uniso.c

/* uniso.c - Unpack ISO9660 File System from a stream

Copyright (C) 2011 Timo Teräs <timo.teras@iki.fi>
All rights reserved.

Permission is hereby granted, free of charge, to any person obtaining
a copy of this software and associated documentation files (the
"Software"), to deal in the Software without restriction, including
without limitation the rights to use, copy, modify, merge, publish,
distribute, sublicense, and/or sell copies of the Software, and to
permit persons to whom the Software is furnished to do so, subject to
the following conditions:

The above copyright notice and this permission notice shall be
included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

*/

/* Compile with:
 * gcc -std=gnu99 -D_GNU_SOURCE -D_BSD_SOURCE -O2 uniso.c -o uniso */

/*
 * TODO:
 * - fix unaligned 16-bit accesses from iso headers (ARM / MIPS)
 * - options, verbose logs
 */

/* needed for SPLICE_F_MOVE */
#define _GNU_SOURCE

#include <errno.h>
#include <getopt.h>
#include <stdio.h>
#include <fcntl.h>
#include <wchar.h>
#include <unistd.h>
#include <stddef.h>
#include <stdlib.h>
#include <string.h>
#include <sys/stat.h>
#include <sys/types.h>

#if defined(__linux__)
# include <malloc.h>
# include <endian.h>
# include <sys/vfs.h>
#elif defined(__APPLE__)
# include <libkern/OSByteOrder.h>
# define be16toh(x) OSSwapBigToHostInt16(x)
#endif

#ifndef EMEDIUMTYPE
#define EMEDIUMTYPE EILSEQ
#endif


/**/

#ifndef container_of
#define container_of(ptr, type, member) ({                      \
        const typeof( ((type *)0)->member ) *__mptr = (ptr);    \
        (type *)( (char *)__mptr - offsetof(type,member) );})
#endif

#define LIST_POISON1 (void *) 0xdeadbeef
#define LIST_POISON2 (void *) 0xabbaabba

struct list_head {
	struct list_head *next, *prev;
};

static inline void list_init(struct list_head *list)
{
	list->next = list;
	list->prev = list;
}

static inline void __list_add(struct list_head *new,
			      struct list_head *prev,
			      struct list_head *next)
{
	next->prev = new;
	new->next = next;
	new->prev = prev;
	prev->next = new;
}

static inline void list_add(struct list_head *new, struct list_head *head)
{
	__list_add(new, head, head->next);
}

static inline void list_add_tail(struct list_head *new, struct list_head *head)
{
	__list_add(new, head->prev, head);
}

static inline void __list_del(struct list_head * prev, struct list_head * next)
{
	next->prev = prev;
	prev->next = next;
}

static inline void list_del(struct list_head *entry)
{
	__list_del(entry->prev, entry->next);
	entry->next = LIST_POISON1;
	entry->prev = LIST_POISON2;
}

static inline int list_hashed(const struct list_head *n)
{
	return n->next != n && n->next != NULL;
}

#define list_entry(ptr, type, member) container_of(ptr,type,member)

/**/

#define ISOFS_BLOCK_SIZE		2048
#define ISOFS_TMPBUF_SIZE		(32*ISOFS_BLOCK_SIZE)

#if __BYTE_ORDER == __LITTLE_ENDIAN
#define endianess	le
#else
#define endianess	be
#endif

struct isofs723 {
	u_int16_t le;
	u_int16_t be;
};

struct isofs731 {
	u_int32_t le;
};

struct isofs732 {
	u_int32_t be;
};

struct isofs733 {
	u_int32_t le;
	u_int32_t be;
};

#define ISOFS_VD_PRIMARY		1
#define ISOFS_VD_SUPPLEMENTARY		2
#define ISOFS_VD_END			255
#define ISOFS_STANDARD_ID		"CD001"

struct isofs_volume_descriptor {
	unsigned char type;
	unsigned char id[5];
	unsigned char version;
	unsigned char data[2040];
};

/* Primary or Supplementary Volume Descriptor
 * (they are about the same, Supplementary just has few more fields
 *  and the date they point to might be encoded differently) */
struct isofs_pri_sup_descriptor {
	unsigned char 	type;
	unsigned char 	id[5];
	unsigned char 	version;
	char		flags;			/* unused1 in primary */
	char		system_id[32];
	char		volume_id[32];
	char		unused2[8];
	struct isofs733	volume_space_size;
	char		escape[32];		/* unused3 in primary */
	struct isofs723	volume_set_size;
	struct isofs723	volume_sequence_number;
	struct isofs723	logical_block_size;
	struct isofs733	path_table_size;
	struct isofs731	type_l_path_table, opt_type_l_path_table;
	struct isofs732	type_m_path_table, opt_type_m_path_table;
	char		root_directory_record[34];
	char		volume_set_id[128];
	char		publisher_id[128];
	char		preparer_id[128];
	char		application_id[128];
	char		copyright_file_id[37];
	char		abstract_file_id[37];
	char		bibliographic_file_id[37];
	char		creation_date[17];
	char		modification_date[17];
	char		expiration_date[17];
	char		effective_date[17];
	char		file_structure_version;
	char		unused4[1];
	char		application_data[512];
	char		unused5[653];
};

struct isofs_directory_record {
	unsigned char	length;
	unsigned char	ext_attr_length;
	struct isofs733	extent;
	struct isofs733	size;
	char		date[7];
	char		flags;
#define ISOFS_DR_FLAG_DIRECTORY		(1<<1)
	unsigned char	file_unit_size;
	unsigned char	interleave;
	struct isofs723	volume_sequence_number;
	unsigned char	name_len;
	char		name[0];
} __attribute__((packed));

struct uniso_reader;
struct uniso_context;

typedef int (*uniso_handler_f)(struct uniso_context *, struct uniso_reader *);

struct uniso_reader {
	struct list_head parser_list;
	off_t offset;
	uniso_handler_f handler;
};

struct uniso_dirent {
	struct uniso_reader reader;
	size_t size;
	u_int32_t flags;
	char name[0];
};

struct uniso_context {
	int stream_fd, null_fd;
	int skip_method, copy_method;
	int joliet_level;
	int loglevel, block_size;
	off_t pos, last_queued_read, disk_free, disk_needed;
	struct list_head parser_head;
	struct uniso_reader volume_desc_reader;
	unsigned char *tmpbuf;
};

#if defined(__linux__)
static int do_splice(struct uniso_context *ctx, int to_fd, size_t bytes)
{
	int r, left;

	for (left = bytes; left; ) {
		r = splice(ctx->stream_fd, NULL, to_fd, NULL, left,
			   SPLICE_F_MOVE);
		if (r < 0) return -errno;
		if (r == 0) return bytes - left;
		left -= r;
		ctx->pos += r;
	}

	return bytes;
}
#else
#define do_splice(ctx,to_fd,bytes) (-1)
#endif

static int do_read(struct uniso_context *ctx, unsigned char *buf, size_t bytes)
{
	int r, left;

	for (left = bytes; left; ) {
		r = read(ctx->stream_fd, buf, left);
		if (r < 0) {
			perror("read");
			return -errno;
		}
		if (r == 0) return bytes - left;
		left -= r;
		buf += r;
		ctx->pos += r;
	}

	return bytes;
}

static int do_write(int tofd, const unsigned char *buf, size_t bytes)
{
	int r, left;

	for (left = bytes; left; ) {
		r = write(tofd, buf, left);
		if (r < 0) {
			perror("write");
			return -errno;
		}
		left -= r;
		buf += r;
	}

	return bytes;
}

static int do_skip(struct uniso_context *ctx, size_t bytes)
{
	int r, left, now;

	switch (ctx->skip_method) {
	case 0:
		if (lseek(ctx->stream_fd, bytes, SEEK_CUR) != (off_t) -1) {
			ctx->pos += bytes;
			return bytes;
		}
		ctx->skip_method = 1;
	case 1:
		r = do_splice(ctx, ctx->null_fd, bytes);
		if (r >= 0) return r;
		ctx->skip_method = 2;
	case 2:
		for (left = bytes; left; ) {
			now = left;
			if (now > ISOFS_TMPBUF_SIZE)
				now = ISOFS_TMPBUF_SIZE;
			r = do_read(ctx, ctx->tmpbuf, now);
			if (r < 0) return r;
			if (r == 0) return bytes - left;
			left -= now;
		}
		return bytes;
	}
	return 0;
}

static int do_copy(struct uniso_context *ctx, int tofd, size_t bytes)
{
	int r, now, left;

	switch (ctx->copy_method) {
	case 0:
		r = do_splice(ctx, tofd, bytes);
		if (r >= 0) return r;
		ctx->copy_method = 1;
	case 1:
		/* FIXME: Use mmaped IO */
		ctx->copy_method = 2;
	case 2:
		for (left = bytes; left; ) {
			now = left;
			if (now > ISOFS_TMPBUF_SIZE)
				now = ISOFS_TMPBUF_SIZE;

			r = do_read(ctx, ctx->tmpbuf, now);
			if (r <= 0) return r < 0 ? r : -EIO;

			now = r;
			r = do_write(tofd, ctx->tmpbuf, now);
			if (r != now) return r < 0 ? r : -EIO;

			left -= now;
		}
		return bytes;
	}
	return 0;
}

static int queue_reader(
		struct uniso_context *ctx, struct uniso_reader *reader,
		off_t offset, uniso_handler_f handler)
{
	struct list_head *n;
	struct uniso_reader *r;

	if (offset < ctx->pos) {
		fprintf(stderr, "ERROR: non-linear reads are not supported "
				"(asked %zx, we are at %zx)\n",
			offset, ctx->last_queued_read);
		return -1;
	}

	reader->offset = offset;
	reader->handler = handler;
	list_init(&reader->parser_list);

	if (offset >= ctx->last_queued_read || !list_hashed(&ctx->parser_head)) {
		ctx->last_queued_read = offset;
		list_add_tail(&reader->parser_list, &ctx->parser_head);
		return 0;
	}

	for (n = ctx->parser_head.prev; n != &ctx->parser_head; n = n->prev) {
		r = list_entry(n, struct uniso_reader, parser_list);
		if (r->offset < offset)
			break;
	}
	list_add(&reader->parser_list, n);

	return 0;
}

static int queue_dirent(struct uniso_context *ctx, void *isode, const char *name);

static int link_or_clone(const char *src, const char *dst, size_t bytes)
{
	int src_fd, dst_fd;
	int r;
	static char buf[ISOFS_TMPBUF_SIZE];

	if (link(src, dst) == 0)
		return 0;

	src_fd = open(src, O_RDONLY);
	if (src_fd < 0)
		return src_fd;

	dst_fd = open(dst, O_WRONLY | O_TRUNC | O_CREAT, 0777);
	if (dst_fd < 0) {
		close(src_fd);
		return dst_fd;
	}

	while (bytes) {
		size_t now = bytes;
		if (now > sizeof(buf))
			now = sizeof(buf);
		r = read(src_fd, buf, now);
		if (r < 0) return r;

		r = write(dst_fd, buf, now);
		if (r < 0) return r;
		bytes -= now;
	}
	close(dst_fd);
	close(src_fd);
	return 0;
}

static int uniso_read_file(struct uniso_context *ctx,
                           struct uniso_reader *rd)
{
	struct uniso_dirent *dir = container_of(rd, struct uniso_dirent, reader);
	int fd, rc;
	static off_t prev_offset = 0;
	static char prev_name[512];
	static size_t prev_size = 0;

	if (ctx->disk_free && ctx->disk_needed > ctx->disk_free) {
		if (ctx->loglevel > 0)
			fprintf(stderr, "ERROR: Disk needed %zd MiB (free %zd MiB)\n",
				(size_t)(ctx->disk_needed * ctx->block_size / (1024*1024)),
				(size_t)(ctx->disk_free * ctx->block_size / (1024*1024)));
		return -ENOSPC;
	}

	// FIXME: seems that hardlinked files get the same file extent
	// shared. need to fix extraction of such files.
	if (ctx->pos > rd->offset) {
		rc = 1;
		if (rd->offset == prev_offset)
			rc = link_or_clone(prev_name, dir->name, prev_size);
		if (ctx->loglevel > 0 && rc != 0)
			fprintf(stderr, "WARNING: Not extracting '%s' as "
					"it's sharing file-extents with "
					"another file\n", dir->name);
		return 0;
	}

	prev_offset = rd->offset;
	strncpy(prev_name, dir->name, sizeof(prev_name));
	prev_size = dir->size;

	fd = open(dir->name, O_WRONLY | O_TRUNC | O_CREAT, 0777);
	if (fd < 0)
		return -errno;

	if (ctx->loglevel > 1)
		fprintf(stderr, "%s : %zd bytes, flags 0x%08x\n",
			dir->name, dir->size, dir->flags);

	if (dir->size) {
		rc = -posix_fallocate(fd, 0, dir->size);
		if (rc) goto err;
	}
	rc = do_copy(ctx, fd, dir->size);
	if (rc == dir->size) fdatasync(fd);
err:
	close(fd);
	if (rc < 0) return rc;
	if (rc != dir->size) return -EIO;
	return 0;
}

static int uniso_read_directory(struct uniso_context *ctx,
				struct uniso_reader *rd)
{
	struct uniso_dirent *dir = container_of(rd, struct uniso_dirent, reader);
	struct isofs_directory_record *ide;
	char buf[512], *name;
	unsigned char *tmp;
	int r, i, offs;

	if (ctx->loglevel > 1)
		fprintf(stderr, "%s : DIR, flags 0x%08x\n",
			dir->name, dir->flags);
	mkdir(dir->name, 0777);

	tmp = malloc(dir->size);
	if (tmp == NULL) {
		r = -ENOMEM;
		goto ret_nomem;
	}

	r = do_read(ctx, tmp, dir->size);
	if (r < 0)
		goto ret;

	strcpy(buf, dir->name);
	strcat(buf, "/");

	for (offs = 0; offs < dir->size; ) {
		ide = (void *) &tmp[offs];

		offs += (ide->length + 1) & ~1;
		if (ide->length == 0) {
			offs &= ~(ISOFS_BLOCK_SIZE-1);
			offs += ISOFS_BLOCK_SIZE;
			continue;
		}

		/* Skip '.' and '..' entries */
		if ((ide->name_len == 0) ||
		    (ide->name_len == 1 && ide->name[0] <= 1))
			continue;

		name = &buf[strlen(dir->name) + 1];
		if (ctx->joliet_level) {
			u_int16_t *wname = (u_int16_t *) ide->name;
			for (i = 0; i < ide->name_len; i += 2, wname++) {
				r = wctomb(name, be16toh(*wname));
				if (r == -1)
					*name++ = '?';
				else
					name += r;
			}
			*name = 0;
		} else {
			memcpy(name, ide->name, ide->name_len);
			name[ide->name_len] = 0;
		}

		queue_dirent(ctx, ide, buf);
	}
	r = 0;

ret:
	free(tmp);
ret_nomem:
	free(dir);
	return r;
}

static int queue_dirent(struct uniso_context *ctx, void *isode, const char *name)
{
	struct isofs_directory_record *ide = isode;
	struct uniso_dirent *dir;

	dir = calloc(1, sizeof(*dir) + strlen(name) + 1);
	if (dir == NULL)
		return -ENOMEM;

	dir->size = ide->size.endianess;
	dir->flags = ide->flags;
	strcpy(dir->name, name);

	if (!(ide->flags & ISOFS_DR_FLAG_DIRECTORY))
		ctx->disk_needed += (dir->size + ctx->block_size - 1) / ctx->block_size;

	return queue_reader(ctx, &dir->reader,
	                    ide->extent.endianess * ISOFS_BLOCK_SIZE,
	                    (ide->flags & ISOFS_DR_FLAG_DIRECTORY) ?
	                    uniso_read_directory : uniso_read_file);
}

static int uniso_read_volume_descriptor(struct uniso_context *ctx,
					struct uniso_reader *rd)
{
	unsigned char buf[ISOFS_BLOCK_SIZE];
	struct isofs_volume_descriptor *vd = (void*) buf;
	struct isofs_pri_sup_descriptor *pd = (void*) buf;
	char root_dir[sizeof(pd->root_directory_record)];
	int r;

	do {
		r = do_read(ctx, buf, ISOFS_BLOCK_SIZE);
		if (r < 0)
			return r;

		if (memcmp(vd->id, ISOFS_STANDARD_ID, sizeof(vd->id)) != 0)
			return -EMEDIUMTYPE;

		switch (vd->type) {
		case ISOFS_VD_PRIMARY:
			if (ctx->joliet_level == 0)
				memcpy(root_dir, pd->root_directory_record, sizeof(root_dir));
			break;
		case ISOFS_VD_SUPPLEMENTARY:
			if (pd->escape[0] == 0x25 && pd->escape[1] == 0x2f) {
				switch (pd->escape[2]) {
				case 0x45:
					ctx->joliet_level++;
				case 0x43:
					ctx->joliet_level++;
				case 0x40:
					ctx->joliet_level++;
					memcpy(root_dir, pd->root_directory_record, sizeof(root_dir));
				}
			}
			break;
		}
	} while (vd->type != ISOFS_VD_END);

	if (ctx->joliet_level && ctx->loglevel > 1)
		fprintf(stderr, "INFO: Microsoft Joliet Level %d\n",
			ctx->joliet_level);

	return queue_dirent(ctx, root_dir, ".");
}

static void get_disk_info(struct uniso_context *ctx)
{
#if defined(__linux__)
	struct statfs sfs;

	if (statfs(".", &sfs) == 0) {
		if (sfs.f_bsize) ctx->block_size = sfs.f_bsize;
		ctx->disk_free = (off_t)sfs.f_bavail;
	}
#endif
}

int uniso(int fd)
{
	struct uniso_context context, *ctx = &context;
	struct uniso_reader *rd;
	ssize_t r, skipped;

	memset(ctx, 0, sizeof(*ctx));
	list_init(&ctx->parser_head);
	ctx->stream_fd = fd;
	ctx->null_fd = open("/dev/null", O_RDWR);
	ctx->tmpbuf = malloc(ISOFS_TMPBUF_SIZE);
	ctx->loglevel = 1;
	ctx->block_size = 4*1024;
	get_disk_info(ctx);

	queue_reader(ctx, &ctx->volume_desc_reader,
	             16 * ISOFS_BLOCK_SIZE,
	             uniso_read_volume_descriptor);

	while (list_hashed(&ctx->parser_head)) {
		rd = list_entry(ctx->parser_head.next, struct uniso_reader, parser_list);
		list_del(&rd->parser_list);

		r = rd->offset - ctx->pos;
		if (r > 0) {
			if ((r = do_skip(ctx, r)) < 0)
				return r;
		}
		r = rd->handler(ctx, rd);
		if (r != 0)
			return r;
	}

	if (ctx->skip_method) {
		skipped = 0;
		do {
			r = do_skip(ctx, ISOFS_TMPBUF_SIZE);
			if (r > 0) skipped += r;
		} while (r == ISOFS_TMPBUF_SIZE);
		if (ctx->loglevel > 1) fprintf(stderr, "Skipped %zd bytes at the end\n", skipped);
	}

	free(ctx->tmpbuf);
	close(ctx->null_fd);
	return 0;
}

static void usage(FILE *out, int rc) {
	fprintf(out,
	"usage: uniso [-h]\n"
	"\n"
	"Unpack ISO9660 File System from STDIN.\n"
	"\n"
	"options:\n"
	" -h  Show this help\n"
	"\n");

	exit(rc);
}

int main(int argc, char *argv[])
{
	while (1) {
		int c = getopt(argc, argv, "h");
		if (c < 0)
			break;

		switch (c) {
		case 'h':
			usage(stdout, 0);
			break;
		default:
			usage(stderr, 1);
		}
	}

	uniso(STDIN_FILENO);
	return 0;
}