[apple/ld64.git] / src / other / rebase.cpp

/* -*- mode: C++; c-basic-offset: 4; tab-width: 4 -*- 
 *
 * Copyright (c) 2006-2008 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 <sys/types.h>
#include <sys/stat.h>
#include <sys/mman.h>
#include <mach/mach.h>
#include <limits.h>
#include <stdarg.h>
#include <stdio.h>
#include <fcntl.h>
#include <errno.h>
#include <unistd.h>
#include <vector>
#include <set>


#include "MachOFileAbstraction.hpp"
#include "Architectures.hpp"

static bool verbose = false;

__attribute__((noreturn))
void throwf(const char* format, ...) 
{
	va_list	list;
	char*	p;
	va_start(list, format);
	vasprintf(&p, format, list);
	va_end(list);
	
	const char*	t = p;
	throw t;
}


class AbstractRebaser
{
public:
	virtual cpu_type_t							getArchitecture() const = 0;
	virtual uint64_t							getBaseAddress() const = 0;
	virtual uint64_t							getVMSize() const = 0;
	virtual void								setBaseAddress(uint64_t) = 0;
};


template <typename A>
class Rebaser : public AbstractRebaser
{
public:
												Rebaser(const void* machHeader);
	virtual										~Rebaser() {}

	virtual cpu_type_t							getArchitecture() const;
	virtual uint64_t							getBaseAddress() const;
	virtual uint64_t							getVMSize() const;
	virtual void								setBaseAddress(uint64_t);

private:
	typedef typename A::P					P;
	typedef typename A::P::E				E;
	typedef typename A::P::uint_t			pint_t;
	
	struct vmmap { pint_t vmaddr; pint_t vmsize; pint_t fileoff; };
	
	void										setRelocBase();
	void										buildSectionTable();
	void										adjustLoadCommands();
	void										adjustSymbolTable();
	void										adjustDATA();
	void										doLocalRelocation(const macho_relocation_info<P>* reloc);
	pint_t*										mappedAddressForVMAddress(uint32_t vmaddress);
	void										rebaseAt(int segIndex, uint64_t offset, uint8_t type);
	
	const macho_header<P>*						fHeader;
	pint_t										fOrignalVMRelocBaseAddress;
	pint_t										fSlide;
	std::vector<vmmap>							fVMMApping;
};


class MultiArchRebaser
{
public:
												MultiArchRebaser(const char* path, bool writable=false);
												~MultiArchRebaser();

	const std::vector<AbstractRebaser*>&		getArchs() const { return fRebasers; }
	void										commit();

private:
	std::vector<AbstractRebaser*>				fRebasers;
	void*										fMappingAddress;
	uint64_t									fFileSize;
};


MultiArchRebaser::MultiArchRebaser(const char* path, bool writable)
 : fMappingAddress(0), fFileSize(0)
{
	// map in whole file
	int fd = ::open(path, (writable ? O_RDWR : O_RDONLY), 0);
	if ( fd == -1 )
		throwf("can't open file %s, errno=%d", path, errno);
	struct stat stat_buf;
	if ( fstat(fd, &stat_buf) == -1)
		throwf("can't stat open file %s, errno=%d", path, errno);
	if ( stat_buf.st_size < 20 )
		throwf("file too small %s", path);
	const int prot = writable ? (PROT_READ | PROT_WRITE) : PROT_READ;
	const int flags = writable ? (MAP_FILE | MAP_SHARED) : (MAP_FILE | MAP_PRIVATE);
	uint8_t* p = (uint8_t*)::mmap(NULL, stat_buf.st_size, prot, flags, fd, 0);
	if ( p == (uint8_t*)(-1) )
		throwf("can't map file %s, errno=%d", path, errno);
	::close(fd);

	// if fat file, process each architecture
	const fat_header* fh = (fat_header*)p;
	const mach_header* mh = (mach_header*)p;
	if ( fh->magic == OSSwapBigToHostInt32(FAT_MAGIC) ) {
		// Fat header is always big-endian
		const struct fat_arch* archs = (struct fat_arch*)(p + sizeof(struct fat_header));
		for (unsigned long i=0; i < OSSwapBigToHostInt32(fh->nfat_arch); ++i) {
			uint32_t fileOffset = OSSwapBigToHostInt32(archs[i].offset);
			try {
				switch ( OSSwapBigToHostInt32(archs[i].cputype) ) {
					case CPU_TYPE_POWERPC:
						fRebasers.push_back(new Rebaser<ppc>(&p[fileOffset]));
						break;
					case CPU_TYPE_POWERPC64:
						fRebasers.push_back(new Rebaser<ppc64>(&p[fileOffset]));
						break;
					case CPU_TYPE_I386:
						fRebasers.push_back(new Rebaser<x86>(&p[fileOffset]));
						break;
					case CPU_TYPE_X86_64:
						fRebasers.push_back(new Rebaser<x86_64>(&p[fileOffset]));
						break;
					case CPU_TYPE_ARM:
						fRebasers.push_back(new Rebaser<arm>(&p[fileOffset]));
						break;
					default:
						throw "unknown file format";
				}
			}
			catch (const char* msg) {
				fprintf(stderr, "rebase warning: %s for %s\n", msg, path);
			}
		}
	}
	else {
		try {
			if ( (OSSwapBigToHostInt32(mh->magic) == MH_MAGIC) && (OSSwapBigToHostInt32(mh->cputype) == CPU_TYPE_POWERPC)) {
				fRebasers.push_back(new Rebaser<ppc>(mh));
			}
			else if ( (OSSwapBigToHostInt32(mh->magic) == MH_MAGIC_64) && (OSSwapBigToHostInt32(mh->cputype) == CPU_TYPE_POWERPC64)) {
				fRebasers.push_back(new Rebaser<ppc64>(mh));
			}
			else if ( (OSSwapLittleToHostInt32(mh->magic) == MH_MAGIC) && (OSSwapLittleToHostInt32(mh->cputype) == CPU_TYPE_I386)) {
				fRebasers.push_back(new Rebaser<x86>(mh));
			}
			else if ( (OSSwapLittleToHostInt32(mh->magic) == MH_MAGIC_64) && (OSSwapLittleToHostInt32(mh->cputype) == CPU_TYPE_X86_64)) {
				fRebasers.push_back(new Rebaser<x86_64>(mh));
			}
			else if ( (OSSwapLittleToHostInt32(mh->magic) == MH_MAGIC) && (OSSwapLittleToHostInt32(mh->cputype) == CPU_TYPE_ARM)) {
				fRebasers.push_back(new Rebaser<arm>(mh));
			}
			else {
				throw "unknown file format";
			}
		}
		catch (const char* msg) {
			fprintf(stderr, "rebase warning: %s for %s\n", msg, path);
		}
	}
	
	fMappingAddress = p;
	fFileSize = stat_buf.st_size;
}


MultiArchRebaser::~MultiArchRebaser()
{
	::munmap(fMappingAddress, fFileSize);
}

void MultiArchRebaser::commit()
{
	::msync(fMappingAddress, fFileSize, MS_ASYNC); 
}


template <typename A>
Rebaser<A>::Rebaser(const void* machHeader)
 : 	fHeader((const macho_header<P>*)machHeader)
{
	switch ( fHeader->filetype() ) {
		case MH_DYLIB:
			if ( (fHeader->flags() & MH_SPLIT_SEGS) != 0 )
				throw "split-seg dylibs cannot be rebased";
			break;
		case MH_BUNDLE:
			break;
		default:
			throw "file is not a dylib or bundle";
	}
		
}

template <> cpu_type_t Rebaser<ppc>::getArchitecture()    const { return CPU_TYPE_POWERPC; }
template <> cpu_type_t Rebaser<ppc64>::getArchitecture()  const { return CPU_TYPE_POWERPC64; }
template <> cpu_type_t Rebaser<x86>::getArchitecture()    const { return CPU_TYPE_I386; }
template <> cpu_type_t Rebaser<x86_64>::getArchitecture() const { return CPU_TYPE_X86_64; }
template <> cpu_type_t Rebaser<arm>::getArchitecture() const { return CPU_TYPE_ARM; }

template <typename A>
uint64_t Rebaser<A>::getBaseAddress() const
{
	uint64_t lowestSegmentAddress = LLONG_MAX;
	const macho_load_command<P>* const cmds = (macho_load_command<P>*)((uint8_t*)fHeader + sizeof(macho_header<P>));
	const uint32_t cmd_count = fHeader->ncmds();
	const macho_load_command<P>* cmd = cmds;
	for (uint32_t i = 0; i < cmd_count; ++i) {
		if ( cmd->cmd() == macho_segment_command<P>::CMD ) {
			const macho_segment_command<P>* segCmd = (const macho_segment_command<P>*)cmd;
			if ( segCmd->vmaddr() < lowestSegmentAddress ) {
				lowestSegmentAddress = segCmd->vmaddr();
			}
		}
		cmd = (const macho_load_command<P>*)(((uint8_t*)cmd)+cmd->cmdsize());
	}
	return lowestSegmentAddress;
}

template <typename A>
uint64_t Rebaser<A>::getVMSize() const
{
	const macho_segment_command<P>* highestSegmentCmd = NULL; 
	const macho_load_command<P>* const cmds = (macho_load_command<P>*)((uint8_t*)fHeader + sizeof(macho_header<P>));
	const uint32_t cmd_count = fHeader->ncmds();
	const macho_load_command<P>* cmd = cmds;
	for (uint32_t i = 0; i < cmd_count; ++i) {
		if ( cmd->cmd() == macho_segment_command<P>::CMD ) {
			const macho_segment_command<P>* segCmd = (const macho_segment_command<P>*)cmd;
			if ( (highestSegmentCmd == NULL) || (segCmd->vmaddr() > highestSegmentCmd->vmaddr()) ) {
				highestSegmentCmd = segCmd;
			}
		}
		cmd = (const macho_load_command<P>*)(((uint8_t*)cmd)+cmd->cmdsize());
	}
	
	return ((highestSegmentCmd->vmaddr() + highestSegmentCmd->vmsize() - this->getBaseAddress() + 4095) & (-4096));
}


template <typename A>
void Rebaser<A>::setBaseAddress(uint64_t addr)
{
	// calculate slide
	fSlide = addr - this->getBaseAddress();
	
	// compute base address for relocations
	this->setRelocBase();
	
	// build cache of section index to section 
	this->buildSectionTable();
		
	// update load commands
	this->adjustLoadCommands();
	
	// update symbol table  
	this->adjustSymbolTable();
	
	// update writable segments that have internal pointers
	this->adjustDATA();
}

template <typename A>
void Rebaser<A>::adjustLoadCommands()
{
	const macho_load_command<P>* const cmds = (macho_load_command<P>*)((uint8_t*)fHeader + sizeof(macho_header<P>));
	const uint32_t cmd_count = fHeader->ncmds();
	const macho_load_command<P>* cmd = cmds;
	for (uint32_t i = 0; i < cmd_count; ++i) {
		switch ( cmd->cmd() ) {
			case LC_ID_DYLIB:
				if ( (fHeader->flags() & MH_PREBOUND) != 0 ) {
					// clear timestamp so that any prebound clients are invalidated
					macho_dylib_command<P>* dylib  = (macho_dylib_command<P>*)cmd;
					dylib->set_timestamp(1);
				}
				break;
			case LC_LOAD_DYLIB:
			case LC_LOAD_WEAK_DYLIB:
			case LC_REEXPORT_DYLIB:
			case LC_LOAD_UPWARD_DYLIB:
				if ( (fHeader->flags() & MH_PREBOUND) != 0 ) {
					// clear expected timestamps so that this image will load with invalid prebinding 
					macho_dylib_command<P>* dylib  = (macho_dylib_command<P>*)cmd;
					dylib->set_timestamp(2);
				}
				break;
			case macho_routines_command<P>::CMD:
				// update -init command
				{
					macho_routines_command<P>* routines = (macho_routines_command<P>*)cmd;
					routines->set_init_address(routines->init_address() + fSlide);
				}
				break;
			case macho_segment_command<P>::CMD:
				// update segment commands
				{
					macho_segment_command<P>* seg = (macho_segment_command<P>*)cmd;
					seg->set_vmaddr(seg->vmaddr() + fSlide);
					macho_section<P>* const sectionsStart = (macho_section<P>*)((char*)seg + sizeof(macho_segment_command<P>));
					macho_section<P>* const sectionsEnd = &sectionsStart[seg->nsects()];
					for(macho_section<P>* sect = sectionsStart; sect < sectionsEnd; ++sect) {
						sect->set_addr(sect->addr() + fSlide);
					}
				}
				break;
		}
		cmd = (const macho_load_command<P>*)(((uint8_t*)cmd)+cmd->cmdsize());
	}
}


template <typename A>
void Rebaser<A>::buildSectionTable()
{
	// build vector of sections
	const macho_load_command<P>* const cmds = (macho_load_command<P>*)((uint8_t*)fHeader + sizeof(macho_header<P>));
	const uint32_t cmd_count = fHeader->ncmds();
	const macho_load_command<P>* cmd = cmds;
	for (uint32_t i = 0; i < cmd_count; ++i) {
		if ( cmd->cmd() == macho_segment_command<P>::CMD ) {
			const macho_segment_command<P>* seg = (macho_segment_command<P>*)cmd;
			vmmap mapping;
			mapping.vmaddr = seg->vmaddr();
			mapping.vmsize = seg->vmsize();
			mapping.fileoff = seg->fileoff();
			fVMMApping.push_back(mapping);
		}
		cmd = (const macho_load_command<P>*)(((uint8_t*)cmd)+cmd->cmdsize());
	}	
}


template <typename A>
void Rebaser<A>::adjustSymbolTable()
{
	const macho_dysymtab_command<P>* dysymtab = NULL;
	macho_nlist<P>* symbolTable = NULL;
	const char* strings = NULL;

	// get symbol table info
	const macho_load_command<P>* const cmds = (macho_load_command<P>*)((uint8_t*)fHeader + sizeof(macho_header<P>));
	const uint32_t cmd_count = fHeader->ncmds();
	const macho_load_command<P>* cmd = cmds;
	for (uint32_t i = 0; i < cmd_count; ++i) {
		switch (cmd->cmd()) {
			case LC_SYMTAB:
				{
					const macho_symtab_command<P>* symtab = (macho_symtab_command<P>*)cmd;
					symbolTable = (macho_nlist<P>*)(((uint8_t*)fHeader) + symtab->symoff());
					strings = (char*)(((uint8_t*)fHeader) + symtab->stroff());
                }    
				break;
			case LC_DYSYMTAB:
				dysymtab = (macho_dysymtab_command<P>*)cmd;
				break;
		}
		cmd = (const macho_load_command<P>*)(((uint8_t*)cmd)+cmd->cmdsize());
	}	

	// walk all exports and slide their n_value
	macho_nlist<P>* lastExport = &symbolTable[dysymtab->iextdefsym()+dysymtab->nextdefsym()];
	for (macho_nlist<P>* entry = &symbolTable[dysymtab->iextdefsym()]; entry < lastExport; ++entry) {
		if ( (entry->n_type() & N_TYPE) == N_SECT )
			entry->set_n_value(entry->n_value() + fSlide);
	}

	// walk all local symbols and slide their n_value
	macho_nlist<P>* lastLocal = &symbolTable[dysymtab->ilocalsym()+dysymtab->nlocalsym()];
	for (macho_nlist<P>* entry = &symbolTable[dysymtab->ilocalsym()]; entry < lastLocal; ++entry) {
		if ( ((entry->n_type() & N_STAB) == 0) && ((entry->n_type() & N_TYPE) == N_SECT) ) {
			entry->set_n_value(entry->n_value() + fSlide);
		}
		else if ( entry->n_type() & N_STAB ) {
			// some stabs need to be slid too
			switch ( entry->n_type() ) {
				case N_FUN:
					// don't slide end-of-function FUN which is FUN with no string
					if ( (entry->n_strx() == 0) || (strings[entry->n_strx()] == '\0') )
						break;
				case N_BNSYM:
				case N_STSYM:
				case N_LCSYM:
					entry->set_n_value(entry->n_value() + fSlide);
					break;
			}
		}
	}
Commit	Line	Data
69a49097 A	1	/* -- mode: C++; c-basic-offset: 4; tab-width: 4 --
69a49097 A	2	*
55e3d2f6	3	* Copyright (c) 2006-2008 Apple Inc. All rights reserved.
69a49097 A	4	*
	5	* @APPLE_LICENSE_HEADER_START@
	6	*
	7	* This file contains Original Code and/or Modifications of Original Code
	8	* as defined in and that are subject to the Apple Public Source License
	9	* Version 2.0 (the 'License'). You may not use this file except in
	10	* compliance with the License. Please obtain a copy of the License at
	11	* http://www.opensource.apple.com/apsl/ and read it before using this
	12	* file.
	13	*
	14	* The Original Code and all software distributed under the License are
	15	* distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
	16	* EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
	17	* INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
	18	* FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
	19	* Please see the License for the specific language governing rights and
	20	* limitations under the License.
	21	*
	22	* @APPLE_LICENSE_HEADER_END@
	23	*/
	24
	25	#include <sys/types.h>
	26	#include <sys/stat.h>
	27	#include <sys/mman.h>
	28	#include <mach/mach.h>
	29	#include <limits.h>
	30	#include <stdarg.h>
	31	#include <stdio.h>
	32	#include <fcntl.h>
	33	#include <errno.h>
	34	#include <unistd.h>
69a49097 A	35	#include <vector>
	36	#include <set>
	37
2f2f92e4	38
69a49097 A	39	#include "MachOFileAbstraction.hpp"
	40	#include "Architectures.hpp"
	41
69a49097 A	42	static bool verbose = false;
	43
	44	__attribute__((noreturn))
	45	void throwf(const char* format, ...)
	46	{
	47	va_list list;
	48	char* p;
	49	va_start(list, format);
	50	vasprintf(&p, format, list);
	51	va_end(list);
	52
	53	const char* t = p;
	54	throw t;
	55	}
	56
	57
	58	class AbstractRebaser
	59	{
	60	public:
	61	virtual cpu_type_t getArchitecture() const = 0;
	62	virtual uint64_t getBaseAddress() const = 0;
	63	virtual uint64_t getVMSize() const = 0;
	64	virtual void setBaseAddress(uint64_t) = 0;
	65	};
	66
	67
	68	template <typename A>
	69	class Rebaser : public AbstractRebaser
	70	{
	71	public:
	72	Rebaser(const void* machHeader);
	73	virtual ~Rebaser() {}
	74
	75	virtual cpu_type_t getArchitecture() const;
	76	virtual uint64_t getBaseAddress() const;
	77	virtual uint64_t getVMSize() const;
	78	virtual void setBaseAddress(uint64_t);
	79
	80	private:
	81	typedef typename A::P P;
	82	typedef typename A::P::E E;
	83	typedef typename A::P::uint_t pint_t;
	84
	85	struct vmmap { pint_t vmaddr; pint_t vmsize; pint_t fileoff; };
	86
	87	void setRelocBase();
	88	void buildSectionTable();
	89	void adjustLoadCommands();
	90	void adjustSymbolTable();
	91	void adjustDATA();
	92	void doLocalRelocation(const macho_relocation_info<P>* reloc);
	93	pint_t* mappedAddressForVMAddress(uint32_t vmaddress);
55e3d2f6	94	void rebaseAt(int segIndex, uint64_t offset, uint8_t type);
69a49097 A	95
	96	const macho_header<P>* fHeader;
	97	pint_t fOrignalVMRelocBaseAddress;
	98	pint_t fSlide;
69a49097 A	99	std::vector<vmmap> fVMMApping;
	100	};
	101
	102
	103
	104	class MultiArchRebaser
	105	{
	106	public:
	107	MultiArchRebaser(const char* path, bool writable=false);
	108	~MultiArchRebaser();
	109
	110	const std::vector<AbstractRebaser*>& getArchs() const { return fRebasers; }
	111	void commit();
	112
	113	private:
	114	std::vector<AbstractRebaser*> fRebasers;
	115	void* fMappingAddress;
	116	uint64_t fFileSize;
	117	};
	118
	119
	120
	121	MultiArchRebaser::MultiArchRebaser(const char* path, bool writable)
	122	: fMappingAddress(0), fFileSize(0)
	123	{
	124	// map in whole file
	125	int fd = ::open(path, (writable ? O_RDWR : O_RDONLY), 0);
	126	if ( fd == -1 )
55e3d2f6	127	throwf("can't open file %s, errno=%d", path, errno);
69a49097 A	128	struct stat stat_buf;
	129	if ( fstat(fd, &stat_buf) == -1)
	130	throwf("can't stat open file %s, errno=%d", path, errno);
	131	if ( stat_buf.st_size < 20 )
	132	throwf("file too small %s", path);
	133	const int prot = writable ? (PROT_READ \| PROT_WRITE) : PROT_READ;
	134	const int flags = writable ? (MAP_FILE \| MAP_SHARED) : (MAP_FILE \| MAP_PRIVATE);
	135	uint8_t* p = (uint8_t*)::mmap(NULL, stat_buf.st_size, prot, flags, fd, 0);
	136	if ( p == (uint8_t*)(-1) )
	137	throwf("can't map file %s, errno=%d", path, errno);
	138	::close(fd);
	139
	140	// if fat file, process each architecture
	141	const fat_header* fh = (fat_header*)p;
	142	const mach_header* mh = (mach_header*)p;
	143	if ( fh->magic == OSSwapBigToHostInt32(FAT_MAGIC) ) {
	144	// Fat header is always big-endian
	145	const struct fat_arch* archs = (struct fat_arch*)(p + sizeof(struct fat_header));
	146	for (unsigned long i=0; i < OSSwapBigToHostInt32(fh->nfat_arch); ++i) {
	147	uint32_t fileOffset = OSSwapBigToHostInt32(archs[i].offset);
	148	try {
	149	switch ( OSSwapBigToHostInt32(archs[i].cputype) ) {
	150	case CPU_TYPE_POWERPC:
	151	fRebasers.push_back(new Rebaser<ppc>(&p[fileOffset]));
	152	break;
	153	case CPU_TYPE_POWERPC64:
	154	fRebasers.push_back(new Rebaser<ppc64>(&p[fileOffset]));
	155	break;
	156	case CPU_TYPE_I386:
	157	fRebasers.push_back(new Rebaser<x86>(&p[fileOffset]));
	158	break;
	159	case CPU_TYPE_X86_64:
	160	fRebasers.push_back(new Rebaser<x86_64>(&p[fileOffset]));
	161	break;
2f2f92e4 A	162	case CPU_TYPE_ARM:
	163	fRebasers.push_back(new Rebaser<arm>(&p[fileOffset]));
	164	break;
69a49097 A	165	default:
	166	throw "unknown file format";
	167	}
	168	}
	169	catch (const char* msg) {
	170	fprintf(stderr, "rebase warning: %s for %s\n", msg, path);
	171	}
	172	}
	173	}
	174	else {
	175	try {
	176	if ( (OSSwapBigToHostInt32(mh->magic) == MH_MAGIC) && (OSSwapBigToHostInt32(mh->cputype) == CPU_TYPE_POWERPC)) {
	177	fRebasers.push_back(new Rebaser<ppc>(mh));
	178	}
	179	else if ( (OSSwapBigToHostInt32(mh->magic) == MH_MAGIC_64) && (OSSwapBigToHostInt32(mh->cputype) == CPU_TYPE_POWERPC64)) {
	180	fRebasers.push_back(new Rebaser<ppc64>(mh));
	181	}
	182	else if ( (OSSwapLittleToHostInt32(mh->magic) == MH_MAGIC) && (OSSwapLittleToHostInt32(mh->cputype) == CPU_TYPE_I386)) {
	183	fRebasers.push_back(new Rebaser<x86>(mh));
	184	}
	185	else if ( (OSSwapLittleToHostInt32(mh->magic) == MH_MAGIC_64) && (OSSwapLittleToHostInt32(mh->cputype) == CPU_TYPE_X86_64)) {
	186	fRebasers.push_back(new Rebaser<x86_64>(mh));
	187	}
2f2f92e4 A	188	else if ( (OSSwapLittleToHostInt32(mh->magic) == MH_MAGIC) && (OSSwapLittleToHostInt32(mh->cputype) == CPU_TYPE_ARM)) {
	189	fRebasers.push_back(new Rebaser<arm>(mh));
	190	}
69a49097 A	191	else {
	192	throw "unknown file format";
	193	}
	194	}
	195	catch (const char* msg) {
	196	fprintf(stderr, "rebase warning: %s for %s\n", msg, path);
	197	}
	198	}
	199
	200	fMappingAddress = p;
	201	fFileSize = stat_buf.st_size;
	202	}
	203
	204
	205	MultiArchRebaser::~MultiArchRebaser()
	206	{
	207	::munmap(fMappingAddress, fFileSize);
	208	}
	209
	210	void MultiArchRebaser::commit()
	211	{
	212	::msync(fMappingAddress, fFileSize, MS_ASYNC);
	213	}
	214
	215
	216
	217	template <typename A>
	218	Rebaser<A>::Rebaser(const void* machHeader)
	219	: fHeader((const macho_header<P>*)machHeader)
	220	{
	221	switch ( fHeader->filetype() ) {
	222	case MH_DYLIB:
	223	if ( (fHeader->flags() & MH_SPLIT_SEGS) != 0 )
	224	throw "split-seg dylibs cannot be rebased";
	225	break;
	226	case MH_BUNDLE:
	227	break;
	228	default:
	229	throw "file is not a dylib or bundle";
	230	}
	231
	232	}
	233
	234	template <> cpu_type_t Rebaser<ppc>::getArchitecture() const { return CPU_TYPE_POWERPC; }
	235	template <> cpu_type_t Rebaser<ppc64>::getArchitecture() const { return CPU_TYPE_POWERPC64; }
	236	template <> cpu_type_t Rebaser<x86>::getArchitecture() const { return CPU_TYPE_I386; }
	237	template <> cpu_type_t Rebaser<x86_64>::getArchitecture() const { return CPU_TYPE_X86_64; }
2f2f92e4	238	template <> cpu_type_t Rebaser<arm>::getArchitecture() const { return CPU_TYPE_ARM; }
69a49097 A	239
	240	template <typename A>
	241	uint64_t Rebaser<A>::getBaseAddress() const
	242	{
	243	uint64_t lowestSegmentAddress = LLONG_MAX;
	244	const macho_load_command<P>* const cmds = (macho_load_command<P>)((uint8_t)fHeader + sizeof(macho_header<P>));
	245	const uint32_t cmd_count = fHeader->ncmds();
	246	const macho_load_command<P>* cmd = cmds;
	247	for (uint32_t i = 0; i < cmd_count; ++i) {
	248	if ( cmd->cmd() == macho_segment_command<P>::CMD ) {
	249	const macho_segment_command<P>* segCmd = (const macho_segment_command<P>*)cmd;
	250	if ( segCmd->vmaddr() < lowestSegmentAddress ) {
	251	lowestSegmentAddress = segCmd->vmaddr();
	252	}
	253	}
	254	cmd = (const macho_load_command<P>)(((uint8_t)cmd)+cmd->cmdsize());
	255	}
	256	return lowestSegmentAddress;
	257	}
	258
	259	template <typename A>
	260	uint64_t Rebaser<A>::getVMSize() const
	261	{
	262	const macho_segment_command<P>* highestSegmentCmd = NULL;
	263	const macho_load_command<P>* const cmds = (macho_load_command<P>)((uint8_t)fHeader + sizeof(macho_header<P>));
	264	const uint32_t cmd_count = fHeader->ncmds();
	265	const macho_load_command<P>* cmd = cmds;
	266	for (uint32_t i = 0; i < cmd_count; ++i) {
	267	if ( cmd->cmd() == macho_segment_command<P>::CMD ) {
	268	const macho_segment_command<P>* segCmd = (const macho_segment_command<P>*)cmd;
	269	if ( (highestSegmentCmd == NULL) \|\| (segCmd->vmaddr() > highestSegmentCmd->vmaddr()) ) {
	270	highestSegmentCmd = segCmd;
	271	}
	272	}
	273	cmd = (const macho_load_command<P>)(((uint8_t)cmd)+cmd->cmdsize());
	274	}
	275
	276	return ((highestSegmentCmd->vmaddr() + highestSegmentCmd->vmsize() - this->getBaseAddress() + 4095) & (-4096));
	277	}
	278
	279
	280	template <typename A>
	281	void Rebaser<A>::setBaseAddress(uint64_t addr)
	282	{
	283	// calculate slide
	284	fSlide = addr - this->getBaseAddress();
	285
	286	// compute base address for relocations
	287	this->setRelocBase();
	288
	289	// build cache of section index to section
	290	this->buildSectionTable();
	291
	292	// update load commands
	293	this->adjustLoadCommands();
	294
	295	// update symbol table
	296	this->adjustSymbolTable();
	297
	298	// update writable segments that have internal pointers
	299	this->adjustDATA();
	300	}
	301
	302	template <typename A>
303	void Rebaser<A>::adjustLoadCommands()
304	{
69a49097 A	305	const macho_load_command<P>* const cmds = (macho_load_command<P>)((uint8_t)fHeader + sizeof(macho_header<P>));
	306	const uint32_t cmd_count = fHeader->ncmds();
	307	const macho_load_command<P>* cmd = cmds;
	308	for (uint32_t i = 0; i < cmd_count; ++i) {
	309	switch ( cmd->cmd() ) {
	310	case LC_ID_DYLIB:
	311	if ( (fHeader->flags() & MH_PREBOUND) != 0 ) {
	312	// clear timestamp so that any prebound clients are invalidated
	313	macho_dylib_command<P>* dylib = (macho_dylib_command<P>*)cmd;
	314	dylib->set_timestamp(1);
	315	}
	316	break;
	317	case LC_LOAD_DYLIB:
	318	case LC_LOAD_WEAK_DYLIB:
a645023d A	319	case LC_REEXPORT_DYLIB:
a645023d A	320	case LC_LOAD_UPWARD_DYLIB:
69a49097 A	321	if ( (fHeader->flags() & MH_PREBOUND) != 0 ) {
	322	// clear expected timestamps so that this image will load with invalid prebinding
	323	macho_dylib_command<P>* dylib = (macho_dylib_command<P>*)cmd;
	324	dylib->set_timestamp(2);
	325	}
	326	break;
	327	case macho_routines_command<P>::CMD:
	328	// update -init command
	329	{
a645023d	330	macho_routines_command<P>* routines = (macho_routines_command<P>*)cmd;
69a49097 A	331	routines->set_init_address(routines->init_address() + fSlide);
	332	}
	333	break;
	334	case macho_segment_command<P>::CMD:
	335	// update segment commands
	336	{
	337	macho_segment_command<P>* seg = (macho_segment_command<P>*)cmd;
	338	seg->set_vmaddr(seg->vmaddr() + fSlide);
	339	macho_section<P>* const sectionsStart = (macho_section<P>)((char)seg + sizeof(macho_segment_command<P>));
	340	macho_section<P>* const sectionsEnd = &sectionsStart[seg->nsects()];
	341	for(macho_section<P>* sect = sectionsStart; sect < sectionsEnd; ++sect) {
	342	sect->set_addr(sect->addr() + fSlide);
	343	}
	344	}
	345	break;
	346	}
	347	cmd = (const macho_load_command<P>)(((uint8_t)cmd)+cmd->cmdsize());
	348	}
	349	}
	350
	351
	352	template <typename A>
	353	void Rebaser<A>::buildSectionTable()
	354	{
	355	// build vector of sections
	356	const macho_load_command<P>* const cmds = (macho_load_command<P>)((uint8_t)fHeader + sizeof(macho_header<P>));
	357	const uint32_t cmd_count = fHeader->ncmds();
	358	const macho_load_command<P>* cmd = cmds;
	359	for (uint32_t i = 0; i < cmd_count; ++i) {
	360	if ( cmd->cmd() == macho_segment_command<P>::CMD ) {
	361	const macho_segment_command<P>* seg = (macho_segment_command<P>*)cmd;
	362	vmmap mapping;
	363	mapping.vmaddr = seg->vmaddr();
	364	mapping.vmsize = seg->vmsize();
	365	mapping.fileoff = seg->fileoff();
	366	fVMMApping.push_back(mapping);
	367	}
	368	cmd = (const macho_load_command<P>)(((uint8_t)cmd)+cmd->cmdsize());
	369	}
	370	}
	371
	372
	373	template <typename A>
	374	void Rebaser<A>::adjustSymbolTable()
	375	{
	376	const macho_dysymtab_command<P>* dysymtab = NULL;
	377	macho_nlist<P>* symbolTable = NULL;
a645023d	378	const char* strings = NULL;
69a49097 A	379
	380	// get symbol table info
	381	const macho_load_command<P>* const cmds = (macho_load_command<P>)((uint8_t)fHeader + sizeof(macho_header<P>));
	382	const uint32_t cmd_count = fHeader->ncmds();
	383	const macho_load_command<P>* cmd = cmds;
	384	for (uint32_t i = 0; i < cmd_count; ++i) {
	385	switch (cmd->cmd()) {
	386	case LC_SYMTAB:
	387	{
	388	const macho_symtab_command<P>* symtab = (macho_symtab_command<P>*)cmd;
	389	symbolTable = (macho_nlist<P>)(((uint8_t)fHeader) + symtab->symoff());
a645023d A	390	strings = (char)(((uint8_t)fHeader) + symtab->stroff());
a645023d A	391	}
69a49097 A	392	break;
	393	case LC_DYSYMTAB:
	394	dysymtab = (macho_dysymtab_command<P>*)cmd;
	395	break;
	396	}
	397	cmd = (const macho_load_command<P>)(((uint8_t)cmd)+cmd->cmdsize());
	398	}
	399
	400	// walk all exports and slide their n_value
	401	macho_nlist<P>* lastExport = &symbolTable[dysymtab->iextdefsym()+dysymtab->nextdefsym()];
	402	for (macho_nlist<P>* entry = &symbolTable[dysymtab->iextdefsym()]; entry < lastExport; ++entry) {
	403	if ( (entry->n_type() & N_TYPE) == N_SECT )
	404	entry->set_n_value(entry->n_value() + fSlide);
	405	}
	406
	407	// walk all local symbols and slide their n_value
	408	macho_nlist<P>* lastLocal = &symbolTable[dysymtab->ilocalsym()+dysymtab->nlocalsym()];
	409	for (macho_nlist<P>* entry = &symbolTable[dysymtab->ilocalsym()]; entry < lastLocal; ++entry) {
a645023d	410	if ( ((entry->n_type() & N_STAB) == 0) && ((entry->n_type() & N_TYPE) == N_SECT) ) {
69a49097	411	entry->set_n_value(entry->n_value() + fSlide);
a645023d A	412	}
	413	else if ( entry->n_type() & N_STAB ) {
	414	// some stabs need to be slid too
	415	switch ( entry->n_type() ) {
	416	case N_FUN:
	417	// don't slide end-of-function FUN which is FUN with no string
	418	if ( (entry->n_strx() == 0) \|\| (strings[entry->n_strx()] == '\0') )
	419	break;
	420	case N_BNSYM:
	421	case N_STSYM:
	422	case N_LCSYM:
	423	entry->set_n_value(entry->n_value() + fSlide);
	424	break;
	425	}
	426	}
69a49097	427	}
74cfe461 A	428
74cfe461 A	429