1OBJCOPY(1) GNU Development Tools OBJCOPY(1)
2
6 objcopy - copy and translate object files
7
9 objcopy [-F bfdname|--target=bfdname]
10 [-I bfdname|--input-target=bfdname]
11 [-O bfdname|--output-target=bfdname]
12 [-B bfdarch|--binary-architecture=bfdarch]
13 [-S|--strip-all]
14 [-g|--strip-debug]
15 [--strip-unneeded]
16 [-K symbolname|--keep-symbol=symbolname]
17 [-N symbolname|--strip-symbol=symbolname]
18 [--strip-unneeded-symbol=symbolname]
19 [-G symbolname|--keep-global-symbol=symbolname]
20 [--localize-hidden]
21 [-L symbolname|--localize-symbol=symbolname]
22 [--globalize-symbol=symbolname]
23 [--globalize-symbols=filename]
24 [-W symbolname|--weaken-symbol=symbolname]
25 [-w|--wildcard]
26 [-x|--discard-all]
27 [-X|--discard-locals]
28 [-b byte|--byte=byte]
29 [-i [breadth]|--interleave[=breadth]]
30 [--interleave-width=width]
31 [-j sectionpattern|--only-section=sectionpattern]
32 [-R sectionpattern|--remove-section=sectionpattern]
33 [--remove-relocations=sectionpattern]
34 [-p|--preserve-dates]
35 [-D|--enable-deterministic-archives]
36 [-U|--disable-deterministic-archives]
37 [--debugging]
38 [--gap-fill=val]
39 [--pad-to=address]
40 [--set-start=val]
41 [--adjust-start=incr]
42 [--change-addresses=incr]
43 [--change-section-address sectionpattern{=,+,-}val]
44 [--change-section-lma sectionpattern{=,+,-}val]
45 [--change-section-vma sectionpattern{=,+,-}val]
46 [--change-warnings] [--no-change-warnings]
47 [--set-section-flags sectionpattern=flags]
48 [--add-section sectionname=filename]
49 [--dump-section sectionname=filename]
50 [--update-section sectionname=filename]
51 [--rename-section oldname=newname[,flags]]
52 [--long-section-names {enable,disable,keep}]
53 [--change-leading-char] [--remove-leading-char]
54 [--reverse-bytes=num]
55 [--srec-len=ival] [--srec-forceS3]
56 [--redefine-sym old=new]
57 [--redefine-syms=filename]
58 [--weaken]
59 [--keep-symbols=filename]
60 [--strip-symbols=filename]
61 [--strip-unneeded-symbols=filename]
62 [--keep-global-symbols=filename]
63 [--localize-symbols=filename]
64 [--weaken-symbols=filename]
65 [--add-symbol name=[section:]value[,flags]]
66 [--alt-machine-code=index]
67 [--prefix-symbols=string]
68 [--prefix-sections=string]
69 [--prefix-alloc-sections=string]
70 [--add-gnu-debuglink=path-to-file]
71 [--keep-file-symbols]
72 [--only-keep-debug]
73 [--strip-dwo]
74 [--extract-dwo]
75 [--extract-symbol]
76 [--writable-text]
77 [--readonly-text]
78 [--pure]
79 [--impure]
80 [--file-alignment=num]
81 [--heap=size]
82 [--image-base=address]
83 [--section-alignment=num]
84 [--stack=size]
85 [--subsystem=which:major.minor]
86 [--compress-debug-sections]
87 [--decompress-debug-sections]
88 [--elf-stt-common=val]
89 [--merge-notes]
90 [--no-merge-notes]
91 [-v|--verbose]
92 [-V|--version]
93 [--help] [--info]
94 infile [outfile]
95
97 The GNU objcopy utility copies the contents of an object file to
98 another. objcopy uses the GNU BFD Library to read and write the object
99 files. It can write the destination object file in a format different
100 from that of the source object file. The exact behavior of objcopy is
101 controlled by command-line options. Note that objcopy should be able
102 to copy a fully linked file between any two formats. However, copying a
103 relocatable object file between any two formats may not work as
104 expected.
105 objcopy creates temporary files to do its translations and deletes them
107 afterward. objcopy uses BFD to do all its translation work; it has
108 access to all the formats described in BFD and thus is able to
109 recognize most formats without being told explicitly.
110 objcopy can be used to generate S-records by using an output target of
112 srec (e.g., use -O srec).
113 objcopy can be used to generate a raw binary file by using an output
115 target of binary (e.g., use -O binary). When objcopy generates a raw
116 binary file, it will essentially produce a memory dump of the contents
117 of the input object file. All symbols and relocation information will
118 be discarded. The memory dump will start at the load address of the
119 lowest section copied into the output file.
120 When generating an S-record or a raw binary file, it may be helpful to
122 use -S to remove sections containing debugging information. In some
123 cases -R will be useful to remove sections which contain information
124 that is not needed by the binary file.
125 Note---objcopy is not able to change the endianness of its input files.
127 If the input format has an endianness (some formats do not), objcopy
128 can only copy the inputs into file formats that have the same
129 endianness or which have no endianness (e.g., srec). (However, see the
130 --reverse-bytes option.)
131
133 infile
134 outfile
135 The input and output files, respectively. If you do not specify
136 outfile, objcopy creates a temporary file and destructively renames
137 the result with the name of infile.
138 -I bfdname
140 --input-target=bfdname
141 Consider the source file's object format to be bfdname, rather than
142 attempting to deduce it.
143 -O bfdname
145 --output-target=bfdname
146 Write the output file using the object format bfdname.
147 -F bfdname
149 --target=bfdname
150 Use bfdname as the object format for both the input and the output
151 file; i.e., simply transfer data from source to destination with no
152 translation.
153 -B bfdarch
155 --binary-architecture=bfdarch
156 Useful when transforming a architecture-less input file into an
157 object file. In this case the output architecture can be set to
158 bfdarch. This option will be ignored if the input file has a known
159 bfdarch. You can access this binary data inside a program by
160 referencing the special symbols that are created by the conversion
161 process. These symbols are called _binary_objfile_start,
162 _binary_objfile_end and _binary_objfile_size. e.g. you can
163 transform a picture file into an object file and then access it in
164 your code using these symbols.
165 -j sectionpattern
167 --only-section=sectionpattern
168 Copy only the indicated sections from the input file to the output
169 file. This option may be given more than once. Note that using
170 this option inappropriately may make the output file unusable.
171 Wildcard characters are accepted in sectionpattern.
172 If the first character of sectionpattern is the exclamation point
174 (!) then matching sections will not be copied, even if earlier use
175 of --only-section on the same command line would otherwise copy it.
176 For example:
177 --only-section=.text.* --only-section=!.text.foo
179 will copy all sectinos maching '.text.*' but not the section
181 '.text.foo'.
182 -R sectionpattern
184 --remove-section=sectionpattern
185 Remove any section matching sectionpattern from the output file.
186 This option may be given more than once. Note that using this
187 option inappropriately may make the output file unusable. Wildcard
188 characters are accepted in sectionpattern. Using both the -j and
189 -R options together results in undefined behaviour.
190 If the first character of sectionpattern is the exclamation point
192 (!) then matching sections will not be removed even if an earlier
193 use of --remove-section on the same command line would otherwise
194 remove it. For example:
195 --remove-section=.text.* --remove-section=!.text.foo
197 will remove all sections matching the pattern '.text.*', but will
199 not remove the section '.text.foo'.
200 --remove-relocations=sectionpattern
202 Remove non-dynamic relocations from the output file for any section
203 matching sectionpattern. This option may be given more than once.
204 Note that using this option inappropriately may make the output
205 file unusable, and attempting to remove a dynamic relocation
206 section such as .rela.plt from an executable or shared library with
207 --remove-relocations=.plt will not work. Wildcard characters are
208 accepted in sectionpattern. For example:
209 --remove-relocations=.text.*
211 will remove the relocations for all sections matching the pattern
213 '.text.*'.
214 If the first character of sectionpattern is the exclamation point
216 (!) then matching sections will not have their relocation removed
217 even if an earlier use of --remove-relocations on the same command
218 line would otherwise cause the relocations to be removed. For
219 example:
220 --remove-relocations=.text.* --remove-relocations=!.text.foo
222 will remove all relocations for sections matching the pattern
224 '.text.*', but will not remove relocations for the section
225 '.text.foo'.
226 -S
228 --strip-all
229 Do not copy relocation and symbol information from the source file.
230 -g
232 --strip-debug
233 Do not copy debugging symbols or sections from the source file.
234 --strip-unneeded
236 Strip all symbols that are not needed for relocation processing.
237 -K symbolname
239 --keep-symbol=symbolname
240 When stripping symbols, keep symbol symbolname even if it would
241 normally be stripped. This option may be given more than once.
242 -N symbolname
244 --strip-symbol=symbolname
245 Do not copy symbol symbolname from the source file. This option
246 may be given more than once.
247 --strip-unneeded-symbol=symbolname
249 Do not copy symbol symbolname from the source file unless it is
250 needed by a relocation. This option may be given more than once.
251 -G symbolname
253 --keep-global-symbol=symbolname
254 Keep only symbol symbolname global. Make all other symbols local
255 to the file, so that they are not visible externally. This option
256 may be given more than once. Note: this option cannot be used in
257 conjunction with the --globalize-symbol or --globalize-symbols
258 options.
259 --localize-hidden
261 In an ELF object, mark all symbols that have hidden or internal
262 visibility as local. This option applies on top of symbol-specific
263 localization options such as -L.
264 -L symbolname
266 --localize-symbol=symbolname
267 Convert a global or weak symbol called symbolname into a local
268 symbol, so that it is not visible externally. This option may be
269 given more than once. Note - unique symbols are not converted.
270 -W symbolname
272 --weaken-symbol=symbolname
273 Make symbol symbolname weak. This option may be given more than
274 once.
275 --globalize-symbol=symbolname
277 Give symbol symbolname global scoping so that it is visible outside
278 of the file in which it is defined. This option may be given more
279 than once. Note: this option cannot be used in conjunction with
280 the -G or --keep-global-symbol options.
281 -w
283 --wildcard
284 Permit regular expressions in symbolnames used in other command
285 line options. The question mark (?), asterisk (*), backslash (\)
286 and square brackets ([]) operators can be used anywhere in the
287 symbol name. If the first character of the symbol name is the
288 exclamation point (!) then the sense of the switch is reversed for
289 that symbol. For example:
290 -w -W !foo -W fo*
292 would cause objcopy to weaken all symbols that start with "fo"
294 except for the symbol "foo".
295 -x
297 --discard-all
298 Do not copy non-global symbols from the source file.
299 -X
301 --discard-locals
302 Do not copy compiler-generated local symbols. (These usually start
303 with L or ..)
304 -b byte
306 --byte=byte
307 If interleaving has been enabled via the --interleave option then
308 start the range of bytes to keep at the byteth byte. byte can be
309 in the range from 0 to breadth-1, where breadth is the value given
310 by the --interleave option.
311 -i [breadth]
313 --interleave[=breadth]
314 Only copy a range out of every breadth bytes. (Header data is not
315 affected). Select which byte in the range begins the copy with the
316 --byte option. Select the width of the range with the
317 --interleave-width option.
318 This option is useful for creating files to program ROM. It is
320 typically used with an "srec" output target. Note that objcopy
321 will complain if you do not specify the --byte option as well.
322 The default interleave breadth is 4, so with --byte set to 0,
324 objcopy would copy the first byte out of every four bytes from the
325 input to the output.
326 --interleave-width=width
328 When used with the --interleave option, copy width bytes at a time.
329 The start of the range of bytes to be copied is set by the --byte
330 option, and the extent of the range is set with the --interleave
331 option.
332 The default value for this option is 1. The value of width plus
334 the byte value set by the --byte option must not exceed the
335 interleave breadth set by the --interleave option.
336 This option can be used to create images for two 16-bit flashes
338 interleaved in a 32-bit bus by passing -b 0 -i 4
339 --interleave-width=2 and -b 2 -i 4 --interleave-width=2 to two
340 objcopy commands. If the input was '12345678' then the outputs
341 would be '1256' and '3478' respectively.
342 -p
344 --preserve-dates
345 Set the access and modification dates of the output file to be the
346 same as those of the input file.
347 -D
349 --enable-deterministic-archives
350 Operate in deterministic mode. When copying archive members and
351 writing the archive index, use zero for UIDs, GIDs, timestamps, and
352 use consistent file modes for all files.
353 If binutils was configured with --enable-deterministic-archives,
355 then this mode is on by default. It can be disabled with the -U
356 option, below.
357 -U
359 --disable-deterministic-archives
360 Do not operate in deterministic mode. This is the inverse of the
361 -D option, above: when copying archive members and writing the
362 archive index, use their actual UID, GID, timestamp, and file mode
363 values.
364 This is the default unless binutils was configured with
366 --enable-deterministic-archives.
367 --debugging
369 Convert debugging information, if possible. This is not the
370 default because only certain debugging formats are supported, and
371 the conversion process can be time consuming.
372 --gap-fill val
374 Fill gaps between sections with val. This operation applies to the
375 load address (LMA) of the sections. It is done by increasing the
376 size of the section with the lower address, and filling in the
377 extra space created with val.
378 --pad-to address
380 Pad the output file up to the load address address. This is done
381 by increasing the size of the last section. The extra space is
382 filled in with the value specified by --gap-fill (default zero).
383 --set-start val
385 Set the start address of the new file to val. Not all object file
386 formats support setting the start address.
387 --change-start incr
389 --adjust-start incr
390 Change the start address by adding incr. Not all object file
391 formats support setting the start address.
392 --change-addresses incr
394 --adjust-vma incr
395 Change the VMA and LMA addresses of all sections, as well as the
396 start address, by adding incr. Some object file formats do not
397 permit section addresses to be changed arbitrarily. Note that this
398 does not relocate the sections; if the program expects sections to
399 be loaded at a certain address, and this option is used to change
400 the sections such that they are loaded at a different address, the
401 program may fail.
402 --change-section-address sectionpattern{=,+,-}val
404 --adjust-section-vma sectionpattern{=,+,-}val
405 Set or change both the VMA address and the LMA address of any
406 section matching sectionpattern. If = is used, the section address
407 is set to val. Otherwise, val is added to or subtracted from the
408 section address. See the comments under --change-addresses, above.
409 If sectionpattern does not match any sections in the input file, a
410 warning will be issued, unless --no-change-warnings is used.
411 --change-section-lma sectionpattern{=,+,-}val
413 Set or change the LMA address of any sections matching
414 sectionpattern. The LMA address is the address where the section
415 will be loaded into memory at program load time. Normally this is
416 the same as the VMA address, which is the address of the section at
417 program run time, but on some systems, especially those where a
418 program is held in ROM, the two can be different. If = is used,
419 the section address is set to val. Otherwise, val is added to or
420 subtracted from the section address. See the comments under
421 --change-addresses, above. If sectionpattern does not match any
422 sections in the input file, a warning will be issued, unless
423 --no-change-warnings is used.
424 --change-section-vma sectionpattern{=,+,-}val
426 Set or change the VMA address of any section matching
427 sectionpattern. The VMA address is the address where the section
428 will be located once the program has started executing. Normally
429 this is the same as the LMA address, which is the address where the
430 section will be loaded into memory, but on some systems, especially
431 those where a program is held in ROM, the two can be different. If
432 = is used, the section address is set to val. Otherwise, val is
433 added to or subtracted from the section address. See the comments
434 under --change-addresses, above. If sectionpattern does not match
435 any sections in the input file, a warning will be issued, unless
436 --no-change-warnings is used.
437 --change-warnings
439 --adjust-warnings
440 If --change-section-address or --change-section-lma or
441 --change-section-vma is used, and the section pattern does not
442 match any sections, issue a warning. This is the default.
443 --no-change-warnings
445 --no-adjust-warnings
446 Do not issue a warning if --change-section-address or
447 --adjust-section-lma or --adjust-section-vma is used, even if the
448 section pattern does not match any sections.
449 --set-section-flags sectionpattern=flags
451 Set the flags for any sections matching sectionpattern. The flags
452 argument is a comma separated string of flag names. The recognized
453 names are alloc, contents, load, noload, readonly, code, data, rom,
454 share, and debug. You can set the contents flag for a section
455 which does not have contents, but it is not meaningful to clear the
456 contents flag of a section which does have contents--just remove
457 the section instead. Not all flags are meaningful for all object
458 file formats.
459 --add-section sectionname=filename
461 Add a new section named sectionname while copying the file. The
462 contents of the new section are taken from the file filename. The
463 size of the section will be the size of the file. This option only
464 works on file formats which can support sections with arbitrary
465 names. Note - it may be necessary to use the --set-section-flags
466 option to set the attributes of the newly created section.
467 --dump-section sectionname=filename
469 Place the contents of section named sectionname into the file
470 filename, overwriting any contents that may have been there
471 previously. This option is the inverse of --add-section. This
472 option is similar to the --only-section option except that it does
473 not create a formatted file, it just dumps the contents as raw
474 binary data, without applying any relocations. The option can be
475 specified more than once.
476 --update-section sectionname=filename
478 Replace the existing contents of a section named sectionname with
479 the contents of file filename. The size of the section will be
480 adjusted to the size of the file. The section flags for
481 sectionname will be unchanged. For ELF format files the section to
482 segment mapping will also remain unchanged, something which is not
483 possible using --remove-section followed by --add-section. The
484 option can be specified more than once.
485 Note - it is possible to use --rename-section and --update-section
487 to both update and rename a section from one command line. In this
488 case, pass the original section name to --update-section, and the
489 original and new section names to --rename-section.
490 --add-symbol name=[section:]value[,flags]
492 Add a new symbol named name while copying the file. This option
493 may be specified multiple times. If the section is given, the
494 symbol will be associated with and relative to that section,
495 otherwise it will be an ABS symbol. Specifying an undefined
496 section will result in a fatal error. There is no check for the
497 value, it will be taken as specified. Symbol flags can be
498 specified and not all flags will be meaningful for all object file
499 formats. By default, the symbol will be global. The special flag
500 'before=othersym' will insert the new symbol in front of the
501 specified othersym, otherwise the symbol(s) will be added at the
502 end of the symbol table in the order they appear.
503 --rename-section oldname=newname[,flags]
505 Rename a section from oldname to newname, optionally changing the
506 section's flags to flags in the process. This has the advantage
507 over using a linker script to perform the rename in that the output
508 stays as an object file and does not become a linked executable.
509 This option is particularly helpful when the input format is
511 binary, since this will always create a section called .data. If
512 for example, you wanted instead to create a section called .rodata
513 containing binary data you could use the following command line to
514 achieve it:
515 objcopy -I binary -O <output_format> -B <architecture> \
517 --rename-section .data=.rodata,alloc,load,readonly,data,contents \
518 <input_binary_file> <output_object_file>
519 --long-section-names {enable,disable,keep}
521 Controls the handling of long section names when processing "COFF"
522 and "PE-COFF" object formats. The default behaviour, keep, is to
523 preserve long section names if any are present in the input file.
524 The enable and disable options forcibly enable or disable the use
525 of long section names in the output object; when disable is in
526 effect, any long section names in the input object will be
527 truncated. The enable option will only emit long section names if
528 any are present in the inputs; this is mostly the same as keep, but
529 it is left undefined whether the enable option might force the
530 creation of an empty string table in the output file.
531 --change-leading-char
533 Some object file formats use special characters at the start of
534 symbols. The most common such character is underscore, which
535 compilers often add before every symbol. This option tells objcopy
536 to change the leading character of every symbol when it converts
537 between object file formats. If the object file formats use the
538 same leading character, this option has no effect. Otherwise, it
539 will add a character, or remove a character, or change a character,
540 as appropriate.
541 --remove-leading-char
543 If the first character of a global symbol is a special symbol
544 leading character used by the object file format, remove the
545 character. The most common symbol leading character is underscore.
546 This option will remove a leading underscore from all global
547 symbols. This can be useful if you want to link together objects
548 of different file formats with different conventions for symbol
549 names. This is different from --change-leading-char because it
550 always changes the symbol name when appropriate, regardless of the
551 object file format of the output file.
552 --reverse-bytes=num
554 Reverse the bytes in a section with output contents. A section
555 length must be evenly divisible by the value given in order for the
556 swap to be able to take place. Reversing takes place before the
557 interleaving is performed.
558 This option is used typically in generating ROM images for
560 problematic target systems. For example, on some target boards,
561 the 32-bit words fetched from 8-bit ROMs are re-assembled in
562 little-endian byte order regardless of the CPU byte order.
563 Depending on the programming model, the endianness of the ROM may
564 need to be modified.
565 Consider a simple file with a section containing the following
567 eight bytes: 12345678.
568 Using --reverse-bytes=2 for the above example, the bytes in the
570 output file would be ordered 21436587.
571 Using --reverse-bytes=4 for the above example, the bytes in the
573 output file would be ordered 43218765.
574 By using --reverse-bytes=2 for the above example, followed by
576 --reverse-bytes=4 on the output file, the bytes in the second
577 output file would be ordered 34127856.
578 --srec-len=ival
580 Meaningful only for srec output. Set the maximum length of the
581 Srecords being produced to ival. This length covers both address,
582 data and crc fields.
583 --srec-forceS3
585 Meaningful only for srec output. Avoid generation of S1/S2
586 records, creating S3-only record format.
587 --redefine-sym old=new
589 Change the name of a symbol old, to new. This can be useful when
590 one is trying link two things together for which you have no
591 source, and there are name collisions.
592 --redefine-syms=filename
594 Apply --redefine-sym to each symbol pair "old new" listed in the
595 file filename. filename is simply a flat file, with one symbol
596 pair per line. Line comments may be introduced by the hash
597 character. This option may be given more than once.
598 --weaken
600 Change all global symbols in the file to be weak. This can be
601 useful when building an object which will be linked against other
602 objects using the -R option to the linker. This option is only
603 effective when using an object file format which supports weak
604 symbols.
605 --keep-symbols=filename
607 Apply --keep-symbol option to each symbol listed in the file
608 filename. filename is simply a flat file, with one symbol name per
609 line. Line comments may be introduced by the hash character. This
610 option may be given more than once.
611 --strip-symbols=filename
613 Apply --strip-symbol option to each symbol listed in the file
614 filename. filename is simply a flat file, with one symbol name per
615 line. Line comments may be introduced by the hash character. This
616 option may be given more than once.
617 --strip-unneeded-symbols=filename
619 Apply --strip-unneeded-symbol option to each symbol listed in the
620 file filename. filename is simply a flat file, with one symbol
621 name per line. Line comments may be introduced by the hash
622 character. This option may be given more than once.
623 --keep-global-symbols=filename
625 Apply --keep-global-symbol option to each symbol listed in the file
626 filename. filename is simply a flat file, with one symbol name per
627 line. Line comments may be introduced by the hash character. This
628 option may be given more than once.
629 --localize-symbols=filename
631 Apply --localize-symbol option to each symbol listed in the file
632 filename. filename is simply a flat file, with one symbol name per
633 line. Line comments may be introduced by the hash character. This
634 option may be given more than once.
635 --globalize-symbols=filename
637 Apply --globalize-symbol option to each symbol listed in the file
638 filename. filename is simply a flat file, with one symbol name per
639 line. Line comments may be introduced by the hash character. This
640 option may be given more than once. Note: this option cannot be
641 used in conjunction with the -G or --keep-global-symbol options.
642 --weaken-symbols=filename
644 Apply --weaken-symbol option to each symbol listed in the file
645 filename. filename is simply a flat file, with one symbol name per
646 line. Line comments may be introduced by the hash character. This
647 option may be given more than once.
648 --alt-machine-code=index
650 If the output architecture has alternate machine codes, use the
651 indexth code instead of the default one. This is useful in case a
652 machine is assigned an official code and the tool-chain adopts the
653 new code, but other applications still depend on the original code
654 being used. For ELF based architectures if the index alternative
655 does not exist then the value is treated as an absolute number to
656 be stored in the e_machine field of the ELF header.
657 --writable-text
659 Mark the output text as writable. This option isn't meaningful for
660 all object file formats.
661 --readonly-text
663 Make the output text write protected. This option isn't meaningful
664 for all object file formats.
665 --pure
667 Mark the output file as demand paged. This option isn't meaningful
668 for all object file formats.
669 --impure
671 Mark the output file as impure. This option isn't meaningful for
672 all object file formats.
673 --prefix-symbols=string
675 Prefix all symbols in the output file with string.
676 --prefix-sections=string
678 Prefix all section names in the output file with string.
679 --prefix-alloc-sections=string
681 Prefix all the names of all allocated sections in the output file
682 with string.
683 --add-gnu-debuglink=path-to-file
685 Creates a .gnu_debuglink section which contains a reference to
686 path-to-file and adds it to the output file. Note: the file at
687 path-to-file must exist. Part of the process of adding the
688 .gnu_debuglink section involves embedding a checksum of the
689 contents of the debug info file into the section.
690 If the debug info file is built in one location but it is going to
692 be installed at a later time into a different location then do not
693 use the path to the installed location. The --add-gnu-debuglink
694 option will fail because the installed file does not exist yet.
695 Instead put the debug info file in the current directory and use
696 the --add-gnu-debuglink option without any directory components,
697 like this:
698 objcopy --add-gnu-debuglink=foo.debug
700 At debug time the debugger will attempt to look for the separate
702 debug info file in a set of known locations. The exact set of
703 these locations varies depending upon the distribution being used,
704 but it typically includes:
705 "* The same directory as the executable."
707 "* A sub-directory of the directory containing the executable"
708 called .debug
709 "* A global debug directory such as /usr/lib/debug."
711 As long as the debug info file has been installed into one of these
713 locations before the debugger is run everything should work
714 correctly.
715 --keep-file-symbols
717 When stripping a file, perhaps with --strip-debug or
718 --strip-unneeded, retain any symbols specifying source file names,
719 which would otherwise get stripped.
720 --only-keep-debug
722 Strip a file, removing contents of any sections that would not be
723 stripped by --strip-debug and leaving the debugging sections
724 intact. In ELF files, this preserves all note sections in the
725 output.
726 Note - the section headers of the stripped sections are preserved,
728 including their sizes, but the contents of the section are
729 discarded. The section headers are preserved so that other tools
730 can match up the debuginfo file with the real executable, even if
731 that executable has been relocated to a different address space.
732 The intention is that this option will be used in conjunction with
734 --add-gnu-debuglink to create a two part executable. One a
735 stripped binary which will occupy less space in RAM and in a
736 distribution and the second a debugging information file which is
737 only needed if debugging abilities are required. The suggested
738 procedure to create these files is as follows:
739 1.<Link the executable as normal. Assuming that it is called>
741 "foo" then...
742 1.<Run "objcopy --only-keep-debug foo foo.dbg" to>
744 create a file containing the debugging info.
745 1.<Run "objcopy --strip-debug foo" to create a>
747 stripped executable.
748 1.<Run "objcopy --add-gnu-debuglink=foo.dbg foo">
750 to add a link to the debugging info into the stripped
751 executable.
752 Note---the choice of ".dbg" as an extension for the debug info file
754 is arbitrary. Also the "--only-keep-debug" step is optional. You
755 could instead do this:
756 1.<Link the executable as normal.>
758 1.<Copy "foo" to "foo.full">
759 1.<Run "objcopy --strip-debug foo">
760 1.<Run "objcopy --add-gnu-debuglink=foo.full foo">
761 i.e., the file pointed to by the --add-gnu-debuglink can be the
763 full executable. It does not have to be a file created by the
764 --only-keep-debug switch.
765 Note---this switch is only intended for use on fully linked files.
767 It does not make sense to use it on object files where the
768 debugging information may be incomplete. Besides the gnu_debuglink
769 feature currently only supports the presence of one filename
770 containing debugging information, not multiple filenames on a one-
771 per-object-file basis.
772 --strip-dwo
774 Remove the contents of all DWARF .dwo sections, leaving the
775 remaining debugging sections and all symbols intact. This option
776 is intended for use by the compiler as part of the -gsplit-dwarf
777 option, which splits debug information between the .o file and a
778 separate .dwo file. The compiler generates all debug information
779 in the same file, then uses the --extract-dwo option to copy the
780 .dwo sections to the .dwo file, then the --strip-dwo option to
781 remove those sections from the original .o file.
782 --extract-dwo
784 Extract the contents of all DWARF .dwo sections. See the
785 --strip-dwo option for more information.
786 --file-alignment num
788 Specify the file alignment. Sections in the file will always begin
789 at file offsets which are multiples of this number. This defaults
790 to 512. [This option is specific to PE targets.]
791 --heap reserve
793 --heap reserve,commit
794 Specify the number of bytes of memory to reserve (and optionally
795 commit) to be used as heap for this program. [This option is
796 specific to PE targets.]
797 --image-base value
799 Use value as the base address of your program or dll. This is the
800 lowest memory location that will be used when your program or dll
801 is loaded. To reduce the need to relocate and improve performance
802 of your dlls, each should have a unique base address and not
803 overlap any other dlls. The default is 0x400000 for executables,
804 and 0x10000000 for dlls. [This option is specific to PE targets.]
805 --section-alignment num
807 Sets the section alignment. Sections in memory will always begin
808 at addresses which are a multiple of this number. Defaults to
809 0x1000. [This option is specific to PE targets.]
810 --stack reserve
812 --stack reserve,commit
813 Specify the number of bytes of memory to reserve (and optionally
814 commit) to be used as stack for this program. [This option is
815 specific to PE targets.]
816 --subsystem which
818 --subsystem which:major
819 --subsystem which:major.minor
820 Specifies the subsystem under which your program will execute. The
821 legal values for which are "native", "windows", "console", "posix",
822 "efi-app", "efi-bsd", "efi-rtd", "sal-rtd", and "xbox". You may
823 optionally set the subsystem version also. Numeric values are also
824 accepted for which. [This option is specific to PE targets.]
825 --extract-symbol
827 Keep the file's section flags and symbols but remove all section
828 data. Specifically, the option:
829 *<removes the contents of all sections;>
831 *<sets the size of every section to zero; and>
832 *<sets the file's start address to zero.>
833 This option is used to build a .sym file for a VxWorks kernel. It
835 can also be a useful way of reducing the size of a --just-symbols
836 linker input file.
837 --compress-debug-sections
839 Compress DWARF debug sections using zlib with SHF_COMPRESSED from
840 the ELF ABI. Note - if compression would actually make a section
841 larger, then it is not compressed.
842 --compress-debug-sections=none
844 --compress-debug-sections=zlib
845 --compress-debug-sections=zlib-gnu
846 --compress-debug-sections=zlib-gabi
847 For ELF files, these options control how DWARF debug sections are
848 compressed. --compress-debug-sections=none is equivalent to
849 --decompress-debug-sections. --compress-debug-sections=zlib and
850 --compress-debug-sections=zlib-gabi are equivalent to
851 --compress-debug-sections. --compress-debug-sections=zlib-gnu
852 compresses DWARF debug sections using zlib. The debug sections are
853 renamed to begin with .zdebug instead of .debug. Note - if
854 compression would actually make a section larger, then it is not
855 compressed nor renamed.
856 --decompress-debug-sections
858 Decompress DWARF debug sections using zlib. The original section
859 names of the compressed sections are restored.
860 --elf-stt-common=yes
862 --elf-stt-common=no
863 For ELF files, these options control whether common symbols should
864 be converted to the "STT_COMMON" or "STT_OBJECT" type.
865 --elf-stt-common=yes converts common symbol type to "STT_COMMON".
866 --elf-stt-common=no converts common symbol type to "STT_OBJECT".
867 --merge-notes
869 --no-merge-notes
870 For ELF files, attempt (or do not attempt) to reduce the size of
871 any SHT_NOTE type sections by removing duplicate notes.
872 -V
874 --version
875 Show the version number of objcopy.
876 -v
878 --verbose
879 Verbose output: list all object files modified. In the case of
880 archives, objcopy -V lists all members of the archive.
881 --help
883 Show a summary of the options to objcopy.
884 --info
886 Display a list showing all architectures and object formats
887 available.
888 @file
890 Read command-line options from file. The options read are inserted
891 in place of the original @file option. If file does not exist, or
892 cannot be read, then the option will be treated literally, and not
893 removed.
894 Options in file are separated by whitespace. A whitespace
896 character may be included in an option by surrounding the entire
897 option in either single or double quotes. Any character (including
898 a backslash) may be included by prefixing the character to be
899 included with a backslash. The file may itself contain additional
900 @file options; any such options will be processed recursively.
901
903 ld(1), objdump(1), and the Info entries for binutils.
904
906 Copyright (c) 1991-2019 Free Software Foundation, Inc.
907 Permission is granted to copy, distribute and/or modify this document
909 under the terms of the GNU Free Documentation License, Version 1.3 or
910 any later version published by the Free Software Foundation; with no
911 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
912 Texts. A copy of the license is included in the section entitled "GNU
913 Free Documentation License".
914binutils-2.31.90 2019-01-19 OBJCOPY(1)