DynamicLoader.cpp 19 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540
  1. /*
  2. * Copyright (c) 2019-2020, Andrew Kaster <andrewdkaster@gmail.com>
  3. * Copyright (c) 2020, Itamar S. <itamar8910@gmail.com>
  4. * Copyright (c) 2021, Andreas Kling <kling@serenityos.org>
  5. * All rights reserved.
  6. *
  7. * Redistribution and use in source and binary forms, with or without
  8. * modification, are permitted provided that the following conditions are met:
  9. *
  10. * 1. Redistributions of source code must retain the above copyright notice, this
  11. * list of conditions and the following disclaimer.
  12. *
  13. * 2. Redistributions in binary form must reproduce the above copyright notice,
  14. * this list of conditions and the following disclaimer in the documentation
  15. * and/or other materials provided with the distribution.
  16. *
  17. * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
  18. * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  19. * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
  20. * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
  21. * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  22. * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
  23. * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
  24. * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
  25. * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
  26. * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  27. */
  28. #include <AK/Debug.h>
  29. #include <AK/Optional.h>
  30. #include <AK/StringBuilder.h>
  31. #include <LibELF/DynamicLinker.h>
  32. #include <LibELF/DynamicLoader.h>
  33. #include <LibELF/Hashes.h>
  34. #include <LibELF/Validation.h>
  35. #include <assert.h>
  36. #include <dlfcn.h>
  37. #include <stdio.h>
  38. #include <stdlib.h>
  39. #include <string.h>
  40. #include <sys/mman.h>
  41. #include <sys/stat.h>
  42. #ifndef __serenity__
  43. static void* mmap_with_name(void* addr, size_t length, int prot, int flags, int fd, off_t offset, const char*)
  44. {
  45. return mmap(addr, length, prot, flags, fd, offset);
  46. }
  47. # define MAP_RANDOMIZED 0
  48. #endif
  49. namespace ELF {
  50. RefPtr<DynamicLoader> DynamicLoader::try_create(int fd, String filename)
  51. {
  52. struct stat stat;
  53. if (fstat(fd, &stat) < 0) {
  54. perror("DynamicLoader::try_create fstat");
  55. return {};
  56. }
  57. VERIFY(stat.st_size >= 0);
  58. auto size = static_cast<size_t>(stat.st_size);
  59. if (size < sizeof(Elf32_Ehdr))
  60. return {};
  61. String file_mmap_name = String::formatted("ELF_DYN: {}", filename);
  62. auto* data = mmap_with_name(nullptr, size, PROT_READ, MAP_PRIVATE, fd, 0, file_mmap_name.characters());
  63. if (data == MAP_FAILED) {
  64. perror("DynamicLoader::try_create mmap");
  65. return {};
  66. }
  67. return adopt(*new DynamicLoader(fd, move(filename), data, size));
  68. }
  69. DynamicLoader::DynamicLoader(int fd, String filename, void* data, size_t size)
  70. : m_filename(move(filename))
  71. , m_file_size(size)
  72. , m_image_fd(fd)
  73. , m_file_data(data)
  74. , m_elf_image((u8*)m_file_data, m_file_size)
  75. {
  76. m_tls_size = calculate_tls_size();
  77. m_valid = validate();
  78. }
  79. DynamicLoader::~DynamicLoader()
  80. {
  81. if (munmap(m_file_data, m_file_size) < 0) {
  82. perror("munmap");
  83. VERIFY_NOT_REACHED();
  84. }
  85. if (close(m_image_fd) < 0) {
  86. perror("close");
  87. VERIFY_NOT_REACHED();
  88. }
  89. }
  90. const DynamicObject& DynamicLoader::dynamic_object() const
  91. {
  92. if (!m_cached_dynamic_object) {
  93. VirtualAddress dynamic_section_address;
  94. m_elf_image.for_each_program_header([&dynamic_section_address](auto program_header) {
  95. if (program_header.type() == PT_DYNAMIC) {
  96. dynamic_section_address = VirtualAddress(program_header.raw_data());
  97. }
  98. return IterationDecision::Continue;
  99. });
  100. VERIFY(!dynamic_section_address.is_null());
  101. m_cached_dynamic_object = ELF::DynamicObject::create(VirtualAddress(m_elf_image.base_address()), dynamic_section_address);
  102. }
  103. return *m_cached_dynamic_object;
  104. }
  105. size_t DynamicLoader::calculate_tls_size() const
  106. {
  107. size_t tls_size = 0;
  108. m_elf_image.for_each_program_header([&tls_size](auto program_header) {
  109. if (program_header.type() == PT_TLS) {
  110. tls_size = program_header.size_in_memory();
  111. }
  112. return IterationDecision::Continue;
  113. });
  114. return tls_size;
  115. }
  116. bool DynamicLoader::validate()
  117. {
  118. if (!m_elf_image.is_valid())
  119. return false;
  120. auto* elf_header = (Elf32_Ehdr*)m_file_data;
  121. if (!validate_elf_header(*elf_header, m_file_size))
  122. return false;
  123. if (!validate_program_headers(*elf_header, m_file_size, (u8*)m_file_data, m_file_size, &m_program_interpreter))
  124. return false;
  125. return true;
  126. }
  127. void* DynamicLoader::symbol_for_name(const StringView& name)
  128. {
  129. auto result = m_dynamic_object->hash_section().lookup_symbol(name, compute_gnu_hash(name), compute_sysv_hash(name));
  130. if (!result.has_value())
  131. return nullptr;
  132. auto symbol = result.value();
  133. if (symbol.is_undefined())
  134. return nullptr;
  135. return m_dynamic_object->base_address().offset(symbol.value()).as_ptr();
  136. }
  137. RefPtr<DynamicObject> DynamicLoader::map()
  138. {
  139. VERIFY(!m_dynamic_object);
  140. if (!m_valid) {
  141. dbgln("DynamicLoader::map failed: image is invalid");
  142. return nullptr;
  143. }
  144. load_program_headers();
  145. m_dynamic_object = DynamicObject::create(m_text_segment_load_address, m_dynamic_section_address);
  146. m_dynamic_object->set_tls_offset(m_tls_offset);
  147. m_dynamic_object->set_tls_size(m_tls_size);
  148. return m_dynamic_object;
  149. }
  150. bool DynamicLoader::link(unsigned flags, size_t total_tls_size)
  151. {
  152. return load_stage_2(flags, total_tls_size);
  153. }
  154. bool DynamicLoader::load_stage_2(unsigned flags, size_t total_tls_size)
  155. {
  156. VERIFY(flags & RTLD_GLOBAL);
  157. if (m_dynamic_object->has_text_relocations()) {
  158. VERIFY(m_text_segment_load_address.get() != 0);
  159. #ifndef AK_OS_MACOS
  160. // Remap this text region as private.
  161. if (mremap(m_text_segment_load_address.as_ptr(), m_text_segment_size, m_text_segment_size, MAP_PRIVATE) == MAP_FAILED) {
  162. perror("mremap .text: MAP_PRIVATE");
  163. return false;
  164. }
  165. #endif
  166. if (0 > mprotect(m_text_segment_load_address.as_ptr(), m_text_segment_size, PROT_READ | PROT_WRITE)) {
  167. perror("mprotect .text: PROT_READ | PROT_WRITE"); // FIXME: dlerror?
  168. return false;
  169. }
  170. }
  171. do_main_relocations(total_tls_size);
  172. return true;
  173. }
  174. void DynamicLoader::do_main_relocations(size_t total_tls_size)
  175. {
  176. auto do_single_relocation = [&](const ELF::DynamicObject::Relocation& relocation) {
  177. switch (do_relocation(total_tls_size, relocation)) {
  178. case RelocationResult::Failed:
  179. dbgln("Loader.so: {} unresolved symbol '{}'", m_filename, relocation.symbol().name());
  180. VERIFY_NOT_REACHED();
  181. case RelocationResult::ResolveLater:
  182. m_unresolved_relocations.append(relocation);
  183. break;
  184. case RelocationResult::Success:
  185. break;
  186. }
  187. return IterationDecision::Continue;
  188. };
  189. m_dynamic_object->relocation_section().for_each_relocation(do_single_relocation);
  190. m_dynamic_object->plt_relocation_section().for_each_relocation(do_single_relocation);
  191. }
  192. RefPtr<DynamicObject> DynamicLoader::load_stage_3(unsigned flags, size_t total_tls_size)
  193. {
  194. do_lazy_relocations(total_tls_size);
  195. if (flags & RTLD_LAZY) {
  196. if (m_dynamic_object->has_plt())
  197. setup_plt_trampoline();
  198. }
  199. if (mprotect(m_text_segment_load_address.as_ptr(), m_text_segment_size, PROT_READ | PROT_EXEC) < 0) {
  200. perror("mprotect .text: PROT_READ | PROT_EXEC"); // FIXME: dlerror?
  201. return nullptr;
  202. }
  203. if (m_relro_segment_size) {
  204. if (mprotect(m_relro_segment_address.as_ptr(), m_relro_segment_size, PROT_READ) < 0) {
  205. perror("mprotect .relro: PROT_READ");
  206. return nullptr;
  207. }
  208. #if __serenity__
  209. if (set_mmap_name(m_relro_segment_address.as_ptr(), m_relro_segment_size, String::formatted("{}: .relro", m_filename).characters()) < 0) {
  210. perror("set_mmap_name .relro");
  211. return nullptr;
  212. }
  213. #endif
  214. }
  215. call_object_init_functions();
  216. return m_dynamic_object;
  217. }
  218. void DynamicLoader::do_lazy_relocations(size_t total_tls_size)
  219. {
  220. for (const auto& relocation : m_unresolved_relocations) {
  221. if (auto res = do_relocation(total_tls_size, relocation); res != RelocationResult::Success) {
  222. dbgln("Loader.so: {} unresolved symbol '{}'", m_filename, relocation.symbol().name());
  223. VERIFY_NOT_REACHED();
  224. }
  225. }
  226. }
  227. void DynamicLoader::load_program_headers()
  228. {
  229. Optional<ProgramHeaderRegion> text_region;
  230. Optional<ProgramHeaderRegion> data_region;
  231. Optional<ProgramHeaderRegion> tls_region;
  232. Optional<ProgramHeaderRegion> relro_region;
  233. VirtualAddress dynamic_region_desired_vaddr;
  234. m_elf_image.for_each_program_header([&](const Image::ProgramHeader& program_header) {
  235. ProgramHeaderRegion region {};
  236. region.set_program_header(program_header.raw_header());
  237. if (region.is_tls_template()) {
  238. VERIFY(!tls_region.has_value());
  239. tls_region = region;
  240. } else if (region.is_load()) {
  241. if (region.is_executable()) {
  242. VERIFY(!text_region.has_value());
  243. text_region = region;
  244. } else {
  245. VERIFY(!data_region.has_value());
  246. data_region = region;
  247. }
  248. } else if (region.is_dynamic()) {
  249. dynamic_region_desired_vaddr = region.desired_load_address();
  250. } else if (region.is_relro()) {
  251. VERIFY(!relro_region.has_value());
  252. relro_region = region;
  253. }
  254. return IterationDecision::Continue;
  255. });
  256. VERIFY(text_region.has_value());
  257. VERIFY(data_region.has_value());
  258. // Process regions in order: .text, .data, .tls
  259. void* requested_load_address = m_elf_image.is_dynamic() ? nullptr : text_region.value().desired_load_address().as_ptr();
  260. int reservation_mmap_flags = MAP_ANON | MAP_PRIVATE | MAP_NORESERVE;
  261. if (m_elf_image.is_dynamic())
  262. reservation_mmap_flags |= MAP_RANDOMIZED;
  263. else
  264. reservation_mmap_flags |= MAP_FIXED;
  265. VERIFY(!text_region.value().is_writable());
  266. // First, we make a dummy reservation mapping, in order to allocate enough VM
  267. // to hold both text+data contiguously in the address space.
  268. FlatPtr ph_text_base = text_region.value().desired_load_address().page_base().get();
  269. FlatPtr ph_text_end = round_up_to_power_of_two(text_region.value().desired_load_address().offset(text_region.value().size_in_memory()).get(), PAGE_SIZE);
  270. FlatPtr ph_data_base = data_region.value().desired_load_address().page_base().get();
  271. FlatPtr ph_data_end = round_up_to_power_of_two(data_region.value().desired_load_address().offset(data_region.value().size_in_memory()).get(), PAGE_SIZE);
  272. size_t total_mapping_size = ph_data_end - ph_text_base;
  273. size_t text_segment_size = ph_text_end - ph_text_base;
  274. size_t data_segment_size = ph_data_end - ph_data_base;
  275. auto* reservation = mmap(requested_load_address, total_mapping_size, PROT_NONE, reservation_mmap_flags, 0, 0);
  276. if (reservation == MAP_FAILED) {
  277. perror("mmap reservation");
  278. VERIFY_NOT_REACHED();
  279. }
  280. // Then we unmap the reservation.
  281. if (munmap(reservation, total_mapping_size) < 0) {
  282. perror("munmap reservation");
  283. VERIFY_NOT_REACHED();
  284. }
  285. // Now we can map the text segment at the reserved address.
  286. auto* text_segment_begin = (u8*)mmap_with_name(
  287. reservation,
  288. text_segment_size,
  289. PROT_READ,
  290. MAP_FILE | MAP_SHARED | MAP_FIXED,
  291. m_image_fd,
  292. text_region.value().offset(),
  293. String::formatted("{}: .text", m_filename).characters());
  294. if (text_segment_begin == MAP_FAILED) {
  295. perror("mmap text");
  296. VERIFY_NOT_REACHED();
  297. }
  298. VERIFY(requested_load_address == nullptr || requested_load_address == text_segment_begin);
  299. m_text_segment_size = text_segment_size;
  300. m_text_segment_load_address = VirtualAddress { (FlatPtr)text_segment_begin };
  301. if (relro_region.has_value()) {
  302. m_relro_segment_size = relro_region->size_in_memory();
  303. m_relro_segment_address = m_text_segment_load_address.offset(relro_region->desired_load_address().get());
  304. }
  305. if (m_elf_image.is_dynamic())
  306. m_dynamic_section_address = dynamic_region_desired_vaddr.offset(m_text_segment_load_address.get());
  307. else
  308. m_dynamic_section_address = dynamic_region_desired_vaddr;
  309. FlatPtr data_segment_offset_from_text = ph_data_base - ph_text_base;
  310. // Finally, we make an anonymous mapping for the data segment. Contents are then copied from the file.
  311. auto* data_segment_address = (u8*)text_segment_begin + data_segment_offset_from_text;
  312. auto* data_segment = (u8*)mmap_with_name(
  313. data_segment_address,
  314. data_segment_size,
  315. PROT_READ | PROT_WRITE,
  316. MAP_ANONYMOUS | MAP_PRIVATE | MAP_FIXED,
  317. 0,
  318. 0,
  319. String::formatted("{}: .data", m_filename).characters());
  320. if (MAP_FAILED == data_segment) {
  321. perror("mmap data");
  322. VERIFY_NOT_REACHED();
  323. }
  324. VirtualAddress data_segment_start;
  325. if (m_elf_image.is_dynamic())
  326. data_segment_start = data_region.value().desired_load_address().offset((FlatPtr)text_segment_begin);
  327. else
  328. data_segment_start = data_region.value().desired_load_address();
  329. memcpy(data_segment_start.as_ptr(), (u8*)m_file_data + data_region.value().offset(), data_region.value().size_in_image());
  330. // FIXME: Initialize the values in the TLS section. Currently, it is zeroed.
  331. }
  332. DynamicLoader::RelocationResult DynamicLoader::do_relocation(size_t total_tls_size, const ELF::DynamicObject::Relocation& relocation)
  333. {
  334. FlatPtr* patch_ptr = nullptr;
  335. if (is_dynamic())
  336. patch_ptr = (FlatPtr*)(m_dynamic_object->base_address().as_ptr() + relocation.offset());
  337. else
  338. patch_ptr = (FlatPtr*)(FlatPtr)relocation.offset();
  339. switch (relocation.type()) {
  340. case R_386_NONE:
  341. // Apparently most loaders will just skip these?
  342. // Seems if the 'link editor' generates one something is funky with your code
  343. break;
  344. case R_386_32: {
  345. auto symbol = relocation.symbol();
  346. auto res = lookup_symbol(symbol);
  347. if (!res.has_value()) {
  348. if (symbol.bind() == STB_WEAK)
  349. return RelocationResult::ResolveLater;
  350. dbgln("ERROR: symbol not found: {}.", symbol.name());
  351. VERIFY_NOT_REACHED();
  352. }
  353. auto symbol_address = res.value().address;
  354. *patch_ptr += symbol_address.get();
  355. break;
  356. }
  357. case R_386_PC32: {
  358. auto symbol = relocation.symbol();
  359. auto result = lookup_symbol(symbol);
  360. VERIFY(result.has_value());
  361. auto relative_offset = result.value().address - m_dynamic_object->base_address().offset(relocation.offset());
  362. *patch_ptr += relative_offset.get();
  363. break;
  364. }
  365. case R_386_GLOB_DAT: {
  366. auto symbol = relocation.symbol();
  367. auto res = lookup_symbol(symbol);
  368. if (!res.has_value()) {
  369. if (symbol.bind() == STB_WEAK)
  370. return RelocationResult::ResolveLater;
  371. // Symbol not found
  372. return RelocationResult::Failed;
  373. }
  374. auto symbol_location = res.value().address;
  375. VERIFY(symbol_location != m_dynamic_object->base_address());
  376. *patch_ptr = symbol_location.get();
  377. break;
  378. }
  379. case R_386_RELATIVE: {
  380. // FIXME: According to the spec, R_386_relative ones must be done first.
  381. // We could explicitly do them first using m_number_of_relocations from DT_RELCOUNT
  382. // However, our compiler is nice enough to put them at the front of the relocations for us :)
  383. *patch_ptr += (FlatPtr)m_dynamic_object->base_address().as_ptr(); // + addend for RelA (addend for Rel is stored at addr)
  384. break;
  385. }
  386. case R_386_TLS_TPOFF32:
  387. case R_386_TLS_TPOFF: {
  388. auto symbol = relocation.symbol();
  389. // For some reason, LibC has a R_386_TLS_TPOFF that refers to the undefined symbol.. huh
  390. if (relocation.symbol_index() == 0)
  391. break;
  392. auto res = lookup_symbol(symbol);
  393. if (!res.has_value())
  394. break;
  395. u32 symbol_value = res.value().value;
  396. auto* dynamic_object_of_symbol = res.value().dynamic_object;
  397. VERIFY(dynamic_object_of_symbol);
  398. size_t offset_of_tls_end = dynamic_object_of_symbol->tls_offset().value() + dynamic_object_of_symbol->tls_size().value();
  399. *patch_ptr = (offset_of_tls_end - total_tls_size - symbol_value - sizeof(Elf32_Addr));
  400. break;
  401. }
  402. case R_386_JMP_SLOT: {
  403. // FIXME: Or BIND_NOW flag passed in?
  404. if (m_dynamic_object->must_bind_now()) {
  405. // Eagerly BIND_NOW the PLT entries, doing all the symbol looking goodness
  406. // The patch method returns the address for the LAZY fixup path, but we don't need it here
  407. m_dynamic_object->patch_plt_entry(relocation.offset_in_section());
  408. } else {
  409. u8* relocation_address = relocation.address().as_ptr();
  410. if (m_elf_image.is_dynamic())
  411. *(u32*)relocation_address += (FlatPtr)m_dynamic_object->base_address().as_ptr();
  412. }
  413. break;
  414. }
  415. default:
  416. // Raise the alarm! Someone needs to implement this relocation type
  417. dbgln("Found a new exciting relocation type {}", relocation.type());
  418. VERIFY_NOT_REACHED();
  419. }
  420. return RelocationResult::Success;
  421. }
  422. // Defined in <arch>/plt_trampoline.S
  423. extern "C" void _plt_trampoline(void) __attribute__((visibility("hidden")));
  424. void DynamicLoader::setup_plt_trampoline()
  425. {
  426. VERIFY(m_dynamic_object);
  427. VERIFY(m_dynamic_object->has_plt());
  428. VirtualAddress got_address = m_dynamic_object->plt_got_base_address();
  429. auto* got_ptr = (FlatPtr*)got_address.as_ptr();
  430. got_ptr[1] = (FlatPtr)m_dynamic_object.ptr();
  431. got_ptr[2] = (FlatPtr)&_plt_trampoline;
  432. }
  433. // Called from our ASM routine _plt_trampoline.
  434. // Tell the compiler that it might be called from other places:
  435. extern "C" FlatPtr _fixup_plt_entry(DynamicObject* object, u32 relocation_offset);
  436. extern "C" FlatPtr _fixup_plt_entry(DynamicObject* object, u32 relocation_offset)
  437. {
  438. return object->patch_plt_entry(relocation_offset).get();
  439. }
  440. void DynamicLoader::call_object_init_functions()
  441. {
  442. typedef void (*InitFunc)();
  443. if (m_dynamic_object->has_init_section()) {
  444. auto init_function = (InitFunc)(m_dynamic_object->init_section().address().as_ptr());
  445. (init_function)();
  446. }
  447. if (m_dynamic_object->has_init_array_section()) {
  448. auto init_array_section = m_dynamic_object->init_array_section();
  449. InitFunc* init_begin = (InitFunc*)(init_array_section.address().as_ptr());
  450. InitFunc* init_end = init_begin + init_array_section.entry_count();
  451. while (init_begin != init_end) {
  452. // Android sources claim that these can be -1, to be ignored.
  453. // 0 definitely shows up. Apparently 0/-1 are valid? Confusing.
  454. if (!*init_begin || ((FlatPtr)*init_begin == (FlatPtr)-1))
  455. continue;
  456. (*init_begin)();
  457. ++init_begin;
  458. }
  459. }
  460. }
  461. Optional<DynamicObject::SymbolLookupResult> DynamicLoader::lookup_symbol(const ELF::DynamicObject::Symbol& symbol)
  462. {
  463. if (symbol.is_undefined() || symbol.bind() == STB_WEAK)
  464. return DynamicLinker::lookup_global_symbol(symbol.name());
  465. return DynamicObject::SymbolLookupResult { symbol.value(), symbol.address(), symbol.bind(), &symbol.object() };
  466. }
  467. } // end namespace ELF