Loading...
--- dyld/dyld-95.3/src/dyldInitialization.cpp
+++ dyld/dyld-852/src/dyldInitialization.cpp
@@ -1,6 +1,6 @@
/* -*- mode: C++; c-basic-offset: 4; tab-width: 4 -*-
*
- * Copyright (c) 2004-2005 Apple Computer, Inc. All rights reserved.
+ * Copyright (c) 2004-2008 Apple Inc. All rights reserved.
*
* @APPLE_LICENSE_HEADER_START@
*
@@ -27,45 +27,23 @@
#include <stddef.h>
#include <string.h>
#include <stdlib.h>
+#include <sys/mman.h>
#include <mach/mach.h>
-#include <mach-o/loader.h>
-#include <mach-o/ldsyms.h>
-#include <mach-o/reloc.h>
-#if __ppc__ || __ppc64__
- #include <mach-o/ppc/reloc.h>
-#endif
-#if __x86_64__
- #include <mach-o/x86_64/reloc.h>
-#endif
-#include "dyld.h"
-#ifndef MH_PIE
- #define MH_PIE 0x200000
-#endif
+#include "dyld2.h"
+#include "dyldSyscallInterface.h"
+#include "MachOAnalyzer.h"
+#include "Tracing.h"
+// from libc.a
+extern "C" void mach_init();
+extern "C" void __guard_setup(const char* apple[]);
+extern "C" void _subsystem_init(const char* apple[]);
-#if __LP64__
- #define macho_header mach_header_64
- #define LC_SEGMENT_COMMAND LC_SEGMENT_64
- #define macho_segment_command segment_command_64
- #define macho_section section_64
- #define RELOC_SIZE 3
-#else
- #define macho_header mach_header
- #define LC_SEGMENT_COMMAND LC_SEGMENT
- #define macho_segment_command segment_command
- #define macho_section section
- #define RELOC_SIZE 2
-#endif
+// from dyld_debugger.cpp
+extern void syncProcessInfo();
-#if __x86_64__
- #define POINTER_RELOC X86_64_RELOC_UNSIGNED
-#else
- #define POINTER_RELOC GENERIC_RELOC_VANILLA
-#endif
-
-// from dyld.cpp
-namespace dyld { extern bool isRosetta(); };
+const dyld::SyscallHelpers* gSyscallHelpers = NULL;
//
@@ -76,7 +54,16 @@
namespace dyldbootstrap {
+// currently dyld has no initializers, but if some come back, set this to non-zero
+#define DYLD_INITIALIZER_SUPPORT 0
+
+
+#if DYLD_INITIALIZER_SUPPORT
+
typedef void (*Initializer)(int argc, const char* argv[], const char* envp[], const char* apple[]);
+
+extern const Initializer inits_start __asm("section$start$__DATA$__mod_init_func");
+extern const Initializer inits_end __asm("section$end$__DATA$__mod_init_func");
//
// For a regular executable, the crt code calls dyld to run the executables initializers.
@@ -84,275 +71,59 @@
// dyld (should be static) but is a dynamic executable and needs this hack to run its own initializers.
// We pass argc, argv, etc in case libc.a uses those arguments
//
-static void runDyldInitializers(const struct macho_header* mh, intptr_t slide, int argc, const char* argv[], const char* envp[], const char* apple[])
+static void runDyldInitializers(int argc, const char* argv[], const char* envp[], const char* apple[])
{
- const uint32_t cmd_count = mh->ncmds;
- const struct load_command* const cmds = (struct load_command*)(((char*)mh)+sizeof(macho_header));
- const struct load_command* cmd = cmds;
- for (uint32_t i = 0; i < cmd_count; ++i) {
- switch (cmd->cmd) {
- case LC_SEGMENT_COMMAND:
- {
- const struct macho_segment_command* seg = (struct macho_segment_command*)cmd;
- const struct macho_section* const sectionsStart = (struct macho_section*)((char*)seg + sizeof(struct macho_segment_command));
- const struct macho_section* const sectionsEnd = §ionsStart[seg->nsects];
- for (const struct macho_section* sect=sectionsStart; sect < sectionsEnd; ++sect) {
- const uint8_t type = sect->flags & SECTION_TYPE;
- if ( type == S_MOD_INIT_FUNC_POINTERS ){
- Initializer* inits = (Initializer*)(sect->addr + slide);
- const uint32_t count = sect->size / sizeof(uintptr_t);
- for (uint32_t i=0; i < count; ++i) {
- Initializer func = inits[i];
- func(argc, argv, envp, apple);
- }
- }
- }
- }
- break;
- }
- cmd = (const struct load_command*)(((char*)cmd)+cmd->cmdsize);
+ for (const Initializer* p = &inits_start; p < &inits_end; ++p) {
+ (*p)(argc, argv, envp, apple);
}
}
-
-//
-// If the kernel does not load dyld at its preferred address, we need to apply
-// fixups to various initialized parts of the __DATA segment
-//
-static void rebaseDyld(const struct macho_header* mh, intptr_t slide)
-{
- // rebase non-lazy pointers (which all point internal to dyld, since dyld uses no shared libraries)
- // and get interesting pointers into dyld
- const uint32_t cmd_count = mh->ncmds;
- const struct load_command* const cmds = (struct load_command*)(((char*)mh)+sizeof(macho_header));
- const struct load_command* cmd = cmds;
- const struct macho_segment_command* linkEditSeg = NULL;
-#if __x86_64__
- const struct macho_segment_command* firstWritableSeg = NULL;
-#endif
- const struct dysymtab_command* dynamicSymbolTable = NULL;
- for (uint32_t i = 0; i < cmd_count; ++i) {
- switch (cmd->cmd) {
- case LC_SEGMENT_COMMAND:
- {
- const struct macho_segment_command* seg = (struct macho_segment_command*)cmd;
- if ( strcmp(seg->segname, "__LINKEDIT") == 0 )
- linkEditSeg = seg;
- const struct macho_section* const sectionsStart = (struct macho_section*)((char*)seg + sizeof(struct macho_segment_command));
- const struct macho_section* const sectionsEnd = §ionsStart[seg->nsects];
- for (const struct macho_section* sect=sectionsStart; sect < sectionsEnd; ++sect) {
- const uint8_t type = sect->flags & SECTION_TYPE;
- if ( type == S_NON_LAZY_SYMBOL_POINTERS ) {
- // rebase non-lazy pointers (which all point internal to dyld, since dyld uses no shared libraries)
- const uint32_t pointerCount = sect->size / sizeof(uintptr_t);
- uintptr_t* const symbolPointers = (uintptr_t*)(sect->addr + slide);
- for (uint32_t j=0; j < pointerCount; ++j) {
- symbolPointers[j] += slide;
- }
- }
- }
-#if __x86_64__
- if ( (firstWritableSeg == NULL) && (seg->initprot & VM_PROT_WRITE) )
- firstWritableSeg = seg;
-#endif
- }
- break;
- case LC_DYSYMTAB:
- dynamicSymbolTable = (struct dysymtab_command *)cmd;
- break;
- }
- cmd = (const struct load_command*)(((char*)cmd)+cmd->cmdsize);
- }
-
- // use reloc's to rebase all random data pointers
-#if __x86_64__
- const uintptr_t relocBase = firstWritableSeg->vmaddr + slide;
-#else
- const uintptr_t relocBase = (uintptr_t)mh;
-#endif
- const relocation_info* const relocsStart = (struct relocation_info*)(linkEditSeg->vmaddr + slide + dynamicSymbolTable->locreloff - linkEditSeg->fileoff);
- const relocation_info* const relocsEnd = &relocsStart[dynamicSymbolTable->nlocrel];
- for (const relocation_info* reloc=relocsStart; reloc < relocsEnd; ++reloc) {
- #if __ppc__ || __ppc64__ || __i36__
- if ( (reloc->r_address & R_SCATTERED) != 0 )
- throw "scattered relocation in dyld";
- #endif
- if ( reloc->r_length != RELOC_SIZE )
- throw "relocation in dyld has wrong size";
-
- if ( reloc->r_type != POINTER_RELOC )
- throw "relocation in dyld has wrong type";
-
- // update pointer by amount dyld slid
- *((uintptr_t*)(reloc->r_address + relocBase)) += slide;
- }
-}
+#endif // DYLD_INITIALIZER_SUPPORT
//
-// For some reason the kernel loads dyld with __TEXT and __LINKEDIT writable
-// rdar://problem/3702311
+// On disk, all pointers in dyld's DATA segment are chained together.
+// They need to be fixed up to be real pointers to run.
//
-static void segmentProtectDyld(const struct macho_header* mh, intptr_t slide)
+static void rebaseDyld(const dyld3::MachOLoaded* dyldMH)
{
- const uint32_t cmd_count = mh->ncmds;
- const struct load_command* const cmds = (struct load_command*)(((char*)mh)+sizeof(macho_header));
- const struct load_command* cmd = cmds;
- for (uint32_t i = 0; i < cmd_count; ++i) {
- switch (cmd->cmd) {
- case LC_SEGMENT_COMMAND:
- {
- const struct macho_segment_command* seg = (struct macho_segment_command*)cmd;
- vm_address_t addr = seg->vmaddr + slide;
- vm_size_t size = seg->vmsize;
- const bool setCurrentPermissions = false;
- vm_protect(mach_task_self(), addr, size, setCurrentPermissions, seg->initprot);
- //dyld::log("dyld: segment %s, 0x%08X -> 0x%08X, set to %d\n", seg->segname, addr, addr+size-1, seg->initprot);
- }
- break;
- }
- cmd = (const struct load_command*)(((char*)cmd)+cmd->cmdsize);
- }
-
+ // walk all fixups chains and rebase dyld
+ const dyld3::MachOAnalyzer* ma = (dyld3::MachOAnalyzer*)dyldMH;
+ assert(ma->hasChainedFixups());
+ uintptr_t slide = (long)ma; // all fixup chain based images have a base address of zero, so slide == load address
+ __block Diagnostics diag;
+ ma->withChainStarts(diag, 0, ^(const dyld_chained_starts_in_image* starts) {
+ ma->fixupAllChainedFixups(diag, starts, slide, dyld3::Array<const void*>(), nullptr);
+ });
+ diag.assertNoError();
+
+ // now that rebasing done, initialize mach/syscall layer
+ mach_init();
+
+ // <rdar://47805386> mark __DATA_CONST segment in dyld as read-only (once fixups are done)
+ ma->forEachSegment(^(const dyld3::MachOFile::SegmentInfo& info, bool& stop) {
+ if ( info.readOnlyData ) {
+ ::mprotect(((uint8_t*)(dyldMH))+info.vmAddr, (size_t)info.vmSize, VM_PROT_READ);
+ }
+ });
}
-
-//
-// re-map the main executable to a new random address
-//
-static const struct mach_header* randomizeExecutableLoadAddress(const struct mach_header* orgMH, uintptr_t* appsSlide)
-{
-#if __ppc__
- // don't slide PIE programs running under rosetta
- if ( dyld::isRosetta() )
- return orgMH;
-#endif
- // count segments
- uint32_t segCount = 0;
- const uint32_t cmd_count = orgMH->ncmds;
- const struct load_command* const cmds = (struct load_command*)(((char*)orgMH)+sizeof(macho_header));
- const struct load_command* cmd = cmds;
- for (uint32_t i = 0; i < cmd_count; ++i) {
- if ( cmd->cmd == LC_SEGMENT_COMMAND ) {
- const struct macho_segment_command* segCmd = (struct macho_segment_command*)cmd;
- // page-zero and custom stacks don't move
- if ( (strcmp(segCmd->segname, "__PAGEZERO") != 0) && (strcmp(segCmd->segname, "__UNIXSTACK") != 0) )
- ++segCount;
- }
- cmd = (const struct load_command*)(((char*)cmd)+cmd->cmdsize);
- }
-
- // make copy of segment info
- macho_segment_command segs[segCount];
- uint32_t index = 0;
- uintptr_t highestAddressUsed = 0;
- uintptr_t lowestAddressUsed = UINTPTR_MAX;
- cmd = cmds;
- for (uint32_t i = 0; i < cmd_count; ++i) {
- if ( cmd->cmd == LC_SEGMENT_COMMAND ) {
- const struct macho_segment_command* segCmd = (struct macho_segment_command*)cmd;
- if ( (strcmp(segCmd->segname, "__PAGEZERO") != 0) && (strcmp(segCmd->segname, "__UNIXSTACK") != 0) ) {
- segs[index++] = *segCmd;
- if ( (segCmd->vmaddr + segCmd->vmsize) > highestAddressUsed )
- highestAddressUsed = ((segCmd->vmaddr + segCmd->vmsize) + 4095) & -4096;
- if ( segCmd->vmaddr < lowestAddressUsed )
- lowestAddressUsed = segCmd->vmaddr;
- // do nothing if kernel has already randomized load address
- if ( (strcmp(segCmd->segname, "__TEXT") == 0) && (segCmd->vmaddr != (uintptr_t)orgMH) )
- return orgMH;
- }
- }
- cmd = (const struct load_command*)(((char*)cmd)+cmd->cmdsize);
- }
-
- // choose a random new base address
-#if __LP64__
- uintptr_t highestAddressPossible = highestAddressUsed + 0x100000000ULL;
-#else
- uintptr_t highestAddressPossible = 0x80000000;
-#endif
- uintptr_t sizeNeeded = highestAddressUsed-lowestAddressUsed;
- if ( (highestAddressPossible-sizeNeeded) < highestAddressUsed ) {
- // new and old segments will overlap
- // need better algorithm for remapping
- // punt and don't re-map
- return orgMH;
- }
- uintptr_t possibleRange = (highestAddressPossible-sizeNeeded) - highestAddressUsed;
- uintptr_t newBaseAddress = highestAddressUsed + ((arc4random() % possibleRange) & -4096);
-
- vm_address_t addr = newBaseAddress;
- // reserve new address range
- if ( vm_allocate(mach_task_self(), &addr, sizeNeeded, VM_FLAGS_FIXED) == KERN_SUCCESS ) {
- // copy each segment to new address
- for (uint32_t i = 0; i < segCount; ++i) {
- uintptr_t newSegAddress = segs[i].vmaddr - lowestAddressUsed + newBaseAddress;
- if ( (vm_copy(mach_task_self(), segs[i].vmaddr, segs[i].vmsize, newSegAddress) != KERN_SUCCESS)
- || (vm_protect(mach_task_self(), newSegAddress, segs[i].vmsize, true, segs[i].maxprot) != KERN_SUCCESS)
- || (vm_protect(mach_task_self(), newSegAddress, segs[i].vmsize, false, segs[i].initprot) != KERN_SUCCESS) ) {
- // can't copy so dealloc new region and run with original base address
- vm_deallocate(mach_task_self(), newBaseAddress, sizeNeeded);
- dyld::warn("could not relocate position independent exectable\n");
- return orgMH;
- }
- }
- // unmap original segments
- vm_deallocate(mach_task_self(), lowestAddressUsed, highestAddressUsed-lowestAddressUsed);
-
- // run with newly mapped executable
- *appsSlide = newBaseAddress - lowestAddressUsed;
- return (const struct mach_header*)newBaseAddress;
- }
-
- // can't get new range, so don't slide to random address
- return orgMH;
-}
-
-
-extern "C" void dyld_exceptions_init(const struct macho_header*, uintptr_t slide); // in dyldExceptions.cpp
-extern "C" void mach_init();
-
-//
-// _pthread_keys is partitioned in a lower part that dyld will use; libSystem
-// will use the upper part. We set __pthread_tsd_first to 1 as the start of
-// the lower part. Libc will take #1 and c++ exceptions will take #2. There
-// is one free key=3 left.
-//
-extern "C" {
- extern int __pthread_tsd_first;
- extern void _pthread_keys_init();
-}
//
// This is code to bootstrap dyld. This work in normally done for a program by dyld and crt.
// In dyld we have to do this manually.
//
-uintptr_t start(const struct mach_header* appsMachHeader, int argc, const char* argv[], intptr_t slide)
+uintptr_t start(const dyld3::MachOLoaded* appsMachHeader, int argc, const char* argv[],
+ const dyld3::MachOLoaded* dyldsMachHeader, uintptr_t* startGlue)
{
- // _mh_dylinker_header is magic symbol defined by static linker (ld), see <mach-o/ldsyms.h>
- const struct macho_header* dyldsMachHeader = (const struct macho_header*)(((char*)&_mh_dylinker_header)+slide);
-
+
+ // Emit kdebug tracepoint to indicate dyld bootstrap has started <rdar://46878536>
+ dyld3::kdebug_trace_dyld_marker(DBG_DYLD_TIMING_BOOTSTRAP_START, 0, 0, 0, 0);
+
// if kernel had to slide dyld, we need to fix up load sensitive locations
// we have to do this before using any global variables
- if ( slide != 0 ) {
- rebaseDyld(dyldsMachHeader, slide);
- }
-
- uintptr_t appsSlide = 0;
-
- // set pthread keys to dyld range
- __pthread_tsd_first = 1;
- _pthread_keys_init();
-
- // enable C++ exceptions to work inside dyld
- dyld_exceptions_init(dyldsMachHeader, slide);
-
- // allow dyld to use mach messaging
- mach_init();
+ rebaseDyld(dyldsMachHeader);
- // set protection on segments (has to be done after mach_init)
- segmentProtectDyld(dyldsMachHeader, slide);
-
// kernel sets up env pointer to be just past end of agv array
const char** envp = &argv[argc+1];
@@ -361,18 +132,51 @@
while(*apple != NULL) { ++apple; }
++apple;
+ // set up random value for stack canary
+ __guard_setup(apple);
+
+#if DYLD_INITIALIZER_SUPPORT
// run all C++ initializers inside dyld
- runDyldInitializers(dyldsMachHeader, slide, argc, argv, envp, apple);
-
- // if main executable was linked -pie, then randomize its load address
- if ( appsMachHeader->flags & MH_PIE )
- appsMachHeader = randomizeExecutableLoadAddress(appsMachHeader, &appsSlide);
-
+ runDyldInitializers(argc, argv, envp, apple);
+#endif
+
+ _subsystem_init(apple);
+
// now that we are done bootstrapping dyld, call dyld's main
- return dyld::_main(appsMachHeader, appsSlide, argc, argv, envp, apple);
+ uintptr_t appsSlide = appsMachHeader->getSlide();
+ return dyld::_main((macho_header*)appsMachHeader, appsSlide, argc, argv, envp, apple, startGlue);
}
+#if TARGET_OS_SIMULATOR
+
+extern "C" uintptr_t start_sim(int argc, const char* argv[], const char* envp[], const char* apple[],
+ const dyld3::MachOLoaded* mainExecutableMH, const dyld3::MachOLoaded* dyldMH, uintptr_t dyldSlide,
+ const dyld::SyscallHelpers*, uintptr_t* startGlue);
+
+
+uintptr_t start_sim(int argc, const char* argv[], const char* envp[], const char* apple[],
+ const dyld3::MachOLoaded* mainExecutableMH, const dyld3::MachOLoaded* dyldSimMH, uintptr_t dyldSlide,
+ const dyld::SyscallHelpers* sc, uintptr_t* startGlue)
+{
+ // save table of syscall pointers
+ gSyscallHelpers = sc;
+
+ // dyld_sim uses chained rebases, so it always need to be fixed up
+ rebaseDyld(dyldSimMH);
+
+ // set up random value for stack canary
+ __guard_setup(apple);
+
+ // setup gProcessInfo to point to host dyld's struct
+ dyld::gProcessInfo = (struct dyld_all_image_infos*)(sc->getProcessInfo());
+ syncProcessInfo();
+
+ // now that we are done bootstrapping dyld, call dyld's main
+ uintptr_t appsSlide = mainExecutableMH->getSlide();
+ return dyld::_main((macho_header*)mainExecutableMH, appsSlide, argc, argv, envp, apple, startGlue);
+}
+#endif
} // end of namespace