view src/unexnext.c @ 4139:0b32ee899a3a

Consistently use the mark bit of the root interval's parent field to say whether or not the interval tree has been visited (and skip it when revisited), and the mark bit of the plist field to say whether or not that interval has been visited (and abort if revisited); don't try to use the plist mark bit for both meanings. * alloc.c (mark_interval_tree): Don't test if the interval tree has already been visited here; let the MARK_INTERVAL_TREE macro do that; avoid function call overhead. Mark the interval tree as having been visited by setting TREE->parent's mark bit. (MARK_INTERVAL_TREE): If the tree has been visited (according to I->parent's mark bit), don't call mark_interval_tree. (gc_sweep): Rebalance the interval trees of those large strings which are still alive. This also clears the mark bits of those trees' root intervals' parent fields. (compact_strings): Rebalance the interval tree of each small strings which is still alive. This also clears the mark bits of that tree's root interval's parent field. Since the string has moved, update the root interval's parent pointer to contain the new address. * lisp.h (struct interval): Doc fix; explain the roles of the mark bits of the parent and plist members.

/* Dump Emacs in macho format.
   Copyright (C) 1990, 1993 Free Software Foundation, Inc.
   Written by Bradley Taylor ([e-mail]).

This file is part of GNU Emacs.

GNU Emacs is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2, or (at your option)
any later version.

GNU Emacs is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with GNU Emacs; see the file COPYING.  If not, write to
the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.  */


#undef __STRICT_BSD__

#include <stdio.h>
#include <stdlib.h>
#include <stdarg.h>
#include <mach/mach.h>
#include <mach-o/loader.h>
#include <sys/file.h>
#include <sys/stat.h>
#include <libc.h>


int malloc_cookie;

/*
 * Kludge: we don't expect any program data beyond VM_HIGHDATA
 * What is really needed is a way to find out from malloc() which
 * pages it vm_allocated and write only those out into the data segment.
 *
 * This kludge may break when we stop using fixed virtual address
 * shared libraries. Actually, emacs will probably continue working, but be 
 * much larger on disk than it needs to be (because non-malloced data will
 * be in the file).
 */
static const unsigned VM_HIGHDATA = 0x2000000;

typedef struct region_t {
	vm_address_t address;
	vm_size_t size;
	vm_prot_t protection;
	vm_prot_t max_protection;
	vm_inherit_t inheritance;
	boolean_t shared;
	port_t object_name;
	vm_offset_t offset;
} region_t;


static void
grow(
     struct load_command ***the_commands,
     unsigned *the_commands_len
     )
{
	if (*the_commands == NULL) {
		*the_commands_len = 1;
		*the_commands = malloc(sizeof(*the_commands));
	} else {
		(*the_commands_len)++;
		*the_commands = realloc(*the_commands, 
					(*the_commands_len *
					 sizeof(**the_commands)));
	}
}


static void
save_command(
	     struct load_command *command,
	     struct load_command ***the_commands,
	     unsigned *the_commands_len
	     )
{
	struct load_command **tmp;

	grow(the_commands, the_commands_len);
	tmp = &(*the_commands)[*the_commands_len - 1];
	*tmp = malloc(command->cmdsize);
	bcopy(command, *tmp, command->cmdsize);
}

static void
fatal_unexec(char *format, ...)
{
	va_list ap;

	va_start(ap, format);
	fprintf(stderr, "unexec: ");
	vfprintf(stderr, format, ap);
	fprintf(stderr, "\n");
	va_end(ap);
}

static int
read_macho(
	   int fd,
	   struct mach_header *the_header,
	   struct load_command ***the_commands,
	   unsigned *the_commands_len
	   )
{
	struct load_command command;
	struct load_command *buf;
	int i;
	int size;

	if (read(fd, the_header, sizeof(*the_header)) != sizeof(*the_header)) {
		fatal_unexec("cannot read macho header");
		return (0);
	}
	for (i = 0; i < the_header->ncmds; i++) {
		if (read(fd, &command, sizeof(struct load_command)) != 
		    sizeof(struct load_command)) {
		  	fatal_unexec("cannot read macho load command header");
			return (0);
		}
		size = command.cmdsize - sizeof(struct load_command);
		if (size < 0) {
		  	fatal_unexec("bogus load command size");
			return (0);
		}
		buf = malloc(command.cmdsize);
		buf->cmd = command.cmd;
		buf->cmdsize = command.cmdsize;
		if (read(fd, ((char *)buf + 
			      sizeof(struct load_command)), 
			 size) != size) {
		  	fatal_unexec("cannot read load command data");
			return (0);
		}
		save_command(buf, the_commands, the_commands_len);
	}
	return (1);
}

static int
filldatagap(
	    vm_address_t start_address,
	    vm_size_t *size,
	    vm_address_t end_address
	    )
{
	vm_address_t address;
	vm_size_t gapsize;

	address = (start_address + *size);
	gapsize = end_address - address;
	*size += gapsize;
	if (vm_allocate(task_self(), &address, gapsize,
			FALSE) != KERN_SUCCESS) {
		fatal_unexec("cannot vm_allocate");
	        return (0);
	}
	return (1);
}

static int
get_data_region(
		vm_address_t *address,
		vm_size_t *size
		)
{
	region_t region;
	kern_return_t ret;
	struct section *sect;

	sect = getsectbyname(SEG_DATA, SECT_DATA);
	region.address = 0;
	*address = 0;
	for (;;) {
		ret = vm_region(task_self(), 
				&region.address, 
				&region.size, 
				&region.protection, 
				&region.max_protection, 
				&region.inheritance,
				&region.shared, 
				&region.object_name, 
				&region.offset);
		if (ret != KERN_SUCCESS || region.address >= VM_HIGHDATA) {
			break;
		}
		if (*address != 0) {
			if (region.address > *address + *size) {
				if (!filldatagap(*address, size, 
						 region.address)) {
					return (0);
				}
			} 
			*size += region.size;
		} else {
			if (region.address == sect->addr) {
				*address = region.address;
				*size = region.size;
			} 
		}
		region.address += region.size;
	}
	return (1);
}

static char *
my_malloc(
	  vm_size_t size
	  )
{
	vm_address_t address;

	if (vm_allocate(task_self(), &address, size, TRUE) != KERN_SUCCESS) {
		return (NULL);
	}
	return ((char *)address);
}

static void
my_free(
	char *buf,
	vm_size_t size
	)
{
	vm_deallocate(task_self(), (vm_address_t)buf, size);
}

static int
unexec_doit(
	    int infd,
	    int outfd
	    )
{
	int i;
	struct load_command **the_commands = NULL;
	unsigned the_commands_len;
	struct mach_header the_header;
	int fgrowth;
	int fdatastart;
	int fdatasize;
	int size;
	struct stat st;
	char *buf;
	vm_address_t data_address;
	vm_size_t data_size;

	struct segment_command *segment;

	if (!read_macho(infd, &the_header, &the_commands, &the_commands_len)) {
		return (0);
	}


	malloc_cookie = malloc_freezedry ();
	if (!get_data_region(&data_address, &data_size)) {
		return (0);
	}


	/*
	 * DO NOT USE MALLOC IN THIS SECTION
	 */
	{
		/*
		 * Fix offsets
		 */
		for (i = 0; i < the_commands_len; i++) {
			switch (the_commands[i]->cmd) {
			case LC_SEGMENT:
				segment = ((struct segment_command *)
					   the_commands[i]);
				if (strcmp(segment->segname, SEG_DATA) == 0) {
					fdatastart = segment->fileoff;
					fdatasize = segment->filesize;
					fgrowth = (data_size - 
						   segment->filesize);
					segment->vmsize = data_size;
					segment->filesize = data_size;
				}
				break;
			case LC_SYMTAB:
				((struct symtab_command *)
				 the_commands[i])->symoff += fgrowth;
				((struct symtab_command *)
				 the_commands[i])->stroff += fgrowth;
				break;
			case LC_SYMSEG:
				((struct symseg_command *)
				 the_commands[i])->offset += fgrowth;
				break;
			default:
				break;
			}
		}
		
		/*
		 * Write header
		 */
		if (write(outfd, &the_header, 
			  sizeof(the_header)) != sizeof(the_header)) {
			fatal_unexec("cannot write output file");
			return (0);
		}
		
		/*
		 * Write commands
		 */
		for (i = 0; i < the_commands_len; i++) {
			if (write(outfd, the_commands[i], 
				  the_commands[i]->cmdsize) != 
			    the_commands[i]->cmdsize) {
			  	fatal_unexec("cannot write output file");
				return (0);
			}
		}
		
		/*
		 * Write original text
		 */
		if (lseek(infd, the_header.sizeofcmds + sizeof(the_header), 
			  L_SET) < 0) {
		  	fatal_unexec("cannot seek input file");
			return (0);
		}
		size = fdatastart - (sizeof(the_header) + 
				     the_header.sizeofcmds);
		buf = my_malloc(size);
		if (read(infd, buf, size) != size) {
			my_free(buf, size);
		  	fatal_unexec("cannot read input file");
		}
		if (write(outfd, buf, size) != size) {
			my_free(buf, size);
			fatal_unexec("cannot write output file");
			return (0);
		}
		my_free(buf, size);
		
		
		/*
		 * Write new data
		 */
		if (write(outfd, (char *)data_address, 
			  data_size) != data_size) {
			fatal_unexec("cannot write output file");
			return (0);
		}
		
	}

	/*
	 * OKAY TO USE MALLOC NOW
	 */

	/*
	 * Write rest of file
	 */
	fstat(infd, &st);
	if (lseek(infd, fdatasize, L_INCR) < 0) {
		fatal_unexec("cannot seek input file");
		return (0);
	}
	size = st.st_size - lseek(infd, 0, L_INCR);

	buf = malloc(size);
	if (read(infd, buf, size) != size) {
		free(buf);
		fatal_unexec("cannot read input file");
		return (0);
	}
	if (write(outfd, buf, size) != size) {
		free(buf);
		fatal_unexec("cannot write output file");
		return (0);
	}
	free(buf);
	return (1);
}

void
unexec(
       char *outfile,
       char *infile
       )
{
	int infd;
	int outfd;
	char tmpbuf[L_tmpnam];
	char *tmpfile;

	infd = open(infile, O_RDONLY, 0);
	if (infd < 0) {
	  	fatal_unexec("cannot open input file `%s'", infile);
		exit(1);
	}
	
	tmpnam(tmpbuf);
	tmpfile = rindex(tmpbuf, '/');
	if (tmpfile == NULL) {
		tmpfile = tmpbuf;
	} else {
		tmpfile++;
	}
	outfd = open(tmpfile, O_WRONLY|O_TRUNC|O_CREAT, 0755);
	if (outfd < 0) {
		close(infd);
		fatal_unexec("cannot open tmp file `%s'", tmpfile);
		exit(1);
	}
	if (!unexec_doit(infd, outfd)) {
		close(infd);
		close(outfd);
		unlink(tmpfile);
		exit(1);
	}
	close(infd);
	close(outfd);
	if (rename(tmpfile, outfile) < 0) {
		unlink(tmpfile);
		fatal_unexec("cannot rename `%s' to `%s'", tmpfile, outfile);
		exit(1);
	}
}