CacheBuilder.cpp   [plain text]

/* -*- 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 <unistd.h>
#include <dirent.h>
#include <sys/errno.h>
#include <sys/fcntl.h>
#include <sys/param.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <mach/mach.h>
#include <mach/mach_time.h>
#include <mach/mach_vm.h>
#include <mach-o/loader.h>
#include <mach-o/fat.h>
#include <mach/shared_region.h>
#include <assert.h>
#include <CommonCrypto/CommonHMAC.h>
#include <CommonCrypto/CommonDigest.h>
#include <CommonCrypto/CommonDigestSPI.h>
#include <pthread/pthread.h>

#include <string>
#include <vector>
#include <unordered_map>
#include <unordered_set>

#include "MachOFileAbstraction.hpp"
#include "CodeSigningTypes.h"
#include "DyldSharedCache.h"
#include "CacheBuilder.h"
#include "FileAbstraction.hpp"
#include "Trie.hpp"
#include "FileUtils.h"
#include "Diagnostics.h"
#include "ClosureBuilder.h"
#include "Closure.h"
#include "StringUtils.h"

#if __has_include("dyld_cache_config.h")
    #include "dyld_cache_config.h"
#else
    #define ARM_SHARED_REGION_START      0x1A000000ULL
    #define ARM_SHARED_REGION_SIZE       0x26000000ULL
    #define ARM64_SHARED_REGION_START   0x180000000ULL
    #define ARM64_SHARED_REGION_SIZE     0x40000000ULL
#endif

#ifndef ARM64_32_SHARED_REGION_START
    #define ARM64_32_SHARED_REGION_START 0x1A000000ULL
    #define ARM64_32_SHARED_REGION_SIZE  0x26000000ULL
#endif

const CacheBuilder::ArchLayout CacheBuilder::_s_archLayout[] = {
    { 0x7FFF20000000ULL,            0xEFE00000ULL,              0x40000000, 0xFFFF000000000000, "x86_64",  0,          0,          0,          12, 2, true,  true  },
    { 0x7FFF20000000ULL,            0xEFE00000ULL,              0x40000000, 0xFFFF000000000000, "x86_64h", 0,          0,          0,          12, 2, true,  true  },
    { SHARED_REGION_BASE_I386,      SHARED_REGION_SIZE_I386,    0x00200000,                0x0, "i386",    0,          0,          0,          12, 0, false, false },
    { ARM64_SHARED_REGION_START,    ARM64_SHARED_REGION_SIZE,   0x02000000, 0x00FFFF0000000000, "arm64",   0x0000C000, 0x00100000, 0x07F00000, 14, 2, false, true  },
#if SUPPORT_ARCH_arm64e
    { ARM64_SHARED_REGION_START,    ARM64_SHARED_REGION_SIZE,   0x02000000, 0x00FFFF0000000000, "arm64e",  0x0000C000, 0x00100000, 0x07F00000, 14, 2, false, true  },
#endif
#if SUPPORT_ARCH_arm64_32
    { ARM64_32_SHARED_REGION_START, ARM64_32_SHARED_REGION_SIZE,0x02000000,         0xC0000000, "arm64_32",0x0000C000, 0x00100000, 0x07F00000, 14, 6, false, false },
#endif
    { ARM_SHARED_REGION_START,      ARM_SHARED_REGION_SIZE,     0x02000000,         0xE0000000, "armv7s",  0,          0,          0,          14, 4, false, false },
    { ARM_SHARED_REGION_START,      ARM_SHARED_REGION_SIZE,     0x00400000,         0xE0000000, "armv7k",  0,          0,          0,          14, 4, false, false },
    { 0x40000000,                   0x40000000,                 0x02000000,                0x0, "sim-x86", 0,          0,          0,          14, 0, false, false }
};


// These are dylibs that may be interposed, so stubs calling into them should never be bypassed
const char* const CacheBuilder::_s_neverStubEliminate[] = {
    "/usr/lib/system/libdispatch.dylib",
    nullptr
};


CacheBuilder::CacheBuilder(const DyldSharedCache::CreateOptions& options, const dyld3::closure::FileSystem& fileSystem)
    : _options(options)
    , _fileSystem(fileSystem)
    , _fullAllocatedBuffer(0)
    , _diagnostics(options.loggingPrefix, options.verbose)
    , _archLayout(nullptr)
    , _aliasCount(0)
    , _slideInfoFileOffset(0)
    , _slideInfoBufferSizeAllocated(0)
    , _allocatedBufferSize(0)
    , _branchPoolsLinkEditStartAddr(0)
{

    std::string targetArch = options.archName;
    if ( options.forSimulator && (options.archName == "i386") )
        targetArch = "sim-x86";

    for (const ArchLayout& layout : _s_archLayout) {
        if ( layout.archName == targetArch ) {
            _archLayout = &layout;
            break;
        }
    }

    if (!_archLayout) {
        _diagnostics.error("Tool was built without support for: '%s'", targetArch.c_str());
    }
}


std::string CacheBuilder::errorMessage()
{
    return _diagnostics.errorMessage();
}

const std::set<std::string> CacheBuilder::warnings()
{
    return _diagnostics.warnings();
}

const std::set<const dyld3::MachOAnalyzer*> CacheBuilder::evictions()
{
    return _evictions;
}

void CacheBuilder::deleteBuffer()
{
    vm_deallocate(mach_task_self(), _fullAllocatedBuffer, _archLayout->sharedMemorySize);
    _fullAllocatedBuffer = 0;
    _allocatedBufferSize = 0;
}


void CacheBuilder::makeSortedDylibs(const std::vector<LoadedMachO>& dylibs, const std::unordered_map<std::string, unsigned> sortOrder)
{
    for (const LoadedMachO& dylib : dylibs) {
        _sortedDylibs.push_back({ &dylib, dylib.mappedFile.runtimePath, {} });
    }

    std::sort(_sortedDylibs.begin(), _sortedDylibs.end(), [&](const DylibInfo& a, const DylibInfo& b) {
        const auto& orderA = sortOrder.find(a.input->mappedFile.runtimePath);
        const auto& orderB = sortOrder.find(b.input->mappedFile.runtimePath);
        bool foundA = (orderA != sortOrder.end());
        bool foundB = (orderB != sortOrder.end());

        // Order all __DATA_DIRTY segments specified in the order file first, in
        // the order specified in the file, followed by any other __DATA_DIRTY
        // segments in lexicographic order.
        if ( foundA && foundB )
            return orderA->second < orderB->second;
        else if ( foundA )
            return true;
        else if ( foundB )
             return false;
        else
             return a.input->mappedFile.runtimePath < b.input->mappedFile.runtimePath;
    });
}


inline uint32_t absolutetime_to_milliseconds(uint64_t abstime)
{
    return (uint32_t)(abstime/1000/1000);
}

struct DylibAndSize
{
    const CacheBuilder::LoadedMachO*    input;
    const char*                         installName;
    uint64_t                            size;
};

uint64_t CacheBuilder::cacheOverflowAmount()
{
    if ( _archLayout->sharedRegionsAreDiscontiguous ) {
        // for macOS x86_64 cache, need to check each region for overflow
        if ( _readExecuteRegion.sizeInUse > 0x60000000 )
            return (_readExecuteRegion.sizeInUse - 0x60000000);

        if ( _readWriteRegion.sizeInUse > 0x40000000 )
            return (_readWriteRegion.sizeInUse - 0x40000000);

        if ( _readOnlyRegion.sizeInUse > 0x3FE00000 )
            return (_readOnlyRegion.sizeInUse - 0x3FE00000);
    }
    else {
        bool alreadyOptimized = (_readOnlyRegion.sizeInUse != _readOnlyRegion.bufferSize);
        uint64_t vmSize = _readOnlyRegion.unslidLoadAddress - _readExecuteRegion.unslidLoadAddress;
        if ( alreadyOptimized )
            vmSize += _readOnlyRegion.sizeInUse;
        else if ( _options.excludeLocalSymbols )
            vmSize += (_readOnlyRegion.sizeInUse * 37/100); // assume locals removal and LINKEDIT optimzation reduces LINKEDITs %25 of original size
        else
            vmSize += (_readOnlyRegion.sizeInUse * 80/100); // assume LINKEDIT optimzation reduces LINKEDITs to %80 of original size
        if ( vmSize > _archLayout->sharedMemorySize )
            return vmSize - _archLayout->sharedMemorySize;
    }
    // fits in shared region
    return 0;
}

size_t CacheBuilder::evictLeafDylibs(uint64_t reductionTarget, std::vector<const LoadedMachO*>& overflowDylibs)
{
    // build count of how many references there are to each dylib
    __block std::map<std::string, unsigned int> referenceCount;
    for (const DylibInfo& dylib : _sortedDylibs) {
        dylib.input->mappedFile.mh->forEachDependentDylib(^(const char* loadPath, bool isWeak, bool isReExport, bool isUpward, uint32_t compatVersion, uint32_t curVersion, bool &stop) {
            referenceCount[loadPath] += 1;
        });
    }

    // find all dylibs not referenced
    std::vector<DylibAndSize> unreferencedDylibs;
    for (const DylibInfo& dylib : _sortedDylibs) {
        const char* installName = dylib.input->mappedFile.mh->installName();
        if ( referenceCount.count(installName) == 0 ) {
            // conservative: sum up all segments except LINKEDIT
            __block uint64_t segsSize = 0;
            dylib.input->mappedFile.mh->forEachSegment(^(const dyld3::MachOFile::SegmentInfo& info, bool& stop) {
                if ( strcmp(info.segName, "__LINKEDIT") != 0 )
                    segsSize += info.vmSize;
            });
            unreferencedDylibs.push_back({ dylib.input, installName, segsSize });
        }
    }
    // sort leaf dylibs by size
    std::sort(unreferencedDylibs.begin(), unreferencedDylibs.end(), [&](const DylibAndSize& a, const DylibAndSize& b) {
        return ( a.size > b.size );
    });

     // build set of dylibs that if removed will allow cache to build
    for (DylibAndSize& dylib : unreferencedDylibs) {
        if ( _options.verbose )
            _diagnostics.warning("to prevent cache overflow, not caching %s", dylib.installName);
        _evictions.insert(dylib.input->mappedFile.mh);
        // Track the evicted dylibs so we can try build "other" dlopen closures for them.
        overflowDylibs.push_back(dylib.input);
        if ( dylib.size > reductionTarget )
            break;
        reductionTarget -= dylib.size;
    }

    // prune _sortedDylibs
    _sortedDylibs.erase(std::remove_if(_sortedDylibs.begin(), _sortedDylibs.end(), [&](const DylibInfo& dylib) {
        return (_evictions.count(dylib.input->mappedFile.mh) != 0);
    }),_sortedDylibs.end());

    return _evictions.size();
}

// Handles building a list of input files to the CacheBuilder itself.
class CacheInputBuilder {
public:
    CacheInputBuilder(const dyld3::closure::FileSystem& fileSystem,
                      std::string reqArchitecture, dyld3::Platform reqPlatform)
    : fileSystem(fileSystem), reqArchitecture(reqArchitecture), reqPlatform(reqPlatform) { }

    // Loads and maps any MachOs in the given list of files.
    void loadMachOs(std::vector<CacheBuilder::InputFile>& inputFiles,
                    std::vector<CacheBuilder::LoadedMachO>& dylibsToCache,
                    std::vector<CacheBuilder::LoadedMachO>& otherDylibs,
                    std::vector<CacheBuilder::LoadedMachO>& executables,
                    std::vector<CacheBuilder::LoadedMachO>& couldNotLoadFiles) {

        std::map<std::string, uint64_t> dylibInstallNameMap;
        for (CacheBuilder::InputFile& inputFile : inputFiles) {
            dyld3::closure::LoadedFileInfo loadedFileInfo = dyld3::MachOAnalyzer::load(inputFile.diag, fileSystem, inputFile.path, reqArchitecture.c_str(), reqPlatform);
            const dyld3::MachOAnalyzer* ma = (const dyld3::MachOAnalyzer*)loadedFileInfo.fileContent;
            if (ma == nullptr) {
                couldNotLoadFiles.emplace_back((CacheBuilder::LoadedMachO){ DyldSharedCache::MappedMachO(), loadedFileInfo, &inputFile });
                continue;
            }

            DyldSharedCache::MappedMachO mappedFile(inputFile.path, ma, loadedFileInfo.sliceLen, false, false,
                                                    loadedFileInfo.sliceOffset, loadedFileInfo.mtime, loadedFileInfo.inode);

            // The file can be loaded with the given slice, but we may still want to exlude it from the cache.
            if (ma->isDylib()) {
                std::string installName = ma->installName();

                // Let the platform exclude the file before we do anything else.
                if (platformExcludesInstallName(installName)) {
                    inputFile.diag.verbose("Platform excluded file\n");
                    fileSystem.unloadFile(loadedFileInfo);
                    continue;
                }

                if (!ma->canBePlacedInDyldCache(inputFile.path, ^(const char* msg) {
                    inputFile.diag.warning("Dylib located at '%s' cannot be placed in cache because: %s", inputFile.path, msg);
                })) {
                    // TODO: Add exclusion lists here?
                    // Probably not as we already applied the dylib exclusion list.
                    otherDylibs.emplace_back((CacheBuilder::LoadedMachO){ mappedFile, loadedFileInfo, &inputFile });
                    continue;
                }

                // Otherwise see if we have another file with this install name
                auto iteratorAndInserted = dylibInstallNameMap.insert(std::make_pair(installName, dylibsToCache.size()));
                if (iteratorAndInserted.second) {
                    // We inserted the dylib so we haven't seen another with this name.
                    if (installName[0] != '@' && installName != inputFile.path) {
                        inputFile.diag.warning("Dylib located at '%s' has installname '%s'", inputFile.path, installName.c_str());
                    }

                    dylibsToCache.emplace_back((CacheBuilder::LoadedMachO){ mappedFile, loadedFileInfo, &inputFile });
                } else {
                    // We didn't insert this one so we've seen it before.
                    CacheBuilder::LoadedMachO& previousLoadedMachO = dylibsToCache[iteratorAndInserted.first->second];
                    inputFile.diag.warning("Multiple dylibs claim installname '%s' ('%s' and '%s')", installName.c_str(), inputFile.path, previousLoadedMachO.mappedFile.runtimePath.c_str());

                    // This is the "Good" one, overwrite
                    if (inputFile.path == installName) {
                        // Unload the old one
                        fileSystem.unloadFile(previousLoadedMachO.loadedFileInfo);

                        // And replace with this one.
                        previousLoadedMachO.mappedFile = mappedFile;
                        previousLoadedMachO.loadedFileInfo = loadedFileInfo;
                    }
                }
            } else if (ma->isBundle()) {
                // TODO: Add exclusion lists here?
                otherDylibs.emplace_back((CacheBuilder::LoadedMachO){ mappedFile, loadedFileInfo, &inputFile });
            } else if (ma->isDynamicExecutable()) {
                if (platformExcludesExecutablePath_macOS(inputFile.path)) {
                    inputFile.diag.verbose("Platform excluded file\n");
                    fileSystem.unloadFile(loadedFileInfo);
                    continue;
                }
                executables.emplace_back((CacheBuilder::LoadedMachO){ mappedFile, loadedFileInfo, &inputFile });
            } else {
                inputFile.diag.verbose("Unsupported mach file type\n");
                fileSystem.unloadFile(loadedFileInfo);
            }
        }
    }

private:



    static bool platformExcludesInstallName_macOS(const std::string& installName) {
        return false;
    }

    static bool platformExcludesInstallName_iOS(const std::string& installName) {
        if ( installName == "/System/Library/Caches/com.apple.xpc/sdk.dylib" )
            return true;
        if ( installName == "/System/Library/Caches/com.apple.xpcd/xpcd_cache.dylib" )
            return true;
        return false;
    }

    static bool platformExcludesInstallName_tvOS(const std::string& installName) {
        return platformExcludesInstallName_iOS(installName);
    }

    static bool platformExcludesInstallName_watchOS(const std::string& installName) {
        return platformExcludesInstallName_iOS(installName);
    }

    static bool platformExcludesInstallName_bridgeOS(const std::string& installName) {
        return platformExcludesInstallName_iOS(installName);
    }

    // Returns true if the current platform requires that this install name be excluded from the shared cache
    // Note that this overrides any exclusion from anywhere else.
    bool platformExcludesInstallName(const std::string& installName) {
        switch (reqPlatform) {
            case dyld3::Platform::unknown:
                return false;
            case dyld3::Platform::macOS:
                return platformExcludesInstallName_macOS(installName);
            case dyld3::Platform::iOS:
                return platformExcludesInstallName_iOS(installName);
            case dyld3::Platform::tvOS:
                return platformExcludesInstallName_tvOS(installName);
            case dyld3::Platform::watchOS:
                return platformExcludesInstallName_watchOS(installName);
            case dyld3::Platform::bridgeOS:
                return platformExcludesInstallName_bridgeOS(installName);
            case dyld3::Platform::iOSMac:
                return false;
            case dyld3::Platform::iOS_simulator:
                return false;
            case dyld3::Platform::tvOS_simulator:
                return false;
            case dyld3::Platform::watchOS_simulator:
                return false;
        }
    }




    static bool platformExcludesExecutablePath_macOS(const std::string& path) {
        return false;
    }

    static bool platformExcludesExecutablePath_iOS(const std::string& path) {
        //HACK exclude all launchd and installd variants until we can do something about xpcd_cache.dylib and friends
        if (path == "/sbin/launchd"
            || path == "/usr/local/sbin/launchd.debug"
            || path == "/usr/local/sbin/launchd.development"
            || path == "/usr/libexec/installd") {
            return true;
        }
        return false;
    }

    static bool platformExcludesExecutablePath_tvOS(const std::string& path) {
        return platformExcludesExecutablePath_iOS(path);
    }

    static bool platformExcludesExecutablePath_watchOS(const std::string& path) {
        return platformExcludesExecutablePath_iOS(path);
    }

    static bool platformExcludesExecutablePath_bridgeOS(const std::string& path) {
        return platformExcludesExecutablePath_iOS(path);
    }

    // Returns true if the current platform requires that this path be excluded from the shared cache
    // Note that this overrides any exclusion from anywhere else.
    bool platformExcludesExecutablePath(const std::string& path) {
        switch (reqPlatform) {
            case dyld3::Platform::unknown:
                return false;
            case dyld3::Platform::macOS:
                return platformExcludesExecutablePath_macOS(path);
            case dyld3::Platform::iOS:
                return platformExcludesExecutablePath_iOS(path);
            case dyld3::Platform::tvOS:
                return platformExcludesExecutablePath_tvOS(path);
            case dyld3::Platform::watchOS:
                return platformExcludesExecutablePath_watchOS(path);
            case dyld3::Platform::bridgeOS:
                return platformExcludesExecutablePath_bridgeOS(path);
            case dyld3::Platform::iOSMac:
                return false;
            case dyld3::Platform::iOS_simulator:
                return false;
            case dyld3::Platform::tvOS_simulator:
                return false;
            case dyld3::Platform::watchOS_simulator:
                return false;
        }
    }

    const dyld3::closure::FileSystem&                   fileSystem;
    std::string                                         reqArchitecture;
    dyld3::Platform                                     reqPlatform;
};

static void verifySelfContained(std::vector<CacheBuilder::LoadedMachO>& dylibsToCache,
                                std::vector<CacheBuilder::LoadedMachO>& otherDylibs,
                                std::vector<CacheBuilder::LoadedMachO>& couldNotLoadFiles)
{
    // build map of dylibs
    __block std::map<std::string, const CacheBuilder::LoadedMachO*> knownDylibs;
    __block std::map<std::string, const CacheBuilder::LoadedMachO*> allDylibs;
    for (const CacheBuilder::LoadedMachO& dylib : dylibsToCache) {
        knownDylibs.insert({ dylib.mappedFile.runtimePath, &dylib });
        allDylibs.insert({ dylib.mappedFile.runtimePath, &dylib });
        if (const char* installName = dylib.mappedFile.mh->installName()) {
            knownDylibs.insert({ installName, &dylib });
            allDylibs.insert({ installName, &dylib });
        }
    }

    for (const CacheBuilder::LoadedMachO& dylib : otherDylibs) {
        allDylibs.insert({ dylib.mappedFile.runtimePath, &dylib });
        if (const char* installName = dylib.mappedFile.mh->installName())
            allDylibs.insert({ installName, &dylib });
    }

    for (const CacheBuilder::LoadedMachO& dylib : couldNotLoadFiles) {
        allDylibs.insert({ dylib.inputFile->path, &dylib });
    }

    // check all dependencies to assure every dylib in cache only depends on other dylibs in cache
    __block std::map<std::string, std::set<std::string>> badDylibs;
    __block bool doAgain = true;
    while ( doAgain ) {
        doAgain = false;
        // scan dylib list making sure all dependents are in dylib list
        for (const CacheBuilder::LoadedMachO& dylib : dylibsToCache) {
            if ( badDylibs.count(dylib.mappedFile.runtimePath) != 0 )
                continue;
            dylib.mappedFile.mh->forEachDependentDylib(^(const char* loadPath, bool isWeak, bool isReExport, bool isUpward, uint32_t compatVersion, uint32_t curVersion, bool& stop) {
                if (isWeak)
                    return;
                if ( knownDylibs.count(loadPath) == 0 ) {
                    badDylibs[dylib.mappedFile.runtimePath].insert(std::string("Could not find dependency '") + loadPath + "'");
                    knownDylibs.erase(dylib.mappedFile.runtimePath);
                    knownDylibs.erase(dylib.mappedFile.mh->installName());
                    doAgain = true;
                }
            });
        }
    }

    // Now walk the dylibs which depend on missing dylibs and see if any of them are required binaries.
    for (auto badDylibsIterator : badDylibs) {
        const std::string& dylibRuntimePath = badDylibsIterator.first;
        auto requiredDylibIterator = allDylibs.find(dylibRuntimePath);
        if (requiredDylibIterator == allDylibs.end())
            continue;
        if (!requiredDylibIterator->second->inputFile->mustBeIncluded())
            continue;
        // This dylib is required so mark all dependencies as requried too
        __block std::vector<const CacheBuilder::LoadedMachO*> worklist;
        worklist.push_back(requiredDylibIterator->second);
        while (!worklist.empty()) {
            const CacheBuilder::LoadedMachO* dylib = worklist.back();
            worklist.pop_back();
            if (!dylib->mappedFile.mh)
                continue;
            dylib->mappedFile.mh->forEachDependentDylib(^(const char* loadPath, bool isWeak, bool isReExport, bool isUpward, uint32_t compatVersion, uint32_t curVersion, bool& stop) {
                if (isWeak)
                    return;
                auto dylibIterator = allDylibs.find(loadPath);
                if (dylibIterator != allDylibs.end()) {
                    if (dylibIterator->second->inputFile->state == CacheBuilder::InputFile::Unset) {
                        dylibIterator->second->inputFile->state = CacheBuilder::InputFile::MustBeIncludedForDependent;
                        worklist.push_back(dylibIterator->second);
                    }
                }
            });
        }
    }

    // FIXME: Make this an option we can pass in
    const bool evictLeafDylibs = true;
    if (evictLeafDylibs) {
        doAgain = true;
        while ( doAgain ) {
            doAgain = false;

            // build count of how many references there are to each dylib
            __block std::set<std::string> referencedDylibs;
            for (const CacheBuilder::LoadedMachO& dylib : dylibsToCache) {
                if ( badDylibs.count(dylib.mappedFile.runtimePath) != 0 )
                    continue;
                dylib.mappedFile.mh->forEachDependentDylib(^(const char* loadPath, bool isWeak, bool isReExport, bool isUpward, uint32_t compatVersion, uint32_t curVersion, bool &stop) {
                    referencedDylibs.insert(loadPath);
                });
            }

            // find all dylibs not referenced
            std::vector<DylibAndSize> unreferencedDylibs;
            for (const CacheBuilder::LoadedMachO& dylib : dylibsToCache) {
                if ( badDylibs.count(dylib.mappedFile.runtimePath) != 0 )
                    continue;
                const char* installName = dylib.mappedFile.mh->installName();
                if ( (referencedDylibs.count(installName) == 0) && (dylib.inputFile->state == CacheBuilder::InputFile::MustBeExcludedIfUnused) ) {
                    badDylibs[dylib.mappedFile.runtimePath].insert(std::string("It has been explicitly excluded as it is unused"));
                    doAgain = true;
                }
            }
        }
    }

    // Move bad dylibs from dylibs to cache to other dylibs.
    for (const CacheBuilder::LoadedMachO& dylib : dylibsToCache) {
        auto i = badDylibs.find(dylib.mappedFile.runtimePath);
        if ( i != badDylibs.end()) {
            otherDylibs.push_back(dylib);
            for (const std::string& reason : i->second )
                otherDylibs.back().inputFile->diag.warning("Dylib located at '%s' not placed in shared cache because: %s", dylib.mappedFile.runtimePath.c_str(), reason.c_str());
        }
    }

    const auto& badDylibsLambdaRef = badDylibs;
    dylibsToCache.erase(std::remove_if(dylibsToCache.begin(), dylibsToCache.end(), [&](const CacheBuilder::LoadedMachO& dylib) {
        if (badDylibsLambdaRef.find(dylib.mappedFile.runtimePath) != badDylibsLambdaRef.end())
            return true;
        return false;
    }), dylibsToCache.end());
}

// This is the new build API which takes the raw files (which could be FAT) and tries to build a cache from them.
// We should remove the other build() method, or make it private so that this can wrap it.
void CacheBuilder::build(std::vector<CacheBuilder::InputFile>& inputFiles,
                         std::vector<DyldSharedCache::FileAlias>& aliases) {
    // First filter down to files which are actually MachO's
    CacheInputBuilder cacheInputBuilder(_fileSystem, _archLayout->archName, _options.platform);

    std::vector<LoadedMachO> dylibsToCache;
    std::vector<LoadedMachO> otherDylibs;
    std::vector<LoadedMachO> executables;
    std::vector<LoadedMachO> couldNotLoadFiles;
    cacheInputBuilder.loadMachOs(inputFiles, dylibsToCache, otherDylibs, executables, couldNotLoadFiles);

    verifySelfContained(dylibsToCache, otherDylibs, couldNotLoadFiles);

    // Check for required binaries before we try to build the cache
    if (!_diagnostics.hasError()) {
        // If we succeeded in building, then now see if there was a missing required file, and if so why its missing.
        std::string errorString;
        for (const LoadedMachO& dylib : otherDylibs) {
            if (dylib.inputFile->mustBeIncluded()) {
                // An error loading a required file must be propagated up to the top level diagnostic handler.
                bool gotWarning = false;
                for (const std::string& warning : dylib.inputFile->diag.warnings()) {
                    gotWarning = true;
                    std::string message = warning;
                    if (message.back() == '\n')
                        message.pop_back();
                    if (!errorString.empty())
                        errorString += "ERROR: ";
                    errorString += "Required binary was not included in the shared cache '" + std::string(dylib.inputFile->path) + "' because: " + message + "\n";
                }
                if (!gotWarning) {
                    if (!errorString.empty())
                        errorString += "ERROR: ";
                    errorString += "Required binary was not included in the shared cache '" + std::string(dylib.inputFile->path) + "' because: 'unknown error.  Please report to dyld'\n";
                }
            }
        }
        for (const LoadedMachO& dylib : couldNotLoadFiles) {
            if (dylib.inputFile->mustBeIncluded()) {
                if (dylib.inputFile->diag.hasError()) {
                    if (!errorString.empty())
                        errorString += "ERROR: ";
                    errorString += "Required binary was not included in the shared cache '" + std::string(dylib.inputFile->path) + "' because: " + dylib.inputFile->diag.errorMessage() + "\n";
                } else {
                    if (!errorString.empty())
                        errorString += "ERROR: ";
                    errorString += "Required binary was not included in the shared cache '" + std::string(dylib.inputFile->path) + "' because: 'unknown error.  Please report to dyld'\n";

                }
            }
        }
        if (!errorString.empty()) {
            _diagnostics.error("%s", errorString.c_str());
        }
    }

    if (!_diagnostics.hasError())
        build(dylibsToCache, otherDylibs, executables, aliases);

    if (!_diagnostics.hasError()) {
        // If we succeeded in building, then now see if there was a missing required file, and if so why its missing.
        std::string errorString;
        for (CacheBuilder::InputFile& inputFile : inputFiles) {
            if (inputFile.mustBeIncluded() && inputFile.diag.hasError()) {
                // An error loading a required file must be propagated up to the top level diagnostic handler.
                std::string message = inputFile.diag.errorMessage();
                if (message.back() == '\n')
                    message.pop_back();
                errorString += "Required binary was not included in the shared cache '" + std::string(inputFile.path) + "' because: " + message + "\n";
            }
        }
        if (!errorString.empty()) {
            _diagnostics.error("%s", errorString.c_str());
        }
    }

    // Add all the warnings from the input files to the top level warnings on the main diagnostics object.
    for (CacheBuilder::InputFile& inputFile : inputFiles) {
        for (const std::string& warning : inputFile.diag.warnings())
            _diagnostics.warning("%s", warning.c_str());
    }

    // Clean up the loaded files
    for (LoadedMachO& loadedMachO : dylibsToCache)
        _fileSystem.unloadFile(loadedMachO.loadedFileInfo);
    for (LoadedMachO& loadedMachO : otherDylibs)
        _fileSystem.unloadFile(loadedMachO.loadedFileInfo);
    for (LoadedMachO& loadedMachO : executables)
        _fileSystem.unloadFile(loadedMachO.loadedFileInfo);
}

void CacheBuilder::build(const std::vector<DyldSharedCache::MappedMachO>& dylibs,
                         const std::vector<DyldSharedCache::MappedMachO>& otherOsDylibsInput,
                         const std::vector<DyldSharedCache::MappedMachO>& osExecutables,
                         std::vector<DyldSharedCache::FileAlias>& aliases) {

    std::vector<LoadedMachO> dylibsToCache;
    std::vector<LoadedMachO> otherDylibs;
    std::vector<LoadedMachO> executables;

    for (const DyldSharedCache::MappedMachO& mappedMachO : dylibs) {
        dyld3::closure::LoadedFileInfo loadedFileInfo;
        loadedFileInfo.fileContent      = mappedMachO.mh;
        loadedFileInfo.fileContentLen   = mappedMachO.length;
        loadedFileInfo.sliceOffset      = mappedMachO.sliceFileOffset;
        loadedFileInfo.sliceLen         = mappedMachO.length;
        loadedFileInfo.inode            = mappedMachO.inode;
        loadedFileInfo.mtime            = mappedMachO.modTime;
        loadedFileInfo.path             = mappedMachO.runtimePath.c_str();
        dylibsToCache.emplace_back((LoadedMachO){ mappedMachO, loadedFileInfo, nullptr });
    }

    for (const DyldSharedCache::MappedMachO& mappedMachO : otherOsDylibsInput) {
        dyld3::closure::LoadedFileInfo loadedFileInfo;
        loadedFileInfo.fileContent      = mappedMachO.mh;
        loadedFileInfo.fileContentLen   = mappedMachO.length;
        loadedFileInfo.sliceOffset      = mappedMachO.sliceFileOffset;
        loadedFileInfo.sliceLen         = mappedMachO.length;
        loadedFileInfo.inode            = mappedMachO.inode;
        loadedFileInfo.mtime            = mappedMachO.modTime;
        loadedFileInfo.path             = mappedMachO.runtimePath.c_str();
        otherDylibs.emplace_back((LoadedMachO){ mappedMachO, loadedFileInfo, nullptr });
    }

    for (const DyldSharedCache::MappedMachO& mappedMachO : osExecutables) {
        dyld3::closure::LoadedFileInfo loadedFileInfo;
        loadedFileInfo.fileContent      = mappedMachO.mh;
        loadedFileInfo.fileContentLen   = mappedMachO.length;
        loadedFileInfo.sliceOffset      = mappedMachO.sliceFileOffset;
        loadedFileInfo.sliceLen         = mappedMachO.length;
        loadedFileInfo.inode            = mappedMachO.inode;
        loadedFileInfo.mtime            = mappedMachO.modTime;
        loadedFileInfo.path             = mappedMachO.runtimePath.c_str();
        executables.emplace_back((LoadedMachO){ mappedMachO, loadedFileInfo, nullptr });
    }

    build(dylibsToCache, otherDylibs, executables, aliases);
}

void CacheBuilder::build(const std::vector<LoadedMachO>& dylibs,
                         const std::vector<LoadedMachO>& otherOsDylibsInput,
                         const std::vector<LoadedMachO>& osExecutables,
                         std::vector<DyldSharedCache::FileAlias>& aliases)
{
    // <rdar://problem/21317611> error out instead of crash if cache has no dylibs
    // FIXME: plist should specify required vs optional dylibs
    if ( dylibs.size() < 30 ) {
        _diagnostics.error("missing required minimum set of dylibs");
        return;
    }
    uint64_t t1 = mach_absolute_time();

    // make copy of dylib list and sort
    makeSortedDylibs(dylibs, _options.dylibOrdering);

    // allocate space used by largest possible cache plus room for LINKEDITS before optimization
    _allocatedBufferSize = _archLayout->sharedMemorySize * 1.50;
    if ( vm_allocate(mach_task_self(), &_fullAllocatedBuffer, _allocatedBufferSize, VM_FLAGS_ANYWHERE) != 0 ) {
        _diagnostics.error("could not allocate buffer");
        return;
    }

    // assign addresses for each segment of each dylib in new cache
    assignSegmentAddresses();
    std::vector<const LoadedMachO*> overflowDylibs;
    while ( cacheOverflowAmount() != 0 ) {
        if ( !_options.evictLeafDylibsOnOverflow ) {
            _diagnostics.error("cache overflow by %lluMB", cacheOverflowAmount() / 1024 / 1024);
            return;
        }
        size_t evictionCount = evictLeafDylibs(cacheOverflowAmount(), overflowDylibs);
        // re-layout cache
        for (DylibInfo& dylib : _sortedDylibs)
            dylib.cacheLocation.clear();
        assignSegmentAddresses();

        _diagnostics.verbose("cache overflow, evicted %lu leaf dylibs\n", evictionCount);
    }
    markPaddingInaccessible();

     // copy all segments into cache
    uint64_t t2 = mach_absolute_time();
    writeCacheHeader();
    copyRawSegments();

    // rebase all dylibs for new location in cache
    uint64_t t3 = mach_absolute_time();
    _aslrTracker.setDataRegion(_readWriteRegion.buffer, _readWriteRegion.sizeInUse);
    adjustAllImagesForNewSegmentLocations();
    if ( _diagnostics.hasError() )
        return;

    // build ImageArray for dyld3, which has side effect of binding all cached dylibs
    uint64_t t4 = mach_absolute_time();
    buildImageArray(aliases);
    if ( _diagnostics.hasError() )
        return;

    // optimize ObjC
    uint64_t t5 = mach_absolute_time();
    DyldSharedCache* dyldCache = (DyldSharedCache*)_readExecuteRegion.buffer;
    if ( _options.optimizeObjC )
        optimizeObjC();
    if ( _diagnostics.hasError() )
        return;


    // optimize away stubs
    uint64_t t6 = mach_absolute_time();
    std::vector<uint64_t> branchPoolOffsets;
    uint64_t cacheStartAddress = _archLayout->sharedMemoryStart;
    if ( _options.optimizeStubs ) {
        std::vector<uint64_t> branchPoolStartAddrs;
        const uint64_t* p = (uint64_t*)((uint8_t*)dyldCache + dyldCache->header.branchPoolsOffset);
        for (uint32_t i=0; i < dyldCache->header.branchPoolsCount; ++i) {
            uint64_t poolAddr = p[i];
            branchPoolStartAddrs.push_back(poolAddr);
            branchPoolOffsets.push_back(poolAddr - cacheStartAddress);
        }
        optimizeAwayStubs(branchPoolStartAddrs, _branchPoolsLinkEditStartAddr);
    }


    // FIPS seal corecrypto, This must be done after stub elimination (so that __TEXT,__text is not changed after sealing)
    fipsSign();

    // merge and compact LINKEDIT segments
    uint64_t t7 = mach_absolute_time();
    optimizeLinkedit(branchPoolOffsets);

    // copy ImageArray to end of read-only region
    addImageArray();
    if ( _diagnostics.hasError() )
        return;

    // compute and add dlopen closures for all other dylibs
    addOtherImageArray(otherOsDylibsInput, overflowDylibs);
    if ( _diagnostics.hasError() )
        return;

    // compute and add launch closures to end of read-only region
    uint64_t t8 = mach_absolute_time();
    addClosures(osExecutables);
    if ( _diagnostics.hasError() )
        return;

    // update final readOnly region size
    dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)(_readExecuteRegion.buffer + dyldCache->header.mappingOffset);
    mappings[2].size = _readOnlyRegion.sizeInUse;
    if ( _options.excludeLocalSymbols )
        dyldCache->header.localSymbolsOffset = _readOnlyRegion.cacheFileOffset + _readOnlyRegion.sizeInUse;

    // record max slide now that final size is established
    if ( _archLayout->sharedRegionsAreDiscontiguous ) {
        // special case x86_64 which has three non-contiguous chunks each in their own 1GB regions
        uint64_t maxSlide0 = 0x60000000 - _readExecuteRegion.sizeInUse; // TEXT region has 1.5GB region
        uint64_t maxSlide1 = 0x40000000 - _readWriteRegion.sizeInUse;
        uint64_t maxSlide2 = 0x3FE00000 - _readOnlyRegion.sizeInUse;
        dyldCache->header.maxSlide = std::min(std::min(maxSlide0, maxSlide1), maxSlide2);
    }
    else {
        dyldCache->header.maxSlide = (_archLayout->sharedMemoryStart + _archLayout->sharedMemorySize) - (_readOnlyRegion.unslidLoadAddress + _readOnlyRegion.sizeInUse);
    }

    uint64_t t9 = mach_absolute_time();
    
    // fill in slide info at start of region[2]
    // do this last because it modifies pointers in DATA segments
    if ( _options.cacheSupportsASLR ) {
#if SUPPORT_ARCH_arm64e
        if ( strcmp(_archLayout->archName, "arm64e") == 0 )
            writeSlideInfoV3(_aslrTracker.bitmap(), _aslrTracker.dataPageCount());
        else
#endif
        if ( _archLayout->is64 )
            writeSlideInfoV2<Pointer64<LittleEndian>>(_aslrTracker.bitmap(), _aslrTracker.dataPageCount());
        else
#if SUPPORT_ARCH_arm64_32
        if ( strcmp(_archLayout->archName, "arm64_32") == 0 )
            writeSlideInfoV4<Pointer32<LittleEndian>>(_aslrTracker.bitmap(), _aslrTracker.dataPageCount());
        else
#endif
            writeSlideInfoV2<Pointer32<LittleEndian>>(_aslrTracker.bitmap(), _aslrTracker.dataPageCount());
    }

    uint64_t t10 = mach_absolute_time();

    // last sanity check on size
    if ( cacheOverflowAmount() != 0 ) {
        _diagnostics.error("cache overflow after optimizations 0x%llX -> 0x%llX", _readExecuteRegion.unslidLoadAddress, _readOnlyRegion.unslidLoadAddress + _readOnlyRegion.sizeInUse);
        return;
    }

    // codesignature is part of file, but is not mapped
    codeSign();
    if ( _diagnostics.hasError() )
        return;

    uint64_t t11 = mach_absolute_time();

    if ( _options.verbose ) {
        fprintf(stderr, "time to layout cache: %ums\n", absolutetime_to_milliseconds(t2-t1));
        fprintf(stderr, "time to copy cached dylibs into buffer: %ums\n", absolutetime_to_milliseconds(t3-t2));
        fprintf(stderr, "time to adjust segments for new split locations: %ums\n", absolutetime_to_milliseconds(t4-t3));
        fprintf(stderr, "time to bind all images: %ums\n", absolutetime_to_milliseconds(t5-t4));
        fprintf(stderr, "time to optimize Objective-C: %ums\n", absolutetime_to_milliseconds(t6-t5));
        fprintf(stderr, "time to do stub elimination: %ums\n", absolutetime_to_milliseconds(t7-t6));
        fprintf(stderr, "time to optimize LINKEDITs: %ums\n", absolutetime_to_milliseconds(t8-t7));
        fprintf(stderr, "time to build %lu closures: %ums\n", osExecutables.size(), absolutetime_to_milliseconds(t9-t8));
        fprintf(stderr, "time to compute slide info: %ums\n", absolutetime_to_milliseconds(t10-t9));
        fprintf(stderr, "time to compute UUID and codesign cache file: %ums\n", absolutetime_to_milliseconds(t11-t10));
    }

    return;
}


void CacheBuilder::writeCacheHeader()
{
    // "dyld_v1" + spaces + archName(), with enough spaces to pad to 15 bytes
    std::string magic = "dyld_v1";
    magic.append(15 - magic.length() - _options.archName.length(), ' ');
    magic.append(_options.archName);
    assert(magic.length() == 15);

    // fill in header
    dyld_cache_header* dyldCacheHeader = (dyld_cache_header*)_readExecuteRegion.buffer;
    memcpy(dyldCacheHeader->magic, magic.c_str(), 16);
    dyldCacheHeader->mappingOffset        = sizeof(dyld_cache_header);
    dyldCacheHeader->mappingCount         = 3;
    dyldCacheHeader->imagesOffset         = (uint32_t)(dyldCacheHeader->mappingOffset + 3*sizeof(dyld_cache_mapping_info) + sizeof(uint64_t)*_branchPoolStarts.size());
    dyldCacheHeader->imagesCount          = (uint32_t)_sortedDylibs.size() + _aliasCount;
    dyldCacheHeader->dyldBaseAddress      = 0;
    dyldCacheHeader->codeSignatureOffset  = 0;
    dyldCacheHeader->codeSignatureSize    = 0;
    dyldCacheHeader->slideInfoOffset      = _slideInfoFileOffset;
    dyldCacheHeader->slideInfoSize        = _slideInfoBufferSizeAllocated;
    dyldCacheHeader->localSymbolsOffset   = 0;
    dyldCacheHeader->localSymbolsSize     = 0;
    dyldCacheHeader->cacheType            = _options.optimizeStubs ? kDyldSharedCacheTypeProduction : kDyldSharedCacheTypeDevelopment;
    dyldCacheHeader->accelerateInfoAddr   = 0;
    dyldCacheHeader->accelerateInfoSize   = 0;
    bzero(dyldCacheHeader->uuid, 16);// overwritten later by recomputeCacheUUID()
    dyldCacheHeader->branchPoolsOffset    = dyldCacheHeader->mappingOffset + 3*sizeof(dyld_cache_mapping_info);
    dyldCacheHeader->branchPoolsCount     = (uint32_t)_branchPoolStarts.size();
    dyldCacheHeader->imagesTextOffset     = dyldCacheHeader->imagesOffset + sizeof(dyld_cache_image_info)*dyldCacheHeader->imagesCount;
    dyldCacheHeader->imagesTextCount      = _sortedDylibs.size();
    dyldCacheHeader->dylibsImageGroupAddr = 0;
    dyldCacheHeader->dylibsImageGroupSize = 0;
    dyldCacheHeader->otherImageGroupAddr  = 0;
    dyldCacheHeader->otherImageGroupSize  = 0;
    dyldCacheHeader->progClosuresAddr     = 0;
    dyldCacheHeader->progClosuresSize     = 0;
    dyldCacheHeader->progClosuresTrieAddr = 0;
    dyldCacheHeader->progClosuresTrieSize = 0;
    dyldCacheHeader->platform             = (uint8_t)_options.platform;
    dyldCacheHeader->formatVersion        = dyld3::closure::kFormatVersion;
    dyldCacheHeader->dylibsExpectedOnDisk = !_options.dylibsRemovedDuringMastering;
    dyldCacheHeader->simulator            = _options.forSimulator;
    dyldCacheHeader->locallyBuiltCache    = _options.isLocallyBuiltCache;
    dyldCacheHeader->formatVersion        = dyld3::closure::kFormatVersion;
    dyldCacheHeader->sharedRegionStart    = _archLayout->sharedMemoryStart;
    dyldCacheHeader->sharedRegionSize     = _archLayout->sharedMemorySize;

   // fill in mappings
    dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)(_readExecuteRegion.buffer + dyldCacheHeader->mappingOffset);
    mappings[0].address    = _readExecuteRegion.unslidLoadAddress;
    mappings[0].fileOffset = 0;
    mappings[0].size       = _readExecuteRegion.sizeInUse;
    mappings[0].maxProt    = VM_PROT_READ | VM_PROT_EXECUTE;
    mappings[0].initProt   = VM_PROT_READ | VM_PROT_EXECUTE;
    mappings[1].address    = _readWriteRegion.unslidLoadAddress;
    mappings[1].fileOffset = _readExecuteRegion.sizeInUse;
    mappings[1].size       = _readWriteRegion.sizeInUse;
    mappings[1].maxProt    = VM_PROT_READ | VM_PROT_WRITE;
    mappings[1].initProt   = VM_PROT_READ | VM_PROT_WRITE;
    mappings[2].address    = _readOnlyRegion.unslidLoadAddress;
    mappings[2].fileOffset = _readExecuteRegion.sizeInUse + _readWriteRegion.sizeInUse;
    mappings[2].size       = _readOnlyRegion.sizeInUse;
    mappings[2].maxProt    = VM_PROT_READ;
    mappings[2].initProt   = VM_PROT_READ;

    // fill in branch pool addresses
    uint64_t* p = (uint64_t*)(_readExecuteRegion.buffer + dyldCacheHeader->branchPoolsOffset);
    for (uint64_t pool : _branchPoolStarts) {
        *p++ = pool;
    }

    // fill in image table
    dyld_cache_image_info* images = (dyld_cache_image_info*)(_readExecuteRegion.buffer + dyldCacheHeader->imagesOffset);
    for (const DylibInfo& dylib : _sortedDylibs) {
        const char* installName = dylib.input->mappedFile.mh->installName();
        images->address = dylib.cacheLocation[0].dstCacheUnslidAddress;
        if ( _options.dylibsRemovedDuringMastering ) {
            images->modTime = 0;
            images->inode   = pathHash(installName);
        }
        else {
            images->modTime = dylib.input->mappedFile.modTime;
            images->inode   = dylib.input->mappedFile.inode;
        }
        uint32_t installNameOffsetInTEXT =  (uint32_t)(installName - (char*)dylib.input->mappedFile.mh);
        images->pathFileOffset = (uint32_t)dylib.cacheLocation[0].dstCacheFileOffset + installNameOffsetInTEXT;
        ++images;
    }
    // append aliases image records and strings
/*
    for (auto &dylib : _dylibs) {
        if (!dylib->installNameAliases.empty()) {
            for (const std::string& alias : dylib->installNameAliases) {
                images->set_address(_segmentMap[dylib][0].address);
                if (_manifest.platform() == "osx") {
                    images->modTime = dylib->lastModTime;
                    images->inode = dylib->inode;
                }
                else {
                    images->modTime = 0;
                    images->inode = pathHash(alias.c_str());
                }
                images->pathFileOffset = offset;
                //fprintf(stderr, "adding alias %s for %s\n", alias.c_str(), dylib->installName.c_str());
                ::strcpy((char*)&_buffer[offset], alias.c_str());
                offset += alias.size() + 1;
                ++images;
            }
        }
    }
*/
    // calculate start of text image array and trailing string pool
    dyld_cache_image_text_info* textImages = (dyld_cache_image_text_info*)(_readExecuteRegion.buffer + dyldCacheHeader->imagesTextOffset);
    uint32_t stringOffset = (uint32_t)(dyldCacheHeader->imagesTextOffset + sizeof(dyld_cache_image_text_info) * _sortedDylibs.size());

    // write text image array and image names pool at same time
    for (const DylibInfo& dylib : _sortedDylibs) {
        dylib.input->mappedFile.mh->getUuid(textImages->uuid);
        textImages->loadAddress     = dylib.cacheLocation[0].dstCacheUnslidAddress;
        textImages->textSegmentSize = (uint32_t)dylib.cacheLocation[0].dstCacheSegmentSize;
        textImages->pathOffset      = stringOffset;
        const char* installName = dylib.input->mappedFile.mh->installName();
        ::strcpy((char*)_readExecuteRegion.buffer + stringOffset, installName);
        stringOffset += (uint32_t)strlen(installName)+1;
        ++textImages;
    }

    // make sure header did not overflow into first mapped image
    const dyld_cache_image_info* firstImage = (dyld_cache_image_info*)(_readExecuteRegion.buffer + dyldCacheHeader->imagesOffset);
    assert(stringOffset <= (firstImage->address - mappings[0].address));
}

void CacheBuilder::copyRawSegments()
{
    const bool log = false;
    dispatch_apply(_sortedDylibs.size(), DISPATCH_APPLY_AUTO, ^(size_t index) {
        const DylibInfo& dylib = _sortedDylibs[index];
        for (const SegmentMappingInfo& info : dylib.cacheLocation) {
            if (log) fprintf(stderr, "copy %s segment %s (0x%08X bytes) from %p to %p (logical addr 0x%llX) for %s\n",
                             _options.archName.c_str(), info.segName, info.copySegmentSize, info.srcSegment, info.dstSegment, info.dstCacheUnslidAddress, dylib.input->mappedFile.runtimePath.c_str());
            ::memcpy(info.dstSegment, info.srcSegment, info.copySegmentSize);
            if (uint64_t paddingSize = info.dstCacheSegmentSize - info.copySegmentSize) {
                ::memset((char*)info.dstSegment + info.copySegmentSize, 0, paddingSize);
            }
        }
    });
}

void CacheBuilder::adjustAllImagesForNewSegmentLocations()
{
    __block std::vector<Diagnostics> diags;
    diags.resize(_sortedDylibs.size());

    if (_options.platform == dyld3::Platform::macOS) {
        dispatch_apply(_sortedDylibs.size(), DISPATCH_APPLY_AUTO, ^(size_t index) {
            const DylibInfo& dylib = _sortedDylibs[index];
            adjustDylibSegments(dylib, diags[index]);
        });
    } else {
        // Note this has to be done in serial because the LOH Tracker isn't thread safe
        for (size_t index = 0; index != _sortedDylibs.size(); ++index) {
            const DylibInfo& dylib = _sortedDylibs[index];
            adjustDylibSegments(dylib, diags[index]);
        }
    }

    for (const Diagnostics& diag : diags) {
        if ( diag.hasError() ) {
            _diagnostics.error("%s", diag.errorMessage().c_str());
            break;
        }
    }
}

void CacheBuilder::assignSegmentAddresses()
{
    // calculate size of header info and where first dylib's mach_header should start
    size_t startOffset = sizeof(dyld_cache_header) + 3*sizeof(dyld_cache_mapping_info);
    size_t maxPoolCount = 0;
    if ( _archLayout->branchReach != 0 )
        maxPoolCount = (_archLayout->sharedMemorySize / _archLayout->branchReach);
    startOffset += maxPoolCount * sizeof(uint64_t);
    startOffset += sizeof(dyld_cache_image_info) * _sortedDylibs.size();
    startOffset += sizeof(dyld_cache_image_text_info) * _sortedDylibs.size();
    for (const DylibInfo& dylib : _sortedDylibs) {
        startOffset += (strlen(dylib.input->mappedFile.mh->installName()) + 1);
    }
    //fprintf(stderr, "%s total header size = 0x%08lX\n", _options.archName.c_str(), startOffset);
    startOffset = align(startOffset, 12);

    _branchPoolStarts.clear();

    // assign TEXT segment addresses
    _readExecuteRegion.buffer               = (uint8_t*)_fullAllocatedBuffer;
    _readExecuteRegion.bufferSize           = 0;
    _readExecuteRegion.sizeInUse            = 0;
    _readExecuteRegion.unslidLoadAddress    = _archLayout->sharedMemoryStart;
    _readExecuteRegion.cacheFileOffset      = 0;
    __block uint64_t addr = _readExecuteRegion.unslidLoadAddress + startOffset; // header
    __block uint64_t lastPoolAddress = addr;
    for (DylibInfo& dylib : _sortedDylibs) {
        __block uint64_t textSegVmAddr = 0;
        dylib.input->mappedFile.mh->forEachSegment(^(const dyld3::MachOFile::SegmentInfo& segInfo, bool& stop) {
            if ( strcmp(segInfo.segName, "__TEXT") == 0 )
                textSegVmAddr = segInfo.vmAddr;
            if ( segInfo.protections != (VM_PROT_READ | VM_PROT_EXECUTE) )
                return;
            // Insert branch island pools every 128MB for arm64
            if ( (_archLayout->branchPoolTextSize != 0) && ((addr + segInfo.vmSize - lastPoolAddress) > _archLayout->branchReach) ) {
                _branchPoolStarts.push_back(addr);
                _diagnostics.verbose("adding branch pool at 0x%llX\n", addr);
                lastPoolAddress = addr;
                addr += _archLayout->branchPoolTextSize;
            }
            // Keep __TEXT segments 4K or more aligned
            addr = align(addr, std::max((int)segInfo.p2align, (int)12));
            uint64_t offsetInRegion = addr - _readExecuteRegion.unslidLoadAddress;
            SegmentMappingInfo loc;
            loc.srcSegment             = (uint8_t*)dylib.input->mappedFile.mh + segInfo.vmAddr - textSegVmAddr;
            loc.segName                = segInfo.segName;
            loc.dstSegment             = _readExecuteRegion.buffer + offsetInRegion;
            loc.dstCacheUnslidAddress  = addr;
            loc.dstCacheFileOffset     = (uint32_t)offsetInRegion;
            loc.dstCacheSegmentSize    = (uint32_t)align(segInfo.sizeOfSections, 12);
            loc.copySegmentSize        = (uint32_t)align(segInfo.sizeOfSections, 12);
            loc.srcSegmentIndex        = segInfo.segIndex;
            dylib.cacheLocation.push_back(loc);
            addr += loc.dstCacheSegmentSize;
        });
    }
    // align TEXT region end
    uint64_t endTextAddress = align(addr, _archLayout->sharedRegionAlignP2);
    _readExecuteRegion.bufferSize = endTextAddress - _readExecuteRegion.unslidLoadAddress;
    _readExecuteRegion.sizeInUse  = _readExecuteRegion.bufferSize;

    // assign __DATA* addresses
    if ( _archLayout->sharedRegionsAreDiscontiguous )
        addr = _archLayout->sharedMemoryStart + 0x60000000;
    else
        addr = align((addr + _archLayout->sharedRegionPadding), _archLayout->sharedRegionAlignP2);
    _readWriteRegion.buffer               = (uint8_t*)_fullAllocatedBuffer + addr - _archLayout->sharedMemoryStart;
    _readWriteRegion.bufferSize           = 0;
    _readWriteRegion.sizeInUse            = 0;
    _readWriteRegion.unslidLoadAddress    = addr;
    _readWriteRegion.cacheFileOffset      = _readExecuteRegion.sizeInUse;

    // layout all __DATA_CONST segments
    __block int dataConstSegmentCount = 0;
    for (DylibInfo& dylib : _sortedDylibs) {
        __block uint64_t textSegVmAddr = 0;
       dylib.input->mappedFile.mh->forEachSegment(^(const dyld3::MachOFile::SegmentInfo& segInfo, bool& stop) {
            if ( strcmp(segInfo.segName, "__TEXT") == 0 )
                textSegVmAddr = segInfo.vmAddr;
            if ( segInfo.protections != (VM_PROT_READ | VM_PROT_WRITE) )
                return;
            if ( strcmp(segInfo.segName, "__DATA_CONST") != 0 )
                return;
            ++dataConstSegmentCount;
            // Pack __DATA_CONST segments
            addr = align(addr, segInfo.p2align);
            size_t copySize = std::min((size_t)segInfo.fileSize, (size_t)segInfo.sizeOfSections);
            uint64_t offsetInRegion = addr - _readWriteRegion.unslidLoadAddress;
            SegmentMappingInfo loc;
            loc.srcSegment             = (uint8_t*)dylib.input->mappedFile.mh + segInfo.vmAddr - textSegVmAddr;
            loc.segName                = segInfo.segName;
            loc.dstSegment             = _readWriteRegion.buffer + offsetInRegion;
            loc.dstCacheUnslidAddress  = addr;
            loc.dstCacheFileOffset     = (uint32_t)(_readWriteRegion.cacheFileOffset + offsetInRegion);
            loc.dstCacheSegmentSize    = (uint32_t)segInfo.sizeOfSections;
            loc.copySegmentSize        = (uint32_t)copySize;
            loc.srcSegmentIndex        = segInfo.segIndex;
            dylib.cacheLocation.push_back(loc);
            addr += loc.dstCacheSegmentSize;
        });
    }

    // layout all __DATA segments (and other r/w non-dirty, non-const) segments
    for (DylibInfo& dylib : _sortedDylibs) {
        __block uint64_t textSegVmAddr = 0;
       dylib.input->mappedFile.mh->forEachSegment(^(const dyld3::MachOFile::SegmentInfo& segInfo, bool& stop) {
            if ( strcmp(segInfo.segName, "__TEXT") == 0 )
                textSegVmAddr = segInfo.vmAddr;
            if ( segInfo.protections != (VM_PROT_READ | VM_PROT_WRITE) )
                return;
            if ( strcmp(segInfo.segName, "__DATA_CONST") == 0 )
                return;
            if ( strcmp(segInfo.segName, "__DATA_DIRTY") == 0 )
                return;
            if ( dataConstSegmentCount > 10 ) {
                // Pack __DATA segments only if we also have __DATA_CONST segments
                addr = align(addr, segInfo.p2align);
            }
            else {
                // Keep __DATA segments 4K or more aligned
                addr = align(addr, std::max((int)segInfo.p2align, (int)12));
            }
            size_t copySize = std::min((size_t)segInfo.fileSize, (size_t)segInfo.sizeOfSections);
            uint64_t offsetInRegion = addr - _readWriteRegion.unslidLoadAddress;
            SegmentMappingInfo loc;
            loc.srcSegment             = (uint8_t*)dylib.input->mappedFile.mh + segInfo.vmAddr - textSegVmAddr;
            loc.segName                = segInfo.segName;
            loc.dstSegment             = _readWriteRegion.buffer + offsetInRegion;
            loc.dstCacheUnslidAddress  = addr;
            loc.dstCacheFileOffset     = (uint32_t)(_readWriteRegion.cacheFileOffset + offsetInRegion);
            loc.dstCacheSegmentSize    = (uint32_t)segInfo.sizeOfSections;
            loc.copySegmentSize        = (uint32_t)copySize;
            loc.srcSegmentIndex        = segInfo.segIndex;
            dylib.cacheLocation.push_back(loc);
            addr += loc.dstCacheSegmentSize;
        });
    }

    // layout all __DATA_DIRTY segments, sorted (FIXME)
	const size_t dylibCount = _sortedDylibs.size();
    uint32_t dirtyDataSortIndexes[dylibCount];
    for (size_t i=0; i < dylibCount; ++i)
        dirtyDataSortIndexes[i] = (uint32_t)i;
    std::sort(&dirtyDataSortIndexes[0], &dirtyDataSortIndexes[dylibCount], [&](const uint32_t& a, const uint32_t& b) {
        const auto& orderA = _options.dirtyDataSegmentOrdering.find(_sortedDylibs[a].input->mappedFile.runtimePath);
        const auto& orderB = _options.dirtyDataSegmentOrdering.find(_sortedDylibs[b].input->mappedFile.runtimePath);
        bool foundA = (orderA != _options.dirtyDataSegmentOrdering.end());
        bool foundB = (orderB != _options.dirtyDataSegmentOrdering.end());

        // Order all __DATA_DIRTY segments specified in the order file first, in the order specified in the file,
        // followed by any other __DATA_DIRTY segments in lexicographic order.
        if ( foundA && foundB )
            return orderA->second < orderB->second;
        else if ( foundA )
            return true;
        else if ( foundB )
             return false;
        else
             return _sortedDylibs[a].input->mappedFile.runtimePath < _sortedDylibs[b].input->mappedFile.runtimePath;
    });
    addr = align(addr, 12);
    for (size_t i=0; i < dylibCount; ++i) {
        DylibInfo& dylib  = _sortedDylibs[dirtyDataSortIndexes[i]];
        __block uint64_t textSegVmAddr = 0;
        dylib.input->mappedFile.mh->forEachSegment(^(const dyld3::MachOFile::SegmentInfo& segInfo, bool& stop) {
            if ( strcmp(segInfo.segName, "__TEXT") == 0 )
                textSegVmAddr = segInfo.vmAddr;
            if ( segInfo.protections != (VM_PROT_READ | VM_PROT_WRITE) )
                return;
            if ( strcmp(segInfo.segName, "__DATA_DIRTY") != 0 )
                return;
            // Pack __DATA_DIRTY segments
            addr = align(addr, segInfo.p2align);
            size_t copySize = std::min((size_t)segInfo.fileSize, (size_t)segInfo.sizeOfSections);
            uint64_t offsetInRegion = addr - _readWriteRegion.unslidLoadAddress;
            SegmentMappingInfo loc;
            loc.srcSegment             = (uint8_t*)dylib.input->mappedFile.mh + segInfo.vmAddr - textSegVmAddr;
            loc.segName                = segInfo.segName;
            loc.dstSegment             = _readWriteRegion.buffer + offsetInRegion;
            loc.dstCacheUnslidAddress  = addr;
            loc.dstCacheFileOffset     = (uint32_t)(_readWriteRegion.cacheFileOffset + offsetInRegion);
            loc.dstCacheSegmentSize    = (uint32_t)segInfo.sizeOfSections;
            loc.copySegmentSize        = (uint32_t)copySize;
            loc.srcSegmentIndex        = segInfo.segIndex;
            dylib.cacheLocation.push_back(loc);
            addr += loc.dstCacheSegmentSize;
        });
    }

    // align DATA region end
    uint64_t endDataAddress = align(addr, _archLayout->sharedRegionAlignP2);
    _readWriteRegion.bufferSize   = endDataAddress - _readWriteRegion.unslidLoadAddress;
    _readWriteRegion.sizeInUse    = _readWriteRegion.bufferSize;

    // start read-only region
    if ( _archLayout->sharedRegionsAreDiscontiguous )
        addr = _archLayout->sharedMemoryStart + 0xA0000000;
    else
        addr = align((addr + _archLayout->sharedRegionPadding), _archLayout->sharedRegionAlignP2);
    _readOnlyRegion.buffer               = (uint8_t*)_fullAllocatedBuffer + addr - _archLayout->sharedMemoryStart;
    _readOnlyRegion.bufferSize           = 0;
    _readOnlyRegion.sizeInUse            = 0;
    _readOnlyRegion.unslidLoadAddress    = addr;
    _readOnlyRegion.cacheFileOffset      = _readWriteRegion.cacheFileOffset + _readWriteRegion.sizeInUse;

    // reserve space for kernel ASLR slide info at start of r/o region
    if ( _options.cacheSupportsASLR ) {
        size_t slideInfoSize = sizeof(dyld_cache_slide_info);
        slideInfoSize = std::max(slideInfoSize, sizeof(dyld_cache_slide_info2));
        slideInfoSize = std::max(slideInfoSize, sizeof(dyld_cache_slide_info3));
        slideInfoSize = std::max(slideInfoSize, sizeof(dyld_cache_slide_info4));
        _slideInfoBufferSizeAllocated = align(slideInfoSize + (_readWriteRegion.sizeInUse/4096) * _archLayout->slideInfoBytesPerPage, _archLayout->sharedRegionAlignP2);
        _slideInfoFileOffset = _readOnlyRegion.cacheFileOffset;
        addr += _slideInfoBufferSizeAllocated;
    }

    // layout all read-only (but not LINKEDIT) segments
    for (DylibInfo& dylib : _sortedDylibs) {
        __block uint64_t textSegVmAddr = 0;
        dylib.input->mappedFile.mh->forEachSegment(^(const dyld3::MachOFile::SegmentInfo& segInfo, bool& stop) {
            if ( strcmp(segInfo.segName, "__TEXT") == 0 )
                textSegVmAddr = segInfo.vmAddr;
            if ( segInfo.protections != VM_PROT_READ )
                return;
            if ( strcmp(segInfo.segName, "__LINKEDIT") == 0 )
                return;
            // Keep segments segments 4K or more aligned
            addr = align(addr, std::max((int)segInfo.p2align, (int)12));
            uint64_t offsetInRegion = addr - _readOnlyRegion.unslidLoadAddress;
            SegmentMappingInfo loc;
            loc.srcSegment             = (uint8_t*)dylib.input->mappedFile.mh + segInfo.vmAddr - textSegVmAddr;
            loc.segName                = segInfo.segName;
            loc.dstSegment             = _readOnlyRegion.buffer + offsetInRegion;
            loc.dstCacheUnslidAddress  = addr;
            loc.dstCacheFileOffset     = (uint32_t)(_readOnlyRegion.cacheFileOffset + offsetInRegion);
            loc.dstCacheSegmentSize    = (uint32_t)align(segInfo.sizeOfSections, 12);
            loc.copySegmentSize        = (uint32_t)segInfo.sizeOfSections;
            loc.srcSegmentIndex        = segInfo.segIndex;
            dylib.cacheLocation.push_back(loc);
            addr += loc.dstCacheSegmentSize;
        });
    }
    // layout all LINKEDIT segments (after other read-only segments), aligned to 16KB
    addr = align(addr, 14);
    _nonLinkEditReadOnlySize =  addr - _readOnlyRegion.unslidLoadAddress;
    for (DylibInfo& dylib : _sortedDylibs) {
        __block uint64_t textSegVmAddr = 0;
        dylib.input->mappedFile.mh->forEachSegment(^(const dyld3::MachOFile::SegmentInfo& segInfo, bool& stop) {
            if ( strcmp(segInfo.segName, "__TEXT") == 0 )
                textSegVmAddr = segInfo.vmAddr;
            if ( segInfo.protections != VM_PROT_READ )
                return;
            if ( strcmp(segInfo.segName, "__LINKEDIT") != 0 )
                return;
            // Keep segments segments 4K or more aligned
            addr = align(addr, std::max((int)segInfo.p2align, (int)12));
            size_t copySize = std::min((size_t)segInfo.fileSize, (size_t)segInfo.sizeOfSections);
            uint64_t offsetInRegion = addr - _readOnlyRegion.unslidLoadAddress;
            SegmentMappingInfo loc;
            loc.srcSegment             = (uint8_t*)dylib.input->mappedFile.mh + segInfo.vmAddr - textSegVmAddr;
            loc.segName                = segInfo.segName;
            loc.dstSegment             = _readOnlyRegion.buffer + offsetInRegion;
            loc.dstCacheUnslidAddress  = addr;
            loc.dstCacheFileOffset     = (uint32_t)(_readOnlyRegion.cacheFileOffset + offsetInRegion);
            loc.dstCacheSegmentSize    = (uint32_t)align(segInfo.sizeOfSections, 12);
            loc.copySegmentSize        = (uint32_t)copySize;
            loc.srcSegmentIndex        = segInfo.segIndex;
            dylib.cacheLocation.push_back(loc);
            addr += loc.dstCacheSegmentSize;
        });
    }
    // add room for branch pool linkedits
    _branchPoolsLinkEditStartAddr = addr;
    addr += (_branchPoolStarts.size() * _archLayout->branchPoolLinkEditSize);

    // align r/o region end
    uint64_t endReadOnlyAddress = align(addr, _archLayout->sharedRegionAlignP2);
    _readOnlyRegion.bufferSize  = endReadOnlyAddress - _readOnlyRegion.unslidLoadAddress;
    _readOnlyRegion.sizeInUse   = _readOnlyRegion.bufferSize;

    //fprintf(stderr, "RX region=%p -> %p, logical addr=0x%llX\n", _readExecuteRegion.buffer, _readExecuteRegion.buffer+_readExecuteRegion.bufferSize, _readExecuteRegion.unslidLoadAddress);
    //fprintf(stderr, "RW region=%p -> %p, logical addr=0x%llX\n", _readWriteRegion.buffer,   _readWriteRegion.buffer+_readWriteRegion.bufferSize, _readWriteRegion.unslidLoadAddress);
    //fprintf(stderr, "RO region=%p -> %p, logical addr=0x%llX\n", _readOnlyRegion.buffer,    _readOnlyRegion.buffer+_readOnlyRegion.bufferSize, _readOnlyRegion.unslidLoadAddress);

    // sort SegmentMappingInfo for each image to be in the same order as original segments
    for (DylibInfo& dylib : _sortedDylibs) {
        std::sort(dylib.cacheLocation.begin(), dylib.cacheLocation.end(), [&](const SegmentMappingInfo& a, const SegmentMappingInfo& b) {
            return a.srcSegmentIndex < b.srcSegmentIndex;
        });
    }
}

void CacheBuilder::markPaddingInaccessible()
{
    // region between RX and RW
    uint8_t* startPad1 = _readExecuteRegion.buffer+_readExecuteRegion.sizeInUse;
    uint8_t* endPad1   = _readWriteRegion.buffer;
    ::vm_protect(mach_task_self(), (vm_address_t)startPad1, endPad1-startPad1, false, 0);

    // region between RW and RO
    uint8_t* startPad2 = _readWriteRegion.buffer+_readWriteRegion.sizeInUse;
    uint8_t* endPad2   = _readOnlyRegion.buffer;
    ::vm_protect(mach_task_self(), (vm_address_t)startPad2, endPad2-startPad2, false, 0);
}


uint64_t CacheBuilder::pathHash(const char* path)
{
    uint64_t sum = 0;
    for (const char* s=path; *s != '\0'; ++s)
        sum += sum*4 + *s;
    return sum;
}


void CacheBuilder::findDylibAndSegment(const void* contentPtr, std::string& foundDylibName, std::string& foundSegName)
{
    foundDylibName = "???";
    foundSegName   = "???";
    uint64_t unslidVmAddr = ((uint8_t*)contentPtr - _readExecuteRegion.buffer) + _readExecuteRegion.unslidLoadAddress;
    const DyldSharedCache* cache = (DyldSharedCache*)_readExecuteRegion.buffer;
    cache->forEachImage(^(const mach_header* mh, const char* installName) {
        ((dyld3::MachOLoaded*)mh)->forEachSegment(^(const dyld3::MachOFile::SegmentInfo& info, bool &stop) {
            if ( (unslidVmAddr >= info.vmAddr) && (unslidVmAddr < (info.vmAddr+info.vmSize)) ) {
                foundDylibName = installName;
                foundSegName   = info.segName;
                stop           = true;
            }
        });
    });
}


template <typename P>
bool CacheBuilder::makeRebaseChainV2(uint8_t* pageContent, uint16_t lastLocationOffset, uint16_t offset, const dyld_cache_slide_info2* info)
{
    typedef typename P::uint_t     pint_t;

    const pint_t   deltaMask    = (pint_t)(info->delta_mask);
    const pint_t   valueMask    = ~deltaMask;
    const pint_t   valueAdd     = (pint_t)(info->value_add);
    const unsigned deltaShift   = __builtin_ctzll(deltaMask) - 2;
    const uint32_t maxDelta     = (uint32_t)(deltaMask >> deltaShift);

    pint_t* lastLoc = (pint_t*)&pageContent[lastLocationOffset+0];
    pint_t lastValue = (pint_t)P::getP(*lastLoc);
    if ( (lastValue - valueAdd) & deltaMask ) {
        std::string dylibName;
        std::string segName;
        findDylibAndSegment((void*)pageContent, dylibName, segName);
        _diagnostics.error("rebase pointer does not point within cache. lastOffset=0x%04X, seg=%s, dylib=%s\n",
                            lastLocationOffset, segName.c_str(), dylibName.c_str());
        return false;
    }
    if ( offset <= (lastLocationOffset+maxDelta) ) {
        // previous location in range, make link from it
        // encode this location into last value
        pint_t delta = offset - lastLocationOffset;
        pint_t newLastValue = ((lastValue - valueAdd) & valueMask) | (delta << deltaShift);
        //warning("  add chain: delta = %d, lastOffset=0x%03X, offset=0x%03X, org value=0x%08lX, new value=0x%08lX",
        //                    offset - lastLocationOffset, lastLocationOffset, offset, (long)lastValue, (long)newLastValue);
        P::setP(*lastLoc, newLastValue);
        return true;
    }
    //fprintf(stderr, "  too big delta = %d, lastOffset=0x%03X, offset=0x%03X\n", offset - lastLocationOffset, lastLocationOffset, offset);

    // distance between rebase locations is too far
    // see if we can make a chain from non-rebase locations
    uint16_t nonRebaseLocationOffsets[1024];
    unsigned nrIndex = 0;
    for (uint16_t i = lastLocationOffset; i < offset-maxDelta; ) {
        nonRebaseLocationOffsets[nrIndex] = 0;
        for (int j=maxDelta; j > 0; j -= 4) {
            pint_t value = (pint_t)P::getP(*(pint_t*)&pageContent[i+j]);
            if ( value == 0 ) {
                // Steal values of 0 to be used in the rebase chain
                nonRebaseLocationOffsets[nrIndex] = i+j;
                break;
            }
        }
        if ( nonRebaseLocationOffsets[nrIndex] == 0 ) {
            lastValue = (pint_t)P::getP(*lastLoc);
            pint_t newValue = ((lastValue - valueAdd) & valueMask);
            //warning("   no way to make non-rebase delta chain, terminate off=0x%03X, old value=0x%08lX, new value=0x%08lX", lastLocationOffset, (long)value, (long)newValue);
            P::setP(*lastLoc, newValue);
            return false;
        }
        i = nonRebaseLocationOffsets[nrIndex];
        ++nrIndex;
    }

    // we can make chain. go back and add each non-rebase location to chain
    uint16_t prevOffset = lastLocationOffset;
    pint_t* prevLoc = (pint_t*)&pageContent[prevOffset];
    for (unsigned n=0; n < nrIndex; ++n) {
        uint16_t nOffset = nonRebaseLocationOffsets[n];
        assert(nOffset != 0);
        pint_t* nLoc = (pint_t*)&pageContent[nOffset];
        uint32_t delta2 = nOffset - prevOffset;
        pint_t value = (pint_t)P::getP(*prevLoc);
        pint_t newValue;
        if ( value == 0 )
            newValue = (delta2 << deltaShift);
        else
            newValue = ((value - valueAdd) & valueMask) | (delta2 << deltaShift);
        //warning("    non-rebase delta = %d, to off=0x%03X, old value=0x%08lX, new value=0x%08lX", delta2, nOffset, (long)value, (long)newValue);
        P::setP(*prevLoc, newValue);
        prevOffset = nOffset;
        prevLoc = nLoc;
    }
    uint32_t delta3 = offset - prevOffset;
    pint_t value = (pint_t)P::getP(*prevLoc);
    pint_t newValue;
    if ( value == 0 )
        newValue = (delta3 << deltaShift);
    else
        newValue = ((value - valueAdd) & valueMask) | (delta3 << deltaShift);
    //warning("    non-rebase delta = %d, to off=0x%03X, old value=0x%08lX, new value=0x%08lX", delta3, offset, (long)value, (long)newValue);
    P::setP(*prevLoc, newValue);

    return true;
}


template <typename P>
void CacheBuilder::addPageStartsV2(uint8_t* pageContent, const bool bitmap[], const dyld_cache_slide_info2* info,
                                std::vector<uint16_t>& pageStarts, std::vector<uint16_t>& pageExtras)
{
    typedef typename P::uint_t     pint_t;

    const pint_t   deltaMask    = (pint_t)(info->delta_mask);
    const pint_t   valueMask    = ~deltaMask;
    const uint32_t pageSize     = info->page_size;
    const pint_t   valueAdd     = (pint_t)(info->value_add);

    uint16_t startValue = DYLD_CACHE_SLIDE_PAGE_ATTR_NO_REBASE;
    uint16_t lastLocationOffset = 0xFFFF;
    for(uint32_t i=0; i < pageSize/4; ++i) {
        unsigned offset = i*4;
        if ( bitmap[i] ) {
            if ( startValue == DYLD_CACHE_SLIDE_PAGE_ATTR_NO_REBASE ) {
                // found first rebase location in page
                startValue = i;
            }
            else if ( !makeRebaseChainV2<P>(pageContent, lastLocationOffset, offset, info) ) {
                // can't record all rebasings in one chain
                if ( (startValue & DYLD_CACHE_SLIDE_PAGE_ATTR_EXTRA) == 0 ) {
                    // switch page_start to "extras" which is a list of chain starts
                    unsigned indexInExtras = (unsigned)pageExtras.size();
                    if ( indexInExtras > 0x3FFF ) {
                        _diagnostics.error("rebase overflow in v2 page extras");
                        return;
                    }
                    pageExtras.push_back(startValue);
                    startValue = indexInExtras | DYLD_CACHE_SLIDE_PAGE_ATTR_EXTRA;
                }
                pageExtras.push_back(i);
            }
            lastLocationOffset = offset;
        }
    }
    if ( lastLocationOffset != 0xFFFF ) {
        // mark end of chain
        pint_t* lastLoc = (pint_t*)&pageContent[lastLocationOffset];
        pint_t lastValue = (pint_t)P::getP(*lastLoc);
        pint_t newValue = ((lastValue - valueAdd) & valueMask);
        P::setP(*lastLoc, newValue);
    }
    if ( startValue & DYLD_CACHE_SLIDE_PAGE_ATTR_EXTRA ) {
        // add end bit to extras
        pageExtras.back() |= DYLD_CACHE_SLIDE_PAGE_ATTR_END;
    }
    pageStarts.push_back(startValue);
}

template <typename P>
void CacheBuilder::writeSlideInfoV2(const bool bitmap[], unsigned dataPageCount)
{
    typedef typename P::uint_t    pint_t;
    typedef typename P::E         E;
    const uint32_t pageSize = 4096;

    // fill in fixed info
    assert(_slideInfoFileOffset != 0);
    dyld_cache_slide_info2* info = (dyld_cache_slide_info2*)_readOnlyRegion.buffer;
    info->version    = 2;
    info->page_size  = pageSize;
    info->delta_mask = _archLayout->pointerDeltaMask;
    info->value_add  = (sizeof(pint_t) == 8) ? 0 : _archLayout->sharedMemoryStart;  // only value_add for 32-bit archs

    // set page starts and extras for each page
    std::vector<uint16_t> pageStarts;
    std::vector<uint16_t> pageExtras;
    pageStarts.reserve(dataPageCount);
    uint8_t* pageContent = _readWriteRegion.buffer;
    const bool* bitmapForPage = bitmap;
    for (unsigned i=0; i < dataPageCount; ++i) {
        //warning("page[%d]", i);
        addPageStartsV2<P>(pageContent, bitmapForPage, info, pageStarts, pageExtras);
        if ( _diagnostics.hasError() ) {
            return;
        }
        pageContent += pageSize;
        bitmapForPage += (sizeof(bool)*(pageSize/4));
    }

    // fill in computed info
    info->page_starts_offset = sizeof(dyld_cache_slide_info2);
    info->page_starts_count  = (unsigned)pageStarts.size();
    info->page_extras_offset = (unsigned)(sizeof(dyld_cache_slide_info2)+pageStarts.size()*sizeof(uint16_t));
    info->page_extras_count  = (unsigned)pageExtras.size();
    uint16_t* pageStartsBuffer = (uint16_t*)((char*)info + info->page_starts_offset);
    uint16_t* pageExtrasBuffer = (uint16_t*)((char*)info + info->page_extras_offset);
    for (unsigned i=0; i < pageStarts.size(); ++i)
        pageStartsBuffer[i] = pageStarts[i];
    for (unsigned i=0; i < pageExtras.size(); ++i)
        pageExtrasBuffer[i] = pageExtras[i];
    // update header with final size
    uint64_t slideInfoSize = align(info->page_extras_offset + pageExtras.size()*sizeof(uint16_t), _archLayout->sharedRegionAlignP2);
    if ( slideInfoSize > _slideInfoBufferSizeAllocated ) {
        _diagnostics.error("kernel slide info overflow buffer");
    }
    ((dyld_cache_header*)_readExecuteRegion.buffer)->slideInfoSize = slideInfoSize;
    //fprintf(stderr, "pageCount=%u, page_starts_count=%lu, page_extras_count=%lu\n", dataPageCount, pageStarts.size(), pageExtras.size());
}

// fits in to int16_t
static bool smallValue(uint64_t value)
{
    uint32_t high = (value & 0xFFFF8000);
    return (high == 0) || (high == 0xFFFF8000);
}

template <typename P>
bool CacheBuilder::makeRebaseChainV4(uint8_t* pageContent, uint16_t lastLocationOffset, uint16_t offset, const dyld_cache_slide_info4* info)
{
    typedef typename P::uint_t     pint_t;

    const pint_t   deltaMask    = (pint_t)(info->delta_mask);
    const pint_t   valueMask    = ~deltaMask;
    const pint_t   valueAdd     = (pint_t)(info->value_add);
    const unsigned deltaShift   = __builtin_ctzll(deltaMask) - 2;
    const uint32_t maxDelta     = (uint32_t)(deltaMask >> deltaShift);

    pint_t* lastLoc = (pint_t*)&pageContent[lastLocationOffset+0];
    pint_t lastValue = (pint_t)P::getP(*lastLoc);
    if ( (lastValue - valueAdd) & deltaMask ) {
        std::string dylibName;
        std::string segName;
        findDylibAndSegment((void*)pageContent, dylibName, segName);
        _diagnostics.error("rebase pointer does not point within cache. lastOffset=0x%04X, seg=%s, dylib=%s\n",
                            lastLocationOffset, segName.c_str(), dylibName.c_str());
        return false;
    }
    if ( offset <= (lastLocationOffset+maxDelta) ) {
        // previous location in range, make link from it
        // encode this location into last value
        pint_t delta = offset - lastLocationOffset;
        pint_t newLastValue = ((lastValue - valueAdd) & valueMask) | (delta << deltaShift);
        //warning("  add chain: delta = %d, lastOffset=0x%03X, offset=0x%03X, org value=0x%08lX, new value=0x%08lX",
        //                    offset - lastLocationOffset, lastLocationOffset, offset, (long)lastValue, (long)newLastValue);
        P::setP(*lastLoc, newLastValue);
        return true;
    }
    //fprintf(stderr, "  too big delta = %d, lastOffset=0x%03X, offset=0x%03X\n", offset - lastLocationOffset, lastLocationOffset, offset);

    // distance between rebase locations is too far
    // see if we can make a chain from non-rebase locations
    uint16_t nonRebaseLocationOffsets[1024];
    unsigned nrIndex = 0;
    for (uint16_t i = lastLocationOffset; i < offset-maxDelta; ) {
        nonRebaseLocationOffsets[nrIndex] = 0;
        for (int j=maxDelta; j > 0; j -= 4) {
            pint_t value = (pint_t)P::getP(*(pint_t*)&pageContent[i+j]);
            if ( smallValue(value) ) {
                // Steal values of 0 to be used in the rebase chain
                nonRebaseLocationOffsets[nrIndex] = i+j;
                break;
            }
        }
        if ( nonRebaseLocationOffsets[nrIndex] == 0 ) {
            lastValue = (pint_t)P::getP(*lastLoc);
            pint_t newValue = ((lastValue - valueAdd) & valueMask);
            //fprintf(stderr, "   no way to make non-rebase delta chain, terminate off=0x%03X, old value=0x%08lX, new value=0x%08lX\n",
            //                lastLocationOffset, (long)lastValue, (long)newValue);
            P::setP(*lastLoc, newValue);
            return false;
        }
        i = nonRebaseLocationOffsets[nrIndex];
        ++nrIndex;
    }

    // we can make chain. go back and add each non-rebase location to chain
    uint16_t prevOffset = lastLocationOffset;
    pint_t* prevLoc = (pint_t*)&pageContent[prevOffset];
    for (unsigned n=0; n < nrIndex; ++n) {
        uint16_t nOffset = nonRebaseLocationOffsets[n];
        assert(nOffset != 0);
        pint_t* nLoc = (pint_t*)&pageContent[nOffset];
        uint32_t delta2 = nOffset - prevOffset;
        pint_t value = (pint_t)P::getP(*prevLoc);
        pint_t newValue;
        if ( smallValue(value) )
            newValue = (value & valueMask) | (delta2 << deltaShift);
        else
            newValue = ((value - valueAdd) & valueMask) | (delta2 << deltaShift);
        //warning("    non-rebase delta = %d, to off=0x%03X, old value=0x%08lX, new value=0x%08lX", delta2, nOffset, (long)value, (long)newValue);
        P::setP(*prevLoc, newValue);
        prevOffset = nOffset;
        prevLoc = nLoc;
    }
    uint32_t delta3 = offset - prevOffset;
    pint_t value = (pint_t)P::getP(*prevLoc);
    pint_t newValue;
    if ( smallValue(value) )
        newValue = (value & valueMask) | (delta3 << deltaShift);
    else
        newValue = ((value - valueAdd) & valueMask) | (delta3 << deltaShift);
    //warning("    non-rebase delta = %d, to off=0x%03X, old value=0x%08lX, new value=0x%08lX", delta3, offset, (long)value, (long)newValue);
    P::setP(*prevLoc, newValue);

    return true;
}


template <typename P>
void CacheBuilder::addPageStartsV4(uint8_t* pageContent, const bool bitmap[], const dyld_cache_slide_info4* info,
                                std::vector<uint16_t>& pageStarts, std::vector<uint16_t>& pageExtras)
{
    typedef typename P::uint_t     pint_t;

    const pint_t   deltaMask    = (pint_t)(info->delta_mask);
    const pint_t   valueMask    = ~deltaMask;
    const uint32_t pageSize     = info->page_size;
    const pint_t   valueAdd     = (pint_t)(info->value_add);

    uint16_t startValue = DYLD_CACHE_SLIDE4_PAGE_NO_REBASE;
    uint16_t lastLocationOffset = 0xFFFF;
    for(uint32_t i=0; i < pageSize/4; ++i) {
        unsigned offset = i*4;
        if ( bitmap[i] ) {
            if ( startValue == DYLD_CACHE_SLIDE4_PAGE_NO_REBASE ) {
                // found first rebase location in page
                startValue = i;
            }
            else if ( !makeRebaseChainV4<P>(pageContent, lastLocationOffset, offset, info) ) {
                // can't record all rebasings in one chain
                if ( (startValue & DYLD_CACHE_SLIDE4_PAGE_USE_EXTRA) == 0 ) {
                    // switch page_start to "extras" which is a list of chain starts
                    unsigned indexInExtras = (unsigned)pageExtras.size();
                    if ( indexInExtras >= DYLD_CACHE_SLIDE4_PAGE_INDEX ) {
                        _diagnostics.error("rebase overflow in v4 page extras");
                        return;
                    }
                    pageExtras.push_back(startValue);
                    startValue = indexInExtras | DYLD_CACHE_SLIDE4_PAGE_USE_EXTRA;
                }
                pageExtras.push_back(i);
            }
            lastLocationOffset = offset;
        }
    }
    if ( lastLocationOffset != 0xFFFF ) {
        // mark end of chain
        pint_t* lastLoc = (pint_t*)&pageContent[lastLocationOffset];
        pint_t lastValue = (pint_t)P::getP(*lastLoc);
        pint_t newValue = ((lastValue - valueAdd) & valueMask);
        P::setP(*lastLoc, newValue);
    }
    if ( startValue & DYLD_CACHE_SLIDE4_PAGE_USE_EXTRA ) {
        // add end bit to extras
        pageExtras.back() |= DYLD_CACHE_SLIDE4_PAGE_EXTRA_END;
    }
    pageStarts.push_back(startValue);
}



template <typename P>
void CacheBuilder::writeSlideInfoV4(const bool bitmap[], unsigned dataPageCount)
{
    typedef typename P::uint_t    pint_t;
    typedef typename P::E         E;
    const uint32_t pageSize = 4096;

    // fill in fixed info
    assert(_slideInfoFileOffset != 0);
    dyld_cache_slide_info4* info = (dyld_cache_slide_info4*)_readOnlyRegion.buffer;
    info->version    = 4;
    info->page_size  = pageSize;
    info->delta_mask = _archLayout->pointerDeltaMask;
    info->value_add  = (sizeof(pint_t) == 8) ? 0 : _archLayout->sharedMemoryStart;  // only value_add for 32-bit archs

    // set page starts and extras for each page
    std::vector<uint16_t> pageStarts;
    std::vector<uint16_t> pageExtras;
    pageStarts.reserve(dataPageCount);
    uint8_t* pageContent = _readWriteRegion.buffer;
    const bool* bitmapForPage = bitmap;
    for (unsigned i=0; i < dataPageCount; ++i) {
        addPageStartsV4<P>(pageContent, bitmapForPage, info, pageStarts, pageExtras);
        if ( _diagnostics.hasError() ) {
            return;
        }
        pageContent += pageSize;
        bitmapForPage += (sizeof(bool)*(pageSize/4));
    }
    // fill in computed info
    info->page_starts_offset = sizeof(dyld_cache_slide_info4);
    info->page_starts_count  = (unsigned)pageStarts.size();
    info->page_extras_offset = (unsigned)(sizeof(dyld_cache_slide_info4)+pageStarts.size()*sizeof(uint16_t));
    info->page_extras_count  = (unsigned)pageExtras.size();
    uint16_t* pageStartsBuffer = (uint16_t*)((char*)info + info->page_starts_offset);
    uint16_t* pageExtrasBuffer = (uint16_t*)((char*)info + info->page_extras_offset);
    for (unsigned i=0; i < pageStarts.size(); ++i)
        pageStartsBuffer[i] = pageStarts[i];
    for (unsigned i=0; i < pageExtras.size(); ++i)
        pageExtrasBuffer[i] = pageExtras[i];
    // update header with final size
    uint64_t slideInfoSize = align(info->page_extras_offset + pageExtras.size()*sizeof(uint16_t), _archLayout->sharedRegionAlignP2);
    if ( slideInfoSize > _slideInfoBufferSizeAllocated ) {
        _diagnostics.error("kernel slide info v4 overflow buffer");
    }
    ((dyld_cache_header*)_readExecuteRegion.buffer)->slideInfoSize = slideInfoSize;
    //fprintf(stderr, "pageCount=%u, page_starts_count=%lu, page_extras_count=%lu\n", dataPageCount, pageStarts.size(), pageExtras.size());
}


/*
void CacheBuilder::writeSlideInfoV1()
{
    // build one 128-byte bitmap per page (4096) of DATA
    uint8_t* const dataStart = (uint8_t*)_buffer.get() + regions[1].fileOffset;
    uint8_t* const dataEnd   = dataStart + regions[1].size;
    const long bitmapSize = (dataEnd - dataStart)/(4*8);
    uint8_t* bitmap = (uint8_t*)calloc(bitmapSize, 1);
    for (void* p : _pointersForASLR) {
        if ( (p < dataStart) || ( p > dataEnd) )
            terminate("DATA pointer for sliding, out of range\n");
        long offset = (long)((uint8_t*)p - dataStart);
        if ( (offset % 4) != 0 )
            terminate("pointer not 4-byte aligned in DATA offset 0x%08lX\n", offset);
        long byteIndex = offset / (4*8);
        long bitInByte =  (offset % 32) >> 2;
        bitmap[byteIndex] |= (1 << bitInByte);
    }

    // allocate worst case size block of all slide info
    const unsigned entry_size = 4096/(8*4); // 8 bits per byte, possible pointer every 4 bytes.
    const unsigned toc_count = (unsigned)bitmapSize/entry_size;
    dyld_cache_slide_info* slideInfo = (dyld_cache_slide_info*)((uint8_t*)_buffer + _slideInfoFileOffset);
    slideInfo->version          = 1;
    slideInfo->toc_offset       = sizeof(dyld_cache_slide_info);
    slideInfo->toc_count        = toc_count;
    slideInfo->entries_offset   = (slideInfo->toc_offset+2*toc_count+127)&(-128);
    slideInfo->entries_count    = 0;
    slideInfo->entries_size     = entry_size;
    // append each unique entry
    const dyldCacheSlideInfoEntry* bitmapAsEntries = (dyldCacheSlideInfoEntry*)bitmap;
    dyldCacheSlideInfoEntry* const entriesInSlidInfo = (dyldCacheSlideInfoEntry*)((char*)slideInfo+slideInfo->entries_offset());
    int entry_count = 0;
    for (int i=0; i < toc_count; ++i) {
        const dyldCacheSlideInfoEntry* thisEntry = &bitmapAsEntries[i];
        // see if it is same as one already added
        bool found = false;
        for (int j=0; j < entry_count; ++j) {
            if ( memcmp(thisEntry, &entriesInSlidInfo[j], entry_size) == 0 ) {
                slideInfo->set_toc(i, j);
                found = true;
                break;
            }
        }
        if ( !found ) {
            // append to end
            memcpy(&entriesInSlidInfo[entry_count], thisEntry, entry_size);
            slideInfo->set_toc(i, entry_count++);
        }
    }
    slideInfo->entries_count  = entry_count;
    ::free((void*)bitmap);

    _buffer.header->slideInfoSize = align(slideInfo->entries_offset + entry_count*entry_size, _archLayout->sharedRegionAlignP2);
}

*/



uint16_t CacheBuilder::pageStartV3(uint8_t* pageContent, uint32_t pageSize, const bool bitmap[])
{
    const int maxPerPage = pageSize / 4;
    uint16_t result = DYLD_CACHE_SLIDE_V3_PAGE_ATTR_NO_REBASE;
    dyld3::MachOLoaded::ChainedFixupPointerOnDisk* lastLoc = nullptr;
    for (int i=0; i < maxPerPage; ++i) {
        if ( bitmap[i] ) {
            if ( result == DYLD_CACHE_SLIDE_V3_PAGE_ATTR_NO_REBASE ) {
                // found first rebase location in page
                result = i * 4;
            }
            dyld3::MachOLoaded::ChainedFixupPointerOnDisk* loc = (dyld3::MachOLoaded::ChainedFixupPointerOnDisk*)(pageContent + i*4);;
            if ( lastLoc != nullptr ) {
                // update chain (original chain may be wrong because of segment packing)
                lastLoc->plainRebase.next = loc - lastLoc;
            }
            lastLoc = loc;
        }
    }
    if ( lastLoc != nullptr ) {
        // mark last one as end of chain
        lastLoc->plainRebase.next = 0;
	}
    return result;
}


void CacheBuilder::writeSlideInfoV3(const bool bitmap[], unsigned dataPageCount)
{
	const uint32_t pageSize = 4096;

    // fill in fixed info
    assert(_slideInfoFileOffset != 0);
    dyld_cache_slide_info3* info = (dyld_cache_slide_info3*)_readOnlyRegion.buffer;
    info->version           = 3;
    info->page_size         = pageSize;
    info->page_starts_count = dataPageCount;
    info->auth_value_add    = _archLayout->sharedMemoryStart;
    
    // fill in per-page starts
    uint8_t* pageContent = _readWriteRegion.buffer;
    const bool* bitmapForPage = bitmap;
    for (unsigned i=0; i < dataPageCount; ++i) {
        info->page_starts[i] = pageStartV3(pageContent, pageSize, bitmapForPage);
        pageContent += pageSize;
        bitmapForPage += (sizeof(bool)*(pageSize/4));
    }

    // update header with final size
    dyld_cache_header* dyldCacheHeader = (dyld_cache_header*)_readExecuteRegion.buffer;
    dyldCacheHeader->slideInfoSize = align(__offsetof(dyld_cache_slide_info3, page_starts[dataPageCount]), _archLayout->sharedRegionAlignP2);
    if ( dyldCacheHeader->slideInfoSize > _slideInfoBufferSizeAllocated ) {
        _diagnostics.error("kernel slide info overflow buffer");
    }
}


void CacheBuilder::fipsSign()
{
    // find libcorecrypto.dylib in cache being built
    DyldSharedCache* dyldCache = (DyldSharedCache*)_readExecuteRegion.buffer;
    __block const dyld3::MachOLoaded* ml = nullptr;
    dyldCache->forEachImage(^(const mach_header* mh, const char* installName) {
        if ( strcmp(installName, "/usr/lib/system/libcorecrypto.dylib") == 0 )
            ml = (dyld3::MachOLoaded*)mh;
    });
    if ( ml == nullptr ) {
        _diagnostics.warning("Could not find libcorecrypto.dylib, skipping FIPS sealing");
        return;
    }

    // find location in libcorecrypto.dylib to store hash of __text section
    uint64_t hashStoreSize;
    const void* hashStoreLocation = ml->findSectionContent("__TEXT", "__fips_hmacs", hashStoreSize);
    if ( hashStoreLocation == nullptr ) {
        _diagnostics.warning("Could not find __TEXT/__fips_hmacs section in libcorecrypto.dylib, skipping FIPS sealing");
        return;
    }
    if ( hashStoreSize != 32 ) {
        _diagnostics.warning("__TEXT/__fips_hmacs section in libcorecrypto.dylib is not 32 bytes in size, skipping FIPS sealing");
        return;
    }

    // compute hmac hash of __text section
    uint64_t textSize;
    const void* textLocation = ml->findSectionContent("__TEXT", "__text", textSize);
    if ( textLocation == nullptr ) {
        _diagnostics.warning("Could not find __TEXT/__text section in libcorecrypto.dylib, skipping FIPS sealing");
        return;
    }
    unsigned char hmac_key = 0;
    CCHmac(kCCHmacAlgSHA256, &hmac_key, 1, textLocation, textSize, (void*)hashStoreLocation); // store hash directly into hashStoreLocation
}

void CacheBuilder::codeSign()
{
    uint8_t  dscHashType;
    uint8_t  dscHashSize;
    uint32_t dscDigestFormat;
    bool agile = false;

    // select which codesigning hash
    switch (_options.codeSigningDigestMode) {
        case DyldSharedCache::Agile:
            agile = true;
            // Fall through to SHA1, because the main code directory remains SHA1 for compatibility.
        case DyldSharedCache::SHA1only:
            dscHashType     = CS_HASHTYPE_SHA1;
            dscHashSize     = CS_HASH_SIZE_SHA1;
            dscDigestFormat = kCCDigestSHA1;
            break;
        case DyldSharedCache::SHA256only:
            dscHashType     = CS_HASHTYPE_SHA256;
            dscHashSize     = CS_HASH_SIZE_SHA256;
            dscDigestFormat = kCCDigestSHA256;
            break;
        default:
            _diagnostics.error("codeSigningDigestMode has unknown, unexpected value %d, bailing out.",
                               _options.codeSigningDigestMode);
            return;
    }

    std::string cacheIdentifier = "com.apple.dyld.cache." + _options.archName;
    if ( _options.dylibsRemovedDuringMastering ) {
        if ( _options.optimizeStubs  )
            cacheIdentifier = "com.apple.dyld.cache." + _options.archName + ".release";
        else
            cacheIdentifier = "com.apple.dyld.cache." + _options.archName + ".development";
    }
    // get pointers into shared cache buffer
    size_t          inBbufferSize = _readExecuteRegion.sizeInUse+_readWriteRegion.sizeInUse+_readOnlyRegion.sizeInUse+_localSymbolsRegion.sizeInUse;

    // layout code signature contents
    uint32_t blobCount     = agile ? 4 : 3;
    size_t   idSize        = cacheIdentifier.size()+1; // +1 for terminating 0
    uint32_t slotCount     = (uint32_t)((inBbufferSize + CS_PAGE_SIZE - 1) / CS_PAGE_SIZE);
    uint32_t xSlotCount    = CSSLOT_REQUIREMENTS;
    size_t   idOffset      = offsetof(CS_CodeDirectory, end_withExecSeg);
    size_t   hashOffset    = idOffset+idSize + dscHashSize*xSlotCount;
    size_t   hash256Offset = idOffset+idSize + CS_HASH_SIZE_SHA256*xSlotCount;
    size_t   cdSize        = hashOffset + (slotCount * dscHashSize);
    size_t   cd256Size     = agile ? hash256Offset + (slotCount * CS_HASH_SIZE_SHA256) : 0;
    size_t   reqsSize      = 12;
    size_t   cmsSize       = sizeof(CS_Blob);
    size_t   cdOffset      = sizeof(CS_SuperBlob) + blobCount*sizeof(CS_BlobIndex);
    size_t   cd256Offset   = cdOffset + cdSize;
    size_t   reqsOffset    = cd256Offset + cd256Size; // equals cdOffset + cdSize if not agile
    size_t   cmsOffset     = reqsOffset + reqsSize;
    size_t   sbSize        = cmsOffset + cmsSize;
    size_t   sigSize       = align(sbSize, 14);       // keep whole cache 16KB aligned

    // allocate space for blob
    vm_address_t codeSigAlloc;
    if ( vm_allocate(mach_task_self(), &codeSigAlloc, sigSize, VM_FLAGS_ANYWHERE) != 0 ) {
        _diagnostics.error("could not allocate code signature buffer");
        return;
    }
    _codeSignatureRegion.buffer     = (uint8_t*)codeSigAlloc;
    _codeSignatureRegion.bufferSize = sigSize;
    _codeSignatureRegion.sizeInUse  = sigSize;

    // create overall code signature which is a superblob
    CS_SuperBlob* sb = reinterpret_cast<CS_SuperBlob*>(_codeSignatureRegion.buffer);
    sb->magic           = htonl(CSMAGIC_EMBEDDED_SIGNATURE);
    sb->length          = htonl(sbSize);
    sb->count           = htonl(blobCount);
    sb->index[0].type   = htonl(CSSLOT_CODEDIRECTORY);
    sb->index[0].offset = htonl(cdOffset);
    sb->index[1].type   = htonl(CSSLOT_REQUIREMENTS);
    sb->index[1].offset = htonl(reqsOffset);
    sb->index[2].type   = htonl(CSSLOT_CMS_SIGNATURE);
    sb->index[2].offset = htonl(cmsOffset);
    if ( agile ) {
        sb->index[3].type = htonl(CSSLOT_ALTERNATE_CODEDIRECTORIES + 0);
        sb->index[3].offset = htonl(cd256Offset);
    }

    // fill in empty requirements
    CS_RequirementsBlob* reqs = (CS_RequirementsBlob*)(((char*)sb)+reqsOffset);
    reqs->magic  = htonl(CSMAGIC_REQUIREMENTS);
    reqs->length = htonl(sizeof(CS_RequirementsBlob));
    reqs->data   = 0;

    // initialize fixed fields of Code Directory
    CS_CodeDirectory* cd = (CS_CodeDirectory*)(((char*)sb)+cdOffset);
    cd->magic           = htonl(CSMAGIC_CODEDIRECTORY);
    cd->length          = htonl(cdSize);
    cd->version         = htonl(0x20400);               // supports exec segment
    cd->flags           = htonl(kSecCodeSignatureAdhoc);
    cd->hashOffset      = htonl(hashOffset);
    cd->identOffset     = htonl(idOffset);
    cd->nSpecialSlots   = htonl(xSlotCount);
    cd->nCodeSlots      = htonl(slotCount);
    cd->codeLimit       = htonl(inBbufferSize);
    cd->hashSize        = dscHashSize;
    cd->hashType        = dscHashType;
    cd->platform        = 0;                            // not platform binary
    cd->pageSize        = __builtin_ctz(CS_PAGE_SIZE);  // log2(CS_PAGE_SIZE);
    cd->spare2          = 0;                            // unused (must be zero)
    cd->scatterOffset   = 0;                            // not supported anymore
    cd->teamOffset      = 0;                            // no team ID
    cd->spare3          = 0;                            // unused (must be zero)
    cd->codeLimit64     = 0;                            // falls back to codeLimit

    // executable segment info
    cd->execSegBase     = htonll(_readExecuteRegion.cacheFileOffset); // base of TEXT segment
    cd->execSegLimit    = htonll(_readExecuteRegion.sizeInUse);       // size of TEXT segment
    cd->execSegFlags    = 0;                                          // not a main binary

    // initialize dynamic fields of Code Directory
    strcpy((char*)cd + idOffset, cacheIdentifier.c_str());

    // add special slot hashes
    uint8_t* hashSlot = (uint8_t*)cd + hashOffset;
    uint8_t* reqsHashSlot = &hashSlot[-CSSLOT_REQUIREMENTS*dscHashSize];
    CCDigest(dscDigestFormat, (uint8_t*)reqs, sizeof(CS_RequirementsBlob), reqsHashSlot);

    CS_CodeDirectory* cd256;
    uint8_t* hash256Slot;
    uint8_t* reqsHash256Slot;
    if ( agile ) {
        // Note that the assumption here is that the size up to the hashes is the same as for
        // sha1 code directory, and that they come last, after everything else.

        cd256 = (CS_CodeDirectory*)(((char*)sb)+cd256Offset);
        cd256->magic           = htonl(CSMAGIC_CODEDIRECTORY);
        cd256->length          = htonl(cd256Size);
        cd256->version         = htonl(0x20400);               // supports exec segment
        cd256->flags           = htonl(kSecCodeSignatureAdhoc);
        cd256->hashOffset      = htonl(hash256Offset);
        cd256->identOffset     = htonl(idOffset);
        cd256->nSpecialSlots   = htonl(xSlotCount);
        cd256->nCodeSlots      = htonl(slotCount);
        cd256->codeLimit       = htonl(inBbufferSize);
        cd256->hashSize        = CS_HASH_SIZE_SHA256;
        cd256->hashType        = CS_HASHTYPE_SHA256;
        cd256->platform        = 0;                            // not platform binary
        cd256->pageSize        = __builtin_ctz(CS_PAGE_SIZE);  // log2(CS_PAGE_SIZE);
        cd256->spare2          = 0;                            // unused (must be zero)
        cd256->scatterOffset   = 0;                            // not supported anymore
        cd256->teamOffset      = 0;                            // no team ID
        cd256->spare3          = 0;                            // unused (must be zero)
        cd256->codeLimit64     = 0;                            // falls back to codeLimit

        // executable segment info
        cd256->execSegBase     = cd->execSegBase;
        cd256->execSegLimit    = cd->execSegLimit;
        cd256->execSegFlags    = cd->execSegFlags;

        // initialize dynamic fields of Code Directory
        strcpy((char*)cd256 + idOffset, cacheIdentifier.c_str());

        // add special slot hashes
        hash256Slot = (uint8_t*)cd256 + hash256Offset;
        reqsHash256Slot = &hash256Slot[-CSSLOT_REQUIREMENTS*CS_HASH_SIZE_SHA256];
        CCDigest(kCCDigestSHA256, (uint8_t*)reqs, sizeof(CS_RequirementsBlob), reqsHash256Slot);
    }
    else {
        cd256 = NULL;
        hash256Slot = NULL;
        reqsHash256Slot = NULL;
    }

    // fill in empty CMS blob for ad-hoc signing
    CS_Blob* cms = (CS_Blob*)(((char*)sb)+cmsOffset);
    cms->magic  = htonl(CSMAGIC_BLOBWRAPPER);
    cms->length = htonl(sizeof(CS_Blob));


    // alter header of cache to record size and location of code signature
    // do this *before* hashing each page
    dyld_cache_header* cache = (dyld_cache_header*)_readExecuteRegion.buffer;
    cache->codeSignatureOffset  = inBbufferSize;
    cache->codeSignatureSize    = sigSize;

    const uint32_t rwSlotStart = (uint32_t)(_readExecuteRegion.sizeInUse / CS_PAGE_SIZE);
    const uint32_t roSlotStart = (uint32_t)(rwSlotStart + _readWriteRegion.sizeInUse / CS_PAGE_SIZE);
    const uint32_t localsSlotStart = (uint32_t)(roSlotStart + _readOnlyRegion.sizeInUse / CS_PAGE_SIZE);
    auto codeSignPage = ^(size_t i) {
        const uint8_t* code = nullptr;
        // move to correct region
        if ( i < rwSlotStart )
            code = _readExecuteRegion.buffer + (i * CS_PAGE_SIZE);
        else if ( i >= rwSlotStart && i < roSlotStart )
            code = _readWriteRegion.buffer + ((i - rwSlotStart) * CS_PAGE_SIZE);
        else if ( i >= roSlotStart && i < localsSlotStart )
            code = _readOnlyRegion.buffer + ((i - roSlotStart) * CS_PAGE_SIZE);
        else
            code = _localSymbolsRegion.buffer + ((i - localsSlotStart) * CS_PAGE_SIZE);

        CCDigest(dscDigestFormat, code, CS_PAGE_SIZE, hashSlot + (i * dscHashSize));

        if ( agile ) {
            CCDigest(kCCDigestSHA256, code, CS_PAGE_SIZE, hash256Slot + (i * CS_HASH_SIZE_SHA256));
        }
    };

    // compute hashes
    dispatch_apply(slotCount, DISPATCH_APPLY_AUTO, ^(size_t i) {
        codeSignPage(i);
    });

    // Now that we have a code signature, compute a UUID from it.

    // Clear existing UUID, then MD5 whole cache buffer.
    {
        uint8_t* uuidLoc = cache->uuid;
        assert(uuid_is_null(uuidLoc));
        static_assert(offsetof(dyld_cache_header, uuid) / CS_PAGE_SIZE == 0, "uuid is expected in the first page of the cache");
        CC_MD5((const void*)cd, (unsigned)cdSize, uuidLoc);
        // <rdar://problem/6723729> uuids should conform to RFC 4122 UUID version 4 & UUID version 5 formats
        uuidLoc[6] = ( uuidLoc[6] & 0x0F ) | ( 3 << 4 );
        uuidLoc[8] = ( uuidLoc[8] & 0x3F ) | 0x80;

        // Now codesign page 0 again
        codeSignPage(0);
    }

    // hash of entire code directory (cdHash) uses same hash as each page
    uint8_t fullCdHash[dscHashSize];
    CCDigest(dscDigestFormat, (const uint8_t*)cd, cdSize, fullCdHash);
    // Note: cdHash is defined as first 20 bytes of hash
    memcpy(_cdHashFirst, fullCdHash, 20);
    if ( agile ) {
        uint8_t fullCdHash256[CS_HASH_SIZE_SHA256];
        CCDigest(kCCDigestSHA256, (const uint8_t*)cd256, cd256Size, fullCdHash256);
        // Note: cdHash is defined as first 20 bytes of hash, even for sha256
        memcpy(_cdHashSecond, fullCdHash256, 20);
    }
    else {
        memset(_cdHashSecond, 0, 20);
    }
}

const bool CacheBuilder::agileSignature()
{
    return _options.codeSigningDigestMode == DyldSharedCache::Agile;
}

static const std::string cdHash(uint8_t hash[20])
{
    char buff[48];
    for (int i = 0; i < 20; ++i)
        sprintf(&buff[2*i], "%2.2x", hash[i]);
    return buff;
}

const std::string CacheBuilder::cdHashFirst()
{
    return cdHash(_cdHashFirst);
}

const std::string CacheBuilder::cdHashSecond()
{
    return cdHash(_cdHashSecond);
}


void CacheBuilder::buildImageArray(std::vector<DyldSharedCache::FileAlias>& aliases)
{
    typedef dyld3::closure::ClosureBuilder::CachedDylibInfo         CachedDylibInfo;
    typedef dyld3::closure::Image::PatchableExport::PatchLocation   PatchLocation;
    typedef uint64_t                                                CacheOffset;

    // convert STL data structures to simple arrays to passe to makeDyldCacheImageArray()
    __block std::vector<CachedDylibInfo> dylibInfos;
    __block std::unordered_map<dyld3::closure::ImageNum, const dyld3::MachOLoaded*> imageNumToML;
    DyldSharedCache* cache = (DyldSharedCache*)_readExecuteRegion.buffer;
    cache->forEachImage(^(const mach_header* mh, const char* installName) {
        uint64_t mtime;
        uint64_t inode;
        cache->getIndexedImageEntry((uint32_t)dylibInfos.size(), mtime, inode);
        CachedDylibInfo entry;
        entry.fileInfo.fileContent  = mh;
        entry.fileInfo.path         = installName;
        entry.fileInfo.sliceOffset  = 0;
        entry.fileInfo.inode        = inode;
        entry.fileInfo.mtime        = mtime;
        dylibInfos.push_back(entry);
        imageNumToML[(dyld3::closure::ImageNum)(dylibInfos.size())] = (dyld3::MachOLoaded*)mh;
    });

    // Convert symlinks from STL to simple char pointers.
    std::vector<dyld3::closure::ClosureBuilder::CachedDylibAlias> dylibAliases;
    dylibAliases.reserve(aliases.size());
    for (const auto& alias : aliases)
        dylibAliases.push_back({ alias.realPath.c_str(), alias.aliasPath.c_str() });


    __block std::unordered_map<const dyld3::MachOLoaded*, std::set<CacheOffset>>    dylibToItsExports;
    __block std::unordered_map<CacheOffset, std::vector<PatchLocation>>             exportsToUses;
    __block std::unordered_map<CacheOffset, const char*>                            exportsToName;

    dyld3::closure::ClosureBuilder::CacheDylibsBindingHandlers handlers;

    handlers.chainedBind = ^(dyld3::closure::ImageNum, const dyld3::MachOLoaded* imageLoadAddress,
                             const dyld3::Array<uint64_t>& starts,
                             const dyld3::Array<dyld3::closure::Image::ResolvedSymbolTarget>& targets,
                             const dyld3::Array<dyld3::closure::ClosureBuilder::ResolvedTargetInfo>& targetInfos) {
        for (uint64_t start : starts) {
            dyld3::closure::Image::forEachChainedFixup((void*)imageLoadAddress, start, ^(uint64_t* fixupLoc, dyld3::MachOLoaded::ChainedFixupPointerOnDisk fixupInfo, bool& stop) {
                 // record location in aslr tracker so kernel can slide this on page-in
                _aslrTracker.add(fixupLoc);

                // if bind, record info for patch table and convert to rebase
                if ( fixupInfo.plainBind.bind ) {
                    dyld3::closure::Image::ResolvedSymbolTarget               target     = targets[fixupInfo.plainBind.ordinal];
                    const dyld3::closure::ClosureBuilder::ResolvedTargetInfo& targetInfo = targetInfos[fixupInfo.plainBind.ordinal];
                    dyld3::MachOLoaded::ChainedFixupPointerOnDisk*            loc;
                    uint64_t                                                  offsetInCache;
                    switch ( target.sharedCache.kind ) {
                        case dyld3::closure::Image::ResolvedSymbolTarget::kindSharedCache:
                            loc = (dyld3::MachOLoaded::ChainedFixupPointerOnDisk*)fixupLoc;
                            offsetInCache = target.sharedCache.offset - targetInfo.addend;
                            dylibToItsExports[targetInfo.foundInDylib].insert(offsetInCache);
                            exportsToName[offsetInCache] = targetInfo.foundSymbolName;
                            if ( fixupInfo.authBind.auth ) {
                                // turn this auth bind into an auth rebase into the cache
                                loc->authRebase.bind = 0;
                                loc->authRebase.target = target.sharedCache.offset;
                                exportsToUses[offsetInCache].push_back(PatchLocation((uint8_t*)fixupLoc - _readExecuteRegion.buffer, targetInfo.addend, *loc));
                            }
                            else {
                                // turn this plain bind into an plain rebase into the cache
                                loc->plainRebase.bind = 0;
                                loc->plainRebase.target = _archLayout->sharedMemoryStart + target.sharedCache.offset;
                                exportsToUses[offsetInCache].push_back(PatchLocation((uint8_t*)fixupLoc - _readExecuteRegion.buffer, targetInfo.addend));
                            }
                            break;
                        case dyld3::closure::Image::ResolvedSymbolTarget::kindAbsolute:
                             if ( _archLayout->is64 )
                                *((uint64_t*)fixupLoc) = target.absolute.value;
                            else
                                *((uint32_t*)fixupLoc) = (uint32_t)(target.absolute.value);
                            // don't record absolute targets for ASLR
                            break;
                        default:
                            assert(0 && "unsupported ResolvedSymbolTarget kind in dyld cache");
                    }
                }
            });
        }
    };

    handlers.rebase = ^(dyld3::closure::ImageNum imageNum, const dyld3::MachOLoaded* imageToFix, uint32_t runtimeOffset) {
        // record location in aslr tracker so kernel can slide this on page-in
        uint8_t* fixupLoc = (uint8_t*)imageToFix+runtimeOffset;
        _aslrTracker.add(fixupLoc);
    };

    handlers.bind = ^(dyld3::closure::ImageNum imageNum, const dyld3::MachOLoaded* mh,
                      uint32_t runtimeOffset, dyld3::closure::Image::ResolvedSymbolTarget target,
                      const dyld3::closure::ClosureBuilder::ResolvedTargetInfo& targetInfo) {
        uint8_t* fixupLoc = (uint8_t*)mh+runtimeOffset;

        // binder is called a second time for weak_bind info, which we ignore when building cache
        const bool weakDefUseAlreadySet = targetInfo.weakBindCoalese && _aslrTracker.has(fixupLoc);

        // do actual bind that sets pointer in image to unslid target address
        uint64_t offsetInCache;
        switch ( target.sharedCache.kind ) {
            case dyld3::closure::Image::ResolvedSymbolTarget::kindSharedCache:
                offsetInCache = target.sharedCache.offset - targetInfo.addend;
                dylibToItsExports[targetInfo.foundInDylib].insert(offsetInCache);
                exportsToUses[offsetInCache].push_back(PatchLocation(fixupLoc - _readExecuteRegion.buffer, targetInfo.addend));
                exportsToName[offsetInCache] = targetInfo.foundSymbolName;
                if ( !weakDefUseAlreadySet ) {
                    if ( _archLayout->is64 )
                        *((uint64_t*)fixupLoc) = _archLayout->sharedMemoryStart + target.sharedCache.offset;
                    else
                        *((uint32_t*)fixupLoc) = (uint32_t)(_archLayout->sharedMemoryStart + target.sharedCache.offset);
                    // record location in aslr tracker so kernel can slide this on page-in
                    _aslrTracker.add(fixupLoc);
                }
                break;
            case dyld3::closure::Image::ResolvedSymbolTarget::kindAbsolute:
                 if ( _archLayout->is64 )
                    *((uint64_t*)fixupLoc) = target.absolute.value;
                else
                    *((uint32_t*)fixupLoc) = (uint32_t)(target.absolute.value);
                // don't record absolute targets for ASLR
                // HACK: Split seg may have added a target.  Remove it
                _aslrTracker.remove(fixupLoc);
                if ( (targetInfo.libOrdinal > 0) && (targetInfo.libOrdinal <= mh->dependentDylibCount()) ) {
                    _missingWeakImports[fixupLoc] = mh->dependentDylibLoadPath(targetInfo.libOrdinal - 1);
                }
                break;
            default:
                assert(0 && "unsupported ResolvedSymbolTarget kind in dyld cache");
        }
    };

    handlers.forEachExportsPatch = ^(dyld3::closure::ImageNum imageNum, void (^handler)(const dyld3::closure::ClosureBuilder::CacheDylibsBindingHandlers::PatchInfo&)) {
        const dyld3::MachOLoaded* ml = imageNumToML[imageNum];
        for (CacheOffset exportCacheOffset : dylibToItsExports[ml]) {
            dyld3::closure::ClosureBuilder::CacheDylibsBindingHandlers::PatchInfo info;
            std::vector<PatchLocation>& uses = exportsToUses[exportCacheOffset];
            uses.erase(std::unique(uses.begin(), uses.end()), uses.end());
            info.exportCacheOffset = (uint32_t)exportCacheOffset;
            info.exportSymbolName  = exportsToName[exportCacheOffset];
            info.usesCount         = (uint32_t)uses.size();
            info.usesArray         = &uses.front();
            handler(info);
        }
    };


    // build ImageArray for all dylibs in dyld cache
    dyld3::closure::PathOverrides pathOverrides;
    dyld3::closure::ClosureBuilder cb(dyld3::closure::kFirstDyldCacheImageNum, _fileSystem, cache, false, pathOverrides, dyld3::closure::ClosureBuilder::AtPath::none, nullptr, _archLayout->archName, _options.platform, &handlers);
    dyld3::Array<CachedDylibInfo> dylibs(&dylibInfos[0], dylibInfos.size(), dylibInfos.size());
    const dyld3::Array<dyld3::closure::ClosureBuilder::CachedDylibAlias> aliasesArray(dylibAliases.data(), dylibAliases.size(), dylibAliases.size());
    _imageArray = cb.makeDyldCacheImageArray(_options.optimizeStubs, dylibs, aliasesArray);
    if ( cb.diagnostics().hasError() ) {
        _diagnostics.error("%s", cb.diagnostics().errorMessage().c_str());
        return;
    }
}

void CacheBuilder::addImageArray()
{
    // build trie of dylib paths
    __block std::vector<DylibIndexTrie::Entry> dylibEntrys;
    _imageArray->forEachImage(^(const dyld3::closure::Image* image, bool& stop) {
        dylibEntrys.push_back(DylibIndexTrie::Entry(image->path(), DylibIndex(image->imageNum()-1)));
        image->forEachAlias(^(const char *aliasPath, bool &innerStop) {
            dylibEntrys.push_back(DylibIndexTrie::Entry(aliasPath, DylibIndex(image->imageNum()-1)));
        });
    });
    DylibIndexTrie dylibsTrie(dylibEntrys);
    std::vector<uint8_t> trieBytes;
    dylibsTrie.emit(trieBytes);
    while ( (trieBytes.size() % 4) != 0 )
        trieBytes.push_back(0);

    // check for fit
    uint64_t imageArraySize = _imageArray->size();
    size_t freeSpace = _readOnlyRegion.bufferSize - _readOnlyRegion.sizeInUse;
    if ( imageArraySize+trieBytes.size() > freeSpace ) {
        _diagnostics.error("cache buffer too small to hold ImageArray and Trie (buffer size=%lldMB, imageArray size=%lldMB, trie size=%luKB, free space=%ldMB)",
                            _allocatedBufferSize/1024/1024, imageArraySize/1024/1024, trieBytes.size()/1024, freeSpace/1024/1024);
        return;
    }

    // copy into cache and update header
    DyldSharedCache* dyldCache = (DyldSharedCache*)_readExecuteRegion.buffer;
    dyldCache->header.dylibsImageArrayAddr = _readOnlyRegion.unslidLoadAddress + _readOnlyRegion.sizeInUse;
    dyldCache->header.dylibsImageArraySize = imageArraySize;
    dyldCache->header.dylibsTrieAddr       = dyldCache->header.dylibsImageArrayAddr + imageArraySize;
    dyldCache->header.dylibsTrieSize       = trieBytes.size();
    ::memcpy(_readOnlyRegion.buffer + _readOnlyRegion.sizeInUse, _imageArray, imageArraySize);
    ::memcpy(_readOnlyRegion.buffer + _readOnlyRegion.sizeInUse + imageArraySize, &trieBytes[0], trieBytes.size());
    _readOnlyRegion.sizeInUse += align(imageArraySize+trieBytes.size(),14);
}

void CacheBuilder::addOtherImageArray(const std::vector<LoadedMachO>& otherDylibsAndBundles, std::vector<const LoadedMachO*>& overflowDylibs)
{
    DyldSharedCache* cache = (DyldSharedCache*)_readExecuteRegion.buffer;
    dyld3::closure::PathOverrides pathOverrides;
    dyld3::closure::ClosureBuilder cb(dyld3::closure::kFirstOtherOSImageNum, _fileSystem, cache, false, pathOverrides, dyld3::closure::ClosureBuilder::AtPath::none, nullptr, _archLayout->archName, _options.platform);

    // make ImageArray for other dylibs and bundles
    STACK_ALLOC_ARRAY(dyld3::closure::LoadedFileInfo, others, otherDylibsAndBundles.size() + overflowDylibs.size());
    for (const LoadedMachO& other : otherDylibsAndBundles) {
        if ( !contains(other.loadedFileInfo.path, ".app/") )
            others.push_back(other.loadedFileInfo);
    }

    for (const LoadedMachO* dylib : overflowDylibs) {
        if (dylib->mappedFile.mh->canHavePrecomputedDlopenClosure(dylib->mappedFile.runtimePath.c_str(), ^(const char*) {}) )
            others.push_back(dylib->loadedFileInfo);
    }

    // Sort the others array by name so that it is deterministic
    std::sort(others.begin(), others.end(),
              [](const dyld3::closure::LoadedFileInfo& a, const dyld3::closure::LoadedFileInfo& b) {
                  return strcmp(a.path, b.path) < 0;
              });

    const dyld3::closure::ImageArray* otherImageArray = cb.makeOtherDylibsImageArray(others, (uint32_t)_sortedDylibs.size());

    // build trie of paths
    __block std::vector<DylibIndexTrie::Entry> otherEntrys;
    otherImageArray->forEachImage(^(const dyld3::closure::Image* image, bool& stop) {
        if ( !image->isInvalid() )
            otherEntrys.push_back(DylibIndexTrie::Entry(image->path(), DylibIndex(image->imageNum())));
    });
    DylibIndexTrie dylibsTrie(otherEntrys);
    std::vector<uint8_t> trieBytes;
    dylibsTrie.emit(trieBytes);
    while ( (trieBytes.size() % 4) != 0 )
        trieBytes.push_back(0);

    // check for fit
    uint64_t imageArraySize = otherImageArray->size();
    size_t freeSpace = _readOnlyRegion.bufferSize - _readOnlyRegion.sizeInUse;
    if ( imageArraySize+trieBytes.size() > freeSpace ) {
        _diagnostics.error("cache buffer too small to hold ImageArray and Trie (buffer size=%lldMB, imageArray size=%lldMB, trie size=%luKB, free space=%ldMB)",
                           _allocatedBufferSize/1024/1024, imageArraySize/1024/1024, trieBytes.size()/1024, freeSpace/1024/1024);
        return;
    }

    // copy into cache and update header
    DyldSharedCache* dyldCache = (DyldSharedCache*)_readExecuteRegion.buffer;
    dyldCache->header.otherImageArrayAddr = _readOnlyRegion.unslidLoadAddress + _readOnlyRegion.sizeInUse;
    dyldCache->header.otherImageArraySize = imageArraySize;
    dyldCache->header.otherTrieAddr       = dyldCache->header.otherImageArrayAddr + imageArraySize;
    dyldCache->header.otherTrieSize       = trieBytes.size();
    ::memcpy(_readOnlyRegion.buffer + _readOnlyRegion.sizeInUse, otherImageArray, imageArraySize);
    ::memcpy(_readOnlyRegion.buffer + _readOnlyRegion.sizeInUse + imageArraySize, &trieBytes[0], trieBytes.size());
    _readOnlyRegion.sizeInUse += align(imageArraySize+trieBytes.size(),14);
}


void CacheBuilder::addClosures(const std::vector<LoadedMachO>& osExecutables)
{
    const DyldSharedCache* dyldCache = (DyldSharedCache*)_readExecuteRegion.buffer;

    __block std::vector<Diagnostics> osExecutablesDiags;
    __block std::vector<const dyld3::closure::LaunchClosure*> osExecutablesClosures;
    osExecutablesDiags.resize(osExecutables.size());
    osExecutablesClosures.resize(osExecutables.size());

    dispatch_apply(osExecutables.size(), DISPATCH_APPLY_AUTO, ^(size_t index) {
        const LoadedMachO& loadedMachO = osExecutables[index];
        // don't pre-build closures for staged apps into dyld cache, since they won't run from that location
        if ( startsWith(loadedMachO.mappedFile.runtimePath, "/private/var/staged_system_apps/") ) {
            return;
        }
        dyld3::closure::PathOverrides pathOverrides;
        dyld3::closure::ClosureBuilder builder(dyld3::closure::kFirstLaunchClosureImageNum, _fileSystem, dyldCache, false, pathOverrides, dyld3::closure::ClosureBuilder::AtPath::all, nullptr, _archLayout->archName, _options.platform, nullptr);
        bool issetuid = false;
        if ( this->_options.platform == dyld3::Platform::macOS )
            _fileSystem.fileExists(loadedMachO.loadedFileInfo.path, nullptr, nullptr, &issetuid);
        const dyld3::closure::LaunchClosure* mainClosure = builder.makeLaunchClosure(loadedMachO.loadedFileInfo, issetuid);
        if ( builder.diagnostics().hasError() ) {
           osExecutablesDiags[index].error("%s", builder.diagnostics().errorMessage().c_str());
        }
        else {
            assert(mainClosure != nullptr);
            osExecutablesClosures[index] = mainClosure;
        }
    });

    std::map<std::string, const dyld3::closure::LaunchClosure*> closures;
    for (uint64_t i = 0, e = osExecutables.size(); i != e; ++i) {
        const LoadedMachO& loadedMachO = osExecutables[i];
        const Diagnostics& diag = osExecutablesDiags[i];
        if (diag.hasError()) {
            if ( _options.verbose ) {
                _diagnostics.warning("building closure for '%s': %s", loadedMachO.mappedFile.runtimePath.c_str(), diag.errorMessage().c_str());
                for (const std::string& warn : diag.warnings() )
                    _diagnostics.warning("%s", warn.c_str());
            }
            if ( loadedMachO.inputFile && (loadedMachO.inputFile->mustBeIncluded()) ) {
                loadedMachO.inputFile->diag.error("%s", diag.errorMessage().c_str());
            }
        } else {
            // Note, a closure could be null here if it has a path we skip.
            if (osExecutablesClosures[i] != nullptr)
                closures[loadedMachO.mappedFile.runtimePath] = osExecutablesClosures[i];
        }
    }

    osExecutablesDiags.clear();
    osExecutablesClosures.clear();

    // preflight space needed
    size_t closuresSpace = 0;
    for (const auto& entry : closures) {
        closuresSpace += entry.second->size();
    }
    size_t freeSpace = _readOnlyRegion.bufferSize - _readOnlyRegion.sizeInUse;
    if ( closuresSpace > freeSpace ) {
        _diagnostics.error("cache buffer too small to hold all closures (buffer size=%lldMB, closures size=%ldMB, free space=%ldMB)",
                            _allocatedBufferSize/1024/1024, closuresSpace/1024/1024, freeSpace/1024/1024);
        return;
    }
    DyldSharedCache* cache = (DyldSharedCache*)_readExecuteRegion.buffer;
    cache->header.progClosuresAddr = _readOnlyRegion.unslidLoadAddress + _readOnlyRegion.sizeInUse;
    uint8_t* closuresBase = _readOnlyRegion.buffer + _readOnlyRegion.sizeInUse;
    std::vector<DylibIndexTrie::Entry> closureEntrys;
    uint32_t currentClosureOffset = 0;
    for (const auto& entry : closures) {
        const dyld3::closure::LaunchClosure* closure = entry.second;
        closureEntrys.push_back(DylibIndexTrie::Entry(entry.first, DylibIndex(currentClosureOffset)));
        size_t size = closure->size();
        assert((size % 4) == 0);
        memcpy(closuresBase+currentClosureOffset, closure, size);
        currentClosureOffset += size;
        freeSpace -= size;
        closure->deallocate();
    }
    cache->header.progClosuresSize = currentClosureOffset;
    _readOnlyRegion.sizeInUse += currentClosureOffset;
    freeSpace = _readOnlyRegion.bufferSize - _readOnlyRegion.sizeInUse;
    // build trie of indexes into closures list
    DylibIndexTrie closureTrie(closureEntrys);
    std::vector<uint8_t> trieBytes;
    closureTrie.emit(trieBytes);
    while ( (trieBytes.size() % 8) != 0 )
        trieBytes.push_back(0);
    if ( trieBytes.size() > freeSpace ) {
        _diagnostics.error("cache buffer too small to hold all closures trie (buffer size=%lldMB, trie size=%ldMB, free space=%ldMB)",
                            _allocatedBufferSize/1024/1024, trieBytes.size()/1024/1024, freeSpace/1024/1024);
        return;
    }
    memcpy(_readOnlyRegion.buffer + _readOnlyRegion.sizeInUse, &trieBytes[0], trieBytes.size());
    cache->header.progClosuresTrieAddr = _readOnlyRegion.unslidLoadAddress + _readOnlyRegion.sizeInUse;
    cache->header.progClosuresTrieSize = trieBytes.size();
    _readOnlyRegion.sizeInUse += trieBytes.size();
    _readOnlyRegion.sizeInUse = align(_readOnlyRegion.sizeInUse, 14);
}


bool CacheBuilder::writeCache(void (^cacheSizeCallback)(uint64_t size), bool (^copyCallback)(const uint8_t* src, uint64_t size, uint64_t dstOffset))
{
    const dyld_cache_header*       cacheHeader = (dyld_cache_header*)_readExecuteRegion.buffer;
    const dyld_cache_mapping_info* mappings = (dyld_cache_mapping_info*)(_readExecuteRegion.buffer + cacheHeader->mappingOffset);
    assert(_readExecuteRegion.sizeInUse       == mappings[0].size);
    assert(_readWriteRegion.sizeInUse         == mappings[1].size);
    assert(_readOnlyRegion.sizeInUse          == mappings[2].size);
    assert(_readExecuteRegion.cacheFileOffset == mappings[0].fileOffset);
    assert(_readWriteRegion.cacheFileOffset   == mappings[1].fileOffset);
    assert(_readOnlyRegion.cacheFileOffset    == mappings[2].fileOffset);
    assert(_codeSignatureRegion.sizeInUse     == cacheHeader->codeSignatureSize);
    assert(cacheHeader->codeSignatureOffset   == mappings[2].fileOffset+_readOnlyRegion.sizeInUse+_localSymbolsRegion.sizeInUse);
    cacheSizeCallback(_readExecuteRegion.sizeInUse+_readWriteRegion.sizeInUse+_readOnlyRegion.sizeInUse+_localSymbolsRegion.sizeInUse+_codeSignatureRegion.sizeInUse);
    bool fullyWritten = copyCallback(_readExecuteRegion.buffer, _readExecuteRegion.sizeInUse, mappings[0].fileOffset);
    fullyWritten &= copyCallback(_readWriteRegion.buffer, _readWriteRegion.sizeInUse, mappings[1].fileOffset);
    fullyWritten &= copyCallback(_readOnlyRegion.buffer, _readOnlyRegion.sizeInUse, mappings[2].fileOffset);
    if ( _localSymbolsRegion.sizeInUse != 0 ) {
        assert(cacheHeader->localSymbolsOffset == mappings[2].fileOffset+_readOnlyRegion.sizeInUse);
        fullyWritten &= copyCallback(_localSymbolsRegion.buffer, _localSymbolsRegion.sizeInUse, cacheHeader->localSymbolsOffset);
    }
    fullyWritten &= copyCallback(_codeSignatureRegion.buffer, _codeSignatureRegion.sizeInUse, cacheHeader->codeSignatureOffset);
    return fullyWritten;
}


void CacheBuilder::writeFile(const std::string& path)
{
    std::string pathTemplate = path + "-XXXXXX";
    size_t templateLen = strlen(pathTemplate.c_str())+2;
    char pathTemplateSpace[templateLen];
    strlcpy(pathTemplateSpace, pathTemplate.c_str(), templateLen);
    int fd = mkstemp(pathTemplateSpace);
    if ( fd != -1 ) {
        auto cacheSizeCallback = ^(uint64_t size) {
            ::ftruncate(fd, size);
        };
        auto copyCallback = ^(const uint8_t* src, uint64_t size, uint64_t dstOffset) {
            uint64_t writtenSize = pwrite(fd, src, size, dstOffset);
            return writtenSize == size;
        };
        bool fullyWritten = writeCache(cacheSizeCallback, copyCallback);
        if ( fullyWritten ) {
            ::fchmod(fd, S_IRUSR|S_IWUSR|S_IRGRP|S_IROTH); // mkstemp() makes file "rw-------", switch it to "rw-r--r--"
            if ( ::rename(pathTemplateSpace, path.c_str()) == 0) {
                ::close(fd);
                return; // success
            }
        }
        else {
            _diagnostics.error("could not write file %s", pathTemplateSpace);
        }
        ::close(fd);
        ::unlink(pathTemplateSpace);
    }
    else {
        _diagnostics.error("could not open file %s", pathTemplateSpace);
    }
}

void CacheBuilder::writeBuffer(uint8_t*& buffer, uint64_t& bufferSize) {
    auto cacheSizeCallback = ^(uint64_t size) {
        buffer = (uint8_t*)malloc(size);
        bufferSize = size;
    };
    auto copyCallback = ^(const uint8_t* src, uint64_t size, uint64_t dstOffset) {
        memcpy(buffer + dstOffset, src, size);
        return true;
    };
    bool fullyWritten = writeCache(cacheSizeCallback, copyCallback);
    assert(fullyWritten);
}

void CacheBuilder::writeMapFile(const std::string& path)
{
    const DyldSharedCache* cache = (DyldSharedCache*)_readExecuteRegion.buffer;
    std::string mapContent = cache->mapFile();
    safeSave(mapContent.c_str(), mapContent.size(), path);
}

void CacheBuilder::writeMapFileBuffer(uint8_t*& buffer, uint64_t& bufferSize)
{
    const DyldSharedCache* cache = (DyldSharedCache*)_readExecuteRegion.buffer;
    std::string mapContent = cache->mapFile();
    buffer = (uint8_t*)malloc(mapContent.size() + 1);
    bufferSize = mapContent.size() + 1;
    memcpy(buffer, mapContent.data(), bufferSize);
}


void CacheBuilder::forEachCacheDylib(void (^callback)(const std::string& path)) {
    for (const DylibInfo& dylibInfo : _sortedDylibs)
        callback(dylibInfo.runtimePath);
}


CacheBuilder::ASLR_Tracker::~ASLR_Tracker()
{
    if ( _bitmap != nullptr )
        ::free(_bitmap);
}

void CacheBuilder::ASLR_Tracker::setDataRegion(const void* rwRegionStart, size_t rwRegionSize)
{
    _pageCount   = (unsigned)(rwRegionSize+_pageSize-1)/_pageSize;
    _regionStart = (uint8_t*)rwRegionStart;
    _endStart    = (uint8_t*)rwRegionStart + rwRegionSize;
    _bitmap      = (bool*)calloc(_pageCount*(_pageSize/4)*sizeof(bool), 1);
}

void CacheBuilder::ASLR_Tracker::add(void* loc)
{
    uint8_t* p = (uint8_t*)loc;
    assert(p >= _regionStart);
    assert(p < _endStart);
    _bitmap[(p-_regionStart)/4] = true;
}

void CacheBuilder::ASLR_Tracker::remove(void* loc)
{
    uint8_t* p = (uint8_t*)loc;
    assert(p >= _regionStart);
    assert(p < _endStart);
    _bitmap[(p-_regionStart)/4] = false;
}

bool CacheBuilder::ASLR_Tracker::has(void* loc)
{
    uint8_t* p = (uint8_t*)loc;
    assert(p >= _regionStart);
    assert(p < _endStart);
    return _bitmap[(p-_regionStart)/4];
}