launch-cache/dyld_shared_cache_util.cpp

   1 /* -*- mode: C++; c-basic-offset: 4; tab-width: 4 -*-
   2  *
   3  * Copyright (c) 2009-2010 Apple Inc. All rights reserved.
   4  *
   5  * @APPLE_LICENSE_HEADER_START@
   6  *
   7  * This file contains Original Code and/or Modifications of Original Code
   8  * as defined in and that are subject to the Apple Public Source License
   9  * Version 2.0 (the 'License'). You may not use this file except in
  10  * compliance with the License. Please obtain a copy of the License at
  11  * http://www.opensource.apple.com/apsl/ and read it before using this
  12  * file.
  13  *
  14  * The Original Code and all software distributed under the License are
  15  * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
  16  * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
  17  * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
  18  * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
  19  * Please see the License for the specific language governing rights and
  20  * limitations under the License.
  21  *
  22  * @APPLE_LICENSE_HEADER_END@
  23  */
  24
  25 #include <stdio.h>
  26 #include <unistd.h>
  27 #include <sys/stat.h>
  28 #include <string.h>
  29 #include <fcntl.h>
  30 #include <stdlib.h>
  31 #include <errno.h>
  32 #include <sys/mman.h>
  33 #include <sys/syslimits.h>
  34 #include <mach-o/arch.h>
  35 #include <mach-o/loader.h>
  36
  37 #include <vector>
  38 #include <map>
  39 #include <set>
  40
  41 #include "dsc_iterator.h"
  42 #include "dyld_cache_format.h"
  43 #include "Architectures.hpp"
  44 #include "MachOFileAbstraction.hpp"
  45 #include "CacheFileAbstraction.hpp"
  46
  47
  48 #define OP_NULL 0
  49 #define OP_LIST_DEPENDENCIES 1
  50 #define OP_LIST_DYLIBS 2
  51 #define OP_LIST_LINKEDIT 3
  52
  53 #define UUID_BYTES 16
  54
  55 // Define this here so we can work with or without block support
  56 typedef void (*segment_callback_t)(const char* dylib, const char* segName, uint64_t offset, uint64_t sizem,
  57                                                                                                         uint64_t mappedddress, uint64_t slide, void* userData);
  58
  59 struct seg_callback_args {
  60         char *target_path;
  61         uint32_t target_found;
  62         void *mapped_cache;
  63         uint32_t op;
  64         uint8_t print_uuids;
  65         uint8_t print_vmaddrs;
  66     uint8_t print_dylib_versions;
  67 };
  68
  69 void usage() {
  70         fprintf(stderr, "Usage: dscutil -list [ -uuid ] [-vmaddr] | -dependents <dylib-path> [ -versions ] | -linkedit [ shared-cache-file ]\n");
  71 }
  72
  73 /*
  74  * Get the path to the native shared cache for this host
  75  */
  76 static const char* default_shared_cache_path() {
  77 #if __i386__
  78         return MACOSX_DYLD_SHARED_CACHE_DIR DYLD_SHARED_CACHE_BASE_NAME "i386";
  79 #elif __x86_64__
  80         return MACOSX_DYLD_SHARED_CACHE_DIR DYLD_SHARED_CACHE_BASE_NAME "x86_64";
  81 #elif __ARM_ARCH_5TEJ__
  82         return IPHONE_DYLD_SHARED_CACHE_DIR DYLD_SHARED_CACHE_BASE_NAME "armv5";
  83 #elif __ARM_ARCH_6K__
  84         return IPHONE_DYLD_SHARED_CACHE_DIR DYLD_SHARED_CACHE_BASE_NAME "armv6";
  85 #elif __ARM_ARCH_7A__
  86         return IPHONE_DYLD_SHARED_CACHE_DIR DYLD_SHARED_CACHE_BASE_NAME "armv7";
  87 #elif __ARM_ARCH_7F__
  88         return IPHONE_DYLD_SHARED_CACHE_DIR DYLD_SHARED_CACHE_BASE_NAME "armv7f";
  89 #elif __ARM_ARCH_7K__
  90         return IPHONE_DYLD_SHARED_CACHE_DIR DYLD_SHARED_CACHE_BASE_NAME "armv7k";
  91 #else
  92         #error unsupported architecture
  93 #endif
  94 }
  95
  96 /*
  97  * Get a vector of all the load commands from the header pointed to by headerAddr
  98  */
  99 template <typename A>
 100 std::vector<macho_load_command<typename A::P>* > get_load_cmds(void *headerAddr) {
 101         typedef typename A::P           P;
 102         typedef typename A::P::E        E;
 103
 104         std::vector<macho_load_command<P>* > cmd_vector;
 105
 106         const macho_header<P>* mh = (const macho_header<P>*)headerAddr;
 107         uint32_t ncmds = mh->ncmds();
 108         const macho_load_command<P>* const cmds = (macho_load_command<P>*)((long)mh + sizeof(macho_header<P>));
 109         const macho_load_command<P>* cmd = cmds;
 110
 111         for (uint32_t i = 0; i < ncmds; i++) {
 112                 cmd_vector.push_back((macho_load_command<P>*)cmd);
 113                 cmd = (const macho_load_command<P>*)(((uint8_t*)cmd)+cmd->cmdsize());
 114         }
 115         return cmd_vector;
 116 }
 117
 118 /*
 119  * List dependencies from the mach-o header at headerAddr
 120  * in the same format as 'otool -L'
 121  */
 122 template <typename A>
 123 void list_dependencies(const char *dylib, void *headerAddr, uint8_t print_dylib_versions) {
 124         typedef typename A::P           P;
 125         typedef typename A::P::E        E;
 126
 127         std::vector< macho_load_command<P>* > cmds;
 128         cmds = get_load_cmds<A>(headerAddr);
 129         for(typename std::vector<macho_load_command<P>*>::iterator it = cmds.begin(); it != cmds.end(); ++it) {
 130                 uint32_t cmdType = (*it)->cmd();
 131                 if (cmdType == LC_LOAD_DYLIB ||
 132                         cmdType == LC_ID_DYLIB ||
 133                         cmdType == LC_LOAD_WEAK_DYLIB ||
 134                         cmdType == LC_REEXPORT_DYLIB ||
 135                         cmdType == LC_LOAD_UPWARD_DYLIB) {
 136                         macho_dylib_command<P>* dylib_cmd = (macho_dylib_command<P>*)*it;
 137                         const char *name = dylib_cmd->name();
 138                         uint32_t compat_vers = dylib_cmd->compatibility_version();
 139                         uint32_t current_vers = dylib_cmd->current_version();
 140
 141             if (print_dylib_versions) {
 142                 printf("\t%s", name);
 143                 if ( compat_vers != 0xFFFFFFFF )
 144                     printf("(compatibility version %u.%u.%u, current version %u.%u.%u)\n",
 145                        (compat_vers >> 16),
 146                        (compat_vers >> 8) & 0xff,
 147                        (compat_vers) & 0xff,
 148                        (current_vers >> 16),
 149                        (current_vers >> 8) & 0xff,
 150                        (current_vers) & 0xff);
 151                 else
 152                     printf("\n");
 153             } else {
 154                 printf("\t%s\n", name);
 155             }
 156                 }
 157         }
 158 }
 159
 160 /*
 161  * Print out a dylib from the shared cache, optionally including the UUID
 162  */
 163 template <typename A>
 164 void print_dylib(const char *dylib, void *headerAddr, uint64_t slide, struct seg_callback_args *args) {
 165         typedef typename A::P           P;
 166         typedef typename A::P::E        E;
 167         char uuid_str[UUID_BYTES*3];
 168         uint8_t got_uuid = 0;
 169         uint64_t vmaddr = 0;
 170
 171         std::vector< macho_load_command<P>* > cmds;
 172         cmds = get_load_cmds<A>(headerAddr);
 173         for(typename std::vector<macho_load_command<P>*>::iterator it = cmds.begin(); it != cmds.end(); ++it) {
 174                 uint32_t cmdType = (*it)->cmd();
 175                 if (cmdType == LC_UUID) {
 176                         macho_uuid_command<P>* uuid_cmd = (macho_uuid_command<P>*)*it;
 177                         const uint8_t *uuid = uuid_cmd->uuid();
 178                         sprintf(uuid_str, "<%02X%02X%02X%02X-%02X%02X-%02X%02X-%02X%02X-%02X%02X%02X%02X%02X%02X>",
 179                                 uuid[0], uuid[1],  uuid[2],  uuid[3],  uuid[4],  uuid[5],  uuid[6], uuid[7],
 180                                 uuid[8], uuid[9], uuid[10], uuid[11], uuid[12], uuid[13], uuid[14], uuid[15]);
 181                         got_uuid = 1;
 182                 }
 183                 else if (cmdType == LC_SEGMENT) {
 184                         macho_segment_command<P>* seg_cmd = (macho_segment_command<P>*)*it;
 185                         if (strcmp(seg_cmd->segname(), "__TEXT") == 0) {
 186                                 vmaddr = seg_cmd->vmaddr();
 187                         }
 188                 }
 189         }
 190
 191         if (args->print_vmaddrs)
 192                 printf("0x%08llX ", vmaddr+slide);
 193         if (args->print_uuids) {
 194                 if (got_uuid)
 195                         printf("%s ", uuid_str);
 196             else
 197                         printf("<     no uuid in dylib               > ");
 198     }
 199         printf("%s\n", dylib);
 200 }
 201
 202
 203 uint64_t                                        sLinkeditBase = 0;
 204 std::map<uint32_t, char*>       sPageToContent;
 205
 206
 207
 208
 209 static void add_linkedit(uint32_t pageStart, uint32_t pageEnd, char* message)
 210 {
 211         for (uint32_t p = pageStart; p <= pageEnd; p += 4096) {
 212                 std::map<uint32_t, char*>::iterator pos = sPageToContent.find(p);
 213                 if ( pos == sPageToContent.end() ) {
 214                         sPageToContent[p] = strdup(message);
 215                 }
 216                 else {
 217                         char* oldMessage = pos->second;
 218                         char* newMesssage;
 219                         asprintf(&newMesssage, "%s, %s", oldMessage, message);
 220                         sPageToContent[p] = newMesssage;
 221                         free(oldMessage);
 222                 }
 223         }
 224 }
 225
 226
 227 /*
 228  * get LINKEDIT info for dylib
 229  */
 230 template <typename A>
 231 void process_linkedit(const char* dylib, void* headerAddr) {
 232         typedef typename A::P           P;
 233         typedef typename A::P::E        E;
 234         // filter out symlinks by only handling first path found for each mach header
 235         static std::set<void*> seenImages;
 236         if ( seenImages.count(headerAddr) != 0 )
 237                 return;
 238         seenImages.insert(headerAddr);
 239         const macho_header<P>* mh = (const macho_header<P>*)headerAddr;
 240         uint32_t ncmds = mh->ncmds();
 241         const macho_load_command<P>* const cmds = (macho_load_command<P>*)((long)mh + sizeof(macho_header<P>));
 242         const macho_load_command<P>* cmd = cmds;
 243         for (uint32_t i = 0; i < ncmds; i++) {
 244                 if ( cmd->cmd() == LC_DYLD_INFO_ONLY ) {
 245                         macho_dyld_info_command<P>* dyldInfo = (macho_dyld_info_command<P>*)cmd;
 246                         char message[1000];
 247                         const char* shortName = strrchr(dylib, '/') + 1;
 248                         // add export trie info
 249                         if ( dyldInfo->export_size() != 0 ) {
 250                                 //printf("export_off=0x%X\n", dyldInfo->export_off());
 251                                 uint32_t exportPageOffsetStart = dyldInfo->export_off() & (-4096);
 252                                 uint32_t exportPageOffsetEnd = (dyldInfo->export_off() + dyldInfo->export_size()) & (-4096);
 253                                 sprintf(message, "exports from %s", shortName);
 254                                 add_linkedit(exportPageOffsetStart, exportPageOffsetEnd, message);
 255                         }
 256                         // add binding info
 257                         if ( dyldInfo->bind_size() != 0 ) {
 258                                 uint32_t bindPageOffsetStart = dyldInfo->bind_off() & (-4096);
 259                                 uint32_t bindPageOffsetEnd = (dyldInfo->bind_off() + dyldInfo->bind_size()) & (-4096);
 260                                 sprintf(message, "bindings from %s", shortName);
 261                                 add_linkedit(bindPageOffsetStart, bindPageOffsetEnd, message);
 262                         }
 263                         // add lazy binding info
 264                         if ( dyldInfo->lazy_bind_size() != 0 ) {
 265                                 uint32_t lazybindPageOffsetStart = dyldInfo->lazy_bind_off() & (-4096);
 266                                 uint32_t lazybindPageOffsetEnd = (dyldInfo->lazy_bind_off() + dyldInfo->lazy_bind_size()) & (-4096);
 267                                 sprintf(message, "lazy bindings from %s", shortName);
 268                                 add_linkedit(lazybindPageOffsetStart, lazybindPageOffsetEnd, message);
 269                         }
 270                         // add weak binding info
 271                         if ( dyldInfo->weak_bind_size() != 0 ) {
 272                                 uint32_t weakbindPageOffsetStart = dyldInfo->weak_bind_off() & (-4096);
 273                                 uint32_t weakbindPageOffsetEnd = (dyldInfo->weak_bind_off() + dyldInfo->weak_bind_size()) & (-4096);
 274                                 sprintf(message, "weak bindings from %s", shortName);
 275                                 add_linkedit(weakbindPageOffsetStart, weakbindPageOffsetEnd, message);
 276                         }
 277                 }
 278                 cmd = (const macho_load_command<P>*)(((uint8_t*)cmd)+cmd->cmdsize());
 279         }
 280 }
 281
 282
 283
 284 /*
 285  * This callback is used with dsc_iterator, and called once for each segment in the target shared cache
 286  */
 287 template <typename A>
 288 void segment_callback(const char *dylib, const char *segName, uint64_t offset, uint64_t sizem,
 289                                                                                 uint64_t mappedAddress, uint64_t slide, void *userData) {
 290         typedef typename A::P    P;
 291         typedef typename A::P::E E;
 292         struct seg_callback_args *args = (struct seg_callback_args *)userData;
 293         if (strncmp(segName, "__TEXT", 6) == 0) {
 294                 int target_match = args->target_path ? (strcmp(args->target_path, dylib) == 0) : 0;
 295                 if (!args->target_path || target_match) {
 296                         if (target_match) {
 297                                 args->target_found = 1;
 298                         }
 299                         void *headerAddr = (void*)((long)args->mapped_cache + (long)offset);
 300                         switch (args->op) {
 301                                 case OP_LIST_DEPENDENCIES:
 302                                         list_dependencies<A>(dylib, headerAddr, args->print_dylib_versions);
 303                                         break;
 304                                 case OP_LIST_DYLIBS:
 305                                         print_dylib<A>(dylib, headerAddr, slide, args);
 306                                         break;
 307                                 case OP_LIST_LINKEDIT:
 308                                         process_linkedit<A>(dylib, headerAddr);
 309                                 default:
 310                                         break;
 311                         }
 312                 }
 313         }
 314         else if (strncmp(segName, "__LINKEDIT", 6) == 0) {
 315                 sLinkeditBase = mappedAddress - offset;
 316         }
 317 }
 318
 319
 320
 321
 322 int main (int argc, char **argv) {
 323         struct seg_callback_args args;
 324         const char *shared_cache_path = NULL;
 325         void *mapped_cache;
 326         struct stat statbuf;
 327         int cache_fd;
 328         char c;
 329         bool print_slide_info = false;
 330
 331     args.target_path = NULL;
 332         args.op = OP_NULL;
 333     args.print_uuids = 0;
 334         args.print_vmaddrs = 0;
 335     args.print_dylib_versions = 0;
 336     args.target_found = 0;
 337
 338     for (uint32_t optind = 1; optind < argc; optind++) {
 339         char *opt = argv[optind];
 340         if (opt[0] == '-') {
 341             if (strcmp(opt, "-list") == 0) {
 342                 if (args.op) {
 343                     fprintf(stderr, "Error: select one of -list or -dependents\n");
 344                     usage();
 345                     exit(1);
 346                 }
 347                 args.op = OP_LIST_DYLIBS;
 348             } else if (strcmp(opt, "-dependents") == 0) {
 349                 if (args.op) {
 350                     fprintf(stderr, "Error: select one of -list or -dependents\n");
 351                     usage();
 352                     exit(1);
 353                 }
 354                 if (!(++optind < argc)) {
 355                     fprintf(stderr, "Error: option -depdendents requires an argument\n");
 356                     usage();
 357                     exit(1);
 358                 }
 359                 args.op = OP_LIST_DEPENDENCIES;
 360                 args.target_path = argv[optind];
 361             } else if (strcmp(opt, "-uuid") == 0) {
 362                 args.print_uuids = 1;
 363             } else if (strcmp(opt, "-versions") == 0) {
 364                 args.print_dylib_versions = 1;
 365             } else if (strcmp(opt, "-vmaddr") == 0) {
 366                                 args.print_vmaddrs = 1;
 367                     } else if (strcmp(opt, "-linkedit") == 0) {
 368                 args.op = OP_LIST_LINKEDIT;
 369                         } else if (strcmp(opt, "-slide_info") == 0) {
 370                                 print_slide_info = true;
 371             } else {
 372                 fprintf(stderr, "Error: unrecognized option %s\n", opt);
 373                 usage();
 374                 exit(1);
 375             }
 376         } else {
 377             shared_cache_path = opt;
 378         }
 379     }
 380
 381         if ( !print_slide_info ) {
 382                 if (args.op == OP_NULL) {
 383                         fprintf(stderr, "Error: select one of -list or -dependents\n");
 384                         usage();
 385                         exit(1);
 386                 }
 387
 388                 if (args.print_uuids && args.op != OP_LIST_DYLIBS)
 389                         fprintf(stderr, "Warning: -uuid option ignored outside of -list mode\n");
 390                 if (args.print_vmaddrs && args.op != OP_LIST_DYLIBS)
 391                         fprintf(stderr, "Warning: -vmaddr option ignored outside of -list mode\n");
 392                 if (args.print_dylib_versions && args.op != OP_LIST_DEPENDENCIES)
 393                         fprintf(stderr, "Warning: -versions option ignored outside of -dependents mode\n");
 394
 395                 if (args.op == OP_LIST_DEPENDENCIES && !args.target_path) {
 396                         fprintf(stderr, "Error: -dependents given, but no dylib path specified\n");
 397                         usage();
 398                         exit(1);
 399                 }
 400         }
 401
 402         if (!shared_cache_path)
 403                 shared_cache_path = default_shared_cache_path();
 404
 405         if (stat(shared_cache_path, &statbuf)) {
 406                 fprintf(stderr, "Error: stat failed for dyld shared cache at %s\n", shared_cache_path);
 407                 exit(1);
 408         }
 409
 410         cache_fd = open(shared_cache_path, O_RDONLY);
 411         if (cache_fd < 0) {
 412                 fprintf(stderr, "Error: failed to open shared cache file at %s\n", shared_cache_path);
 413                 exit(1);
 414         }
 415         mapped_cache = mmap(NULL, statbuf.st_size, PROT_READ, MAP_PRIVATE, cache_fd, 0);
 416         if (mapped_cache == MAP_FAILED) {
 417                 fprintf(stderr, "Error: mmap() for shared cache at %s failed, errno=%d\n", shared_cache_path, errno);
 418                 exit(1);
 419         }
 420
 421         if ( print_slide_info ) {
 422                 const dyldCacheHeader<LittleEndian>* header = (dyldCacheHeader<LittleEndian>*)mapped_cache;
 423                 if (   (strcmp(header->magic(), "dyld_v1  x86_64") != 0)
 424                         && (strcmp(header->magic(), "dyld_v1   armv6") != 0)
 425                         && (strcmp(header->magic(), "dyld_v1   armv7") != 0)
 426                         && (strcmp(header->magic(), "dyld_v1  armv7f") != 0)
 427                         && (strcmp(header->magic(), "dyld_v1  armv7k") != 0) ) {
 428                         fprintf(stderr, "Error: unrecognized dyld shared cache magic or arch does not support sliding\n");
 429                         exit(1);
 430                 }
 431                 if ( header->slideInfoOffset() == 0 ) {
 432                         fprintf(stderr, "Error: dyld shared cache does not contain slide info\n");
 433                         exit(1);
 434                 }
 435                 const dyldCacheFileMapping<LittleEndian>* mappings = (dyldCacheFileMapping<LittleEndian>*)((char*)mapped_cache + header->mappingOffset());
 436                 const dyldCacheFileMapping<LittleEndian>* dataMapping = &mappings[1];
 437                 uint64_t dataStartAddress = dataMapping->address();
 438                 uint64_t dataSize = dataMapping->size();
 439                 const dyldCacheSlideInfo<LittleEndian>* slideInfoHeader = (dyldCacheSlideInfo<LittleEndian>*)((char*)mapped_cache+header->slideInfoOffset());
 440                 printf("slide info version=%d\n", slideInfoHeader->version());
 441                 printf("toc_count=%d, data page count=%lld\n", slideInfoHeader->toc_count(), dataSize/4096);
 442                 const dyldCacheSlideInfoEntry* entries = (dyldCacheSlideInfoEntry*)((char*)slideInfoHeader + slideInfoHeader->entries_offset());
 443                 for(int i=0; i < slideInfoHeader->toc_count(); ++i) {
 444                         printf("0x%08llX: [% 5d,% 5d] ", dataStartAddress + i*4096, i, slideInfoHeader->toc(i));
 445                         const dyldCacheSlideInfoEntry* entry = &entries[slideInfoHeader->toc(i)];
 446                         for(int j=0; j < slideInfoHeader->entries_size(); ++j)
 447                                 printf("%02X", entry->bits[j]);
 448                         printf("\n");
 449                 }
 450
 451
 452         }
 453         else {
 454                 segment_callback_t callback;
 455                 if ( strcmp((char*)mapped_cache, "dyld_v1    i386") == 0 )
 456                         callback = segment_callback<x86>;
 457                 else if ( strcmp((char*)mapped_cache, "dyld_v1  x86_64") == 0 )
 458                         callback = segment_callback<x86_64>;
 459                 else if ( strcmp((char*)mapped_cache, "dyld_v1   armv5") == 0 )
 460                         callback = segment_callback<arm>;
 461                 else if ( strcmp((char*)mapped_cache, "dyld_v1   armv6") == 0 )
 462                         callback = segment_callback<arm>;
 463                 else if ( strcmp((char*)mapped_cache, "dyld_v1   armv7") == 0 )
 464                         callback = segment_callback<arm>;
 465                 else if ( strcmp((char*)mapped_cache, "dyld_v1  armv7f") == 0 )
 466                         callback = segment_callback<arm>;
 467                 else if ( strcmp((char*)mapped_cache, "dyld_v1  armv7k") == 0 )
 468                         callback = segment_callback<arm>;
 469                 else {
 470                         fprintf(stderr, "Error: unrecognized dyld shared cache magic.\n");
 471                         exit(1);
 472                 }
 473
 474                 args.mapped_cache = mapped_cache;
 475
 476         #if __BLOCKS__
 477                 // Shim to allow building for the host
 478                 void *argsPtr = &args;
 479                 dyld_shared_cache_iterate_segments_with_slide(mapped_cache,
 480                                                                                    ^(const char* dylib, const char* segName, uint64_t offset, uint64_t size, uint64_t mappedddress, uint64_t slide ) {
 481                                                                                            (callback)(dylib, segName, offset, size, mappedddress, slide, argsPtr);
 482                                                                                    });
 483         #else
 484                 dyld_shared_cache_iterate_segments_with_slide_nb(mapped_cache, callback, &args);
 485         #endif
 486
 487                 if (args.op == OP_LIST_LINKEDIT) {
 488                         // dump -linkedit information
 489                         for (std::map<uint32_t, char*>::iterator it = sPageToContent.begin(); it != sPageToContent.end(); ++it) {
 490                                 printf("0x%0llX %s\n", sLinkeditBase+it->first, it->second);
 491                         }
 492                 }
 493
 494
 495                 if (args.target_path && !args.target_found) {
 496                         fprintf(stderr, "Error: could not find '%s' in the shared cache at\n  %s\n", args.target_path, shared_cache_path);
 497                         exit(1);
 498                 }
 499         }
 500         return 0;
 501 }