1LLVM-PDBUTIL(1) LLVM LLVM-PDBUTIL(1)
2
6 llvm-pdbutil - PDB File forensics and diagnostics
7 • Synopsis
9 • Description
11 • Subcommands
13 • pretty
15 • Summary
17 • Options
19 • Filtering and Sorting Options
21 • Symbol Type Options
23 • Other Options
25 • dump
27 • Summary
29 • Options
31 • MSF Container Options
33 • Module & File Options
35 • Symbol Options
37 • Type Record Options
39 • Miscellaneous Options
41 • bytes
43 • Summary
45 • Options
47 • MSF File Options
49 • PDB Stream Options
51 • DBI Stream Options
53 • Module Options
55 • Type Record Options
57 • pdb2yaml
59 • Summary
61 • Options
63 • yaml2pdb
65 • Summary
67 • Options
69 • merge
71 • Summary
73 • Options
75
77 llvm-pdbutil [subcommand] [options]
78
80 Display types, symbols, CodeView records, and other information from a
81 PDB file, as well as manipulate and create PDB files. llvm-pdbutil is
82 normally used by FileCheck-based tests to test LLVM's PDB reading and
83 writing functionality, but can also be used for general PDB file inves‐
84 tigation and forensics, or as a replacement for cvdump.
85
87 llvm-pdbutil is separated into several subcommands each tailored to a
88 different purpose. A brief summary of each command follows, with more
89 detail in the sections that follow.
90 • pretty - Dump symbol and type information in a format that tries
92 to look as much like the original source code as possible.
93 • dump - Dump low level types and structures from the PDB file, in‐
95 cluding CodeView records, hash tables, PDB streams, etc.
96 • bytes - Dump data from the PDB file's streams, records, types,
98 symbols, etc as raw bytes.
99 • yaml2pdb - Given a yaml description of a PDB file, produce a valid
101 PDB file that matches that description.
102 • pdb2yaml - For a given PDB file, produce a YAML description of
104 some or all of the file in a way that the PDB can be recon‐
105 structed.
106 • merge - Given two PDBs, produce a third PDB that is the result of
108 merging the two input PDBs.
109 pretty
111 IMPORTANT:
112 The pretty subcommand is built on the Windows DIA SDK, and as such
113 is not supported on non-Windows platforms.
114 USAGE: llvm-pdbutil pretty [options] <input PDB file>
116 Summary
118 The pretty subcommand displays a very high level representation of your
119 program's debug info. Since it is built on the Windows DIA SDK which
120 is the standard API that Windows tools and debuggers query debug infor‐
121 mation, it presents a more authoritative view of how a debugger is go‐
122 ing to interpret your debug information than a mode which displays
123 low-level CodeView records.
124 Options
126 Filtering and Sorting Options
127 NOTE:
128 exclude filters take priority over include filters. So if a filter
129 matches both an include and an exclude rule, then it is excluded.
130 -exclude-compilands=<string>
132 When dumping compilands, compiland source-file contributions, or
133 per-compiland symbols, this option instructs llvm-pdbutil to
134 omit any compilands that match the specified regular expression.
135 -exclude-symbols=<string>
137 When dumping global, public, or per-compiland symbols, this op‐
138 tion instructs llvm-pdbutil to omit any symbols that match the
139 specified regular expression.
140 -exclude-types=<string>
142 When dumping types, this option instructs llvm-pdbutil to omit
143 any types that match the specified regular expression.
144 -include-compilands=<string>
146 When dumping compilands, compiland source-file contributions, or
147 per-compiland symbols, limit the initial search to only those
148 compilands that match the specified regular expression.
149 -include-symbols=<string>
151 When dumping global, public, or per-compiland symbols, limit the
152 initial search to only those symbols that match the specified
153 regular expression.
154 -include-types=<string>
156 When dumping types, limit the initial search to only those types
157 that match the specified regular expression.
158 -min-class-padding=<uint>
160 Only display types that have at least the specified amount of
161 alignment padding, accounting for padding in base classes and
162 aggregate field members.
163 -min-class-padding-imm=<uint>
165 Only display types that have at least the specified amount of
166 alignment padding, ignoring padding in base classes and aggre‐
167 gate field members.
168 -min-type-size=<uint>
170 Only display types T where sizeof(T) is greater than or equal to
171 the specified amount.
172 -no-compiler-generated
174 Don't show compiler generated types and symbols
175 -no-enum-definitions
177 When dumping an enum, don't show the full enum (e.g. the indi‐
178 vidual enumerator values).
179 -no-system-libs
181 Don't show symbols from system libraries
182 Symbol Type Options
184 -all Implies all other options in this category.
185 -class-definitions=<format>
187 Displays class definitions in the specified format.
188 =all - Display all class members including data, constants, typedefs, functions, etc (default)
190 =layout - Only display members that contribute to class size.
191 =none - Don't display class definitions (e.g. only display the name and base list)
192 -class-order
194 Displays classes in the specified order.
195 =none - Undefined / no particular sort order (default)
197 =name - Sort classes by name
198 =size - Sort classes by size
199 =padding - Sort classes by amount of padding
200 =padding-pct - Sort classes by percentage of space consumed by padding
201 =padding-imm - Sort classes by amount of immediate padding
202 =padding-pct-imm - Sort classes by percentage of space consumed by immediate padding
203 -class-recurse-depth=<uint>
205 When dumping class definitions, stop after recursing the speci‐
206 fied number of times. The default is 0, which is no limit.
207 -classes
209 Display classes
210 -compilands
212 Display compilands (e.g. object files)
213 -enums Display enums
215 -externals
217 Dump external (e.g. exported) symbols
218 -globals
220 Dump global symbols
221 -lines Dump the mappings between source lines and code addresses.
223 -module-syms
225 Display symbols (variables, functions, etc) for each compiland
226 -sym-types=<types>
228 Type of symbols to dump when -globals, -externals, or -mod‐
229 ule-syms is specified. (default all)
230 =thunks - Display thunk symbols
232 =data - Display data symbols
233 =funcs - Display function symbols
234 =all - Display all symbols (default)
235 -symbol-order=<order>
237 For symbols dumped via the -module-syms, -globals, or -externals
238 options, sort the results in specified order.
239 =none - Undefined / no particular sort order
241 =name - Sort symbols by name
242 =size - Sort symbols by size
243 -typedefs
245 Display typedef types
246 -types Display all types (implies -classes, -enums, -typedefs)
248 Other Options
250 -color-output
251 Force color output on or off. By default, color if used if out‐
252 putting to a terminal.
253 -load-address=<uint>
255 When displaying relative virtual addresses, assume the process
256 is loaded at the given address and display what would be the ab‐
257 solute address.
258 dump
260 USAGE: llvm-pdbutil dump [options] <input PDB file>
261 Summary
263 The dump subcommand displays low level information about the structure
264 of a PDB file. It is used heavily by LLVM's testing infrastructure,
265 but can also be used for PDB forensics. It serves a role similar to
266 that of Microsoft's cvdump tool.
267 NOTE:
269 The dump subcommand exposes internal details of the file format. As
270 such, the reader should be familiar with /PDB/index before using
271 this command.
272 Options
274 MSF Container Options
275 -streams
276 dump a summary of all of the streams in the PDB file.
277 -stream-blocks
279 In conjunction with -streams, add information to the output
280 about what blocks the specified stream occupies.
281 -summary
283 Dump MSF and PDB header information.
284 Module & File Options
286 -modi=<uint>
287 For all options that dump information from each module/compi‐
288 land, limit to the specified module.
289 -files Dump the source files that contribute to each displayed module.
291 -il Dump inlinee line information (DEBUG_S_INLINEELINES CodeView
293 subsection)
294 -l Dump line information (DEBUG_S_LINES CodeView subsection)
296 -modules
298 Dump compiland information
299 -xme Dump cross module exports (DEBUG_S_CROSSSCOPEEXPORTS CodeView
301 subsection)
302 -xmi Dump cross module imports (DEBUG_S_CROSSSCOPEIMPORTS CodeView
304 subsection)
305 Symbol Options
307 -globals
308 dump global symbol records
309 -global-extras
311 dump additional information about the globals, such as hash
312 buckets and hash values.
313 -publics
315 dump public symbol records
316 -public-extras
318 dump additional information about the publics, such as hash
319 buckets and hash values.
320 -symbols
322 dump symbols (functions, variables, etc) for each module dumped.
323 -sym-data
325 For each symbol record dumped as a result of the -symbols op‐
326 tion, display the full bytes of the record in binary as well.
327 Type Record Options
329 -types Dump CodeView type records from TPI stream
330 -type-extras
332 Dump additional information from the TPI stream, such as hashes
333 and the type index offsets array.
334 -type-data
336 For each type record dumped, display the full bytes of the
337 record in binary as well.
338 -type-index=<uint>
340 Only dump types with the specified type index.
341 -ids Dump CodeView type records from IPI stream.
343 -id-extras
345 Dump additional information from the IPI stream, such as hashes
346 and the type index offsets array.
347 -id-data
349 For each ID record dumped, display the full bytes of the record
350 in binary as well.
351 -id-index=<uint>
353 only dump ID records with the specified hexadecimal type index.
354 -dependents
356 When used in conjunction with -type-index or -id-index, dumps
357 the entire dependency graph for the specified index instead of
358 just the single record with the specified index. For example,
359 if type index 0x4000 is a function whose return type has index
360 0x3000, and you specify -dependents=0x4000, then this would dump
361 both records (as well as any other dependents in the tree).
362 Miscellaneous Options
364 -all Implies most other options.
365 -section-contribs
367 Dump section contributions.
368 -section-headers
370 Dump image section headers.
371 -section-map
373 Dump section map.
374 -string-table
376 Dump PDB string table.
377 bytes
379 USAGE: llvm-pdbutil bytes [options] <input PDB file>
380 Summary
382 Like the dump subcommand, the bytes subcommand displays low level in‐
383 formation about the structure of a PDB file, but it is used for even
384 deeper forensics. The bytes subcommand finds various structures in a
385 PDB file based on the command line options specified, and dumps them in
386 hex. Someone working on support for emitting PDBs would use this heav‐
387 ily, for example, to compare one PDB against another PDB to ensure
388 byte-for-byte compatibility. It is not enough to simply compare the
389 bytes of an entire file, or an entire stream because it's perfectly
390 fine for the same structure to exist at different locations in two dif‐
391 ferent PDBs, and "finding" the structure is half the battle.
392 Options
394 MSF File Options
395 -block-range=<start[-end]>
396 Dump binary data from specified range of MSF file blocks.
397 -byte-range=<start[-end]>
399 Dump binary data from specified range of bytes in the file.
400 -fpm Dump the MSF free page map.
402 -stream-data=<string>
404 Dump binary data from the specified streams. Format is
405 SN[:Start][@Size]. For example, -stream-data=7:3@12 dumps 12
406 bytes from stream 7, starting at offset 3 in the stream.
407 PDB Stream Options
409 -name-map
410 Dump bytes of PDB Name Map
411 DBI Stream Options
413 -ec Dump the edit and continue map substream of the DBI stream.
414 -files Dump the file info substream of the DBI stream.
416 -modi Dump the modi substream of the DBI stream.
418 -sc Dump section contributions substream of the DBI stream.
420 -sm Dump the section map from the DBI stream.
422 -type-server
424 Dump the type server map from the DBI stream.
425 Module Options
427 -mod=<uint>
428 Limit all options in this category to the specified module in‐
429 dex. By default, options in this category will dump bytes from
430 all modules.
431 -chunks
433 Dump the bytes of each module's C13 debug subsection.
434 -split-chunks
436 When specified with -chunks, split the C13 debug subsection into
437 a separate chunk for each subsection type, and dump them sepa‐
438 rately.
439 -syms Dump the symbol record substream from each module.
441 Type Record Options
443 -id=<uint>
444 Dump the record from the IPI stream with the given type index.
445 -type=<uint>
447 Dump the record from the TPI stream with the given type index.
448 pdb2yaml
450 USAGE: llvm-pdbutil pdb2yaml [options] <input PDB file>
451 Summary
453 Options
454 yaml2pdb
455 USAGE: llvm-pdbutil yaml2pdb [options] <input YAML file>
456 Summary
458 Generate a PDB file from a YAML description. The YAML syntax is not
459 described here. Instead, use llvm-pdbutil pdb2yaml and examine the
460 output for an example starting point.
461 Options
463 -pdb=<file-name>
464 Write the resulting PDB to the specified file.
466 merge
468 USAGE: llvm-pdbutil merge [options] <input PDB file 1> <input PDB file
469 2>
470 Summary
472 Merge two PDB files into a single file.
473 Options
475 -pdb=<file-name>
476 Write the resulting PDB to the specified file.
478
480 Maintained by the LLVM Team (https://llvm.org/).
481
483 2003-2021, LLVM Project
48412 2021-05-17 LLVM-PDBUTIL(1)