
17797       -mbranch-cost=n
17798           Set the branch costs to roughly "n" instructions.  The default is
17799           1.
17800
17801       -msched-prolog
17802       -mno-sched-prolog
17803           Permit scheduling of function prologue and epilogue sequences.
17804           Using this option can result in code that is not compliant with the
17805           C-SKY V2 ABI prologue requirements and that cannot be debugged or
17806           backtraced.  It is disabled by default.
17807
17808       Darwin Options
17809
17810       These options are defined for all architectures running the Darwin
17811       operating system.
17812
17813       FSF GCC on Darwin does not create "fat" object files; it creates an
17814       object file for the single architecture that GCC was built to target.
17815       Apple's GCC on Darwin does create "fat" files if multiple -arch options
17816       are used; it does so by running the compiler or linker multiple times
17817       and joining the results together with lipo.
17818
17819       The subtype of the file created (like ppc7400 or ppc970 or i686) is
17820       determined by the flags that specify the ISA that GCC is targeting,
17821       like -mcpu or -march.  The -force_cpusubtype_ALL option can be used to
17822       override this.
17823
17824       The Darwin tools vary in their behavior when presented with an ISA
17825       mismatch.  The assembler, as, only permits instructions to be used that
17826       are valid for the subtype of the file it is generating, so you cannot
17827       put 64-bit instructions in a ppc750 object file.  The linker for shared
17828       libraries, /usr/bin/libtool, fails and prints an error if asked to
17829       create a shared library with a less restrictive subtype than its input
17830       files (for instance, trying to put a ppc970 object file in a ppc7400
17831       library).  The linker for executables, ld, quietly gives the executable
17832       the most restrictive subtype of any of its input files.
17833
17834       -Fdir
17835           Add the framework directory dir to the head of the list of
17836           directories to be searched for header files.  These directories are
17837           interleaved with those specified by -I options and are scanned in a
17838           left-to-right order.
17839
17840           A framework directory is a directory with frameworks in it.  A
17841           framework is a directory with a Headers and/or PrivateHeaders
17842           directory contained directly in it that ends in .framework.  The
17843           name of a framework is the name of this directory excluding the
17844           .framework.  Headers associated with the framework are found in one
17845           of those two directories, with Headers being searched first.  A
17846           subframework is a framework directory that is in a framework's
17847           Frameworks directory.  Includes of subframework headers can only
17848           appear in a header of a framework that contains the subframework,
17849           or in a sibling subframework header.  Two subframeworks are
17850           siblings if they occur in the same framework.  A subframework
17851           should not have the same name as a framework; a warning is issued
17852           if this is violated.  Currently a subframework cannot have
17853           subframeworks; in the future, the mechanism may be extended to
17854           support this.  The standard frameworks can be found in
17855           /System/Library/Frameworks and /Library/Frameworks.  An example
17856           include looks like "#include <Framework/header.h>", where Framework
17857           denotes the name of the framework and header.h is found in the
17858           PrivateHeaders or Headers directory.
17859
17860       -iframeworkdir
17861           Like -F except the directory is a treated as a system directory.
17862           The main difference between this -iframework and -F is that with
17863           -iframework the compiler does not warn about constructs contained
17864           within header files found via dir.  This option is valid only for
17865           the C family of languages.
17866
17867       -gused
17868           Emit debugging information for symbols that are used.  For stabs
17869           debugging format, this enables -feliminate-unused-debug-symbols.
17870           This is by default ON.
17871
17872       -gfull
17873           Emit debugging information for all symbols and types.
17874
17875       -mmacosx-version-min=version
17876           The earliest version of MacOS X that this executable will run on is
17877           version.  Typical values of version include 10.1, 10.2, and 10.3.9.
17878
17879           If the compiler was built to use the system's headers by default,
17880           then the default for this option is the system version on which the
17881           compiler is running, otherwise the default is to make choices that
17882           are compatible with as many systems and code bases as possible.
17883
17884       -mkernel
17885           Enable kernel development mode.  The -mkernel option sets -static,
17886           -fno-common, -fno-use-cxa-atexit, -fno-exceptions,
17887           -fno-non-call-exceptions, -fapple-kext, -fno-weak and -fno-rtti
17888           where applicable.  This mode also sets -mno-altivec, -msoft-float,
17889           -fno-builtin and -mlong-branch for PowerPC targets.
17890
17891       -mone-byte-bool
17892           Override the defaults for "bool" so that "sizeof(bool)==1".  By
17893           default "sizeof(bool)" is 4 when compiling for Darwin/PowerPC and 1
17894           when compiling for Darwin/x86, so this option has no effect on x86.
17895
17896           Warning: The -mone-byte-bool switch causes GCC to generate code
17897           that is not binary compatible with code generated without that
17898           switch.  Using this switch may require recompiling all other
17899           modules in a program, including system libraries.  Use this switch
17900           to conform to a non-default data model.
17901
17902       -mfix-and-continue
17903       -ffix-and-continue
17904       -findirect-data
17905           Generate code suitable for fast turnaround development, such as to
17906           allow GDB to dynamically load .o files into already-running
17907           programs.  -findirect-data and -ffix-and-continue are provided for
17908           backwards compatibility.
17909
17910       -all_load
17911           Loads all members of static archive libraries.  See man ld(1) for
17912           more information.
17913
17914       -arch_errors_fatal
17915           Cause the errors having to do with files that have the wrong
17916           architecture to be fatal.
17917
17918       -bind_at_load
17919           Causes the output file to be marked such that the dynamic linker
17920           will bind all undefined references when the file is loaded or
17921           launched.
17922
17923       -bundle
17924           Produce a Mach-o bundle format file.  See man ld(1) for more
17925           information.
17926
17927       -bundle_loader executable
17928           This option specifies the executable that will load the build
17929           output file being linked.  See man ld(1) for more information.
17930
17931       -dynamiclib
17932           When passed this option, GCC produces a dynamic library instead of
17933           an executable when linking, using the Darwin libtool command.
17934
17935       -force_cpusubtype_ALL
17936           This causes GCC's output file to have the ALL subtype, instead of
17937           one controlled by the -mcpu or -march option.
17938
17939       -allowable_client  client_name
17940       -client_name
17941       -compatibility_version
17942       -current_version
17943       -dead_strip
17944       -dependency-file
17945       -dylib_file
17946       -dylinker_install_name
17947       -dynamic
17948       -exported_symbols_list
17949       -filelist
17950       -flat_namespace
17951       -force_flat_namespace
17952       -headerpad_max_install_names
17953       -image_base
17954       -init
17955       -install_name
17956       -keep_private_externs
17957       -multi_module
17958       -multiply_defined
17959       -multiply_defined_unused
17960       -noall_load
17961       -no_dead_strip_inits_and_terms
17962       -nofixprebinding
17963       -nomultidefs
17964       -noprebind
17965       -noseglinkedit
17966       -pagezero_size
17967       -prebind
17968       -prebind_all_twolevel_modules
17969       -private_bundle
17970       -read_only_relocs
17971       -sectalign
17972       -sectobjectsymbols
17973       -whyload
17974       -seg1addr
17975       -sectcreate
17976       -sectobjectsymbols
17977       -sectorder
17978       -segaddr
17979       -segs_read_only_addr
17980       -segs_read_write_addr
17981       -seg_addr_table
17982       -seg_addr_table_filename
17983       -seglinkedit
17984       -segprot
17985       -segs_read_only_addr
17986       -segs_read_write_addr
17987       -single_module
17988       -static
17989       -sub_library
17990       -sub_umbrella
17991       -twolevel_namespace
17992       -umbrella
17993       -undefined
17994       -unexported_symbols_list
17995       -weak_reference_mismatches
17996       -whatsloaded
17997           These options are passed to the Darwin linker.  The Darwin linker
17998           man page describes them in detail.
17999
18000       DEC Alpha Options
18001
18002       These -m options are defined for the DEC Alpha implementations:
18003
18004       -mno-soft-float
18005       -msoft-float
18006           Use (do not use) the hardware floating-point instructions for
18007           floating-point operations.  When -msoft-float is specified,
18008           functions in libgcc.a are used to perform floating-point
18009           operations.  Unless they are replaced by routines that emulate the
18010           floating-point operations, or compiled in such a way as to call
18011           such emulations routines, these routines issue floating-point
18012           operations.   If you are compiling for an Alpha without floating-
18013           point operations, you must ensure that the library is built so as
18014           not to call them.
18015
18016           Note that Alpha implementations without floating-point operations
18017           are required to have floating-point registers.
18018
18019       -mfp-reg
18020       -mno-fp-regs
18021           Generate code that uses (does not use) the floating-point register
18022           set.  -mno-fp-regs implies -msoft-float.  If the floating-point
18023           register set is not used, floating-point operands are passed in
18024           integer registers as if they were integers and floating-point
18025           results are passed in $0 instead of $f0.  This is a non-standard
18026           calling sequence, so any function with a floating-point argument or
18027           return value called by code compiled with -mno-fp-regs must also be
18028           compiled with that option.
18029
18030           A typical use of this option is building a kernel that does not
18031           use, and hence need not save and restore, any floating-point
18032           registers.
18033
18034       -mieee
18035           The Alpha architecture implements floating-point hardware optimized
18036           for maximum performance.  It is mostly compliant with the IEEE
18037           floating-point standard.  However, for full compliance, software
18038           assistance is required.  This option generates code fully IEEE-
18039           compliant code except that the inexact-flag is not maintained (see
18040           below).  If this option is turned on, the preprocessor macro
18041           "_IEEE_FP" is defined during compilation.  The resulting code is
18042           less efficient but is able to correctly support denormalized
18043           numbers and exceptional IEEE values such as not-a-number and
18044           plus/minus infinity.  Other Alpha compilers call this option
18045           -ieee_with_no_inexact.
18046
18047       -mieee-with-inexact
18048           This is like -mieee except the generated code also maintains the
18049           IEEE inexact-flag.  Turning on this option causes the generated
18050           code to implement fully-compliant IEEE math.  In addition to
18051           "_IEEE_FP", "_IEEE_FP_EXACT" is defined as a preprocessor macro.
18052           On some Alpha implementations the resulting code may execute
18053           significantly slower than the code generated by default.  Since
18054           there is very little code that depends on the inexact-flag, you
18055           should normally not specify this option.  Other Alpha compilers
18056           call this option -ieee_with_inexact.
18057
18058       -mfp-trap-mode=trap-mode
18059           This option controls what floating-point related traps are enabled.
18060           Other Alpha compilers call this option -fptm trap-mode.  The trap
18061           mode can be set to one of four values:
18062
18063           n   This is the default (normal) setting.  The only traps that are
18064               enabled are the ones that cannot be disabled in software (e.g.,
18065               division by zero trap).
18066
18067           u   In addition to the traps enabled by n, underflow traps are
18068               enabled as well.
18069
18070           su  Like u, but the instructions are marked to be safe for software
18071               completion (see Alpha architecture manual for details).
18072
18073           sui Like su, but inexact traps are enabled as well.
18074
18075       -mfp-rounding-mode=rounding-mode
18076           Selects the IEEE rounding mode.  Other Alpha compilers call this
18077           option -fprm rounding-mode.  The rounding-mode can be one of:
18078
18079           n   Normal IEEE rounding mode.  Floating-point numbers are rounded
18080               towards the nearest machine number or towards the even machine
18081               number in case of a tie.
18082
18083           m   Round towards minus infinity.
18084
18085           c   Chopped rounding mode.  Floating-point numbers are rounded
18086               towards zero.
18087
18088           d   Dynamic rounding mode.  A field in the floating-point control
18089               register (fpcr, see Alpha architecture reference manual)
18090               controls the rounding mode in effect.  The C library
18091               initializes this register for rounding towards plus infinity.
18092               Thus, unless your program modifies the fpcr, d corresponds to
18093               round towards plus infinity.
18094
18095       -mtrap-precision=trap-precision
18096           In the Alpha architecture, floating-point traps are imprecise.
18097           This means without software assistance it is impossible to recover
18098           from a floating trap and program execution normally needs to be
18099           terminated.  GCC can generate code that can assist operating system
18100           trap handlers in determining the exact location that caused a
18101           floating-point trap.  Depending on the requirements of an
18102           application, different levels of precisions can be selected:
18103
18104           p   Program precision.  This option is the default and means a trap
18105               handler can only identify which program caused a floating-point
18106               exception.
18107
18108           f   Function precision.  The trap handler can determine the
18109               function that caused a floating-point exception.
18110
18111           i   Instruction precision.  The trap handler can determine the
18112               exact instruction that caused a floating-point exception.
18113
18114           Other Alpha compilers provide the equivalent options called
18115           -scope_safe and -resumption_safe.
18116
18117       -mieee-conformant
18118           This option marks the generated code as IEEE conformant.  You must
18119           not use this option unless you also specify -mtrap-precision=i and
18120           either -mfp-trap-mode=su or -mfp-trap-mode=sui.  Its only effect is
18121           to emit the line .eflag 48 in the function prologue of the
18122           generated assembly file.
18123
18124       -mbuild-constants
18125           Normally GCC examines a 32- or 64-bit integer constant to see if it
18126           can construct it from smaller constants in two or three
18127           instructions.  If it cannot, it outputs the constant as a literal
18128           and generates code to load it from the data segment at run time.
18129
18130           Use this option to require GCC to construct all integer constants
18131           using code, even if it takes more instructions (the maximum is
18132           six).
18133
18134           You typically use this option to build a shared library dynamic
18135           loader.  Itself a shared library, it must relocate itself in memory
18136           before it can find the variables and constants in its own data
18137           segment.
18138
18139       -mbwx
18140       -mno-bwx
18141       -mcix
18142       -mno-cix
18143       -mfix
18144       -mno-fix
18145       -mmax
18146       -mno-max
18147           Indicate whether GCC should generate code to use the optional BWX,
18148           CIX, FIX and MAX instruction sets.  The default is to use the
18149           instruction sets supported by the CPU type specified via -mcpu=
18150           option or that of the CPU on which GCC was built if none is
18151           specified.
18152
18153       -mfloat-vax
18154       -mfloat-ieee
18155           Generate code that uses (does not use) VAX F and G floating-point
18156           arithmetic instead of IEEE single and double precision.
18157
18158       -mexplicit-relocs
18159       -mno-explicit-relocs
18160           Older Alpha assemblers provided no way to generate symbol
18161           relocations except via assembler macros.  Use of these macros does
18162           not allow optimal instruction scheduling.  GNU binutils as of
18163           version 2.12 supports a new syntax that allows the compiler to
18164           explicitly mark which relocations should apply to which
18165           instructions.  This option is mostly useful for debugging, as GCC
18166           detects the capabilities of the assembler when it is built and sets
18167           the default accordingly.
18168
18169       -msmall-data
18170       -mlarge-data
18171           When -mexplicit-relocs is in effect, static data is accessed via
18172           gp-relative relocations.  When -msmall-data is used, objects 8
18173           bytes long or smaller are placed in a small data area (the ".sdata"
18174           and ".sbss" sections) and are accessed via 16-bit relocations off
18175           of the $gp register.  This limits the size of the small data area
18176           to 64KB, but allows the variables to be directly accessed via a
18177           single instruction.
18178
18179           The default is -mlarge-data.  With this option the data area is
18180           limited to just below 2GB.  Programs that require more than 2GB of
18181           data must use "malloc" or "mmap" to allocate the data in the heap
18182           instead of in the program's data segment.
18183
18184           When generating code for shared libraries, -fpic implies
18185           -msmall-data and -fPIC implies -mlarge-data.
18186
18187       -msmall-text
18188       -mlarge-text
18189           When -msmall-text is used, the compiler assumes that the code of
18190           the entire program (or shared library) fits in 4MB, and is thus
18191           reachable with a branch instruction.  When -msmall-data is used,
18192           the compiler can assume that all local symbols share the same $gp
18193           value, and thus reduce the number of instructions required for a
18194           function call from 4 to 1.
18195
18196           The default is -mlarge-text.
18197
18198       -mcpu=cpu_type
18199           Set the instruction set and instruction scheduling parameters for
18200           machine type cpu_type.  You can specify either the EV style name or
18201           the corresponding chip number.  GCC supports scheduling parameters
18202           for the EV4, EV5 and EV6 family of processors and chooses the
18203           default values for the instruction set from the processor you
18204           specify.  If you do not specify a processor type, GCC defaults to
18205           the processor on which the compiler was built.
18206
18207           Supported values for cpu_type are
18208
18209           ev4
18210           ev45
18211           21064
18212               Schedules as an EV4 and has no instruction set extensions.
18213
18214           ev5
18215           21164
18216               Schedules as an EV5 and has no instruction set extensions.
18217
18218           ev56
18219           21164a
18220               Schedules as an EV5 and supports the BWX extension.
18221
18222           pca56
18223           21164pc
18224           21164PC
18225               Schedules as an EV5 and supports the BWX and MAX extensions.
18226
18227           ev6
18228           21264
18229               Schedules as an EV6 and supports the BWX, FIX, and MAX
18230               extensions.
18231
18232           ev67
18233           21264a
18234               Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX
18235               extensions.
18236
18237           Native toolchains also support the value native, which selects the
18238           best architecture option for the host processor.  -mcpu=native has
18239           no effect if GCC does not recognize the processor.
18240
18241       -mtune=cpu_type
18242           Set only the instruction scheduling parameters for machine type
18243           cpu_type.  The instruction set is not changed.
18244
18245           Native toolchains also support the value native, which selects the
18246           best architecture option for the host processor.  -mtune=native has
18247           no effect if GCC does not recognize the processor.
18248
18249       -mmemory-latency=time
18250           Sets the latency the scheduler should assume for typical memory
18251           references as seen by the application.  This number is highly
18252           dependent on the memory access patterns used by the application and
18253           the size of the external cache on the machine.
18254
18255           Valid options for time are
18256
18257           number
18258               A decimal number representing clock cycles.
18259
18260           L1
18261           L2
18262           L3
18263           main
18264               The compiler contains estimates of the number of clock cycles
18265               for "typical" EV4 & EV5 hardware for the Level 1, 2 & 3 caches
18266               (also called Dcache, Scache, and Bcache), as well as to main
18267               memory.  Note that L3 is only valid for EV5.
18268
18269       eBPF Options
18270
18271       -mframe-limit=bytes
18272           This specifies the hard limit for frame sizes, in bytes.
18273           Currently, the value that can be specified should be less than or
18274           equal to 32767.  Defaults to whatever limit is imposed by the
18275           version of the Linux kernel targeted.
18276
18277       -mkernel=version
18278           This specifies the minimum version of the kernel that will run the
18279           compiled program.  GCC uses this version to determine which
18280           instructions to use, what kernel helpers to allow, etc.  Currently,
18281           version can be one of 4.0, 4.1, 4.2, 4.3, 4.4, 4.5, 4.6, 4.7, 4.8,
18282           4.9, 4.10, 4.11, 4.12, 4.13, 4.14, 4.15, 4.16, 4.17, 4.18, 4.19,
18283           4.20, 5.0, 5.1, 5.2, latest and native.
18284
18285       -mbig-endian
18286           Generate code for a big-endian target.
18287
18288       -mlittle-endian
18289           Generate code for a little-endian target.  This is the default.
18290
18291       FR30 Options
18292
18293       These options are defined specifically for the FR30 port.
18294
18295       -msmall-model
18296           Use the small address space model.  This can produce smaller code,
18297           but it does assume that all symbolic values and addresses fit into
18298           a 20-bit range.
18299
18300       -mno-lsim
18301           Assume that runtime support has been provided and so there is no
18302           need to include the simulator library (libsim.a) on the linker
18303           command line.
18304
18305       FT32 Options
18306
18307       These options are defined specifically for the FT32 port.
18308
18309       -msim
18310           Specifies that the program will be run on the simulator.  This
18311           causes an alternate runtime startup and library to be linked.  You
18312           must not use this option when generating programs that will run on
18313           real hardware; you must provide your own runtime library for
18314           whatever I/O functions are needed.
18315
18316       -mlra
18317           Enable Local Register Allocation.  This is still experimental for
18318           FT32, so by default the compiler uses standard reload.
18319
18320       -mnodiv
18321           Do not use div and mod instructions.
18322
18323       -mft32b
18324           Enable use of the extended instructions of the FT32B processor.
18325
18326       -mcompress
18327           Compress all code using the Ft32B code compression scheme.
18328
18329       -mnopm
18330           Do not generate code that reads program memory.
18331
18332       FRV Options
18333
18334       -mgpr-32
18335           Only use the first 32 general-purpose registers.
18336
18337       -mgpr-64
18338           Use all 64 general-purpose registers.
18339
18340       -mfpr-32
18341           Use only the first 32 floating-point registers.
18342
18343       -mfpr-64
18344           Use all 64 floating-point registers.
18345
18346       -mhard-float
18347           Use hardware instructions for floating-point operations.
18348
18349       -msoft-float
18350           Use library routines for floating-point operations.
18351
18352       -malloc-cc
18353           Dynamically allocate condition code registers.
18354
18355       -mfixed-cc
18356           Do not try to dynamically allocate condition code registers, only
18357           use "icc0" and "fcc0".
18358
18359       -mdword
18360           Change ABI to use double word insns.
18361
18362       -mno-dword
18363           Do not use double word instructions.
18364
18365       -mdouble
18366           Use floating-point double instructions.
18367
18368       -mno-double
18369           Do not use floating-point double instructions.
18370
18371       -mmedia
18372           Use media instructions.
18373
18374       -mno-media
18375           Do not use media instructions.
18376
18377       -mmuladd
18378           Use multiply and add/subtract instructions.
18379
18380       -mno-muladd
18381           Do not use multiply and add/subtract instructions.
18382
18383       -mfdpic
18384           Select the FDPIC ABI, which uses function descriptors to represent
18385           pointers to functions.  Without any PIC/PIE-related options, it
18386           implies -fPIE.  With -fpic or -fpie, it assumes GOT entries and
18387           small data are within a 12-bit range from the GOT base address;
18388           with -fPIC or -fPIE, GOT offsets are computed with 32 bits.  With a
18389           bfin-elf target, this option implies -msim.
18390
18391       -minline-plt
18392           Enable inlining of PLT entries in function calls to functions that
18393           are not known to bind locally.  It has no effect without -mfdpic.
18394           It's enabled by default if optimizing for speed and compiling for
18395           shared libraries (i.e., -fPIC or -fpic), or when an optimization
18396           option such as -O3 or above is present in the command line.
18397
18398       -mTLS
18399           Assume a large TLS segment when generating thread-local code.
18400
18401       -mtls
18402           Do not assume a large TLS segment when generating thread-local
18403           code.
18404
18405       -mgprel-ro
18406           Enable the use of "GPREL" relocations in the FDPIC ABI for data
18407           that is known to be in read-only sections.  It's enabled by
18408           default, except for -fpic or -fpie: even though it may help make
18409           the global offset table smaller, it trades 1 instruction for 4.
18410           With -fPIC or -fPIE, it trades 3 instructions for 4, one of which
18411           may be shared by multiple symbols, and it avoids the need for a GOT
18412           entry for the referenced symbol, so it's more likely to be a win.
18413           If it is not, -mno-gprel-ro can be used to disable it.
18414
18415       -multilib-library-pic
18416           Link with the (library, not FD) pic libraries.  It's implied by
18417           -mlibrary-pic, as well as by -fPIC and -fpic without -mfdpic.  You
18418           should never have to use it explicitly.
18419
18420       -mlinked-fp
18421           Follow the EABI requirement of always creating a frame pointer
18422           whenever a stack frame is allocated.  This option is enabled by
18423           default and can be disabled with -mno-linked-fp.
18424
18425       -mlong-calls
18426           Use indirect addressing to call functions outside the current
18427           compilation unit.  This allows the functions to be placed anywhere
18428           within the 32-bit address space.
18429
18430       -malign-labels
18431           Try to align labels to an 8-byte boundary by inserting NOPs into
18432           the previous packet.  This option only has an effect when VLIW
18433           packing is enabled.  It doesn't create new packets; it merely adds
18434           NOPs to existing ones.
18435
18436       -mlibrary-pic
18437           Generate position-independent EABI code.
18438
18439       -macc-4
18440           Use only the first four media accumulator registers.
18441
18442       -macc-8
18443           Use all eight media accumulator registers.
18444
18445       -mpack
18446           Pack VLIW instructions.
18447
18448       -mno-pack
18449           Do not pack VLIW instructions.
18450
18451       -mno-eflags
18452           Do not mark ABI switches in e_flags.
18453
18454       -mcond-move
18455           Enable the use of conditional-move instructions (default).
18456
18457           This switch is mainly for debugging the compiler and will likely be
18458           removed in a future version.
18459
18460       -mno-cond-move
18461           Disable the use of conditional-move instructions.
18462
18463           This switch is mainly for debugging the compiler and will likely be
18464           removed in a future version.
18465
18466       -mscc
18467           Enable the use of conditional set instructions (default).
18468
18469           This switch is mainly for debugging the compiler and will likely be
18470           removed in a future version.
18471
18472       -mno-scc
18473           Disable the use of conditional set instructions.
18474
18475           This switch is mainly for debugging the compiler and will likely be
18476           removed in a future version.
18477
18478       -mcond-exec
18479           Enable the use of conditional execution (default).
18480
18481           This switch is mainly for debugging the compiler and will likely be
18482           removed in a future version.
18483
18484       -mno-cond-exec
18485           Disable the use of conditional execution.
18486
18487           This switch is mainly for debugging the compiler and will likely be
18488           removed in a future version.
18489
18490       -mvliw-branch
18491           Run a pass to pack branches into VLIW instructions (default).
18492
18493           This switch is mainly for debugging the compiler and will likely be
18494           removed in a future version.
18495
18496       -mno-vliw-branch
18497           Do not run a pass to pack branches into VLIW instructions.
18498
18499           This switch is mainly for debugging the compiler and will likely be
18500           removed in a future version.
18501
18502       -mmulti-cond-exec
18503           Enable optimization of "&&" and "||" in conditional execution
18504           (default).
18505
18506           This switch is mainly for debugging the compiler and will likely be
18507           removed in a future version.
18508
18509       -mno-multi-cond-exec
18510           Disable optimization of "&&" and "||" in conditional execution.
18511
18512           This switch is mainly for debugging the compiler and will likely be
18513           removed in a future version.
18514
18515       -mnested-cond-exec
18516           Enable nested conditional execution optimizations (default).
18517
18518           This switch is mainly for debugging the compiler and will likely be
18519           removed in a future version.
18520
18521       -mno-nested-cond-exec
18522           Disable nested conditional execution optimizations.
18523
18524           This switch is mainly for debugging the compiler and will likely be
18525           removed in a future version.
18526
18527       -moptimize-membar
18528           This switch removes redundant "membar" instructions from the
18529           compiler-generated code.  It is enabled by default.
18530
18531       -mno-optimize-membar
18532           This switch disables the automatic removal of redundant "membar"
18533           instructions from the generated code.
18534
18535       -mtomcat-stats
18536           Cause gas to print out tomcat statistics.
18537
18538       -mcpu=cpu
18539           Select the processor type for which to generate code.  Possible
18540           values are frv, fr550, tomcat, fr500, fr450, fr405, fr400, fr300
18541           and simple.
18542
18543       GNU/Linux Options
18544
18545       These -m options are defined for GNU/Linux targets:
18546
18547       -mglibc
18548           Use the GNU C library.  This is the default except on
18549           *-*-linux-*uclibc*, *-*-linux-*musl* and *-*-linux-*android*
18550           targets.
18551
18552       -muclibc
18553           Use uClibc C library.  This is the default on *-*-linux-*uclibc*
18554           targets.
18555
18556       -mmusl
18557           Use the musl C library.  This is the default on *-*-linux-*musl*
18558           targets.
18559
18560       -mbionic
18561           Use Bionic C library.  This is the default on *-*-linux-*android*
18562           targets.
18563
18564       -mandroid
18565           Compile code compatible with Android platform.  This is the default
18566           on *-*-linux-*android* targets.
18567
18568           When compiling, this option enables -mbionic, -fPIC,
18569           -fno-exceptions and -fno-rtti by default.  When linking, this
18570           option makes the GCC driver pass Android-specific options to the
18571           linker.  Finally, this option causes the preprocessor macro
18572           "__ANDROID__" to be defined.
18573
18574       -tno-android-cc
18575           Disable compilation effects of -mandroid, i.e., do not enable
18576           -mbionic, -fPIC, -fno-exceptions and -fno-rtti by default.
18577
18578       -tno-android-ld
18579           Disable linking effects of -mandroid, i.e., pass standard Linux
18580           linking options to the linker.
18581
18582       H8/300 Options
18583
18584       These -m options are defined for the H8/300 implementations:
18585
18586       -mrelax
18587           Shorten some address references at link time, when possible; uses
18588           the linker option -relax.
18589
18590       -mh Generate code for the H8/300H.
18591
18592       -ms Generate code for the H8S.
18593
18594       -mn Generate code for the H8S and H8/300H in the normal mode.  This
18595           switch must be used either with -mh or -ms.
18596
18597       -ms2600
18598           Generate code for the H8S/2600.  This switch must be used with -ms.
18599
18600       -mexr
18601           Extended registers are stored on stack before execution of function
18602           with monitor attribute. Default option is -mexr.  This option is
18603           valid only for H8S targets.
18604
18605       -mno-exr
18606           Extended registers are not stored on stack before execution of
18607           function with monitor attribute. Default option is -mno-exr.  This
18608           option is valid only for H8S targets.
18609
18610       -mint32
18611           Make "int" data 32 bits by default.
18612
18613       -malign-300
18614           On the H8/300H and H8S, use the same alignment rules as for the
18615           H8/300.  The default for the H8/300H and H8S is to align longs and
18616           floats on 4-byte boundaries.  -malign-300 causes them to be aligned
18617           on 2-byte boundaries.  This option has no effect on the H8/300.
18618
18619       HPPA Options
18620
18621       These -m options are defined for the HPPA family of computers:
18622
18623       -march=architecture-type
18624           Generate code for the specified architecture.  The choices for
18625           architecture-type are 1.0 for PA 1.0, 1.1 for PA 1.1, and 2.0 for
18626           PA 2.0 processors.  Refer to /usr/lib/sched.models on an HP-UX
18627           system to determine the proper architecture option for your
18628           machine.  Code compiled for lower numbered architectures runs on
18629           higher numbered architectures, but not the other way around.
18630
18631       -mpa-risc-1-0
18632       -mpa-risc-1-1
18633       -mpa-risc-2-0
18634           Synonyms for -march=1.0, -march=1.1, and -march=2.0 respectively.
18635
18636       -mcaller-copies
18637           The caller copies function arguments passed by hidden reference.
18638           This option should be used with care as it is not compatible with
18639           the default 32-bit runtime.  However, only aggregates larger than
18640           eight bytes are passed by hidden reference and the option provides
18641           better compatibility with OpenMP.
18642
18643       -mjump-in-delay
18644           This option is ignored and provided for compatibility purposes
18645           only.
18646
18647       -mdisable-fpregs
18648           Prevent floating-point registers from being used in any manner.
18649           This is necessary for compiling kernels that perform lazy context
18650           switching of floating-point registers.  If you use this option and
18651           attempt to perform floating-point operations, the compiler aborts.
18652
18653       -mdisable-indexing
18654           Prevent the compiler from using indexing address modes.  This
18655           avoids some rather obscure problems when compiling MIG generated
18656           code under MACH.
18657
18658       -mno-space-regs
18659           Generate code that assumes the target has no space registers.  This
18660           allows GCC to generate faster indirect calls and use unscaled index
18661           address modes.
18662
18663           Such code is suitable for level 0 PA systems and kernels.
18664
18665       -mfast-indirect-calls
18666           Generate code that assumes calls never cross space boundaries.
18667           This allows GCC to emit code that performs faster indirect calls.
18668
18669           This option does not work in the presence of shared libraries or
18670           nested functions.
18671
18672       -mfixed-range=register-range
18673           Generate code treating the given register range as fixed registers.
18674           A fixed register is one that the register allocator cannot use.
18675           This is useful when compiling kernel code.  A register range is
18676           specified as two registers separated by a dash.  Multiple register
18677           ranges can be specified separated by a comma.
18678
18679       -mlong-load-store
18680           Generate 3-instruction load and store sequences as sometimes
18681           required by the HP-UX 10 linker.  This is equivalent to the +k
18682           option to the HP compilers.
18683
18684       -mportable-runtime
18685           Use the portable calling conventions proposed by HP for ELF
18686           systems.
18687
18688       -mgas
18689           Enable the use of assembler directives only GAS understands.
18690
18691       -mschedule=cpu-type
18692           Schedule code according to the constraints for the machine type
18693           cpu-type.  The choices for cpu-type are 700 7100, 7100LC, 7200,
18694           7300 and 8000.  Refer to /usr/lib/sched.models on an HP-UX system
18695           to determine the proper scheduling option for your machine.  The
18696           default scheduling is 8000.
18697
18698       -mlinker-opt
18699           Enable the optimization pass in the HP-UX linker.  Note this makes
18700           symbolic debugging impossible.  It also triggers a bug in the HP-UX
18701           8 and HP-UX 9 linkers in which they give bogus error messages when
18702           linking some programs.
18703
18704       -msoft-float
18705           Generate output containing library calls for floating point.
18706           Warning: the requisite libraries are not available for all HPPA
18707           targets.  Normally the facilities of the machine's usual C compiler
18708           are used, but this cannot be done directly in cross-compilation.
18709           You must make your own arrangements to provide suitable library
18710           functions for cross-compilation.
18711
18712           -msoft-float changes the calling convention in the output file;
18713           therefore, it is only useful if you compile all of a program with
18714           this option.  In particular, you need to compile libgcc.a, the
18715           library that comes with GCC, with -msoft-float in order for this to
18716           work.
18717
18718       -msio
18719           Generate the predefine, "_SIO", for server IO.  The default is
18720           -mwsio.  This generates the predefines, "__hp9000s700",
18721           "__hp9000s700__" and "_WSIO", for workstation IO.  These options
18722           are available under HP-UX and HI-UX.
18723
18724       -mgnu-ld
18725           Use options specific to GNU ld.  This passes -shared to ld when
18726           building a shared library.  It is the default when GCC is
18727           configured, explicitly or implicitly, with the GNU linker.  This
18728           option does not affect which ld is called; it only changes what
18729           parameters are passed to that ld.  The ld that is called is
18730           determined by the --with-ld configure option, GCC's program search
18731           path, and finally by the user's PATH.  The linker used by GCC can
18732           be printed using which `gcc -print-prog-name=ld`.  This option is
18733           only available on the 64-bit HP-UX GCC, i.e. configured with
18734           hppa*64*-*-hpux*.
18735
18736       -mhp-ld
18737           Use options specific to HP ld.  This passes -b to ld when building
18738           a shared library and passes +Accept TypeMismatch to ld on all
18739           links.  It is the default when GCC is configured, explicitly or
18740           implicitly, with the HP linker.  This option does not affect which
18741           ld is called; it only changes what parameters are passed to that
18742           ld.  The ld that is called is determined by the --with-ld configure
18743           option, GCC's program search path, and finally by the user's PATH.
18744           The linker used by GCC can be printed using which `gcc
18745           -print-prog-name=ld`.  This option is only available on the 64-bit
18746           HP-UX GCC, i.e. configured with hppa*64*-*-hpux*.
18747
18748       -mlong-calls
18749           Generate code that uses long call sequences.  This ensures that a
18750           call is always able to reach linker generated stubs.  The default
18751           is to generate long calls only when the distance from the call site
18752           to the beginning of the function or translation unit, as the case
18753           may be, exceeds a predefined limit set by the branch type being
18754           used.  The limits for normal calls are 7,600,000 and 240,000 bytes,
18755           respectively for the PA 2.0 and PA 1.X architectures.  Sibcalls are
18756           always limited at 240,000 bytes.
18757
18758           Distances are measured from the beginning of functions when using
18759           the -ffunction-sections option, or when using the -mgas and
18760           -mno-portable-runtime options together under HP-UX with the SOM
18761           linker.
18762
18763           It is normally not desirable to use this option as it degrades
18764           performance.  However, it may be useful in large applications,
18765           particularly when partial linking is used to build the application.
18766
18767           The types of long calls used depends on the capabilities of the
18768           assembler and linker, and the type of code being generated.  The
18769           impact on systems that support long absolute calls, and long pic
18770           symbol-difference or pc-relative calls should be relatively small.
18771           However, an indirect call is used on 32-bit ELF systems in pic code
18772           and it is quite long.
18773
18774       -munix=unix-std
18775           Generate compiler predefines and select a startfile for the
18776           specified UNIX standard.  The choices for unix-std are 93, 95 and
18777           98.  93 is supported on all HP-UX versions.  95 is available on HP-
18778           UX 10.10 and later.  98 is available on HP-UX 11.11 and later.  The
18779           default values are 93 for HP-UX 10.00, 95 for HP-UX 10.10 though to
18780           11.00, and 98 for HP-UX 11.11 and later.
18781
18782           -munix=93 provides the same predefines as GCC 3.3 and 3.4.
18783           -munix=95 provides additional predefines for "XOPEN_UNIX" and
18784           "_XOPEN_SOURCE_EXTENDED", and the startfile unix95.o.  -munix=98
18785           provides additional predefines for "_XOPEN_UNIX",
18786           "_XOPEN_SOURCE_EXTENDED", "_INCLUDE__STDC_A1_SOURCE" and
18787           "_INCLUDE_XOPEN_SOURCE_500", and the startfile unix98.o.
18788
18789           It is important to note that this option changes the interfaces for
18790           various library routines.  It also affects the operational behavior
18791           of the C library.  Thus, extreme care is needed in using this
18792           option.
18793
18794           Library code that is intended to operate with more than one UNIX
18795           standard must test, set and restore the variable
18796           "__xpg4_extended_mask" as appropriate.  Most GNU software doesn't
18797           provide this capability.
18798
18799       -nolibdld
18800           Suppress the generation of link options to search libdld.sl when
18801           the -static option is specified on HP-UX 10 and later.
18802
18803       -static
18804           The HP-UX implementation of setlocale in libc has a dependency on
18805           libdld.sl.  There isn't an archive version of libdld.sl.  Thus,
18806           when the -static option is specified, special link options are
18807           needed to resolve this dependency.
18808
18809           On HP-UX 10 and later, the GCC driver adds the necessary options to
18810           link with libdld.sl when the -static option is specified.  This
18811           causes the resulting binary to be dynamic.  On the 64-bit port, the
18812           linkers generate dynamic binaries by default in any case.  The
18813           -nolibdld option can be used to prevent the GCC driver from adding
18814           these link options.
18815
18816       -threads
18817           Add support for multithreading with the dce thread library under
18818           HP-UX.  This option sets flags for both the preprocessor and
18819           linker.
18820
18821       IA-64 Options
18822
18823       These are the -m options defined for the Intel IA-64 architecture.
18824
18825       -mbig-endian
18826           Generate code for a big-endian target.  This is the default for HP-
18827           UX.
18828
18829       -mlittle-endian
18830           Generate code for a little-endian target.  This is the default for
18831           AIX5 and GNU/Linux.
18832
18833       -mgnu-as
18834       -mno-gnu-as
18835           Generate (or don't) code for the GNU assembler.  This is the
18836           default.
18837
18838       -mgnu-ld
18839       -mno-gnu-ld
18840           Generate (or don't) code for the GNU linker.  This is the default.
18841
18842       -mno-pic
18843           Generate code that does not use a global pointer register.  The
18844           result is not position independent code, and violates the IA-64
18845           ABI.
18846
18847       -mvolatile-asm-stop
18848       -mno-volatile-asm-stop
18849           Generate (or don't) a stop bit immediately before and after
18850           volatile asm statements.
18851
18852       -mregister-names
18853       -mno-register-names
18854           Generate (or don't) in, loc, and out register names for the stacked
18855           registers.  This may make assembler output more readable.
18856
18857       -mno-sdata
18858       -msdata
18859           Disable (or enable) optimizations that use the small data section.
18860           This may be useful for working around optimizer bugs.
18861
18862       -mconstant-gp
18863           Generate code that uses a single constant global pointer value.
18864           This is useful when compiling kernel code.
18865
18866       -mauto-pic
18867           Generate code that is self-relocatable.  This implies
18868           -mconstant-gp.  This is useful when compiling firmware code.
18869
18870       -minline-float-divide-min-latency
18871           Generate code for inline divides of floating-point values using the
18872           minimum latency algorithm.
18873
18874       -minline-float-divide-max-throughput
18875           Generate code for inline divides of floating-point values using the
18876           maximum throughput algorithm.
18877
18878       -mno-inline-float-divide
18879           Do not generate inline code for divides of floating-point values.
18880
18881       -minline-int-divide-min-latency
18882           Generate code for inline divides of integer values using the
18883           minimum latency algorithm.
18884
18885       -minline-int-divide-max-throughput
18886           Generate code for inline divides of integer values using the
18887           maximum throughput algorithm.
18888
18889       -mno-inline-int-divide
18890           Do not generate inline code for divides of integer values.
18891
18892       -minline-sqrt-min-latency
18893           Generate code for inline square roots using the minimum latency
18894           algorithm.
18895
18896       -minline-sqrt-max-throughput
18897           Generate code for inline square roots using the maximum throughput
18898           algorithm.
18899
18900       -mno-inline-sqrt
18901           Do not generate inline code for "sqrt".
18902
18903       -mfused-madd
18904       -mno-fused-madd
18905           Do (don't) generate code that uses the fused multiply/add or
18906           multiply/subtract instructions.  The default is to use these
18907           instructions.
18908
18909       -mno-dwarf2-asm
18910       -mdwarf2-asm
18911           Don't (or do) generate assembler code for the DWARF line number
18912           debugging info.  This may be useful when not using the GNU
18913           assembler.
18914
18915       -mearly-stop-bits
18916       -mno-early-stop-bits
18917           Allow stop bits to be placed earlier than immediately preceding the
18918           instruction that triggered the stop bit.  This can improve
18919           instruction scheduling, but does not always do so.
18920
18921       -mfixed-range=register-range
18922           Generate code treating the given register range as fixed registers.
18923           A fixed register is one that the register allocator cannot use.
18924           This is useful when compiling kernel code.  A register range is
18925           specified as two registers separated by a dash.  Multiple register
18926           ranges can be specified separated by a comma.
18927
18928       -mtls-size=tls-size
18929           Specify bit size of immediate TLS offsets.  Valid values are 14,
18930           22, and 64.
18931
18932       -mtune=cpu-type
18933           Tune the instruction scheduling for a particular CPU, Valid values
18934           are itanium, itanium1, merced, itanium2, and mckinley.
18935
18936       -milp32
18937       -mlp64
18938           Generate code for a 32-bit or 64-bit environment.  The 32-bit
18939           environment sets int, long and pointer to 32 bits.  The 64-bit
18940           environment sets int to 32 bits and long and pointer to 64 bits.
18941           These are HP-UX specific flags.
18942
18943       -mno-sched-br-data-spec
18944       -msched-br-data-spec
18945           (Dis/En)able data speculative scheduling before reload.  This
18946           results in generation of "ld.a" instructions and the corresponding
18947           check instructions ("ld.c" / "chk.a").  The default setting is
18948           disabled.
18949
18950       -msched-ar-data-spec
18951       -mno-sched-ar-data-spec
18952           (En/Dis)able data speculative scheduling after reload.  This
18953           results in generation of "ld.a" instructions and the corresponding
18954           check instructions ("ld.c" / "chk.a").  The default setting is
18955           enabled.
18956
18957       -mno-sched-control-spec
18958       -msched-control-spec
18959           (Dis/En)able control speculative scheduling.  This feature is
18960           available only during region scheduling (i.e. before reload).  This
18961           results in generation of the "ld.s" instructions and the
18962           corresponding check instructions "chk.s".  The default setting is
18963           disabled.
18964
18965       -msched-br-in-data-spec
18966       -mno-sched-br-in-data-spec
18967           (En/Dis)able speculative scheduling of the instructions that are
18968           dependent on the data speculative loads before reload.  This is
18969           effective only with -msched-br-data-spec enabled.  The default
18970           setting is enabled.
18971
18972       -msched-ar-in-data-spec
18973       -mno-sched-ar-in-data-spec
18974           (En/Dis)able speculative scheduling of the instructions that are
18975           dependent on the data speculative loads after reload.  This is
18976           effective only with -msched-ar-data-spec enabled.  The default
18977           setting is enabled.
18978
18979       -msched-in-control-spec
18980       -mno-sched-in-control-spec
18981           (En/Dis)able speculative scheduling of the instructions that are
18982           dependent on the control speculative loads.  This is effective only
18983           with -msched-control-spec enabled.  The default setting is enabled.
18984
18985       -mno-sched-prefer-non-data-spec-insns
18986       -msched-prefer-non-data-spec-insns
18987           If enabled, data-speculative instructions are chosen for schedule
18988           only if there are no other choices at the moment.  This makes the
18989           use of the data speculation much more conservative.  The default
18990           setting is disabled.
18991
18992       -mno-sched-prefer-non-control-spec-insns
18993       -msched-prefer-non-control-spec-insns
18994           If enabled, control-speculative instructions are chosen for
18995           schedule only if there are no other choices at the moment.  This
18996           makes the use of the control speculation much more conservative.
18997           The default setting is disabled.
18998
18999       -mno-sched-count-spec-in-critical-path
19000       -msched-count-spec-in-critical-path
19001           If enabled, speculative dependencies are considered during
19002           computation of the instructions priorities.  This makes the use of
19003           the speculation a bit more conservative.  The default setting is
19004           disabled.
19005
19006       -msched-spec-ldc
19007           Use a simple data speculation check.  This option is on by default.
19008
19009       -msched-control-spec-ldc
19010           Use a simple check for control speculation.  This option is on by
19011           default.
19012
19013       -msched-stop-bits-after-every-cycle
19014           Place a stop bit after every cycle when scheduling.  This option is
19015           on by default.
19016
19017       -msched-fp-mem-deps-zero-cost
19018           Assume that floating-point stores and loads are not likely to cause
19019           a conflict when placed into the same instruction group.  This
19020           option is disabled by default.
19021
19022       -msel-sched-dont-check-control-spec
19023           Generate checks for control speculation in selective scheduling.
19024           This flag is disabled by default.
19025
19026       -msched-max-memory-insns=max-insns
19027           Limit on the number of memory insns per instruction group, giving
19028           lower priority to subsequent memory insns attempting to schedule in
19029           the same instruction group. Frequently useful to prevent cache bank
19030           conflicts.  The default value is 1.
19031
19032       -msched-max-memory-insns-hard-limit
19033           Makes the limit specified by msched-max-memory-insns a hard limit,
19034           disallowing more than that number in an instruction group.
19035           Otherwise, the limit is "soft", meaning that non-memory operations
19036           are preferred when the limit is reached, but memory operations may
19037           still be scheduled.
19038
19039       LM32 Options
19040
19041       These -m options are defined for the LatticeMico32 architecture:
19042
19043       -mbarrel-shift-enabled
19044           Enable barrel-shift instructions.
19045
19046       -mdivide-enabled
19047           Enable divide and modulus instructions.
19048
19049       -mmultiply-enabled
19050           Enable multiply instructions.
19051
19052       -msign-extend-enabled
19053           Enable sign extend instructions.
19054
19055       -muser-enabled
19056           Enable user-defined instructions.
19057
19058       M32C Options
19059
19060       -mcpu=name
19061           Select the CPU for which code is generated.  name may be one of r8c
19062           for the R8C/Tiny series, m16c for the M16C (up to /60) series,
19063           m32cm for the M16C/80 series, or m32c for the M32C/80 series.
19064
19065       -msim
19066           Specifies that the program will be run on the simulator.  This
19067           causes an alternate runtime library to be linked in which supports,
19068           for example, file I/O.  You must not use this option when
19069           generating programs that will run on real hardware; you must
19070           provide your own runtime library for whatever I/O functions are
19071           needed.
19072
19073       -memregs=number
19074           Specifies the number of memory-based pseudo-registers GCC uses
19075           during code generation.  These pseudo-registers are used like real
19076           registers, so there is a tradeoff between GCC's ability to fit the
19077           code into available registers, and the performance penalty of using
19078           memory instead of registers.  Note that all modules in a program
19079           must be compiled with the same value for this option.  Because of
19080           that, you must not use this option with GCC's default runtime
19081           libraries.
19082
19083       M32R/D Options
19084
19085       These -m options are defined for Renesas M32R/D architectures:
19086
19087       -m32r2
19088           Generate code for the M32R/2.
19089
19090       -m32rx
19091           Generate code for the M32R/X.
19092
19093       -m32r
19094           Generate code for the M32R.  This is the default.
19095
19096       -mmodel=small
19097           Assume all objects live in the lower 16MB of memory (so that their
19098           addresses can be loaded with the "ld24" instruction), and assume
19099           all subroutines are reachable with the "bl" instruction.  This is
19100           the default.
19101
19102           The addressability of a particular object can be set with the
19103           "model" attribute.
19104
19105       -mmodel=medium
19106           Assume objects may be anywhere in the 32-bit address space (the
19107           compiler generates "seth/add3" instructions to load their
19108           addresses), and assume all subroutines are reachable with the "bl"
19109           instruction.
19110
19111       -mmodel=large
19112           Assume objects may be anywhere in the 32-bit address space (the
19113           compiler generates "seth/add3" instructions to load their
19114           addresses), and assume subroutines may not be reachable with the
19115           "bl" instruction (the compiler generates the much slower
19116           "seth/add3/jl" instruction sequence).
19117
19118       -msdata=none
19119           Disable use of the small data area.  Variables are put into one of
19120           ".data", ".bss", or ".rodata" (unless the "section" attribute has
19121           been specified).  This is the default.
19122
19123           The small data area consists of sections ".sdata" and ".sbss".
19124           Objects may be explicitly put in the small data area with the
19125           "section" attribute using one of these sections.
19126
19127       -msdata=sdata
19128           Put small global and static data in the small data area, but do not
19129           generate special code to reference them.
19130
19131       -msdata=use
19132           Put small global and static data in the small data area, and
19133           generate special instructions to reference them.
19134
19135       -G num
19136           Put global and static objects less than or equal to num bytes into
19137           the small data or BSS sections instead of the normal data or BSS
19138           sections.  The default value of num is 8.  The -msdata option must
19139           be set to one of sdata or use for this option to have any effect.
19140
19141           All modules should be compiled with the same -G num value.
19142           Compiling with different values of num may or may not work; if it
19143           doesn't the linker gives an error message---incorrect code is not
19144           generated.
19145
19146       -mdebug
19147           Makes the M32R-specific code in the compiler display some
19148           statistics that might help in debugging programs.
19149
19150       -malign-loops
19151           Align all loops to a 32-byte boundary.
19152
19153       -mno-align-loops
19154           Do not enforce a 32-byte alignment for loops.  This is the default.
19155
19156       -missue-rate=number
19157           Issue number instructions per cycle.  number can only be 1 or 2.
19158
19159       -mbranch-cost=number
19160           number can only be 1 or 2.  If it is 1 then branches are preferred
19161           over conditional code, if it is 2, then the opposite applies.
19162
19163       -mflush-trap=number
19164           Specifies the trap number to use to flush the cache.  The default
19165           is 12.  Valid numbers are between 0 and 15 inclusive.
19166
19167       -mno-flush-trap
19168           Specifies that the cache cannot be flushed by using a trap.
19169
19170       -mflush-func=name
19171           Specifies the name of the operating system function to call to
19172           flush the cache.  The default is _flush_cache, but a function call
19173           is only used if a trap is not available.
19174
19175       -mno-flush-func
19176           Indicates that there is no OS function for flushing the cache.
19177
19178       M680x0 Options
19179
19180       These are the -m options defined for M680x0 and ColdFire processors.
19181       The default settings depend on which architecture was selected when the
19182       compiler was configured; the defaults for the most common choices are
19183       given below.
19184
19185       -march=arch
19186           Generate code for a specific M680x0 or ColdFire instruction set
19187           architecture.  Permissible values of arch for M680x0 architectures
19188           are: 68000, 68010, 68020, 68030, 68040, 68060 and cpu32.  ColdFire
19189           architectures are selected according to Freescale's ISA
19190           classification and the permissible values are: isaa, isaaplus, isab
19191           and isac.
19192
19193           GCC defines a macro "__mcfarch__" whenever it is generating code
19194           for a ColdFire target.  The arch in this macro is one of the -march
19195           arguments given above.
19196
19197           When used together, -march and -mtune select code that runs on a
19198           family of similar processors but that is optimized for a particular
19199           microarchitecture.
19200
19201       -mcpu=cpu
19202           Generate code for a specific M680x0 or ColdFire processor.  The
19203           M680x0 cpus are: 68000, 68010, 68020, 68030, 68040, 68060, 68302,
19204           68332 and cpu32.  The ColdFire cpus are given by the table below,
19205           which also classifies the CPUs into families:
19206
19207           Family : -mcpu arguments
19208           51 : 51 51ac 51ag 51cn 51em 51je 51jf 51jg 51jm 51mm 51qe 51qm
19209           5206 : 5202 5204 5206
19210           5206e : 5206e
19211           5208 : 5207 5208
19212           5211a : 5210a 5211a
19213           5213 : 5211 5212 5213
19214           5216 : 5214 5216
19215           52235 : 52230 52231 52232 52233 52234 52235
19216           5225 : 5224 5225
19217           52259 : 52252 52254 52255 52256 52258 52259
19218           5235 : 5232 5233 5234 5235 523x
19219           5249 : 5249
19220           5250 : 5250
19221           5271 : 5270 5271
19222           5272 : 5272
19223           5275 : 5274 5275
19224           5282 : 5280 5281 5282 528x
19225           53017 : 53011 53012 53013 53014 53015 53016 53017
19226           5307 : 5307
19227           5329 : 5327 5328 5329 532x
19228           5373 : 5372 5373 537x
19229           5407 : 5407
19230           5475 : 5470 5471 5472 5473 5474 5475 547x 5480 5481 5482 5483 5484
19231           5485
19232
19233           -mcpu=cpu overrides -march=arch if arch is compatible with cpu.
19234           Other combinations of -mcpu and -march are rejected.
19235
19236           GCC defines the macro "__mcf_cpu_cpu" when ColdFire target cpu is
19237           selected.  It also defines "__mcf_family_family", where the value
19238           of family is given by the table above.
19239
19240       -mtune=tune
19241           Tune the code for a particular microarchitecture within the
19242           constraints set by -march and -mcpu.  The M680x0 microarchitectures
19243           are: 68000, 68010, 68020, 68030, 68040, 68060 and cpu32.  The
19244           ColdFire microarchitectures are: cfv1, cfv2, cfv3, cfv4 and cfv4e.
19245
19246           You can also use -mtune=68020-40 for code that needs to run
19247           relatively well on 68020, 68030 and 68040 targets.  -mtune=68020-60
19248           is similar but includes 68060 targets as well.  These two options
19249           select the same tuning decisions as -m68020-40 and -m68020-60
19250           respectively.
19251
19252           GCC defines the macros "__mcarch" and "__mcarch__" when tuning for
19253           680x0 architecture arch.  It also defines "mcarch" unless either
19254           -ansi or a non-GNU -std option is used.  If GCC is tuning for a
19255           range of architectures, as selected by -mtune=68020-40 or
19256           -mtune=68020-60, it defines the macros for every architecture in
19257           the range.
19258
19259           GCC also defines the macro "__muarch__" when tuning for ColdFire
19260           microarchitecture uarch, where uarch is one of the arguments given
19261           above.
19262
19263       -m68000
19264       -mc68000
19265           Generate output for a 68000.  This is the default when the compiler
19266           is configured for 68000-based systems.  It is equivalent to
19267           -march=68000.
19268
19269           Use this option for microcontrollers with a 68000 or EC000 core,
19270           including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
19271
19272       -m68010
19273           Generate output for a 68010.  This is the default when the compiler
19274           is configured for 68010-based systems.  It is equivalent to
19275           -march=68010.
19276
19277       -m68020
19278       -mc68020
19279           Generate output for a 68020.  This is the default when the compiler
19280           is configured for 68020-based systems.  It is equivalent to
19281           -march=68020.
19282
19283       -m68030
19284           Generate output for a 68030.  This is the default when the compiler
19285           is configured for 68030-based systems.  It is equivalent to
19286           -march=68030.
19287
19288       -m68040
19289           Generate output for a 68040.  This is the default when the compiler
19290           is configured for 68040-based systems.  It is equivalent to
19291           -march=68040.
19292
19293           This option inhibits the use of 68881/68882 instructions that have
19294           to be emulated by software on the 68040.  Use this option if your
19295           68040 does not have code to emulate those instructions.
19296
19297       -m68060
19298           Generate output for a 68060.  This is the default when the compiler
19299           is configured for 68060-based systems.  It is equivalent to
19300           -march=68060.
19301
19302           This option inhibits the use of 68020 and 68881/68882 instructions
19303           that have to be emulated by software on the 68060.  Use this option
19304           if your 68060 does not have code to emulate those instructions.
19305
19306       -mcpu32
19307           Generate output for a CPU32.  This is the default when the compiler
19308           is configured for CPU32-based systems.  It is equivalent to
19309           -march=cpu32.
19310
19311           Use this option for microcontrollers with a CPU32 or CPU32+ core,
19312           including the 68330, 68331, 68332, 68333, 68334, 68336, 68340,
19313           68341, 68349 and 68360.
19314
19315       -m5200
19316           Generate output for a 520X ColdFire CPU.  This is the default when
19317           the compiler is configured for 520X-based systems.  It is
19318           equivalent to -mcpu=5206, and is now deprecated in favor of that
19319           option.
19320
19321           Use this option for microcontroller with a 5200 core, including the
19322           MCF5202, MCF5203, MCF5204 and MCF5206.
19323
19324       -m5206e
19325           Generate output for a 5206e ColdFire CPU.  The option is now
19326           deprecated in favor of the equivalent -mcpu=5206e.
19327
19328       -m528x
19329           Generate output for a member of the ColdFire 528X family.  The
19330           option is now deprecated in favor of the equivalent -mcpu=528x.
19331
19332       -m5307
19333           Generate output for a ColdFire 5307 CPU.  The option is now
19334           deprecated in favor of the equivalent -mcpu=5307.
19335
19336       -m5407
19337           Generate output for a ColdFire 5407 CPU.  The option is now
19338           deprecated in favor of the equivalent -mcpu=5407.
19339
19340       -mcfv4e
19341           Generate output for a ColdFire V4e family CPU (e.g. 547x/548x).
19342           This includes use of hardware floating-point instructions.  The
19343           option is equivalent to -mcpu=547x, and is now deprecated in favor
19344           of that option.
19345
19346       -m68020-40
19347           Generate output for a 68040, without using any of the new
19348           instructions.  This results in code that can run relatively
19349           efficiently on either a 68020/68881 or a 68030 or a 68040.  The
19350           generated code does use the 68881 instructions that are emulated on
19351           the 68040.
19352
19353           The option is equivalent to -march=68020 -mtune=68020-40.
19354
19355       -m68020-60
19356           Generate output for a 68060, without using any of the new
19357           instructions.  This results in code that can run relatively
19358           efficiently on either a 68020/68881 or a 68030 or a 68040.  The
19359           generated code does use the 68881 instructions that are emulated on
19360           the 68060.
19361
19362           The option is equivalent to -march=68020 -mtune=68020-60.
19363
19364       -mhard-float
19365       -m68881
19366           Generate floating-point instructions.  This is the default for
19367           68020 and above, and for ColdFire devices that have an FPU.  It
19368           defines the macro "__HAVE_68881__" on M680x0 targets and
19369           "__mcffpu__" on ColdFire targets.
19370
19371       -msoft-float
19372           Do not generate floating-point instructions; use library calls
19373           instead.  This is the default for 68000, 68010, and 68832 targets.
19374           It is also the default for ColdFire devices that have no FPU.
19375
19376       -mdiv
19377       -mno-div
19378           Generate (do not generate) ColdFire hardware divide and remainder
19379           instructions.  If -march is used without -mcpu, the default is "on"
19380           for ColdFire architectures and "off" for M680x0 architectures.
19381           Otherwise, the default is taken from the target CPU (either the
19382           default CPU, or the one specified by -mcpu).  For example, the
19383           default is "off" for -mcpu=5206 and "on" for -mcpu=5206e.
19384
19385           GCC defines the macro "__mcfhwdiv__" when this option is enabled.
19386
19387       -mshort
19388           Consider type "int" to be 16 bits wide, like "short int".
19389           Additionally, parameters passed on the stack are also aligned to a
19390           16-bit boundary even on targets whose API mandates promotion to
19391           32-bit.
19392
19393       -mno-short
19394           Do not consider type "int" to be 16 bits wide.  This is the
19395           default.
19396
19397       -mnobitfield
19398       -mno-bitfield
19399           Do not use the bit-field instructions.  The -m68000, -mcpu32 and
19400           -m5200 options imply -mnobitfield.
19401
19402       -mbitfield
19403           Do use the bit-field instructions.  The -m68020 option implies
19404           -mbitfield.  This is the default if you use a configuration
19405           designed for a 68020.
19406
19407       -mrtd
19408           Use a different function-calling convention, in which functions
19409           that take a fixed number of arguments return with the "rtd"
19410           instruction, which pops their arguments while returning.  This
19411           saves one instruction in the caller since there is no need to pop
19412           the arguments there.
19413
19414           This calling convention is incompatible with the one normally used
19415           on Unix, so you cannot use it if you need to call libraries
19416           compiled with the Unix compiler.
19417
19418           Also, you must provide function prototypes for all functions that
19419           take variable numbers of arguments (including "printf"); otherwise
19420           incorrect code is generated for calls to those functions.
19421
19422           In addition, seriously incorrect code results if you call a
19423           function with too many arguments.  (Normally, extra arguments are
19424           harmlessly ignored.)
19425
19426           The "rtd" instruction is supported by the 68010, 68020, 68030,
19427           68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
19428
19429           The default is -mno-rtd.
19430
19431       -malign-int
19432       -mno-align-int
19433           Control whether GCC aligns "int", "long", "long long", "float",
19434           "double", and "long double" variables on a 32-bit boundary
19435           (-malign-int) or a 16-bit boundary (-mno-align-int).  Aligning
19436           variables on 32-bit boundaries produces code that runs somewhat
19437           faster on processors with 32-bit busses at the expense of more
19438           memory.
19439
19440           Warning: if you use the -malign-int switch, GCC aligns structures
19441           containing the above types differently than most published
19442           application binary interface specifications for the m68k.
19443
19444           Use the pc-relative addressing mode of the 68000 directly, instead
19445           of using a global offset table.  At present, this option implies
19446           -fpic, allowing at most a 16-bit offset for pc-relative addressing.
19447           -fPIC is not presently supported with -mpcrel, though this could be
19448           supported for 68020 and higher processors.
19449
19450       -mno-strict-align
19451       -mstrict-align
19452           Do not (do) assume that unaligned memory references are handled by
19453           the system.
19454
19455       -msep-data
19456           Generate code that allows the data segment to be located in a
19457           different area of memory from the text segment.  This allows for
19458           execute-in-place in an environment without virtual memory
19459           management.  This option implies -fPIC.
19460
19461       -mno-sep-data
19462           Generate code that assumes that the data segment follows the text
19463           segment.  This is the default.
19464
19465       -mid-shared-library
19466           Generate code that supports shared libraries via the library ID
19467           method.  This allows for execute-in-place and shared libraries in
19468           an environment without virtual memory management.  This option
19469           implies -fPIC.
19470
19471       -mno-id-shared-library
19472           Generate code that doesn't assume ID-based shared libraries are
19473           being used.  This is the default.
19474
19475       -mshared-library-id=n
19476           Specifies the identification number of the ID-based shared library
19477           being compiled.  Specifying a value of 0 generates more compact
19478           code; specifying other values forces the allocation of that number
19479           to the current library, but is no more space- or time-efficient
19480           than omitting this option.
19481
19482       -mxgot
19483       -mno-xgot
19484           When generating position-independent code for ColdFire, generate
19485           code that works if the GOT has more than 8192 entries.  This code
19486           is larger and slower than code generated without this option.  On
19487           M680x0 processors, this option is not needed; -fPIC suffices.
19488
19489           GCC normally uses a single instruction to load values from the GOT.
19490           While this is relatively efficient, it only works if the GOT is
19491           smaller than about 64k.  Anything larger causes the linker to
19492           report an error such as:
19493
19494                   relocation truncated to fit: R_68K_GOT16O foobar
19495
19496           If this happens, you should recompile your code with -mxgot.  It
19497           should then work with very large GOTs.  However, code generated
19498           with -mxgot is less efficient, since it takes 4 instructions to
19499           fetch the value of a global symbol.
19500
19501           Note that some linkers, including newer versions of the GNU linker,
19502           can create multiple GOTs and sort GOT entries.  If you have such a
19503           linker, you should only need to use -mxgot when compiling a single
19504           object file that accesses more than 8192 GOT entries.  Very few do.
19505
19506           These options have no effect unless GCC is generating position-
19507           independent code.
19508
19509       -mlong-jump-table-offsets
19510           Use 32-bit offsets in "switch" tables.  The default is to use
19511           16-bit offsets.
19512
19513       MCore Options
19514
19515       These are the -m options defined for the Motorola M*Core processors.
19516
19517       -mhardlit
19518       -mno-hardlit
19519           Inline constants into the code stream if it can be done in two
19520           instructions or less.
19521
19522       -mdiv
19523       -mno-div
19524           Use the divide instruction.  (Enabled by default).
19525
19526       -mrelax-immediate
19527       -mno-relax-immediate
19528           Allow arbitrary-sized immediates in bit operations.
19529
19530       -mwide-bitfields
19531       -mno-wide-bitfields
19532           Always treat bit-fields as "int"-sized.
19533
19534       -m4byte-functions
19535       -mno-4byte-functions
19536           Force all functions to be aligned to a 4-byte boundary.
19537
19538       -mcallgraph-data
19539       -mno-callgraph-data
19540           Emit callgraph information.
19541
19542       -mslow-bytes
19543       -mno-slow-bytes
19544           Prefer word access when reading byte quantities.
19545
19546       -mlittle-endian
19547       -mbig-endian
19548           Generate code for a little-endian target.
19549
19550       -m210
19551       -m340
19552           Generate code for the 210 processor.
19553
19554       -mno-lsim
19555           Assume that runtime support has been provided and so omit the
19556           simulator library (libsim.a) from the linker command line.
19557
19558       -mstack-increment=size
19559           Set the maximum amount for a single stack increment operation.
19560           Large values can increase the speed of programs that contain
19561           functions that need a large amount of stack space, but they can
19562           also trigger a segmentation fault if the stack is extended too
19563           much.  The default value is 0x1000.
19564
19565       MeP Options
19566
19567       -mabsdiff
19568           Enables the "abs" instruction, which is the absolute difference
19569           between two registers.
19570
19571       -mall-opts
19572           Enables all the optional instructions---average, multiply, divide,
19573           bit operations, leading zero, absolute difference, min/max, clip,
19574           and saturation.
19575
19576       -maverage
19577           Enables the "ave" instruction, which computes the average of two
19578           registers.
19579
19580       -mbased=n
19581           Variables of size n bytes or smaller are placed in the ".based"
19582           section by default.  Based variables use the $tp register as a base
19583           register, and there is a 128-byte limit to the ".based" section.
19584
19585       -mbitops
19586           Enables the bit operation instructions---bit test ("btstm"), set
19587           ("bsetm"), clear ("bclrm"), invert ("bnotm"), and test-and-set
19588           ("tas").
19589
19590       -mc=name
19591           Selects which section constant data is placed in.  name may be
19592           tiny, near, or far.
19593
19594       -mclip
19595           Enables the "clip" instruction.  Note that -mclip is not useful
19596           unless you also provide -mminmax.
19597
19598       -mconfig=name
19599           Selects one of the built-in core configurations.  Each MeP chip has
19600           one or more modules in it; each module has a core CPU and a variety
19601           of coprocessors, optional instructions, and peripherals.  The
19602           "MeP-Integrator" tool, not part of GCC, provides these
19603           configurations through this option; using this option is the same
19604           as using all the corresponding command-line options.  The default
19605           configuration is default.
19606
19607       -mcop
19608           Enables the coprocessor instructions.  By default, this is a 32-bit
19609           coprocessor.  Note that the coprocessor is normally enabled via the
19610           -mconfig= option.
19611
19612       -mcop32
19613           Enables the 32-bit coprocessor's instructions.
19614
19615       -mcop64
19616           Enables the 64-bit coprocessor's instructions.
19617
19618       -mivc2
19619           Enables IVC2 scheduling.  IVC2 is a 64-bit VLIW coprocessor.
19620
19621       -mdc
19622           Causes constant variables to be placed in the ".near" section.
19623
19624       -mdiv
19625           Enables the "div" and "divu" instructions.
19626
19627       -meb
19628           Generate big-endian code.
19629
19630       -mel
19631           Generate little-endian code.
19632
19633       -mio-volatile
19634           Tells the compiler that any variable marked with the "io" attribute
19635           is to be considered volatile.
19636
19637       -ml Causes variables to be assigned to the ".far" section by default.
19638
19639       -mleadz
19640           Enables the "leadz" (leading zero) instruction.
19641
19642       -mm Causes variables to be assigned to the ".near" section by default.
19643
19644       -mminmax
19645           Enables the "min" and "max" instructions.
19646
19647       -mmult
19648           Enables the multiplication and multiply-accumulate instructions.
19649
19650       -mno-opts
19651           Disables all the optional instructions enabled by -mall-opts.
19652
19653       -mrepeat
19654           Enables the "repeat" and "erepeat" instructions, used for low-
19655           overhead looping.
19656
19657       -ms Causes all variables to default to the ".tiny" section.  Note that
19658           there is a 65536-byte limit to this section.  Accesses to these
19659           variables use the %gp base register.
19660
19661       -msatur
19662           Enables the saturation instructions.  Note that the compiler does
19663           not currently generate these itself, but this option is included
19664           for compatibility with other tools, like "as".
19665
19666       -msdram
19667           Link the SDRAM-based runtime instead of the default ROM-based
19668           runtime.
19669
19670       -msim
19671           Link the simulator run-time libraries.
19672
19673       -msimnovec
19674           Link the simulator runtime libraries, excluding built-in support
19675           for reset and exception vectors and tables.
19676
19677       -mtf
19678           Causes all functions to default to the ".far" section.  Without
19679           this option, functions default to the ".near" section.
19680
19681       -mtiny=n
19682           Variables that are n bytes or smaller are allocated to the ".tiny"
19683           section.  These variables use the $gp base register.  The default
19684           for this option is 4, but note that there's a 65536-byte limit to
19685           the ".tiny" section.
19686
19687       MicroBlaze Options
19688
19689       -msoft-float
19690           Use software emulation for floating point (default).
19691
19692       -mhard-float
19693           Use hardware floating-point instructions.
19694
19695       -mmemcpy
19696           Do not optimize block moves, use "memcpy".
19697
19698       -mno-clearbss
19699           This option is deprecated.  Use -fno-zero-initialized-in-bss
19700           instead.
19701
19702       -mcpu=cpu-type
19703           Use features of, and schedule code for, the given CPU.  Supported
19704           values are in the format vX.YY.Z, where X is a major version, YY is
19705           the minor version, and Z is compatibility code.  Example values are
19706           v3.00.a, v4.00.b, v5.00.a, v5.00.b, v6.00.a.
19707
19708       -mxl-soft-mul
19709           Use software multiply emulation (default).
19710
19711       -mxl-soft-div
19712           Use software emulation for divides (default).
19713
19714       -mxl-barrel-shift
19715           Use the hardware barrel shifter.
19716
19717       -mxl-pattern-compare
19718           Use pattern compare instructions.
19719
19720       -msmall-divides
19721           Use table lookup optimization for small signed integer divisions.
19722
19723       -mxl-stack-check
19724           This option is deprecated.  Use -fstack-check instead.
19725
19726       -mxl-gp-opt
19727           Use GP-relative ".sdata"/".sbss" sections.
19728
19729       -mxl-multiply-high
19730           Use multiply high instructions for high part of 32x32 multiply.
19731
19732       -mxl-float-convert
19733           Use hardware floating-point conversion instructions.
19734
19735       -mxl-float-sqrt
19736           Use hardware floating-point square root instruction.
19737
19738       -mbig-endian
19739           Generate code for a big-endian target.
19740
19741       -mlittle-endian
19742           Generate code for a little-endian target.
19743
19744       -mxl-reorder
19745           Use reorder instructions (swap and byte reversed load/store).
19746
19747       -mxl-mode-app-model
19748           Select application model app-model.  Valid models are
19749
19750           executable
19751               normal executable (default), uses startup code crt0.o.
19752
19753           -mpic-data-is-text-relative
19754               Assume that the displacement between the text and data segments
19755               is fixed at static link time.  This allows data to be
19756               referenced by offset from start of text address instead of GOT
19757               since PC-relative addressing is not supported.
19758
19759           xmdstub
19760               for use with Xilinx Microprocessor Debugger (XMD) based
19761               software intrusive debug agent called xmdstub. This uses
19762               startup file crt1.o and sets the start address of the program
19763               to 0x800.
19764
19765           bootstrap
19766               for applications that are loaded using a bootloader.  This
19767               model uses startup file crt2.o which does not contain a
19768               processor reset vector handler. This is suitable for
19769               transferring control on a processor reset to the bootloader
19770               rather than the application.
19771
19772           novectors
19773               for applications that do not require any of the MicroBlaze
19774               vectors. This option may be useful for applications running
19775               within a monitoring application. This model uses crt3.o as a
19776               startup file.
19777
19778           Option -xl-mode-app-model is a deprecated alias for -mxl-mode-app-
19779           model.
19780
19781       MIPS Options
19782
19783       -EB Generate big-endian code.
19784
19785       -EL Generate little-endian code.  This is the default for mips*el-*-*
19786           configurations.
19787
19788       -march=arch
19789           Generate code that runs on arch, which can be the name of a generic
19790           MIPS ISA, or the name of a particular processor.  The ISA names
19791           are: mips1, mips2, mips3, mips4, mips32, mips32r2, mips32r3,
19792           mips32r5, mips32r6, mips64, mips64r2, mips64r3, mips64r5 and
19793           mips64r6.  The processor names are: 4kc, 4km, 4kp, 4ksc, 4kec,
19794           4kem, 4kep, 4ksd, 5kc, 5kf, 20kc, 24kc, 24kf2_1, 24kf1_1, 24kec,
19795           24kef2_1, 24kef1_1, 34kc, 34kf2_1, 34kf1_1, 34kn, 74kc, 74kf2_1,
19796           74kf1_1, 74kf3_2, 1004kc, 1004kf2_1, 1004kf1_1, i6400, i6500,
19797           interaptiv, loongson2e, loongson2f, loongson3a, gs464, gs464e,
19798           gs264e, m4k, m14k, m14kc, m14ke, m14kec, m5100, m5101, octeon,
19799           octeon+, octeon2, octeon3, orion, p5600, p6600, r2000, r3000,
19800           r3900, r4000, r4400, r4600, r4650, r4700, r5900, r6000, r8000,
19801           rm7000, rm9000, r10000, r12000, r14000, r16000, sb1, sr71000,
19802           vr4100, vr4111, vr4120, vr4130, vr4300, vr5000, vr5400, vr5500, xlr
19803           and xlp.  The special value from-abi selects the most compatible
19804           architecture for the selected ABI (that is, mips1 for 32-bit ABIs
19805           and mips3 for 64-bit ABIs).
19806
19807           The native Linux/GNU toolchain also supports the value native,
19808           which selects the best architecture option for the host processor.
19809           -march=native has no effect if GCC does not recognize the
19810           processor.
19811
19812           In processor names, a final 000 can be abbreviated as k (for
19813           example, -march=r2k).  Prefixes are optional, and vr may be written
19814           r.
19815
19816           Names of the form nf2_1 refer to processors with FPUs clocked at
19817           half the rate of the core, names of the form nf1_1 refer to
19818           processors with FPUs clocked at the same rate as the core, and
19819           names of the form nf3_2 refer to processors with FPUs clocked a
19820           ratio of 3:2 with respect to the core.  For compatibility reasons,
19821           nf is accepted as a synonym for nf2_1 while nx and bfx are accepted
19822           as synonyms for nf1_1.
19823
19824           GCC defines two macros based on the value of this option.  The
19825           first is "_MIPS_ARCH", which gives the name of target architecture,
19826           as a string.  The second has the form "_MIPS_ARCH_foo", where foo
19827           is the capitalized value of "_MIPS_ARCH".  For example,
19828           -march=r2000 sets "_MIPS_ARCH" to "r2000" and defines the macro
19829           "_MIPS_ARCH_R2000".
19830
19831           Note that the "_MIPS_ARCH" macro uses the processor names given
19832           above.  In other words, it has the full prefix and does not
19833           abbreviate 000 as k.  In the case of from-abi, the macro names the
19834           resolved architecture (either "mips1" or "mips3").  It names the
19835           default architecture when no -march option is given.
19836
19837       -mtune=arch
19838           Optimize for arch.  Among other things, this option controls the
19839           way instructions are scheduled, and the perceived cost of
19840           arithmetic operations.  The list of arch values is the same as for
19841           -march.
19842
19843           When this option is not used, GCC optimizes for the processor
19844           specified by -march.  By using -march and -mtune together, it is
19845           possible to generate code that runs on a family of processors, but
19846           optimize the code for one particular member of that family.
19847
19848           -mtune defines the macros "_MIPS_TUNE" and "_MIPS_TUNE_foo", which
19849           work in the same way as the -march ones described above.
19850
19851       -mips1
19852           Equivalent to -march=mips1.
19853
19854       -mips2
19855           Equivalent to -march=mips2.
19856
19857       -mips3
19858           Equivalent to -march=mips3.
19859
19860       -mips4
19861           Equivalent to -march=mips4.
19862
19863       -mips32
19864           Equivalent to -march=mips32.
19865
19866       -mips32r3
19867           Equivalent to -march=mips32r3.
19868
19869       -mips32r5
19870           Equivalent to -march=mips32r5.
19871
19872       -mips32r6
19873           Equivalent to -march=mips32r6.
19874
19875       -mips64
19876           Equivalent to -march=mips64.
19877
19878       -mips64r2
19879           Equivalent to -march=mips64r2.
19880
19881       -mips64r3
19882           Equivalent to -march=mips64r3.
19883
19884       -mips64r5
19885           Equivalent to -march=mips64r5.
19886
19887       -mips64r6
19888           Equivalent to -march=mips64r6.
19889
19890       -mips16
19891       -mno-mips16
19892           Generate (do not generate) MIPS16 code.  If GCC is targeting a
19893           MIPS32 or MIPS64 architecture, it makes use of the MIPS16e ASE.
19894
19895           MIPS16 code generation can also be controlled on a per-function
19896           basis by means of "mips16" and "nomips16" attributes.
19897
19898       -mflip-mips16
19899           Generate MIPS16 code on alternating functions.  This option is
19900           provided for regression testing of mixed MIPS16/non-MIPS16 code
19901           generation, and is not intended for ordinary use in compiling user
19902           code.
19903
19904       -minterlink-compressed
19905       -mno-interlink-compressed
19906           Require (do not require) that code using the standard
19907           (uncompressed) MIPS ISA be link-compatible with MIPS16 and
19908           microMIPS code, and vice versa.
19909
19910           For example, code using the standard ISA encoding cannot jump
19911           directly to MIPS16 or microMIPS code; it must either use a call or
19912           an indirect jump.  -minterlink-compressed therefore disables direct
19913           jumps unless GCC knows that the target of the jump is not
19914           compressed.
19915
19916       -minterlink-mips16
19917       -mno-interlink-mips16
19918           Aliases of -minterlink-compressed and -mno-interlink-compressed.
19919           These options predate the microMIPS ASE and are retained for
19920           backwards compatibility.
19921
19922       -mabi=32
19923       -mabi=o64
19924       -mabi=n32
19925       -mabi=64
19926       -mabi=eabi
19927           Generate code for the given ABI.
19928
19929           Note that the EABI has a 32-bit and a 64-bit variant.  GCC normally
19930           generates 64-bit code when you select a 64-bit architecture, but
19931           you can use -mgp32 to get 32-bit code instead.
19932
19933           For information about the O64 ABI, see
19934           <http://gcc.gnu.org/projects/mipso64-abi.html>.
19935
19936           GCC supports a variant of the o32 ABI in which floating-point
19937           registers are 64 rather than 32 bits wide.  You can select this
19938           combination with -mabi=32 -mfp64.  This ABI relies on the "mthc1"
19939           and "mfhc1" instructions and is therefore only supported for
19940           MIPS32R2, MIPS32R3 and MIPS32R5 processors.
19941
19942           The register assignments for arguments and return values remain the
19943           same, but each scalar value is passed in a single 64-bit register
19944           rather than a pair of 32-bit registers.  For example, scalar
19945           floating-point values are returned in $f0 only, not a $f0/$f1 pair.
19946           The set of call-saved registers also remains the same in that the
19947           even-numbered double-precision registers are saved.
19948
19949           Two additional variants of the o32 ABI are supported to enable a
19950           transition from 32-bit to 64-bit registers.  These are FPXX
19951           (-mfpxx) and FP64A (-mfp64 -mno-odd-spreg).  The FPXX extension
19952           mandates that all code must execute correctly when run using 32-bit
19953           or 64-bit registers.  The code can be interlinked with either FP32
19954           or FP64, but not both.  The FP64A extension is similar to the FP64
19955           extension but forbids the use of odd-numbered single-precision
19956           registers.  This can be used in conjunction with the "FRE" mode of
19957           FPUs in MIPS32R5 processors and allows both FP32 and FP64A code to
19958           interlink and run in the same process without changing FPU modes.
19959
19960       -mabicalls
19961       -mno-abicalls
19962           Generate (do not generate) code that is suitable for SVR4-style
19963           dynamic objects.  -mabicalls is the default for SVR4-based systems.
19964
19965       -mshared
19966       -mno-shared
19967           Generate (do not generate) code that is fully position-independent,
19968           and that can therefore be linked into shared libraries.  This
19969           option only affects -mabicalls.
19970
19971           All -mabicalls code has traditionally been position-independent,
19972           regardless of options like -fPIC and -fpic.  However, as an
19973           extension, the GNU toolchain allows executables to use absolute
19974           accesses for locally-binding symbols.  It can also use shorter GP
19975           initialization sequences and generate direct calls to locally-
19976           defined functions.  This mode is selected by -mno-shared.
19977
19978           -mno-shared depends on binutils 2.16 or higher and generates
19979           objects that can only be linked by the GNU linker.  However, the
19980           option does not affect the ABI of the final executable; it only
19981           affects the ABI of relocatable objects.  Using -mno-shared
19982           generally makes executables both smaller and quicker.
19983
19984           -mshared is the default.
19985
19986       -mplt
19987       -mno-plt
19988           Assume (do not assume) that the static and dynamic linkers support
19989           PLTs and copy relocations.  This option only affects -mno-shared
19990           -mabicalls.  For the n64 ABI, this option has no effect without
19991           -msym32.
19992
19993           You can make -mplt the default by configuring GCC with
19994           --with-mips-plt.  The default is -mno-plt otherwise.
19995
19996       -mxgot
19997       -mno-xgot
19998           Lift (do not lift) the usual restrictions on the size of the global
19999           offset table.
20000
20001           GCC normally uses a single instruction to load values from the GOT.
20002           While this is relatively efficient, it only works if the GOT is
20003           smaller than about 64k.  Anything larger causes the linker to
20004           report an error such as:
20005
20006                   relocation truncated to fit: R_MIPS_GOT16 foobar
20007
20008           If this happens, you should recompile your code with -mxgot.  This
20009           works with very large GOTs, although the code is also less
20010           efficient, since it takes three instructions to fetch the value of
20011           a global symbol.
20012
20013           Note that some linkers can create multiple GOTs.  If you have such
20014           a linker, you should only need to use -mxgot when a single object
20015           file accesses more than 64k's worth of GOT entries.  Very few do.
20016
20017           These options have no effect unless GCC is generating position
20018           independent code.
20019
20020       -mgp32
20021           Assume that general-purpose registers are 32 bits wide.
20022
20023       -mgp64
20024           Assume that general-purpose registers are 64 bits wide.
20025
20026       -mfp32
20027           Assume that floating-point registers are 32 bits wide.
20028
20029       -mfp64
20030           Assume that floating-point registers are 64 bits wide.
20031
20032       -mfpxx
20033           Do not assume the width of floating-point registers.
20034
20035       -mhard-float
20036           Use floating-point coprocessor instructions.
20037
20038       -msoft-float
20039           Do not use floating-point coprocessor instructions.  Implement
20040           floating-point calculations using library calls instead.
20041
20042       -mno-float
20043           Equivalent to -msoft-float, but additionally asserts that the
20044           program being compiled does not perform any floating-point
20045           operations.  This option is presently supported only by some bare-
20046           metal MIPS configurations, where it may select a special set of
20047           libraries that lack all floating-point support (including, for
20048           example, the floating-point "printf" formats).  If code compiled
20049           with -mno-float accidentally contains floating-point operations, it
20050           is likely to suffer a link-time or run-time failure.
20051
20052       -msingle-float
20053           Assume that the floating-point coprocessor only supports single-
20054           precision operations.
20055
20056       -mdouble-float
20057           Assume that the floating-point coprocessor supports double-
20058           precision operations.  This is the default.
20059
20060       -modd-spreg
20061       -mno-odd-spreg
20062           Enable the use of odd-numbered single-precision floating-point
20063           registers for the o32 ABI.  This is the default for processors that
20064           are known to support these registers.  When using the o32 FPXX ABI,
20065           -mno-odd-spreg is set by default.
20066
20067       -mabs=2008
20068       -mabs=legacy
20069           These options control the treatment of the special not-a-number
20070           (NaN) IEEE 754 floating-point data with the "abs.fmt" and "neg.fmt"
20071           machine instructions.
20072
20073           By default or when -mabs=legacy is used the legacy treatment is
20074           selected.  In this case these instructions are considered
20075           arithmetic and avoided where correct operation is required and the
20076           input operand might be a NaN.  A longer sequence of instructions
20077           that manipulate the sign bit of floating-point datum manually is
20078           used instead unless the -ffinite-math-only option has also been
20079           specified.
20080
20081           The -mabs=2008 option selects the IEEE 754-2008 treatment.  In this
20082           case these instructions are considered non-arithmetic and therefore
20083           operating correctly in all cases, including in particular where the
20084           input operand is a NaN.  These instructions are therefore always
20085           used for the respective operations.
20086
20087       -mnan=2008
20088       -mnan=legacy
20089           These options control the encoding of the special not-a-number
20090           (NaN) IEEE 754 floating-point data.
20091
20092           The -mnan=legacy option selects the legacy encoding.  In this case
20093           quiet NaNs (qNaNs) are denoted by the first bit of their trailing
20094           significand field being 0, whereas signaling NaNs (sNaNs) are
20095           denoted by the first bit of their trailing significand field being
20096           1.
20097
20098           The -mnan=2008 option selects the IEEE 754-2008 encoding.  In this
20099           case qNaNs are denoted by the first bit of their trailing
20100           significand field being 1, whereas sNaNs are denoted by the first
20101           bit of their trailing significand field being 0.
20102
20103           The default is -mnan=legacy unless GCC has been configured with
20104           --with-nan=2008.
20105
20106       -mllsc
20107       -mno-llsc
20108           Use (do not use) ll, sc, and sync instructions to implement atomic
20109           memory built-in functions.  When neither option is specified, GCC
20110           uses the instructions if the target architecture supports them.
20111
20112           -mllsc is useful if the runtime environment can emulate the
20113           instructions and -mno-llsc can be useful when compiling for
20114           nonstandard ISAs.  You can make either option the default by
20115           configuring GCC with --with-llsc and --without-llsc respectively.
20116           --with-llsc is the default for some configurations; see the
20117           installation documentation for details.
20118
20119       -mdsp
20120       -mno-dsp
20121           Use (do not use) revision 1 of the MIPS DSP ASE.
20122             This option defines the preprocessor macro "__mips_dsp".  It also
20123           defines "__mips_dsp_rev" to 1.
20124
20125       -mdspr2
20126       -mno-dspr2
20127           Use (do not use) revision 2 of the MIPS DSP ASE.
20128             This option defines the preprocessor macros "__mips_dsp" and
20129           "__mips_dspr2".  It also defines "__mips_dsp_rev" to 2.
20130
20131       -msmartmips
20132       -mno-smartmips
20133           Use (do not use) the MIPS SmartMIPS ASE.
20134
20135       -mpaired-single
20136       -mno-paired-single
20137           Use (do not use) paired-single floating-point instructions.
20138             This option requires hardware floating-point support to be
20139           enabled.
20140
20141       -mdmx
20142       -mno-mdmx
20143           Use (do not use) MIPS Digital Media Extension instructions.  This
20144           option can only be used when generating 64-bit code and requires
20145           hardware floating-point support to be enabled.
20146
20147       -mips3d
20148       -mno-mips3d
20149           Use (do not use) the MIPS-3D ASE.  The option -mips3d implies
20150           -mpaired-single.
20151
20152       -mmicromips
20153       -mno-micromips
20154           Generate (do not generate) microMIPS code.
20155
20156           MicroMIPS code generation can also be controlled on a per-function
20157           basis by means of "micromips" and "nomicromips" attributes.
20158
20159       -mmt
20160       -mno-mt
20161           Use (do not use) MT Multithreading instructions.
20162
20163       -mmcu
20164       -mno-mcu
20165           Use (do not use) the MIPS MCU ASE instructions.
20166
20167       -meva
20168       -mno-eva
20169           Use (do not use) the MIPS Enhanced Virtual Addressing instructions.
20170
20171       -mvirt
20172       -mno-virt
20173           Use (do not use) the MIPS Virtualization (VZ) instructions.
20174
20175       -mxpa
20176       -mno-xpa
20177           Use (do not use) the MIPS eXtended Physical Address (XPA)
20178           instructions.
20179
20180       -mcrc
20181       -mno-crc
20182           Use (do not use) the MIPS Cyclic Redundancy Check (CRC)
20183           instructions.
20184
20185       -mginv
20186       -mno-ginv
20187           Use (do not use) the MIPS Global INValidate (GINV) instructions.
20188
20189       -mloongson-mmi
20190       -mno-loongson-mmi
20191           Use (do not use) the MIPS Loongson MultiMedia extensions
20192           Instructions (MMI).
20193
20194       -mloongson-ext
20195       -mno-loongson-ext
20196           Use (do not use) the MIPS Loongson EXTensions (EXT) instructions.
20197
20198       -mloongson-ext2
20199       -mno-loongson-ext2
20200           Use (do not use) the MIPS Loongson EXTensions r2 (EXT2)
20201           instructions.
20202
20203       -mlong64
20204           Force "long" types to be 64 bits wide.  See -mlong32 for an
20205           explanation of the default and the way that the pointer size is
20206           determined.
20207
20208       -mlong32
20209           Force "long", "int", and pointer types to be 32 bits wide.
20210
20211           The default size of "int"s, "long"s and pointers depends on the
20212           ABI.  All the supported ABIs use 32-bit "int"s.  The n64 ABI uses
20213           64-bit "long"s, as does the 64-bit EABI; the others use 32-bit
20214           "long"s.  Pointers are the same size as "long"s, or the same size
20215           as integer registers, whichever is smaller.
20216
20217       -msym32
20218       -mno-sym32
20219           Assume (do not assume) that all symbols have 32-bit values,
20220           regardless of the selected ABI.  This option is useful in
20221           combination with -mabi=64 and -mno-abicalls because it allows GCC
20222           to generate shorter and faster references to symbolic addresses.
20223
20224       -G num
20225           Put definitions of externally-visible data in a small data section
20226           if that data is no bigger than num bytes.  GCC can then generate
20227           more efficient accesses to the data; see -mgpopt for details.
20228
20229           The default -G option depends on the configuration.
20230
20231       -mlocal-sdata
20232       -mno-local-sdata
20233           Extend (do not extend) the -G behavior to local data too, such as
20234           to static variables in C.  -mlocal-sdata is the default for all
20235           configurations.
20236
20237           If the linker complains that an application is using too much small
20238           data, you might want to try rebuilding the less performance-
20239           critical parts with -mno-local-sdata.  You might also want to build
20240           large libraries with -mno-local-sdata, so that the libraries leave
20241           more room for the main program.
20242
20243       -mextern-sdata
20244       -mno-extern-sdata
20245           Assume (do not assume) that externally-defined data is in a small
20246           data section if the size of that data is within the -G limit.
20247           -mextern-sdata is the default for all configurations.
20248
20249           If you compile a module Mod with -mextern-sdata -G num -mgpopt, and
20250           Mod references a variable Var that is no bigger than num bytes, you
20251           must make sure that Var is placed in a small data section.  If Var
20252           is defined by another module, you must either compile that module
20253           with a high-enough -G setting or attach a "section" attribute to
20254           Var's definition.  If Var is common, you must link the application
20255           with a high-enough -G setting.
20256
20257           The easiest way of satisfying these restrictions is to compile and
20258           link every module with the same -G option.  However, you may wish
20259           to build a library that supports several different small data
20260           limits.  You can do this by compiling the library with the highest
20261           supported -G setting and additionally using -mno-extern-sdata to
20262           stop the library from making assumptions about externally-defined
20263           data.
20264
20265       -mgpopt
20266       -mno-gpopt
20267           Use (do not use) GP-relative accesses for symbols that are known to
20268           be in a small data section; see -G, -mlocal-sdata and
20269           -mextern-sdata.  -mgpopt is the default for all configurations.
20270
20271           -mno-gpopt is useful for cases where the $gp register might not
20272           hold the value of "_gp".  For example, if the code is part of a
20273           library that might be used in a boot monitor, programs that call
20274           boot monitor routines pass an unknown value in $gp.  (In such
20275           situations, the boot monitor itself is usually compiled with -G0.)
20276
20277           -mno-gpopt implies -mno-local-sdata and -mno-extern-sdata.
20278
20279       -membedded-data
20280       -mno-embedded-data
20281           Allocate variables to the read-only data section first if possible,
20282           then next in the small data section if possible, otherwise in data.
20283           This gives slightly slower code than the default, but reduces the
20284           amount of RAM required when executing, and thus may be preferred
20285           for some embedded systems.
20286
20287       -muninit-const-in-rodata
20288       -mno-uninit-const-in-rodata
20289           Put uninitialized "const" variables in the read-only data section.
20290           This option is only meaningful in conjunction with -membedded-data.
20291
20292       -mcode-readable=setting
20293           Specify whether GCC may generate code that reads from executable
20294           sections.  There are three possible settings:
20295
20296           -mcode-readable=yes
20297               Instructions may freely access executable sections.  This is
20298               the default setting.
20299
20300           -mcode-readable=pcrel
20301               MIPS16 PC-relative load instructions can access executable
20302               sections, but other instructions must not do so.  This option
20303               is useful on 4KSc and 4KSd processors when the code TLBs have
20304               the Read Inhibit bit set.  It is also useful on processors that
20305               can be configured to have a dual instruction/data SRAM
20306               interface and that, like the M4K, automatically redirect PC-
20307               relative loads to the instruction RAM.
20308
20309           -mcode-readable=no
20310               Instructions must not access executable sections.  This option
20311               can be useful on targets that are configured to have a dual
20312               instruction/data SRAM interface but that (unlike the M4K) do
20313               not automatically redirect PC-relative loads to the instruction
20314               RAM.
20315
20316       -msplit-addresses
20317       -mno-split-addresses
20318           Enable (disable) use of the "%hi()" and "%lo()" assembler
20319           relocation operators.  This option has been superseded by
20320           -mexplicit-relocs but is retained for backwards compatibility.
20321
20322       -mexplicit-relocs
20323       -mno-explicit-relocs
20324           Use (do not use) assembler relocation operators when dealing with
20325           symbolic addresses.  The alternative, selected by
20326           -mno-explicit-relocs, is to use assembler macros instead.
20327
20328           -mexplicit-relocs is the default if GCC was configured to use an
20329           assembler that supports relocation operators.
20330
20331       -mcheck-zero-division
20332       -mno-check-zero-division
20333           Trap (do not trap) on integer division by zero.
20334
20335           The default is -mcheck-zero-division.
20336
20337       -mdivide-traps
20338       -mdivide-breaks
20339           MIPS systems check for division by zero by generating either a
20340           conditional trap or a break instruction.  Using traps results in
20341           smaller code, but is only supported on MIPS II and later.  Also,
20342           some versions of the Linux kernel have a bug that prevents trap
20343           from generating the proper signal ("SIGFPE").  Use -mdivide-traps
20344           to allow conditional traps on architectures that support them and
20345           -mdivide-breaks to force the use of breaks.
20346
20347           The default is usually -mdivide-traps, but this can be overridden
20348           at configure time using --with-divide=breaks.  Divide-by-zero
20349           checks can be completely disabled using -mno-check-zero-division.
20350
20351       -mload-store-pairs
20352       -mno-load-store-pairs
20353           Enable (disable) an optimization that pairs consecutive load or
20354           store instructions to enable load/store bonding.  This option is
20355           enabled by default but only takes effect when the selected
20356           architecture is known to support bonding.
20357
20358       -mmemcpy
20359       -mno-memcpy
20360           Force (do not force) the use of "memcpy" for non-trivial block
20361           moves.  The default is -mno-memcpy, which allows GCC to inline most
20362           constant-sized copies.
20363
20364       -mlong-calls
20365       -mno-long-calls
20366           Disable (do not disable) use of the "jal" instruction.  Calling
20367           functions using "jal" is more efficient but requires the caller and
20368           callee to be in the same 256 megabyte segment.
20369
20370           This option has no effect on abicalls code.  The default is
20371           -mno-long-calls.
20372
20373       -mmad
20374       -mno-mad
20375           Enable (disable) use of the "mad", "madu" and "mul" instructions,
20376           as provided by the R4650 ISA.
20377
20378       -mimadd
20379       -mno-imadd
20380           Enable (disable) use of the "madd" and "msub" integer instructions.
20381           The default is -mimadd on architectures that support "madd" and
20382           "msub" except for the 74k architecture where it was found to
20383           generate slower code.
20384
20385       -mfused-madd
20386       -mno-fused-madd
20387           Enable (disable) use of the floating-point multiply-accumulate
20388           instructions, when they are available.  The default is
20389           -mfused-madd.
20390
20391           On the R8000 CPU when multiply-accumulate instructions are used,
20392           the intermediate product is calculated to infinite precision and is
20393           not subject to the FCSR Flush to Zero bit.  This may be undesirable
20394           in some circumstances.  On other processors the result is
20395           numerically identical to the equivalent computation using separate
20396           multiply, add, subtract and negate instructions.
20397
20398       -nocpp
20399           Tell the MIPS assembler to not run its preprocessor over user
20400           assembler files (with a .s suffix) when assembling them.
20401
20402       -mfix-24k
20403       -mno-fix-24k
20404           Work around the 24K E48 (lost data on stores during refill) errata.
20405           The workarounds are implemented by the assembler rather than by
20406           GCC.
20407
20408       -mfix-r4000
20409       -mno-fix-r4000
20410           Work around certain R4000 CPU errata:
20411
20412           -   A double-word or a variable shift may give an incorrect result
20413               if executed immediately after starting an integer division.
20414
20415           -   A double-word or a variable shift may give an incorrect result
20416               if executed while an integer multiplication is in progress.
20417
20418           -   An integer division may give an incorrect result if started in
20419               a delay slot of a taken branch or a jump.
20420
20421       -mfix-r4400
20422       -mno-fix-r4400
20423           Work around certain R4400 CPU errata:
20424
20425           -   A double-word or a variable shift may give an incorrect result
20426               if executed immediately after starting an integer division.
20427
20428       -mfix-r10000
20429       -mno-fix-r10000
20430           Work around certain R10000 errata:
20431
20432           -   "ll"/"sc" sequences may not behave atomically on revisions
20433               prior to 3.0.  They may deadlock on revisions 2.6 and earlier.
20434
20435           This option can only be used if the target architecture supports
20436           branch-likely instructions.  -mfix-r10000 is the default when
20437           -march=r10000 is used; -mno-fix-r10000 is the default otherwise.
20438
20439       -mfix-r5900
20440       -mno-fix-r5900
20441           Do not attempt to schedule the preceding instruction into the delay
20442           slot of a branch instruction placed at the end of a short loop of
20443           six instructions or fewer and always schedule a "nop" instruction
20444           there instead.  The short loop bug under certain conditions causes
20445           loops to execute only once or twice, due to a hardware bug in the
20446           R5900 chip.  The workaround is implemented by the assembler rather
20447           than by GCC.
20448
20449       -mfix-rm7000
20450       -mno-fix-rm7000
20451           Work around the RM7000 "dmult"/"dmultu" errata.  The workarounds
20452           are implemented by the assembler rather than by GCC.
20453
20454       -mfix-vr4120
20455       -mno-fix-vr4120
20456           Work around certain VR4120 errata:
20457
20458           -   "dmultu" does not always produce the correct result.
20459
20460           -   "div" and "ddiv" do not always produce the correct result if
20461               one of the operands is negative.
20462
20463           The workarounds for the division errata rely on special functions
20464           in libgcc.a.  At present, these functions are only provided by the
20465           "mips64vr*-elf" configurations.
20466
20467           Other VR4120 errata require a NOP to be inserted between certain
20468           pairs of instructions.  These errata are handled by the assembler,
20469           not by GCC itself.
20470
20471       -mfix-vr4130
20472           Work around the VR4130 "mflo"/"mfhi" errata.  The workarounds are
20473           implemented by the assembler rather than by GCC, although GCC
20474           avoids using "mflo" and "mfhi" if the VR4130 "macc", "macchi",
20475           "dmacc" and "dmacchi" instructions are available instead.
20476
20477       -mfix-sb1
20478       -mno-fix-sb1
20479           Work around certain SB-1 CPU core errata.  (This flag currently
20480           works around the SB-1 revision 2 "F1" and "F2" floating-point
20481           errata.)
20482
20483       -mr10k-cache-barrier=setting
20484           Specify whether GCC should insert cache barriers to avoid the side
20485           effects of speculation on R10K processors.
20486
20487           In common with many processors, the R10K tries to predict the
20488           outcome of a conditional branch and speculatively executes
20489           instructions from the "taken" branch.  It later aborts these
20490           instructions if the predicted outcome is wrong.  However, on the
20491           R10K, even aborted instructions can have side effects.
20492
20493           This problem only affects kernel stores and, depending on the
20494           system, kernel loads.  As an example, a speculatively-executed
20495           store may load the target memory into cache and mark the cache line
20496           as dirty, even if the store itself is later aborted.  If a DMA
20497           operation writes to the same area of memory before the "dirty" line
20498           is flushed, the cached data overwrites the DMA-ed data.  See the
20499           R10K processor manual for a full description, including other
20500           potential problems.
20501
20502           One workaround is to insert cache barrier instructions before every
20503           memory access that might be speculatively executed and that might
20504           have side effects even if aborted.  -mr10k-cache-barrier=setting
20505           controls GCC's implementation of this workaround.  It assumes that
20506           aborted accesses to any byte in the following regions does not have
20507           side effects:
20508
20509           1.  the memory occupied by the current function's stack frame;
20510
20511           2.  the memory occupied by an incoming stack argument;
20512
20513           3.  the memory occupied by an object with a link-time-constant
20514               address.
20515
20516           It is the kernel's responsibility to ensure that speculative
20517           accesses to these regions are indeed safe.
20518
20519           If the input program contains a function declaration such as:
20520
20521                   void foo (void);
20522
20523           then the implementation of "foo" must allow "j foo" and "jal foo"
20524           to be executed speculatively.  GCC honors this restriction for
20525           functions it compiles itself.  It expects non-GCC functions (such
20526           as hand-written assembly code) to do the same.
20527
20528           The option has three forms:
20529
20530           -mr10k-cache-barrier=load-store
20531               Insert a cache barrier before a load or store that might be
20532               speculatively executed and that might have side effects even if
20533               aborted.
20534
20535           -mr10k-cache-barrier=store
20536               Insert a cache barrier before a store that might be
20537               speculatively executed and that might have side effects even if
20538               aborted.
20539
20540           -mr10k-cache-barrier=none
20541               Disable the insertion of cache barriers.  This is the default
20542               setting.
20543
20544       -mflush-func=func
20545       -mno-flush-func
20546           Specifies the function to call to flush the I and D caches, or to
20547           not call any such function.  If called, the function must take the
20548           same arguments as the common "_flush_func", that is, the address of
20549           the memory range for which the cache is being flushed, the size of
20550           the memory range, and the number 3 (to flush both caches).  The
20551           default depends on the target GCC was configured for, but commonly
20552           is either "_flush_func" or "__cpu_flush".
20553
20554       mbranch-cost=num
20555           Set the cost of branches to roughly num "simple" instructions.
20556           This cost is only a heuristic and is not guaranteed to produce
20557           consistent results across releases.  A zero cost redundantly
20558           selects the default, which is based on the -mtune setting.
20559
20560       -mbranch-likely
20561       -mno-branch-likely
20562           Enable or disable use of Branch Likely instructions, regardless of
20563           the default for the selected architecture.  By default, Branch
20564           Likely instructions may be generated if they are supported by the
20565           selected architecture.  An exception is for the MIPS32 and MIPS64
20566           architectures and processors that implement those architectures;
20567           for those, Branch Likely instructions are not be generated by
20568           default because the MIPS32 and MIPS64 architectures specifically
20569           deprecate their use.
20570
20571       -mcompact-branches=never
20572       -mcompact-branches=optimal
20573       -mcompact-branches=always
20574           These options control which form of branches will be generated.
20575           The default is -mcompact-branches=optimal.
20576
20577           The -mcompact-branches=never option ensures that compact branch
20578           instructions will never be generated.
20579
20580           The -mcompact-branches=always option ensures that a compact branch
20581           instruction will be generated if available.  If a compact branch
20582           instruction is not available, a delay slot form of the branch will
20583           be used instead.
20584
20585           This option is supported from MIPS Release 6 onwards.
20586
20587           The -mcompact-branches=optimal option will cause a delay slot
20588           branch to be used if one is available in the current ISA and the
20589           delay slot is successfully filled.  If the delay slot is not
20590           filled, a compact branch will be chosen if one is available.
20591
20592       -mfp-exceptions
20593       -mno-fp-exceptions
20594           Specifies whether FP exceptions are enabled.  This affects how FP
20595           instructions are scheduled for some processors.  The default is
20596           that FP exceptions are enabled.
20597
20598           For instance, on the SB-1, if FP exceptions are disabled, and we
20599           are emitting 64-bit code, then we can use both FP pipes.
20600           Otherwise, we can only use one FP pipe.
20601
20602       -mvr4130-align
20603       -mno-vr4130-align
20604           The VR4130 pipeline is two-way superscalar, but can only issue two
20605           instructions together if the first one is 8-byte aligned.  When
20606           this option is enabled, GCC aligns pairs of instructions that it
20607           thinks should execute in parallel.
20608
20609           This option only has an effect when optimizing for the VR4130.  It
20610           normally makes code faster, but at the expense of making it bigger.
20611           It is enabled by default at optimization level -O3.
20612
20613       -msynci
20614       -mno-synci
20615           Enable (disable) generation of "synci" instructions on
20616           architectures that support it.  The "synci" instructions (if
20617           enabled) are generated when "__builtin___clear_cache" is compiled.
20618
20619           This option defaults to -mno-synci, but the default can be
20620           overridden by configuring GCC with --with-synci.
20621
20622           When compiling code for single processor systems, it is generally
20623           safe to use "synci".  However, on many multi-core (SMP) systems, it
20624           does not invalidate the instruction caches on all cores and may
20625           lead to undefined behavior.
20626
20627       -mrelax-pic-calls
20628       -mno-relax-pic-calls
20629           Try to turn PIC calls that are normally dispatched via register $25
20630           into direct calls.  This is only possible if the linker can resolve
20631           the destination at link time and if the destination is within range
20632           for a direct call.
20633
20634           -mrelax-pic-calls is the default if GCC was configured to use an
20635           assembler and a linker that support the ".reloc" assembly directive
20636           and -mexplicit-relocs is in effect.  With -mno-explicit-relocs,
20637           this optimization can be performed by the assembler and the linker
20638           alone without help from the compiler.
20639
20640       -mmcount-ra-address
20641       -mno-mcount-ra-address
20642           Emit (do not emit) code that allows "_mcount" to modify the calling
20643           function's return address.  When enabled, this option extends the
20644           usual "_mcount" interface with a new ra-address parameter, which
20645           has type "intptr_t *" and is passed in register $12.  "_mcount" can
20646           then modify the return address by doing both of the following:
20647
20648           *   Returning the new address in register $31.
20649
20650           *   Storing the new address in "*ra-address", if ra-address is
20651               nonnull.
20652
20653           The default is -mno-mcount-ra-address.
20654
20655       -mframe-header-opt
20656       -mno-frame-header-opt
20657           Enable (disable) frame header optimization in the o32 ABI.  When
20658           using the o32 ABI, calling functions will allocate 16 bytes on the
20659           stack for the called function to write out register arguments.
20660           When enabled, this optimization will suppress the allocation of the
20661           frame header if it can be determined that it is unused.
20662
20663           This optimization is off by default at all optimization levels.
20664
20665       -mlxc1-sxc1
20666       -mno-lxc1-sxc1
20667           When applicable, enable (disable) the generation of "lwxc1",
20668           "swxc1", "ldxc1", "sdxc1" instructions.  Enabled by default.
20669
20670       -mmadd4
20671       -mno-madd4
20672           When applicable, enable (disable) the generation of 4-operand
20673           "madd.s", "madd.d" and related instructions.  Enabled by default.
20674
20675       MMIX Options
20676
20677       These options are defined for the MMIX:
20678
20679       -mlibfuncs
20680       -mno-libfuncs
20681           Specify that intrinsic library functions are being compiled,
20682           passing all values in registers, no matter the size.
20683
20684       -mepsilon
20685       -mno-epsilon
20686           Generate floating-point comparison instructions that compare with
20687           respect to the "rE" epsilon register.
20688
20689       -mabi=mmixware
20690       -mabi=gnu
20691           Generate code that passes function parameters and return values
20692           that (in the called function) are seen as registers $0 and up, as
20693           opposed to the GNU ABI which uses global registers $231 and up.
20694
20695       -mzero-extend
20696       -mno-zero-extend
20697           When reading data from memory in sizes shorter than 64 bits, use
20698           (do not use) zero-extending load instructions by default, rather
20699           than sign-extending ones.
20700
20701       -mknuthdiv
20702       -mno-knuthdiv
20703           Make the result of a division yielding a remainder have the same
20704           sign as the divisor.  With the default, -mno-knuthdiv, the sign of
20705           the remainder follows the sign of the dividend.  Both methods are
20706           arithmetically valid, the latter being almost exclusively used.
20707
20708       -mtoplevel-symbols
20709       -mno-toplevel-symbols
20710           Prepend (do not prepend) a : to all global symbols, so the assembly
20711           code can be used with the "PREFIX" assembly directive.
20712
20713       -melf
20714           Generate an executable in the ELF format, rather than the default
20715           mmo format used by the mmix simulator.
20716
20717       -mbranch-predict
20718       -mno-branch-predict
20719           Use (do not use) the probable-branch instructions, when static
20720           branch prediction indicates a probable branch.
20721
20722       -mbase-addresses
20723       -mno-base-addresses
20724           Generate (do not generate) code that uses base addresses.  Using a
20725           base address automatically generates a request (handled by the
20726           assembler and the linker) for a constant to be set up in a global
20727           register.  The register is used for one or more base address
20728           requests within the range 0 to 255 from the value held in the
20729           register.  The generally leads to short and fast code, but the
20730           number of different data items that can be addressed is limited.
20731           This means that a program that uses lots of static data may require
20732           -mno-base-addresses.
20733
20734       -msingle-exit
20735       -mno-single-exit
20736           Force (do not force) generated code to have a single exit point in
20737           each function.
20738
20739       MN10300 Options
20740
20741       These -m options are defined for Matsushita MN10300 architectures:
20742
20743       -mmult-bug
20744           Generate code to avoid bugs in the multiply instructions for the
20745           MN10300 processors.  This is the default.
20746
20747       -mno-mult-bug
20748           Do not generate code to avoid bugs in the multiply instructions for
20749           the MN10300 processors.
20750
20751       -mam33
20752           Generate code using features specific to the AM33 processor.
20753
20754       -mno-am33
20755           Do not generate code using features specific to the AM33 processor.
20756           This is the default.
20757
20758       -mam33-2
20759           Generate code using features specific to the AM33/2.0 processor.
20760
20761       -mam34
20762           Generate code using features specific to the AM34 processor.
20763
20764       -mtune=cpu-type
20765           Use the timing characteristics of the indicated CPU type when
20766           scheduling instructions.  This does not change the targeted
20767           processor type.  The CPU type must be one of mn10300, am33, am33-2
20768           or am34.
20769
20770       -mreturn-pointer-on-d0
20771           When generating a function that returns a pointer, return the
20772           pointer in both "a0" and "d0".  Otherwise, the pointer is returned
20773           only in "a0", and attempts to call such functions without a
20774           prototype result in errors.  Note that this option is on by
20775           default; use -mno-return-pointer-on-d0 to disable it.
20776
20777       -mno-crt0
20778           Do not link in the C run-time initialization object file.
20779
20780       -mrelax
20781           Indicate to the linker that it should perform a relaxation
20782           optimization pass to shorten branches, calls and absolute memory
20783           addresses.  This option only has an effect when used on the command
20784           line for the final link step.
20785
20786           This option makes symbolic debugging impossible.
20787
20788       -mliw
20789           Allow the compiler to generate Long Instruction Word instructions
20790           if the target is the AM33 or later.  This is the default.  This
20791           option defines the preprocessor macro "__LIW__".
20792
20793       -mno-liw
20794           Do not allow the compiler to generate Long Instruction Word
20795           instructions.  This option defines the preprocessor macro
20796           "__NO_LIW__".
20797
20798       -msetlb
20799           Allow the compiler to generate the SETLB and Lcc instructions if
20800           the target is the AM33 or later.  This is the default.  This option
20801           defines the preprocessor macro "__SETLB__".
20802
20803       -mno-setlb
20804           Do not allow the compiler to generate SETLB or Lcc instructions.
20805           This option defines the preprocessor macro "__NO_SETLB__".
20806
20807       Moxie Options
20808
20809       -meb
20810           Generate big-endian code.  This is the default for moxie-*-*
20811           configurations.
20812
20813       -mel
20814           Generate little-endian code.
20815
20816       -mmul.x
20817           Generate mul.x and umul.x instructions.  This is the default for
20818           moxiebox-*-* configurations.
20819
20820       -mno-crt0
20821           Do not link in the C run-time initialization object file.
20822
20823       MSP430 Options
20824
20825       These options are defined for the MSP430:
20826
20827       -masm-hex
20828           Force assembly output to always use hex constants.  Normally such
20829           constants are signed decimals, but this option is available for
20830           testsuite and/or aesthetic purposes.
20831
20832       -mmcu=
20833           Select the MCU to target.  This is used to create a C preprocessor
20834           symbol based upon the MCU name, converted to upper case and pre-
20835           and post-fixed with __.  This in turn is used by the msp430.h
20836           header file to select an MCU-specific supplementary header file.
20837
20838           The option also sets the ISA to use.  If the MCU name is one that
20839           is known to only support the 430 ISA then that is selected,
20840           otherwise the 430X ISA is selected.  A generic MCU name of msp430
20841           can also be used to select the 430 ISA.  Similarly the generic
20842           msp430x MCU name selects the 430X ISA.
20843
20844           In addition an MCU-specific linker script is added to the linker
20845           command line.  The script's name is the name of the MCU with .ld
20846           appended.  Thus specifying -mmcu=xxx on the gcc command line
20847           defines the C preprocessor symbol "__XXX__" and cause the linker to
20848           search for a script called xxx.ld.
20849
20850           The ISA and hardware multiply supported for the different MCUs is
20851           hard-coded into GCC.  However, an external devices.csv file can be
20852           used to extend device support beyond those that have been hard-
20853           coded.
20854
20855           GCC searches for the devices.csv file using the following methods
20856           in the given precedence order, where the first method takes
20857           precendence over the second which takes precedence over the third.
20858
20859           Include path specified with "-I" and "-L"
20860               devices.csv will be searched for in each of the directories
20861               specified by include paths and linker library search paths.
20862
20863           Path specified by the environment variable MSP430_GCC_INCLUDE_DIR
20864               Define the value of the global environment variable
20865               MSP430_GCC_INCLUDE_DIR to the full path to the directory
20866               containing devices.csv, and GCC will search this directory for
20867               devices.csv.  If devices.csv is found, this directory will also
20868               be registered as an include path, and linker library path.
20869               Header files and linker scripts in this directory can therefore
20870               be used without manually specifying "-I" and "-L" on the
20871               command line.
20872
20873           The msp430-elf{,bare}/include/devices directory
20874               Finally, GCC will examine msp430-elf{,bare}/include/devices
20875               from the toolchain root directory.  This directory does not
20876               exist in a default installation, but if the user has created it
20877               and copied devices.csv there, then the MCU data will be read.
20878               As above, this directory will also be registered as an include
20879               path, and linker library path.
20880
20881           If none of the above search methods find devices.csv, then the
20882           hard-coded MCU data is used.
20883
20884       -mwarn-mcu
20885       -mno-warn-mcu
20886           This option enables or disables warnings about conflicts between
20887           the MCU name specified by the -mmcu option and the ISA set by the
20888           -mcpu option and/or the hardware multiply support set by the
20889           -mhwmult option.  It also toggles warnings about unrecognized MCU
20890           names.  This option is on by default.
20891
20892       -mcpu=
20893           Specifies the ISA to use.  Accepted values are msp430, msp430x and
20894           msp430xv2.  This option is deprecated.  The -mmcu= option should be
20895           used to select the ISA.
20896
20897       -msim
20898           Link to the simulator runtime libraries and linker script.
20899           Overrides any scripts that would be selected by the -mmcu= option.
20900
20901       -mlarge
20902           Use large-model addressing (20-bit pointers, 32-bit "size_t").
20903
20904       -msmall
20905           Use small-model addressing (16-bit pointers, 16-bit "size_t").
20906
20907       -mrelax
20908           This option is passed to the assembler and linker, and allows the
20909           linker to perform certain optimizations that cannot be done until
20910           the final link.
20911
20912       mhwmult=
20913           Describes the type of hardware multiply supported by the target.
20914           Accepted values are none for no hardware multiply, 16bit for the
20915           original 16-bit-only multiply supported by early MCUs.  32bit for
20916           the 16/32-bit multiply supported by later MCUs and f5series for the
20917           16/32-bit multiply supported by F5-series MCUs.  A value of auto
20918           can also be given.  This tells GCC to deduce the hardware multiply
20919           support based upon the MCU name provided by the -mmcu option.  If
20920           no -mmcu option is specified or if the MCU name is not recognized
20921           then no hardware multiply support is assumed.  "auto" is the
20922           default setting.
20923
20924           Hardware multiplies are normally performed by calling a library
20925           routine.  This saves space in the generated code.  When compiling
20926           at -O3 or higher however the hardware multiplier is invoked inline.
20927           This makes for bigger, but faster code.
20928
20929           The hardware multiply routines disable interrupts whilst running
20930           and restore the previous interrupt state when they finish.  This
20931           makes them safe to use inside interrupt handlers as well as in
20932           normal code.
20933
20934       -minrt
20935           Enable the use of a minimum runtime environment - no static
20936           initializers or constructors.  This is intended for memory-
20937           constrained devices.  The compiler includes special symbols in some
20938           objects that tell the linker and runtime which code fragments are
20939           required.
20940
20941       -mtiny-printf
20942           Enable reduced code size "printf" and "puts" library functions.
20943           The tiny implementations of these functions are not reentrant, so
20944           must be used with caution in multi-threaded applications.
20945
20946           Support for streams has been removed and the string to be printed
20947           will always be sent to stdout via the "write" syscall.  The string
20948           is not buffered before it is sent to write.
20949
20950           This option requires Newlib Nano IO, so GCC must be configured with
20951           --enable-newlib-nano-formatted-io.
20952
20953       -mcode-region=
20954       -mdata-region=
20955           These options tell the compiler where to place functions and data
20956           that do not have one of the "lower", "upper", "either" or "section"
20957           attributes.  Possible values are "lower", "upper", "either" or
20958           "any".  The first three behave like the corresponding attribute.
20959           The fourth possible value - "any" - is the default.  It leaves
20960           placement entirely up to the linker script and how it assigns the
20961           standard sections (".text", ".data", etc) to the memory regions.
20962
20963       -msilicon-errata=
20964           This option passes on a request to assembler to enable the fixes
20965           for the named silicon errata.
20966
20967       -msilicon-errata-warn=
20968           This option passes on a request to the assembler to enable warning
20969           messages when a silicon errata might need to be applied.
20970
20971       -mwarn-devices-csv
20972       -mno-warn-devices-csv
20973           Warn if devices.csv is not found or there are problem parsing it
20974           (default: on).
20975
20976       NDS32 Options
20977
20978       These options are defined for NDS32 implementations:
20979
20980       -mbig-endian
20981           Generate code in big-endian mode.
20982
20983       -mlittle-endian
20984           Generate code in little-endian mode.
20985
20986       -mreduced-regs
20987           Use reduced-set registers for register allocation.
20988
20989       -mfull-regs
20990           Use full-set registers for register allocation.
20991
20992       -mcmov
20993           Generate conditional move instructions.
20994
20995       -mno-cmov
20996           Do not generate conditional move instructions.
20997
20998       -mext-perf
20999           Generate performance extension instructions.
21000
21001       -mno-ext-perf
21002           Do not generate performance extension instructions.
21003
21004       -mext-perf2
21005           Generate performance extension 2 instructions.
21006
21007       -mno-ext-perf2
21008           Do not generate performance extension 2 instructions.
21009
21010       -mext-string
21011           Generate string extension instructions.
21012
21013       -mno-ext-string
21014           Do not generate string extension instructions.
21015
21016       -mv3push
21017           Generate v3 push25/pop25 instructions.
21018
21019       -mno-v3push
21020           Do not generate v3 push25/pop25 instructions.
21021
21022       -m16-bit
21023           Generate 16-bit instructions.
21024
21025       -mno-16-bit
21026           Do not generate 16-bit instructions.
21027
21028       -misr-vector-size=num
21029           Specify the size of each interrupt vector, which must be 4 or 16.
21030
21031       -mcache-block-size=num
21032           Specify the size of each cache block, which must be a power of 2
21033           between 4 and 512.
21034
21035       -march=arch
21036           Specify the name of the target architecture.
21037
21038       -mcmodel=code-model
21039           Set the code model to one of
21040
21041           small
21042               All the data and read-only data segments must be within 512KB
21043               addressing space.  The text segment must be within 16MB
21044               addressing space.
21045
21046           medium
21047               The data segment must be within 512KB while the read-only data
21048               segment can be within 4GB addressing space.  The text segment
21049               should be still within 16MB addressing space.
21050
21051           large
21052               All the text and data segments can be within 4GB addressing
21053               space.
21054
21055       -mctor-dtor
21056           Enable constructor/destructor feature.
21057
21058       -mrelax
21059           Guide linker to relax instructions.
21060
21061       Nios II Options
21062
21063       These are the options defined for the Altera Nios II processor.
21064
21065       -G num
21066           Put global and static objects less than or equal to num bytes into
21067           the small data or BSS sections instead of the normal data or BSS
21068           sections.  The default value of num is 8.
21069
21070       -mgpopt=option
21071       -mgpopt
21072       -mno-gpopt
21073           Generate (do not generate) GP-relative accesses.  The following
21074           option names are recognized:
21075
21076           none
21077               Do not generate GP-relative accesses.
21078
21079           local
21080               Generate GP-relative accesses for small data objects that are
21081               not external, weak, or uninitialized common symbols.  Also use
21082               GP-relative addressing for objects that have been explicitly
21083               placed in a small data section via a "section" attribute.
21084
21085           global
21086               As for local, but also generate GP-relative accesses for small
21087               data objects that are external, weak, or common.  If you use
21088               this option, you must ensure that all parts of your program
21089               (including libraries) are compiled with the same -G setting.
21090
21091           data
21092               Generate GP-relative accesses for all data objects in the
21093               program.  If you use this option, the entire data and BSS
21094               segments of your program must fit in 64K of memory and you must
21095               use an appropriate linker script to allocate them within the
21096               addressable range of the global pointer.
21097
21098           all Generate GP-relative addresses for function pointers as well as
21099               data pointers.  If you use this option, the entire text, data,
21100               and BSS segments of your program must fit in 64K of memory and
21101               you must use an appropriate linker script to allocate them
21102               within the addressable range of the global pointer.
21103
21104           -mgpopt is equivalent to -mgpopt=local, and -mno-gpopt is
21105           equivalent to -mgpopt=none.
21106
21107           The default is -mgpopt except when -fpic or -fPIC is specified to
21108           generate position-independent code.  Note that the Nios II ABI does
21109           not permit GP-relative accesses from shared libraries.
21110
21111           You may need to specify -mno-gpopt explicitly when building
21112           programs that include large amounts of small data, including large
21113           GOT data sections.  In this case, the 16-bit offset for GP-relative
21114           addressing may not be large enough to allow access to the entire
21115           small data section.
21116
21117       -mgprel-sec=regexp
21118           This option specifies additional section names that can be accessed
21119           via GP-relative addressing.  It is most useful in conjunction with
21120           "section" attributes on variable declarations and a custom linker
21121           script.  The regexp is a POSIX Extended Regular Expression.
21122
21123           This option does not affect the behavior of the -G option, and the
21124           specified sections are in addition to the standard ".sdata" and
21125           ".sbss" small-data sections that are recognized by -mgpopt.
21126
21127       -mr0rel-sec=regexp
21128           This option specifies names of sections that can be accessed via a
21129           16-bit offset from "r0"; that is, in the low 32K or high 32K of the
21130           32-bit address space.  It is most useful in conjunction with
21131           "section" attributes on variable declarations and a custom linker
21132           script.  The regexp is a POSIX Extended Regular Expression.
21133
21134           In contrast to the use of GP-relative addressing for small data,
21135           zero-based addressing is never generated by default and there are
21136           no conventional section names used in standard linker scripts for
21137           sections in the low or high areas of memory.
21138
21139       -mel
21140       -meb
21141           Generate little-endian (default) or big-endian (experimental) code,
21142           respectively.
21143
21144       -march=arch
21145           This specifies the name of the target Nios II architecture.  GCC
21146           uses this name to determine what kind of instructions it can emit
21147           when generating assembly code.  Permissible names are: r1, r2.
21148
21149           The preprocessor macro "__nios2_arch__" is available to programs,
21150           with value 1 or 2, indicating the targeted ISA level.
21151
21152       -mbypass-cache
21153       -mno-bypass-cache
21154           Force all load and store instructions to always bypass cache by
21155           using I/O variants of the instructions. The default is not to
21156           bypass the cache.
21157
21158       -mno-cache-volatile
21159       -mcache-volatile
21160           Volatile memory access bypass the cache using the I/O variants of
21161           the load and store instructions. The default is not to bypass the
21162           cache.
21163
21164       -mno-fast-sw-div
21165       -mfast-sw-div
21166           Do not use table-based fast divide for small numbers. The default
21167           is to use the fast divide at -O3 and above.
21168
21169       -mno-hw-mul
21170       -mhw-mul
21171       -mno-hw-mulx
21172       -mhw-mulx
21173       -mno-hw-div
21174       -mhw-div
21175           Enable or disable emitting "mul", "mulx" and "div" family of
21176           instructions by the compiler. The default is to emit "mul" and not
21177           emit "div" and "mulx".
21178
21179       -mbmx
21180       -mno-bmx
21181       -mcdx
21182       -mno-cdx
21183           Enable or disable generation of Nios II R2 BMX (bit manipulation)
21184           and CDX (code density) instructions.  Enabling these instructions
21185           also requires -march=r2.  Since these instructions are optional
21186           extensions to the R2 architecture, the default is not to emit them.
21187
21188       -mcustom-insn=N
21189       -mno-custom-insn
21190           Each -mcustom-insn=N option enables use of a custom instruction
21191           with encoding N when generating code that uses insn.  For example,
21192           -mcustom-fadds=253 generates custom instruction 253 for single-
21193           precision floating-point add operations instead of the default
21194           behavior of using a library call.
21195
21196           The following values of insn are supported.  Except as otherwise
21197           noted, floating-point operations are expected to be implemented
21198           with normal IEEE 754 semantics and correspond directly to the C
21199           operators or the equivalent GCC built-in functions.
21200
21201           Single-precision floating point:
21202
21203           fadds, fsubs, fdivs, fmuls
21204               Binary arithmetic operations.
21205
21206           fnegs
21207               Unary negation.
21208
21209           fabss
21210               Unary absolute value.
21211
21212           fcmpeqs, fcmpges, fcmpgts, fcmples, fcmplts, fcmpnes
21213               Comparison operations.
21214
21215           fmins, fmaxs
21216               Floating-point minimum and maximum.  These instructions are
21217               only generated if -ffinite-math-only is specified.
21218
21219           fsqrts
21220               Unary square root operation.
21221
21222           fcoss, fsins, ftans, fatans, fexps, flogs
21223               Floating-point trigonometric and exponential functions.  These
21224               instructions are only generated if -funsafe-math-optimizations
21225               is also specified.
21226
21227           Double-precision floating point:
21228
21229           faddd, fsubd, fdivd, fmuld
21230               Binary arithmetic operations.
21231
21232           fnegd
21233               Unary negation.
21234
21235           fabsd
21236               Unary absolute value.
21237
21238           fcmpeqd, fcmpged, fcmpgtd, fcmpled, fcmpltd, fcmpned
21239               Comparison operations.
21240
21241           fmind, fmaxd
21242               Double-precision minimum and maximum.  These instructions are
21243               only generated if -ffinite-math-only is specified.
21244
21245           fsqrtd
21246               Unary square root operation.
21247
21248           fcosd, fsind, ftand, fatand, fexpd, flogd
21249               Double-precision trigonometric and exponential functions.
21250               These instructions are only generated if
21251               -funsafe-math-optimizations is also specified.
21252
21253           Conversions:
21254
21255           fextsd
21256               Conversion from single precision to double precision.
21257
21258           ftruncds
21259               Conversion from double precision to single precision.
21260
21261           fixsi, fixsu, fixdi, fixdu
21262               Conversion from floating point to signed or unsigned integer
21263               types, with truncation towards zero.
21264
21265           round
21266               Conversion from single-precision floating point to signed
21267               integer, rounding to the nearest integer and ties away from
21268               zero.  This corresponds to the "__builtin_lroundf" function
21269               when -fno-math-errno is used.
21270
21271           floatis, floatus, floatid, floatud
21272               Conversion from signed or unsigned integer types to floating-
21273               point types.
21274
21275           In addition, all of the following transfer instructions for
21276           internal registers X and Y must be provided to use any of the
21277           double-precision floating-point instructions.  Custom instructions
21278           taking two double-precision source operands expect the first
21279           operand in the 64-bit register X.  The other operand (or only
21280           operand of a unary operation) is given to the custom arithmetic
21281           instruction with the least significant half in source register src1
21282           and the most significant half in src2.  A custom instruction that
21283           returns a double-precision result returns the most significant 32
21284           bits in the destination register and the other half in 32-bit
21285           register Y.  GCC automatically generates the necessary code
21286           sequences to write register X and/or read register Y when double-
21287           precision floating-point instructions are used.
21288
21289           fwrx
21290               Write src1 into the least significant half of X and src2 into
21291               the most significant half of X.
21292
21293           fwry
21294               Write src1 into Y.
21295
21296           frdxhi, frdxlo
21297               Read the most or least (respectively) significant half of X and
21298               store it in dest.
21299
21300           frdy
21301               Read the value of Y and store it into dest.
21302
21303           Note that you can gain more local control over generation of Nios
21304           II custom instructions by using the "target("custom-insn=N")" and
21305           "target("no-custom-insn")" function attributes or pragmas.
21306
21307       -mcustom-fpu-cfg=name
21308           This option enables a predefined, named set of custom instruction
21309           encodings (see -mcustom-insn above).  Currently, the following sets
21310           are defined:
21311
21312           -mcustom-fpu-cfg=60-1 is equivalent to: -mcustom-fmuls=252
21313           -mcustom-fadds=253 -mcustom-fsubs=254 -fsingle-precision-constant
21314
21315           -mcustom-fpu-cfg=60-2 is equivalent to: -mcustom-fmuls=252
21316           -mcustom-fadds=253 -mcustom-fsubs=254 -mcustom-fdivs=255
21317           -fsingle-precision-constant
21318
21319           -mcustom-fpu-cfg=72-3 is equivalent to: -mcustom-floatus=243
21320           -mcustom-fixsi=244 -mcustom-floatis=245 -mcustom-fcmpgts=246
21321           -mcustom-fcmples=249 -mcustom-fcmpeqs=250 -mcustom-fcmpnes=251
21322           -mcustom-fmuls=252 -mcustom-fadds=253 -mcustom-fsubs=254
21323           -mcustom-fdivs=255 -fsingle-precision-constant
21324
21325           Custom instruction assignments given by individual -mcustom-insn=
21326           options override those given by -mcustom-fpu-cfg=, regardless of
21327           the order of the options on the command line.
21328
21329           Note that you can gain more local control over selection of a FPU
21330           configuration by using the "target("custom-fpu-cfg=name")" function
21331           attribute or pragma.
21332
21333       These additional -m options are available for the Altera Nios II ELF
21334       (bare-metal) target:
21335
21336       -mhal
21337           Link with HAL BSP.  This suppresses linking with the GCC-provided C
21338           runtime startup and termination code, and is typically used in
21339           conjunction with -msys-crt0= to specify the location of the
21340           alternate startup code provided by the HAL BSP.
21341
21342       -msmallc
21343           Link with a limited version of the C library, -lsmallc, rather than
21344           Newlib.
21345
21346       -msys-crt0=startfile
21347           startfile is the file name of the startfile (crt0) to use when
21348           linking.  This option is only useful in conjunction with -mhal.
21349
21350       -msys-lib=systemlib
21351           systemlib is the library name of the library that provides low-
21352           level system calls required by the C library, e.g. "read" and
21353           "write".  This option is typically used to link with a library
21354           provided by a HAL BSP.
21355
21356       Nvidia PTX Options
21357
21358       These options are defined for Nvidia PTX:
21359
21360       -m32
21361       -m64
21362           Generate code for 32-bit or 64-bit ABI.
21363
21364       -misa=ISA-string
21365           Generate code for given the specified PTX ISA (e.g. sm_35).  ISA
21366           strings must be lower-case.  Valid ISA strings include sm_30 and
21367           sm_35.  The default ISA is sm_30.
21368
21369       -mmainkernel
21370           Link in code for a __main kernel.  This is for stand-alone instead
21371           of offloading execution.
21372
21373       -moptimize
21374           Apply partitioned execution optimizations.  This is the default
21375           when any level of optimization is selected.
21376
21377       -msoft-stack
21378           Generate code that does not use ".local" memory directly for stack
21379           storage. Instead, a per-warp stack pointer is maintained
21380           explicitly. This enables variable-length stack allocation (with
21381           variable-length arrays or "alloca"), and when global memory is used
21382           for underlying storage, makes it possible to access automatic
21383           variables from other threads, or with atomic instructions. This
21384           code generation variant is used for OpenMP offloading, but the
21385           option is exposed on its own for the purpose of testing the
21386           compiler; to generate code suitable for linking into programs using
21387           OpenMP offloading, use option -mgomp.
21388
21389       -muniform-simt
21390           Switch to code generation variant that allows to execute all
21391           threads in each warp, while maintaining memory state and side
21392           effects as if only one thread in each warp was active outside of
21393           OpenMP SIMD regions.  All atomic operations and calls to runtime
21394           (malloc, free, vprintf) are conditionally executed (iff current
21395           lane index equals the master lane index), and the register being
21396           assigned is copied via a shuffle instruction from the master lane.
21397           Outside of SIMD regions lane 0 is the master; inside, each thread
21398           sees itself as the master.  Shared memory array "int __nvptx_uni[]"
21399           stores all-zeros or all-ones bitmasks for each warp, indicating
21400           current mode (0 outside of SIMD regions).  Each thread can bitwise-
21401           and the bitmask at position "tid.y" with current lane index to
21402           compute the master lane index.
21403
21404       -mgomp
21405           Generate code for use in OpenMP offloading: enables -msoft-stack
21406           and -muniform-simt options, and selects corresponding multilib
21407           variant.
21408
21409       OpenRISC Options
21410
21411       These options are defined for OpenRISC:
21412
21413       -mboard=name
21414           Configure a board specific runtime.  This will be passed to the
21415           linker for newlib board library linking.  The default is "or1ksim".
21416
21417       -mnewlib
21418           This option is ignored; it is for compatibility purposes only.
21419           This used to select linker and preprocessor options for use with
21420           newlib.
21421
21422       -msoft-div
21423       -mhard-div
21424           Select software or hardware divide ("l.div", "l.divu")
21425           instructions.  This default is hardware divide.
21426
21427       -msoft-mul
21428       -mhard-mul
21429           Select software or hardware multiply ("l.mul", "l.muli")
21430           instructions.  This default is hardware multiply.
21431
21432       -msoft-float
21433       -mhard-float
21434           Select software or hardware for floating point operations.  The
21435           default is software.
21436
21437       -mdouble-float
21438           When -mhard-float is selected, enables generation of double-
21439           precision floating point instructions.  By default functions from
21440           libgcc are used to perform double-precision floating point
21441           operations.
21442
21443       -munordered-float
21444           When -mhard-float is selected, enables generation of unordered
21445           floating point compare and set flag ("lf.sfun*") instructions.  By
21446           default functions from libgcc are used to perform unordered
21447           floating point compare and set flag operations.
21448
21449       -mcmov
21450           Enable generation of conditional move ("l.cmov") instructions.  By
21451           default the equivalent will be generated using set and branch.
21452
21453       -mror
21454           Enable generation of rotate right ("l.ror") instructions.  By
21455           default functions from libgcc are used to perform rotate right
21456           operations.
21457
21458       -mrori
21459           Enable generation of rotate right with immediate ("l.rori")
21460           instructions.  By default functions from libgcc are used to perform
21461           rotate right with immediate operations.
21462
21463       -msext
21464           Enable generation of sign extension ("l.ext*") instructions.  By
21465           default memory loads are used to perform sign extension.
21466
21467       -msfimm
21468           Enable generation of compare and set flag with immediate ("l.sf*i")
21469           instructions.  By default extra instructions will be generated to
21470           store the immediate to a register first.
21471
21472       -mshftimm
21473           Enable generation of shift with immediate ("l.srai", "l.srli",
21474           "l.slli") instructions.  By default extra instructions will be
21475           generated to store the immediate to a register first.
21476
21477       PDP-11 Options
21478
21479       These options are defined for the PDP-11:
21480
21481       -mfpu
21482           Use hardware FPP floating point.  This is the default.  (FIS
21483           floating point on the PDP-11/40 is not supported.)  Implies -m45.
21484
21485       -msoft-float
21486           Do not use hardware floating point.
21487
21488       -mac0
21489           Return floating-point results in ac0 (fr0 in Unix assembler
21490           syntax).
21491
21492       -mno-ac0
21493           Return floating-point results in memory.  This is the default.
21494
21495       -m40
21496           Generate code for a PDP-11/40.  Implies -msoft-float -mno-split.
21497
21498       -m45
21499           Generate code for a PDP-11/45.  This is the default.
21500
21501       -m10
21502           Generate code for a PDP-11/10.  Implies -msoft-float -mno-split.
21503
21504       -mint16
21505       -mno-int32
21506           Use 16-bit "int".  This is the default.
21507
21508       -mint32
21509       -mno-int16
21510           Use 32-bit "int".
21511
21512       -msplit
21513           Target has split instruction and data space.  Implies -m45.
21514
21515       -munix-asm
21516           Use Unix assembler syntax.
21517
21518       -mdec-asm
21519           Use DEC assembler syntax.
21520
21521       -mgnu-asm
21522           Use GNU assembler syntax.  This is the default.
21523
21524       -mlra
21525           Use the new LRA register allocator.  By default, the old "reload"
21526           allocator is used.
21527
21528       picoChip Options
21529
21530       These -m options are defined for picoChip implementations:
21531
21532       -mae=ae_type
21533           Set the instruction set, register set, and instruction scheduling
21534           parameters for array element type ae_type.  Supported values for
21535           ae_type are ANY, MUL, and MAC.
21536
21537           -mae=ANY selects a completely generic AE type.  Code generated with
21538           this option runs on any of the other AE types.  The code is not as
21539           efficient as it would be if compiled for a specific AE type, and
21540           some types of operation (e.g., multiplication) do not work properly
21541           on all types of AE.
21542
21543           -mae=MUL selects a MUL AE type.  This is the most useful AE type
21544           for compiled code, and is the default.
21545
21546           -mae=MAC selects a DSP-style MAC AE.  Code compiled with this
21547           option may suffer from poor performance of byte (char)
21548           manipulation, since the DSP AE does not provide hardware support
21549           for byte load/stores.
21550
21551       -msymbol-as-address
21552           Enable the compiler to directly use a symbol name as an address in
21553           a load/store instruction, without first loading it into a register.
21554           Typically, the use of this option generates larger programs, which
21555           run faster than when the option isn't used.  However, the results
21556           vary from program to program, so it is left as a user option,
21557           rather than being permanently enabled.
21558
21559       -mno-inefficient-warnings
21560           Disables warnings about the generation of inefficient code.  These
21561           warnings can be generated, for example, when compiling code that
21562           performs byte-level memory operations on the MAC AE type.  The MAC
21563           AE has no hardware support for byte-level memory operations, so all
21564           byte load/stores must be synthesized from word load/store
21565           operations.  This is inefficient and a warning is generated to
21566           indicate that you should rewrite the code to avoid byte operations,
21567           or to target an AE type that has the necessary hardware support.
21568           This option disables these warnings.
21569
21570       PowerPC Options
21571
21572       These are listed under
21573
21574       PRU Options
21575
21576       These command-line options are defined for PRU target:
21577
21578       -minrt
21579           Link with a minimum runtime environment, with no support for static
21580           initializers and constructors.  Using this option can significantly
21581           reduce the size of the final ELF binary.  Beware that the compiler
21582           could still generate code with static initializers and
21583           constructors.  It is up to the programmer to ensure that the source
21584           program will not use those features.
21585
21586       -mmcu=mcu
21587           Specify the PRU MCU variant to use.  Check Newlib for the exact
21588           list of supported MCUs.
21589
21590       -mno-relax
21591           Make GCC pass the --no-relax command-line option to the linker
21592           instead of the --relax option.
21593
21594       -mloop
21595           Allow (or do not allow) GCC to use the LOOP instruction.
21596
21597       -mabi=variant
21598           Specify the ABI variant to output code for.  -mabi=ti selects the
21599           unmodified TI ABI while -mabi=gnu selects a GNU variant that copes
21600           more naturally with certain GCC assumptions.  These are the
21601           differences:
21602
21603           Function Pointer Size
21604               TI ABI specifies that function (code) pointers are 16-bit,
21605               whereas GNU supports only 32-bit data and code pointers.
21606
21607           Optional Return Value Pointer
21608               Function return values larger than 64 bits are passed by using
21609               a hidden pointer as the first argument of the function.  TI
21610               ABI, though, mandates that the pointer can be NULL in case the
21611               caller is not using the returned value.  GNU always passes and
21612               expects a valid return value pointer.
21613
21614           The current -mabi=ti implementation simply raises a compile error
21615           when any of the above code constructs is detected.  As a
21616           consequence the standard C library cannot be built and it is
21617           omitted when linking with -mabi=ti.
21618
21619           Relaxation is a GNU feature and for safety reasons is disabled when
21620           using -mabi=ti.  The TI toolchain does not emit relocations for
21621           QBBx instructions, so the GNU linker cannot adjust them when
21622           shortening adjacent LDI32 pseudo instructions.
21623
21624       RISC-V Options
21625
21626       These command-line options are defined for RISC-V targets:
21627
21628       -mbranch-cost=n
21629           Set the cost of branches to roughly n instructions.
21630
21631       -mplt
21632       -mno-plt
21633           When generating PIC code, do or don't allow the use of PLTs.
21634           Ignored for non-PIC.  The default is -mplt.
21635
21636       -mabi=ABI-string
21637           Specify integer and floating-point calling convention.  ABI-string
21638           contains two parts: the size of integer types and the registers
21639           used for floating-point types.  For example -march=rv64ifd
21640           -mabi=lp64d means that long and pointers are 64-bit (implicitly
21641           defining int to be 32-bit), and that floating-point values up to 64
21642           bits wide are passed in F registers.  Contrast this with
21643           -march=rv64ifd -mabi=lp64f, which still allows the compiler to
21644           generate code that uses the F and D extensions but only allows
21645           floating-point values up to 32 bits long to be passed in registers;
21646           or -march=rv64ifd -mabi=lp64, in which no floating-point arguments
21647           will be passed in registers.
21648
21649           The default for this argument is system dependent, users who want a
21650           specific calling convention should specify one explicitly.  The
21651           valid calling conventions are: ilp32, ilp32f, ilp32d, lp64, lp64f,
21652           and lp64d.  Some calling conventions are impossible to implement on
21653           some ISAs: for example, -march=rv32if -mabi=ilp32d is invalid
21654           because the ABI requires 64-bit values be passed in F registers,
21655           but F registers are only 32 bits wide.  There is also the ilp32e
21656           ABI that can only be used with the rv32e architecture.  This ABI is
21657           not well specified at present, and is subject to change.
21658
21659       -mfdiv
21660       -mno-fdiv
21661           Do or don't use hardware floating-point divide and square root
21662           instructions.  This requires the F or D extensions for floating-
21663           point registers.  The default is to use them if the specified
21664           architecture has these instructions.
21665
21666       -mdiv
21667       -mno-div
21668           Do or don't use hardware instructions for integer division.  This
21669           requires the M extension.  The default is to use them if the
21670           specified architecture has these instructions.
21671
21672       -march=ISA-string
21673           Generate code for given RISC-V ISA (e.g. rv64im).  ISA strings must
21674           be lower-case.  Examples include rv64i, rv32g, rv32e, and rv32imaf.
21675
21676       -mtune=processor-string
21677           Optimize the output for the given processor, specified by
21678           microarchitecture name.  Permissible values for this option are:
21679           rocket, sifive-3-series, sifive-5-series, sifive-7-series, and
21680           size.
21681
21682           When -mtune= is not specified, the default is rocket.
21683
21684           The size choice is not intended for use by end-users.  This is used
21685           when -Os is specified.  It overrides the instruction cost info
21686           provided by -mtune=, but does not override the pipeline info.  This
21687           helps reduce code size while still giving good performance.
21688
21689       -mpreferred-stack-boundary=num
21690           Attempt to keep the stack boundary aligned to a 2 raised to num
21691           byte boundary.  If -mpreferred-stack-boundary is not specified, the
21692           default is 4 (16 bytes or 128-bits).
21693
21694           Warning: If you use this switch, then you must build all modules
21695           with the same value, including any libraries.  This includes the
21696           system libraries and startup modules.
21697
21698       -msmall-data-limit=n
21699           Put global and static data smaller than n bytes into a special
21700           section (on some targets).
21701
21702       -msave-restore
21703       -mno-save-restore
21704           Do or don't use smaller but slower prologue and epilogue code that
21705           uses library function calls.  The default is to use fast inline
21706           prologues and epilogues.
21707
21708       -mstrict-align
21709       -mno-strict-align
21710           Do not or do generate unaligned memory accesses.  The default is
21711           set depending on whether the processor we are optimizing for
21712           supports fast unaligned access or not.
21713
21714       -mcmodel=medlow
21715           Generate code for the medium-low code model. The program and its
21716           statically defined symbols must lie within a single 2 GiB address
21717           range and must lie between absolute addresses -2 GiB and +2 GiB.
21718           Programs can be statically or dynamically linked. This is the
21719           default code model.
21720
21721       -mcmodel=medany
21722           Generate code for the medium-any code model. The program and its
21723           statically defined symbols must be within any single 2 GiB address
21724           range. Programs can be statically or dynamically linked.
21725
21726       -mexplicit-relocs
21727       -mno-exlicit-relocs
21728           Use or do not use assembler relocation operators when dealing with
21729           symbolic addresses.  The alternative is to use assembler macros
21730           instead, which may limit optimization.
21731
21732       -mrelax
21733       -mno-relax
21734           Take advantage of linker relaxations to reduce the number of
21735           instructions required to materialize symbol addresses. The default
21736           is to take advantage of linker relaxations.
21737
21738       -memit-attribute
21739       -mno-emit-attribute
21740           Emit (do not emit) RISC-V attribute to record extra information
21741           into ELF objects.  This feature requires at least binutils 2.32.
21742
21743       -malign-data=type
21744           Control how GCC aligns variables and constants of array, structure,
21745           or union types.  Supported values for type are xlen which uses x
21746           register width as the alignment value, and natural which uses
21747           natural alignment.  xlen is the default.
21748
21749       RL78 Options
21750
21751       -msim
21752           Links in additional target libraries to support operation within a
21753           simulator.
21754
21755       -mmul=none
21756       -mmul=g10
21757       -mmul=g13
21758       -mmul=g14
21759       -mmul=rl78
21760           Specifies the type of hardware multiplication and division support
21761           to be used.  The simplest is "none", which uses software for both
21762           multiplication and division.  This is the default.  The "g13" value
21763           is for the hardware multiply/divide peripheral found on the
21764           RL78/G13 (S2 core) targets.  The "g14" value selects the use of the
21765           multiplication and division instructions supported by the RL78/G14
21766           (S3 core) parts.  The value "rl78" is an alias for "g14" and the
21767           value "mg10" is an alias for "none".
21768
21769           In addition a C preprocessor macro is defined, based upon the
21770           setting of this option.  Possible values are: "__RL78_MUL_NONE__",
21771           "__RL78_MUL_G13__" or "__RL78_MUL_G14__".
21772
21773       -mcpu=g10
21774       -mcpu=g13
21775       -mcpu=g14
21776       -mcpu=rl78
21777           Specifies the RL78 core to target.  The default is the G14 core,
21778           also known as an S3 core or just RL78.  The G13 or S2 core does not
21779           have multiply or divide instructions, instead it uses a hardware
21780           peripheral for these operations.  The G10 or S1 core does not have
21781           register banks, so it uses a different calling convention.
21782
21783           If this option is set it also selects the type of hardware multiply
21784           support to use, unless this is overridden by an explicit -mmul=none
21785           option on the command line.  Thus specifying -mcpu=g13 enables the
21786           use of the G13 hardware multiply peripheral and specifying
21787           -mcpu=g10 disables the use of hardware multiplications altogether.
21788
21789           Note, although the RL78/G14 core is the default target, specifying
21790           -mcpu=g14 or -mcpu=rl78 on the command line does change the
21791           behavior of the toolchain since it also enables G14 hardware
21792           multiply support.  If these options are not specified on the
21793           command line then software multiplication routines will be used
21794           even though the code targets the RL78 core.  This is for backwards
21795           compatibility with older toolchains which did not have hardware
21796           multiply and divide support.
21797
21798           In addition a C preprocessor macro is defined, based upon the
21799           setting of this option.  Possible values are: "__RL78_G10__",
21800           "__RL78_G13__" or "__RL78_G14__".
21801
21802       -mg10
21803       -mg13
21804       -mg14
21805       -mrl78
21806           These are aliases for the corresponding -mcpu= option.  They are
21807           provided for backwards compatibility.
21808
21809       -mallregs
21810           Allow the compiler to use all of the available registers.  By
21811           default registers "r24..r31" are reserved for use in interrupt
21812           handlers.  With this option enabled these registers can be used in
21813           ordinary functions as well.
21814
21815       -m64bit-doubles
21816       -m32bit-doubles
21817           Make the "double" data type be 64 bits (-m64bit-doubles) or 32 bits
21818           (-m32bit-doubles) in size.  The default is -m32bit-doubles.
21819
21820       -msave-mduc-in-interrupts
21821       -mno-save-mduc-in-interrupts
21822           Specifies that interrupt handler functions should preserve the MDUC
21823           registers.  This is only necessary if normal code might use the
21824           MDUC registers, for example because it performs multiplication and
21825           division operations.  The default is to ignore the MDUC registers
21826           as this makes the interrupt handlers faster.  The target option
21827           -mg13 needs to be passed for this to work as this feature is only
21828           available on the G13 target (S2 core).  The MDUC registers will
21829           only be saved if the interrupt handler performs a multiplication or
21830           division operation or it calls another function.
21831
21832       IBM RS/6000 and PowerPC Options
21833
21834       These -m options are defined for the IBM RS/6000 and PowerPC:
21835
21836       -mpowerpc-gpopt
21837       -mno-powerpc-gpopt
21838       -mpowerpc-gfxopt
21839       -mno-powerpc-gfxopt
21840       -mpowerpc64
21841       -mno-powerpc64
21842       -mmfcrf
21843       -mno-mfcrf
21844       -mpopcntb
21845       -mno-popcntb
21846       -mpopcntd
21847       -mno-popcntd
21848       -mfprnd
21849       -mno-fprnd
21850       -mcmpb
21851       -mno-cmpb
21852       -mhard-dfp
21853       -mno-hard-dfp
21854           You use these options to specify which instructions are available
21855           on the processor you are using.  The default value of these options
21856           is determined when configuring GCC.  Specifying the -mcpu=cpu_type
21857           overrides the specification of these options.  We recommend you use
21858           the -mcpu=cpu_type option rather than the options listed above.
21859
21860           Specifying -mpowerpc-gpopt allows GCC to use the optional PowerPC
21861           architecture instructions in the General Purpose group, including
21862           floating-point square root.  Specifying -mpowerpc-gfxopt allows GCC
21863           to use the optional PowerPC architecture instructions in the
21864           Graphics group, including floating-point select.
21865
21866           The -mmfcrf option allows GCC to generate the move from condition
21867           register field instruction implemented on the POWER4 processor and
21868           other processors that support the PowerPC V2.01 architecture.  The
21869           -mpopcntb option allows GCC to generate the popcount and double-
21870           precision FP reciprocal estimate instruction implemented on the
21871           POWER5 processor and other processors that support the PowerPC
21872           V2.02 architecture.  The -mpopcntd option allows GCC to generate
21873           the popcount instruction implemented on the POWER7 processor and
21874           other processors that support the PowerPC V2.06 architecture.  The
21875           -mfprnd option allows GCC to generate the FP round to integer
21876           instructions implemented on the POWER5+ processor and other
21877           processors that support the PowerPC V2.03 architecture.  The -mcmpb
21878           option allows GCC to generate the compare bytes instruction
21879           implemented on the POWER6 processor and other processors that
21880           support the PowerPC V2.05 architecture.  The -mhard-dfp option
21881           allows GCC to generate the decimal floating-point instructions
21882           implemented on some POWER processors.
21883
21884           The -mpowerpc64 option allows GCC to generate the additional 64-bit
21885           instructions that are found in the full PowerPC64 architecture and
21886           to treat GPRs as 64-bit, doubleword quantities.  GCC defaults to
21887           -mno-powerpc64.
21888
21889       -mcpu=cpu_type
21890           Set architecture type, register usage, and instruction scheduling
21891           parameters for machine type cpu_type.  Supported values for
21892           cpu_type are 401, 403, 405, 405fp, 440, 440fp, 464, 464fp, 476,
21893           476fp, 505, 601, 602, 603, 603e, 604, 604e, 620, 630, 740, 7400,
21894           7450, 750, 801, 821, 823, 860, 970, 8540, a2, e300c2, e300c3,
21895           e500mc, e500mc64, e5500, e6500, ec603e, G3, G4, G5, titan, power3,
21896           power4, power5, power5+, power6, power6x, power7, power8, power9,
21897           future, powerpc, powerpc64, powerpc64le, rs64, and native.
21898
21899           -mcpu=powerpc, -mcpu=powerpc64, and -mcpu=powerpc64le specify pure
21900           32-bit PowerPC (either endian), 64-bit big endian PowerPC and
21901           64-bit little endian PowerPC architecture machine types, with an
21902           appropriate, generic processor model assumed for scheduling
21903           purposes.
21904
21905           Specifying native as cpu type detects and selects the architecture
21906           option that corresponds to the host processor of the system
21907           performing the compilation.  -mcpu=native has no effect if GCC does
21908           not recognize the processor.
21909
21910           The other options specify a specific processor.  Code generated
21911           under those options runs best on that processor, and may not run at
21912           all on others.
21913
21914           The -mcpu options automatically enable or disable the following
21915           options:
21916
21917           -maltivec  -mfprnd  -mhard-float  -mmfcrf  -mmultiple -mpopcntb
21918           -mpopcntd  -mpowerpc64 -mpowerpc-gpopt  -mpowerpc-gfxopt -mmulhw
21919           -mdlmzb  -mmfpgpr  -mvsx -mcrypto  -mhtm  -mpower8-fusion
21920           -mpower8-vector -mquad-memory  -mquad-memory-atomic  -mfloat128
21921           -mfloat128-hardware -mprefixed -mpcrel -mmma
21922
21923           The particular options set for any particular CPU varies between
21924           compiler versions, depending on what setting seems to produce
21925           optimal code for that CPU; it doesn't necessarily reflect the
21926           actual hardware's capabilities.  If you wish to set an individual
21927           option to a particular value, you may specify it after the -mcpu
21928           option, like -mcpu=970 -mno-altivec.
21929
21930           On AIX, the -maltivec and -mpowerpc64 options are not enabled or
21931           disabled by the -mcpu option at present because AIX does not have
21932           full support for these options.  You may still enable or disable
21933           them individually if you're sure it'll work in your environment.
21934
21935       -mtune=cpu_type
21936           Set the instruction scheduling parameters for machine type
21937           cpu_type, but do not set the architecture type or register usage,
21938           as -mcpu=cpu_type does.  The same values for cpu_type are used for
21939           -mtune as for -mcpu.  If both are specified, the code generated
21940           uses the architecture and registers set by -mcpu, but the
21941           scheduling parameters set by -mtune.
21942
21943       -mcmodel=small
21944           Generate PowerPC64 code for the small model: The TOC is limited to
21945           64k.
21946
21947       -mcmodel=medium
21948           Generate PowerPC64 code for the medium model: The TOC and other
21949           static data may be up to a total of 4G in size.  This is the
21950           default for 64-bit Linux.
21951
21952       -mcmodel=large
21953           Generate PowerPC64 code for the large model: The TOC may be up to
21954           4G in size.  Other data and code is only limited by the 64-bit
21955           address space.
21956
21957       -maltivec
21958       -mno-altivec
21959           Generate code that uses (does not use) AltiVec instructions, and
21960           also enable the use of built-in functions that allow more direct
21961           access to the AltiVec instruction set.  You may also need to set
21962           -mabi=altivec to adjust the current ABI with AltiVec ABI
21963           enhancements.
21964
21965           When -maltivec is used, the element order for AltiVec intrinsics
21966           such as "vec_splat", "vec_extract", and "vec_insert" match array
21967           element order corresponding to the endianness of the target.  That
21968           is, element zero identifies the leftmost element in a vector
21969           register when targeting a big-endian platform, and identifies the
21970           rightmost element in a vector register when targeting a little-
21971           endian platform.
21972
21973       -mvrsave
21974       -mno-vrsave
21975           Generate VRSAVE instructions when generating AltiVec code.
21976
21977       -msecure-plt
21978           Generate code that allows ld and ld.so to build executables and
21979           shared libraries with non-executable ".plt" and ".got" sections.
21980           This is a PowerPC 32-bit SYSV ABI option.
21981
21982       -mbss-plt
21983           Generate code that uses a BSS ".plt" section that ld.so fills in,
21984           and requires ".plt" and ".got" sections that are both writable and
21985           executable.  This is a PowerPC 32-bit SYSV ABI option.
21986
21987       -misel
21988       -mno-isel
21989           This switch enables or disables the generation of ISEL
21990           instructions.
21991
21992       -mvsx
21993       -mno-vsx
21994           Generate code that uses (does not use) vector/scalar (VSX)
21995           instructions, and also enable the use of built-in functions that
21996           allow more direct access to the VSX instruction set.
21997
21998       -mcrypto
21999       -mno-crypto
22000           Enable the use (disable) of the built-in functions that allow
22001           direct access to the cryptographic instructions that were added in
22002           version 2.07 of the PowerPC ISA.
22003
22004       -mhtm
22005       -mno-htm
22006           Enable (disable) the use of the built-in functions that allow
22007           direct access to the Hardware Transactional Memory (HTM)
22008           instructions that were added in version 2.07 of the PowerPC ISA.
22009
22010       -mpower8-fusion
22011       -mno-power8-fusion
22012           Generate code that keeps (does not keeps) some integer operations
22013           adjacent so that the instructions can be fused together on power8
22014           and later processors.
22015
22016       -mpower8-vector
22017       -mno-power8-vector
22018           Generate code that uses (does not use) the vector and scalar
22019           instructions that were added in version 2.07 of the PowerPC ISA.
22020           Also enable the use of built-in functions that allow more direct
22021           access to the vector instructions.
22022
22023       -mquad-memory
22024       -mno-quad-memory
22025           Generate code that uses (does not use) the non-atomic quad word
22026           memory instructions.  The -mquad-memory option requires use of
22027           64-bit mode.
22028
22029       -mquad-memory-atomic
22030       -mno-quad-memory-atomic
22031           Generate code that uses (does not use) the atomic quad word memory
22032           instructions.  The -mquad-memory-atomic option requires use of
22033           64-bit mode.
22034
22035       -mfloat128
22036       -mno-float128
22037           Enable/disable the __float128 keyword for IEEE 128-bit floating
22038           point and use either software emulation for IEEE 128-bit floating
22039           point or hardware instructions.
22040
22041           The VSX instruction set (-mvsx, -mcpu=power7, -mcpu=power8), or
22042           -mcpu=power9 must be enabled to use the IEEE 128-bit floating point
22043           support.  The IEEE 128-bit floating point support only works on
22044           PowerPC Linux systems.
22045
22046           The default for -mfloat128 is enabled on PowerPC Linux systems
22047           using the VSX instruction set, and disabled on other systems.
22048
22049           If you use the ISA 3.0 instruction set (-mpower9-vector or
22050           -mcpu=power9) on a 64-bit system, the IEEE 128-bit floating point
22051           support will also enable the generation of ISA 3.0 IEEE 128-bit
22052           floating point instructions.  Otherwise, if you do not specify to
22053           generate ISA 3.0 instructions or you are targeting a 32-bit big
22054           endian system, IEEE 128-bit floating point will be done with
22055           software emulation.
22056
22057       -mfloat128-hardware
22058       -mno-float128-hardware
22059           Enable/disable using ISA 3.0 hardware instructions to support the
22060           __float128 data type.
22061
22062           The default for -mfloat128-hardware is enabled on PowerPC Linux
22063           systems using the ISA 3.0 instruction set, and disabled on other
22064           systems.
22065
22066       -m32
22067       -m64
22068           Generate code for 32-bit or 64-bit environments of Darwin and SVR4
22069           targets (including GNU/Linux).  The 32-bit environment sets int,
22070           long and pointer to 32 bits and generates code that runs on any
22071           PowerPC variant.  The 64-bit environment sets int to 32 bits and
22072           long and pointer to 64 bits, and generates code for PowerPC64, as
22073           for -mpowerpc64.
22074
22075       -mfull-toc
22076       -mno-fp-in-toc
22077       -mno-sum-in-toc
22078       -mminimal-toc
22079           Modify generation of the TOC (Table Of Contents), which is created
22080           for every executable file.  The -mfull-toc option is selected by
22081           default.  In that case, GCC allocates at least one TOC entry for
22082           each unique non-automatic variable reference in your program.  GCC
22083           also places floating-point constants in the TOC.  However, only
22084           16,384 entries are available in the TOC.
22085
22086           If you receive a linker error message that saying you have
22087           overflowed the available TOC space, you can reduce the amount of
22088           TOC space used with the -mno-fp-in-toc and -mno-sum-in-toc options.
22089           -mno-fp-in-toc prevents GCC from putting floating-point constants
22090           in the TOC and -mno-sum-in-toc forces GCC to generate code to
22091           calculate the sum of an address and a constant at run time instead
22092           of putting that sum into the TOC.  You may specify one or both of
22093           these options.  Each causes GCC to produce very slightly slower and
22094           larger code at the expense of conserving TOC space.
22095
22096           If you still run out of space in the TOC even when you specify both
22097           of these options, specify -mminimal-toc instead.  This option
22098           causes GCC to make only one TOC entry for every file.  When you
22099           specify this option, GCC produces code that is slower and larger
22100           but which uses extremely little TOC space.  You may wish to use
22101           this option only on files that contain less frequently-executed
22102           code.
22103
22104       -maix64
22105       -maix32
22106           Enable 64-bit AIX ABI and calling convention: 64-bit pointers,
22107           64-bit "long" type, and the infrastructure needed to support them.
22108           Specifying -maix64 implies -mpowerpc64, while -maix32 disables the
22109           64-bit ABI and implies -mno-powerpc64.  GCC defaults to -maix32.
22110
22111       -mxl-compat
22112       -mno-xl-compat
22113           Produce code that conforms more closely to IBM XL compiler
22114           semantics when using AIX-compatible ABI.  Pass floating-point
22115           arguments to prototyped functions beyond the register save area
22116           (RSA) on the stack in addition to argument FPRs.  Do not assume
22117           that most significant double in 128-bit long double value is
22118           properly rounded when comparing values and converting to double.
22119           Use XL symbol names for long double support routines.
22120
22121           The AIX calling convention was extended but not initially
22122           documented to handle an obscure K&R C case of calling a function
22123           that takes the address of its arguments with fewer arguments than
22124           declared.  IBM XL compilers access floating-point arguments that do
22125           not fit in the RSA from the stack when a subroutine is compiled
22126           without optimization.  Because always storing floating-point
22127           arguments on the stack is inefficient and rarely needed, this
22128           option is not enabled by default and only is necessary when calling
22129           subroutines compiled by IBM XL compilers without optimization.
22130
22131       -mpe
22132           Support IBM RS/6000 SP Parallel Environment (PE).  Link an
22133           application written to use message passing with special startup
22134           code to enable the application to run.  The system must have PE
22135           installed in the standard location (/usr/lpp/ppe.poe/), or the
22136           specs file must be overridden with the -specs= option to specify
22137           the appropriate directory location.  The Parallel Environment does
22138           not support threads, so the -mpe option and the -pthread option are
22139           incompatible.
22140
22141       -malign-natural
22142       -malign-power
22143           On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
22144           -malign-natural overrides the ABI-defined alignment of larger
22145           types, such as floating-point doubles, on their natural size-based
22146           boundary.  The option -malign-power instructs GCC to follow the
22147           ABI-specified alignment rules.  GCC defaults to the standard
22148           alignment defined in the ABI.
22149
22150           On 64-bit Darwin, natural alignment is the default, and
22151           -malign-power is not supported.
22152
22153       -msoft-float
22154       -mhard-float
22155           Generate code that does not use (uses) the floating-point register
22156           set.  Software floating-point emulation is provided if you use the
22157           -msoft-float option, and pass the option to GCC when linking.
22158
22159       -mmultiple
22160       -mno-multiple
22161           Generate code that uses (does not use) the load multiple word
22162           instructions and the store multiple word instructions.  These
22163           instructions are generated by default on POWER systems, and not
22164           generated on PowerPC systems.  Do not use -mmultiple on little-
22165           endian PowerPC systems, since those instructions do not work when
22166           the processor is in little-endian mode.  The exceptions are PPC740
22167           and PPC750 which permit these instructions in little-endian mode.
22168
22169       -mupdate
22170       -mno-update
22171           Generate code that uses (does not use) the load or store
22172           instructions that update the base register to the address of the
22173           calculated memory location.  These instructions are generated by
22174           default.  If you use -mno-update, there is a small window between
22175           the time that the stack pointer is updated and the address of the
22176           previous frame is stored, which means code that walks the stack
22177           frame across interrupts or signals may get corrupted data.
22178
22179       -mavoid-indexed-addresses
22180       -mno-avoid-indexed-addresses
22181           Generate code that tries to avoid (not avoid) the use of indexed
22182           load or store instructions. These instructions can incur a
22183           performance penalty on Power6 processors in certain situations,
22184           such as when stepping through large arrays that cross a 16M
22185           boundary.  This option is enabled by default when targeting Power6
22186           and disabled otherwise.
22187
22188       -mfused-madd
22189       -mno-fused-madd
22190           Generate code that uses (does not use) the floating-point multiply
22191           and accumulate instructions.  These instructions are generated by
22192           default if hardware floating point is used.  The machine-dependent
22193           -mfused-madd option is now mapped to the machine-independent
22194           -ffp-contract=fast option, and -mno-fused-madd is mapped to
22195           -ffp-contract=off.
22196
22197       -mmulhw
22198       -mno-mulhw
22199           Generate code that uses (does not use) the half-word multiply and
22200           multiply-accumulate instructions on the IBM 405, 440, 464 and 476
22201           processors.  These instructions are generated by default when
22202           targeting those processors.
22203
22204       -mdlmzb
22205       -mno-dlmzb
22206           Generate code that uses (does not use) the string-search dlmzb
22207           instruction on the IBM 405, 440, 464 and 476 processors.  This
22208           instruction is generated by default when targeting those
22209           processors.
22210
22211       -mno-bit-align
22212       -mbit-align
22213           On System V.4 and embedded PowerPC systems do not (do) force
22214           structures and unions that contain bit-fields to be aligned to the
22215           base type of the bit-field.
22216
22217           For example, by default a structure containing nothing but 8
22218           "unsigned" bit-fields of length 1 is aligned to a 4-byte boundary
22219           and has a size of 4 bytes.  By using -mno-bit-align, the structure
22220           is aligned to a 1-byte boundary and is 1 byte in size.
22221
22222       -mno-strict-align
22223       -mstrict-align
22224           On System V.4 and embedded PowerPC systems do not (do) assume that
22225           unaligned memory references are handled by the system.
22226
22227       -mrelocatable
22228       -mno-relocatable
22229           Generate code that allows (does not allow) a static executable to
22230           be relocated to a different address at run time.  A simple embedded
22231           PowerPC system loader should relocate the entire contents of
22232           ".got2" and 4-byte locations listed in the ".fixup" section, a
22233           table of 32-bit addresses generated by this option.  For this to
22234           work, all objects linked together must be compiled with
22235           -mrelocatable or -mrelocatable-lib.  -mrelocatable code aligns the
22236           stack to an 8-byte boundary.
22237
22238       -mrelocatable-lib
22239       -mno-relocatable-lib
22240           Like -mrelocatable, -mrelocatable-lib generates a ".fixup" section
22241           to allow static executables to be relocated at run time, but
22242           -mrelocatable-lib does not use the smaller stack alignment of
22243           -mrelocatable.  Objects compiled with -mrelocatable-lib may be
22244           linked with objects compiled with any combination of the
22245           -mrelocatable options.
22246
22247       -mno-toc
22248       -mtoc
22249           On System V.4 and embedded PowerPC systems do not (do) assume that
22250           register 2 contains a pointer to a global area pointing to the
22251           addresses used in the program.
22252
22253       -mlittle
22254       -mlittle-endian
22255           On System V.4 and embedded PowerPC systems compile code for the
22256           processor in little-endian mode.  The -mlittle-endian option is the
22257           same as -mlittle.
22258
22259       -mbig
22260       -mbig-endian
22261           On System V.4 and embedded PowerPC systems compile code for the
22262           processor in big-endian mode.  The -mbig-endian option is the same
22263           as -mbig.
22264
22265       -mdynamic-no-pic
22266           On Darwin and Mac OS X systems, compile code so that it is not
22267           relocatable, but that its external references are relocatable.  The
22268           resulting code is suitable for applications, but not shared
22269           libraries.
22270
22271       -msingle-pic-base
22272           Treat the register used for PIC addressing as read-only, rather
22273           than loading it in the prologue for each function.  The runtime
22274           system is responsible for initializing this register with an
22275           appropriate value before execution begins.
22276
22277       -mprioritize-restricted-insns=priority
22278           This option controls the priority that is assigned to dispatch-slot
22279           restricted instructions during the second scheduling pass.  The
22280           argument priority takes the value 0, 1, or 2 to assign no, highest,
22281           or second-highest (respectively) priority to dispatch-slot
22282           restricted instructions.
22283
22284       -msched-costly-dep=dependence_type
22285           This option controls which dependences are considered costly by the
22286           target during instruction scheduling.  The argument dependence_type
22287           takes one of the following values:
22288
22289           no  No dependence is costly.
22290
22291           all All dependences are costly.
22292
22293           true_store_to_load
22294               A true dependence from store to load is costly.
22295
22296           store_to_load
22297               Any dependence from store to load is costly.
22298
22299           number
22300               Any dependence for which the latency is greater than or equal
22301               to number is costly.
22302
22303       -minsert-sched-nops=scheme
22304           This option controls which NOP insertion scheme is used during the
22305           second scheduling pass.  The argument scheme takes one of the
22306           following values:
22307
22308           no  Don't insert NOPs.
22309
22310           pad Pad with NOPs any dispatch group that has vacant issue slots,
22311               according to the scheduler's grouping.
22312
22313           regroup_exact
22314               Insert NOPs to force costly dependent insns into separate
22315               groups.  Insert exactly as many NOPs as needed to force an insn
22316               to a new group, according to the estimated processor grouping.
22317
22318           number
22319               Insert NOPs to force costly dependent insns into separate
22320               groups.  Insert number NOPs to force an insn to a new group.
22321
22322       -mcall-sysv
22323           On System V.4 and embedded PowerPC systems compile code using
22324           calling conventions that adhere to the March 1995 draft of the
22325           System V Application Binary Interface, PowerPC processor
22326           supplement.  This is the default unless you configured GCC using
22327           powerpc-*-eabiaix.
22328
22329       -mcall-sysv-eabi
22330       -mcall-eabi
22331           Specify both -mcall-sysv and -meabi options.
22332
22333       -mcall-sysv-noeabi
22334           Specify both -mcall-sysv and -mno-eabi options.
22335
22336       -mcall-aixdesc
22337           On System V.4 and embedded PowerPC systems compile code for the AIX
22338           operating system.
22339
22340       -mcall-linux
22341           On System V.4 and embedded PowerPC systems compile code for the
22342           Linux-based GNU system.
22343
22344       -mcall-freebsd
22345           On System V.4 and embedded PowerPC systems compile code for the
22346           FreeBSD operating system.
22347
22348       -mcall-netbsd
22349           On System V.4 and embedded PowerPC systems compile code for the
22350           NetBSD operating system.
22351
22352       -mcall-openbsd
22353           On System V.4 and embedded PowerPC systems compile code for the
22354           OpenBSD operating system.
22355
22356       -mtraceback=traceback_type
22357           Select the type of traceback table. Valid values for traceback_type
22358           are full, part, and no.
22359
22360       -maix-struct-return
22361           Return all structures in memory (as specified by the AIX ABI).
22362
22363       -msvr4-struct-return
22364           Return structures smaller than 8 bytes in registers (as specified
22365           by the SVR4 ABI).
22366
22367       -mabi=abi-type
22368           Extend the current ABI with a particular extension, or remove such
22369           extension.  Valid values are altivec, no-altivec, ibmlongdouble,
22370           ieeelongdouble, elfv1, elfv2.
22371
22372       -mabi=ibmlongdouble
22373           Change the current ABI to use IBM extended-precision long double.
22374           This is not likely to work if your system defaults to using IEEE
22375           extended-precision long double.  If you change the long double type
22376           from IEEE extended-precision, the compiler will issue a warning
22377           unless you use the -Wno-psabi option.  Requires -mlong-double-128
22378           to be enabled.
22379
22380       -mabi=ieeelongdouble
22381           Change the current ABI to use IEEE extended-precision long double.
22382           This is not likely to work if your system defaults to using IBM
22383           extended-precision long double.  If you change the long double type
22384           from IBM extended-precision, the compiler will issue a warning
22385           unless you use the -Wno-psabi option.  Requires -mlong-double-128
22386           to be enabled.
22387
22388       -mabi=elfv1
22389           Change the current ABI to use the ELFv1 ABI.  This is the default
22390           ABI for big-endian PowerPC 64-bit Linux.  Overriding the default
22391           ABI requires special system support and is likely to fail in
22392           spectacular ways.
22393
22394       -mabi=elfv2
22395           Change the current ABI to use the ELFv2 ABI.  This is the default
22396           ABI for little-endian PowerPC 64-bit Linux.  Overriding the default
22397           ABI requires special system support and is likely to fail in
22398           spectacular ways.
22399
22400       -mgnu-attribute
22401       -mno-gnu-attribute
22402           Emit .gnu_attribute assembly directives to set tag/value pairs in a
22403           .gnu.attributes section that specify ABI variations in function
22404           parameters or return values.
22405
22406       -mprototype
22407       -mno-prototype
22408           On System V.4 and embedded PowerPC systems assume that all calls to
22409           variable argument functions are properly prototyped.  Otherwise,
22410           the compiler must insert an instruction before every non-prototyped
22411           call to set or clear bit 6 of the condition code register ("CR") to
22412           indicate whether floating-point values are passed in the floating-
22413           point registers in case the function takes variable arguments.
22414           With -mprototype, only calls to prototyped variable argument
22415           functions set or clear the bit.
22416
22417       -msim
22418           On embedded PowerPC systems, assume that the startup module is
22419           called sim-crt0.o and that the standard C libraries are libsim.a
22420           and libc.a.  This is the default for powerpc-*-eabisim
22421           configurations.
22422
22423       -mmvme
22424           On embedded PowerPC systems, assume that the startup module is
22425           called crt0.o and the standard C libraries are libmvme.a and
22426           libc.a.
22427
22428       -mads
22429           On embedded PowerPC systems, assume that the startup module is
22430           called crt0.o and the standard C libraries are libads.a and libc.a.
22431
22432       -myellowknife
22433           On embedded PowerPC systems, assume that the startup module is
22434           called crt0.o and the standard C libraries are libyk.a and libc.a.
22435
22436       -mvxworks
22437           On System V.4 and embedded PowerPC systems, specify that you are
22438           compiling for a VxWorks system.
22439
22440       -memb
22441           On embedded PowerPC systems, set the "PPC_EMB" bit in the ELF flags
22442           header to indicate that eabi extended relocations are used.
22443
22444       -meabi
22445       -mno-eabi
22446           On System V.4 and embedded PowerPC systems do (do not) adhere to
22447           the Embedded Applications Binary Interface (EABI), which is a set
22448           of modifications to the System V.4 specifications.  Selecting
22449           -meabi means that the stack is aligned to an 8-byte boundary, a
22450           function "__eabi" is called from "main" to set up the EABI
22451           environment, and the -msdata option can use both "r2" and "r13" to
22452           point to two separate small data areas.  Selecting -mno-eabi means
22453           that the stack is aligned to a 16-byte boundary, no EABI
22454           initialization function is called from "main", and the -msdata
22455           option only uses "r13" to point to a single small data area.  The
22456           -meabi option is on by default if you configured GCC using one of
22457           the powerpc*-*-eabi* options.
22458
22459       -msdata=eabi
22460           On System V.4 and embedded PowerPC systems, put small initialized
22461           "const" global and static data in the ".sdata2" section, which is
22462           pointed to by register "r2".  Put small initialized non-"const"
22463           global and static data in the ".sdata" section, which is pointed to
22464           by register "r13".  Put small uninitialized global and static data
22465           in the ".sbss" section, which is adjacent to the ".sdata" section.
22466           The -msdata=eabi option is incompatible with the -mrelocatable
22467           option.  The -msdata=eabi option also sets the -memb option.
22468
22469       -msdata=sysv
22470           On System V.4 and embedded PowerPC systems, put small global and
22471           static data in the ".sdata" section, which is pointed to by
22472           register "r13".  Put small uninitialized global and static data in
22473           the ".sbss" section, which is adjacent to the ".sdata" section.
22474           The -msdata=sysv option is incompatible with the -mrelocatable
22475           option.
22476
22477       -msdata=default
22478       -msdata
22479           On System V.4 and embedded PowerPC systems, if -meabi is used,
22480           compile code the same as -msdata=eabi, otherwise compile code the
22481           same as -msdata=sysv.
22482
22483       -msdata=data
22484           On System V.4 and embedded PowerPC systems, put small global data
22485           in the ".sdata" section.  Put small uninitialized global data in
22486           the ".sbss" section.  Do not use register "r13" to address small
22487           data however.  This is the default behavior unless other -msdata
22488           options are used.
22489
22490       -msdata=none
22491       -mno-sdata
22492           On embedded PowerPC systems, put all initialized global and static
22493           data in the ".data" section, and all uninitialized data in the
22494           ".bss" section.
22495
22496       -mreadonly-in-sdata
22497           Put read-only objects in the ".sdata" section as well.  This is the
22498           default.
22499
22500       -mblock-move-inline-limit=num
22501           Inline all block moves (such as calls to "memcpy" or structure
22502           copies) less than or equal to num bytes.  The minimum value for num
22503           is 32 bytes on 32-bit targets and 64 bytes on 64-bit targets.  The
22504           default value is target-specific.
22505
22506       -mblock-compare-inline-limit=num
22507           Generate non-looping inline code for all block compares (such as
22508           calls to "memcmp" or structure compares) less than or equal to num
22509           bytes. If num is 0, all inline expansion (non-loop and loop) of
22510           block compare is disabled. The default value is target-specific.
22511
22512       -mblock-compare-inline-loop-limit=num
22513           Generate an inline expansion using loop code for all block compares
22514           that are less than or equal to num bytes, but greater than the
22515           limit for non-loop inline block compare expansion. If the block
22516           length is not constant, at most num bytes will be compared before
22517           "memcmp" is called to compare the remainder of the block. The
22518           default value is target-specific.
22519
22520       -mstring-compare-inline-limit=num
22521           Compare at most num string bytes with inline code.  If the
22522           difference or end of string is not found at the end of the inline
22523           compare a call to "strcmp" or "strncmp" will take care of the rest
22524           of the comparison. The default is 64 bytes.
22525
22526       -G num
22527           On embedded PowerPC systems, put global and static items less than
22528           or equal to num bytes into the small data or BSS sections instead
22529           of the normal data or BSS section.  By default, num is 8.  The -G
22530           num switch is also passed to the linker.  All modules should be
22531           compiled with the same -G num value.
22532
22533       -mregnames
22534       -mno-regnames
22535           On System V.4 and embedded PowerPC systems do (do not) emit
22536           register names in the assembly language output using symbolic
22537           forms.
22538
22539       -mlongcall
22540       -mno-longcall
22541           By default assume that all calls are far away so that a longer and
22542           more expensive calling sequence is required.  This is required for
22543           calls farther than 32 megabytes (33,554,432 bytes) from the current
22544           location.  A short call is generated if the compiler knows the call
22545           cannot be that far away.  This setting can be overridden by the
22546           "shortcall" function attribute, or by "#pragma longcall(0)".
22547
22548           Some linkers are capable of detecting out-of-range calls and
22549           generating glue code on the fly.  On these systems, long calls are
22550           unnecessary and generate slower code.  As of this writing, the AIX
22551           linker can do this, as can the GNU linker for PowerPC/64.  It is
22552           planned to add this feature to the GNU linker for 32-bit PowerPC
22553           systems as well.
22554
22555           On PowerPC64 ELFv2 and 32-bit PowerPC systems with newer GNU
22556           linkers, GCC can generate long calls using an inline PLT call
22557           sequence (see -mpltseq).  PowerPC with -mbss-plt and PowerPC64
22558           ELFv1 (big-endian) do not support inline PLT calls.
22559
22560           On Darwin/PPC systems, "#pragma longcall" generates "jbsr callee,
22561           L42", plus a branch island (glue code).  The two target addresses
22562           represent the callee and the branch island.  The Darwin/PPC linker
22563           prefers the first address and generates a "bl callee" if the PPC
22564           "bl" instruction reaches the callee directly; otherwise, the linker
22565           generates "bl L42" to call the branch island.  The branch island is
22566           appended to the body of the calling function; it computes the full
22567           32-bit address of the callee and jumps to it.
22568
22569           On Mach-O (Darwin) systems, this option directs the compiler emit
22570           to the glue for every direct call, and the Darwin linker decides
22571           whether to use or discard it.
22572
22573           In the future, GCC may ignore all longcall specifications when the
22574           linker is known to generate glue.
22575
22576       -mpltseq
22577       -mno-pltseq
22578           Implement (do not implement) -fno-plt and long calls using an
22579           inline PLT call sequence that supports lazy linking and long calls
22580           to functions in dlopen'd shared libraries.  Inline PLT calls are
22581           only supported on PowerPC64 ELFv2 and 32-bit PowerPC systems with
22582           newer GNU linkers, and are enabled by default if the support is
22583           detected when configuring GCC, and, in the case of 32-bit PowerPC,
22584           if GCC is configured with --enable-secureplt.  -mpltseq code and
22585           -mbss-plt 32-bit PowerPC relocatable objects may not be linked
22586           together.
22587
22588       -mtls-markers
22589       -mno-tls-markers
22590           Mark (do not mark) calls to "__tls_get_addr" with a relocation
22591           specifying the function argument.  The relocation allows the linker
22592           to reliably associate function call with argument setup
22593           instructions for TLS optimization, which in turn allows GCC to
22594           better schedule the sequence.
22595
22596       -mrecip
22597       -mno-recip
22598           This option enables use of the reciprocal estimate and reciprocal
22599           square root estimate instructions with additional Newton-Raphson
22600           steps to increase precision instead of doing a divide or square
22601           root and divide for floating-point arguments.  You should use the
22602           -ffast-math option when using -mrecip (or at least
22603           -funsafe-math-optimizations, -ffinite-math-only, -freciprocal-math
22604           and -fno-trapping-math).  Note that while the throughput of the
22605           sequence is generally higher than the throughput of the non-
22606           reciprocal instruction, the precision of the sequence can be
22607           decreased by up to 2 ulp (i.e. the inverse of 1.0 equals
22608           0.99999994) for reciprocal square roots.
22609
22610       -mrecip=opt
22611           This option controls which reciprocal estimate instructions may be
22612           used.  opt is a comma-separated list of options, which may be
22613           preceded by a "!" to invert the option:
22614
22615           all Enable all estimate instructions.
22616
22617           default
22618               Enable the default instructions, equivalent to -mrecip.
22619
22620           none
22621               Disable all estimate instructions, equivalent to -mno-recip.
22622
22623           div Enable the reciprocal approximation instructions for both
22624               single and double precision.
22625
22626           divf
22627               Enable the single-precision reciprocal approximation
22628               instructions.
22629
22630           divd
22631               Enable the double-precision reciprocal approximation
22632               instructions.
22633
22634           rsqrt
22635               Enable the reciprocal square root approximation instructions
22636               for both single and double precision.
22637
22638           rsqrtf
22639               Enable the single-precision reciprocal square root
22640               approximation instructions.
22641
22642           rsqrtd
22643               Enable the double-precision reciprocal square root
22644               approximation instructions.
22645
22646           So, for example, -mrecip=all,!rsqrtd enables all of the reciprocal
22647           estimate instructions, except for the "FRSQRTE", "XSRSQRTEDP", and
22648           "XVRSQRTEDP" instructions which handle the double-precision
22649           reciprocal square root calculations.
22650
22651       -mrecip-precision
22652       -mno-recip-precision
22653           Assume (do not assume) that the reciprocal estimate instructions
22654           provide higher-precision estimates than is mandated by the PowerPC
22655           ABI.  Selecting -mcpu=power6, -mcpu=power7 or -mcpu=power8
22656           automatically selects -mrecip-precision.  The double-precision
22657           square root estimate instructions are not generated by default on
22658           low-precision machines, since they do not provide an estimate that
22659           converges after three steps.
22660
22661       -mveclibabi=type
22662           Specifies the ABI type to use for vectorizing intrinsics using an
22663           external library.  The only type supported at present is mass,
22664           which specifies to use IBM's Mathematical Acceleration Subsystem
22665           (MASS) libraries for vectorizing intrinsics using external
22666           libraries.  GCC currently emits calls to "acosd2", "acosf4",
22667           "acoshd2", "acoshf4", "asind2", "asinf4", "asinhd2", "asinhf4",
22668           "atan2d2", "atan2f4", "atand2", "atanf4", "atanhd2", "atanhf4",
22669           "cbrtd2", "cbrtf4", "cosd2", "cosf4", "coshd2", "coshf4", "erfcd2",
22670           "erfcf4", "erfd2", "erff4", "exp2d2", "exp2f4", "expd2", "expf4",
22671           "expm1d2", "expm1f4", "hypotd2", "hypotf4", "lgammad2", "lgammaf4",
22672           "log10d2", "log10f4", "log1pd2", "log1pf4", "log2d2", "log2f4",
22673           "logd2", "logf4", "powd2", "powf4", "sind2", "sinf4", "sinhd2",
22674           "sinhf4", "sqrtd2", "sqrtf4", "tand2", "tanf4", "tanhd2", and
22675           "tanhf4" when generating code for power7.  Both -ftree-vectorize
22676           and -funsafe-math-optimizations must also be enabled.  The MASS
22677           libraries must be specified at link time.
22678
22679       -mfriz
22680       -mno-friz
22681           Generate (do not generate) the "friz" instruction when the
22682           -funsafe-math-optimizations option is used to optimize rounding of
22683           floating-point values to 64-bit integer and back to floating point.
22684           The "friz" instruction does not return the same value if the
22685           floating-point number is too large to fit in an integer.
22686
22687       -mpointers-to-nested-functions
22688       -mno-pointers-to-nested-functions
22689           Generate (do not generate) code to load up the static chain
22690           register ("r11") when calling through a pointer on AIX and 64-bit
22691           Linux systems where a function pointer points to a 3-word
22692           descriptor giving the function address, TOC value to be loaded in
22693           register "r2", and static chain value to be loaded in register
22694           "r11".  The -mpointers-to-nested-functions is on by default.  You
22695           cannot call through pointers to nested functions or pointers to
22696           functions compiled in other languages that use the static chain if
22697           you use -mno-pointers-to-nested-functions.
22698
22699       -msave-toc-indirect
22700       -mno-save-toc-indirect
22701           Generate (do not generate) code to save the TOC value in the
22702           reserved stack location in the function prologue if the function
22703           calls through a pointer on AIX and 64-bit Linux systems.  If the
22704           TOC value is not saved in the prologue, it is saved just before the
22705           call through the pointer.  The -mno-save-toc-indirect option is the
22706           default.
22707
22708       -mcompat-align-parm
22709       -mno-compat-align-parm
22710           Generate (do not generate) code to pass structure parameters with a
22711           maximum alignment of 64 bits, for compatibility with older versions
22712           of GCC.
22713
22714           Older versions of GCC (prior to 4.9.0) incorrectly did not align a
22715           structure parameter on a 128-bit boundary when that structure
22716           contained a member requiring 128-bit alignment.  This is corrected
22717           in more recent versions of GCC.  This option may be used to
22718           generate code that is compatible with functions compiled with older
22719           versions of GCC.
22720
22721           The -mno-compat-align-parm option is the default.
22722
22723       -mstack-protector-guard=guard
22724       -mstack-protector-guard-reg=reg
22725       -mstack-protector-guard-offset=offset
22726       -mstack-protector-guard-symbol=symbol
22727           Generate stack protection code using canary at guard.  Supported
22728           locations are global for global canary or tls for per-thread canary
22729           in the TLS block (the default with GNU libc version 2.4 or later).
22730
22731           With the latter choice the options -mstack-protector-guard-reg=reg
22732           and -mstack-protector-guard-offset=offset furthermore specify which
22733           register to use as base register for reading the canary, and from
22734           what offset from that base register. The default for those is as
22735           specified in the relevant ABI.
22736           -mstack-protector-guard-symbol=symbol overrides the offset with a
22737           symbol reference to a canary in the TLS block.
22738
22739       -mpcrel
22740       -mno-pcrel
22741           Generate (do not generate) pc-relative addressing when the option
22742           -mcpu=future is used.  The -mpcrel option requires that the medium
22743           code model (-mcmodel=medium) and prefixed addressing (-mprefixed)
22744           options are enabled.
22745
22746       -mprefixed
22747       -mno-prefixed
22748           Generate (do not generate) addressing modes using prefixed load and
22749           store instructions when the option -mcpu=future is used.
22750
22751       -mmma
22752       -mno-mma
22753           Generate (do not generate) the MMA instructions when the option
22754           -mcpu=future is used.
22755
22756       RX Options
22757
22758       These command-line options are defined for RX targets:
22759
22760       -m64bit-doubles
22761       -m32bit-doubles
22762           Make the "double" data type be 64 bits (-m64bit-doubles) or 32 bits
22763           (-m32bit-doubles) in size.  The default is -m32bit-doubles.  Note
22764           RX floating-point hardware only works on 32-bit values, which is
22765           why the default is -m32bit-doubles.
22766
22767       -fpu
22768       -nofpu
22769           Enables (-fpu) or disables (-nofpu) the use of RX floating-point
22770           hardware.  The default is enabled for the RX600 series and disabled
22771           for the RX200 series.
22772
22773           Floating-point instructions are only generated for 32-bit floating-
22774           point values, however, so the FPU hardware is not used for doubles
22775           if the -m64bit-doubles option is used.
22776
22777           Note If the -fpu option is enabled then -funsafe-math-optimizations
22778           is also enabled automatically.  This is because the RX FPU
22779           instructions are themselves unsafe.
22780
22781       -mcpu=name
22782           Selects the type of RX CPU to be targeted.  Currently three types
22783           are supported, the generic RX600 and RX200 series hardware and the
22784           specific RX610 CPU.  The default is RX600.
22785
22786           The only difference between RX600 and RX610 is that the RX610 does
22787           not support the "MVTIPL" instruction.
22788
22789           The RX200 series does not have a hardware floating-point unit and
22790           so -nofpu is enabled by default when this type is selected.
22791
22792       -mbig-endian-data
22793       -mlittle-endian-data
22794           Store data (but not code) in the big-endian format.  The default is
22795           -mlittle-endian-data, i.e. to store data in the little-endian
22796           format.
22797
22798       -msmall-data-limit=N
22799           Specifies the maximum size in bytes of global and static variables
22800           which can be placed into the small data area.  Using the small data
22801           area can lead to smaller and faster code, but the size of area is
22802           limited and it is up to the programmer to ensure that the area does
22803           not overflow.  Also when the small data area is used one of the
22804           RX's registers (usually "r13") is reserved for use pointing to this
22805           area, so it is no longer available for use by the compiler.  This
22806           could result in slower and/or larger code if variables are pushed
22807           onto the stack instead of being held in this register.
22808
22809           Note, common variables (variables that have not been initialized)
22810           and constants are not placed into the small data area as they are
22811           assigned to other sections in the output executable.
22812
22813           The default value is zero, which disables this feature.  Note, this
22814           feature is not enabled by default with higher optimization levels
22815           (-O2 etc) because of the potentially detrimental effects of
22816           reserving a register.  It is up to the programmer to experiment and
22817           discover whether this feature is of benefit to their program.  See
22818           the description of the -mpid option for a description of how the
22819           actual register to hold the small data area pointer is chosen.
22820
22821       -msim
22822       -mno-sim
22823           Use the simulator runtime.  The default is to use the libgloss
22824           board-specific runtime.
22825
22826       -mas100-syntax
22827       -mno-as100-syntax
22828           When generating assembler output use a syntax that is compatible
22829           with Renesas's AS100 assembler.  This syntax can also be handled by
22830           the GAS assembler, but it has some restrictions so it is not
22831           generated by default.
22832
22833       -mmax-constant-size=N
22834           Specifies the maximum size, in bytes, of a constant that can be
22835           used as an operand in a RX instruction.  Although the RX
22836           instruction set does allow constants of up to 4 bytes in length to
22837           be used in instructions, a longer value equates to a longer
22838           instruction.  Thus in some circumstances it can be beneficial to
22839           restrict the size of constants that are used in instructions.
22840           Constants that are too big are instead placed into a constant pool
22841           and referenced via register indirection.
22842
22843           The value N can be between 0 and 4.  A value of 0 (the default) or
22844           4 means that constants of any size are allowed.
22845
22846       -mrelax
22847           Enable linker relaxation.  Linker relaxation is a process whereby
22848           the linker attempts to reduce the size of a program by finding
22849           shorter versions of various instructions.  Disabled by default.
22850
22851       -mint-register=N
22852           Specify the number of registers to reserve for fast interrupt
22853           handler functions.  The value N can be between 0 and 4.  A value of
22854           1 means that register "r13" is reserved for the exclusive use of
22855           fast interrupt handlers.  A value of 2 reserves "r13" and "r12".  A
22856           value of 3 reserves "r13", "r12" and "r11", and a value of 4
22857           reserves "r13" through "r10".  A value of 0, the default, does not
22858           reserve any registers.
22859
22860       -msave-acc-in-interrupts
22861           Specifies that interrupt handler functions should preserve the
22862           accumulator register.  This is only necessary if normal code might
22863           use the accumulator register, for example because it performs
22864           64-bit multiplications.  The default is to ignore the accumulator
22865           as this makes the interrupt handlers faster.
22866
22867       -mpid
22868       -mno-pid
22869           Enables the generation of position independent data.  When enabled
22870           any access to constant data is done via an offset from a base
22871           address held in a register.  This allows the location of constant
22872           data to be determined at run time without requiring the executable
22873           to be relocated, which is a benefit to embedded applications with
22874           tight memory constraints.  Data that can be modified is not
22875           affected by this option.
22876
22877           Note, using this feature reserves a register, usually "r13", for
22878           the constant data base address.  This can result in slower and/or
22879           larger code, especially in complicated functions.
22880
22881           The actual register chosen to hold the constant data base address
22882           depends upon whether the -msmall-data-limit and/or the
22883           -mint-register command-line options are enabled.  Starting with
22884           register "r13" and proceeding downwards, registers are allocated
22885           first to satisfy the requirements of -mint-register, then -mpid and
22886           finally -msmall-data-limit.  Thus it is possible for the small data
22887           area register to be "r8" if both -mint-register=4 and -mpid are
22888           specified on the command line.
22889
22890           By default this feature is not enabled.  The default can be
22891           restored via the -mno-pid command-line option.
22892
22893       -mno-warn-multiple-fast-interrupts
22894       -mwarn-multiple-fast-interrupts
22895           Prevents GCC from issuing a warning message if it finds more than
22896           one fast interrupt handler when it is compiling a file.  The
22897           default is to issue a warning for each extra fast interrupt handler
22898           found, as the RX only supports one such interrupt.
22899
22900       -mallow-string-insns
22901       -mno-allow-string-insns
22902           Enables or disables the use of the string manipulation instructions
22903           "SMOVF", "SCMPU", "SMOVB", "SMOVU", "SUNTIL" "SWHILE" and also the
22904           "RMPA" instruction.  These instructions may prefetch data, which is
22905           not safe to do if accessing an I/O register.  (See section 12.2.7
22906           of the RX62N Group User's Manual for more information).
22907
22908           The default is to allow these instructions, but it is not possible
22909           for GCC to reliably detect all circumstances where a string
22910           instruction might be used to access an I/O register, so their use
22911           cannot be disabled automatically.  Instead it is reliant upon the
22912           programmer to use the -mno-allow-string-insns option if their
22913           program accesses I/O space.
22914
22915           When the instructions are enabled GCC defines the C preprocessor
22916           symbol "__RX_ALLOW_STRING_INSNS__", otherwise it defines the symbol
22917           "__RX_DISALLOW_STRING_INSNS__".
22918
22919       -mjsr
22920       -mno-jsr
22921           Use only (or not only) "JSR" instructions to access functions.
22922           This option can be used when code size exceeds the range of "BSR"
22923           instructions.  Note that -mno-jsr does not mean to not use "JSR"
22924           but instead means that any type of branch may be used.
22925
22926       Note: The generic GCC command-line option -ffixed-reg has special
22927       significance to the RX port when used with the "interrupt" function
22928       attribute.  This attribute indicates a function intended to process
22929       fast interrupts.  GCC ensures that it only uses the registers "r10",
22930       "r11", "r12" and/or "r13" and only provided that the normal use of the
22931       corresponding registers have been restricted via the -ffixed-reg or
22932       -mint-register command-line options.
22933
22934       S/390 and zSeries Options
22935
22936       These are the -m options defined for the S/390 and zSeries
22937       architecture.
22938
22939       -mhard-float
22940       -msoft-float
22941           Use (do not use) the hardware floating-point instructions and
22942           registers for floating-point operations.  When -msoft-float is
22943           specified, functions in libgcc.a are used to perform floating-point
22944           operations.  When -mhard-float is specified, the compiler generates
22945           IEEE floating-point instructions.  This is the default.
22946
22947       -mhard-dfp
22948       -mno-hard-dfp
22949           Use (do not use) the hardware decimal-floating-point instructions
22950           for decimal-floating-point operations.  When -mno-hard-dfp is
22951           specified, functions in libgcc.a are used to perform decimal-
22952           floating-point operations.  When -mhard-dfp is specified, the
22953           compiler generates decimal-floating-point hardware instructions.
22954           This is the default for -march=z9-ec or higher.
22955
22956       -mlong-double-64
22957       -mlong-double-128
22958           These switches control the size of "long double" type. A size of 64
22959           bits makes the "long double" type equivalent to the "double" type.
22960           This is the default.
22961
22962       -mbackchain
22963       -mno-backchain
22964           Store (do not store) the address of the caller's frame as backchain
22965           pointer into the callee's stack frame.  A backchain may be needed
22966           to allow debugging using tools that do not understand DWARF call
22967           frame information.  When -mno-packed-stack is in effect, the
22968           backchain pointer is stored at the bottom of the stack frame; when
22969           -mpacked-stack is in effect, the backchain is placed into the
22970           topmost word of the 96/160 byte register save area.
22971
22972           In general, code compiled with -mbackchain is call-compatible with
22973           code compiled with -mmo-backchain; however, use of the backchain
22974           for debugging purposes usually requires that the whole binary is
22975           built with -mbackchain.  Note that the combination of -mbackchain,
22976           -mpacked-stack and -mhard-float is not supported.  In order to
22977           build a linux kernel use -msoft-float.
22978
22979           The default is to not maintain the backchain.
22980
22981       -mpacked-stack
22982       -mno-packed-stack
22983           Use (do not use) the packed stack layout.  When -mno-packed-stack
22984           is specified, the compiler uses the all fields of the 96/160 byte
22985           register save area only for their default purpose; unused fields
22986           still take up stack space.  When -mpacked-stack is specified,
22987           register save slots are densely packed at the top of the register
22988           save area; unused space is reused for other purposes, allowing for
22989           more efficient use of the available stack space.  However, when
22990           -mbackchain is also in effect, the topmost word of the save area is
22991           always used to store the backchain, and the return address register
22992           is always saved two words below the backchain.
22993
22994           As long as the stack frame backchain is not used, code generated
22995           with -mpacked-stack is call-compatible with code generated with
22996           -mno-packed-stack.  Note that some non-FSF releases of GCC 2.95 for
22997           S/390 or zSeries generated code that uses the stack frame backchain
22998           at run time, not just for debugging purposes.  Such code is not
22999           call-compatible with code compiled with -mpacked-stack.  Also, note
23000           that the combination of -mbackchain, -mpacked-stack and
23001           -mhard-float is not supported.  In order to build a linux kernel
23002           use -msoft-float.
23003
23004           The default is to not use the packed stack layout.
23005
23006       -msmall-exec
23007       -mno-small-exec
23008           Generate (or do not generate) code using the "bras" instruction to
23009           do subroutine calls.  This only works reliably if the total
23010           executable size does not exceed 64k.  The default is to use the
23011           "basr" instruction instead, which does not have this limitation.
23012
23013       -m64
23014       -m31
23015           When -m31 is specified, generate code compliant to the GNU/Linux
23016           for S/390 ABI.  When -m64 is specified, generate code compliant to
23017           the GNU/Linux for zSeries ABI.  This allows GCC in particular to
23018           generate 64-bit instructions.  For the s390 targets, the default is
23019           -m31, while the s390x targets default to -m64.
23020
23021       -mzarch
23022       -mesa
23023           When -mzarch is specified, generate code using the instructions
23024           available on z/Architecture.  When -mesa is specified, generate
23025           code using the instructions available on ESA/390.  Note that -mesa
23026           is not possible with -m64.  When generating code compliant to the
23027           GNU/Linux for S/390 ABI, the default is -mesa.  When generating
23028           code compliant to the GNU/Linux for zSeries ABI, the default is
23029           -mzarch.
23030
23031       -mhtm
23032       -mno-htm
23033           The -mhtm option enables a set of builtins making use of
23034           instructions available with the transactional execution facility
23035           introduced with the IBM zEnterprise EC12 machine generation S/390
23036           System z Built-in Functions.  -mhtm is enabled by default when
23037           using -march=zEC12.
23038
23039       -mvx
23040       -mno-vx
23041           When -mvx is specified, generate code using the instructions
23042           available with the vector extension facility introduced with the
23043           IBM z13 machine generation.  This option changes the ABI for some
23044           vector type values with regard to alignment and calling
23045           conventions.  In case vector type values are being used in an ABI-
23046           relevant context a GAS .gnu_attribute command will be added to mark
23047           the resulting binary with the ABI used.  -mvx is enabled by default
23048           when using -march=z13.
23049
23050       -mzvector
23051       -mno-zvector
23052           The -mzvector option enables vector language extensions and
23053           builtins using instructions available with the vector extension
23054           facility introduced with the IBM z13 machine generation.  This
23055           option adds support for vector to be used as a keyword to define
23056           vector type variables and arguments.  vector is only available when
23057           GNU extensions are enabled.  It will not be expanded when
23058           requesting strict standard compliance e.g. with -std=c99.  In
23059           addition to the GCC low-level builtins -mzvector enables a set of
23060           builtins added for compatibility with AltiVec-style implementations
23061           like Power and Cell.  In order to make use of these builtins the
23062           header file vecintrin.h needs to be included.  -mzvector is
23063           disabled by default.
23064
23065       -mmvcle
23066       -mno-mvcle
23067           Generate (or do not generate) code using the "mvcle" instruction to
23068           perform block moves.  When -mno-mvcle is specified, use a "mvc"
23069           loop instead.  This is the default unless optimizing for size.
23070
23071       -mdebug
23072       -mno-debug
23073           Print (or do not print) additional debug information when
23074           compiling.  The default is to not print debug information.
23075
23076       -march=cpu-type
23077           Generate code that runs on cpu-type, which is the name of a system
23078           representing a certain processor type.  Possible values for cpu-
23079           type are z900/arch5, z990/arch6, z9-109, z9-ec/arch7, z10/arch8,
23080           z196/arch9, zEC12, z13/arch11, z14/arch12, and native.
23081
23082           The default is -march=z900.
23083
23084           Specifying native as cpu type can be used to select the best
23085           architecture option for the host processor.  -march=native has no
23086           effect if GCC does not recognize the processor.
23087
23088       -mtune=cpu-type
23089           Tune to cpu-type everything applicable about the generated code,
23090           except for the ABI and the set of available instructions.  The list
23091           of cpu-type values is the same as for -march.  The default is the
23092           value used for -march.
23093
23094       -mtpf-trace
23095       -mno-tpf-trace
23096           Generate code that adds (does not add) in TPF OS specific branches
23097           to trace routines in the operating system.  This option is off by
23098           default, even when compiling for the TPF OS.
23099
23100       -mtpf-trace-skip
23101       -mno-tpf-trace-skip
23102           Generate code that changes (does not change) the default branch
23103           targets enabled by -mtpf-trace to point to specialized trace
23104           routines providing the ability of selectively skipping function
23105           trace entries for the TPF OS.  This option is off by default, even
23106           when compiling for the TPF OS and specifying -mtpf-trace.
23107
23108       -mfused-madd
23109       -mno-fused-madd
23110           Generate code that uses (does not use) the floating-point multiply
23111           and accumulate instructions.  These instructions are generated by
23112           default if hardware floating point is used.
23113
23114       -mwarn-framesize=framesize
23115           Emit a warning if the current function exceeds the given frame
23116           size.  Because this is a compile-time check it doesn't need to be a
23117           real problem when the program runs.  It is intended to identify
23118           functions that most probably cause a stack overflow.  It is useful
23119           to be used in an environment with limited stack size e.g. the linux
23120           kernel.
23121
23122       -mwarn-dynamicstack
23123           Emit a warning if the function calls "alloca" or uses dynamically-
23124           sized arrays.  This is generally a bad idea with a limited stack
23125           size.
23126
23127       -mstack-guard=stack-guard
23128       -mstack-size=stack-size
23129           If these options are provided the S/390 back end emits additional
23130           instructions in the function prologue that trigger a trap if the
23131           stack size is stack-guard bytes above the stack-size (remember that
23132           the stack on S/390 grows downward).  If the stack-guard option is
23133           omitted the smallest power of 2 larger than the frame size of the
23134           compiled function is chosen.  These options are intended to be used
23135           to help debugging stack overflow problems.  The additionally
23136           emitted code causes only little overhead and hence can also be used
23137           in production-like systems without greater performance degradation.
23138           The given values have to be exact powers of 2 and stack-size has to
23139           be greater than stack-guard without exceeding 64k.  In order to be
23140           efficient the extra code makes the assumption that the stack starts
23141           at an address aligned to the value given by stack-size.  The stack-
23142           guard option can only be used in conjunction with stack-size.
23143
23144       -mhotpatch=pre-halfwords,post-halfwords
23145           If the hotpatch option is enabled, a "hot-patching" function
23146           prologue is generated for all functions in the compilation unit.
23147           The funtion label is prepended with the given number of two-byte
23148           NOP instructions (pre-halfwords, maximum 1000000).  After the
23149           label, 2 * post-halfwords bytes are appended, using the largest NOP
23150           like instructions the architecture allows (maximum 1000000).
23151
23152           If both arguments are zero, hotpatching is disabled.
23153
23154           This option can be overridden for individual functions with the
23155           "hotpatch" attribute.
23156
23157       Score Options
23158
23159       These options are defined for Score implementations:
23160
23161       -meb
23162           Compile code for big-endian mode.  This is the default.
23163
23164       -mel
23165           Compile code for little-endian mode.
23166
23167       -mnhwloop
23168           Disable generation of "bcnz" instructions.
23169
23170       -muls
23171           Enable generation of unaligned load and store instructions.
23172
23173       -mmac
23174           Enable the use of multiply-accumulate instructions. Disabled by
23175           default.
23176
23177       -mscore5
23178           Specify the SCORE5 as the target architecture.
23179
23180       -mscore5u
23181           Specify the SCORE5U of the target architecture.
23182
23183       -mscore7
23184           Specify the SCORE7 as the target architecture. This is the default.
23185
23186       -mscore7d
23187           Specify the SCORE7D as the target architecture.
23188
23189       SH Options
23190
23191       These -m options are defined for the SH implementations:
23192
23193       -m1 Generate code for the SH1.
23194
23195       -m2 Generate code for the SH2.
23196
23197       -m2e
23198           Generate code for the SH2e.
23199
23200       -m2a-nofpu
23201           Generate code for the SH2a without FPU, or for a SH2a-FPU in such a
23202           way that the floating-point unit is not used.
23203
23204       -m2a-single-only
23205           Generate code for the SH2a-FPU, in such a way that no double-
23206           precision floating-point operations are used.
23207
23208       -m2a-single
23209           Generate code for the SH2a-FPU assuming the floating-point unit is
23210           in single-precision mode by default.
23211
23212       -m2a
23213           Generate code for the SH2a-FPU assuming the floating-point unit is
23214           in double-precision mode by default.
23215
23216       -m3 Generate code for the SH3.
23217
23218       -m3e
23219           Generate code for the SH3e.
23220
23221       -m4-nofpu
23222           Generate code for the SH4 without a floating-point unit.
23223
23224       -m4-single-only
23225           Generate code for the SH4 with a floating-point unit that only
23226           supports single-precision arithmetic.
23227
23228       -m4-single
23229           Generate code for the SH4 assuming the floating-point unit is in
23230           single-precision mode by default.
23231
23232       -m4 Generate code for the SH4.
23233
23234       -m4-100
23235           Generate code for SH4-100.
23236
23237       -m4-100-nofpu
23238           Generate code for SH4-100 in such a way that the floating-point
23239           unit is not used.
23240
23241       -m4-100-single
23242           Generate code for SH4-100 assuming the floating-point unit is in
23243           single-precision mode by default.
23244
23245       -m4-100-single-only
23246           Generate code for SH4-100 in such a way that no double-precision
23247           floating-point operations are used.
23248
23249       -m4-200
23250           Generate code for SH4-200.
23251
23252       -m4-200-nofpu
23253           Generate code for SH4-200 without in such a way that the floating-
23254           point unit is not used.
23255
23256       -m4-200-single
23257           Generate code for SH4-200 assuming the floating-point unit is in
23258           single-precision mode by default.
23259
23260       -m4-200-single-only
23261           Generate code for SH4-200 in such a way that no double-precision
23262           floating-point operations are used.
23263
23264       -m4-300
23265           Generate code for SH4-300.
23266
23267       -m4-300-nofpu
23268           Generate code for SH4-300 without in such a way that the floating-
23269           point unit is not used.
23270
23271       -m4-300-single
23272           Generate code for SH4-300 in such a way that no double-precision
23273           floating-point operations are used.
23274
23275       -m4-300-single-only
23276           Generate code for SH4-300 in such a way that no double-precision
23277           floating-point operations are used.
23278
23279       -m4-340
23280           Generate code for SH4-340 (no MMU, no FPU).
23281
23282       -m4-500
23283           Generate code for SH4-500 (no FPU).  Passes -isa=sh4-nofpu to the
23284           assembler.
23285
23286       -m4a-nofpu
23287           Generate code for the SH4al-dsp, or for a SH4a in such a way that
23288           the floating-point unit is not used.
23289
23290       -m4a-single-only
23291           Generate code for the SH4a, in such a way that no double-precision
23292           floating-point operations are used.
23293
23294       -m4a-single
23295           Generate code for the SH4a assuming the floating-point unit is in
23296           single-precision mode by default.
23297
23298       -m4a
23299           Generate code for the SH4a.
23300
23301       -m4al
23302           Same as -m4a-nofpu, except that it implicitly passes -dsp to the
23303           assembler.  GCC doesn't generate any DSP instructions at the
23304           moment.
23305
23306       -mb Compile code for the processor in big-endian mode.
23307
23308       -ml Compile code for the processor in little-endian mode.
23309
23310       -mdalign
23311           Align doubles at 64-bit boundaries.  Note that this changes the
23312           calling conventions, and thus some functions from the standard C
23313           library do not work unless you recompile it first with -mdalign.
23314
23315       -mrelax
23316           Shorten some address references at link time, when possible; uses
23317           the linker option -relax.
23318
23319       -mbigtable
23320           Use 32-bit offsets in "switch" tables.  The default is to use
23321           16-bit offsets.
23322
23323       -mbitops
23324           Enable the use of bit manipulation instructions on SH2A.
23325
23326       -mfmovd
23327           Enable the use of the instruction "fmovd".  Check -mdalign for
23328           alignment constraints.
23329
23330       -mrenesas
23331           Comply with the calling conventions defined by Renesas.
23332
23333       -mno-renesas
23334           Comply with the calling conventions defined for GCC before the
23335           Renesas conventions were available.  This option is the default for
23336           all targets of the SH toolchain.
23337
23338       -mnomacsave
23339           Mark the "MAC" register as call-clobbered, even if -mrenesas is
23340           given.
23341
23342       -mieee
23343       -mno-ieee
23344           Control the IEEE compliance of floating-point comparisons, which
23345           affects the handling of cases where the result of a comparison is
23346           unordered.  By default -mieee is implicitly enabled.  If
23347           -ffinite-math-only is enabled -mno-ieee is implicitly set, which
23348           results in faster floating-point greater-equal and less-equal
23349           comparisons.  The implicit settings can be overridden by specifying
23350           either -mieee or -mno-ieee.
23351
23352       -minline-ic_invalidate
23353           Inline code to invalidate instruction cache entries after setting
23354           up nested function trampolines.  This option has no effect if
23355           -musermode is in effect and the selected code generation option
23356           (e.g. -m4) does not allow the use of the "icbi" instruction.  If
23357           the selected code generation option does not allow the use of the
23358           "icbi" instruction, and -musermode is not in effect, the inlined
23359           code manipulates the instruction cache address array directly with
23360           an associative write.  This not only requires privileged mode at
23361           run time, but it also fails if the cache line had been mapped via
23362           the TLB and has become unmapped.
23363
23364       -misize
23365           Dump instruction size and location in the assembly code.
23366
23367       -mpadstruct
23368           This option is deprecated.  It pads structures to multiple of 4
23369           bytes, which is incompatible with the SH ABI.
23370
23371       -matomic-model=model
23372           Sets the model of atomic operations and additional parameters as a
23373           comma separated list.  For details on the atomic built-in functions
23374           see __atomic Builtins.  The following models and parameters are
23375           supported:
23376
23377           none
23378               Disable compiler generated atomic sequences and emit library
23379               calls for atomic operations.  This is the default if the target
23380               is not "sh*-*-linux*".
23381
23382           soft-gusa
23383               Generate GNU/Linux compatible gUSA software atomic sequences
23384               for the atomic built-in functions.  The generated atomic
23385               sequences require additional support from the
23386               interrupt/exception handling code of the system and are only
23387               suitable for SH3* and SH4* single-core systems.  This option is
23388               enabled by default when the target is "sh*-*-linux*" and SH3*
23389               or SH4*.  When the target is SH4A, this option also partially
23390               utilizes the hardware atomic instructions "movli.l" and
23391               "movco.l" to create more efficient code, unless strict is
23392               specified.
23393
23394           soft-tcb
23395               Generate software atomic sequences that use a variable in the
23396               thread control block.  This is a variation of the gUSA
23397               sequences which can also be used on SH1* and SH2* targets.  The
23398               generated atomic sequences require additional support from the
23399               interrupt/exception handling code of the system and are only
23400               suitable for single-core systems.  When using this model, the
23401               gbr-offset= parameter has to be specified as well.
23402
23403           soft-imask
23404               Generate software atomic sequences that temporarily disable
23405               interrupts by setting "SR.IMASK = 1111".  This model works only
23406               when the program runs in privileged mode and is only suitable
23407               for single-core systems.  Additional support from the
23408               interrupt/exception handling code of the system is not
23409               required.  This model is enabled by default when the target is
23410               "sh*-*-linux*" and SH1* or SH2*.
23411
23412           hard-llcs
23413               Generate hardware atomic sequences using the "movli.l" and
23414               "movco.l" instructions only.  This is only available on SH4A
23415               and is suitable for multi-core systems.  Since the hardware
23416               instructions support only 32 bit atomic variables access to 8
23417               or 16 bit variables is emulated with 32 bit accesses.  Code
23418               compiled with this option is also compatible with other
23419               software atomic model interrupt/exception handling systems if
23420               executed on an SH4A system.  Additional support from the
23421               interrupt/exception handling code of the system is not required
23422               for this model.
23423
23424           gbr-offset=
23425               This parameter specifies the offset in bytes of the variable in
23426               the thread control block structure that should be used by the
23427               generated atomic sequences when the soft-tcb model has been
23428               selected.  For other models this parameter is ignored.  The
23429               specified value must be an integer multiple of four and in the
23430               range 0-1020.
23431
23432           strict
23433               This parameter prevents mixed usage of multiple atomic models,
23434               even if they are compatible, and makes the compiler generate
23435               atomic sequences of the specified model only.
23436
23437       -mtas
23438           Generate the "tas.b" opcode for "__atomic_test_and_set".  Notice
23439           that depending on the particular hardware and software
23440           configuration this can degrade overall performance due to the
23441           operand cache line flushes that are implied by the "tas.b"
23442           instruction.  On multi-core SH4A processors the "tas.b" instruction
23443           must be used with caution since it can result in data corruption
23444           for certain cache configurations.
23445
23446       -mprefergot
23447           When generating position-independent code, emit function calls
23448           using the Global Offset Table instead of the Procedure Linkage
23449           Table.
23450
23451       -musermode
23452       -mno-usermode
23453           Don't allow (allow) the compiler generating privileged mode code.
23454           Specifying -musermode also implies -mno-inline-ic_invalidate if the
23455           inlined code would not work in user mode.  -musermode is the
23456           default when the target is "sh*-*-linux*".  If the target is SH1*
23457           or SH2* -musermode has no effect, since there is no user mode.
23458
23459       -multcost=number
23460           Set the cost to assume for a multiply insn.
23461
23462       -mdiv=strategy
23463           Set the division strategy to be used for integer division
23464           operations.  strategy can be one of:
23465
23466           call-div1
23467               Calls a library function that uses the single-step division
23468               instruction "div1" to perform the operation.  Division by zero
23469               calculates an unspecified result and does not trap.  This is
23470               the default except for SH4, SH2A and SHcompact.
23471
23472           call-fp
23473               Calls a library function that performs the operation in double
23474               precision floating point.  Division by zero causes a floating-
23475               point exception.  This is the default for SHcompact with FPU.
23476               Specifying this for targets that do not have a double precision
23477               FPU defaults to "call-div1".
23478
23479           call-table
23480               Calls a library function that uses a lookup table for small
23481               divisors and the "div1" instruction with case distinction for
23482               larger divisors.  Division by zero calculates an unspecified
23483               result and does not trap.  This is the default for SH4.
23484               Specifying this for targets that do not have dynamic shift
23485               instructions defaults to "call-div1".
23486
23487           When a division strategy has not been specified the default
23488           strategy is selected based on the current target.  For SH2A the
23489           default strategy is to use the "divs" and "divu" instructions
23490           instead of library function calls.
23491
23492       -maccumulate-outgoing-args
23493           Reserve space once for outgoing arguments in the function prologue
23494           rather than around each call.  Generally beneficial for performance
23495           and size.  Also needed for unwinding to avoid changing the stack
23496           frame around conditional code.
23497
23498       -mdivsi3_libfunc=name
23499           Set the name of the library function used for 32-bit signed
23500           division to name.  This only affects the name used in the call
23501           division strategies, and the compiler still expects the same sets
23502           of input/output/clobbered registers as if this option were not
23503           present.
23504
23505       -mfixed-range=register-range
23506           Generate code treating the given register range as fixed registers.
23507           A fixed register is one that the register allocator cannot use.
23508           This is useful when compiling kernel code.  A register range is
23509           specified as two registers separated by a dash.  Multiple register
23510           ranges can be specified separated by a comma.
23511
23512       -mbranch-cost=num
23513           Assume num to be the cost for a branch instruction.  Higher numbers
23514           make the compiler try to generate more branch-free code if
23515           possible.  If not specified the value is selected depending on the
23516           processor type that is being compiled for.
23517
23518       -mzdcbranch
23519       -mno-zdcbranch
23520           Assume (do not assume) that zero displacement conditional branch
23521           instructions "bt" and "bf" are fast.  If -mzdcbranch is specified,
23522           the compiler prefers zero displacement branch code sequences.  This
23523           is enabled by default when generating code for SH4 and SH4A.  It
23524           can be explicitly disabled by specifying -mno-zdcbranch.
23525
23526       -mcbranch-force-delay-slot
23527           Force the usage of delay slots for conditional branches, which
23528           stuffs the delay slot with a "nop" if a suitable instruction cannot
23529           be found.  By default this option is disabled.  It can be enabled
23530           to work around hardware bugs as found in the original SH7055.
23531
23532       -mfused-madd
23533       -mno-fused-madd
23534           Generate code that uses (does not use) the floating-point multiply
23535           and accumulate instructions.  These instructions are generated by
23536           default if hardware floating point is used.  The machine-dependent
23537           -mfused-madd option is now mapped to the machine-independent
23538           -ffp-contract=fast option, and -mno-fused-madd is mapped to
23539           -ffp-contract=off.
23540
23541       -mfsca
23542       -mno-fsca
23543           Allow or disallow the compiler to emit the "fsca" instruction for
23544           sine and cosine approximations.  The option -mfsca must be used in
23545           combination with -funsafe-math-optimizations.  It is enabled by
23546           default when generating code for SH4A.  Using -mno-fsca disables
23547           sine and cosine approximations even if -funsafe-math-optimizations
23548           is in effect.
23549
23550       -mfsrra
23551       -mno-fsrra
23552           Allow or disallow the compiler to emit the "fsrra" instruction for
23553           reciprocal square root approximations.  The option -mfsrra must be
23554           used in combination with -funsafe-math-optimizations and
23555           -ffinite-math-only.  It is enabled by default when generating code
23556           for SH4A.  Using -mno-fsrra disables reciprocal square root
23557           approximations even if -funsafe-math-optimizations and
23558           -ffinite-math-only are in effect.
23559
23560       -mpretend-cmove
23561           Prefer zero-displacement conditional branches for conditional move
23562           instruction patterns.  This can result in faster code on the SH4
23563           processor.
23564
23565       -mfdpic
23566           Generate code using the FDPIC ABI.
23567
23568       Solaris 2 Options
23569
23570       These -m options are supported on Solaris 2:
23571
23572       -mclear-hwcap
23573           -mclear-hwcap tells the compiler to remove the hardware
23574           capabilities generated by the Solaris assembler.  This is only
23575           necessary when object files use ISA extensions not supported by the
23576           current machine, but check at runtime whether or not to use them.
23577
23578       -mimpure-text
23579           -mimpure-text, used in addition to -shared, tells the compiler to
23580           not pass -z text to the linker when linking a shared object.  Using
23581           this option, you can link position-dependent code into a shared
23582           object.
23583
23584           -mimpure-text suppresses the "relocations remain against
23585           allocatable but non-writable sections" linker error message.
23586           However, the necessary relocations trigger copy-on-write, and the
23587           shared object is not actually shared across processes.  Instead of
23588           using -mimpure-text, you should compile all source code with -fpic
23589           or -fPIC.
23590
23591       These switches are supported in addition to the above on Solaris 2:
23592
23593       -pthreads
23594           This is a synonym for -pthread.
23595
23596       SPARC Options
23597
23598       These -m options are supported on the SPARC:
23599
23600       -mno-app-regs
23601       -mapp-regs
23602           Specify -mapp-regs to generate output using the global registers 2
23603           through 4, which the SPARC SVR4 ABI reserves for applications.
23604           Like the global register 1, each global register 2 through 4 is
23605           then treated as an allocable register that is clobbered by function
23606           calls.  This is the default.
23607
23608           To be fully SVR4 ABI-compliant at the cost of some performance
23609           loss, specify -mno-app-regs.  You should compile libraries and
23610           system software with this option.
23611
23612       -mflat
23613       -mno-flat
23614           With -mflat, the compiler does not generate save/restore
23615           instructions and uses a "flat" or single register window model.
23616           This model is compatible with the regular register window model.
23617           The local registers and the input registers (0--5) are still
23618           treated as "call-saved" registers and are saved on the stack as
23619           needed.
23620
23621           With -mno-flat (the default), the compiler generates save/restore
23622           instructions (except for leaf functions).  This is the normal
23623           operating mode.
23624
23625       -mfpu
23626       -mhard-float
23627           Generate output containing floating-point instructions.  This is
23628           the default.
23629
23630       -mno-fpu
23631       -msoft-float
23632           Generate output containing library calls for floating point.
23633           Warning: the requisite libraries are not available for all SPARC
23634           targets.  Normally the facilities of the machine's usual C compiler
23635           are used, but this cannot be done directly in cross-compilation.
23636           You must make your own arrangements to provide suitable library
23637           functions for cross-compilation.  The embedded targets sparc-*-aout
23638           and sparclite-*-* do provide software floating-point support.
23639
23640           -msoft-float changes the calling convention in the output file;
23641           therefore, it is only useful if you compile all of a program with
23642           this option.  In particular, you need to compile libgcc.a, the
23643           library that comes with GCC, with -msoft-float in order for this to
23644           work.
23645
23646       -mhard-quad-float
23647           Generate output containing quad-word (long double) floating-point
23648           instructions.
23649
23650       -msoft-quad-float
23651           Generate output containing library calls for quad-word (long
23652           double) floating-point instructions.  The functions called are
23653           those specified in the SPARC ABI.  This is the default.
23654
23655           As of this writing, there are no SPARC implementations that have
23656           hardware support for the quad-word floating-point instructions.
23657           They all invoke a trap handler for one of these instructions, and
23658           then the trap handler emulates the effect of the instruction.
23659           Because of the trap handler overhead, this is much slower than
23660           calling the ABI library routines.  Thus the -msoft-quad-float
23661           option is the default.
23662
23663       -mno-unaligned-doubles
23664       -munaligned-doubles
23665           Assume that doubles have 8-byte alignment.  This is the default.
23666
23667           With -munaligned-doubles, GCC assumes that doubles have 8-byte
23668           alignment only if they are contained in another type, or if they
23669           have an absolute address.  Otherwise, it assumes they have 4-byte
23670           alignment.  Specifying this option avoids some rare compatibility
23671           problems with code generated by other compilers.  It is not the
23672           default because it results in a performance loss, especially for
23673           floating-point code.
23674
23675       -muser-mode
23676       -mno-user-mode
23677           Do not generate code that can only run in supervisor mode.  This is
23678           relevant only for the "casa" instruction emitted for the LEON3
23679           processor.  This is the default.
23680
23681       -mfaster-structs
23682       -mno-faster-structs
23683           With -mfaster-structs, the compiler assumes that structures should
23684           have 8-byte alignment.  This enables the use of pairs of "ldd" and
23685           "std" instructions for copies in structure assignment, in place of
23686           twice as many "ld" and "st" pairs.  However, the use of this
23687           changed alignment directly violates the SPARC ABI.  Thus, it's
23688           intended only for use on targets where the developer acknowledges
23689           that their resulting code is not directly in line with the rules of
23690           the ABI.
23691
23692       -mstd-struct-return
23693       -mno-std-struct-return
23694           With -mstd-struct-return, the compiler generates checking code in
23695           functions returning structures or unions to detect size mismatches
23696           between the two sides of function calls, as per the 32-bit ABI.
23697
23698           The default is -mno-std-struct-return.  This option has no effect
23699           in 64-bit mode.
23700
23701       -mlra
23702       -mno-lra
23703           Enable Local Register Allocation.  This is the default for SPARC
23704           since GCC 7 so -mno-lra needs to be passed to get old Reload.
23705
23706       -mcpu=cpu_type
23707           Set the instruction set, register set, and instruction scheduling
23708           parameters for machine type cpu_type.  Supported values for
23709           cpu_type are v7, cypress, v8, supersparc, hypersparc, leon, leon3,
23710           leon3v7, sparclite, f930, f934, sparclite86x, sparclet, tsc701, v9,
23711           ultrasparc, ultrasparc3, niagara, niagara2, niagara3, niagara4,
23712           niagara7 and m8.
23713
23714           Native Solaris and GNU/Linux toolchains also support the value
23715           native, which selects the best architecture option for the host
23716           processor.  -mcpu=native has no effect if GCC does not recognize
23717           the processor.
23718
23719           Default instruction scheduling parameters are used for values that
23720           select an architecture and not an implementation.  These are v7,
23721           v8, sparclite, sparclet, v9.
23722
23723           Here is a list of each supported architecture and their supported
23724           implementations.
23725
23726           v7  cypress, leon3v7
23727
23728           v8  supersparc, hypersparc, leon, leon3
23729
23730           sparclite
23731               f930, f934, sparclite86x
23732
23733           sparclet
23734               tsc701
23735
23736           v9  ultrasparc, ultrasparc3, niagara, niagara2, niagara3, niagara4,
23737               niagara7, m8
23738
23739           By default (unless configured otherwise), GCC generates code for
23740           the V7 variant of the SPARC architecture.  With -mcpu=cypress, the
23741           compiler additionally optimizes it for the Cypress CY7C602 chip, as
23742           used in the SPARCStation/SPARCServer 3xx series.  This is also
23743           appropriate for the older SPARCStation 1, 2, IPX etc.
23744
23745           With -mcpu=v8, GCC generates code for the V8 variant of the SPARC
23746           architecture.  The only difference from V7 code is that the
23747           compiler emits the integer multiply and integer divide instructions
23748           which exist in SPARC-V8 but not in SPARC-V7.  With
23749           -mcpu=supersparc, the compiler additionally optimizes it for the
23750           SuperSPARC chip, as used in the SPARCStation 10, 1000 and 2000
23751           series.
23752
23753           With -mcpu=sparclite, GCC generates code for the SPARClite variant
23754           of the SPARC architecture.  This adds the integer multiply, integer
23755           divide step and scan ("ffs") instructions which exist in SPARClite
23756           but not in SPARC-V7.  With -mcpu=f930, the compiler additionally
23757           optimizes it for the Fujitsu MB86930 chip, which is the original
23758           SPARClite, with no FPU.  With -mcpu=f934, the compiler additionally
23759           optimizes it for the Fujitsu MB86934 chip, which is the more recent
23760           SPARClite with FPU.
23761
23762           With -mcpu=sparclet, GCC generates code for the SPARClet variant of
23763           the SPARC architecture.  This adds the integer multiply,
23764           multiply/accumulate, integer divide step and scan ("ffs")
23765           instructions which exist in SPARClet but not in SPARC-V7.  With
23766           -mcpu=tsc701, the compiler additionally optimizes it for the TEMIC
23767           SPARClet chip.
23768
23769           With -mcpu=v9, GCC generates code for the V9 variant of the SPARC
23770           architecture.  This adds 64-bit integer and floating-point move
23771           instructions, 3 additional floating-point condition code registers
23772           and conditional move instructions.  With -mcpu=ultrasparc, the
23773           compiler additionally optimizes it for the Sun UltraSPARC I/II/IIi
23774           chips.  With -mcpu=ultrasparc3, the compiler additionally optimizes
23775           it for the Sun UltraSPARC III/III+/IIIi/IIIi+/IV/IV+ chips.  With
23776           -mcpu=niagara, the compiler additionally optimizes it for Sun
23777           UltraSPARC T1 chips.  With -mcpu=niagara2, the compiler
23778           additionally optimizes it for Sun UltraSPARC T2 chips. With
23779           -mcpu=niagara3, the compiler additionally optimizes it for Sun
23780           UltraSPARC T3 chips.  With -mcpu=niagara4, the compiler
23781           additionally optimizes it for Sun UltraSPARC T4 chips.  With
23782           -mcpu=niagara7, the compiler additionally optimizes it for Oracle
23783           SPARC M7 chips.  With -mcpu=m8, the compiler additionally optimizes
23784           it for Oracle M8 chips.
23785
23786       -mtune=cpu_type
23787           Set the instruction scheduling parameters for machine type
23788           cpu_type, but do not set the instruction set or register set that
23789           the option -mcpu=cpu_type does.
23790
23791           The same values for -mcpu=cpu_type can be used for -mtune=cpu_type,
23792           but the only useful values are those that select a particular CPU
23793           implementation.  Those are cypress, supersparc, hypersparc, leon,
23794           leon3, leon3v7, f930, f934, sparclite86x, tsc701, ultrasparc,
23795           ultrasparc3, niagara, niagara2, niagara3, niagara4, niagara7 and
23796           m8.  With native Solaris and GNU/Linux toolchains, native can also
23797           be used.
23798
23799       -mv8plus
23800       -mno-v8plus
23801           With -mv8plus, GCC generates code for the SPARC-V8+ ABI.  The
23802           difference from the V8 ABI is that the global and out registers are
23803           considered 64 bits wide.  This is enabled by default on Solaris in
23804           32-bit mode for all SPARC-V9 processors.
23805
23806       -mvis
23807       -mno-vis
23808           With -mvis, GCC generates code that takes advantage of the
23809           UltraSPARC Visual Instruction Set extensions.  The default is
23810           -mno-vis.
23811
23812       -mvis2
23813       -mno-vis2
23814           With -mvis2, GCC generates code that takes advantage of version 2.0
23815           of the UltraSPARC Visual Instruction Set extensions.  The default
23816           is -mvis2 when targeting a cpu that supports such instructions,
23817           such as UltraSPARC-III and later.  Setting -mvis2 also sets -mvis.
23818
23819       -mvis3
23820       -mno-vis3
23821           With -mvis3, GCC generates code that takes advantage of version 3.0
23822           of the UltraSPARC Visual Instruction Set extensions.  The default
23823           is -mvis3 when targeting a cpu that supports such instructions,
23824           such as niagara-3 and later.  Setting -mvis3 also sets -mvis2 and
23825           -mvis.
23826
23827       -mvis4
23828       -mno-vis4
23829           With -mvis4, GCC generates code that takes advantage of version 4.0
23830           of the UltraSPARC Visual Instruction Set extensions.  The default
23831           is -mvis4 when targeting a cpu that supports such instructions,
23832           such as niagara-7 and later.  Setting -mvis4 also sets -mvis3,
23833           -mvis2 and -mvis.
23834
23835       -mvis4b
23836       -mno-vis4b
23837           With -mvis4b, GCC generates code that takes advantage of version
23838           4.0 of the UltraSPARC Visual Instruction Set extensions, plus the
23839           additional VIS instructions introduced in the Oracle SPARC
23840           Architecture 2017.  The default is -mvis4b when targeting a cpu
23841           that supports such instructions, such as m8 and later.  Setting
23842           -mvis4b also sets -mvis4, -mvis3, -mvis2 and -mvis.
23843
23844       -mcbcond
23845       -mno-cbcond
23846           With -mcbcond, GCC generates code that takes advantage of the
23847           UltraSPARC Compare-and-Branch-on-Condition instructions.  The
23848           default is -mcbcond when targeting a CPU that supports such
23849           instructions, such as Niagara-4 and later.
23850
23851       -mfmaf
23852       -mno-fmaf
23853           With -mfmaf, GCC generates code that takes advantage of the
23854           UltraSPARC Fused Multiply-Add Floating-point instructions.  The
23855           default is -mfmaf when targeting a CPU that supports such
23856           instructions, such as Niagara-3 and later.
23857
23858       -mfsmuld
23859       -mno-fsmuld
23860           With -mfsmuld, GCC generates code that takes advantage of the
23861           Floating-point Multiply Single to Double (FsMULd) instruction.  The
23862           default is -mfsmuld when targeting a CPU supporting the
23863           architecture versions V8 or V9 with FPU except -mcpu=leon.
23864
23865       -mpopc
23866       -mno-popc
23867           With -mpopc, GCC generates code that takes advantage of the
23868           UltraSPARC Population Count instruction.  The default is -mpopc
23869           when targeting a CPU that supports such an instruction, such as
23870           Niagara-2 and later.
23871
23872       -msubxc
23873       -mno-subxc
23874           With -msubxc, GCC generates code that takes advantage of the
23875           UltraSPARC Subtract-Extended-with-Carry instruction.  The default
23876           is -msubxc when targeting a CPU that supports such an instruction,
23877           such as Niagara-7 and later.
23878
23879       -mfix-at697f
23880           Enable the documented workaround for the single erratum of the
23881           Atmel AT697F processor (which corresponds to erratum #13 of the
23882           AT697E processor).
23883
23884       -mfix-ut699
23885           Enable the documented workarounds for the floating-point errata and
23886           the data cache nullify errata of the UT699 processor.
23887
23888       -mfix-ut700
23889           Enable the documented workaround for the back-to-back store errata
23890           of the UT699E/UT700 processor.
23891
23892       -mfix-gr712rc
23893           Enable the documented workaround for the back-to-back store errata
23894           of the GR712RC processor.
23895
23896       These -m options are supported in addition to the above on SPARC-V9
23897       processors in 64-bit environments:
23898
23899       -m32
23900       -m64
23901           Generate code for a 32-bit or 64-bit environment.  The 32-bit
23902           environment sets int, long and pointer to 32 bits.  The 64-bit
23903           environment sets int to 32 bits and long and pointer to 64 bits.
23904
23905       -mcmodel=which
23906           Set the code model to one of
23907
23908           medlow
23909               The Medium/Low code model: 64-bit addresses, programs must be
23910               linked in the low 32 bits of memory.  Programs can be
23911               statically or dynamically linked.
23912
23913           medmid
23914               The Medium/Middle code model: 64-bit addresses, programs must
23915               be linked in the low 44 bits of memory, the text and data
23916               segments must be less than 2GB in size and the data segment
23917               must be located within 2GB of the text segment.
23918
23919           medany
23920               The Medium/Anywhere code model: 64-bit addresses, programs may
23921               be linked anywhere in memory, the text and data segments must
23922               be less than 2GB in size and the data segment must be located
23923               within 2GB of the text segment.
23924
23925           embmedany
23926               The Medium/Anywhere code model for embedded systems: 64-bit
23927               addresses, the text and data segments must be less than 2GB in
23928               size, both starting anywhere in memory (determined at link
23929               time).  The global register %g4 points to the base of the data
23930               segment.  Programs are statically linked and PIC is not
23931               supported.
23932
23933       -mmemory-model=mem-model
23934           Set the memory model in force on the processor to one of
23935
23936           default
23937               The default memory model for the processor and operating
23938               system.
23939
23940           rmo Relaxed Memory Order
23941
23942           pso Partial Store Order
23943
23944           tso Total Store Order
23945
23946           sc  Sequential Consistency
23947
23948           These memory models are formally defined in Appendix D of the
23949           SPARC-V9 architecture manual, as set in the processor's "PSTATE.MM"
23950           field.
23951
23952       -mstack-bias
23953       -mno-stack-bias
23954           With -mstack-bias, GCC assumes that the stack pointer, and frame
23955           pointer if present, are offset by -2047 which must be added back
23956           when making stack frame references.  This is the default in 64-bit
23957           mode.  Otherwise, assume no such offset is present.
23958
23959       Options for System V
23960
23961       These additional options are available on System V Release 4 for
23962       compatibility with other compilers on those systems:
23963
23964       -G  Create a shared object.  It is recommended that -symbolic or
23965           -shared be used instead.
23966
23967       -Qy Identify the versions of each tool used by the compiler, in a
23968           ".ident" assembler directive in the output.
23969
23970       -Qn Refrain from adding ".ident" directives to the output file (this is
23971           the default).
23972
23973       -YP,dirs
23974           Search the directories dirs, and no others, for libraries specified
23975           with -l.
23976
23977       -Ym,dir
23978           Look in the directory dir to find the M4 preprocessor.  The
23979           assembler uses this option.
23980
23981       TILE-Gx Options
23982
23983       These -m options are supported on the TILE-Gx:
23984
23985       -mcmodel=small
23986           Generate code for the small model.  The distance for direct calls
23987           is limited to 500M in either direction.  PC-relative addresses are
23988           32 bits.  Absolute addresses support the full address range.
23989
23990       -mcmodel=large
23991           Generate code for the large model.  There is no limitation on call
23992           distance, pc-relative addresses, or absolute addresses.
23993
23994       -mcpu=name
23995           Selects the type of CPU to be targeted.  Currently the only
23996           supported type is tilegx.
23997
23998       -m32
23999       -m64
24000           Generate code for a 32-bit or 64-bit environment.  The 32-bit
24001           environment sets int, long, and pointer to 32 bits.  The 64-bit
24002           environment sets int to 32 bits and long and pointer to 64 bits.
24003
24004       -mbig-endian
24005       -mlittle-endian
24006           Generate code in big/little endian mode, respectively.
24007
24008       TILEPro Options
24009
24010       These -m options are supported on the TILEPro:
24011
24012       -mcpu=name
24013           Selects the type of CPU to be targeted.  Currently the only
24014           supported type is tilepro.
24015
24016       -m32
24017           Generate code for a 32-bit environment, which sets int, long, and
24018           pointer to 32 bits.  This is the only supported behavior so the
24019           flag is essentially ignored.
24020
24021       V850 Options
24022
24023       These -m options are defined for V850 implementations:
24024
24025       -mlong-calls
24026       -mno-long-calls
24027           Treat all calls as being far away (near).  If calls are assumed to
24028           be far away, the compiler always loads the function's address into
24029           a register, and calls indirect through the pointer.
24030
24031       -mno-ep
24032       -mep
24033           Do not optimize (do optimize) basic blocks that use the same index
24034           pointer 4 or more times to copy pointer into the "ep" register, and
24035           use the shorter "sld" and "sst" instructions.  The -mep option is
24036           on by default if you optimize.
24037
24038       -mno-prolog-function
24039       -mprolog-function
24040           Do not use (do use) external functions to save and restore
24041           registers at the prologue and epilogue of a function.  The external
24042           functions are slower, but use less code space if more than one
24043           function saves the same number of registers.  The -mprolog-function
24044           option is on by default if you optimize.
24045
24046       -mspace
24047           Try to make the code as small as possible.  At present, this just
24048           turns on the -mep and -mprolog-function options.
24049
24050       -mtda=n
24051           Put static or global variables whose size is n bytes or less into
24052           the tiny data area that register "ep" points to.  The tiny data
24053           area can hold up to 256 bytes in total (128 bytes for byte
24054           references).
24055
24056       -msda=n
24057           Put static or global variables whose size is n bytes or less into
24058           the small data area that register "gp" points to.  The small data
24059           area can hold up to 64 kilobytes.
24060
24061       -mzda=n
24062           Put static or global variables whose size is n bytes or less into
24063           the first 32 kilobytes of memory.
24064
24065       -mv850
24066           Specify that the target processor is the V850.
24067
24068       -mv850e3v5
24069           Specify that the target processor is the V850E3V5.  The
24070           preprocessor constant "__v850e3v5__" is defined if this option is
24071           used.
24072
24073       -mv850e2v4
24074           Specify that the target processor is the V850E3V5.  This is an
24075           alias for the -mv850e3v5 option.
24076
24077       -mv850e2v3
24078           Specify that the target processor is the V850E2V3.  The
24079           preprocessor constant "__v850e2v3__" is defined if this option is
24080           used.
24081
24082       -mv850e2
24083           Specify that the target processor is the V850E2.  The preprocessor
24084           constant "__v850e2__" is defined if this option is used.
24085
24086       -mv850e1
24087           Specify that the target processor is the V850E1.  The preprocessor
24088           constants "__v850e1__" and "__v850e__" are defined if this option
24089           is used.
24090
24091       -mv850es
24092           Specify that the target processor is the V850ES.  This is an alias
24093           for the -mv850e1 option.
24094
24095       -mv850e
24096           Specify that the target processor is the V850E.  The preprocessor
24097           constant "__v850e__" is defined if this option is used.
24098
24099           If neither -mv850 nor -mv850e nor -mv850e1 nor -mv850e2 nor
24100           -mv850e2v3 nor -mv850e3v5 are defined then a default target
24101           processor is chosen and the relevant __v850*__ preprocessor
24102           constant is defined.
24103
24104           The preprocessor constants "__v850" and "__v851__" are always
24105           defined, regardless of which processor variant is the target.
24106
24107       -mdisable-callt
24108       -mno-disable-callt
24109           This option suppresses generation of the "CALLT" instruction for
24110           the v850e, v850e1, v850e2, v850e2v3 and v850e3v5 flavors of the
24111           v850 architecture.
24112
24113           This option is enabled by default when the RH850 ABI is in use (see
24114           -mrh850-abi), and disabled by default when the GCC ABI is in use.
24115           If "CALLT" instructions are being generated then the C preprocessor
24116           symbol "__V850_CALLT__" is defined.
24117
24118       -mrelax
24119       -mno-relax
24120           Pass on (or do not pass on) the -mrelax command-line option to the
24121           assembler.
24122
24123       -mlong-jumps
24124       -mno-long-jumps
24125           Disable (or re-enable) the generation of PC-relative jump
24126           instructions.
24127
24128       -msoft-float
24129       -mhard-float
24130           Disable (or re-enable) the generation of hardware floating point
24131           instructions.  This option is only significant when the target
24132           architecture is V850E2V3 or higher.  If hardware floating point
24133           instructions are being generated then the C preprocessor symbol
24134           "__FPU_OK__" is defined, otherwise the symbol "__NO_FPU__" is
24135           defined.
24136
24137       -mloop
24138           Enables the use of the e3v5 LOOP instruction.  The use of this
24139           instruction is not enabled by default when the e3v5 architecture is
24140           selected because its use is still experimental.
24141
24142       -mrh850-abi
24143       -mghs
24144           Enables support for the RH850 version of the V850 ABI.  This is the
24145           default.  With this version of the ABI the following rules apply:
24146
24147           *   Integer sized structures and unions are returned via a memory
24148               pointer rather than a register.
24149
24150           *   Large structures and unions (more than 8 bytes in size) are
24151               passed by value.
24152
24153           *   Functions are aligned to 16-bit boundaries.
24154
24155           *   The -m8byte-align command-line option is supported.
24156
24157           *   The -mdisable-callt command-line option is enabled by default.
24158               The -mno-disable-callt command-line option is not supported.
24159
24160           When this version of the ABI is enabled the C preprocessor symbol
24161           "__V850_RH850_ABI__" is defined.
24162
24163       -mgcc-abi
24164           Enables support for the old GCC version of the V850 ABI.  With this
24165           version of the ABI the following rules apply:
24166
24167           *   Integer sized structures and unions are returned in register
24168               "r10".
24169
24170           *   Large structures and unions (more than 8 bytes in size) are
24171               passed by reference.
24172
24173           *   Functions are aligned to 32-bit boundaries, unless optimizing
24174               for size.
24175
24176           *   The -m8byte-align command-line option is not supported.
24177
24178           *   The -mdisable-callt command-line option is supported but not
24179               enabled by default.
24180
24181           When this version of the ABI is enabled the C preprocessor symbol
24182           "__V850_GCC_ABI__" is defined.
24183
24184       -m8byte-align
24185       -mno-8byte-align
24186           Enables support for "double" and "long long" types to be aligned on
24187           8-byte boundaries.  The default is to restrict the alignment of all
24188           objects to at most 4-bytes.  When -m8byte-align is in effect the C
24189           preprocessor symbol "__V850_8BYTE_ALIGN__" is defined.
24190
24191       -mbig-switch
24192           Generate code suitable for big switch tables.  Use this option only
24193           if the assembler/linker complain about out of range branches within
24194           a switch table.
24195
24196       -mapp-regs
24197           This option causes r2 and r5 to be used in the code generated by
24198           the compiler.  This setting is the default.
24199
24200       -mno-app-regs
24201           This option causes r2 and r5 to be treated as fixed registers.
24202
24203       VAX Options
24204
24205       These -m options are defined for the VAX:
24206
24207       -munix
24208           Do not output certain jump instructions ("aobleq" and so on) that
24209           the Unix assembler for the VAX cannot handle across long ranges.
24210
24211       -mgnu
24212           Do output those jump instructions, on the assumption that the GNU
24213           assembler is being used.
24214
24215       -mg Output code for G-format floating-point numbers instead of
24216           D-format.
24217
24218       Visium Options
24219
24220       -mdebug
24221           A program which performs file I/O and is destined to run on an MCM
24222           target should be linked with this option.  It causes the libraries
24223           libc.a and libdebug.a to be linked.  The program should be run on
24224           the target under the control of the GDB remote debugging stub.
24225
24226       -msim
24227           A program which performs file I/O and is destined to run on the
24228           simulator should be linked with option.  This causes libraries
24229           libc.a and libsim.a to be linked.
24230
24231       -mfpu
24232       -mhard-float
24233           Generate code containing floating-point instructions.  This is the
24234           default.
24235
24236       -mno-fpu
24237       -msoft-float
24238           Generate code containing library calls for floating-point.
24239
24240           -msoft-float changes the calling convention in the output file;
24241           therefore, it is only useful if you compile all of a program with
24242           this option.  In particular, you need to compile libgcc.a, the
24243           library that comes with GCC, with -msoft-float in order for this to
24244           work.
24245
24246       -mcpu=cpu_type
24247           Set the instruction set, register set, and instruction scheduling
24248           parameters for machine type cpu_type.  Supported values for
24249           cpu_type are mcm, gr5 and gr6.
24250
24251           mcm is a synonym of gr5 present for backward compatibility.
24252
24253           By default (unless configured otherwise), GCC generates code for
24254           the GR5 variant of the Visium architecture.
24255
24256           With -mcpu=gr6, GCC generates code for the GR6 variant of the
24257           Visium architecture.  The only difference from GR5 code is that the
24258           compiler will generate block move instructions.
24259
24260       -mtune=cpu_type
24261           Set the instruction scheduling parameters for machine type
24262           cpu_type, but do not set the instruction set or register set that
24263           the option -mcpu=cpu_type would.
24264
24265       -msv-mode
24266           Generate code for the supervisor mode, where there are no
24267           restrictions on the access to general registers.  This is the
24268           default.
24269
24270       -muser-mode
24271           Generate code for the user mode, where the access to some general
24272           registers is forbidden: on the GR5, registers r24 to r31 cannot be
24273           accessed in this mode; on the GR6, only registers r29 to r31 are
24274           affected.
24275
24276       VMS Options
24277
24278       These -m options are defined for the VMS implementations:
24279
24280       -mvms-return-codes
24281           Return VMS condition codes from "main". The default is to return
24282           POSIX-style condition (e.g. error) codes.
24283
24284       -mdebug-main=prefix
24285           Flag the first routine whose name starts with prefix as the main
24286           routine for the debugger.
24287
24288       -mmalloc64
24289           Default to 64-bit memory allocation routines.
24290
24291       -mpointer-size=size
24292           Set the default size of pointers. Possible options for size are 32
24293           or short for 32 bit pointers, 64 or long for 64 bit pointers, and
24294           no for supporting only 32 bit pointers.  The later option disables
24295           "pragma pointer_size".
24296
24297       VxWorks Options
24298
24299       The options in this section are defined for all VxWorks targets.
24300       Options specific to the target hardware are listed with the other
24301       options for that target.
24302
24303       -mrtp
24304           GCC can generate code for both VxWorks kernels and real time
24305           processes (RTPs).  This option switches from the former to the
24306           latter.  It also defines the preprocessor macro "__RTP__".
24307
24308       -non-static
24309           Link an RTP executable against shared libraries rather than static
24310           libraries.  The options -static and -shared can also be used for
24311           RTPs; -static is the default.
24312
24313       -Bstatic
24314       -Bdynamic
24315           These options are passed down to the linker.  They are defined for
24316           compatibility with Diab.
24317
24318       -Xbind-lazy
24319           Enable lazy binding of function calls.  This option is equivalent
24320           to -Wl,-z,now and is defined for compatibility with Diab.
24321
24322       -Xbind-now
24323           Disable lazy binding of function calls.  This option is the default
24324           and is defined for compatibility with Diab.
24325
24326       x86 Options
24327
24328       These -m options are defined for the x86 family of computers.
24329
24330       -march=cpu-type
24331           Generate instructions for the machine type cpu-type.  In contrast
24332           to -mtune=cpu-type, which merely tunes the generated code for the
24333           specified cpu-type, -march=cpu-type allows GCC to generate code
24334           that may not run at all on processors other than the one indicated.
24335           Specifying -march=cpu-type implies -mtune=cpu-type.
24336
24337           The choices for cpu-type are:
24338
24339           native
24340               This selects the CPU to generate code for at compilation time
24341               by determining the processor type of the compiling machine.
24342               Using -march=native enables all instruction subsets supported
24343               by the local machine (hence the result might not run on
24344               different machines).  Using -mtune=native produces code
24345               optimized for the local machine under the constraints of the
24346               selected instruction set.
24347
24348           x86-64
24349               A generic CPU with 64-bit extensions.
24350
24351           i386
24352               Original Intel i386 CPU.
24353
24354           i486
24355               Intel i486 CPU.  (No scheduling is implemented for this chip.)
24356
24357           i586
24358           pentium
24359               Intel Pentium CPU with no MMX support.
24360
24361           lakemont
24362               Intel Lakemont MCU, based on Intel Pentium CPU.
24363
24364           pentium-mmx
24365               Intel Pentium MMX CPU, based on Pentium core with MMX
24366               instruction set support.
24367
24368           pentiumpro
24369               Intel Pentium Pro CPU.
24370
24371           i686
24372               When used with -march, the Pentium Pro instruction set is used,
24373               so the code runs on all i686 family chips.  When used with
24374               -mtune, it has the same meaning as generic.
24375
24376           pentium2
24377               Intel Pentium II CPU, based on Pentium Pro core with MMX
24378               instruction set support.
24379
24380           pentium3
24381           pentium3m
24382               Intel Pentium III CPU, based on Pentium Pro core with MMX and
24383               SSE instruction set support.
24384
24385           pentium-m
24386               Intel Pentium M; low-power version of Intel Pentium III CPU
24387               with MMX, SSE and SSE2 instruction set support.  Used by
24388               Centrino notebooks.
24389
24390           pentium4
24391           pentium4m
24392               Intel Pentium 4 CPU with MMX, SSE and SSE2 instruction set
24393               support.
24394
24395           prescott
24396               Improved version of Intel Pentium 4 CPU with MMX, SSE, SSE2 and
24397               SSE3 instruction set support.
24398
24399           nocona
24400               Improved version of Intel Pentium 4 CPU with 64-bit extensions,
24401               MMX, SSE, SSE2 and SSE3 instruction set support.
24402
24403           core2
24404               Intel Core 2 CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3
24405               and SSSE3 instruction set support.
24406
24407           nehalem
24408               Intel Nehalem CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3,
24409               SSSE3, SSE4.1, SSE4.2 and POPCNT instruction set support.
24410
24411           westmere
24412               Intel Westmere CPU with 64-bit extensions, MMX, SSE, SSE2,
24413               SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, AES and PCLMUL instruction
24414               set support.
24415
24416           sandybridge
24417               Intel Sandy Bridge CPU with 64-bit extensions, MMX, SSE, SSE2,
24418               SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, AVX, AES and PCLMUL
24419               instruction set support.
24420
24421           ivybridge
24422               Intel Ivy Bridge CPU with 64-bit extensions, MMX, SSE, SSE2,
24423               SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, AVX, AES, PCLMUL,
24424               FSGSBASE, RDRND and F16C instruction set support.
24425
24426           haswell
24427               Intel Haswell CPU with 64-bit extensions, MOVBE, MMX, SSE,
24428               SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES,
24429               PCLMUL, FSGSBASE, RDRND, FMA, BMI, BMI2 and F16C instruction
24430               set support.
24431
24432           broadwell
24433               Intel Broadwell CPU with 64-bit extensions, MOVBE, MMX, SSE,
24434               SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES,
24435               PCLMUL, FSGSBASE, RDRND, FMA, BMI, BMI2, F16C, RDSEED, ADCX and
24436               PREFETCHW instruction set support.
24437
24438           skylake
24439               Intel Skylake CPU with 64-bit extensions, MOVBE, MMX, SSE,
24440               SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES,
24441               PCLMUL, FSGSBASE, RDRND, FMA, BMI, BMI2, F16C, RDSEED, ADCX,
24442               PREFETCHW, CLFLUSHOPT, XSAVEC and XSAVES instruction set
24443               support.
24444
24445           bonnell
24446               Intel Bonnell CPU with 64-bit extensions, MOVBE, MMX, SSE,
24447               SSE2, SSE3 and SSSE3 instruction set support.
24448
24449           silvermont
24450               Intel Silvermont CPU with 64-bit extensions, MOVBE, MMX, SSE,
24451               SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, AES, PCLMUL and
24452               RDRND instruction set support.
24453
24454           goldmont
24455               Intel Goldmont CPU with 64-bit extensions, MOVBE, MMX, SSE,
24456               SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, AES, PCLMUL, RDRND,
24457               XSAVE, XSAVEOPT and FSGSBASE instruction set support.
24458
24459           goldmont-plus
24460               Intel Goldmont Plus CPU with 64-bit extensions, MOVBE, MMX,
24461               SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, AES, PCLMUL,
24462               RDRND, XSAVE, XSAVEOPT, FSGSBASE, PTWRITE, RDPID, SGX and UMIP
24463               instruction set support.
24464
24465           tremont
24466               Intel Tremont CPU with 64-bit extensions, MOVBE, MMX, SSE,
24467               SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, AES, PCLMUL, RDRND,
24468               XSAVE, XSAVEOPT, FSGSBASE, PTWRITE, RDPID, SGX, UMIP, GFNI-SSE,
24469               CLWB and ENCLV instruction set support.
24470
24471           knl Intel Knight's Landing CPU with 64-bit extensions, MOVBE, MMX,
24472               SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES,
24473               PCLMUL, FSGSBASE, RDRND, FMA, BMI, BMI2, F16C, RDSEED, ADCX,
24474               PREFETCHW, AVX512F, AVX512PF, AVX512ER and AVX512CD instruction
24475               set support.
24476
24477           knm Intel Knights Mill CPU with 64-bit extensions, MOVBE, MMX, SSE,
24478               SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES,
24479               PCLMUL, FSGSBASE, RDRND, FMA, BMI, BMI2, F16C, RDSEED, ADCX,
24480               PREFETCHW, AVX512F, AVX512PF, AVX512ER, AVX512CD, AVX5124VNNIW,
24481               AVX5124FMAPS and AVX512VPOPCNTDQ instruction set support.
24482
24483           skylake-avx512
24484               Intel Skylake Server CPU with 64-bit extensions, MOVBE, MMX,
24485               SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, PKU, AVX, AVX2,
24486               AES, PCLMUL, FSGSBASE, RDRND, FMA, BMI, BMI2, F16C, RDSEED,
24487               ADCX, PREFETCHW, CLFLUSHOPT, XSAVEC, XSAVES, AVX512F, CLWB,
24488               AVX512VL, AVX512BW, AVX512DQ and AVX512CD instruction set
24489               support.
24490
24491           cannonlake
24492               Intel Cannonlake Server CPU with 64-bit extensions, MOVBE, MMX,
24493               SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, PKU, AVX, AVX2,
24494               AES, PCLMUL, FSGSBASE, RDRND, FMA, BMI, BMI2, F16C, RDSEED,
24495               ADCX, PREFETCHW, CLFLUSHOPT, XSAVEC, XSAVES, AVX512F, AVX512VL,
24496               AVX512BW, AVX512DQ, AVX512CD, AVX512VBMI, AVX512IFMA, SHA and
24497               UMIP instruction set support.
24498
24499           icelake-client
24500               Intel Icelake Client CPU with 64-bit extensions, MOVBE, MMX,
24501               SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, PKU, AVX, AVX2,
24502               AES, PCLMUL, FSGSBASE, RDRND, FMA, BMI, BMI2, F16C, RDSEED,
24503               ADCX, PREFETCHW, CLFLUSHOPT, XSAVEC, XSAVES, AVX512F, AVX512VL,
24504               AVX512BW, AVX512DQ, AVX512CD, AVX512VBMI, AVX512IFMA, SHA,
24505               CLWB, UMIP, RDPID, GFNI, AVX512VBMI2, AVX512VPOPCNTDQ,
24506               AVX512BITALG, AVX512VNNI, VPCLMULQDQ, VAES instruction set
24507               support.
24508
24509           icelake-server
24510               Intel Icelake Server CPU with 64-bit extensions, MOVBE, MMX,
24511               SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, PKU, AVX, AVX2,
24512               AES, PCLMUL, FSGSBASE, RDRND, FMA, BMI, BMI2, F16C, RDSEED,
24513               ADCX, PREFETCHW, CLFLUSHOPT, XSAVEC, XSAVES, AVX512F, AVX512VL,
24514               AVX512BW, AVX512DQ, AVX512CD, AVX512VBMI, AVX512IFMA, SHA,
24515               CLWB, UMIP, RDPID, GFNI, AVX512VBMI2, AVX512VPOPCNTDQ,
24516               AVX512BITALG, AVX512VNNI, VPCLMULQDQ, VAES, PCONFIG and
24517               WBNOINVD instruction set support.
24518
24519           cascadelake
24520               Intel Cascadelake CPU with 64-bit extensions, MOVBE, MMX, SSE,
24521               SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, PKU, AVX, AVX2, AES,
24522               PCLMUL, FSGSBASE, RDRND, FMA, BMI, BMI2, F16C, RDSEED, ADCX,
24523               PREFETCHW, CLFLUSHOPT, XSAVEC, XSAVES, AVX512F, CLWB, AVX512VL,
24524               AVX512BW, AVX512DQ, AVX512CD and AVX512VNNI instruction set
24525               support.
24526
24527           cooperlake
24528               Intel cooperlake CPU with 64-bit extensions, MOVBE, MMX, SSE,
24529               SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, PKU, AVX, AVX2, AES,
24530               PCLMUL, FSGSBASE, RDRND, FMA, BMI, BMI2, F16C, RDSEED, ADCX,
24531               PREFETCHW, CLFLUSHOPT, XSAVEC, XSAVES, AVX512F, CLWB, AVX512VL,
24532               AVX512BW, AVX512DQ, AVX512CD, AVX512VNNI and AVX512BF16
24533               instruction set support.
24534
24535           tigerlake
24536               Intel Tigerlake CPU with 64-bit extensions, MOVBE, MMX, SSE,
24537               SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, PKU, AVX, AVX2, AES,
24538               PCLMUL, FSGSBASE, RDRND, FMA, BMI, BMI2, F16C, RDSEED, ADCX,
24539               PREFETCHW, CLFLUSHOPT, XSAVEC, XSAVES, AVX512F, AVX512VL,
24540               AVX512BW, AVX512DQ, AVX512CD, AVX512VBMI, AVX512IFMA, SHA,
24541               CLWB, UMIP, RDPID, GFNI, AVX512VBMI2, AVX512VPOPCNTDQ,
24542               AVX512BITALG, AVX512VNNI, VPCLMULQDQ, VAES, PCONFIG, WBNOINVD,
24543               MOVDIRI, MOVDIR64B and  AVX512VP2INTERSECT instruction set
24544               support.
24545
24546           k6  AMD K6 CPU with MMX instruction set support.
24547
24548           k6-2
24549           k6-3
24550               Improved versions of AMD K6 CPU with MMX and 3DNow! instruction
24551               set support.
24552
24553           athlon
24554           athlon-tbird
24555               AMD Athlon CPU with MMX, 3dNOW!, enhanced 3DNow! and SSE
24556               prefetch instructions support.
24557
24558           athlon-4
24559           athlon-xp
24560           athlon-mp
24561               Improved AMD Athlon CPU with MMX, 3DNow!, enhanced 3DNow! and
24562               full SSE instruction set support.
24563
24564           k8
24565           opteron
24566           athlon64
24567           athlon-fx
24568               Processors based on the AMD K8 core with x86-64 instruction set
24569               support, including the AMD Opteron, Athlon 64, and Athlon 64 FX
24570               processors.  (This supersets MMX, SSE, SSE2, 3DNow!, enhanced
24571               3DNow! and 64-bit instruction set extensions.)
24572
24573           k8-sse3
24574           opteron-sse3
24575           athlon64-sse3
24576               Improved versions of AMD K8 cores with SSE3 instruction set
24577               support.
24578
24579           amdfam10
24580           barcelona
24581               CPUs based on AMD Family 10h cores with x86-64 instruction set
24582               support.  (This supersets MMX, SSE, SSE2, SSE3, SSE4A, 3DNow!,
24583               enhanced 3DNow!, ABM and 64-bit instruction set extensions.)
24584
24585           bdver1
24586               CPUs based on AMD Family 15h cores with x86-64 instruction set
24587               support.  (This supersets FMA4, AVX, XOP, LWP, AES, PCLMUL,
24588               CX16, MMX, SSE, SSE2, SSE3, SSE4A, SSSE3, SSE4.1, SSE4.2, ABM
24589               and 64-bit instruction set extensions.)
24590
24591           bdver2
24592               AMD Family 15h core based CPUs with x86-64 instruction set
24593               support.  (This supersets BMI, TBM, F16C, FMA, FMA4, AVX, XOP,
24594               LWP, AES, PCLMUL, CX16, MMX, SSE, SSE2, SSE3, SSE4A, SSSE3,
24595               SSE4.1, SSE4.2, ABM and 64-bit instruction set extensions.)
24596
24597           bdver3
24598               AMD Family 15h core based CPUs with x86-64 instruction set
24599               support.  (This supersets BMI, TBM, F16C, FMA, FMA4, FSGSBASE,
24600               AVX, XOP, LWP, AES, PCLMUL, CX16, MMX, SSE, SSE2, SSE3, SSE4A,
24601               SSSE3, SSE4.1, SSE4.2, ABM and 64-bit instruction set
24602               extensions.)
24603
24604           bdver4
24605               AMD Family 15h core based CPUs with x86-64 instruction set
24606               support.  (This supersets BMI, BMI2, TBM, F16C, FMA, FMA4,
24607               FSGSBASE, AVX, AVX2, XOP, LWP, AES, PCLMUL, CX16, MOVBE, MMX,
24608               SSE, SSE2, SSE3, SSE4A, SSSE3, SSE4.1, SSE4.2, ABM and 64-bit
24609               instruction set extensions.)
24610
24611           znver1
24612               AMD Family 17h core based CPUs with x86-64 instruction set
24613               support.  (This supersets BMI, BMI2, F16C, FMA, FSGSBASE, AVX,
24614               AVX2, ADCX, RDSEED, MWAITX, SHA, CLZERO, AES, PCLMUL, CX16,
24615               MOVBE, MMX, SSE, SSE2, SSE3, SSE4A, SSSE3, SSE4.1, SSE4.2, ABM,
24616               XSAVEC, XSAVES, CLFLUSHOPT, POPCNT, and 64-bit instruction set
24617               extensions.)
24618
24619           znver2
24620               AMD Family 17h core based CPUs with x86-64 instruction set
24621               support. (This supersets BMI, BMI2, CLWB, F16C, FMA, FSGSBASE,
24622               AVX, AVX2, ADCX, RDSEED, MWAITX, SHA, CLZERO, AES, PCLMUL,
24623               CX16, MOVBE, MMX, SSE, SSE2, SSE3, SSE4A, SSSE3, SSE4.1,
24624               SSE4.2, ABM, XSAVEC, XSAVES, CLFLUSHOPT, POPCNT, RDPID,
24625               WBNOINVD, and 64-bit instruction set extensions.)
24626
24627           btver1
24628               CPUs based on AMD Family 14h cores with x86-64 instruction set
24629               support.  (This supersets MMX, SSE, SSE2, SSE3, SSSE3, SSE4A,
24630               CX16, ABM and 64-bit instruction set extensions.)
24631
24632           btver2
24633               CPUs based on AMD Family 16h cores with x86-64 instruction set
24634               support. This includes MOVBE, F16C, BMI, AVX, PCLMUL, AES,
24635               SSE4.2, SSE4.1, CX16, ABM, SSE4A, SSSE3, SSE3, SSE2, SSE, MMX
24636               and 64-bit instruction set extensions.
24637
24638           winchip-c6
24639               IDT WinChip C6 CPU, dealt in same way as i486 with additional
24640               MMX instruction set support.
24641
24642           winchip2
24643               IDT WinChip 2 CPU, dealt in same way as i486 with additional
24644               MMX and 3DNow!  instruction set support.
24645
24646           c3  VIA C3 CPU with MMX and 3DNow! instruction set support.  (No
24647               scheduling is implemented for this chip.)
24648
24649           c3-2
24650               VIA C3-2 (Nehemiah/C5XL) CPU with MMX and SSE instruction set
24651               support.  (No scheduling is implemented for this chip.)
24652
24653           c7  VIA C7 (Esther) CPU with MMX, SSE, SSE2 and SSE3 instruction
24654               set support.  (No scheduling is implemented for this chip.)
24655
24656           samuel-2
24657               VIA Eden Samuel 2 CPU with MMX and 3DNow! instruction set
24658               support.  (No scheduling is implemented for this chip.)
24659
24660           nehemiah
24661               VIA Eden Nehemiah CPU with MMX and SSE instruction set support.
24662               (No scheduling is implemented for this chip.)
24663
24664           esther
24665               VIA Eden Esther CPU with MMX, SSE, SSE2 and SSE3 instruction
24666               set support.  (No scheduling is implemented for this chip.)
24667
24668           eden-x2
24669               VIA Eden X2 CPU with x86-64, MMX, SSE, SSE2 and SSE3
24670               instruction set support.  (No scheduling is implemented for
24671               this chip.)
24672
24673           eden-x4
24674               VIA Eden X4 CPU with x86-64, MMX, SSE, SSE2, SSE3, SSSE3,
24675               SSE4.1, SSE4.2, AVX and AVX2 instruction set support.  (No
24676               scheduling is implemented for this chip.)
24677
24678           nano
24679               Generic VIA Nano CPU with x86-64, MMX, SSE, SSE2, SSE3 and
24680               SSSE3 instruction set support.  (No scheduling is implemented
24681               for this chip.)
24682
24683           nano-1000
24684               VIA Nano 1xxx CPU with x86-64, MMX, SSE, SSE2, SSE3 and SSSE3
24685               instruction set support.  (No scheduling is implemented for
24686               this chip.)
24687
24688           nano-2000
24689               VIA Nano 2xxx CPU with x86-64, MMX, SSE, SSE2, SSE3 and SSSE3
24690               instruction set support.  (No scheduling is implemented for
24691               this chip.)
24692
24693           nano-3000
24694               VIA Nano 3xxx CPU with x86-64, MMX, SSE, SSE2, SSE3, SSSE3 and
24695               SSE4.1 instruction set support.  (No scheduling is implemented
24696               for this chip.)
24697
24698           nano-x2
24699               VIA Nano Dual Core CPU with x86-64, MMX, SSE, SSE2, SSE3, SSSE3
24700               and SSE4.1 instruction set support.  (No scheduling is
24701               implemented for this chip.)
24702
24703           nano-x4
24704               VIA Nano Quad Core CPU with x86-64, MMX, SSE, SSE2, SSE3, SSSE3
24705               and SSE4.1 instruction set support.  (No scheduling is
24706               implemented for this chip.)
24707
24708           geode
24709               AMD Geode embedded processor with MMX and 3DNow! instruction
24710               set support.
24711
24712       -mtune=cpu-type
24713           Tune to cpu-type everything applicable about the generated code,
24714           except for the ABI and the set of available instructions.  While
24715           picking a specific cpu-type schedules things appropriately for that
24716           particular chip, the compiler does not generate any code that
24717           cannot run on the default machine type unless you use a -march=cpu-
24718           type option.  For example, if GCC is configured for
24719           i686-pc-linux-gnu then -mtune=pentium4 generates code that is tuned
24720           for Pentium 4 but still runs on i686 machines.
24721
24722           The choices for cpu-type are the same as for -march.  In addition,
24723           -mtune supports 2 extra choices for cpu-type:
24724
24725           generic
24726               Produce code optimized for the most common IA32/AMD64/EM64T
24727               processors.  If you know the CPU on which your code will run,
24728               then you should use the corresponding -mtune or -march option
24729               instead of -mtune=generic.  But, if you do not know exactly
24730               what CPU users of your application will have, then you should
24731               use this option.
24732
24733               As new processors are deployed in the marketplace, the behavior
24734               of this option will change.  Therefore, if you upgrade to a
24735               newer version of GCC, code generation controlled by this option
24736               will change to reflect the processors that are most common at
24737               the time that version of GCC is released.
24738
24739               There is no -march=generic option because -march indicates the
24740               instruction set the compiler can use, and there is no generic
24741               instruction set applicable to all processors.  In contrast,
24742               -mtune indicates the processor (or, in this case, collection of
24743               processors) for which the code is optimized.
24744
24745           intel
24746               Produce code optimized for the most current Intel processors,
24747               which are Haswell and Silvermont for this version of GCC.  If
24748               you know the CPU on which your code will run, then you should
24749               use the corresponding -mtune or -march option instead of
24750               -mtune=intel.  But, if you want your application performs
24751               better on both Haswell and Silvermont, then you should use this
24752               option.
24753
24754               As new Intel processors are deployed in the marketplace, the
24755               behavior of this option will change.  Therefore, if you upgrade
24756               to a newer version of GCC, code generation controlled by this
24757               option will change to reflect the most current Intel processors
24758               at the time that version of GCC is released.
24759
24760               There is no -march=intel option because -march indicates the
24761               instruction set the compiler can use, and there is no common
24762               instruction set applicable to all processors.  In contrast,
24763               -mtune indicates the processor (or, in this case, collection of
24764               processors) for which the code is optimized.
24765
24766       -mcpu=cpu-type
24767           A deprecated synonym for -mtune.
24768
24769       -mfpmath=unit
24770           Generate floating-point arithmetic for selected unit unit.  The
24771           choices for unit are:
24772
24773           387 Use the standard 387 floating-point coprocessor present on the
24774               majority of chips and emulated otherwise.  Code compiled with
24775               this option runs almost everywhere.  The temporary results are
24776               computed in 80-bit precision instead of the precision specified
24777               by the type, resulting in slightly different results compared
24778               to most of other chips.  See -ffloat-store for more detailed
24779               description.
24780
24781               This is the default choice for non-Darwin x86-32 targets.
24782
24783           sse Use scalar floating-point instructions present in the SSE
24784               instruction set.  This instruction set is supported by Pentium
24785               III and newer chips, and in the AMD line by Athlon-4, Athlon XP
24786               and Athlon MP chips.  The earlier version of the SSE
24787               instruction set supports only single-precision arithmetic, thus
24788               the double and extended-precision arithmetic are still done
24789               using 387.  A later version, present only in Pentium 4 and AMD
24790               x86-64 chips, supports double-precision arithmetic too.
24791
24792               For the x86-32 compiler, you must use -march=cpu-type, -msse or
24793               -msse2 switches to enable SSE extensions and make this option
24794               effective.  For the x86-64 compiler, these extensions are
24795               enabled by default.
24796
24797               The resulting code should be considerably faster in the
24798               majority of cases and avoid the numerical instability problems
24799               of 387 code, but may break some existing code that expects
24800               temporaries to be 80 bits.
24801
24802               This is the default choice for the x86-64 compiler, Darwin
24803               x86-32 targets, and the default choice for x86-32 targets with
24804               the SSE2 instruction set when -ffast-math is enabled.
24805
24806           sse,387
24807           sse+387
24808           both
24809               Attempt to utilize both instruction sets at once.  This
24810               effectively doubles the amount of available registers, and on
24811               chips with separate execution units for 387 and SSE the
24812               execution resources too.  Use this option with care, as it is
24813               still experimental, because the GCC register allocator does not
24814               model separate functional units well, resulting in unstable
24815               performance.
24816
24817       -masm=dialect
24818           Output assembly instructions using selected dialect.  Also affects
24819           which dialect is used for basic "asm" and extended "asm". Supported
24820           choices (in dialect order) are att or intel. The default is att.
24821           Darwin does not support intel.
24822
24823       -mieee-fp
24824       -mno-ieee-fp
24825           Control whether or not the compiler uses IEEE floating-point
24826           comparisons.  These correctly handle the case where the result of a
24827           comparison is unordered.
24828
24829       -m80387
24830       -mhard-float
24831           Generate output containing 80387 instructions for floating point.
24832
24833       -mno-80387
24834       -msoft-float
24835           Generate output containing library calls for floating point.
24836
24837           Warning: the requisite libraries are not part of GCC.  Normally the
24838           facilities of the machine's usual C compiler are used, but this
24839           cannot be done directly in cross-compilation.  You must make your
24840           own arrangements to provide suitable library functions for cross-
24841           compilation.
24842
24843           On machines where a function returns floating-point results in the
24844           80387 register stack, some floating-point opcodes may be emitted
24845           even if -msoft-float is used.
24846
24847       -mno-fp-ret-in-387
24848           Do not use the FPU registers for return values of functions.
24849
24850           The usual calling convention has functions return values of types
24851           "float" and "double" in an FPU register, even if there is no FPU.
24852           The idea is that the operating system should emulate an FPU.
24853
24854           The option -mno-fp-ret-in-387 causes such values to be returned in
24855           ordinary CPU registers instead.
24856
24857       -mno-fancy-math-387
24858           Some 387 emulators do not support the "sin", "cos" and "sqrt"
24859           instructions for the 387.  Specify this option to avoid generating
24860           those instructions.  This option is overridden when -march
24861           indicates that the target CPU always has an FPU and so the
24862           instruction does not need emulation.  These instructions are not
24863           generated unless you also use the -funsafe-math-optimizations
24864           switch.
24865
24866       -malign-double
24867       -mno-align-double
24868           Control whether GCC aligns "double", "long double", and "long long"
24869           variables on a two-word boundary or a one-word boundary.  Aligning
24870           "double" variables on a two-word boundary produces code that runs
24871           somewhat faster on a Pentium at the expense of more memory.
24872
24873           On x86-64, -malign-double is enabled by default.
24874
24875           Warning: if you use the -malign-double switch, structures
24876           containing the above types are aligned differently than the
24877           published application binary interface specifications for the
24878           x86-32 and are not binary compatible with structures in code
24879           compiled without that switch.
24880
24881       -m96bit-long-double
24882       -m128bit-long-double
24883           These switches control the size of "long double" type.  The x86-32
24884           application binary interface specifies the size to be 96 bits, so
24885           -m96bit-long-double is the default in 32-bit mode.
24886
24887           Modern architectures (Pentium and newer) prefer "long double" to be
24888           aligned to an 8- or 16-byte boundary.  In arrays or structures
24889           conforming to the ABI, this is not possible.  So specifying
24890           -m128bit-long-double aligns "long double" to a 16-byte boundary by
24891           padding the "long double" with an additional 32-bit zero.
24892
24893           In the x86-64 compiler, -m128bit-long-double is the default choice
24894           as its ABI specifies that "long double" is aligned on 16-byte
24895           boundary.
24896
24897           Notice that neither of these options enable any extra precision
24898           over the x87 standard of 80 bits for a "long double".
24899
24900           Warning: if you override the default value for your target ABI,
24901           this changes the size of structures and arrays containing "long
24902           double" variables, as well as modifying the function calling
24903           convention for functions taking "long double".  Hence they are not
24904           binary-compatible with code compiled without that switch.
24905
24906       -mlong-double-64
24907       -mlong-double-80
24908       -mlong-double-128
24909           These switches control the size of "long double" type. A size of 64
24910           bits makes the "long double" type equivalent to the "double" type.
24911           This is the default for 32-bit Bionic C library.  A size of 128
24912           bits makes the "long double" type equivalent to the "__float128"
24913           type. This is the default for 64-bit Bionic C library.
24914
24915           Warning: if you override the default value for your target ABI,
24916           this changes the size of structures and arrays containing "long
24917           double" variables, as well as modifying the function calling
24918           convention for functions taking "long double".  Hence they are not
24919           binary-compatible with code compiled without that switch.
24920
24921       -malign-data=type
24922           Control how GCC aligns variables.  Supported values for type are
24923           compat uses increased alignment value compatible uses GCC 4.8 and
24924           earlier, abi uses alignment value as specified by the psABI, and
24925           cacheline uses increased alignment value to match the cache line
24926           size.  compat is the default.
24927
24928       -mlarge-data-threshold=threshold
24929           When -mcmodel=medium is specified, data objects larger than
24930           threshold are placed in the large data section.  This value must be
24931           the same across all objects linked into the binary, and defaults to
24932           65535.
24933
24934       -mrtd
24935           Use a different function-calling convention, in which functions
24936           that take a fixed number of arguments return with the "ret num"
24937           instruction, which pops their arguments while returning.  This
24938           saves one instruction in the caller since there is no need to pop
24939           the arguments there.
24940
24941           You can specify that an individual function is called with this
24942           calling sequence with the function attribute "stdcall".  You can
24943           also override the -mrtd option by using the function attribute
24944           "cdecl".
24945
24946           Warning: this calling convention is incompatible with the one
24947           normally used on Unix, so you cannot use it if you need to call
24948           libraries compiled with the Unix compiler.
24949
24950           Also, you must provide function prototypes for all functions that
24951           take variable numbers of arguments (including "printf"); otherwise
24952           incorrect code is generated for calls to those functions.
24953
24954           In addition, seriously incorrect code results if you call a
24955           function with too many arguments.  (Normally, extra arguments are
24956           harmlessly ignored.)
24957
24958       -mregparm=num
24959           Control how many registers are used to pass integer arguments.  By
24960           default, no registers are used to pass arguments, and at most 3
24961           registers can be used.  You can control this behavior for a
24962           specific function by using the function attribute "regparm".
24963
24964           Warning: if you use this switch, and num is nonzero, then you must
24965           build all modules with the same value, including any libraries.
24966           This includes the system libraries and startup modules.
24967
24968       -msseregparm
24969           Use SSE register passing conventions for float and double arguments
24970           and return values.  You can control this behavior for a specific
24971           function by using the function attribute "sseregparm".
24972
24973           Warning: if you use this switch then you must build all modules
24974           with the same value, including any libraries.  This includes the
24975           system libraries and startup modules.
24976
24977       -mvect8-ret-in-mem
24978           Return 8-byte vectors in memory instead of MMX registers.  This is
24979           the default on VxWorks to match the ABI of the Sun Studio compilers
24980           until version 12.  Only use this option if you need to remain
24981           compatible with existing code produced by those previous compiler
24982           versions or older versions of GCC.
24983
24984       -mpc32
24985       -mpc64
24986       -mpc80
24987           Set 80387 floating-point precision to 32, 64 or 80 bits.  When
24988           -mpc32 is specified, the significands of results of floating-point
24989           operations are rounded to 24 bits (single precision); -mpc64 rounds
24990           the significands of results of floating-point operations to 53 bits
24991           (double precision) and -mpc80 rounds the significands of results of
24992           floating-point operations to 64 bits (extended double precision),
24993           which is the default.  When this option is used, floating-point
24994           operations in higher precisions are not available to the programmer
24995           without setting the FPU control word explicitly.
24996
24997           Setting the rounding of floating-point operations to less than the
24998           default 80 bits can speed some programs by 2% or more.  Note that
24999           some mathematical libraries assume that extended-precision (80-bit)
25000           floating-point operations are enabled by default; routines in such
25001           libraries could suffer significant loss of accuracy, typically
25002           through so-called "catastrophic cancellation", when this option is
25003           used to set the precision to less than extended precision.
25004
25005       -mstackrealign
25006           Realign the stack at entry.  On the x86, the -mstackrealign option
25007           generates an alternate prologue and epilogue that realigns the run-
25008           time stack if necessary.  This supports mixing legacy codes that
25009           keep 4-byte stack alignment with modern codes that keep 16-byte
25010           stack alignment for SSE compatibility.  See also the attribute
25011           "force_align_arg_pointer", applicable to individual functions.
25012
25013       -mpreferred-stack-boundary=num
25014           Attempt to keep the stack boundary aligned to a 2 raised to num
25015           byte boundary.  If -mpreferred-stack-boundary is not specified, the
25016           default is 4 (16 bytes or 128 bits).
25017
25018           Warning: When generating code for the x86-64 architecture with SSE
25019           extensions disabled, -mpreferred-stack-boundary=3 can be used to
25020           keep the stack boundary aligned to 8 byte boundary.  Since x86-64
25021           ABI require 16 byte stack alignment, this is ABI incompatible and
25022           intended to be used in controlled environment where stack space is
25023           important limitation.  This option leads to wrong code when
25024           functions compiled with 16 byte stack alignment (such as functions
25025           from a standard library) are called with misaligned stack.  In this
25026           case, SSE instructions may lead to misaligned memory access traps.
25027           In addition, variable arguments are handled incorrectly for 16 byte
25028           aligned objects (including x87 long double and __int128), leading
25029           to wrong results.  You must build all modules with
25030           -mpreferred-stack-boundary=3, including any libraries.  This
25031           includes the system libraries and startup modules.
25032
25033       -mincoming-stack-boundary=num
25034           Assume the incoming stack is aligned to a 2 raised to num byte
25035           boundary.  If -mincoming-stack-boundary is not specified, the one
25036           specified by -mpreferred-stack-boundary is used.
25037
25038           On Pentium and Pentium Pro, "double" and "long double" values
25039           should be aligned to an 8-byte boundary (see -malign-double) or
25040           suffer significant run time performance penalties.  On Pentium III,
25041           the Streaming SIMD Extension (SSE) data type "__m128" may not work
25042           properly if it is not 16-byte aligned.
25043
25044           To ensure proper alignment of this values on the stack, the stack
25045           boundary must be as aligned as that required by any value stored on
25046           the stack.  Further, every function must be generated such that it
25047           keeps the stack aligned.  Thus calling a function compiled with a
25048           higher preferred stack boundary from a function compiled with a
25049           lower preferred stack boundary most likely misaligns the stack.  It
25050           is recommended that libraries that use callbacks always use the
25051           default setting.
25052
25053           This extra alignment does consume extra stack space, and generally
25054           increases code size.  Code that is sensitive to stack space usage,
25055           such as embedded systems and operating system kernels, may want to
25056           reduce the preferred alignment to -mpreferred-stack-boundary=2.
25057
25058       -mmmx
25059       -msse
25060       -msse2
25061       -msse3
25062       -mssse3
25063       -msse4
25064       -msse4a
25065       -msse4.1
25066       -msse4.2
25067       -mavx
25068       -mavx2
25069       -mavx512f
25070       -mavx512pf
25071       -mavx512er
25072       -mavx512cd
25073       -mavx512vl
25074       -mavx512bw
25075       -mavx512dq
25076       -mavx512ifma
25077       -mavx512vbmi
25078       -msha
25079       -maes
25080       -mpclmul
25081       -mclflushopt
25082       -mclwb
25083       -mfsgsbase
25084       -mptwrite
25085       -mrdrnd
25086       -mf16c
25087       -mfma
25088       -mpconfig
25089       -mwbnoinvd
25090       -mfma4
25091       -mprfchw
25092       -mrdpid
25093       -mprefetchwt1
25094       -mrdseed
25095       -msgx
25096       -mxop
25097       -mlwp
25098       -m3dnow
25099       -m3dnowa
25100       -mpopcnt
25101       -mabm
25102       -madx
25103       -mbmi
25104       -mbmi2
25105       -mlzcnt
25106       -mfxsr
25107       -mxsave
25108       -mxsaveopt
25109       -mxsavec
25110       -mxsaves
25111       -mrtm
25112       -mhle
25113       -mtbm
25114       -mmwaitx
25115       -mclzero
25116       -mpku
25117       -mavx512vbmi2
25118       -mavx512bf16
25119       -mgfni
25120       -mvaes
25121       -mwaitpkg
25122       -mvpclmulqdq
25123       -mavx512bitalg
25124       -mmovdiri
25125       -mmovdir64b
25126       -menqcmd
25127       -mavx512vpopcntdq
25128       -mavx512vp2intersect
25129       -mavx5124fmaps
25130       -mavx512vnni
25131       -mavx5124vnniw
25132       -mcldemote
25133           These switches enable the use of instructions in the MMX, SSE,
25134           SSE2, SSE3, SSSE3, SSE4, SSE4A, SSE4.1, SSE4.2, AVX, AVX2, AVX512F,
25135           AVX512PF, AVX512ER, AVX512CD, AVX512VL, AVX512BW, AVX512DQ,
25136           AVX512IFMA, AVX512VBMI, SHA, AES, PCLMUL, CLFLUSHOPT, CLWB,
25137           FSGSBASE, PTWRITE, RDRND, F16C, FMA, PCONFIG, WBNOINVD, FMA4,
25138           PREFETCHW, RDPID, PREFETCHWT1, RDSEED, SGX, XOP, LWP, 3DNow!,
25139           enhanced 3DNow!, POPCNT, ABM, ADX, BMI, BMI2, LZCNT, FXSR, XSAVE,
25140           XSAVEOPT, XSAVEC, XSAVES, RTM, HLE, TBM, MWAITX, CLZERO, PKU,
25141           AVX512VBMI2, GFNI, VAES, WAITPKG, VPCLMULQDQ, AVX512BITALG,
25142           MOVDIRI, MOVDIR64B, AVX512BF16, ENQCMD, AVX512VPOPCNTDQ,
25143           AVX5124FMAPS, AVX512VNNI, AVX5124VNNIW, or CLDEMOTE extended
25144           instruction sets.  Each has a corresponding -mno- option to disable
25145           use of these instructions.
25146
25147           These extensions are also available as built-in functions: see x86
25148           Built-in Functions, for details of the functions enabled and
25149           disabled by these switches.
25150
25151           To generate SSE/SSE2 instructions automatically from floating-point
25152           code (as opposed to 387 instructions), see -mfpmath=sse.
25153
25154           GCC depresses SSEx instructions when -mavx is used. Instead, it
25155           generates new AVX instructions or AVX equivalence for all SSEx
25156           instructions when needed.
25157
25158           These options enable GCC to use these extended instructions in
25159           generated code, even without -mfpmath=sse.  Applications that
25160           perform run-time CPU detection must compile separate files for each
25161           supported architecture, using the appropriate flags.  In
25162           particular, the file containing the CPU detection code should be
25163           compiled without these options.
25164
25165       -mdump-tune-features
25166           This option instructs GCC to dump the names of the x86 performance
25167           tuning features and default settings. The names can be used in
25168           -mtune-ctrl=feature-list.
25169
25170       -mtune-ctrl=feature-list
25171           This option is used to do fine grain control of x86 code generation
25172           features.  feature-list is a comma separated list of feature names.
25173           See also -mdump-tune-features. When specified, the feature is
25174           turned on if it is not preceded with ^, otherwise, it is turned
25175           off.  -mtune-ctrl=feature-list is intended to be used by GCC
25176           developers. Using it may lead to code paths not covered by testing
25177           and can potentially result in compiler ICEs or runtime errors.
25178
25179       -mno-default
25180           This option instructs GCC to turn off all tunable features. See
25181           also -mtune-ctrl=feature-list and -mdump-tune-features.
25182
25183       -mcld
25184           This option instructs GCC to emit a "cld" instruction in the
25185           prologue of functions that use string instructions.  String
25186           instructions depend on the DF flag to select between autoincrement
25187           or autodecrement mode.  While the ABI specifies the DF flag to be
25188           cleared on function entry, some operating systems violate this
25189           specification by not clearing the DF flag in their exception
25190           dispatchers.  The exception handler can be invoked with the DF flag
25191           set, which leads to wrong direction mode when string instructions
25192           are used.  This option can be enabled by default on 32-bit x86
25193           targets by configuring GCC with the --enable-cld configure option.
25194           Generation of "cld" instructions can be suppressed with the
25195           -mno-cld compiler option in this case.
25196
25197       -mvzeroupper
25198           This option instructs GCC to emit a "vzeroupper" instruction before
25199           a transfer of control flow out of the function to minimize the AVX
25200           to SSE transition penalty as well as remove unnecessary "zeroupper"
25201           intrinsics.
25202
25203       -mprefer-avx128
25204           This option instructs GCC to use 128-bit AVX instructions instead
25205           of 256-bit AVX instructions in the auto-vectorizer.
25206
25207       -mprefer-vector-width=opt
25208           This option instructs GCC to use opt-bit vector width in
25209           instructions instead of default on the selected platform.
25210
25211           none
25212               No extra limitations applied to GCC other than defined by the
25213               selected platform.
25214
25215           128 Prefer 128-bit vector width for instructions.
25216
25217           256 Prefer 256-bit vector width for instructions.
25218
25219           512 Prefer 512-bit vector width for instructions.
25220
25221       -mcx16
25222           This option enables GCC to generate "CMPXCHG16B" instructions in
25223           64-bit code to implement compare-and-exchange operations on 16-byte
25224           aligned 128-bit objects.  This is useful for atomic updates of data
25225           structures exceeding one machine word in size.  The compiler uses
25226           this instruction to implement __sync Builtins.  However, for
25227           __atomic Builtins operating on 128-bit integers, a library call is
25228           always used.
25229
25230       -msahf
25231           This option enables generation of "SAHF" instructions in 64-bit
25232           code.  Early Intel Pentium 4 CPUs with Intel 64 support, prior to
25233           the introduction of Pentium 4 G1 step in December 2005, lacked the
25234           "LAHF" and "SAHF" instructions which are supported by AMD64.  These
25235           are load and store instructions, respectively, for certain status
25236           flags.  In 64-bit mode, the "SAHF" instruction is used to optimize
25237           "fmod", "drem", and "remainder" built-in functions; see Other
25238           Builtins for details.
25239
25240       -mmovbe
25241           This option enables use of the "movbe" instruction to implement
25242           "__builtin_bswap32" and "__builtin_bswap64".
25243
25244       -mshstk
25245           The -mshstk option enables shadow stack built-in functions from x86
25246           Control-flow Enforcement Technology (CET).
25247
25248       -mcrc32
25249           This option enables built-in functions "__builtin_ia32_crc32qi",
25250           "__builtin_ia32_crc32hi", "__builtin_ia32_crc32si" and
25251           "__builtin_ia32_crc32di" to generate the "crc32" machine
25252           instruction.
25253
25254       -mrecip
25255           This option enables use of "RCPSS" and "RSQRTSS" instructions (and
25256           their vectorized variants "RCPPS" and "RSQRTPS") with an additional
25257           Newton-Raphson step to increase precision instead of "DIVSS" and
25258           "SQRTSS" (and their vectorized variants) for single-precision
25259           floating-point arguments.  These instructions are generated only
25260           when -funsafe-math-optimizations is enabled together with
25261           -ffinite-math-only and -fno-trapping-math.  Note that while the
25262           throughput of the sequence is higher than the throughput of the
25263           non-reciprocal instruction, the precision of the sequence can be
25264           decreased by up to 2 ulp (i.e. the inverse of 1.0 equals
25265           0.99999994).
25266
25267           Note that GCC implements "1.0f/sqrtf(x)" in terms of "RSQRTSS" (or
25268           "RSQRTPS") already with -ffast-math (or the above option
25269           combination), and doesn't need -mrecip.
25270
25271           Also note that GCC emits the above sequence with additional Newton-
25272           Raphson step for vectorized single-float division and vectorized
25273           "sqrtf(x)" already with -ffast-math (or the above option
25274           combination), and doesn't need -mrecip.
25275
25276       -mrecip=opt
25277           This option controls which reciprocal estimate instructions may be
25278           used.  opt is a comma-separated list of options, which may be
25279           preceded by a ! to invert the option:
25280
25281           all Enable all estimate instructions.
25282
25283           default
25284               Enable the default instructions, equivalent to -mrecip.
25285
25286           none
25287               Disable all estimate instructions, equivalent to -mno-recip.
25288
25289           div Enable the approximation for scalar division.
25290
25291           vec-div
25292               Enable the approximation for vectorized division.
25293
25294           sqrt
25295               Enable the approximation for scalar square root.
25296
25297           vec-sqrt
25298               Enable the approximation for vectorized square root.
25299
25300           So, for example, -mrecip=all,!sqrt enables all of the reciprocal
25301           approximations, except for square root.
25302
25303       -mveclibabi=type
25304           Specifies the ABI type to use for vectorizing intrinsics using an
25305           external library.  Supported values for type are svml for the Intel
25306           short vector math library and acml for the AMD math core library.
25307           To use this option, both -ftree-vectorize and
25308           -funsafe-math-optimizations have to be enabled, and an SVML or ACML
25309           ABI-compatible library must be specified at link time.
25310
25311           GCC currently emits calls to "vmldExp2", "vmldLn2", "vmldLog102",
25312           "vmldPow2", "vmldTanh2", "vmldTan2", "vmldAtan2", "vmldAtanh2",
25313           "vmldCbrt2", "vmldSinh2", "vmldSin2", "vmldAsinh2", "vmldAsin2",
25314           "vmldCosh2", "vmldCos2", "vmldAcosh2", "vmldAcos2", "vmlsExp4",
25315           "vmlsLn4", "vmlsLog104", "vmlsPow4", "vmlsTanh4", "vmlsTan4",
25316           "vmlsAtan4", "vmlsAtanh4", "vmlsCbrt4", "vmlsSinh4", "vmlsSin4",
25317           "vmlsAsinh4", "vmlsAsin4", "vmlsCosh4", "vmlsCos4", "vmlsAcosh4"
25318           and "vmlsAcos4" for corresponding function type when
25319           -mveclibabi=svml is used, and "__vrd2_sin", "__vrd2_cos",
25320           "__vrd2_exp", "__vrd2_log", "__vrd2_log2", "__vrd2_log10",
25321           "__vrs4_sinf", "__vrs4_cosf", "__vrs4_expf", "__vrs4_logf",
25322           "__vrs4_log2f", "__vrs4_log10f" and "__vrs4_powf" for the
25323           corresponding function type when -mveclibabi=acml is used.
25324
25325       -mabi=name
25326           Generate code for the specified calling convention.  Permissible
25327           values are sysv for the ABI used on GNU/Linux and other systems,
25328           and ms for the Microsoft ABI.  The default is to use the Microsoft
25329           ABI when targeting Microsoft Windows and the SysV ABI on all other
25330           systems.  You can control this behavior for specific functions by
25331           using the function attributes "ms_abi" and "sysv_abi".
25332
25333       -mforce-indirect-call
25334           Force all calls to functions to be indirect. This is useful when
25335           using Intel Processor Trace where it generates more precise timing
25336           information for function calls.
25337
25338       -mmanual-endbr
25339           Insert ENDBR instruction at function entry only via the "cf_check"
25340           function attribute. This is useful when used with the option
25341           -fcf-protection=branch to control ENDBR insertion at the function
25342           entry.
25343
25344       -mcall-ms2sysv-xlogues
25345           Due to differences in 64-bit ABIs, any Microsoft ABI function that
25346           calls a System V ABI function must consider RSI, RDI and XMM6-15 as
25347           clobbered.  By default, the code for saving and restoring these
25348           registers is emitted inline, resulting in fairly lengthy prologues
25349           and epilogues.  Using -mcall-ms2sysv-xlogues emits prologues and
25350           epilogues that use stubs in the static portion of libgcc to perform
25351           these saves and restores, thus reducing function size at the cost
25352           of a few extra instructions.
25353
25354       -mtls-dialect=type
25355           Generate code to access thread-local storage using the gnu or gnu2
25356           conventions.  gnu is the conservative default; gnu2 is more
25357           efficient, but it may add compile- and run-time requirements that
25358           cannot be satisfied on all systems.
25359
25360       -mpush-args
25361       -mno-push-args
25362           Use PUSH operations to store outgoing parameters.  This method is
25363           shorter and usually equally fast as method using SUB/MOV operations
25364           and is enabled by default.  In some cases disabling it may improve
25365           performance because of improved scheduling and reduced
25366           dependencies.
25367
25368       -maccumulate-outgoing-args
25369           If enabled, the maximum amount of space required for outgoing
25370           arguments is computed in the function prologue.  This is faster on
25371           most modern CPUs because of reduced dependencies, improved
25372           scheduling and reduced stack usage when the preferred stack
25373           boundary is not equal to 2.  The drawback is a notable increase in
25374           code size.  This switch implies -mno-push-args.
25375
25376       -mthreads
25377           Support thread-safe exception handling on MinGW.  Programs that
25378           rely on thread-safe exception handling must compile and link all
25379           code with the -mthreads option.  When compiling, -mthreads defines
25380           -D_MT; when linking, it links in a special thread helper library
25381           -lmingwthrd which cleans up per-thread exception-handling data.
25382
25383       -mms-bitfields
25384       -mno-ms-bitfields
25385           Enable/disable bit-field layout compatible with the native
25386           Microsoft Windows compiler.
25387
25388           If "packed" is used on a structure, or if bit-fields are used, it
25389           may be that the Microsoft ABI lays out the structure differently
25390           than the way GCC normally does.  Particularly when moving packed
25391           data between functions compiled with GCC and the native Microsoft
25392           compiler (either via function call or as data in a file), it may be
25393           necessary to access either format.
25394
25395           This option is enabled by default for Microsoft Windows targets.
25396           This behavior can also be controlled locally by use of variable or
25397           type attributes.  For more information, see x86 Variable Attributes
25398           and x86 Type Attributes.
25399
25400           The Microsoft structure layout algorithm is fairly simple with the
25401           exception of the bit-field packing.  The padding and alignment of
25402           members of structures and whether a bit-field can straddle a
25403           storage-unit boundary are determine by these rules:
25404
25405           1. Structure members are stored sequentially in the order in which
25406           they are
25407               declared: the first member has the lowest memory address and
25408               the last member the highest.
25409
25410           2. Every data object has an alignment requirement.  The alignment
25411           requirement
25412               for all data except structures, unions, and arrays is either
25413               the size of the object or the current packing size (specified
25414               with either the "aligned" attribute or the "pack" pragma),
25415               whichever is less.  For structures, unions, and arrays, the
25416               alignment requirement is the largest alignment requirement of
25417               its members.  Every object is allocated an offset so that:
25418
25419                       offset % alignment_requirement == 0
25420
25421           3. Adjacent bit-fields are packed into the same 1-, 2-, or 4-byte
25422           allocation
25423               unit if the integral types are the same size and if the next
25424               bit-field fits into the current allocation unit without
25425               crossing the boundary imposed by the common alignment
25426               requirements of the bit-fields.
25427
25428           MSVC interprets zero-length bit-fields in the following ways:
25429
25430           1. If a zero-length bit-field is inserted between two bit-fields
25431           that
25432               are normally coalesced, the bit-fields are not coalesced.
25433
25434               For example:
25435
25436                       struct
25437                        {
25438                          unsigned long bf_1 : 12;
25439                          unsigned long : 0;
25440                          unsigned long bf_2 : 12;
25441                        } t1;
25442
25443               The size of "t1" is 8 bytes with the zero-length bit-field.  If
25444               the zero-length bit-field were removed, "t1"'s size would be 4
25445               bytes.
25446
25447           2. If a zero-length bit-field is inserted after a bit-field, "foo",
25448           and the
25449               alignment of the zero-length bit-field is greater than the
25450               member that follows it, "bar", "bar" is aligned as the type of
25451               the zero-length bit-field.
25452
25453               For example:
25454
25455                       struct
25456                        {
25457                          char foo : 4;
25458                          short : 0;
25459                          char bar;
25460                        } t2;
25461
25462                       struct
25463                        {
25464                          char foo : 4;
25465                          short : 0;
25466                          double bar;
25467                        } t3;
25468
25469               For "t2", "bar" is placed at offset 2, rather than offset 1.
25470               Accordingly, the size of "t2" is 4.  For "t3", the zero-length
25471               bit-field does not affect the alignment of "bar" or, as a
25472               result, the size of the structure.
25473
25474               Taking this into account, it is important to note the
25475               following:
25476
25477               1. If a zero-length bit-field follows a normal bit-field, the
25478               type of the
25479                   zero-length bit-field may affect the alignment of the
25480                   structure as whole. For example, "t2" has a size of 4
25481                   bytes, since the zero-length bit-field follows a normal
25482                   bit-field, and is of type short.
25483
25484               2. Even if a zero-length bit-field is not followed by a normal
25485               bit-field, it may
25486                   still affect the alignment of the structure:
25487
25488                           struct
25489                            {
25490                              char foo : 6;
25491                              long : 0;
25492                            } t4;
25493
25494                   Here, "t4" takes up 4 bytes.
25495
25496           3. Zero-length bit-fields following non-bit-field members are
25497           ignored:
25498                       struct
25499                        {
25500                          char foo;
25501                          long : 0;
25502                          char bar;
25503                        } t5;
25504
25505               Here, "t5" takes up 2 bytes.
25506
25507       -mno-align-stringops
25508           Do not align the destination of inlined string operations.  This
25509           switch reduces code size and improves performance in case the
25510           destination is already aligned, but GCC doesn't know about it.
25511
25512       -minline-all-stringops
25513           By default GCC inlines string operations only when the destination
25514           is known to be aligned to least a 4-byte boundary.  This enables
25515           more inlining and increases code size, but may improve performance
25516           of code that depends on fast "memcpy" and "memset" for short
25517           lengths.  The option enables inline expansion of "strlen" for all
25518           pointer alignments.
25519
25520       -minline-stringops-dynamically
25521           For string operations of unknown size, use run-time checks with
25522           inline code for small blocks and a library call for large blocks.
25523
25524       -mstringop-strategy=alg
25525           Override the internal decision heuristic for the particular
25526           algorithm to use for inlining string operations.  The allowed
25527           values for alg are:
25528
25529           rep_byte
25530           rep_4byte
25531           rep_8byte
25532               Expand using i386 "rep" prefix of the specified size.
25533
25534           byte_loop
25535           loop
25536           unrolled_loop
25537               Expand into an inline loop.
25538
25539           libcall
25540               Always use a library call.
25541
25542       -mmemcpy-strategy=strategy
25543           Override the internal decision heuristic to decide if
25544           "__builtin_memcpy" should be inlined and what inline algorithm to
25545           use when the expected size of the copy operation is known. strategy
25546           is a comma-separated list of alg:max_size:dest_align triplets.  alg
25547           is specified in -mstringop-strategy, max_size specifies the max
25548           byte size with which inline algorithm alg is allowed.  For the last
25549           triplet, the max_size must be "-1". The max_size of the triplets in
25550           the list must be specified in increasing order.  The minimal byte
25551           size for alg is 0 for the first triplet and "max_size + 1" of the
25552           preceding range.
25553
25554       -mmemset-strategy=strategy
25555           The option is similar to -mmemcpy-strategy= except that it is to
25556           control "__builtin_memset" expansion.
25557
25558       -momit-leaf-frame-pointer
25559           Don't keep the frame pointer in a register for leaf functions.
25560           This avoids the instructions to save, set up, and restore frame
25561           pointers and makes an extra register available in leaf functions.
25562           The option -fomit-leaf-frame-pointer removes the frame pointer for
25563           leaf functions, which might make debugging harder.
25564
25565       -mtls-direct-seg-refs
25566       -mno-tls-direct-seg-refs
25567           Controls whether TLS variables may be accessed with offsets from
25568           the TLS segment register (%gs for 32-bit, %fs for 64-bit), or
25569           whether the thread base pointer must be added.  Whether or not this
25570           is valid depends on the operating system, and whether it maps the
25571           segment to cover the entire TLS area.
25572
25573           For systems that use the GNU C Library, the default is on.
25574
25575       -msse2avx
25576       -mno-sse2avx
25577           Specify that the assembler should encode SSE instructions with VEX
25578           prefix.  The option -mavx turns this on by default.
25579
25580       -mfentry
25581       -mno-fentry
25582           If profiling is active (-pg), put the profiling counter call before
25583           the prologue.  Note: On x86 architectures the attribute
25584           "ms_hook_prologue" isn't possible at the moment for -mfentry and
25585           -pg.
25586
25587       -mrecord-mcount
25588       -mno-record-mcount
25589           If profiling is active (-pg), generate a __mcount_loc section that
25590           contains pointers to each profiling call. This is useful for
25591           automatically patching and out calls.
25592
25593       -mnop-mcount
25594       -mno-nop-mcount
25595           If profiling is active (-pg), generate the calls to the profiling
25596           functions as NOPs. This is useful when they should be patched in
25597           later dynamically. This is likely only useful together with
25598           -mrecord-mcount.
25599
25600       -minstrument-return=type
25601           Instrument function exit in -pg -mfentry instrumented functions
25602           with call to specified function. This only instruments true returns
25603           ending with ret, but not sibling calls ending with jump. Valid
25604           types are none to not instrument, call to generate a call to
25605           __return__, or nop5 to generate a 5 byte nop.
25606
25607       -mrecord-return
25608       -mno-record-return
25609           Generate a __return_loc section pointing to all return
25610           instrumentation code.
25611
25612       -mfentry-name=name
25613           Set name of __fentry__ symbol called at function entry for -pg
25614           -mfentry functions.
25615
25616       -mfentry-section=name
25617           Set name of section to record -mrecord-mcount calls (default
25618           __mcount_loc).
25619
25620       -mskip-rax-setup
25621       -mno-skip-rax-setup
25622           When generating code for the x86-64 architecture with SSE
25623           extensions disabled, -mskip-rax-setup can be used to skip setting
25624           up RAX register when there are no variable arguments passed in
25625           vector registers.
25626
25627           Warning: Since RAX register is used to avoid unnecessarily saving
25628           vector registers on stack when passing variable arguments, the
25629           impacts of this option are callees may waste some stack space,
25630           misbehave or jump to a random location.  GCC 4.4 or newer don't
25631           have those issues, regardless the RAX register value.
25632
25633       -m8bit-idiv
25634       -mno-8bit-idiv
25635           On some processors, like Intel Atom, 8-bit unsigned integer divide
25636           is much faster than 32-bit/64-bit integer divide.  This option
25637           generates a run-time check.  If both dividend and divisor are
25638           within range of 0 to 255, 8-bit unsigned integer divide is used
25639           instead of 32-bit/64-bit integer divide.
25640
25641       -mavx256-split-unaligned-load
25642       -mavx256-split-unaligned-store
25643           Split 32-byte AVX unaligned load and store.
25644
25645       -mstack-protector-guard=guard
25646       -mstack-protector-guard-reg=reg
25647       -mstack-protector-guard-offset=offset
25648           Generate stack protection code using canary at guard.  Supported
25649           locations are global for global canary or tls for per-thread canary
25650           in the TLS block (the default).  This option has effect only when
25651           -fstack-protector or -fstack-protector-all is specified.
25652
25653           With the latter choice the options -mstack-protector-guard-reg=reg
25654           and -mstack-protector-guard-offset=offset furthermore specify which
25655           segment register (%fs or %gs) to use as base register for reading
25656           the canary, and from what offset from that base register.  The
25657           default for those is as specified in the relevant ABI.
25658
25659       -mgeneral-regs-only
25660           Generate code that uses only the general-purpose registers.  This
25661           prevents the compiler from using floating-point, vector, mask and
25662           bound registers.
25663
25664       -mindirect-branch=choice
25665           Convert indirect call and jump with choice.  The default is keep,
25666           which keeps indirect call and jump unmodified.  thunk converts
25667           indirect call and jump to call and return thunk.  thunk-inline
25668           converts indirect call and jump to inlined call and return thunk.
25669           thunk-extern converts indirect call and jump to external call and
25670           return thunk provided in a separate object file.  You can control
25671           this behavior for a specific function by using the function
25672           attribute "indirect_branch".
25673
25674           Note that -mcmodel=large is incompatible with
25675           -mindirect-branch=thunk and -mindirect-branch=thunk-extern since
25676           the thunk function may not be reachable in the large code model.
25677
25678           Note that -mindirect-branch=thunk-extern is compatible with
25679           -fcf-protection=branch since the external thunk can be made to
25680           enable control-flow check.
25681
25682       -mfunction-return=choice
25683           Convert function return with choice.  The default is keep, which
25684           keeps function return unmodified.  thunk converts function return
25685           to call and return thunk.  thunk-inline converts function return to
25686           inlined call and return thunk.  thunk-extern converts function
25687           return to external call and return thunk provided in a separate
25688           object file.  You can control this behavior for a specific function
25689           by using the function attribute "function_return".
25690
25691           Note that -mindirect-return=thunk-extern is compatible with
25692           -fcf-protection=branch since the external thunk can be made to
25693           enable control-flow check.
25694
25695           Note that -mcmodel=large is incompatible with
25696           -mfunction-return=thunk and -mfunction-return=thunk-extern since
25697           the thunk function may not be reachable in the large code model.
25698
25699       -mindirect-branch-register
25700           Force indirect call and jump via register.
25701
25702       These -m switches are supported in addition to the above on x86-64
25703       processors in 64-bit environments.
25704
25705       -m32
25706       -m64
25707       -mx32
25708       -m16
25709       -miamcu
25710           Generate code for a 16-bit, 32-bit or 64-bit environment.  The -m32
25711           option sets "int", "long", and pointer types to 32 bits, and
25712           generates code that runs on any i386 system.
25713
25714           The -m64 option sets "int" to 32 bits and "long" and pointer types
25715           to 64 bits, and generates code for the x86-64 architecture.  For
25716           Darwin only the -m64 option also turns off the -fno-pic and
25717           -mdynamic-no-pic options.
25718
25719           The -mx32 option sets "int", "long", and pointer types to 32 bits,
25720           and generates code for the x86-64 architecture.
25721
25722           The -m16 option is the same as -m32, except for that it outputs the
25723           ".code16gcc" assembly directive at the beginning of the assembly
25724           output so that the binary can run in 16-bit mode.
25725
25726           The -miamcu option generates code which conforms to Intel MCU
25727           psABI.  It requires the -m32 option to be turned on.
25728
25729       -mno-red-zone
25730           Do not use a so-called "red zone" for x86-64 code.  The red zone is
25731           mandated by the x86-64 ABI; it is a 128-byte area beyond the
25732           location of the stack pointer that is not modified by signal or
25733           interrupt handlers and therefore can be used for temporary data
25734           without adjusting the stack pointer.  The flag -mno-red-zone
25735           disables this red zone.
25736
25737       -mcmodel=small
25738           Generate code for the small code model: the program and its symbols
25739           must be linked in the lower 2 GB of the address space.  Pointers
25740           are 64 bits.  Programs can be statically or dynamically linked.
25741           This is the default code model.
25742
25743       -mcmodel=kernel
25744           Generate code for the kernel code model.  The kernel runs in the
25745           negative 2 GB of the address space.  This model has to be used for
25746           Linux kernel code.
25747
25748       -mcmodel=medium
25749           Generate code for the medium model: the program is linked in the
25750           lower 2 GB of the address space.  Small symbols are also placed
25751           there.  Symbols with sizes larger than -mlarge-data-threshold are
25752           put into large data or BSS sections and can be located above 2GB.
25753           Programs can be statically or dynamically linked.
25754
25755       -mcmodel=large
25756           Generate code for the large model.  This model makes no assumptions
25757           about addresses and sizes of sections.
25758
25759       -maddress-mode=long
25760           Generate code for long address mode.  This is only supported for
25761           64-bit and x32 environments.  It is the default address mode for
25762           64-bit environments.
25763
25764       -maddress-mode=short
25765           Generate code for short address mode.  This is only supported for
25766           32-bit and x32 environments.  It is the default address mode for
25767           32-bit and x32 environments.
25768
25769       x86 Windows Options
25770
25771       These additional options are available for Microsoft Windows targets:
25772
25773       -mconsole
25774           This option specifies that a console application is to be
25775           generated, by instructing the linker to set the PE header subsystem
25776           type required for console applications.  This option is available
25777           for Cygwin and MinGW targets and is enabled by default on those
25778           targets.
25779
25780       -mdll
25781           This option is available for Cygwin and MinGW targets.  It
25782           specifies that a DLL---a dynamic link library---is to be generated,
25783           enabling the selection of the required runtime startup object and
25784           entry point.
25785
25786       -mnop-fun-dllimport
25787           This option is available for Cygwin and MinGW targets.  It
25788           specifies that the "dllimport" attribute should be ignored.
25789
25790       -mthread
25791           This option is available for MinGW targets. It specifies that
25792           MinGW-specific thread support is to be used.
25793
25794       -municode
25795           This option is available for MinGW-w64 targets.  It causes the
25796           "UNICODE" preprocessor macro to be predefined, and chooses Unicode-
25797           capable runtime startup code.
25798
25799       -mwin32
25800           This option is available for Cygwin and MinGW targets.  It
25801           specifies that the typical Microsoft Windows predefined macros are
25802           to be set in the pre-processor, but does not influence the choice
25803           of runtime library/startup code.
25804
25805       -mwindows
25806           This option is available for Cygwin and MinGW targets.  It
25807           specifies that a GUI application is to be generated by instructing
25808           the linker to set the PE header subsystem type appropriately.
25809
25810       -fno-set-stack-executable
25811           This option is available for MinGW targets. It specifies that the
25812           executable flag for the stack used by nested functions isn't set.
25813           This is necessary for binaries running in kernel mode of Microsoft
25814           Windows, as there the User32 API, which is used to set executable
25815           privileges, isn't available.
25816
25817       -fwritable-relocated-rdata
25818           This option is available for MinGW and Cygwin targets.  It
25819           specifies that relocated-data in read-only section is put into the
25820           ".data" section.  This is a necessary for older runtimes not
25821           supporting modification of ".rdata" sections for pseudo-relocation.
25822
25823       -mpe-aligned-commons
25824           This option is available for Cygwin and MinGW targets.  It
25825           specifies that the GNU extension to the PE file format that permits
25826           the correct alignment of COMMON variables should be used when
25827           generating code.  It is enabled by default if GCC detects that the
25828           target assembler found during configuration supports the feature.
25829
25830       See also under x86 Options for standard options.
25831
25832       Xstormy16 Options
25833
25834       These options are defined for Xstormy16:
25835
25836       -msim
25837           Choose startup files and linker script suitable for the simulator.
25838
25839       Xtensa Options
25840
25841       These options are supported for Xtensa targets:
25842
25843       -mconst16
25844       -mno-const16
25845           Enable or disable use of "CONST16" instructions for loading
25846           constant values.  The "CONST16" instruction is currently not a
25847           standard option from Tensilica.  When enabled, "CONST16"
25848           instructions are always used in place of the standard "L32R"
25849           instructions.  The use of "CONST16" is enabled by default only if
25850           the "L32R" instruction is not available.
25851
25852       -mfused-madd
25853       -mno-fused-madd
25854           Enable or disable use of fused multiply/add and multiply/subtract
25855           instructions in the floating-point option.  This has no effect if
25856           the floating-point option is not also enabled.  Disabling fused
25857           multiply/add and multiply/subtract instructions forces the compiler
25858           to use separate instructions for the multiply and add/subtract
25859           operations.  This may be desirable in some cases where strict IEEE
25860           754-compliant results are required: the fused multiply add/subtract
25861           instructions do not round the intermediate result, thereby
25862           producing results with more bits of precision than specified by the
25863           IEEE standard.  Disabling fused multiply add/subtract instructions
25864           also ensures that the program output is not sensitive to the
25865           compiler's ability to combine multiply and add/subtract operations.
25866
25867       -mserialize-volatile
25868       -mno-serialize-volatile
25869           When this option is enabled, GCC inserts "MEMW" instructions before
25870           "volatile" memory references to guarantee sequential consistency.
25871           The default is -mserialize-volatile.  Use -mno-serialize-volatile
25872           to omit the "MEMW" instructions.
25873
25874       -mforce-no-pic
25875           For targets, like GNU/Linux, where all user-mode Xtensa code must
25876           be position-independent code (PIC), this option disables PIC for
25877           compiling kernel code.
25878
25879       -mtext-section-literals
25880       -mno-text-section-literals
25881           These options control the treatment of literal pools.  The default
25882           is -mno-text-section-literals, which places literals in a separate
25883           section in the output file.  This allows the literal pool to be
25884           placed in a data RAM/ROM, and it also allows the linker to combine
25885           literal pools from separate object files to remove redundant
25886           literals and improve code size.  With -mtext-section-literals, the
25887           literals are interspersed in the text section in order to keep them
25888           as close as possible to their references.  This may be necessary
25889           for large assembly files.  Literals for each function are placed
25890           right before that function.
25891
25892       -mauto-litpools
25893       -mno-auto-litpools
25894           These options control the treatment of literal pools.  The default
25895           is -mno-auto-litpools, which places literals in a separate section
25896           in the output file unless -mtext-section-literals is used.  With
25897           -mauto-litpools the literals are interspersed in the text section
25898           by the assembler.  Compiler does not produce explicit ".literal"
25899           directives and loads literals into registers with "MOVI"
25900           instructions instead of "L32R" to let the assembler do relaxation
25901           and place literals as necessary.  This option allows assembler to
25902           create several literal pools per function and assemble very big
25903           functions, which may not be possible with -mtext-section-literals.
25904
25905       -mtarget-align
25906       -mno-target-align
25907           When this option is enabled, GCC instructs the assembler to
25908           automatically align instructions to reduce branch penalties at the
25909           expense of some code density.  The assembler attempts to widen
25910           density instructions to align branch targets and the instructions
25911           following call instructions.  If there are not enough preceding
25912           safe density instructions to align a target, no widening is
25913           performed.  The default is -mtarget-align.  These options do not
25914           affect the treatment of auto-aligned instructions like "LOOP",
25915           which the assembler always aligns, either by widening density
25916           instructions or by inserting NOP instructions.
25917
25918       -mlongcalls
25919       -mno-longcalls
25920           When this option is enabled, GCC instructs the assembler to
25921           translate direct calls to indirect calls unless it can determine
25922           that the target of a direct call is in the range allowed by the
25923           call instruction.  This translation typically occurs for calls to
25924           functions in other source files.  Specifically, the assembler
25925           translates a direct "CALL" instruction into an "L32R" followed by a
25926           "CALLX" instruction.  The default is -mno-longcalls.  This option
25927           should be used in programs where the call target can potentially be
25928           out of range.  This option is implemented in the assembler, not the
25929           compiler, so the assembly code generated by GCC still shows direct
25930           call instructions---look at the disassembled object code to see the
25931           actual instructions.  Note that the assembler uses an indirect call
25932           for every cross-file call, not just those that really are out of
25933           range.
25934
25935       zSeries Options
25936
25937       These are listed under
25938

ENVIRONMENT

25940       This section describes several environment variables that affect how
25941       GCC operates.  Some of them work by specifying directories or prefixes
25942       to use when searching for various kinds of files.  Some are used to
25943       specify other aspects of the compilation environment.
25944
25945       Note that you can also specify places to search using options such as
25946       -B, -I and -L.  These take precedence over places specified using
25947       environment variables, which in turn take precedence over those
25948       specified by the configuration of GCC.
25949
25950       LANG
25951       LC_CTYPE
25952       LC_MESSAGES
25953       LC_ALL
25954           These environment variables control the way that GCC uses
25955           localization information which allows GCC to work with different
25956           national conventions.  GCC inspects the locale categories LC_CTYPE
25957           and LC_MESSAGES if it has been configured to do so.  These locale
25958           categories can be set to any value supported by your installation.
25959           A typical value is en_GB.UTF-8 for English in the United Kingdom
25960           encoded in UTF-8.
25961
25962           The LC_CTYPE environment variable specifies character
25963           classification.  GCC uses it to determine the character boundaries
25964           in a string; this is needed for some multibyte encodings that
25965           contain quote and escape characters that are otherwise interpreted
25966           as a string end or escape.
25967
25968           The LC_MESSAGES environment variable specifies the language to use
25969           in diagnostic messages.
25970
25971           If the LC_ALL environment variable is set, it overrides the value
25972           of LC_CTYPE and LC_MESSAGES; otherwise, LC_CTYPE and LC_MESSAGES
25973           default to the value of the LANG environment variable.  If none of
25974           these variables are set, GCC defaults to traditional C English
25975           behavior.
25976
25977       TMPDIR
25978           If TMPDIR is set, it specifies the directory to use for temporary
25979           files.  GCC uses temporary files to hold the output of one stage of
25980           compilation which is to be used as input to the next stage: for
25981           example, the output of the preprocessor, which is the input to the
25982           compiler proper.
25983
25984       GCC_COMPARE_DEBUG
25985           Setting GCC_COMPARE_DEBUG is nearly equivalent to passing
25986           -fcompare-debug to the compiler driver.  See the documentation of
25987           this option for more details.
25988
25989       GCC_EXEC_PREFIX
25990           If GCC_EXEC_PREFIX is set, it specifies a prefix to use in the
25991           names of the subprograms executed by the compiler.  No slash is
25992           added when this prefix is combined with the name of a subprogram,
25993           but you can specify a prefix that ends with a slash if you wish.
25994
25995           If GCC_EXEC_PREFIX is not set, GCC attempts to figure out an
25996           appropriate prefix to use based on the pathname it is invoked with.
25997
25998           If GCC cannot find the subprogram using the specified prefix, it
25999           tries looking in the usual places for the subprogram.
26000
26001           The default value of GCC_EXEC_PREFIX is prefix/lib/gcc/ where
26002           prefix is the prefix to the installed compiler. In many cases
26003           prefix is the value of "prefix" when you ran the configure script.
26004
26005           Other prefixes specified with -B take precedence over this prefix.
26006
26007           This prefix is also used for finding files such as crt0.o that are
26008           used for linking.
26009
26010           In addition, the prefix is used in an unusual way in finding the
26011           directories to search for header files.  For each of the standard
26012           directories whose name normally begins with /usr/local/lib/gcc
26013           (more precisely, with the value of GCC_INCLUDE_DIR), GCC tries
26014           replacing that beginning with the specified prefix to produce an
26015           alternate directory name.  Thus, with -Bfoo/, GCC searches foo/bar
26016           just before it searches the standard directory /usr/local/lib/bar.
26017           If a standard directory begins with the configured prefix then the
26018           value of prefix is replaced by GCC_EXEC_PREFIX when looking for
26019           header files.
26020
26021       COMPILER_PATH
26022           The value of COMPILER_PATH is a colon-separated list of
26023           directories, much like PATH.  GCC tries the directories thus
26024           specified when searching for subprograms, if it cannot find the
26025           subprograms using GCC_EXEC_PREFIX.
26026
26027       LIBRARY_PATH
26028           The value of LIBRARY_PATH is a colon-separated list of directories,
26029           much like PATH.  When configured as a native compiler, GCC tries
26030           the directories thus specified when searching for special linker
26031           files, if it cannot find them using GCC_EXEC_PREFIX.  Linking using
26032           GCC also uses these directories when searching for ordinary
26033           libraries for the -l option (but directories specified with -L come
26034           first).
26035
26036       LANG
26037           This variable is used to pass locale information to the compiler.
26038           One way in which this information is used is to determine the
26039           character set to be used when character literals, string literals
26040           and comments are parsed in C and C++.  When the compiler is
26041           configured to allow multibyte characters, the following values for
26042           LANG are recognized:
26043
26044           C-JIS
26045               Recognize JIS characters.
26046
26047           C-SJIS
26048               Recognize SJIS characters.
26049
26050           C-EUCJP
26051               Recognize EUCJP characters.
26052
26053           If LANG is not defined, or if it has some other value, then the
26054           compiler uses "mblen" and "mbtowc" as defined by the default locale
26055           to recognize and translate multibyte characters.
26056
26057       Some additional environment variables affect the behavior of the
26058       preprocessor.
26059
26060       CPATH
26061       C_INCLUDE_PATH
26062       CPLUS_INCLUDE_PATH
26063       OBJC_INCLUDE_PATH
26064           Each variable's value is a list of directories separated by a
26065           special character, much like PATH, in which to look for header
26066           files.  The special character, "PATH_SEPARATOR", is target-
26067           dependent and determined at GCC build time.  For Microsoft Windows-
26068           based targets it is a semicolon, and for almost all other targets
26069           it is a colon.
26070
26071           CPATH specifies a list of directories to be searched as if
26072           specified with -I, but after any paths given with -I options on the
26073           command line.  This environment variable is used regardless of
26074           which language is being preprocessed.
26075
26076           The remaining environment variables apply only when preprocessing
26077           the particular language indicated.  Each specifies a list of
26078           directories to be searched as if specified with -isystem, but after
26079           any paths given with -isystem options on the command line.
26080
26081           In all these variables, an empty element instructs the compiler to
26082           search its current working directory.  Empty elements can appear at
26083           the beginning or end of a path.  For instance, if the value of
26084           CPATH is ":/special/include", that has the same effect as
26085           -I. -I/special/include.
26086
26087       DEPENDENCIES_OUTPUT
26088           If this variable is set, its value specifies how to output
26089           dependencies for Make based on the non-system header files
26090           processed by the compiler.  System header files are ignored in the
26091           dependency output.
26092
26093           The value of DEPENDENCIES_OUTPUT can be just a file name, in which
26094           case the Make rules are written to that file, guessing the target
26095           name from the source file name.  Or the value can have the form
26096           file target, in which case the rules are written to file file using
26097           target as the target name.
26098
26099           In other words, this environment variable is equivalent to
26100           combining the options -MM and -MF, with an optional -MT switch too.
26101
26102       SUNPRO_DEPENDENCIES
26103           This variable is the same as DEPENDENCIES_OUTPUT (see above),
26104           except that system header files are not ignored, so it implies -M
26105           rather than -MM.  However, the dependence on the main input file is
26106           omitted.
26107
26108       SOURCE_DATE_EPOCH
26109           If this variable is set, its value specifies a UNIX timestamp to be
26110           used in replacement of the current date and time in the "__DATE__"
26111           and "__TIME__" macros, so that the embedded timestamps become
26112           reproducible.
26113
26114           The value of SOURCE_DATE_EPOCH must be a UNIX timestamp, defined as
26115           the number of seconds (excluding leap seconds) since 01 Jan 1970
26116           00:00:00 represented in ASCII; identical to the output of
26117           @command{date +%s} on GNU/Linux and other systems that support the
26118           %s extension in the "date" command.
26119
26120           The value should be a known timestamp such as the last modification
26121           time of the source or package and it should be set by the build
26122           process.
26123

BUGS

26125       For instructions on reporting bugs, see
26126       <http://bugzilla.redhat.com/bugzilla>.
26127

FOOTNOTES

26129       1.  On some systems, gcc -shared needs to build supplementary stub code
26130           for constructors to work.  On multi-libbed systems, gcc -shared
26131           must select the correct support libraries to link against.  Failing
26132           to supply the correct flags may lead to subtle defects.  Supplying
26133           them in cases where they are not necessary is innocuous.
26134

SEE ALSO

26136       gpl(7), gfdl(7), fsf-funding(7), cpp(1), gcov(1), as(1), ld(1), gdb(1),
26137       dbx(1) and the Info entries for gcc, cpp, as, ld, binutils and gdb.
26138

AUTHOR

26140       See the Info entry for gcc, or
26141       <http://gcc.gnu.org/onlinedocs/gcc/Contributors.html>, for contributors
26142       to GCC.
26143
26145       Copyright (c) 1988-2020 Free Software Foundation, Inc.
26146
26147       Permission is granted to copy, distribute and/or modify this document
26148       under the terms of the GNU Free Documentation License, Version 1.3 or
26149       any later version published by the Free Software Foundation; with the
26150       Invariant Sections being "GNU General Public License" and "Funding Free
26151       Software", the Front-Cover texts being (a) (see below), and with the
26152       Back-Cover Texts being (b) (see below).  A copy of the license is
26153       included in the gfdl(7) man page.
26154
26155       (a) The FSF's Front-Cover Text is:
26156
26157            A GNU Manual
26158
26159       (b) The FSF's Back-Cover Text is:
26160
26161            You have freedom to copy and modify this GNU Manual, like GNU
26162            software.  Copies published by the Free Software Foundation raise
26163            funds for GNU development.
26164
26165
26166
26167gcc-10.2.1                        2020-07-23                            GCC(1)
Impressum