dyld3/shared-cache/make_ios_dyld_cache.cpp

   1 /* -*- mode: C++; c-basic-offset: 4; indent-tabs-mode: nil -*-
   2  *
   3  * Copyright (c) 2016 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 <sys/types.h>
  26 #include <sys/stat.h>
  27 #include <sys/mman.h>
  28 #include <mach/mach.h>
  29 #include <mach/mach_time.h>
  30 #include <limits.h>
  31 #include <stdarg.h>
  32 #include <stdio.h>
  33 #include <stdlib.h>
  34 #include <math.h>
  35 #include <fcntl.h>
  36 #include <dlfcn.h>
  37 #include <signal.h>
  38 #include <errno.h>
  39 #include <assert.h>
  40 #include <sys/uio.h>
  41 #include <unistd.h>
  42 #include <sys/param.h>
  43 #include <sys/sysctl.h>
  44 #include <sys/resource.h>
  45 #include <dirent.h>
  46 #include <rootless.h>
  47 #include <dscsym.h>
  48 #include <dispatch/dispatch.h>
  49 #include <pthread/pthread.h>
  50
  51 #include <algorithm>
  52 #include <vector>
  53 #include <unordered_set>
  54 #include <unordered_set>
  55 #include <iostream>
  56 #include <fstream>
  57
  58 #include "MachOParser.h"
  59 #include "FileUtils.h"
  60 #include "StringUtils.h"
  61 #include "DyldSharedCache.h"
  62
  63
  64
  65 struct MappedMachOsByCategory
  66 {
  67     std::string                                 archName;
  68     std::vector<DyldSharedCache::MappedMachO>   dylibsForCache;
  69     std::vector<DyldSharedCache::MappedMachO>   otherDylibsAndBundles;
  70     std::vector<DyldSharedCache::MappedMachO>   mainExecutables;
  71 };
  72
  73 static bool verbose = false;
  74
  75
  76 static bool addIfMachO(const std::string& buildRootPath, const std::string& runtimePath, const struct stat& statBuf, dyld3::Platform platform, std::vector<MappedMachOsByCategory>& files)
  77 {
  78     // read start of file to determine if it is mach-o or a fat file
  79     std::string fullPath = buildRootPath + runtimePath;
  80     int fd = ::open(fullPath.c_str(), O_RDONLY);
  81     if ( fd < 0 )
  82         return false;
  83     bool result = false;
  84     const void* wholeFile = ::mmap(NULL, (size_t)statBuf.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
  85     if ( wholeFile != MAP_FAILED ) {
  86         Diagnostics diag;
  87         bool usedWholeFile = false;
  88         for (MappedMachOsByCategory& file : files) {
  89             size_t sliceOffset;
  90             size_t sliceLength;
  91             bool fatButMissingSlice;
  92             const void* slice = MAP_FAILED;
  93             if ( dyld3::FatUtil::isFatFileWithSlice(diag, wholeFile, statBuf.st_size, file.archName, sliceOffset, sliceLength, fatButMissingSlice) ) {
  94                 slice = ::mmap(NULL, sliceLength, PROT_READ, MAP_PRIVATE, fd, sliceOffset);
  95                 if ( slice != MAP_FAILED ) {
  96                     //fprintf(stderr, "mapped slice at %p size=0x%0lX, offset=0x%0lX for %s\n", p, len, offset, fullPath.c_str());
  97                     if ( !dyld3::MachOParser::isValidMachO(diag, file.archName, platform, slice, sliceLength, fullPath.c_str(), false) ) {
  98                         ::munmap((void*)slice, sliceLength);
  99                         slice = MAP_FAILED;
 100                     }
 101                 }
 102             }
 103             else if ( !fatButMissingSlice && dyld3::MachOParser::isValidMachO(diag, file.archName, platform, wholeFile, statBuf.st_size, fullPath.c_str(), false) ) {
 104                 slice           = wholeFile;
 105                 sliceLength     = statBuf.st_size;
 106                 sliceOffset     = 0;
 107                 usedWholeFile   = true;
 108                 //fprintf(stderr, "mapped whole file at %p size=0x%0lX for %s\n", p, len, inputPath.c_str());
 109             }
 110             if ( slice != MAP_FAILED ) {
 111                 const mach_header* mh = (mach_header*)slice;
 112                 dyld3::MachOParser parser(mh);
 113                 if ( parser.platform() != platform ) {
 114                     fprintf(stderr, "skipped wrong platform binary: %s\n", fullPath.c_str());
 115                     result = false;
 116                 }
 117                 else {
 118                     bool sip = true; // assume anything found in the simulator runtime is a platform binary
 119                     if ( parser.isDynamicExecutable() ) {
 120                         bool issetuid = (statBuf.st_mode & (S_ISUID|S_ISGID));
 121                         file.mainExecutables.emplace_back(runtimePath, mh, sliceLength, issetuid, sip, sliceOffset, statBuf.st_mtime, statBuf.st_ino);
 122                     }
 123                     else {
 124                         if ( parser.canBePlacedInDyldCache(runtimePath) ) {
 125                             file.dylibsForCache.emplace_back(runtimePath, mh, sliceLength, false, sip, sliceOffset, statBuf.st_mtime, statBuf.st_ino);
 126                         }
 127                         else {
 128                             file.otherDylibsAndBundles.emplace_back(runtimePath, mh, sliceLength, false, sip, sliceOffset, statBuf.st_mtime, statBuf.st_ino);
 129                         }
 130                     }
 131                     result = true;
 132                 }
 133             }
 134         }
 135         if ( !usedWholeFile )
 136             ::munmap((void*)wholeFile, statBuf.st_size);
 137     }
 138     ::close(fd);
 139     return result;
 140 }
 141
 142
 143 static bool parsePathsFile(const std::string& filePath, std::vector<std::string>& paths) {
 144     std::ifstream myfile( filePath );
 145     if ( myfile.is_open() ) {
 146     std::string line;
 147         while ( std::getline(myfile, line) ) {
 148             size_t pos = line.find('#');
 149             if ( pos != std::string::npos )
 150                 line.resize(pos);
 151             while ( line.size() != 0 && isspace(line.back()) ) {
 152                 line.pop_back();
 153             }
 154             if ( !line.empty() )
 155                 paths.push_back(line);
 156         }
 157         myfile.close();
 158         return true;
 159     }
 160     return false;
 161 }
 162
 163
 164 static void mapAllFiles(const std::string& dylibsRootDir, const std::vector<std::string>& paths, dyld3::Platform platform, std::vector<MappedMachOsByCategory>& files)
 165 {
 166     for (const std::string& runtimePath : paths) {
 167         std::string fullPath = dylibsRootDir + runtimePath;
 168         struct stat statBuf;
 169         if ( (stat(fullPath.c_str(), &statBuf) != 0) || !addIfMachO(dylibsRootDir, runtimePath, statBuf, platform, files) )
 170              fprintf(stderr, "could not load: %s\n", fullPath.c_str());
 171     }
 172 }
 173
 174
 175
 176 inline uint32_t absolutetime_to_milliseconds(uint64_t abstime)
 177 {
 178     return (uint32_t)(abstime/1000/1000);
 179 }
 180
 181
 182 #define TERMINATE_IF_LAST_ARG( s )      \
 183     do {                                \
 184         if ( i == argc - 1 ) {          \
 185             fprintf(stderr, s );        \
 186             return 1;                   \
 187         }                               \
 188     } while ( 0 )
 189
 190 int main(int argc, const char* argv[])
 191 {
 192     std::string                     rootPath;
 193     std::string                     dylibListFile;
 194     bool                            force = false;
 195     std::string                     cacheDir;
 196     std::string                     dylibsList;
 197     std::unordered_set<std::string> archStrs;
 198
 199     dyld3::Platform platform = dyld3::Platform::iOS;
 200
 201     // parse command line options
 202     for (int i = 1; i < argc; ++i) {
 203         const char* arg = argv[i];
 204         if (strcmp(arg, "-debug") == 0) {
 205             verbose = true;
 206         }
 207         else if (strcmp(arg, "-verbose") == 0) {
 208             verbose = true;
 209         }
 210         else if (strcmp(arg, "-tvOS") == 0) {
 211             platform = dyld3::Platform::tvOS;
 212         }
 213         else if (strcmp(arg, "-iOS") == 0) {
 214             platform = dyld3::Platform::iOS;
 215         }
 216         else if (strcmp(arg, "-watchOS") == 0) {
 217             platform = dyld3::Platform::watchOS;
 218         }
 219         else if ( strcmp(arg, "-root") == 0 ) {
 220             TERMINATE_IF_LAST_ARG("-root missing path argument\n");
 221             rootPath = argv[++i];
 222         }
 223         else if ( strcmp(arg, "-dylibs_list") == 0 ) {
 224             TERMINATE_IF_LAST_ARG("-dylibs_list missing path argument\n");
 225             dylibsList = argv[++i];
 226         }
 227         else if (strcmp(arg, "-cache_dir") == 0) {
 228             TERMINATE_IF_LAST_ARG("-cache_dir missing path argument\n");
 229             cacheDir = argv[++i];
 230         }
 231         else if (strcmp(arg, "-arch") == 0) {
 232             TERMINATE_IF_LAST_ARG("-arch missing argument\n");
 233             archStrs.insert(argv[++i]);
 234         }
 235         else if (strcmp(arg, "-force") == 0) {
 236             force = true;
 237         }
 238         else {
 239             //usage();
 240             fprintf(stderr, "update_dyld_sim_shared_cache: unknown option: %s\n", arg);
 241             return 1;
 242         }
 243     }
 244
 245     if ( cacheDir.empty() ) {
 246         fprintf(stderr, "missing -cache_dir <path> option to specify directory in which to write cache file(s)\n");
 247         return 1;
 248     }
 249
 250     if ( rootPath.empty() ) {
 251         fprintf(stderr, "missing -runtime_dir <path> option to specify directory which is root of simulator runtime)\n");
 252         return 1;
 253     }
 254     else {
 255         // canonicalize rootPath
 256         char resolvedPath[PATH_MAX];
 257         if ( realpath(rootPath.c_str(), resolvedPath) != NULL ) {
 258             rootPath = resolvedPath;
 259         }
 260         if ( rootPath.back() != '/' )
 261             rootPath = rootPath + "/";
 262     }
 263
 264     int err = mkpath_np(cacheDir.c_str(), S_IRWXU | S_IRGRP|S_IXGRP | S_IROTH|S_IXOTH);
 265     if ( (err != 0) && (err != EEXIST) ) {
 266         fprintf(stderr, "mkpath_np fail: %d", err);
 267         return 1;
 268     }
 269
 270     if ( archStrs.empty() ) {
 271         switch ( platform ) {
 272             case dyld3::Platform::iOS:
 273             case dyld3::Platform::tvOS:
 274                 archStrs.insert("arm64");
 275                 break;
 276             case dyld3::Platform::watchOS:
 277                 archStrs.insert("armv7k");
 278                 break;
 279              case dyld3::Platform::unknown:
 280              case dyld3::Platform::macOS:
 281                 assert(0 && "macOS not support with this tool");
 282                 break;
 283         }
 284     }
 285
 286     uint64_t t1 = mach_absolute_time();
 287
 288     // find all mach-o files for requested architectures
 289     std::vector<MappedMachOsByCategory> allFileSets;
 290     if ( archStrs.count("arm64") )
 291         allFileSets.push_back({"arm64"});
 292     if ( archStrs.count("armv7k") )
 293         allFileSets.push_back({"armv7k"});
 294     std::vector<std::string> paths;
 295     parsePathsFile(dylibsList, paths);
 296     mapAllFiles(rootPath, paths, platform, allFileSets);
 297
 298     uint64_t t2 = mach_absolute_time();
 299
 300     fprintf(stderr, "time to scan file system and construct lists of mach-o files: %ums\n", absolutetime_to_milliseconds(t2-t1));
 301
 302     // build all caches in parallel
 303     __block bool cacheBuildFailure = false;
 304     dispatch_apply(allFileSets.size(), dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0), ^(size_t index) {
 305         const MappedMachOsByCategory& fileSet = allFileSets[index];
 306         const std::string outFile = cacheDir + "/dyld_shared_cache_" + fileSet.archName;
 307
 308         fprintf(stderr, "make %s cache with %lu dylibs, %lu other dylibs, %lu programs\n", fileSet.archName.c_str(), fileSet.dylibsForCache.size(), fileSet.otherDylibsAndBundles.size(), fileSet.mainExecutables.size());
 309
 310         // build cache new cache file
 311         DyldSharedCache::CreateOptions options;
 312         options.archName                     = fileSet.archName;
 313         options.platform                     = platform;
 314         options.excludeLocalSymbols          = true;
 315         options.optimizeStubs                = false;
 316         options.optimizeObjC                 = true;
 317         options.codeSigningDigestMode = (platform() == dyld3::Platform::watchOS) ?
 318                                         DyldSharedCache::Agile : DyldSharedCache::SHA256only;
 319         options.dylibsRemovedDuringMastering = true;
 320         options.inodesAreSameAsRuntime       = false;
 321         options.cacheSupportsASLR            = true;
 322         options.forSimulator                 = false;
 323         options.verbose                      = verbose;
 324         options.evictLeafDylibsOnOverflow    = false;
 325         options.pathPrefixes                 = { rootPath };
 326         DyldSharedCache::CreateResults results = DyldSharedCache::create(options, fileSet.dylibsForCache, fileSet.otherDylibsAndBundles, fileSet.mainExecutables);
 327
 328         // print any warnings
 329         for (const std::string& warn : results.warnings) {
 330             fprintf(stderr, "update_dyld_sim_shared_cache: warning: %s %s\n", fileSet.archName.c_str(), warn.c_str());
 331         }
 332         if ( !results.errorMessage.empty() ) {
 333             // print error (if one)
 334             fprintf(stderr, "update_dyld_sim_shared_cache: %s\n", results.errorMessage.c_str());
 335             cacheBuildFailure = true;
 336         }
 337         else {
 338             // save new cache file to disk and write new .map file
 339             assert(results.cacheContent != nullptr);
 340             if ( !safeSave(results.cacheContent, results.cacheLength, outFile) )
 341                 cacheBuildFailure = true;
 342             if ( !cacheBuildFailure ) {
 343                 std::string mapStr = results.cacheContent->mapFile();
 344                 std::string outFileMap = cacheDir + "/dyld_shared_cache_" + fileSet.archName + ".map";
 345                 safeSave(mapStr.c_str(), mapStr.size(), outFileMap);
 346             }
 347             // free created cache buffer
 348             vm_deallocate(mach_task_self(), (vm_address_t)results.cacheContent, results.cacheLength);
 349         }
 350     });
 351
 352     // we could unmap all input files, but tool is about to quit
 353
 354     return (cacheBuildFailure ? 1 : 0);
 355 }
 356