Objdump.pm 15 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549
  1. # Copyright © 2007-2010 Raphaël Hertzog <hertzog@debian.org>
  2. #
  3. # This program is free software; you can redistribute it and/or modify
  4. # it under the terms of the GNU General Public License as published by
  5. # the Free Software Foundation; either version 2 of the License, or
  6. # (at your option) any later version.
  7. #
  8. # This program is distributed in the hope that it will be useful,
  9. # but WITHOUT ANY WARRANTY; without even the implied warranty of
  10. # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  11. # GNU General Public License for more details.
  12. #
  13. # You should have received a copy of the GNU General Public License
  14. # along with this program. If not, see <https://www.gnu.org/licenses/>.
  15. package Dpkg::Shlibs::Objdump;
  16. use strict;
  17. use warnings;
  18. use feature qw(state);
  19. our $VERSION = '0.01';
  20. use Dpkg::Gettext;
  21. use Dpkg::ErrorHandling;
  22. use Dpkg::Path qw(find_command);
  23. use Dpkg::Arch qw(debarch_to_gnutriplet get_build_arch get_host_arch);
  24. use Dpkg::IPC;
  25. # Decide which objdump to call
  26. our $OBJDUMP = 'objdump';
  27. if (get_build_arch() ne get_host_arch()) {
  28. my $od = debarch_to_gnutriplet(get_host_arch()) . '-objdump';
  29. $OBJDUMP = $od if find_command($od);
  30. }
  31. sub new {
  32. my $this = shift;
  33. my $class = ref($this) || $this;
  34. my $self = { objects => {} };
  35. bless $self, $class;
  36. return $self;
  37. }
  38. sub add_object {
  39. my ($self, $obj) = @_;
  40. my $id = $obj->get_id;
  41. if ($id) {
  42. $self->{objects}{$id} = $obj;
  43. }
  44. return $id;
  45. }
  46. sub analyze {
  47. my ($self, $file) = @_;
  48. my $obj = Dpkg::Shlibs::Objdump::Object->new($file);
  49. return $self->add_object($obj);
  50. }
  51. sub locate_symbol {
  52. my ($self, $name) = @_;
  53. foreach my $obj (values %{$self->{objects}}) {
  54. my $sym = $obj->get_symbol($name);
  55. if (defined($sym) && $sym->{defined}) {
  56. return $sym;
  57. }
  58. }
  59. return;
  60. }
  61. sub get_object {
  62. my ($self, $objid) = @_;
  63. if ($self->has_object($objid)) {
  64. return $self->{objects}{$objid};
  65. }
  66. return;
  67. }
  68. sub has_object {
  69. my ($self, $objid) = @_;
  70. return exists $self->{objects}{$objid};
  71. }
  72. use constant {
  73. ELF_BITS_NONE => 0,
  74. ELF_BITS_32 => 1,
  75. ELF_BITS_64 => 2,
  76. ELF_ORDER_NONE => 0,
  77. ELF_ORDER_2LSB => 1,
  78. ELF_ORDER_2MSB => 2,
  79. ELF_MACH_SPARC => 2,
  80. ELF_MACH_MIPS => 8,
  81. ELF_MACH_SPARC64_OLD => 11,
  82. ELF_MACH_SPARC32PLUS => 18,
  83. ELF_MACH_PPC64 => 21,
  84. ELF_MACH_S390 => 22,
  85. ELF_MACH_ARM => 40,
  86. ELF_MACH_ALPHA_OLD => 41,
  87. ELF_MACH_SH => 42,
  88. ELF_MACH_SPARC64 => 43,
  89. ELF_MACH_IA64 => 50,
  90. ELF_MACH_AVR => 83,
  91. ELF_MACH_M32R => 88,
  92. ELF_MACH_MN10300 => 89,
  93. ELF_MACH_MN10200 => 90,
  94. ELF_MACH_OR1K => 92,
  95. ELF_MACH_XTENSA => 94,
  96. ELF_MACH_MICROBLAZE => 189,
  97. ELF_MACH_AVR_OLD => 0x1057,
  98. ELF_MACH_OR1K_OLD => 0x8472,
  99. ELF_MACH_ALPHA => 0x9026,
  100. ELF_MACH_M32R_CYGNUS => 0x9041,
  101. ELF_MACH_S390_OLD => 0xa390,
  102. ELF_MACH_XTENSA_OLD => 0xabc7,
  103. ELF_MACH_MICROBLAZE_OLD => 0xbaab,
  104. ELF_MACH_MN10300_CYGNUS => 0xbeef,
  105. ELF_MACH_MN10200_CYGNUS => 0xdead,
  106. ELF_VERSION_NONE => 0,
  107. ELF_VERSION_CURRENT => 1,
  108. # List of processor flags that might influence the ABI.
  109. ELF_FLAG_ARM_ALIGN8 => 0x00000040,
  110. ELF_FLAG_ARM_NEW_ABI => 0x00000080,
  111. ELF_FLAG_ARM_OLD_ABI => 0x00000100,
  112. ELF_FLAG_ARM_SOFT_FLOAT => 0x00000200,
  113. ELF_FLAG_ARM_HARD_FLOAT => 0x00000400,
  114. ELF_FLAG_ARM_EABI_MASK => 0xff000000,
  115. ELF_FLAG_IA64_ABI64 => 0x00000010,
  116. ELF_FLAG_MIPS_ABI2 => 0x00000020,
  117. ELF_FLAG_MIPS_32BIT => 0x00000100,
  118. ELF_FLAG_MIPS_FP64 => 0x00000200,
  119. ELF_FLAG_MIPS_NAN2008 => 0x00000400,
  120. ELF_FLAG_MIPS_ABI_MASK => 0x0000f000,
  121. ELF_FLAG_MIPS_ARCH_MASK => 0xf0000000,
  122. ELF_FLAG_PPC64_ABI64 => 0x00000003,
  123. ELF_FLAG_SH_MACH_MASK => 0x0000001f,
  124. };
  125. # These map alternative or old machine IDs to their canonical form.
  126. my %elf_mach_map = (
  127. ELF_MACH_ALPHA_OLD() => ELF_MACH_ALPHA,
  128. ELF_MACH_AVR_OLD() => ELF_MACH_AVR,
  129. ELF_MACH_M32R_CYGNUS() => ELF_MACH_M32R,
  130. ELF_MACH_MICROBLAZE_OLD() => ELF_MACH_MICROBLAZE,
  131. ELF_MACH_MN10200_CYGNUS() => ELF_MACH_MN10200,
  132. ELF_MACH_MN10300_CYGNUS() => ELF_MACH_MN10300,
  133. ELF_MACH_OR1K_OLD() => ELF_MACH_OR1K,
  134. ELF_MACH_S390_OLD() => ELF_MACH_S390,
  135. ELF_MACH_SPARC32PLUS() => ELF_MACH_SPARC,
  136. ELF_MACH_SPARC64_OLD() => ELF_MACH_SPARC64,
  137. ELF_MACH_XTENSA_OLD() => ELF_MACH_XTENSA,
  138. );
  139. # These masks will try to expose processor flags that are ABI incompatible,
  140. # and as such are part of defining the architecture ABI. If uncertain it is
  141. # always better to not mask a flag, because that preserves the historical
  142. # behavior, and we do not drop dependencies.
  143. my %elf_flags_mask = (
  144. ELF_MACH_IA64() => ELF_FLAG_IA64_ABI64,
  145. ELF_MACH_MIPS() => ELF_FLAG_MIPS_ABI_MASK | ELF_FLAG_MIPS_ABI2,
  146. ELF_MACH_PPC64() => ELF_FLAG_PPC64_ABI64,
  147. );
  148. sub get_format {
  149. my ($file) = @_;
  150. state %format;
  151. return $format{$file} if exists $format{$file};
  152. my $header;
  153. open my $fh, '<', $file or syserr(g_('cannot read %s'), $file);
  154. my $rc = read $fh, $header, 64;
  155. if (not defined $rc) {
  156. syserr(g_('cannot read %s'), $file);
  157. } elsif ($rc != 64) {
  158. return;
  159. }
  160. close $fh;
  161. my %elf;
  162. # Unpack the identifier field.
  163. @elf{qw(magic bits endian vertype osabi verabi)} = unpack 'a4C5', $header;
  164. return unless $elf{magic} eq "\x7fELF";
  165. return unless $elf{vertype} == ELF_VERSION_CURRENT;
  166. my ($elf_word, $elf_endian);
  167. if ($elf{bits} == ELF_BITS_32) {
  168. $elf_word = 'L';
  169. } elsif ($elf{bits} == ELF_BITS_64) {
  170. $elf_word = 'Q';
  171. } elsif ($elf{bits} == ELF_BITS_NONE) {
  172. return;
  173. }
  174. if ($elf{endian} == ELF_ORDER_2LSB) {
  175. $elf_endian = '<';
  176. } elsif ($elf{endian} == ELF_ORDER_2MSB) {
  177. $elf_endian = '>';
  178. } elsif ($elf{endian} == ELF_ORDER_NONE) {
  179. return;
  180. }
  181. # Unpack the endianness and size dependent fields.
  182. my $tmpl = "x16(S2Lx[${elf_word}3]L)${elf_endian}";
  183. @elf{qw(type mach version flags)} = unpack $tmpl, $header;
  184. # Canonicalize the machine ID.
  185. $elf{mach} = $elf_mach_map{$elf{mach}} // $elf{mach};
  186. # Mask any processor flags that might not change the architecture ABI.
  187. $elf{flags} &= $elf_flags_mask{$elf{mach}} // 0;
  188. # Repack for easy comparison, as a big-endian byte stream, so that
  189. # unpacking for output gives meaningful results.
  190. $format{$file} = pack 'C2(SL)>', @elf{qw(bits endian mach flags)};
  191. return $format{$file};
  192. }
  193. sub is_elf {
  194. my $file = shift;
  195. open(my $file_fh, '<', $file) or syserr(g_('cannot read %s'), $file);
  196. my ($header, $result) = ('', 0);
  197. if (read($file_fh, $header, 4) == 4) {
  198. $result = 1 if ($header =~ /^\177ELF$/);
  199. }
  200. close($file_fh);
  201. return $result;
  202. }
  203. package Dpkg::Shlibs::Objdump::Object;
  204. use strict;
  205. use warnings;
  206. use Dpkg::Gettext;
  207. use Dpkg::ErrorHandling;
  208. sub new {
  209. my $this = shift;
  210. my $file = shift // '';
  211. my $class = ref($this) || $this;
  212. my $self = {};
  213. bless $self, $class;
  214. $self->reset;
  215. if ($file) {
  216. $self->analyze($file);
  217. }
  218. return $self;
  219. }
  220. sub reset {
  221. my $self = shift;
  222. $self->{file} = '';
  223. $self->{id} = '';
  224. $self->{SONAME} = '';
  225. $self->{HASH} = '';
  226. $self->{GNU_HASH} = '';
  227. $self->{SONAME} = '';
  228. $self->{NEEDED} = [];
  229. $self->{RPATH} = [];
  230. $self->{dynsyms} = {};
  231. $self->{flags} = {};
  232. $self->{dynrelocs} = {};
  233. return $self;
  234. }
  235. sub analyze {
  236. my ($self, $file) = @_;
  237. $file ||= $self->{file};
  238. return unless $file;
  239. $self->reset;
  240. $self->{file} = $file;
  241. $self->{exec_abi} = Dpkg::Shlibs::Objdump::get_format($file);
  242. if (not defined $self->{exec_abi}) {
  243. warning(g_("unknown executable format in file '%s'"), $file);
  244. return;
  245. }
  246. local $ENV{LC_ALL} = 'C';
  247. open(my $objdump, '-|', $OBJDUMP, '-w', '-f', '-p', '-T', '-R', $file)
  248. or syserr(g_('cannot fork for %s'), $OBJDUMP);
  249. my $ret = $self->parse_objdump_output($objdump);
  250. close($objdump);
  251. return $ret;
  252. }
  253. sub parse_objdump_output {
  254. my ($self, $fh) = @_;
  255. my $section = 'none';
  256. while (<$fh>) {
  257. s/\s*$//;
  258. next if length == 0;
  259. if (/^DYNAMIC SYMBOL TABLE:/) {
  260. $section = 'dynsym';
  261. next;
  262. } elsif (/^DYNAMIC RELOCATION RECORDS/) {
  263. $section = 'dynreloc';
  264. $_ = <$fh>; # Skip header
  265. next;
  266. } elsif (/^Dynamic Section:/) {
  267. $section = 'dyninfo';
  268. next;
  269. } elsif (/^Program Header:/) {
  270. $section = 'header';
  271. next;
  272. } elsif (/^Version definitions:/) {
  273. $section = 'verdef';
  274. next;
  275. } elsif (/^Version References:/) {
  276. $section = 'verref';
  277. next;
  278. }
  279. if ($section eq 'dynsym') {
  280. $self->parse_dynamic_symbol($_);
  281. } elsif ($section eq 'dynreloc') {
  282. if (/^\S+\s+(\S+)\s+(.+)$/) {
  283. $self->{dynrelocs}{$2} = $1;
  284. } else {
  285. warning(g_("couldn't parse dynamic relocation record: %s"), $_);
  286. }
  287. } elsif ($section eq 'dyninfo') {
  288. if (/^\s*NEEDED\s+(\S+)/) {
  289. push @{$self->{NEEDED}}, $1;
  290. } elsif (/^\s*SONAME\s+(\S+)/) {
  291. $self->{SONAME} = $1;
  292. } elsif (/^\s*HASH\s+(\S+)/) {
  293. $self->{HASH} = $1;
  294. } elsif (/^\s*GNU_HASH\s+(\S+)/) {
  295. $self->{GNU_HASH} = $1;
  296. } elsif (/^\s*RUNPATH\s+(\S+)/) {
  297. # RUNPATH takes precedence over RPATH but is
  298. # considered after LD_LIBRARY_PATH while RPATH
  299. # is considered before (if RUNPATH is not set).
  300. my $runpath = $1;
  301. $self->{RPATH} = [ split /:/, $runpath ];
  302. } elsif (/^\s*RPATH\s+(\S+)/) {
  303. my $rpath = $1;
  304. unless (scalar(@{$self->{RPATH}})) {
  305. $self->{RPATH} = [ split /:/, $rpath ];
  306. }
  307. }
  308. } elsif ($section eq 'none') {
  309. if (/^\s*.+:\s*file\s+format\s+(\S+)$/) {
  310. $self->{format} = $1;
  311. } elsif (/^architecture:\s*\S+,\s*flags\s*\S+:$/) {
  312. # Parse 2 lines of "-f"
  313. # architecture: i386, flags 0x00000112:
  314. # EXEC_P, HAS_SYMS, D_PAGED
  315. # start address 0x08049b50
  316. $_ = <$fh>;
  317. chomp;
  318. $self->{flags}{$_} = 1 foreach (split(/,\s*/));
  319. }
  320. }
  321. }
  322. # Update status of dynamic symbols given the relocations that have
  323. # been parsed after the symbols...
  324. $self->apply_relocations();
  325. return $section ne 'none';
  326. }
  327. # Output format of objdump -w -T
  328. #
  329. # /lib/libc.so.6: file format elf32-i386
  330. #
  331. # DYNAMIC SYMBOL TABLE:
  332. # 00056ef0 g DF .text 000000db GLIBC_2.2 getwchar
  333. # 00000000 g DO *ABS* 00000000 GCC_3.0 GCC_3.0
  334. # 00069960 w DF .text 0000001e GLIBC_2.0 bcmp
  335. # 00000000 w D *UND* 00000000 _pthread_cleanup_pop_restore
  336. # 0000b788 g DF .text 0000008e Base .protected xine_close
  337. # 0000b788 g DF .text 0000008e .hidden IA__g_free
  338. # | ||||||| | | | |
  339. # | ||||||| | | Version str (.visibility) + Symbol name
  340. # | ||||||| | Alignment
  341. # | ||||||| Section name (or *UND* for an undefined symbol)
  342. # | ||||||F=Function,f=file,O=object
  343. # | |||||d=debugging,D=dynamic
  344. # | ||||I=Indirect
  345. # | |||W=warning
  346. # | ||C=constructor
  347. # | |w=weak
  348. # | g=global,l=local,!=both global/local
  349. # Size of the symbol
  350. #
  351. # GLIBC_2.2 is the version string associated to the symbol
  352. # (GLIBC_2.2) is the same but the symbol is hidden, a newer version of the
  353. # symbol exist
  354. my $vis_re = qr/(\.protected|\.hidden|\.internal|0x\S+)/;
  355. my $dynsym_re = qr<
  356. ^
  357. [0-9a-f]+ # Symbol size
  358. \ (.{7}) # Flags
  359. \s+(\S+) # Section name
  360. \s+[0-9a-f]+ # Alignment
  361. (?:\s+(\S+))? # Version string
  362. (?:\s+$vis_re)? # Visibility
  363. \s+(.+) # Symbol name
  364. >x;
  365. sub parse_dynamic_symbol {
  366. my ($self, $line) = @_;
  367. if ($line =~ $dynsym_re) {
  368. my ($flags, $sect, $ver, $vis, $name) = ($1, $2, $3, $4, $5);
  369. # Special case if version is missing but extra visibility
  370. # attribute replaces it in the match
  371. if (defined($ver) and $ver =~ /^$vis_re$/) {
  372. $vis = $ver;
  373. $ver = '';
  374. }
  375. # Cleanup visibility field
  376. $vis =~ s/^\.// if defined($vis);
  377. my $symbol = {
  378. name => $name,
  379. version => $ver // '',
  380. section => $sect,
  381. dynamic => substr($flags, 5, 1) eq 'D',
  382. debug => substr($flags, 5, 1) eq 'd',
  383. type => substr($flags, 6, 1),
  384. weak => substr($flags, 1, 1) eq 'w',
  385. local => substr($flags, 0, 1) eq 'l',
  386. global => substr($flags, 0, 1) eq 'g',
  387. visibility => $vis // '',
  388. hidden => '',
  389. defined => $sect ne '*UND*'
  390. };
  391. # Handle hidden symbols
  392. if (defined($ver) and $ver =~ /^\((.*)\)$/) {
  393. $ver = $1;
  394. $symbol->{version} = $1;
  395. $symbol->{hidden} = 1;
  396. }
  397. # Register symbol
  398. $self->add_dynamic_symbol($symbol);
  399. } elsif ($line =~ /^[0-9a-f]+ (.{7})\s+(\S+)\s+[0-9a-f]+/) {
  400. # Same start but no version and no symbol ... just ignore
  401. } elsif ($line =~ /^REG_G\d+\s+/) {
  402. # Ignore some s390-specific output like
  403. # REG_G6 g R *UND* 0000000000000000 #scratch
  404. } else {
  405. warning(g_("couldn't parse dynamic symbol definition: %s"), $line);
  406. }
  407. }
  408. sub apply_relocations {
  409. my $self = shift;
  410. foreach my $sym (values %{$self->{dynsyms}}) {
  411. # We want to mark as undefined symbols those which are currently
  412. # defined but that depend on a copy relocation
  413. next if not $sym->{defined};
  414. next if not exists $self->{dynrelocs}{$sym->{name}};
  415. if ($self->{dynrelocs}{$sym->{name}} =~ /^R_.*_COPY$/) {
  416. $sym->{defined} = 0;
  417. }
  418. }
  419. }
  420. sub add_dynamic_symbol {
  421. my ($self, $symbol) = @_;
  422. $symbol->{objid} = $symbol->{soname} = $self->get_id();
  423. $symbol->{soname} =~ s{^.*/}{} unless $self->{SONAME};
  424. if ($symbol->{version}) {
  425. $self->{dynsyms}{$symbol->{name} . '@' . $symbol->{version}} = $symbol;
  426. } else {
  427. $self->{dynsyms}{$symbol->{name} . '@Base'} = $symbol;
  428. }
  429. }
  430. sub get_id {
  431. my $self = shift;
  432. return $self->{SONAME} || $self->{file};
  433. }
  434. sub get_symbol {
  435. my ($self, $name) = @_;
  436. if (exists $self->{dynsyms}{$name}) {
  437. return $self->{dynsyms}{$name};
  438. }
  439. if ($name !~ /@/) {
  440. if (exists $self->{dynsyms}{$name . '@Base'}) {
  441. return $self->{dynsyms}{$name . '@Base'};
  442. }
  443. }
  444. return;
  445. }
  446. sub get_exported_dynamic_symbols {
  447. my $self = shift;
  448. return grep { $_->{defined} && $_->{dynamic} && !$_->{local} }
  449. values %{$self->{dynsyms}};
  450. }
  451. sub get_undefined_dynamic_symbols {
  452. my $self = shift;
  453. return grep { (!$_->{defined}) && $_->{dynamic} }
  454. values %{$self->{dynsyms}};
  455. }
  456. sub get_needed_libraries {
  457. my $self = shift;
  458. return @{$self->{NEEDED}};
  459. }
  460. sub is_executable {
  461. my $self = shift;
  462. return exists $self->{flags}{EXEC_P} && $self->{flags}{EXEC_P};
  463. }
  464. sub is_public_library {
  465. my $self = shift;
  466. return exists $self->{flags}{DYNAMIC} && $self->{flags}{DYNAMIC}
  467. && exists $self->{SONAME} && $self->{SONAME};
  468. }
  469. 1;