Loading...
common/DyldSharedCache.cpp /dev/null dyld-941.4
--- /dev/null
+++ dyld/dyld-941.4/common/DyldSharedCache.cpp
@@ -0,0 +1,2235 @@
+/* -*- mode: C++; c-basic-offset: 4; indent-tabs-mode: nil -*-
+ *
+ * Copyright (c) 2014 Apple Inc. All rights reserved.
+ *
+ * @APPLE_LICENSE_HEADER_START@
+ *
+ * This file contains Original Code and/or Modifications of Original Code
+ * as defined in and that are subject to the Apple Public Source License
+ * Version 2.0 (the 'License'). You may not use this file except in
+ * compliance with the License. Please obtain a copy of the License at
+ * http://www.opensource.apple.com/apsl/ and read it before using this
+ * file.
+ *
+ * The Original Code and all software distributed under the License are
+ * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
+ * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
+ * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
+ * Please see the License for the specific language governing rights and
+ * limitations under the License.
+ *
+ * @APPLE_LICENSE_HEADER_END@
+ */
+
+
+#include <dirent.h>
+#include <sys/errno.h>
+#include <sys/fcntl.h>
+#include <sys/mman.h>
+#include <sys/param.h>
+#include <mach/mach.h>
+#include <mach-o/loader.h>
+#include <mach-o/fat.h>
+#include <mach-o/dyld_priv.h>
+#include <assert.h>
+#include <unistd.h>
+#include <dlfcn.h>
+
+#if BUILDING_CACHE_BUILDER
+#include <set>
+#include <string>
+#include <vector>
+#include <unordered_map>
+#include <unordered_set>
+#include "SharedCacheBuilder.h"
+#include "FileUtils.h"
+#endif
+
+#define NO_ULEB
+#include "MachOLoaded.h"
+#include "ClosureFileSystemPhysical.h"
+#include "DyldSharedCache.h"
+#include "Trie.hpp"
+#include "StringUtils.h"
+#include "PrebuiltLoader.h"
+
+#include "objc-shared-cache.h"
+
+#if !(BUILDING_LIBDYLD || BUILDING_DYLD)
+#include "JSONWriter.h"
+#include <sstream>
+#endif
+
+using dyld4::PrebuiltLoader;
+using dyld4::PrebuiltLoaderSet;
+#if (BUILDING_LIBDYLD || BUILDING_DYLD)
+VIS_HIDDEN bool gEnableSharedCacheDataConst = false;
+#endif
+
+
+#if BUILDING_CACHE_BUILDER
+DyldSharedCache::CreateResults DyldSharedCache::create(const CreateOptions&               options,
+                                                       const dyld3::closure::FileSystem&  fileSystem,
+                                                       const std::vector<MappedMachO>&    dylibsToCache,
+                                                       const std::vector<MappedMachO>&    otherOsDylibs,
+                                                       const std::vector<MappedMachO>&    osExecutables)
+{
+    CreateResults       results;
+    SharedCacheBuilder  cache(options, fileSystem);
+    if (!cache.errorMessage().empty()) {
+        results.errorMessage = cache.errorMessage();
+        return results;
+    }
+
+    std::vector<FileAlias> aliases;
+    switch ( options.platform ) {
+        case dyld3::Platform::iOS:
+        case dyld3::Platform::watchOS:
+        case dyld3::Platform::tvOS:
+            // FIXME: embedded cache builds should be getting aliases from manifest
+            aliases.push_back({"/System/Library/Frameworks/IOKit.framework/Versions/A/IOKit", "/System/Library/Frameworks/IOKit.framework/IOKit"});
+            aliases.push_back({"/usr/lib/libstdc++.6.dylib",                                  "/usr/lib/libstdc++.dylib"});
+            aliases.push_back({"/usr/lib/libstdc++.6.dylib",                                  "/usr/lib/libstdc++.6.0.9.dylib"});
+            aliases.push_back({"/usr/lib/libz.1.dylib",                                       "/usr/lib/libz.dylib"});
+            aliases.push_back({"/usr/lib/libSystem.B.dylib",                                  "/usr/lib/libSystem.dylib"});
+            aliases.push_back({"/System/Library/Frameworks/Foundation.framework/Foundation",  "/usr/lib/libextension.dylib"}); // <rdar://44315703>
+            break;
+        default:
+            break;
+    }
+
+    cache.build(dylibsToCache, otherOsDylibs, osExecutables, aliases);
+
+    results.warnings       = cache.warnings();
+    results.evictions      = cache.evictions();
+    if ( cache.errorMessage().empty() ) {
+        if ( !options.outputFilePath.empty() )  {
+            // write cache file, if path non-empty
+            cache.writeFile(options.outputFilePath);
+        }
+        if ( !options.outputMapFilePath.empty() ) {
+            // write map file, if path non-empty
+            cache.writeMapFile(options.outputMapFilePath);
+        }
+    }
+    results.errorMessage = cache.errorMessage();
+    cache.deleteBuffer();
+    return results;
+}
+
+bool DyldSharedCache::verifySelfContained(std::vector<MappedMachO>& dylibsToCache,
+                                          std::unordered_set<std::string>& badZippered,
+                                          MappedMachO (^loader)(const std::string& runtimePath, Diagnostics& diag),
+                                          std::vector<std::pair<DyldSharedCache::MappedMachO, std::set<std::string>>>& rejected)
+{
+    // build map of dylibs
+    __block std::map<std::string, std::set<std::string>> badDylibs;
+    __block std::set<std::string> knownDylibs;
+    for (const DyldSharedCache::MappedMachO& dylib : dylibsToCache) {
+        std::set<std::string> reasons;
+        if ( dylib.mh->canBePlacedInDyldCache(dylib.runtimePath.c_str(), ^(const char* msg) { badDylibs[dylib.runtimePath].insert(msg);}) ) {
+            knownDylibs.insert(dylib.runtimePath);
+            knownDylibs.insert(dylib.mh->installName());
+        } else {
+            badDylibs[dylib.runtimePath].insert("");
+        }
+    }
+
+    // check all dependencies to assure every dylib in cache only depends on other dylibs in cache
+    __block std::set<std::string> missingWeakDylibs;
+    __block bool doAgain = true;
+    while ( doAgain ) {
+        __block std::vector<DyldSharedCache::MappedMachO> foundMappings;
+        doAgain = false;
+        // scan dylib list making sure all dependents are in dylib list
+        for (const DyldSharedCache::MappedMachO& dylib : dylibsToCache) {
+            if ( badDylibs.count(dylib.runtimePath) != 0 )
+                continue;
+            dylib.mh->forEachDependentDylib(^(const char* loadPath, bool isWeak, bool isReExport, bool isUpward, uint32_t compatVersion, uint32_t curVersion, bool& stop) {
+                if ( isWeak && (missingWeakDylibs.count(loadPath) != 0) )
+                    return;
+                if ( knownDylibs.count(loadPath) == 0 ) {
+                    doAgain = true;
+                    if ( badZippered.count(loadPath) != 0 ) {
+                        badDylibs[dylib.runtimePath].insert("");
+                        knownDylibs.erase(dylib.runtimePath);
+                        knownDylibs.erase(dylib.mh->installName());
+                        badZippered.insert(dylib.runtimePath);
+                        badZippered.insert(dylib.mh->installName());
+                        return;
+                    }
+                    Diagnostics diag;
+                    MappedMachO foundMapping;
+                    if ( badDylibs.count(loadPath) == 0 )
+                        foundMapping = loader(loadPath, diag);
+                    if ( foundMapping.length == 0 ) {
+                        // We allow weakly linked dylibs to be missing only if they are not present on disk
+                        // The shared cache doesn't contain enough information to patch them in later if they are
+                        // found on disk, so we don't want to pull something in to cache and cut it off from a dylib it
+                        // could have used.
+                        if ( isWeak ) {
+                            missingWeakDylibs.insert(loadPath);
+                            return;
+                        }
+
+                        if (diag.hasError())
+                            badDylibs[dylib.runtimePath].insert(diag.errorMessage());
+                        else
+                            badDylibs[dylib.runtimePath].insert(std::string("Could not find dependency '") + loadPath +"'");
+                        knownDylibs.erase(dylib.runtimePath);
+                        knownDylibs.erase(dylib.mh->installName());
+                    }
+                    else {
+                        std::set<std::string> reasons;
+                        if ( foundMapping.mh->canBePlacedInDyldCache(foundMapping.runtimePath.c_str(), ^(const char* msg) { badDylibs[foundMapping.runtimePath].insert(msg);})) {
+                            // see if existing mapping was returned
+                            bool alreadyInVector = false;
+                            for (const MappedMachO& existing : dylibsToCache) {
+                                if ( existing.mh == foundMapping.mh ) {
+                                    alreadyInVector = true;
+                                    break;
+                                }
+                            }
+                            if ( !alreadyInVector )
+                                foundMappings.push_back(foundMapping);
+                            knownDylibs.insert(loadPath);
+                            knownDylibs.insert(foundMapping.runtimePath);
+                            knownDylibs.insert(foundMapping.mh->installName());
+                        } else {
+                            badDylibs[dylib.runtimePath].insert("");
+                        }
+                   }
+                }
+            });
+        }
+        dylibsToCache.insert(dylibsToCache.end(), foundMappings.begin(), foundMappings.end());
+        // remove bad dylibs
+        const auto badDylibsCopy = badDylibs;
+        dylibsToCache.erase(std::remove_if(dylibsToCache.begin(), dylibsToCache.end(), [&](const DyldSharedCache::MappedMachO& dylib) {
+            auto i = badDylibsCopy.find(dylib.runtimePath);
+            if ( i !=  badDylibsCopy.end()) {
+                // Only add the warning if we are not a bad zippered dylib
+                if ( badZippered.count(dylib.runtimePath) == 0 )
+                    rejected.push_back(std::make_pair(dylib, i->second));
+                return true;
+             }
+             else {
+                return false;
+             }
+        }), dylibsToCache.end());
+    }
+
+    return badDylibs.empty();
+}
+#endif
+
+template<typename T>
+const T DyldSharedCache::getAddrField(uint64_t addr) const {
+    uint64_t slide = (uint64_t)this - unslidLoadAddress();
+    return (const T)(addr + slide);
+}
+
+uint64_t DyldSharedCache::getCodeSignAddress() const
+{
+    auto mappings = (const dyld_cache_mapping_info*)((uint8_t*)this + header.mappingOffset);
+    return mappings[header.mappingCount-1].address + mappings[header.mappingCount-1].size;
+}
+
+void DyldSharedCache::forEachRegion(void (^handler)(const void* content, uint64_t vmAddr, uint64_t size,
+                                                    uint32_t initProt, uint32_t maxProt, uint64_t flags,
+                                                    bool& stopRegion)) const
+{
+    // <rdar://problem/49875993> sanity check cache header
+    if ( strncmp(header.magic, "dyld_v1", 7) != 0 )
+        return;
+    if ( header.mappingOffset > 1024 )
+        return;
+    if ( header.mappingCount > 20 )
+        return;
+    if ( header.mappingOffset <= __offsetof(dyld_cache_header, mappingWithSlideOffset) ) {
+        const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+        const dyld_cache_mapping_info* mappingsEnd = &mappings[header.mappingCount];
+        for (const dyld_cache_mapping_info* m=mappings; m < mappingsEnd; ++m) {
+            bool stop = false;
+            handler((char*)this + m->fileOffset, m->address, m->size, m->initProt, m->maxProt, 0, stop);
+            if ( stop )
+                return;
+        }
+    } else {
+        const dyld_cache_mapping_and_slide_info* mappings = (const dyld_cache_mapping_and_slide_info*)((char*)this + header.mappingWithSlideOffset);
+        const dyld_cache_mapping_and_slide_info* mappingsEnd = &mappings[header.mappingCount];
+        for (const dyld_cache_mapping_and_slide_info* m=mappings; m < mappingsEnd; ++m) {
+            bool stop = false;
+            handler((char*)this + m->fileOffset, m->address, m->size, m->initProt, m->maxProt, m->flags, stop);
+            if ( stop )
+                return;
+        }
+    }
+}
+
+void DyldSharedCache::forEachRange(void (^handler)(const char* mappingName,
+                                                   uint64_t unslidVMAddr, uint64_t vmSize,
+                                                   uint32_t cacheFileIndex, uint64_t fileOffset,
+                                                   uint32_t initProt, uint32_t maxProt,
+                                                   bool& stopRange),
+                                   void (^subCacheHandler)(const DyldSharedCache* subCache, uint32_t cacheFileIndex)) const
+{
+    __block uint32_t cacheFileIndex = 0;
+    forEachCache(^(const DyldSharedCache *cache, bool& stopCache) {
+        cache->forEachRegion(^(const void *content, uint64_t unslidVMAddr, uint64_t size,
+                        uint32_t initProt, uint32_t maxProt, uint64_t flags, bool& stopRegion) {
+            const char* mappingName = "";
+            if ( maxProt & VM_PROT_EXECUTE ) {
+                mappingName = "__TEXT";
+            } else if ( maxProt & VM_PROT_WRITE ) {
+                if ( flags & DYLD_CACHE_MAPPING_AUTH_DATA ) {
+                    if ( flags & DYLD_CACHE_MAPPING_DIRTY_DATA )
+                        mappingName = "__AUTH_DIRTY";
+                    else if ( flags & DYLD_CACHE_MAPPING_CONST_DATA )
+                        mappingName = "__AUTH_CONST";
+                    else
+                        mappingName = "__AUTH";
+                } else {
+                    if ( flags & DYLD_CACHE_MAPPING_DIRTY_DATA )
+                        mappingName = "__DATA_DIRTY";
+                    else if ( flags & DYLD_CACHE_MAPPING_CONST_DATA )
+                        mappingName = "__DATA_CONST";
+                    else
+                        mappingName = "__DATA";
+                }
+            }
+            else if ( maxProt & VM_PROT_READ ) {
+                mappingName = "__LINKEDIT";
+            } else {
+                mappingName = "*unknown*";
+            }
+            uint64_t fileOffset = (uint8_t*)content - (uint8_t*)cache;
+            bool stop = false;
+            handler(mappingName, unslidVMAddr, size, cacheFileIndex, fileOffset, initProt, maxProt, stop);
+            if ( stop ) {
+                stopRegion = true;
+                stopCache = true;
+                return;
+            }
+        });
+
+        if ( stopCache )
+            return;
+
+        if ( subCacheHandler != nullptr )
+            subCacheHandler(cache, cacheFileIndex);
+
+        ++cacheFileIndex;
+    });
+}
+
+void DyldSharedCache::forEachCache(void (^handler)(const DyldSharedCache *cache, bool& stopCache)) const
+{
+    // Always start with the current file
+    bool stop = false;
+    handler(this, stop);
+    if ( stop )
+        return;
+
+    // We may or may not be followed by sub caches.
+    if ( header.mappingOffset <= __offsetof(dyld_cache_header, subCacheArrayCount) )
+        return;
+
+    const dyld_subcache_entry* subCaches = (const dyld_subcache_entry*)((uint8_t*)this + header.subCacheArrayOffset);
+    for (uint32_t i = 0; i != header.subCacheArrayCount; ++i) {
+        const DyldSharedCache* cache = (const DyldSharedCache*)((uint8_t*)this + subCaches[i].cacheVMOffset);
+        handler(cache, stop);
+        if ( stop )
+            return;
+    }
+}
+
+uint32_t DyldSharedCache::numSubCaches() const {
+    // We may or may not be followed by sub caches.
+    if ( header.mappingOffset <= __offsetof(dyld_cache_header, subCacheArrayCount) )
+        return 0;
+
+    return header.subCacheArrayCount;
+}
+
+bool DyldSharedCache::inCache(const void* addr, size_t length, bool& readOnly) const
+{
+    // quick out if before start of cache
+    if ( addr < this )
+        return false;
+
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    uintptr_t slide = (uintptr_t)this - (uintptr_t)(mappings[0].address);
+    uintptr_t unslidStart = (uintptr_t)addr - slide;
+
+    // quick out if after end of cache
+    const dyld_cache_mapping_info* lastMapping = &mappings[header.mappingCount - 1];
+    if ( unslidStart > (lastMapping->address + lastMapping->size) )
+        return false;
+
+    // walk cache regions
+    const dyld_cache_mapping_info* mappingsEnd = &mappings[header.mappingCount];
+    uintptr_t unslidEnd = unslidStart + length;
+    for (const dyld_cache_mapping_info* m=mappings; m < mappingsEnd; ++m) {
+        if ( (unslidStart >= m->address) && (unslidEnd < (m->address+m->size)) ) {
+            readOnly = ((m->initProt & VM_PROT_WRITE) == 0);
+            return true;
+        }
+    }
+
+    return false;
+}
+
+bool DyldSharedCache::isAlias(const char* path) const {
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    uintptr_t slide = (uintptr_t)this - (uintptr_t)(mappings[0].address);
+    // paths for aliases are store between cache header and first segment
+    return path < ((char*)mappings[0].address + slide);
+}
+
+uint32_t DyldSharedCache::imagesCount() const {
+    if ( header.mappingOffset >= __offsetof(dyld_cache_header, imagesCount) ) {
+        return header.imagesCount;
+    }
+    return header.imagesCountOld;
+}
+
+const dyld_cache_image_info* DyldSharedCache::images() const {
+    if ( header.mappingOffset >= __offsetof(dyld_cache_header, imagesCount) ) {
+        return (dyld_cache_image_info*)((char*)this + header.imagesOffset);
+    }
+    return (dyld_cache_image_info*)((char*)this + header.imagesOffsetOld);
+}
+
+void DyldSharedCache::forEachImage(void (^handler)(const mach_header* mh, const char* installName)) const
+{
+    const dyld_cache_image_info*   dylibs   = images();
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    if ( mappings[0].fileOffset != 0 )
+        return;
+    uint64_t firstImageOffset = 0;
+    uint64_t firstRegionAddress = mappings[0].address;
+    for (uint32_t i=0; i < imagesCount(); ++i) {
+        const char* dylibPath  = (char*)this + dylibs[i].pathFileOffset;
+        uint64_t offset = dylibs[i].address - firstRegionAddress;
+        if ( firstImageOffset == 0 )
+            firstImageOffset = offset;
+        // skip over aliases.  This is no longer valid in newer caches.  They store aliases only in the trie
+#if 0
+        if ( dylibs[i].pathFileOffset < firstImageOffset)
+            continue;
+#endif
+        const mach_header* mh = (mach_header*)((char*)this + offset);
+        handler(mh, dylibPath);
+    }
+}
+
+void DyldSharedCache::forEachDylib(void (^handler)(const dyld3::MachOAnalyzer* ma, const char* installName, uint32_t imageIndex, uint64_t inode, uint64_t mtime, bool& stop)) const
+{
+    const dyld_cache_image_info*   dylibs   = (dyld_cache_image_info*)((char*)this + header.imagesOffset);
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    if ( mappings[0].fileOffset != 0 )
+        return;
+    uint64_t firstImageOffset = 0;
+    uint64_t firstRegionAddress = mappings[0].address;
+    for (uint32_t i=0; i < header.imagesCount; ++i) {
+        uint64_t offset = dylibs[i].address - firstRegionAddress;
+        if ( firstImageOffset == 0 )
+            firstImageOffset = offset;
+        const char* dylibPath  = (char*)this + dylibs[i].pathFileOffset;
+        const mach_header* mh = (mach_header*)((char*)this + offset);
+        bool stop = false;
+        handler((const dyld3::MachOAnalyzer*)mh, dylibPath, i, dylibs[i].inode, dylibs[i].modTime, stop);
+        if ( stop )
+            break;
+    }
+}
+
+void DyldSharedCache::forEachImageEntry(void (^handler)(const char* path, uint64_t mTime, uint64_t inode)) const
+{
+    const dyld_cache_image_info*   dylibs   = images();
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    if ( mappings[0].fileOffset != 0 )
+        return;
+    uint64_t firstImageOffset = 0;
+    uint64_t firstRegionAddress = mappings[0].address;
+    for (uint32_t i=0; i < imagesCount(); ++i) {
+        const char* dylibPath  = (char*)this + dylibs[i].pathFileOffset;
+        uint64_t offset = dylibs[i].address - firstRegionAddress;
+        if ( firstImageOffset == 0 )
+            firstImageOffset = offset;
+        // skip over aliases.  This is no longer valid in newer caches.  They store aliases only in the trie
+#if 0
+        if ( dylibs[i].pathFileOffset < firstImageOffset)
+            continue;
+#endif
+        handler(dylibPath, dylibs[i].modTime, dylibs[i].inode);
+    }
+}
+
+const bool DyldSharedCache::hasLocalSymbolsInfo() const
+{
+    return (header.localSymbolsOffset != 0 && header.mappingOffset > offsetof(dyld_cache_header,localSymbolsSize));
+}
+
+const bool DyldSharedCache::hasLocalSymbolsInfoFile() const
+{
+    if ( header.mappingOffset > __offsetof(dyld_cache_header, symbolFileUUID) )
+        return !uuid_is_null(header.symbolFileUUID);
+
+    // Old cache file
+    return false;
+}
+
+const void* DyldSharedCache::getLocalNlistEntries(const dyld_cache_local_symbols_info* localInfo) {
+    return (uint8_t*)localInfo + localInfo->nlistOffset;
+}
+
+const void* DyldSharedCache::getLocalNlistEntries() const
+{
+    // check for cache without local symbols info
+    if (!this->hasLocalSymbolsInfo())
+        return nullptr;
+    const auto localInfo = (dyld_cache_local_symbols_info*)((uint8_t*)this + header.localSymbolsOffset);
+    return getLocalNlistEntries(localInfo);
+}
+
+const uint32_t DyldSharedCache::getLocalNlistCount() const
+{
+    // check for cache without local symbols info
+     if (!this->hasLocalSymbolsInfo())
+        return 0;
+    const auto localInfo = (dyld_cache_local_symbols_info*)((uint8_t*)this + header.localSymbolsOffset);
+    return localInfo->nlistCount;
+}
+
+const char* DyldSharedCache::getLocalStrings(const dyld_cache_local_symbols_info* localInfo)
+{
+    return (char*)localInfo + localInfo->stringsOffset;
+}
+
+const char* DyldSharedCache::getLocalStrings() const
+{
+    // check for cache without local symbols info
+     if (!this->hasLocalSymbolsInfo())
+        return nullptr;
+    const auto localInfo = (dyld_cache_local_symbols_info*)((uint8_t*)this + header.localSymbolsOffset);
+    return getLocalStrings(localInfo);
+}
+
+const uint32_t DyldSharedCache::getLocalStringsSize() const
+{
+    // check for cache without local symbols info
+     if (!this->hasLocalSymbolsInfo())
+        return 0;
+    const auto localInfo = (dyld_cache_local_symbols_info*)((uint8_t*)this + header.localSymbolsOffset);
+    return localInfo->stringsSize;
+}
+
+const dyld3::MachOFile* DyldSharedCache::getImageFromPath(const char* dylibPath) const
+{
+    const dyld_cache_image_info*   dylibs   = images();
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    uint32_t dyldCacheImageIndex;
+    if ( hasImagePath(dylibPath, dyldCacheImageIndex) )
+        return (dyld3::MachOFile*)((uint8_t*)this + dylibs[dyldCacheImageIndex].address - mappings[0].address);
+    return nullptr;
+}
+
+void DyldSharedCache::forEachLocalSymbolEntry(void (^handler)(uint64_t dylibOffset, uint32_t nlistStartIndex, uint32_t nlistCount, bool& stop)) const
+{
+    // check for cache without local symbols info
+    if (!this->hasLocalSymbolsInfo())
+        return;
+    const auto localInfo = (dyld_cache_local_symbols_info*)((uint8_t*)this + header.localSymbolsOffset);
+
+    if ( header.mappingOffset >= __offsetof(dyld_cache_header, symbolFileUUID) ) {
+        // On new caches, the dylibOffset is 64-bits, and is a VM offset
+        const auto localEntries = (dyld_cache_local_symbols_entry_64*)((uint8_t*)localInfo + localInfo->entriesOffset);
+        bool stop = false;
+        for (uint32_t i = 0; i < localInfo->entriesCount; i++) {
+            const dyld_cache_local_symbols_entry_64& localEntry = localEntries[i];
+            handler(localEntry.dylibOffset, localEntry.nlistStartIndex, localEntry.nlistCount, stop);
+        }
+    } else {
+        // On old caches, the dylibOffset is 64-bits, and is a file offset
+        // Note, as we are only looking for mach_header's, a file offset is a VM offset in this case
+        const auto localEntries = (dyld_cache_local_symbols_entry*)((uint8_t*)localInfo + localInfo->entriesOffset);
+        bool stop = false;
+        for (uint32_t i = 0; i < localInfo->entriesCount; i++) {
+            const dyld_cache_local_symbols_entry& localEntry = localEntries[i];
+            handler(localEntry.dylibOffset, localEntry.nlistStartIndex, localEntry.nlistCount, stop);
+        }
+    }
+}
+
+const mach_header* DyldSharedCache::getIndexedImageEntry(uint32_t index, uint64_t& mTime, uint64_t& inode) const
+{
+    const dyld_cache_image_info*   dylibs   = images();
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    mTime = dylibs[index].modTime;
+    inode = dylibs[index].inode;
+    return (mach_header*)((uint8_t*)this + dylibs[index].address - mappings[0].address);
+}
+
+
+ const char* DyldSharedCache::getIndexedImagePath(uint32_t index) const
+{
+    auto dylibs = images();
+    return (char*)this + dylibs[index].pathFileOffset;
+}
+
+void DyldSharedCache::forEachImageTextSegment(void (^handler)(uint64_t loadAddressUnslid, uint64_t textSegmentSize, const uuid_t dylibUUID, const char* installName, bool& stop)) const
+{
+    // check for old cache without imagesText array
+    if ( (header.mappingOffset <= __offsetof(dyld_cache_header, imagesTextOffset)) || (header.imagesTextCount == 0) )
+        return;
+
+    // walk imageText table and call callback for each entry
+    const dyld_cache_image_text_info* imagesText = (dyld_cache_image_text_info*)((char*)this + header.imagesTextOffset);
+    const dyld_cache_image_text_info* imagesTextEnd = &imagesText[header.imagesTextCount];
+    bool stop = false;
+    for (const dyld_cache_image_text_info* p=imagesText; p < imagesTextEnd && !stop; ++p) {
+        handler(p->loadAddress, p->textSegmentSize, p->uuid, (char*)this + p->pathOffset, stop);
+    }
+}
+
+bool DyldSharedCache::addressInText(uint64_t cacheOffset, uint32_t* imageIndex) const
+{
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    uint64_t targetAddr = mappings[0].address + cacheOffset;
+    // walk imageText table and call callback for each entry
+    const dyld_cache_image_text_info* imagesText = (dyld_cache_image_text_info*)((char*)this + header.imagesTextOffset);
+    const dyld_cache_image_text_info* imagesTextEnd = &imagesText[header.imagesTextCount];
+    for (const dyld_cache_image_text_info* p=imagesText; p < imagesTextEnd; ++p) {
+        if ( (p->loadAddress <= targetAddr) && (targetAddr < p->loadAddress+p->textSegmentSize) ) {
+            *imageIndex = (uint32_t)(p-imagesText);
+            return true;
+        }
+    }
+    return false;
+}
+
+const char* DyldSharedCache::archName() const
+{
+    const char* archSubString = ((char*)this) + 7;
+    while (*archSubString == ' ')
+        ++archSubString;
+    return archSubString;
+}
+
+
+dyld3::Platform DyldSharedCache::platform() const
+{
+    return (dyld3::Platform)header.platform;
+}
+
+#if BUILDING_CACHE_BUILDER
+std::string DyldSharedCache::mapFile() const
+{
+    __block std::string             result;
+    __block std::vector<uint64_t>   regionStartAddresses;
+    __block std::vector<uint64_t>   regionSizes;
+    __block std::vector<uint64_t>   regionFileOffsets;
+
+    result.reserve(256*1024);
+    forEachRegion(^(const void* content, uint64_t vmAddr, uint64_t size,
+                    uint32_t initProt, uint32_t maxProt, uint64_t flags, bool& stopRegion) {
+        regionStartAddresses.push_back(vmAddr);
+        regionSizes.push_back(size);
+        regionFileOffsets.push_back((uint8_t*)content - (uint8_t*)this);
+        char lineBuffer[256];
+        const char* prot = "RW";
+        if ( maxProt == (VM_PROT_EXECUTE|VM_PROT_READ) )
+            prot = "EX";
+        else if ( maxProt == VM_PROT_READ )
+            prot = "RO";
+        if ( size > 1024*1024 )
+            sprintf(lineBuffer, "mapping  %s %4lluMB 0x%0llX -> 0x%0llX\n", prot, size/(1024*1024), vmAddr, vmAddr+size);
+        else
+            sprintf(lineBuffer, "mapping  %s %4lluKB 0x%0llX -> 0x%0llX\n", prot, size/1024,        vmAddr, vmAddr+size);
+        result += lineBuffer;
+    });
+
+    // TODO:  add linkedit breakdown
+    result += "\n\n";
+
+    forEachImage(^(const mach_header* mh, const char* installName) {
+        result += std::string(installName) + "\n";
+        const dyld3::MachOFile* mf = (dyld3::MachOFile*)mh;
+        mf->forEachSegment(^(const dyld3::MachOFile::SegmentInfo& info, bool& stop) {
+            char lineBuffer[256];
+            sprintf(lineBuffer, "\t%16s 0x%08llX -> 0x%08llX\n", info.segName, info.vmAddr, info.vmAddr+info.vmSize);
+            result += lineBuffer;
+        });
+        result += "\n";
+    });
+
+    return result;
+}
+#endif
+
+
+uint64_t DyldSharedCache::unslidLoadAddress() const
+{
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    return mappings[0].address;
+}
+
+void DyldSharedCache::getUUID(uuid_t uuid) const
+{
+    memcpy(uuid, header.uuid, sizeof(uuid_t));
+}
+
+uint64_t DyldSharedCache::mappedSize() const
+{
+    // If we have sub caches, then the cache header itself tells us how much space we need to cover all caches
+    if ( header.mappingOffset >= __offsetof(dyld_cache_header, subCacheArrayCount) ) {
+        return header.sharedRegionSize;
+    } else {
+        __block uint64_t startAddr = 0;
+        __block uint64_t endAddr = 0;
+        forEachRegion(^(const void* content, uint64_t vmAddr, uint64_t size,
+                        uint32_t initProt, uint32_t maxProt, uint64_t flags, bool& stopRegion) {
+            if ( startAddr == 0 )
+                startAddr = vmAddr;
+            uint64_t end = vmAddr+size;
+            if ( end > endAddr )
+                endAddr = end;
+        });
+        return (endAddr - startAddr);
+    }
+}
+
+bool DyldSharedCache::findMachHeaderImageIndex(const mach_header* mh, uint32_t& imageIndex) const
+{
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    uintptr_t slide = (uintptr_t)this - (uintptr_t)(mappings[0].address);
+    uint64_t unslidMh = (uintptr_t)mh - slide;
+    const dyld_cache_image_info* dylibs = images();
+    for (uint32_t i=0; i < imagesCount(); ++i) {
+        if ( dylibs[i].address == unslidMh ) {
+            imageIndex = i;
+            return true;
+        }
+    }
+    return false;
+}
+
+bool DyldSharedCache::hasImagePath(const char* dylibPath, uint32_t& imageIndex) const
+{
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    if ( mappings[0].fileOffset != 0 )
+        return false;
+    if ( header.mappingOffset >= 0x118 ) {
+        uintptr_t      slide           = (uintptr_t)this - (uintptr_t)(mappings[0].address);
+        const uint8_t* dylibTrieStart  = (uint8_t*)(this->header.dylibsTrieAddr + slide);
+        const uint8_t* dylibTrieEnd    = dylibTrieStart + this->header.dylibsTrieSize;
+
+        Diagnostics diag;
+        const uint8_t* imageNode = dyld3::MachOLoaded::trieWalk(diag, dylibTrieStart, dylibTrieEnd, dylibPath);
+        if ( imageNode != NULL ) {
+            imageIndex = (uint32_t)dyld3::MachOFile::read_uleb128(diag, imageNode, dylibTrieEnd);
+            return true;
+        }
+    }
+    else {
+        const dyld_cache_image_info* dylibs = images();
+        uint64_t firstImageOffset = 0;
+        uint64_t firstRegionAddress = mappings[0].address;
+        for (uint32_t i=0; i < imagesCount(); ++i) {
+            const char* aPath  = (char*)this + dylibs[i].pathFileOffset;
+            if ( strcmp(aPath, dylibPath) == 0 ) {
+                imageIndex = i;
+                return true;
+            }
+            uint64_t offset = dylibs[i].address - firstRegionAddress;
+            if ( firstImageOffset == 0 )
+                firstImageOffset = offset;
+            // skip over aliases.  This is no longer valid in newer caches.  They store aliases only in the trie
+#if 0
+            if ( dylibs[i].pathFileOffset < firstImageOffset)
+                continue;
+#endif
+        }
+    }
+
+    return false;
+}
+
+bool DyldSharedCache::isOverridablePath(const char* dylibPath) const
+{
+    // all dylibs in customer dyld cache cannot be overridden except libdispatch.dylib
+    if ( header.cacheType == kDyldSharedCacheTypeProduction ) {
+        return (strcmp(dylibPath, "/usr/lib/system/libdispatch.dylib") == 0);
+    }
+    // in dev caches we can override all paths
+    return true;
+}
+
+bool DyldSharedCache::hasNonOverridablePath(const char* dylibPath) const
+{
+    // all dylibs in customer dyld cache cannot be overridden except libdispatch.dylib
+    bool pathIsInDyldCacheWhichCannotBeOverridden = false;
+    if ( header.cacheType == kDyldSharedCacheTypeProduction ) {
+        uint32_t imageIndex;
+        pathIsInDyldCacheWhichCannotBeOverridden = this->hasImagePath(dylibPath, imageIndex);
+        if ( pathIsInDyldCacheWhichCannotBeOverridden && isOverridablePath(dylibPath) )
+            pathIsInDyldCacheWhichCannotBeOverridden = false;
+    }
+    return pathIsInDyldCacheWhichCannotBeOverridden;
+}
+
+intptr_t DyldSharedCache::slide() const
+{
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    return (intptr_t)this - (intptr_t)(mappings[0].address);
+}
+
+const PrebuiltLoader* DyldSharedCache::findPrebuiltLoader(const char* path) const
+{
+    if ( header.mappingOffset < __offsetof(dyld_cache_header, programTrieSize) )
+        return nullptr;
+    uint32_t imageIndex;
+    if ( !this->hasImagePath(path, imageIndex) )
+        return nullptr;
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    if ( mappings[0].fileOffset != 0 )
+        return nullptr;
+    if ( header.mappingOffset < __offsetof(dyld_cache_header, dylibsPBLSetAddr) )
+        return nullptr;
+    if ( header.dylibsPBLSetAddr == 0 )
+        return nullptr;
+    uintptr_t                slide        = (uintptr_t)this - (uintptr_t)(mappings[0].address);
+    const PrebuiltLoaderSet* pbLoaderSet  = (PrebuiltLoaderSet*)(this->header.dylibsPBLSetAddr + slide);
+    return pbLoaderSet->atIndex(imageIndex);
+}
+
+void DyldSharedCache::forEachLaunchLoaderSet(void (^handler)(const char* executableRuntimePath, const PrebuiltLoaderSet* pbls)) const
+{
+    if ( header.mappingOffset < __offsetof(dyld_cache_header, programTrieSize) )
+        return;
+    if ( this->header.programTrieAddr == 0 )
+        return;
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    uintptr_t      slide                = (uintptr_t)this - (uintptr_t)(mappings[0].address);
+    const uint8_t* executableTrieStart  = (uint8_t*)(this->header.programTrieAddr + slide);
+    const uint8_t* executableTrieEnd    = executableTrieStart + this->header.programTrieSize;
+    const uint8_t* poolStart            = (uint8_t*)(this->header.programsPBLSetPoolAddr + slide);
+
+    std::vector<DylibIndexTrie::Entry> loaderSetEntries;
+    if ( Trie<DylibIndex>::parseTrie(executableTrieStart, executableTrieEnd, loaderSetEntries) ) {
+        for (const DylibIndexTrie::Entry& entry : loaderSetEntries ) {
+            uint32_t offset = entry.info.index;
+            if ( offset < this->header.programsPBLSetPoolSize )
+                handler(entry.name.c_str(), (const PrebuiltLoaderSet*)(poolStart+offset));
+        }
+    }
+}
+
+const PrebuiltLoaderSet* DyldSharedCache::findLaunchLoaderSet(const char* executablePath) const
+{
+    if ( header.mappingOffset < __offsetof(dyld_cache_header, programTrieSize) )
+        return nullptr;
+    if ( this->header.programTrieAddr == 0 )
+        return nullptr;
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    uintptr_t      slide                = (uintptr_t)this - (uintptr_t)(mappings[0].address);
+    const uint8_t* executableTrieStart  = (uint8_t*)(this->header.programTrieAddr + slide);
+    const uint8_t* executableTrieEnd    = executableTrieStart + this->header.programTrieSize;
+    const uint8_t* poolStart            = (uint8_t*)(this->header.programsPBLSetPoolAddr + slide);
+
+    Diagnostics diag;
+    if ( const uint8_t* imageNode = dyld3::MachOLoaded::trieWalk(diag, executableTrieStart, executableTrieEnd, executablePath) ) {
+        uint32_t poolOffset = (uint32_t)dyld3::MachOFile::read_uleb128(diag, imageNode, executableTrieEnd);
+        if ( poolOffset < this->header.programsPBLSetPoolSize ) {
+            return (PrebuiltLoaderSet*)((uint8_t*)poolStart + poolOffset);
+        }
+    }
+
+    return nullptr;
+}
+
+bool DyldSharedCache::hasLaunchLoaderSetWithCDHash(const char* cdHashString) const
+{
+    if ( cdHashString == nullptr )
+        return false;
+
+    // Check source doesn't overflow buffer.  strncat unfortunately isn't available
+    if ( strlen(cdHashString) >= 128 )
+        return false;
+
+    char buffer[128] = { '\0' };
+    strcat(buffer, "/cdhash/");
+    strlcat(buffer, cdHashString, sizeof(buffer));
+    return findLaunchLoaderSet(buffer) != nullptr;
+}
+
+
+#if 0 //!BUILDING_LIBDSC
+const dyld3::closure::Image* DyldSharedCache::findDlopenOtherImage(const char* path) const
+{
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    if ( mappings[0].fileOffset != 0 )
+        return nullptr;
+    if ( header.mappingOffset < __offsetof(dyld_cache_header, otherImageArrayAddr) )
+        return nullptr;
+    if ( header.otherImageArrayAddr == 0 )
+        return nullptr;
+    uintptr_t      slide           = (uintptr_t)this - (uintptr_t)(mappings[0].address);
+    const uint8_t* dylibTrieStart  = (uint8_t*)(this->header.otherTrieAddr + slide);
+    const uint8_t* dylibTrieEnd    = dylibTrieStart + this->header.otherTrieSize;
+
+    Diagnostics diag;
+    const uint8_t* imageNode = dyld3::MachOLoaded::trieWalk(diag, dylibTrieStart, dylibTrieEnd, path);
+    if ( imageNode != NULL ) {
+        dyld3::closure::ImageNum imageNum = (uint32_t)dyld3::MachOFile::read_uleb128(diag, imageNode, dylibTrieEnd);
+        uint64_t arrayAddrOffset = header.otherImageArrayAddr - mappings[0].address;
+        const dyld3::closure::ImageArray* otherImageArray = (dyld3::closure::ImageArray*)((char*)this + arrayAddrOffset);
+        return otherImageArray->imageForNum(imageNum);
+    }
+
+    return nullptr;
+}
+
+const dyld3::closure::LaunchClosure* DyldSharedCache::findClosure(const char* executablePath) const
+{
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    uintptr_t      slide                = (uintptr_t)this - (uintptr_t)(mappings[0].address);
+    const uint8_t* executableTrieStart  = (uint8_t*)(this->header.progClosuresTrieAddr + slide);
+    const uint8_t* executableTrieEnd    = executableTrieStart + this->header.progClosuresTrieSize;
+    const uint8_t* closuresStart        = (uint8_t*)(this->header.progClosuresAddr + slide);
+
+    Diagnostics diag;
+    const uint8_t* imageNode = dyld3::MachOLoaded::trieWalk(diag, executableTrieStart, executableTrieEnd, executablePath);
+    if ( (imageNode == NULL) && (strncmp(executablePath, "/System/", 8) == 0) ) {
+        // anything in /System/ should have a closure.  Perhaps it was launched via symlink path
+        char realPath[PATH_MAX];
+        if ( realpath(executablePath, realPath) != NULL )
+            imageNode = dyld3::MachOLoaded::trieWalk(diag, executableTrieStart, executableTrieEnd, realPath);
+    }
+    if ( imageNode != NULL ) {
+        uint32_t closureOffset = (uint32_t)dyld3::MachOFile::read_uleb128(diag, imageNode, executableTrieEnd);
+        if ( closureOffset < this->header.progClosuresSize )
+            return (dyld3::closure::LaunchClosure*)((uint8_t*)closuresStart + closureOffset);
+    }
+
+    return nullptr;
+}
+
+#if !BUILDING_LIBDYLD && !BUILDING_DYLD
+void DyldSharedCache::forEachLaunchClosure(void (^handler)(const char* executableRuntimePath, const dyld3::closure::LaunchClosure* closure)) const
+{
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    uintptr_t      slide                = (uintptr_t)this - (uintptr_t)(mappings[0].address);
+    const uint8_t* executableTrieStart  = (uint8_t*)(this->header.progClosuresTrieAddr + slide);
+    const uint8_t* executableTrieEnd    = executableTrieStart + this->header.progClosuresTrieSize;
+    const uint8_t* closuresStart        = (uint8_t*)(this->header.progClosuresAddr + slide);
+
+    std::vector<DylibIndexTrie::Entry> closureEntries;
+    if ( Trie<DylibIndex>::parseTrie(executableTrieStart, executableTrieEnd, closureEntries) ) {
+        for (DylibIndexTrie::Entry& entry : closureEntries ) {
+            uint32_t offset = entry.info.index;
+            if ( offset < this->header.progClosuresSize )
+                handler(entry.name.c_str(), (const dyld3::closure::LaunchClosure*)(closuresStart+offset));
+        }
+    }
+}
+
+void DyldSharedCache::forEachDlopenImage(void (^handler)(const char* runtimePath, const dyld3::closure::Image* image)) const
+{
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    uintptr_t      slide           = (uintptr_t)this - (uintptr_t)(mappings[0].address);
+    const uint8_t* otherTrieStart  = (uint8_t*)(this->header.otherTrieAddr + slide);
+    const uint8_t* otherTrieEnd    = otherTrieStart + this->header.otherTrieSize;
+
+    std::vector<DylibIndexTrie::Entry> otherEntries;
+    if ( Trie<DylibIndex>::parseTrie(otherTrieStart, otherTrieEnd, otherEntries) ) {
+        for (const DylibIndexTrie::Entry& entry : otherEntries ) {
+            dyld3::closure::ImageNum imageNum = entry.info.index;
+            uint64_t arrayAddrOffset = header.otherImageArrayAddr - mappings[0].address;
+            const dyld3::closure::ImageArray* otherImageArray = (dyld3::closure::ImageArray*)((char*)this + arrayAddrOffset);
+            handler(entry.name.c_str(), otherImageArray->imageForNum(imageNum));
+        }
+    }
+}
+#endif // !BUILDING_LIBDYLD && !BUILDING_DYLD
+
+
+const dyld3::closure::ImageArray* DyldSharedCache::cachedDylibsImageArray() const
+{
+    // check for old cache without imagesArray
+    if ( header.mappingOffset < 0x100 )
+        return nullptr;
+
+    if ( header.dylibsImageArrayAddr == 0 )
+        return nullptr;
+        
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    uint64_t arrayAddrOffset = header.dylibsImageArrayAddr - mappings[0].address;
+    return (dyld3::closure::ImageArray*)((char*)this + arrayAddrOffset);
+}
+
+const dyld3::closure::ImageArray* DyldSharedCache::otherOSImageArray() const
+{
+    // check for old cache without imagesArray
+    if ( header.mappingOffset < __offsetof(dyld_cache_header, otherImageArrayAddr) )
+        return nullptr;
+
+    if ( header.otherImageArrayAddr == 0 )
+        return nullptr;
+
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    uint64_t arrayAddrOffset = header.otherImageArrayAddr - mappings[0].address;
+    return (dyld3::closure::ImageArray*)((char*)this + arrayAddrOffset);
+}
+#endif // !BUILDING_LIBDSC
+
+void DyldSharedCache::forEachDylibPath(void (^handler)(const char* dylibPath, uint32_t index)) const
+{
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    uintptr_t      slide                = (uintptr_t)this - (uintptr_t)(mappings[0].address);
+    const uint8_t* dylibTrieStart       = (uint8_t*)(this->header.dylibsTrieAddr + slide);
+    const uint8_t* dylibTrieEnd         = dylibTrieStart + this->header.dylibsTrieSize;
+
+   std::vector<DylibIndexTrie::Entry> dylibEntries;
+    if ( Trie<DylibIndex>::parseTrie(dylibTrieStart, dylibTrieEnd, dylibEntries) ) {
+        for (DylibIndexTrie::Entry& entry : dylibEntries ) {
+            handler(entry.name.c_str(), entry.info.index);
+        }
+    }
+}
+
+uint32_t DyldSharedCache::patchInfoVersion() const {
+    if ( header.mappingOffset <= __offsetof(dyld_cache_header, swiftOptsSize) ) {
+        return 1;
+    }
+
+    const dyld_cache_patch_info_v2* patchInfo = getAddrField<dyld_cache_patch_info_v2*>(header.patchInfoAddr);
+    return patchInfo->patchTableVersion;
+}
+
+uint32_t DyldSharedCache::patchableExportCount(uint32_t imageIndex) const {
+    if ( header.patchInfoAddr == 0 )
+        return 0;
+
+    uint32_t patchVersion = patchInfoVersion();
+
+    if ( patchVersion == 1 ) {
+        // Old cache.  The patch table uses the V1 structs
+
+        const dyld_cache_patch_info_v1* patchInfo = getAddrField<dyld_cache_patch_info_v1*>(header.patchInfoAddr);
+        const dyld_cache_image_patches_v1* patchArray = getAddrField<dyld_cache_image_patches_v1*>(patchInfo->patchTableArrayAddr);
+        if (imageIndex > patchInfo->patchTableArrayCount)
+            return 0;
+        return patchArray[imageIndex].patchExportsCount;
+    }
+
+    // V2 and newer structs
+    if ( patchVersion == 2 ) {
+        const dyld_cache_patch_info_v2* patchInfo = getAddrField<dyld_cache_patch_info_v2*>(header.patchInfoAddr);
+        const dyld_cache_image_patches_v2* patchArray = getAddrField<dyld_cache_image_patches_v2*>(patchInfo->patchTableArrayAddr);
+
+        // FIXME: This shouldn't be possible.  The cache builder always adds 1 dyld_cache_image_patches_v2 per-image.  But for now
+        // be conservative to match V1.
+        if (imageIndex > patchInfo->patchTableArrayCount)
+            return 0;
+
+        const dyld_cache_image_patches_v2& imagePatches = patchArray[imageIndex];
+        return imagePatches.patchExportsCount;
+    }
+
+    // Unknown version
+    assert(false);
+}
+
+void DyldSharedCache::forEachPatchableExport(uint32_t imageIndex, void (^handler)(uint32_t dylibVMOffsetOfImpl, const char* exportName)) const {
+    if ( header.patchInfoAddr == 0 )
+        return;
+
+    uint32_t patchVersion = patchInfoVersion();
+
+    if ( patchVersion == 1 ) {
+        // Old cache.  The patch table uses the V1 structs
+
+        // This patch table uses cache offsets, so convert from cache offset to "image + offset"
+        uint64_t mTime;
+        uint64_t inode;
+        const dyld3::MachOAnalyzer* imageMA = (dyld3::MachOAnalyzer*)(this->getIndexedImageEntry(imageIndex, mTime, inode));
+        if ( imageMA == nullptr )
+            return;
+
+        uint64_t imageLoadAddress = imageMA->preferredLoadAddress();
+        uint64_t cacheUnslidAddress = unslidLoadAddress();
+
+        const dyld_cache_patch_info_v1* patchInfo = getAddrField<dyld_cache_patch_info_v1*>(header.patchInfoAddr);
+        const dyld_cache_image_patches_v1* patchArray = getAddrField<dyld_cache_image_patches_v1*>(patchInfo->patchTableArrayAddr);
+        if (imageIndex > patchInfo->patchTableArrayCount)
+            return;
+        const dyld_cache_image_patches_v1& patch = patchArray[imageIndex];
+        if ( (patch.patchExportsStartIndex + patch.patchExportsCount) > patchInfo->patchExportArrayCount )
+            return;
+        const dyld_cache_patchable_export_v1* patchExports = getAddrField<dyld_cache_patchable_export_v1*>(patchInfo->patchExportArrayAddr);
+        const char* exportNames = getAddrField<char*>(patchInfo->patchExportNamesAddr);
+        for (uint64_t exportIndex = 0; exportIndex != patch.patchExportsCount; ++exportIndex) {
+            const dyld_cache_patchable_export_v1& patchExport = patchExports[patch.patchExportsStartIndex + exportIndex];
+            const char* exportName = ( patchExport.exportNameOffset < patchInfo->patchExportNamesSize ) ? &exportNames[patchExport.exportNameOffset] : "";
+
+            // Convert from a cache offset to an offset from the input image
+            uint32_t imageOffset = (uint32_t)((cacheUnslidAddress + patchExport.cacheOffsetOfImpl) - imageLoadAddress);
+            handler(imageOffset, exportName);
+        }
+
+        return;
+    }
+
+    // V2 and newer structs
+    if ( patchVersion == 2 ) {
+        const auto* patchInfo = getAddrField<dyld_cache_patch_info_v2*>(header.patchInfoAddr);
+        const auto* patchArray = getAddrField<dyld_cache_image_patches_v2*>(patchInfo->patchTableArrayAddr);
+        const auto* imageExports = getAddrField<dyld_cache_image_export_v2*>(patchInfo->patchImageExportsArrayAddr);
+        const char* exportNames = getAddrField<char*>(patchInfo->patchExportNamesAddr);
+
+        // FIXME: This shouldn't be possible.  The cache builder always adds 1 dyld_cache_image_patches_v2 per-image.  But for now
+        // be conservative to match V1.
+        if (imageIndex > patchInfo->patchTableArrayCount)
+            return;
+
+        const dyld_cache_image_patches_v2& imagePatches = patchArray[imageIndex];
+        // FIXME: This shouldn't be possible.  We should catch this in the builder instead
+        if ( (imagePatches.patchExportsStartIndex + imagePatches.patchExportsCount) > patchInfo->patchImageExportsArrayCount )
+            return;
+
+        for (uint64_t exportIndex = 0; exportIndex != imagePatches.patchExportsCount; ++exportIndex) {
+            const dyld_cache_image_export_v2& imageExport = imageExports[imagePatches.patchExportsStartIndex + exportIndex];
+            const char* exportName = ( imageExport.exportNameOffset < patchInfo->patchExportNamesSize ) ? &exportNames[imageExport.exportNameOffset] : "";
+            handler(imageExport.dylibOffsetOfImpl, exportName);
+        }
+        return;
+    }
+
+    assert(false);
+}
+
+void DyldSharedCache::forEachPatchableUseOfExport(uint32_t imageIndex, uint32_t dylibVMOffsetOfImpl,
+                                                  void (^handler)(uint32_t userImageIndex, uint32_t userVMOffset,
+                                                                  MachOLoaded::PointerMetaData pmd, uint64_t addend)) const {
+    if ( header.patchInfoAddr == 0 )
+        return;
+
+    uint32_t patchVersion = patchInfoVersion();
+
+    if ( patchVersion == 1 ) {
+        // Old cache.  The patch table uses the V1 structs
+
+        // This patch table uses cache offsets, so convert from "image + offset" to cache offset
+        uint64_t mTime;
+        uint64_t inode;
+        const dyld3::MachOAnalyzer* imageMA = (dyld3::MachOAnalyzer*)(this->getIndexedImageEntry(imageIndex, mTime, inode));
+        if ( imageMA == nullptr )
+            return;
+
+        uint64_t cacheUnslidAddress = unslidLoadAddress();
+        uint32_t cacheOffsetOfImpl = (uint32_t)((imageMA->preferredLoadAddress() - cacheUnslidAddress) + dylibVMOffsetOfImpl);
+
+        // Loading a new cache so get the data from the cache header
+        const dyld_cache_patch_info_v1* patchInfo = getAddrField<dyld_cache_patch_info_v1*>(header.patchInfoAddr);
+        const dyld_cache_image_patches_v1* patchArray = getAddrField<dyld_cache_image_patches_v1*>(patchInfo->patchTableArrayAddr);
+        if (imageIndex > patchInfo->patchTableArrayCount)
+            return;
+        const dyld_cache_image_patches_v1& patch = patchArray[imageIndex];
+        if ( (patch.patchExportsStartIndex + patch.patchExportsCount) > patchInfo->patchExportArrayCount )
+            return;
+
+        // V1 doesn't know which patch location corresponds to which dylib.  This is expensive, but temporary, so find the dylib for
+        // each patch
+        struct DataRange
+        {
+            uint64_t cacheOffsetStart;
+            uint64_t cacheOffsetEnd;
+        };
+        STACK_ALLOC_OVERFLOW_SAFE_ARRAY(DataRange, dataRanges, 8);
+        __block const dyld3::MachOAnalyzer* userDylib = nullptr;
+        __block uint32_t userImageIndex = ~0U;
+
+        const dyld_cache_patchable_export_v1* patchExports = getAddrField<dyld_cache_patchable_export_v1*>(patchInfo->patchExportArrayAddr);
+        const dyld_cache_patchable_location_v1* patchLocations = getAddrField<dyld_cache_patchable_location_v1*>(patchInfo->patchLocationArrayAddr);
+        for (uint64_t exportIndex = 0; exportIndex != patch.patchExportsCount; ++exportIndex) {
+            const dyld_cache_patchable_export_v1& patchExport = patchExports[patch.patchExportsStartIndex + exportIndex];
+            if ( patchExport.cacheOffsetOfImpl != cacheOffsetOfImpl )
+                continue;
+            if ( (patchExport.patchLocationsStartIndex + patchExport.patchLocationsCount) > patchInfo->patchLocationArrayCount )
+                return;
+            for (uint64_t locationIndex = 0; locationIndex != patchExport.patchLocationsCount; ++locationIndex) {
+                const dyld_cache_patchable_location_v1& patchLocation = patchLocations[patchExport.patchLocationsStartIndex + locationIndex];
+
+                bool computeNewRanges = false;
+                if ( userDylib == nullptr ) {
+                    computeNewRanges = true;
+                } else {
+                    bool inRange = false;
+                    for ( const DataRange& range : dataRanges ) {
+                        if ( (patchLocation.cacheOffset >= range.cacheOffsetStart) && (patchLocation.cacheOffset < range.cacheOffsetEnd) ) {
+                            inRange = true;
+                            break;
+                        }
+                    }
+                    if ( !inRange )
+                        computeNewRanges = true;
+                }
+
+                if ( computeNewRanges ) {
+                    userDylib = nullptr;
+                    userImageIndex = ~0U;
+                    dataRanges.clear();
+                    forEachDylib(^(const dyld3::MachOAnalyzer* ma, const char* dylibPath, uint32_t cacheImageIndex, uint64_t, uint64_t, bool& stopImage) {
+                        ma->forEachSegment(^(const MachOAnalyzer::SegmentInfo& info, bool& stopSegment) {
+                            if ( info.writable() )
+                                dataRanges.push_back({ info.vmAddr - cacheUnslidAddress, info.vmAddr + info.vmSize - cacheUnslidAddress });
+                        });
+
+                        bool inRange = false;
+                        for ( const DataRange& range : dataRanges ) {
+                            if ( (patchLocation.cacheOffset >= range.cacheOffsetStart) && (patchLocation.cacheOffset < range.cacheOffsetEnd) ) {
+                                inRange = true;
+                                break;
+                            }
+                        }
+                        if ( inRange ) {
+                            // This is dylib we want.  So we can keep these ranges, and record this mach-header
+                            userDylib = ma;
+                            userImageIndex = cacheImageIndex;
+                            stopImage = true;
+                        } else {
+                            // These ranges don't work.  Clear them and move on to the next dylib
+                            dataRanges.clear();
+                        }
+                    });
+                }
+
+                assert(userDylib != nullptr);
+                assert(userImageIndex != ~0U);
+                assert(!dataRanges.empty());
+
+                uint32_t userVMOffset = (uint32_t)((cacheUnslidAddress + patchLocation.cacheOffset) - userDylib->preferredLoadAddress());
+                dyld3::MachOLoaded::PointerMetaData pmd;
+                pmd.diversity         = patchLocation.discriminator;
+                pmd.high8             = patchLocation.high7 << 1;
+                pmd.authenticated     = patchLocation.authenticated;
+                pmd.key               = patchLocation.key;
+                pmd.usesAddrDiversity = patchLocation.usesAddressDiversity;
+
+                handler(userImageIndex, userVMOffset, pmd, patchLocation.getAddend());
+            }
+        }
+        return;
+    }
+
+    // V2 and newer structs
+    if ( patchVersion == 2 ) {
+        const auto* patchInfo = getAddrField<dyld_cache_patch_info_v2*>(header.patchInfoAddr);
+        const auto* patchArray = getAddrField<dyld_cache_image_patches_v2*>(patchInfo->patchTableArrayAddr);
+        const auto* imageExports = getAddrField<dyld_cache_image_export_v2*>(patchInfo->patchImageExportsArrayAddr);
+        const auto* clientArray = getAddrField<dyld_cache_image_clients_v2*>(patchInfo->patchClientsArrayAddr);
+        const auto* clientExports = getAddrField<dyld_cache_patchable_export_v2*>(patchInfo->patchClientExportsArrayAddr);
+        const auto* patchLocations = getAddrField<dyld_cache_patchable_location_v2*>(patchInfo->patchLocationArrayAddr);
+
+        // FIXME: This shouldn't be possible.  The cache builder always adds 1 dyld_cache_image_patches_v2 per-image.  But for now
+        // be conservative to match V1.
+        if (imageIndex > patchInfo->patchTableArrayCount)
+            return;
+
+        const dyld_cache_image_patches_v2& imagePatches = patchArray[imageIndex];
+        // FIXME: This shouldn't be possible.  We should catch this in the builder instead
+        if ( (imagePatches.patchClientsStartIndex + imagePatches.patchClientsCount) > patchInfo->patchClientsArrayCount )
+            return;
+
+        for (uint64_t clientIndex = 0; clientIndex != imagePatches.patchClientsCount; ++clientIndex) {
+            const dyld_cache_image_clients_v2& client = clientArray[imagePatches.patchClientsStartIndex + clientIndex];
+
+            // FIXME: This shouldn't be possible.  We should catch this in the builder instead
+            if ( (client.patchExportsStartIndex + client.patchExportsCount) > patchInfo->patchClientExportsArrayCount )
+                return;
+
+            for (uint64_t exportIndex = 0; exportIndex != client.patchExportsCount; ++exportIndex) {
+                const dyld_cache_patchable_export_v2& clientExport = clientExports[client.patchExportsStartIndex + exportIndex];
+
+                // The client export points to an image export from the image we are rooting
+                if ( clientExport.imageExportIndex > patchInfo->patchImageExportsArrayCount )
+                    return;
+
+                const dyld_cache_image_export_v2& imageExport = imageExports[clientExport.imageExportIndex];
+
+                if ( imageExport.dylibOffsetOfImpl != dylibVMOffsetOfImpl )
+                    continue;
+
+                if ( (clientExport.patchLocationsStartIndex + clientExport.patchLocationsCount) > patchInfo->patchLocationArrayCount )
+                    return;
+
+                for (uint64_t locationIndex = 0; locationIndex != clientExport.patchLocationsCount; ++locationIndex) {
+                    const dyld_cache_patchable_location_v2& patchLocation = patchLocations[clientExport.patchLocationsStartIndex + locationIndex];
+
+                    dyld3::MachOLoaded::PointerMetaData pmd;
+                    pmd.diversity         = patchLocation.discriminator;
+                    pmd.high8             = patchLocation.high7 << 1;
+                    pmd.authenticated     = patchLocation.authenticated;
+                    pmd.key               = patchLocation.key;
+                    pmd.usesAddrDiversity = patchLocation.usesAddressDiversity;
+
+                    handler(client.clientDylibIndex, patchLocation.dylibOffsetOfUse, pmd, patchLocation.getAddend());
+                }
+            }
+        }
+        return;
+    }
+
+    assert(false);
+}
+
+bool DyldSharedCache::shouldPatchClientOfImage(uint32_t imageIndex, uint32_t userImageIndex) const {
+    if ( header.patchInfoAddr == 0 )
+        return false;
+
+    uint32_t patchVersion = patchInfoVersion();
+
+    if ( patchVersion == 1 ) {
+        // Old cache.  The patch table uses the V1 structs
+        // Only dyld uses this method and is on at least v2, so we don't implement this
+        return false;
+    }
+
+    // V2 and newer structs
+    if ( patchVersion == 2 ) {
+        const auto* patchInfo = getAddrField<dyld_cache_patch_info_v2*>(header.patchInfoAddr);
+        const auto* patchArray = getAddrField<dyld_cache_image_patches_v2*>(patchInfo->patchTableArrayAddr);
+        const auto* clientArray = getAddrField<dyld_cache_image_clients_v2*>(patchInfo->patchClientsArrayAddr);
+
+        // FIXME: This shouldn't be possible.  The cache builder always adds 1 dyld_cache_image_patches_v2 per-image.  But for now
+        // be conservative to match V1.
+        if (imageIndex > patchInfo->patchTableArrayCount)
+            return false;
+
+        const dyld_cache_image_patches_v2& imagePatches = patchArray[imageIndex];
+        // FIXME: This shouldn't be possible.  We should catch this in the builder instead
+        if ( (imagePatches.patchClientsStartIndex + imagePatches.patchClientsCount) > patchInfo->patchClientsArrayCount )
+            return false;
+
+        for (uint64_t clientIndex = 0; clientIndex != imagePatches.patchClientsCount; ++clientIndex) {
+            const dyld_cache_image_clients_v2& client = clientArray[imagePatches.patchClientsStartIndex + clientIndex];
+
+            // We only want fixups in a specific image.  Skip any others
+            if ( client.clientDylibIndex == userImageIndex )
+                return true;
+        }
+        return false;
+    }
+
+    assert(false);
+}
+
+void DyldSharedCache::forEachPatchableUseOfExportInImage(uint32_t imageIndex, uint32_t dylibVMOffsetOfImpl, uint32_t userImageIndex,
+                                                         void (^handler)(uint32_t userVMOffset, MachOLoaded::PointerMetaData pmd, uint64_t addend)) const {
+    if ( header.patchInfoAddr == 0 )
+        return;
+
+    uint32_t patchVersion = patchInfoVersion();
+
+    if ( patchVersion == 1 ) {
+        // Old cache.  The patch table uses the V1 structs
+
+        // This patch table uses cache offsets, so convert from "image + offset" to cache offset
+        uint64_t mTime;
+        uint64_t inode;
+        const dyld3::MachOAnalyzer* imageMA = (dyld3::MachOAnalyzer*)(this->getIndexedImageEntry(imageIndex, mTime, inode));
+        if ( imageMA == nullptr )
+            return;
+
+        uint64_t cacheUnslidAddress = unslidLoadAddress();
+        uint32_t cacheOffsetOfImpl = (uint32_t)((imageMA->preferredLoadAddress() - cacheUnslidAddress) + dylibVMOffsetOfImpl);
+
+        // Loading a new cache so get the data from the cache header
+        const dyld_cache_patch_info_v1* patchInfo = getAddrField<dyld_cache_patch_info_v1*>(header.patchInfoAddr);
+        const dyld_cache_image_patches_v1* patchArray = getAddrField<dyld_cache_image_patches_v1*>(patchInfo->patchTableArrayAddr);
+        if (imageIndex > patchInfo->patchTableArrayCount)
+            return;
+        const dyld_cache_image_patches_v1& patch = patchArray[imageIndex];
+        if ( (patch.patchExportsStartIndex + patch.patchExportsCount) > patchInfo->patchExportArrayCount )
+            return;
+
+        // V1 doesn't know which patch location corresponds to which dylib.  This is expensive, but temporary, so find the dylib for
+        // each patch
+        struct DataRange
+        {
+            uint64_t cacheOffsetStart;
+            uint64_t cacheOffsetEnd;
+        };
+        STACK_ALLOC_OVERFLOW_SAFE_ARRAY(DataRange, dataRanges, 8);
+        __block const dyld3::MachOAnalyzer* userDylib = nullptr;
+        __block uint32_t userDylibImageIndex = ~0U;
+
+        const dyld_cache_patchable_export_v1* patchExports = getAddrField<dyld_cache_patchable_export_v1*>(patchInfo->patchExportArrayAddr);
+        const dyld_cache_patchable_location_v1* patchLocations = getAddrField<dyld_cache_patchable_location_v1*>(patchInfo->patchLocationArrayAddr);
+        for (uint64_t exportIndex = 0; exportIndex != patch.patchExportsCount; ++exportIndex) {
+            const dyld_cache_patchable_export_v1& patchExport = patchExports[patch.patchExportsStartIndex + exportIndex];
+            if ( patchExport.cacheOffsetOfImpl != cacheOffsetOfImpl )
+                continue;
+            if ( (patchExport.patchLocationsStartIndex + patchExport.patchLocationsCount) > patchInfo->patchLocationArrayCount )
+                return;
+            for (uint64_t locationIndex = 0; locationIndex != patchExport.patchLocationsCount; ++locationIndex) {
+                const dyld_cache_patchable_location_v1& patchLocation = patchLocations[patchExport.patchLocationsStartIndex + locationIndex];
+
+                bool computeNewRanges = false;
+                if ( userDylib == nullptr ) {
+                    computeNewRanges = true;
+                } else {
+                    bool inRange = false;
+                    for ( const DataRange& range : dataRanges ) {
+                        if ( (patchLocation.cacheOffset >= range.cacheOffsetStart) && (patchLocation.cacheOffset < range.cacheOffsetEnd) ) {
+                            inRange = true;
+                            break;
+                        }
+                    }
+                    if ( !inRange )
+                        computeNewRanges = true;
+                }
+
+                if ( computeNewRanges ) {
+                    userDylib = nullptr;
+                    userDylibImageIndex = ~0U;
+                    dataRanges.clear();
+                    forEachDylib(^(const dyld3::MachOAnalyzer* ma, const char* dylibPath, uint32_t cacheImageIndex, uint64_t, uint64_t, bool& stopImage) {
+                        ma->forEachSegment(^(const MachOAnalyzer::SegmentInfo& info, bool& stopSegment) {
+                            if ( info.writable() )
+                                dataRanges.push_back({ info.vmAddr - cacheUnslidAddress, info.vmAddr + info.vmSize - cacheUnslidAddress });
+                        });
+
+                        bool inRange = false;
+                        for ( const DataRange& range : dataRanges ) {
+                            if ( (patchLocation.cacheOffset >= range.cacheOffsetStart) && (patchLocation.cacheOffset < range.cacheOffsetEnd) ) {
+                                inRange = true;
+                                break;
+                            }
+                        }
+                        if ( inRange ) {
+                            // This is dylib we want.  So we can keep these ranges, and record this mach-header
+                            userDylib = ma;
+                            userDylibImageIndex = cacheImageIndex;
+                            stopImage = true;
+                        } else {
+                            // These ranges don't work.  Clear them and move on to the next dylib
+                            dataRanges.clear();
+                        }
+                    });
+                }
+
+                assert(userDylib != nullptr);
+                assert(userDylibImageIndex != ~0U);
+                assert(!dataRanges.empty());
+
+                // We only want fixups in a specific image.  Skip any others
+                if ( userDylibImageIndex == userImageIndex ) {
+                    uint32_t userVMOffset = (uint32_t)((cacheUnslidAddress + patchLocation.cacheOffset) - userDylib->preferredLoadAddress());
+                    dyld3::MachOLoaded::PointerMetaData pmd;
+                    pmd.diversity         = patchLocation.discriminator;
+                    pmd.high8             = patchLocation.high7 << 1;
+                    pmd.authenticated     = patchLocation.authenticated;
+                    pmd.key               = patchLocation.key;
+                    pmd.usesAddrDiversity = patchLocation.usesAddressDiversity;
+
+                    handler(userVMOffset, pmd, patchLocation.getAddend());
+                }
+            }
+        }
+        return;
+    }
+
+    // V2 and newer structs
+    if ( patchVersion == 2 ) {
+        const auto* patchInfo = getAddrField<dyld_cache_patch_info_v2*>(header.patchInfoAddr);
+        const auto* patchArray = getAddrField<dyld_cache_image_patches_v2*>(patchInfo->patchTableArrayAddr);
+        const auto* imageExports = getAddrField<dyld_cache_image_export_v2*>(patchInfo->patchImageExportsArrayAddr);
+        const auto* clientArray = getAddrField<dyld_cache_image_clients_v2*>(patchInfo->patchClientsArrayAddr);
+        const auto* clientExports = getAddrField<dyld_cache_patchable_export_v2*>(patchInfo->patchClientExportsArrayAddr);
+        const auto* patchLocations = getAddrField<dyld_cache_patchable_location_v2*>(patchInfo->patchLocationArrayAddr);
+
+        // FIXME: This shouldn't be possible.  The cache builder always adds 1 dyld_cache_image_patches_v2 per-image.  But for now
+        // be conservative to match V1.
+        if (imageIndex > patchInfo->patchTableArrayCount)
+            return;
+
+        const dyld_cache_image_patches_v2& imagePatches = patchArray[imageIndex];
+        // FIXME: This shouldn't be possible.  We should catch this in the builder instead
+        if ( (imagePatches.patchClientsStartIndex + imagePatches.patchClientsCount) > patchInfo->patchClientsArrayCount )
+            return;
+
+        for (uint64_t clientIndex = 0; clientIndex != imagePatches.patchClientsCount; ++clientIndex) {
+            const dyld_cache_image_clients_v2& client = clientArray[imagePatches.patchClientsStartIndex + clientIndex];
+
+            // We only want fixups in a specific image.  Skip any others
+            if ( client.clientDylibIndex != userImageIndex )
+                continue;
+
+            // FIXME: This shouldn't be possible.  We should catch this in the builder instead
+            if ( (client.patchExportsStartIndex + client.patchExportsCount) > patchInfo->patchClientExportsArrayCount )
+                return;
+
+            for (uint64_t exportIndex = 0; exportIndex != client.patchExportsCount; ++exportIndex) {
+                const dyld_cache_patchable_export_v2& clientExport = clientExports[client.patchExportsStartIndex + exportIndex];
+
+                // The client export points to an image export from the image we are rooting
+                if ( clientExport.imageExportIndex > patchInfo->patchImageExportsArrayCount )
+                    return;
+
+                const dyld_cache_image_export_v2& imageExport = imageExports[clientExport.imageExportIndex];
+
+                if ( imageExport.dylibOffsetOfImpl != dylibVMOffsetOfImpl )
+                    continue;
+
+                if ( (clientExport.patchLocationsStartIndex + clientExport.patchLocationsCount) > patchInfo->patchLocationArrayCount )
+                    return;
+
+                for (uint64_t locationIndex = 0; locationIndex != clientExport.patchLocationsCount; ++locationIndex) {
+                    const dyld_cache_patchable_location_v2& patchLocation = patchLocations[clientExport.patchLocationsStartIndex + locationIndex];
+
+                    dyld3::MachOLoaded::PointerMetaData pmd;
+                    pmd.diversity         = patchLocation.discriminator;
+                    pmd.high8             = patchLocation.high7 << 1;
+                    pmd.authenticated     = patchLocation.authenticated;
+                    pmd.key               = patchLocation.key;
+                    pmd.usesAddrDiversity = patchLocation.usesAddressDiversity;
+
+                    handler(patchLocation.dylibOffsetOfUse, pmd, patchLocation.getAddend());
+                }
+            }
+
+            // We only wanted to process this image.  We are now done
+            break;
+        }
+        return;
+    }
+
+    assert(false);
+}
+
+void DyldSharedCache::forEachPatchableUseOfExport(uint32_t imageIndex, uint32_t dylibVMOffsetOfImpl,
+                                                  void (^handler)(uint64_t cacheVMOffset,
+                                                                  MachOLoaded::PointerMetaData pmd, uint64_t addend)) const {
+    if ( header.patchInfoAddr == 0 )
+        return;
+
+    uint32_t patchVersion = patchInfoVersion();
+
+    if ( patchVersion == 1 ) {
+        // Old cache.  The patch table uses the V1 structs
+
+        // This patch table uses cache offsets, so convert from "image + offset" to cache offset
+        uint64_t mTime;
+        uint64_t inode;
+        const dyld3::MachOAnalyzer* imageMA = (dyld3::MachOAnalyzer*)(this->getIndexedImageEntry(imageIndex, mTime, inode));
+        if ( imageMA == nullptr )
+            return;
+
+        uint64_t cacheUnslidAddress = unslidLoadAddress();
+        uint32_t cacheOffsetOfImpl = (uint32_t)((imageMA->preferredLoadAddress() - cacheUnslidAddress) + dylibVMOffsetOfImpl);
+
+        // Loading a new cache so get the data from the cache header
+        const dyld_cache_patch_info_v1* patchInfo = getAddrField<dyld_cache_patch_info_v1*>(header.patchInfoAddr);
+        const dyld_cache_image_patches_v1* patchArray = getAddrField<dyld_cache_image_patches_v1*>(patchInfo->patchTableArrayAddr);
+        if (imageIndex > patchInfo->patchTableArrayCount)
+            return;
+        const dyld_cache_image_patches_v1& patch = patchArray[imageIndex];
+        if ( (patch.patchExportsStartIndex + patch.patchExportsCount) > patchInfo->patchExportArrayCount )
+            return;
+        const dyld_cache_patchable_export_v1* patchExports = getAddrField<dyld_cache_patchable_export_v1*>(patchInfo->patchExportArrayAddr);
+        const dyld_cache_patchable_location_v1* patchLocations = getAddrField<dyld_cache_patchable_location_v1*>(patchInfo->patchLocationArrayAddr);
+        for (uint64_t exportIndex = 0; exportIndex != patch.patchExportsCount; ++exportIndex) {
+            const dyld_cache_patchable_export_v1& patchExport = patchExports[patch.patchExportsStartIndex + exportIndex];
+            if ( patchExport.cacheOffsetOfImpl != cacheOffsetOfImpl )
+                continue;
+            if ( (patchExport.patchLocationsStartIndex + patchExport.patchLocationsCount) > patchInfo->patchLocationArrayCount )
+                return;
+            for (uint64_t locationIndex = 0; locationIndex != patchExport.patchLocationsCount; ++locationIndex) {
+                const dyld_cache_patchable_location_v1& patchLocation = patchLocations[patchExport.patchLocationsStartIndex + locationIndex];
+
+                dyld3::MachOLoaded::PointerMetaData pmd;
+                pmd.diversity         = patchLocation.discriminator;
+                pmd.high8             = patchLocation.high7 << 1;
+                pmd.authenticated     = patchLocation.authenticated;
+                pmd.key               = patchLocation.key;
+                pmd.usesAddrDiversity = patchLocation.usesAddressDiversity;
+
+                handler(patchLocation.cacheOffset, pmd, patchLocation.getAddend());
+            }
+        }
+        return;
+    }
+
+    // V2 and newer structs
+    if ( patchVersion == 2 ) {
+        const auto* patchInfo = getAddrField<dyld_cache_patch_info_v2*>(header.patchInfoAddr);
+        const auto* patchArray = getAddrField<dyld_cache_image_patches_v2*>(patchInfo->patchTableArrayAddr);
+        const auto* imageExports = getAddrField<dyld_cache_image_export_v2*>(patchInfo->patchImageExportsArrayAddr);
+        const auto* clientArray = getAddrField<dyld_cache_image_clients_v2*>(patchInfo->patchClientsArrayAddr);
+        const auto* clientExports = getAddrField<dyld_cache_patchable_export_v2*>(patchInfo->patchClientExportsArrayAddr);
+        const auto* patchLocations = getAddrField<dyld_cache_patchable_location_v2*>(patchInfo->patchLocationArrayAddr);
+
+        // FIXME: This shouldn't be possible.  The cache builder always adds 1 dyld_cache_image_patches_v2 per-image.  But for now
+        // be conservative to match V1.
+        if (imageIndex > patchInfo->patchTableArrayCount)
+            return;
+
+        const dyld_cache_image_patches_v2& imagePatches = patchArray[imageIndex];
+        // FIXME: This shouldn't be possible.  We should catch this in the builder instead
+        if ( (imagePatches.patchClientsStartIndex + imagePatches.patchClientsCount) > patchInfo->patchClientsArrayCount )
+            return;
+
+        uint64_t cacheUnslidAddress = unslidLoadAddress();
+
+        for (uint64_t clientIndex = 0; clientIndex != imagePatches.patchClientsCount; ++clientIndex) {
+            const dyld_cache_image_clients_v2& client = clientArray[imagePatches.patchClientsStartIndex + clientIndex];
+
+            // FIXME: This shouldn't be possible.  We should catch this in the builder instead
+            if ( (client.patchExportsStartIndex + client.patchExportsCount) > patchInfo->patchClientExportsArrayCount )
+                return;
+
+            // The callback wants a cache offset, so convert from "image + address" to a cache offset
+            uint64_t clientModTime;
+            uint64_t clientINode;
+            const dyld3::MachOAnalyzer* clientMA = (dyld3::MachOAnalyzer*)(this->getIndexedImageEntry(client.clientDylibIndex, clientModTime, clientINode));
+            if ( clientMA == nullptr )
+                return;
+
+            uint64_t clientUnslidAddress = clientMA->preferredLoadAddress();
+
+            for (uint64_t exportIndex = 0; exportIndex != client.patchExportsCount; ++exportIndex) {
+                const dyld_cache_patchable_export_v2& clientExport = clientExports[client.patchExportsStartIndex + exportIndex];
+
+                // The client export points to an image export from the image we are rooting
+                if ( clientExport.imageExportIndex > patchInfo->patchImageExportsArrayCount )
+                    return;
+
+                const dyld_cache_image_export_v2& imageExport = imageExports[clientExport.imageExportIndex];
+
+                if ( imageExport.dylibOffsetOfImpl != dylibVMOffsetOfImpl )
+                    continue;
+
+                if ( (clientExport.patchLocationsStartIndex + clientExport.patchLocationsCount) > patchInfo->patchLocationArrayCount )
+                    return;
+
+                for (uint64_t locationIndex = 0; locationIndex != clientExport.patchLocationsCount; ++locationIndex) {
+                    const dyld_cache_patchable_location_v2& patchLocation = patchLocations[clientExport.patchLocationsStartIndex + locationIndex];
+
+                    dyld3::MachOLoaded::PointerMetaData pmd;
+                    pmd.diversity         = patchLocation.discriminator;
+                    pmd.high8             = patchLocation.high7 << 1;
+                    pmd.authenticated     = patchLocation.authenticated;
+                    pmd.key               = patchLocation.key;
+                    pmd.usesAddrDiversity = patchLocation.usesAddressDiversity;
+
+                    uint64_t cacheOffset = (clientUnslidAddress + patchLocation.dylibOffsetOfUse) - cacheUnslidAddress;
+                    handler(cacheOffset, pmd, patchLocation.getAddend());
+                }
+            }
+        }
+        return;
+    }
+
+    assert(false);
+}
+
+#if !(BUILDING_LIBDYLD || BUILDING_DYLD)
+// MRM map file generator
+std::string DyldSharedCache::generateJSONMap(const char* disposition) const {
+    dyld3::json::Node cacheNode;
+
+    cacheNode.map["version"].value = "1";
+    cacheNode.map["disposition"].value = disposition;
+    cacheNode.map["base-address"].value = dyld3::json::hex(unslidLoadAddress());
+    uuid_t cache_uuid;
+    getUUID(cache_uuid);
+    uuid_string_t cache_uuidStr;
+    uuid_unparse(cache_uuid, cache_uuidStr);
+    cacheNode.map["uuid"].value = cache_uuidStr;
+
+    __block dyld3::json::Node imagesNode;
+    forEachImage(^(const mach_header *mh, const char *installName) {
+        dyld3::json::Node imageNode;
+        imageNode.map["path"].value = installName;
+        dyld3::MachOAnalyzer* ma = (dyld3::MachOAnalyzer*)mh;
+        uuid_t uuid;
+        if (ma->getUuid(uuid)) {
+            uuid_string_t uuidStr;
+            uuid_unparse(uuid, uuidStr);
+            imageNode.map["uuid"].value = uuidStr;
+        }
+
+        __block dyld3::json::Node segmentsNode;
+        ma->forEachSegment(^(const dyld3::MachOAnalyzer::SegmentInfo &info, bool &stop) {
+            dyld3::json::Node segmentNode;
+            segmentNode.map["name"].value = info.segName;
+            segmentNode.map["start-vmaddr"].value = dyld3::json::hex(info.vmAddr);
+            segmentNode.map["end-vmaddr"].value = dyld3::json::hex(info.vmAddr + info.vmSize);
+            segmentsNode.array.push_back(segmentNode);
+        });
+        imageNode.map["segments"] = segmentsNode;
+        imagesNode.array.push_back(imageNode);
+    });
+
+    cacheNode.map["images"] = imagesNode;
+
+    std::stringstream stream;
+    printJSON(cacheNode, 0, stream);
+
+    return stream.str();
+}
+
+std::string DyldSharedCache::generateJSONDependents() const {
+    std::unordered_map<std::string, std::set<std::string>> dependents;
+    computeTransitiveDependents(dependents);
+
+    std::stringstream stream;
+
+    stream << "{";
+    bool first = true;
+    for (auto p : dependents) {
+        if (!first) stream << "," << std::endl;
+        first = false;
+
+        stream << "\"" << p.first << "\" : [" << std::endl;
+        bool firstDependent = true;
+        for (const std::string & dependent : p.second) {
+            if (!firstDependent) stream << "," << std::endl;
+            firstDependent = false;
+            stream << "  \"" << dependent << "\"";
+        }
+        stream << "]" <<  std::endl;
+    }
+    stream << "}" << std::endl;
+    return stream.str();
+}
+
+#endif
+
+#if !(BUILDING_LIBDYLD || BUILDING_DYLD)
+dyld3::MachOAnalyzer::VMAddrConverter DyldSharedCache::makeVMAddrConverter(bool contentRebased) const {
+    typedef dyld3::MachOAnalyzer::VMAddrConverter VMAddrConverter;
+
+    __block VMAddrConverter::SharedCacheFormat pointerFormat = VMAddrConverter::SharedCacheFormat::none;
+    __block uint64_t pointerValueAdd = 0;
+    // With subCaches, the first cache file might not have any slide info.  In that case, walk all the files
+    // until we find one with slide info
+    forEachCache(^(const DyldSharedCache *cache, bool& stopCache) {
+        cache->forEachSlideInfo(^(uint64_t mappingStartAddress, uint64_t mappingSize, const uint8_t *mappingPagesStart, uint64_t slideInfoOffset, uint64_t slideInfoSize, const dyld_cache_slide_info *slideInfoHeader) {
+            assert(slideInfoHeader->version >= 2);
+            if ( slideInfoHeader->version == 2 ) {
+                const dyld_cache_slide_info2* slideInfo = (dyld_cache_slide_info2*)(slideInfoHeader);
+                assert(slideInfo->delta_mask == 0x00FFFF0000000000);
+                pointerFormat   = VMAddrConverter::SharedCacheFormat::v2_x86_64_tbi;
+                pointerValueAdd = slideInfo->value_add;
+            } else if ( slideInfoHeader->version == 3 ) {
+                pointerFormat   = VMAddrConverter::SharedCacheFormat::v3;
+                pointerValueAdd = unslidLoadAddress();
+            } else if ( slideInfoHeader->version == 4 ) {
+                const dyld_cache_slide_info4* slideInfo = (dyld_cache_slide_info4*)(slideInfoHeader);
+                assert(slideInfo->delta_mask == 0x00000000C0000000);
+                pointerFormat   = VMAddrConverter::SharedCacheFormat::v4;
+                pointerValueAdd = slideInfo->value_add;
+            } else {
+                assert(false);
+            }
+        });
+    });
+
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    uintptr_t slide = (uintptr_t)this - (uintptr_t)(mappings[0].address);
+
+    VMAddrConverter vmAddrConverter;
+    vmAddrConverter.preferredLoadAddress            = pointerValueAdd;
+    vmAddrConverter.slide                           = slide;
+    vmAddrConverter.chainedPointerFormat            = 0;
+    vmAddrConverter.sharedCacheChainedPointerFormat = pointerFormat;
+    vmAddrConverter.contentRebased                  = contentRebased;
+
+    return vmAddrConverter;
+}
+#endif
+
+bool DyldSharedCache::inDyldCache(const DyldSharedCache* cache, const dyld3::MachOFile* mf) {
+    // FIXME: The cache PBLS has a null state.config.dyldCache.addr when built, so we can't do the
+    // range check.  We might be able to remove the #if below if we can instead set the cache in the config.
+#if BUILDING_CACHE_BUILDER
+    return mf->inDyldCache();
+#else
+    return mf->inDyldCache() && (cache != nullptr) && ((uintptr_t)mf >= (uintptr_t)cache) && ((uintptr_t)mf < ((uintptr_t)cache + cache->mappedSize()));
+#endif
+}
+
+#if !(BUILDING_LIBDYLD || BUILDING_DYLD)
+// mmap() an shared cache file read/only but laid out like it would be at runtime
+const DyldSharedCache* DyldSharedCache::mapCacheFile(const char* path,
+                                           uint64_t baseCacheUnslidAddress,
+                                           uint8_t* buffer)
+{
+    struct stat statbuf;
+    if ( ::stat(path, &statbuf) ) {
+        fprintf(stderr, "Error: stat failed for dyld shared cache at %s\n", path);
+        return nullptr;
+    }
+
+    int cache_fd = ::open(path, O_RDONLY);
+    if (cache_fd < 0) {
+        fprintf(stderr, "Error: failed to open shared cache file at %s\n", path);
+        return nullptr;
+    }
+
+    uint8_t  firstPage[4096];
+    if ( ::pread(cache_fd, firstPage, 4096, 0) != 4096 ) {
+        fprintf(stderr, "Error: failed to read shared cache file at %s\n", path);
+        return nullptr;
+    }
+    const dyld_cache_header*       header   = (dyld_cache_header*)firstPage;
+   if ( header->mappingCount == 0 ) {
+        fprintf(stderr, "Error: No mapping in shared cache file at %s\n", path);
+        return nullptr;
+    }
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)(firstPage + header->mappingOffset);
+    const dyld_cache_mapping_info* lastMapping = &mappings[header->mappingCount - 1];
+
+    // Allocate enough space for the cache and all subCaches
+    uint64_t subCacheBufferOffset = 0;
+    if ( baseCacheUnslidAddress == 0 ) {
+        size_t vmSize = (size_t)header->sharedRegionSize;
+        // If the size is 0, then we might be looking directly at a sub cache.  In that case just allocate a buffer large
+        // enough for its mappings.
+        if ( vmSize == 0 ) {
+            vmSize = (size_t)(lastMapping->address + lastMapping->size - mappings[0].address);
+        }
+        vm_address_t result;
+        kern_return_t r = ::vm_allocate(mach_task_self(), &result, vmSize, VM_FLAGS_ANYWHERE);
+        if ( r != KERN_SUCCESS ) {
+            fprintf(stderr, "Error: failed to allocate space to load shared cache file at %s\n", path);
+            return nullptr;
+        }
+        buffer = (uint8_t*)result;
+    } else {
+        subCacheBufferOffset = mappings[0].address - baseCacheUnslidAddress;
+    }
+
+    for (uint32_t i=0; i < header->mappingCount; ++i) {
+        uint64_t mappingAddressOffset = mappings[i].address - mappings[0].address;
+        void* mapped_cache = ::mmap((void*)(buffer + mappingAddressOffset + subCacheBufferOffset), (size_t)mappings[i].size,
+                                    mappings[i].maxProt, MAP_FIXED | MAP_PRIVATE, cache_fd, mappings[i].fileOffset);
+        if (mapped_cache == MAP_FAILED) {
+            fprintf(stderr, "Error: mmap() for shared cache at %s failed, errno=%d\n", path, errno);
+            return nullptr;
+        }
+    }
+    ::close(cache_fd);
+
+    return (DyldSharedCache*)(buffer + subCacheBufferOffset);
+}
+
+std::vector<const DyldSharedCache*> DyldSharedCache::mapCacheFiles(const char* path)
+{
+    const DyldSharedCache* cache = DyldSharedCache::mapCacheFile(path, 0, nullptr);
+    if ( cache == nullptr )
+        return {};
+
+    std::vector<const DyldSharedCache*> caches;
+    caches.push_back(cache);
+
+    // Load all subcaches, if we have them
+    if ( cache->header.mappingOffset >= __offsetof(dyld_cache_header, subCacheArrayCount) ) {
+        if ( cache->header.subCacheArrayCount != 0 ) {
+            const dyld_subcache_entry* subCacheEntries = (dyld_subcache_entry*)((uint8_t*)cache + cache->header.subCacheArrayOffset);
+
+            for (uint32_t i = 0; i != cache->header.subCacheArrayCount; ++i) {
+                std::string subCachePath = std::string(path) + "." + dyld3::json::decimal(i + 1);
+                const DyldSharedCache* subCache = DyldSharedCache::mapCacheFile(subCachePath.c_str(), cache->unslidLoadAddress(), (uint8_t*)cache);
+                if ( subCache == nullptr )
+                    return {};
+
+                if ( memcmp(subCache->header.uuid, subCacheEntries[i].uuid, 16) != 0 ) {
+                    uuid_string_t expectedUUIDString;
+                    uuid_unparse_upper(subCacheEntries[i].uuid, expectedUUIDString);
+                    uuid_string_t foundUUIDString;
+                    uuid_unparse_upper(subCache->header.uuid, foundUUIDString);
+                    fprintf(stderr, "Error: SubCache[%i] UUID mismatch.  Expected %s, got %s\n", i, expectedUUIDString, foundUUIDString);
+                    return {};
+                }
+
+                caches.push_back(subCache);
+            }
+        }
+    }
+
+    return caches;
+}
+
+void DyldSharedCache::applyCacheRebases() const {
+    auto rebaseChainV4 = ^(uint8_t* pageContent, uint16_t startOffset, const dyld_cache_slide_info4* slideInfo)
+    {
+        const uintptr_t   deltaMask    = (uintptr_t)(slideInfo->delta_mask);
+        const uintptr_t   valueMask    = ~deltaMask;
+        //const uintptr_t   valueAdd     = (uintptr_t)(slideInfo->value_add);
+        const unsigned    deltaShift   = __builtin_ctzll(deltaMask) - 2;
+
+        uint32_t pageOffset = startOffset;
+        uint32_t delta = 1;
+        while ( delta != 0 ) {
+            uint8_t* loc = pageContent + pageOffset;
+            uintptr_t rawValue = *((uintptr_t*)loc);
+            delta = (uint32_t)((rawValue & deltaMask) >> deltaShift);
+            pageOffset += delta;
+            uintptr_t value = (rawValue & valueMask);
+            if ( (value & 0xFFFF8000) == 0 ) {
+               // small positive non-pointer, use as-is
+            }
+            else if ( (value & 0x3FFF8000) == 0x3FFF8000 ) {
+               // small negative non-pointer
+               value |= 0xC0000000;
+            }
+            else {
+                // We don't want to fix up pointers, just the stolen integer slots above
+                // value += valueAdd;
+                // value += slideAmount;
+                continue;
+            }
+            *((uintptr_t*)loc) = value;
+            //dyld::log("         pageOffset=0x%03X, loc=%p, org value=0x%08llX, new value=0x%08llX, delta=0x%X\n", pageOffset, loc, (uint64_t)rawValue, (uint64_t)value, delta);
+        }
+    };
+
+    // On watchOS, the slide info v4 format steals high bits of integers.  We need to undo these
+    this->forEachCache(^(const DyldSharedCache *subCache, bool& stopCache) {
+        subCache->forEachSlideInfo(^(uint64_t mappingStartAddress, uint64_t mappingSize, const uint8_t *dataPagesStart,
+                                      uint64_t slideInfoOffset, uint64_t slideInfoSize, const dyld_cache_slide_info *slideInfo) {
+            if ( slideInfo->version == 4) {
+                const dyld_cache_slide_info4* slideHeader = (dyld_cache_slide_info4*)slideInfo;
+                const uint32_t  page_size = slideHeader->page_size;
+                const uint16_t* page_starts = (uint16_t*)((long)(slideInfo) + slideHeader->page_starts_offset);
+                const uint16_t* page_extras = (uint16_t*)((long)(slideInfo) + slideHeader->page_extras_offset);
+                for (int i=0; i < slideHeader->page_starts_count; ++i) {
+                    uint8_t* page = (uint8_t*)(long)(dataPagesStart + (page_size*i));
+                    uint16_t pageEntry = page_starts[i];
+                    //dyld::log("page[%d]: page_starts[i]=0x%04X\n", i, pageEntry);
+                    if ( pageEntry == DYLD_CACHE_SLIDE4_PAGE_NO_REBASE )
+                        continue;
+                    if ( pageEntry & DYLD_CACHE_SLIDE4_PAGE_USE_EXTRA ) {
+                        uint16_t chainIndex = (pageEntry & DYLD_CACHE_SLIDE4_PAGE_INDEX);
+                        bool done = false;
+                        while ( !done ) {
+                            uint16_t pInfo = page_extras[chainIndex];
+                            uint16_t pageStartOffset = (pInfo & DYLD_CACHE_SLIDE4_PAGE_INDEX)*4;
+                            //dyld::log("     chain[%d] pageOffset=0x%03X\n", chainIndex, pageStartOffset);
+                            rebaseChainV4(page, pageStartOffset, slideHeader);
+                            done = (pInfo & DYLD_CACHE_SLIDE4_PAGE_EXTRA_END);
+                            ++chainIndex;
+                        }
+                    }
+                    else {
+                        uint32_t pageOffset = pageEntry * 4;
+                        //dyld::log("     start pageOffset=0x%03X\n", pageOffset);
+                        rebaseChainV4(page, pageOffset, slideHeader);
+                    }
+                }
+            }
+        });
+    });
+}
+
+#endif
+const dyld_cache_slide_info* DyldSharedCache::legacyCacheSlideInfo() const
+{
+    assert(header.mappingOffset <= __offsetof(dyld_cache_header, mappingWithSlideOffset));
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    uintptr_t slide = (uintptr_t)this - (uintptr_t)(mappings[0].address);
+
+    uint64_t offsetInLinkEditRegion = (header.slideInfoOffsetUnused - mappings[2].fileOffset);
+    return (dyld_cache_slide_info*)((uint8_t*)(mappings[2].address) + slide + offsetInLinkEditRegion);
+}
+
+const dyld_cache_mapping_info* DyldSharedCache::legacyCacheDataRegionMapping() const
+{
+    assert(header.mappingOffset <= __offsetof(dyld_cache_header, mappingWithSlideOffset));
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    return &mappings[1];
+}
+
+const uint8_t* DyldSharedCache::legacyCacheDataRegionBuffer() const
+{
+    assert(header.mappingOffset <= __offsetof(dyld_cache_header, mappingWithSlideOffset));
+    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)((char*)this + header.mappingOffset);
+    uintptr_t slide = (uintptr_t)this - (uintptr_t)(mappings[0].address);
+    
+    return (uint8_t*)(legacyCacheDataRegionMapping()->address) + slide;
+}
+
+const objc_opt::objc_opt_t* DyldSharedCache::objcOpt() const
+{
+    // Find the objc image
+    __block const dyld3::MachOAnalyzer* objcMA = nullptr;
+    uint32_t imageIndex;
+    if ( this->hasImagePath("/usr/lib/libobjc.A.dylib", imageIndex) ) {
+        uint64_t mTime;
+        uint64_t inode;
+        objcMA = (dyld3::MachOAnalyzer*)(this->getIndexedImageEntry(imageIndex, mTime, inode));
+    } else {
+#if BUILDING_CACHE_BUILDER
+        // The cache builder might call this before prior to the trie being written.
+        // In this case, we can still use the image list which is always available
+        forEachImage(^(const mach_header *mh, const char *installName) {
+            if ( !strcmp(installName, "/usr/lib/libobjc.A.dylib") )
+                objcMA = (const dyld3::MachOAnalyzer*)mh;
+        });
+#endif
+    }
+
+    if ( objcMA == nullptr )
+        return nullptr;
+
+    // If we found the objc image, then try to find the read-only data inside.
+    __block const objc_opt::objc_opt_t* objcROContent = nullptr;
+    int64_t slide = objcMA->getSlide();
+    objcMA->forEachSection(^(const dyld3::MachOAnalyzer::SectionInfo& info, bool malformedSectionRange, bool& stop) {
+        if (strcmp(info.segInfo.segName, "__TEXT") != 0)
+            return;
+        if (strcmp(info.sectName, "__objc_opt_ro") != 0)
+            return;
+        if ( malformedSectionRange ) {
+            stop = true;
+            return;
+        }
+        objcROContent = (objc_opt::objc_opt_t*)(info.sectAddr + slide);
+    });
+    if ( objcROContent == nullptr )
+        return nullptr;
+
+    // FIXME: We should fix this once objc and dyld are both in-sync with Large Caches changes
+    if ( objcROContent->version == objc_opt::VERSION || (objcROContent->version == 15) )
+        return objcROContent;
+
+    return nullptr;
+}
+
+const void* DyldSharedCache::objcOptPtrs() const
+{
+    // Find the objc image
+    const dyld3::MachOAnalyzer* objcMA = nullptr;
+    uint32_t imageIndex;
+    if ( this->hasImagePath("/usr/lib/libobjc.A.dylib", imageIndex) ) {
+        uint64_t mTime;
+        uint64_t inode;
+        objcMA = (dyld3::MachOAnalyzer*)this->getIndexedImageEntry(imageIndex, mTime, inode);
+    }
+    else {
+        return nullptr;
+    }
+
+    // If we found the objc image, then try to find the read-only data inside.
+    __block const void* objcPointersContent = nullptr;
+    int64_t slide = objcMA->getSlide();
+    uint32_t pointerSize = objcMA->pointerSize();
+    objcMA->forEachSection(^(const dyld3::MachOAnalyzer::SectionInfo& info, bool malformedSectionRange, bool& stop) {
+        if ( (strncmp(info.segInfo.segName, "__DATA", 6) != 0) && (strncmp(info.segInfo.segName, "__AUTH", 6) != 0) )
+            return;
+        if (strcmp(info.sectName, "__objc_opt_ptrs") != 0)
+            return;
+        if ( info.sectSize != pointerSize ) {
+            stop = true;
+            return;
+        }
+        if ( malformedSectionRange ) {
+            stop = true;
+            return;
+        }
+        objcPointersContent = (uint8_t*)(info.sectAddr + slide);
+    });
+
+    return objcPointersContent;
+}
+
+#if !(BUILDING_LIBDYLD || BUILDING_DYLD)
+uint64_t DyldSharedCache::sharedCacheRelativeSelectorBaseVMAddress() const {
+    if ( header.mappingOffset <= __offsetof(dyld_cache_header, symbolFileUUID) )
+        return 0;
+
+    // In newer shared caches, relative method list selectors are offsets from the magic selector in libobjc
+    __block uint64_t sharedCacheRelativeSelectorBaseVMAddress = 0;
+    constexpr std::string_view magicSelector = "\xf0\x9f\xa4\xaf";
+    dyld3::MachOAnalyzer::VMAddrConverter vmAddrConverter = makeVMAddrConverter(false);
+    uint64_t sharedCacheSlide = (uint64_t)this - unslidLoadAddress();
+    forEachImage(^(const mach_header *mh, const char *installName) {
+        if ( !strcmp(installName, "/usr/lib/libobjc.A.dylib") ) {
+            const dyld3::MachOAnalyzer* ma = (const dyld3::MachOAnalyzer*)mh;
+            Diagnostics diag;
+            ma->forEachObjCSelectorReference(diag, vmAddrConverter,
+                                             ^(uint64_t selRefVMAddr, uint64_t selRefTargetVMAddr, bool &stop) {
+                const char* selValue = (const char*)(selRefTargetVMAddr + sharedCacheSlide);
+                if ( selValue == magicSelector ) {
+                    sharedCacheRelativeSelectorBaseVMAddress = selRefTargetVMAddr;
+                    stop = true;
+                }
+            });
+        }
+    });
+    return sharedCacheRelativeSelectorBaseVMAddress;
+}
+#endif
+
+const SwiftOptimizationHeader* DyldSharedCache::swiftOpt() const {
+    // check for old cache without imagesArray
+    if ( header.mappingOffset <= __offsetof(dyld_cache_header, swiftOptsSize) )
+        return nullptr;
+
+    if ( header.swiftOptsOffset == 0 )
+        return nullptr;
+
+    return (SwiftOptimizationHeader*)((char*)this + header.swiftOptsOffset);
+}
+
+std::pair<const void*, uint64_t> DyldSharedCache::getObjCConstantRange() const
+{
+    uint32_t imageIndex;
+    if ( this->hasImagePath("/usr/lib/system/libdyld.dylib", imageIndex) ) {
+        uint64_t mTime;
+        uint64_t inode;
+        const dyld3::MachOAnalyzer* libDyldMA = (dyld3::MachOAnalyzer*)this->getIndexedImageEntry(imageIndex, mTime, inode);
+        std::pair<const void*, uint64_t> ranges = { nullptr, 0 };
+#if TARGET_OS_OSX
+        ranges.first = libDyldMA->findSectionContent("__DATA", "__objc_ranges", ranges.second);
+#else
+        ranges.first = libDyldMA->findSectionContent("__DATA_CONST", "__objc_ranges", ranges.second);
+#endif
+        return ranges;
+    }
+    return { nullptr, 0 };
+}
+
+bool DyldSharedCache::hasSlideInfo() const {
+    if ( header.mappingOffset <= __offsetof(dyld_cache_header, mappingWithSlideOffset) ) {
+        return header.slideInfoSizeUnused != 0;
+    } else {
+        const dyld_cache_mapping_and_slide_info* slidableMappings = (const dyld_cache_mapping_and_slide_info*)((char*)this + header.mappingWithSlideOffset);
+        for (uint32_t i = 0; i != header.mappingWithSlideCount; ++i) {
+            if ( slidableMappings[i].slideInfoFileSize != 0 ) {
+                return true;
+            }
+        }
+    }
+    return false;
+}
+
+void DyldSharedCache::forEachSlideInfo(void (^handler)(uint64_t mappingStartAddress, uint64_t mappingSize,
+                                                       const uint8_t* mappingPagesStart,
+                                                       uint64_t slideInfoOffset, uint64_t slideInfoSize,
+                                                       const dyld_cache_slide_info* slideInfoHeader)) const {
+    if ( header.mappingOffset <= __offsetof(dyld_cache_header, mappingWithSlideOffset) ) {
+        // Old caches should get the slide info from the cache header and assume a single data region.
+        const dyld_cache_mapping_info* dataMapping = legacyCacheDataRegionMapping();
+        uint64_t dataStartAddress = dataMapping->address;
+        uint64_t dataSize = dataMapping->size;
+        const uint8_t* dataPagesStart = legacyCacheDataRegionBuffer();
+        const dyld_cache_slide_info* slideInfoHeader = legacyCacheSlideInfo();
+
+        handler(dataStartAddress, dataSize, dataPagesStart,
+                header.slideInfoOffsetUnused, header.slideInfoSizeUnused, slideInfoHeader);
+    } else {
+        const dyld_cache_mapping_and_slide_info* slidableMappings = (const dyld_cache_mapping_and_slide_info*)((char*)this + header.mappingWithSlideOffset);
+        const dyld_cache_mapping_and_slide_info* linkeditMapping = &slidableMappings[header.mappingWithSlideCount - 1];
+        uint64_t sharedCacheSlide = (uint64_t)this - unslidLoadAddress();
+
+        for (uint32_t i = 0; i != header.mappingWithSlideCount; ++i) {
+            if ( slidableMappings[i].slideInfoFileOffset != 0 ) {
+                // Get the data pages
+                uint64_t dataStartAddress = slidableMappings[i].address;
+                uint64_t dataSize = slidableMappings[i].size;
+                const uint8_t* dataPagesStart = (uint8_t*)dataStartAddress + sharedCacheSlide;
+
+                // Get the slide info
+                uint64_t offsetInLinkEditRegion = (slidableMappings[i].slideInfoFileOffset - linkeditMapping->fileOffset);
+                const dyld_cache_slide_info* slideInfoHeader = (dyld_cache_slide_info*)((uint8_t*)(linkeditMapping->address) + sharedCacheSlide + offsetInLinkEditRegion);
+                handler(dataStartAddress, dataSize, dataPagesStart,
+                        slidableMappings[i].slideInfoFileOffset, slidableMappings[i].slideInfoFileSize, slideInfoHeader);
+            }
+        }
+    }
+}
+
+const char* DyldSharedCache::getCanonicalPath(const char *path) const
+{
+    uint32_t dyldCacheImageIndex;
+    if ( hasImagePath(path, dyldCacheImageIndex) )
+        return getIndexedImagePath(dyldCacheImageIndex);
+    return nullptr;
+}
+
+#if !(BUILDING_LIBDYLD || BUILDING_DYLD)
+void DyldSharedCache::fillMachOAnalyzersMap(std::unordered_map<std::string,dyld3::MachOAnalyzer*> & dylibAnalyzers) const {
+    forEachImage(^(const mach_header *mh, const char *iteratedInstallName) {
+        dylibAnalyzers[std::string(iteratedInstallName)] = (dyld3::MachOAnalyzer*)mh;
+    });
+}
+
+void DyldSharedCache::computeReverseDependencyMapForDylib(std::unordered_map<std::string, std::set<std::string>> &reverseDependencyMap, const std::unordered_map<std::string,dyld3::MachOAnalyzer*> & dylibAnalyzers, const std::string &loadPath) const {
+    dyld3::MachOAnalyzer *ma = dylibAnalyzers.at(loadPath);
+    if (reverseDependencyMap.find(loadPath) != reverseDependencyMap.end()) return;
+    reverseDependencyMap[loadPath] = std::set<std::string>();
+
+    ma->forEachDependentDylib(^(const char *dependencyLoadPath, bool isWeak, bool isReExport, bool isUpward, uint32_t compatVersion, uint32_t curVersion, bool &stop) {
+        if (isUpward) return;
+        std::string dependencyLoadPathString = std::string(dependencyLoadPath);
+        computeReverseDependencyMapForDylib(reverseDependencyMap, dylibAnalyzers, dependencyLoadPathString);
+        reverseDependencyMap[dependencyLoadPathString].insert(loadPath);
+    });
+}
+
+// Walks the shared cache and construct the reverse dependency graph (if dylib A depends on B,
+// constructs the graph with B -> A edges)
+void DyldSharedCache::computeReverseDependencyMap(std::unordered_map<std::string, std::set<std::string>> &reverseDependencyMap) const {
+    std::unordered_map<std::string,dyld3::MachOAnalyzer*> dylibAnalyzers;
+
+    fillMachOAnalyzersMap(dylibAnalyzers);
+    forEachImage(^(const mach_header *mh, const char *installName) {
+        computeReverseDependencyMapForDylib(reverseDependencyMap, dylibAnalyzers, std::string(installName));
+    });
+}
+
+// uses the reverse dependency graph constructed above to find the recursive set of dependents for each dylib
+void DyldSharedCache::findDependentsRecursively(std::unordered_map<std::string, std::set<std::string>> &transitiveDependents, const std::unordered_map<std::string, std::set<std::string>> &reverseDependencyMap, std::set<std::string> & visited, const std::string &loadPath) const {
+
+    if (transitiveDependents.find(loadPath) != transitiveDependents.end()) {
+        return;
+    }
+
+    if (visited.find(loadPath) != visited.end()) {
+        return;
+    }
+
+    visited.insert(loadPath);
+
+    std::set<std::string> dependents;
+
+    for (const std::string & dependent : reverseDependencyMap.at(loadPath)) {
+        findDependentsRecursively(transitiveDependents, reverseDependencyMap, visited, dependent);
+        if (transitiveDependents.find(dependent) != transitiveDependents.end()) {
+            std::set<std::string> & theseTransitiveDependents = transitiveDependents.at(dependent);
+            dependents.insert(theseTransitiveDependents.begin(), theseTransitiveDependents.end());
+        }
+        dependents.insert(dependent);
+    }
+
+    transitiveDependents[loadPath] = dependents;
+}
+
+// Fills a map from each install name N to the set of install names depending on N
+void DyldSharedCache::computeTransitiveDependents(std::unordered_map<std::string, std::set<std::string>> & transitiveDependents) const {
+    std::unordered_map<std::string, std::set<std::string>> reverseDependencyMap;
+    computeReverseDependencyMap(reverseDependencyMap);
+    forEachImage(^(const mach_header *mh, const char *installName) {
+        std::set<std::string> visited;
+        findDependentsRecursively(transitiveDependents, reverseDependencyMap, visited, std::string(installName));
+    });
+}
+#endif