diff options
author | Imre Kaloz <kaloz@openwrt.org> | 2008-01-28 13:36:09 +0000 |
---|---|---|
committer | Imre Kaloz <kaloz@openwrt.org> | 2008-01-28 13:36:09 +0000 |
commit | 5303deecf170c291df307f23480590b53f8287e5 (patch) | |
tree | c19d3bb7b0ac1d963bf2a2875e3456f5872d9ae5 /target/linux/generic-2.6/patches-2.6.24/001-squashfs.patch | |
parent | f66f06c1894b7becb24cad683f10201e04d60724 (diff) | |
download | upstream-5303deecf170c291df307f23480590b53f8287e5.tar.gz upstream-5303deecf170c291df307f23480590b53f8287e5.tar.bz2 upstream-5303deecf170c291df307f23480590b53f8287e5.zip |
generic 2.6.24 support
SVN-Revision: 10291
Diffstat (limited to 'target/linux/generic-2.6/patches-2.6.24/001-squashfs.patch')
-rw-r--r-- | target/linux/generic-2.6/patches-2.6.24/001-squashfs.patch | 4192 |
1 files changed, 4192 insertions, 0 deletions
diff --git a/target/linux/generic-2.6/patches-2.6.24/001-squashfs.patch b/target/linux/generic-2.6/patches-2.6.24/001-squashfs.patch new file mode 100644 index 0000000000..6d2cb28b98 --- /dev/null +++ b/target/linux/generic-2.6/patches-2.6.24/001-squashfs.patch @@ -0,0 +1,4192 @@ +Index: linux-2.6.23/fs/Kconfig +=================================================================== +--- linux-2.6.23.orig/fs/Kconfig 2007-10-10 13:52:12.000000000 +0800 ++++ linux-2.6.23/fs/Kconfig 2007-10-10 13:52:14.000000000 +0800 +@@ -1364,6 +1364,71 @@ + + If unsure, say N. + ++config SQUASHFS ++ tristate "SquashFS 3.0 - Squashed file system support" ++ select ZLIB_INFLATE ++ help ++ Saying Y here includes support for SquashFS 3.0 (a Compressed Read-Only File ++ System). Squashfs is a highly compressed read-only filesystem for Linux. ++ It uses zlib compression to compress both files, inodes and directories. ++ Inodes in the system are very small and all blocks are packed to minimise ++ data overhead. Block sizes greater than 4K are supported up to a maximum of 64K. ++ SquashFS 3.0 supports 64 bit filesystems and files (larger than 4GB), full ++ uid/gid information, hard links and timestamps. ++ ++ Squashfs is intended for general read-only filesystem use, for archival ++ use (i.e. in cases where a .tar.gz file may be used), and in embedded ++ systems where low overhead is needed. Further information and filesystem tools ++ are available from http://squashfs.sourceforge.net. ++ ++ If you want to compile this as a module ( = code which can be ++ inserted in and removed from the running kernel whenever you want), ++ say M here and read <file:Documentation/modules.txt>. The module ++ will be called squashfs. Note that the root file system (the one ++ containing the directory /) cannot be compiled as a module. ++ ++ If unsure, say N. ++ ++config SQUASHFS_EMBEDDED ++ ++ bool "Additional options for memory-constrained systems" ++ depends on SQUASHFS ++ default n ++ help ++ Saying Y here allows you to specify cache sizes and how Squashfs ++ allocates memory. This is only intended for memory constrained ++ systems. ++ ++ If unsure, say N. ++ ++config SQUASHFS_FRAGMENT_CACHE_SIZE ++ int "Number of fragments cached" if SQUASHFS_EMBEDDED ++ depends on SQUASHFS ++ default "3" ++ help ++ By default SquashFS caches the last 3 fragments read from ++ the filesystem. Increasing this amount may mean SquashFS ++ has to re-read fragments less often from disk, at the expense ++ of extra system memory. Decreasing this amount will mean ++ SquashFS uses less memory at the expense of extra reads from disk. ++ ++ Note there must be at least one cached fragment. Anything ++ much more than three will probably not make much difference. ++ ++config SQUASHFS_VMALLOC ++ bool "Use Vmalloc rather than Kmalloc" if SQUASHFS_EMBEDDED ++ depends on SQUASHFS ++ default n ++ help ++ By default SquashFS uses kmalloc to obtain fragment cache memory. ++ Kmalloc memory is the standard kernel allocator, but it can fail ++ on memory constrained systems. Because of the way Vmalloc works, ++ Vmalloc can succeed when kmalloc fails. Specifying this option ++ will make SquashFS always use Vmalloc to allocate the ++ fragment cache memory. ++ ++ If unsure, say N. ++ + config VXFS_FS + tristate "FreeVxFS file system support (VERITAS VxFS(TM) compatible)" + depends on BLOCK +Index: linux-2.6.23/fs/Makefile +=================================================================== +--- linux-2.6.23.orig/fs/Makefile 2007-10-10 13:52:12.000000000 +0800 ++++ linux-2.6.23/fs/Makefile 2007-10-10 13:52:14.000000000 +0800 +@@ -72,6 +72,7 @@ + obj-$(CONFIG_JBD2) += jbd2/ + obj-$(CONFIG_EXT2_FS) += ext2/ + obj-$(CONFIG_CRAMFS) += cramfs/ ++obj-$(CONFIG_SQUASHFS) += squashfs/ + obj-y += ramfs/ + obj-$(CONFIG_HUGETLBFS) += hugetlbfs/ + obj-$(CONFIG_CODA_FS) += coda/ +Index: linux-2.6.23/fs/squashfs/inode.c +=================================================================== +--- /dev/null 1970-01-01 00:00:00.000000000 +0000 ++++ linux-2.6.23/fs/squashfs/inode.c 2007-10-10 13:52:14.000000000 +0800 +@@ -0,0 +1,2122 @@ ++/* ++ * Squashfs - a compressed read only filesystem for Linux ++ * ++ * Copyright (c) 2002, 2003, 2004, 2005, 2006 ++ * Phillip Lougher <phillip@lougher.org.uk> ++ * ++ * This program 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. ++ * ++ * This program 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 this program; if not, write to the Free Software ++ * Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. ++ * ++ * inode.c ++ */ ++ ++#include <linux/types.h> ++#include <linux/squashfs_fs.h> ++#include <linux/module.h> ++#include <linux/errno.h> ++#include <linux/slab.h> ++#include <linux/fs.h> ++#include <linux/smp_lock.h> ++#include <linux/slab.h> ++#include <linux/squashfs_fs_sb.h> ++#include <linux/squashfs_fs_i.h> ++#include <linux/buffer_head.h> ++#include <linux/vfs.h> ++#include <linux/init.h> ++#include <linux/dcache.h> ++#include <linux/wait.h> ++#include <linux/zlib.h> ++#include <linux/blkdev.h> ++#include <linux/vmalloc.h> ++#include <asm/uaccess.h> ++#include <asm/semaphore.h> ++ ++#include "squashfs.h" ++ ++static void squashfs_put_super(struct super_block *); ++static int squashfs_statfs(struct dentry *, struct kstatfs *); ++static int squashfs_symlink_readpage(struct file *file, struct page *page); ++static int squashfs_readpage(struct file *file, struct page *page); ++static int squashfs_readpage4K(struct file *file, struct page *page); ++static int squashfs_readdir(struct file *, void *, filldir_t); ++static struct inode *squashfs_alloc_inode(struct super_block *sb); ++static void squashfs_destroy_inode(struct inode *inode); ++static int init_inodecache(void); ++static void destroy_inodecache(void); ++static struct dentry *squashfs_lookup(struct inode *, struct dentry *, ++ struct nameidata *); ++static struct inode *squashfs_iget(struct super_block *s, squashfs_inode_t inode); ++static long long read_blocklist(struct inode *inode, int index, ++ int readahead_blks, char *block_list, ++ unsigned short **block_p, unsigned int *bsize); ++static int squashfs_get_sb(struct file_system_type *, int, ++ const char *, void *, struct vfsmount *); ++ ++ ++static z_stream stream; ++ ++static struct file_system_type squashfs_fs_type = { ++ .owner = THIS_MODULE, ++ .name = "squashfs", ++ .get_sb = squashfs_get_sb, ++ .kill_sb = kill_block_super, ++ .fs_flags = FS_REQUIRES_DEV ++}; ++ ++static unsigned char squashfs_filetype_table[] = { ++ DT_UNKNOWN, DT_DIR, DT_REG, DT_LNK, DT_BLK, DT_CHR, DT_FIFO, DT_SOCK ++}; ++ ++static struct super_operations squashfs_ops = { ++ .alloc_inode = squashfs_alloc_inode, ++ .destroy_inode = squashfs_destroy_inode, ++ .statfs = squashfs_statfs, ++ .put_super = squashfs_put_super, ++}; ++ ++SQSH_EXTERN struct address_space_operations squashfs_symlink_aops = { ++ .readpage = squashfs_symlink_readpage ++}; ++ ++SQSH_EXTERN struct address_space_operations squashfs_aops = { ++ .readpage = squashfs_readpage ++}; ++ ++SQSH_EXTERN struct address_space_operations squashfs_aops_4K = { ++ .readpage = squashfs_readpage4K ++}; ++ ++static struct file_operations squashfs_dir_ops = { ++ .read = generic_read_dir, ++ .readdir = squashfs_readdir ++}; ++ ++SQSH_EXTERN struct inode_operations squashfs_dir_inode_ops = { ++ .lookup = squashfs_lookup ++}; ++ ++ ++static struct buffer_head *get_block_length(struct super_block *s, ++ int *cur_index, int *offset, int *c_byte) ++{ ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ unsigned short temp; ++ struct buffer_head *bh; ++ ++ if (!(bh = sb_bread(s, *cur_index))) ++ goto out; ++ ++ if (msblk->devblksize - *offset == 1) { ++ if (msblk->swap) ++ ((unsigned char *) &temp)[1] = *((unsigned char *) ++ (bh->b_data + *offset)); ++ else ++ ((unsigned char *) &temp)[0] = *((unsigned char *) ++ (bh->b_data + *offset)); ++ brelse(bh); ++ if (!(bh = sb_bread(s, ++(*cur_index)))) ++ goto out; ++ if (msblk->swap) ++ ((unsigned char *) &temp)[0] = *((unsigned char *) ++ bh->b_data); ++ else ++ ((unsigned char *) &temp)[1] = *((unsigned char *) ++ bh->b_data); ++ *c_byte = temp; ++ *offset = 1; ++ } else { ++ if (msblk->swap) { ++ ((unsigned char *) &temp)[1] = *((unsigned char *) ++ (bh->b_data + *offset)); ++ ((unsigned char *) &temp)[0] = *((unsigned char *) ++ (bh->b_data + *offset + 1)); ++ } else { ++ ((unsigned char *) &temp)[0] = *((unsigned char *) ++ (bh->b_data + *offset)); ++ ((unsigned char *) &temp)[1] = *((unsigned char *) ++ (bh->b_data + *offset + 1)); ++ } ++ *c_byte = temp; ++ *offset += 2; ++ } ++ ++ if (SQUASHFS_CHECK_DATA(msblk->sblk.flags)) { ++ if (*offset == msblk->devblksize) { ++ brelse(bh); ++ if (!(bh = sb_bread(s, ++(*cur_index)))) ++ goto out; ++ *offset = 0; ++ } ++ if (*((unsigned char *) (bh->b_data + *offset)) != ++ SQUASHFS_MARKER_BYTE) { ++ ERROR("Metadata block marker corrupt @ %x\n", ++ *cur_index); ++ brelse(bh); ++ goto out; ++ } ++ (*offset)++; ++ } ++ return bh; ++ ++out: ++ return NULL; ++} ++ ++ ++SQSH_EXTERN unsigned int squashfs_read_data(struct super_block *s, char *buffer, ++ long long index, unsigned int length, ++ long long *next_index) ++{ ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ struct buffer_head *bh[((SQUASHFS_FILE_MAX_SIZE - 1) >> ++ msblk->devblksize_log2) + 2]; ++ unsigned int offset = index & ((1 << msblk->devblksize_log2) - 1); ++ unsigned int cur_index = index >> msblk->devblksize_log2; ++ int bytes, avail_bytes, b = 0, k; ++ char *c_buffer; ++ unsigned int compressed; ++ unsigned int c_byte = length; ++ ++ if (c_byte) { ++ bytes = msblk->devblksize - offset; ++ compressed = SQUASHFS_COMPRESSED_BLOCK(c_byte); ++ c_buffer = compressed ? msblk->read_data : buffer; ++ c_byte = SQUASHFS_COMPRESSED_SIZE_BLOCK(c_byte); ++ ++ TRACE("Block @ 0x%llx, %scompressed size %d\n", index, compressed ++ ? "" : "un", (unsigned int) c_byte); ++ ++ if (!(bh[0] = sb_getblk(s, cur_index))) ++ goto block_release; ++ ++ for (b = 1; bytes < c_byte; b++) { ++ if (!(bh[b] = sb_getblk(s, ++cur_index))) ++ goto block_release; ++ bytes += msblk->devblksize; ++ } ++ ll_rw_block(READ, b, bh); ++ } else { ++ if (!(bh[0] = get_block_length(s, &cur_index, &offset, ++ &c_byte))) ++ goto read_failure; ++ ++ bytes = msblk->devblksize - offset; ++ compressed = SQUASHFS_COMPRESSED(c_byte); ++ c_buffer = compressed ? msblk->read_data : buffer; ++ c_byte = SQUASHFS_COMPRESSED_SIZE(c_byte); ++ ++ TRACE("Block @ 0x%llx, %scompressed size %d\n", index, compressed ++ ? "" : "un", (unsigned int) c_byte); ++ ++ for (b = 1; bytes < c_byte; b++) { ++ if (!(bh[b] = sb_getblk(s, ++cur_index))) ++ goto block_release; ++ bytes += msblk->devblksize; ++ } ++ ll_rw_block(READ, b - 1, bh + 1); ++ } ++ ++ if (compressed) ++ down(&msblk->read_data_mutex); ++ ++ for (bytes = 0, k = 0; k < b; k++) { ++ avail_bytes = (c_byte - bytes) > (msblk->devblksize - offset) ? ++ msblk->devblksize - offset : ++ c_byte - bytes; ++ wait_on_buffer(bh[k]); ++ if (!buffer_uptodate(bh[k])) ++ goto block_release; ++ memcpy(c_buffer + bytes, bh[k]->b_data + offset, avail_bytes); ++ bytes += avail_bytes; ++ offset = 0; ++ brelse(bh[k]); ++ } ++ ++ /* ++ * uncompress block ++ */ ++ if (compressed) { ++ int zlib_err; ++ ++ stream.next_in = c_buffer; ++ stream.avail_in = c_byte; ++ stream.next_out = buffer; ++ stream.avail_out = msblk->read_size; ++ ++ if (((zlib_err = zlib_inflateInit(&stream)) != Z_OK) || ++ ((zlib_err = zlib_inflate(&stream, Z_FINISH)) ++ != Z_STREAM_END) || ((zlib_err = ++ zlib_inflateEnd(&stream)) != Z_OK)) { ++ ERROR("zlib_fs returned unexpected result 0x%x\n", ++ zlib_err); ++ bytes = 0; ++ } else ++ bytes = stream.total_out; ++ ++ up(&msblk->read_data_mutex); ++ } ++ ++ if (next_index) ++ *next_index = index + c_byte + (length ? 0 : ++ (SQUASHFS_CHECK_DATA(msblk->sblk.flags) ++ ? 3 : 2)); ++ return bytes; ++ ++block_release: ++ while (--b >= 0) ++ brelse(bh[b]); ++ ++read_failure: ++ ERROR("sb_bread failed reading block 0x%x\n", cur_index); ++ return 0; ++} ++ ++ ++SQSH_EXTERN int squashfs_get_cached_block(struct super_block *s, char *buffer, ++ long long block, unsigned int offset, ++ int length, long long *next_block, ++ unsigned int *next_offset) ++{ ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ int n, i, bytes, return_length = length; ++ long long next_index; ++ ++ TRACE("Entered squashfs_get_cached_block [%llx:%x]\n", block, offset); ++ ++ while ( 1 ) { ++ for (i = 0; i < SQUASHFS_CACHED_BLKS; i++) ++ if (msblk->block_cache[i].block == block) ++ break; ++ ++ down(&msblk->block_cache_mutex); ++ ++ if (i == SQUASHFS_CACHED_BLKS) { ++ /* read inode header block */ ++ for (i = msblk->next_cache, n = SQUASHFS_CACHED_BLKS; ++ n ; n --, i = (i + 1) % ++ SQUASHFS_CACHED_BLKS) ++ if (msblk->block_cache[i].block != ++ SQUASHFS_USED_BLK) ++ break; ++ ++ if (n == 0) { ++ wait_queue_t wait; ++ ++ init_waitqueue_entry(&wait, current); ++ add_wait_queue(&msblk->waitq, &wait); ++ set_current_state(TASK_UNINTERRUPTIBLE); ++ up(&msblk->block_cache_mutex); ++ schedule(); ++ set_current_state(TASK_RUNNING); ++ remove_wait_queue(&msblk->waitq, &wait); ++ continue; ++ } ++ msblk->next_cache = (i + 1) % SQUASHFS_CACHED_BLKS; ++ ++ if (msblk->block_cache[i].block == ++ SQUASHFS_INVALID_BLK) { ++ if (!(msblk->block_cache[i].data = ++ kmalloc(SQUASHFS_METADATA_SIZE, ++ GFP_KERNEL))) { ++ ERROR("Failed to allocate cache" ++ "block\n"); ++ up(&msblk->block_cache_mutex); ++ goto out; ++ } ++ } ++ ++ msblk->block_cache[i].block = SQUASHFS_USED_BLK; ++ up(&msblk->block_cache_mutex); ++ ++ if (!(msblk->block_cache[i].length = ++ squashfs_read_data(s, ++ msblk->block_cache[i].data, ++ block, 0, &next_index))) { ++ ERROR("Unable to read cache block [%llx:%x]\n", ++ block, offset); ++ goto out; ++ } ++ ++ down(&msblk->block_cache_mutex); ++ wake_up(&msblk->waitq); ++ msblk->block_cache[i].block = block; ++ msblk->block_cache[i].next_index = next_index; ++ TRACE("Read cache block [%llx:%x]\n", block, offset); ++ } ++ ++ if (msblk->block_cache[i].block != block) { ++ up(&msblk->block_cache_mutex); ++ continue; ++ } ++ ++ if ((bytes = msblk->block_cache[i].length - offset) >= length) { ++ if (buffer) ++ memcpy(buffer, msblk->block_cache[i].data + ++ offset, length); ++ if (msblk->block_cache[i].length - offset == length) { ++ *next_block = msblk->block_cache[i].next_index; ++ *next_offset = 0; ++ } else { ++ *next_block = block; ++ *next_offset = offset + length; ++ } ++ up(&msblk->block_cache_mutex); ++ goto finish; ++ } else { ++ if (buffer) { ++ memcpy(buffer, msblk->block_cache[i].data + ++ offset, bytes); ++ buffer += bytes; ++ } ++ block = msblk->block_cache[i].next_index; ++ up(&msblk->block_cache_mutex); ++ length -= bytes; ++ offset = 0; ++ } ++ } ++ ++finish: ++ return return_length; ++out: ++ return 0; ++} ++ ++ ++static int get_fragment_location(struct super_block *s, unsigned int fragment, ++ long long *fragment_start_block, ++ unsigned int *fragment_size) ++{ ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ long long start_block = ++ msblk->fragment_index[SQUASHFS_FRAGMENT_INDEX(fragment)]; ++ int offset = SQUASHFS_FRAGMENT_INDEX_OFFSET(fragment); ++ struct squashfs_fragment_entry fragment_entry; ++ ++ if (msblk->swap) { ++ struct squashfs_fragment_entry sfragment_entry; ++ ++ if (!squashfs_get_cached_block(s, (char *) &sfragment_entry, ++ start_block, offset, ++ sizeof(sfragment_entry), &start_block, ++ &offset)) ++ goto out; ++ SQUASHFS_SWAP_FRAGMENT_ENTRY(&fragment_entry, &sfragment_entry); ++ } else ++ if (!squashfs_get_cached_block(s, (char *) &fragment_entry, ++ start_block, offset, ++ sizeof(fragment_entry), &start_block, ++ &offset)) ++ goto out; ++ ++ *fragment_start_block = fragment_entry.start_block; ++ *fragment_size = fragment_entry.size; ++ ++ return 1; ++ ++out: ++ return 0; ++} ++ ++ ++SQSH_EXTERN void release_cached_fragment(struct squashfs_sb_info *msblk, struct ++ squashfs_fragment_cache *fragment) ++{ ++ down(&msblk->fragment_mutex); ++ fragment->locked --; ++ wake_up(&msblk->fragment_wait_queue); ++ up(&msblk->fragment_mutex); ++} ++ ++ ++SQSH_EXTERN struct squashfs_fragment_cache *get_cached_fragment(struct super_block ++ *s, long long start_block, ++ int length) ++{ ++ int i, n; ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ ++ while ( 1 ) { ++ down(&msblk->fragment_mutex); ++ ++ for (i = 0; i < SQUASHFS_CACHED_FRAGMENTS && ++ msblk->fragment[i].block != start_block; i++); ++ ++ if (i == SQUASHFS_CACHED_FRAGMENTS) { ++ for (i = msblk->next_fragment, n = ++ SQUASHFS_CACHED_FRAGMENTS; n && ++ msblk->fragment[i].locked; n--, i = (i + 1) % ++ SQUASHFS_CACHED_FRAGMENTS); ++ ++ if (n == 0) { ++ wait_queue_t wait; ++ ++ init_waitqueue_entry(&wait, current); ++ add_wait_queue(&msblk->fragment_wait_queue, ++ &wait); ++ set_current_state(TASK_UNINTERRUPTIBLE); ++ up(&msblk->fragment_mutex); ++ schedule(); ++ set_current_state(TASK_RUNNING); ++ remove_wait_queue(&msblk->fragment_wait_queue, ++ &wait); ++ continue; ++ } ++ msblk->next_fragment = (msblk->next_fragment + 1) % ++ SQUASHFS_CACHED_FRAGMENTS; ++ ++ if (msblk->fragment[i].data == NULL) ++ if (!(msblk->fragment[i].data = SQUASHFS_ALLOC ++ (SQUASHFS_FILE_MAX_SIZE))) { ++ ERROR("Failed to allocate fragment " ++ "cache block\n"); ++ up(&msblk->fragment_mutex); ++ goto out; ++ } ++ ++ msblk->fragment[i].block = SQUASHFS_INVALID_BLK; ++ msblk->fragment[i].locked = 1; ++ up(&msblk->fragment_mutex); ++ ++ if (!(msblk->fragment[i].length = squashfs_read_data(s, ++ msblk->fragment[i].data, ++ start_block, length, NULL))) { ++ ERROR("Unable to read fragment cache block " ++ "[%llx]\n", start_block); ++ msblk->fragment[i].locked = 0; ++ goto out; ++ } ++ ++ msblk->fragment[i].block = start_block; ++ TRACE("New fragment %d, start block %lld, locked %d\n", ++ i, msblk->fragment[i].block, ++ msblk->fragment[i].locked); ++ break; ++ } ++ ++ msblk->fragment[i].locked++; ++ up(&msblk->fragment_mutex); ++ TRACE("Got fragment %d, start block %lld, locked %d\n", i, ++ msblk->fragment[i].block, ++ msblk->fragment[i].locked); ++ break; ++ } ++ ++ return &msblk->fragment[i]; ++ ++out: ++ return NULL; ++} ++ ++ ++static struct inode *squashfs_new_inode(struct super_block *s, ++ struct squashfs_base_inode_header *inodeb) ++{ ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ struct inode *i = new_inode(s); ++ ++ if (i) { ++ i->i_ino = inodeb->inode_number; ++ i->i_mtime.tv_sec = inodeb->mtime; ++ i->i_atime.tv_sec = inodeb->mtime; ++ i->i_ctime.tv_sec = inodeb->mtime; ++ i->i_uid = msblk->uid[inodeb->uid]; ++ i->i_mode = inodeb->mode; ++ i->i_size = 0; ++ if (inodeb->guid == SQUASHFS_GUIDS) ++ i->i_gid = i->i_uid; ++ else ++ i->i_gid = msblk->guid[inodeb->guid]; ++ } ++ ++ return i; ++} ++ ++ ++static struct inode *squashfs_iget(struct super_block *s, squashfs_inode_t inode) ++{ ++ struct inode *i; ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ long long block = SQUASHFS_INODE_BLK(inode) + ++ sblk->inode_table_start; ++ unsigned int offset = SQUASHFS_INODE_OFFSET(inode); ++ long long next_block; ++ unsigned int next_offset; ++ union squashfs_inode_header id, sid; ++ struct squashfs_base_inode_header *inodeb = &id.base, ++ *sinodeb = &sid.base; ++ ++ TRACE("Entered squashfs_iget\n"); ++ ++ if (msblk->swap) { ++ if (!squashfs_get_cached_block(s, (char *) sinodeb, block, ++ offset, sizeof(*sinodeb), &next_block, ++ &next_offset)) ++ goto failed_read; ++ SQUASHFS_SWAP_BASE_INODE_HEADER(inodeb, sinodeb, ++ sizeof(*sinodeb)); ++ } else ++ if (!squashfs_get_cached_block(s, (char *) inodeb, block, ++ offset, sizeof(*inodeb), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ switch(inodeb->inode_type) { ++ case SQUASHFS_FILE_TYPE: { ++ unsigned int frag_size; ++ long long frag_blk; ++ struct squashfs_reg_inode_header *inodep = &id.reg; ++ struct squashfs_reg_inode_header *sinodep = &sid.reg; ++ ++ if (msblk->swap) { ++ if (!squashfs_get_cached_block(s, (char *) ++ sinodep, block, offset, ++ sizeof(*sinodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ SQUASHFS_SWAP_REG_INODE_HEADER(inodep, sinodep); ++ } else ++ if (!squashfs_get_cached_block(s, (char *) ++ inodep, block, offset, ++ sizeof(*inodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ frag_blk = SQUASHFS_INVALID_BLK; ++ if (inodep->fragment != SQUASHFS_INVALID_FRAG && ++ !get_fragment_location(s, ++ inodep->fragment, &frag_blk, &frag_size)) ++ goto failed_read; ++ ++ if((i = squashfs_new_inode(s, inodeb)) == NULL) ++ goto failed_read1; ++ ++ i->i_nlink = 1; ++ i->i_size = inodep->file_size; ++ i->i_fop = &generic_ro_fops; ++ i->i_mode |= S_IFREG; ++ i->i_blocks = ((i->i_size - 1) >> 9) + 1; ++ SQUASHFS_I(i)->u.s1.fragment_start_block = frag_blk; ++ SQUASHFS_I(i)->u.s1.fragment_size = frag_size; ++ SQUASHFS_I(i)->u.s1.fragment_offset = inodep->offset; ++ SQUASHFS_I(i)->start_block = inodep->start_block; ++ SQUASHFS_I(i)->u.s1.block_list_start = next_block; ++ SQUASHFS_I(i)->offset = next_offset; ++ if (sblk->block_size > 4096) ++ i->i_data.a_ops = &squashfs_aops; ++ else ++ i->i_data.a_ops = &squashfs_aops_4K; ++ ++ TRACE("File inode %x:%x, start_block %llx, " ++ "block_list_start %llx, offset %x\n", ++ SQUASHFS_INODE_BLK(inode), offset, ++ inodep->start_block, next_block, ++ next_offset); ++ break; ++ } ++ case SQUASHFS_LREG_TYPE: { ++ unsigned int frag_size; ++ long long frag_blk; ++ struct squashfs_lreg_inode_header *inodep = &id.lreg; ++ struct squashfs_lreg_inode_header *sinodep = &sid.lreg; ++ ++ if (msblk->swap) { ++ if (!squashfs_get_cached_block(s, (char *) ++ sinodep, block, offset, ++ sizeof(*sinodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ SQUASHFS_SWAP_LREG_INODE_HEADER(inodep, sinodep); ++ } else ++ if (!squashfs_get_cached_block(s, (char *) ++ inodep, block, offset, ++ sizeof(*inodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ frag_blk = SQUASHFS_INVALID_BLK; ++ if (inodep->fragment != SQUASHFS_INVALID_FRAG && ++ !get_fragment_location(s, ++ inodep->fragment, &frag_blk, &frag_size)) ++ goto failed_read; ++ ++ if((i = squashfs_new_inode(s, inodeb)) == NULL) ++ goto failed_read1; ++ ++ i->i_nlink = inodep->nlink; ++ i->i_size = inodep->file_size; ++ i->i_fop = &generic_ro_fops; ++ i->i_mode |= S_IFREG; ++ i->i_blocks = ((i->i_size - 1) >> 9) + 1; ++ SQUASHFS_I(i)->u.s1.fragment_start_block = frag_blk; ++ SQUASHFS_I(i)->u.s1.fragment_size = frag_size; ++ SQUASHFS_I(i)->u.s1.fragment_offset = inodep->offset; ++ SQUASHFS_I(i)->start_block = inodep->start_block; ++ SQUASHFS_I(i)->u.s1.block_list_start = next_block; ++ SQUASHFS_I(i)->offset = next_offset; ++ if (sblk->block_size > 4096) ++ i->i_data.a_ops = &squashfs_aops; ++ else ++ i->i_data.a_ops = &squashfs_aops_4K; ++ ++ TRACE("File inode %x:%x, start_block %llx, " ++ "block_list_start %llx, offset %x\n", ++ SQUASHFS_INODE_BLK(inode), offset, ++ inodep->start_block, next_block, ++ next_offset); ++ break; ++ } ++ case SQUASHFS_DIR_TYPE: { ++ struct squashfs_dir_inode_header *inodep = &id.dir; ++ struct squashfs_dir_inode_header *sinodep = &sid.dir; ++ ++ if (msblk->swap) { ++ if (!squashfs_get_cached_block(s, (char *) ++ sinodep, block, offset, ++ sizeof(*sinodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ SQUASHFS_SWAP_DIR_INODE_HEADER(inodep, sinodep); ++ } else ++ if (!squashfs_get_cached_block(s, (char *) ++ inodep, block, offset, ++ sizeof(*inodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ if((i = squashfs_new_inode(s, inodeb)) == NULL) ++ goto failed_read1; ++ ++ i->i_nlink = inodep->nlink; ++ i->i_size = inodep->file_size; ++ i->i_op = &squashfs_dir_inode_ops; ++ i->i_fop = &squashfs_dir_ops; ++ i->i_mode |= S_IFDIR; ++ SQUASHFS_I(i)->start_block = inodep->start_block; ++ SQUASHFS_I(i)->offset = inodep->offset; ++ SQUASHFS_I(i)->u.s2.directory_index_count = 0; ++ SQUASHFS_I(i)->u.s2.parent_inode = inodep->parent_inode; ++ ++ TRACE("Directory inode %x:%x, start_block %x, offset " ++ "%x\n", SQUASHFS_INODE_BLK(inode), ++ offset, inodep->start_block, ++ inodep->offset); ++ break; ++ } ++ case SQUASHFS_LDIR_TYPE: { ++ struct squashfs_ldir_inode_header *inodep = &id.ldir; ++ struct squashfs_ldir_inode_header *sinodep = &sid.ldir; ++ ++ if (msblk->swap) { ++ if (!squashfs_get_cached_block(s, (char *) ++ sinodep, block, offset, ++ sizeof(*sinodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ SQUASHFS_SWAP_LDIR_INODE_HEADER(inodep, ++ sinodep); ++ } else ++ if (!squashfs_get_cached_block(s, (char *) ++ inodep, block, offset, ++ sizeof(*inodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ if((i = squashfs_new_inode(s, inodeb)) == NULL) ++ goto failed_read1; ++ ++ i->i_nlink = inodep->nlink; ++ i->i_size = inodep->file_size; ++ i->i_op = &squashfs_dir_inode_ops; ++ i->i_fop = &squashfs_dir_ops; ++ i->i_mode |= S_IFDIR; ++ SQUASHFS_I(i)->start_block = inodep->start_block; ++ SQUASHFS_I(i)->offset = inodep->offset; ++ SQUASHFS_I(i)->u.s2.directory_index_start = next_block; ++ SQUASHFS_I(i)->u.s2.directory_index_offset = ++ next_offset; ++ SQUASHFS_I(i)->u.s2.directory_index_count = ++ inodep->i_count; ++ SQUASHFS_I(i)->u.s2.parent_inode = inodep->parent_inode; ++ ++ TRACE("Long directory inode %x:%x, start_block %x, " ++ "offset %x\n", ++ SQUASHFS_INODE_BLK(inode), offset, ++ inodep->start_block, inodep->offset); ++ break; ++ } ++ case SQUASHFS_SYMLINK_TYPE: { ++ struct squashfs_symlink_inode_header *inodep = ++ &id.symlink; ++ struct squashfs_symlink_inode_header *sinodep = ++ &sid.symlink; ++ ++ if (msblk->swap) { ++ if (!squashfs_get_cached_block(s, (char *) ++ sinodep, block, offset, ++ sizeof(*sinodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ SQUASHFS_SWAP_SYMLINK_INODE_HEADER(inodep, ++ sinodep); ++ } else ++ if (!squashfs_get_cached_block(s, (char *) ++ inodep, block, offset, ++ sizeof(*inodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ if((i = squashfs_new_inode(s, inodeb)) == NULL) ++ goto failed_read1; ++ ++ i->i_nlink = inodep->nlink; ++ i->i_size = inodep->symlink_size; ++ i->i_op = &page_symlink_inode_operations; ++ i->i_data.a_ops = &squashfs_symlink_aops; ++ i->i_mode |= S_IFLNK; ++ SQUASHFS_I(i)->start_block = next_block; ++ SQUASHFS_I(i)->offset = next_offset; ++ ++ TRACE("Symbolic link inode %x:%x, start_block %llx, " ++ "offset %x\n", ++ SQUASHFS_INODE_BLK(inode), offset, ++ next_block, next_offset); ++ break; ++ } ++ case SQUASHFS_BLKDEV_TYPE: ++ case SQUASHFS_CHRDEV_TYPE: { ++ struct squashfs_dev_inode_header *inodep = &id.dev; ++ struct squashfs_dev_inode_header *sinodep = &sid.dev; ++ ++ if (msblk->swap) { ++ if (!squashfs_get_cached_block(s, (char *) ++ sinodep, block, offset, ++ sizeof(*sinodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ SQUASHFS_SWAP_DEV_INODE_HEADER(inodep, sinodep); ++ } else ++ if (!squashfs_get_cached_block(s, (char *) ++ inodep, block, offset, ++ sizeof(*inodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ if ((i = squashfs_new_inode(s, inodeb)) == NULL) ++ goto failed_read1; ++ ++ i->i_nlink = inodep->nlink; ++ i->i_mode |= (inodeb->inode_type == ++ SQUASHFS_CHRDEV_TYPE) ? S_IFCHR : ++ S_IFBLK; ++ init_special_inode(i, i->i_mode, ++ old_decode_dev(inodep->rdev)); ++ ++ TRACE("Device inode %x:%x, rdev %x\n", ++ SQUASHFS_INODE_BLK(inode), offset, ++ inodep->rdev); ++ break; ++ } ++ case SQUASHFS_FIFO_TYPE: ++ case SQUASHFS_SOCKET_TYPE: { ++ struct squashfs_ipc_inode_header *inodep = &id.ipc; ++ struct squashfs_ipc_inode_header *sinodep = &sid.ipc; ++ ++ if (msblk->swap) { ++ if (!squashfs_get_cached_block(s, (char *) ++ sinodep, block, offset, ++ sizeof(*sinodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ SQUASHFS_SWAP_IPC_INODE_HEADER(inodep, sinodep); ++ } else ++ if (!squashfs_get_cached_block(s, (char *) ++ inodep, block, offset, ++ sizeof(*inodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ if ((i = squashfs_new_inode(s, inodeb)) == NULL) ++ goto failed_read1; ++ ++ i->i_nlink = inodep->nlink; ++ i->i_mode |= (inodeb->inode_type == SQUASHFS_FIFO_TYPE) ++ ? S_IFIFO : S_IFSOCK; ++ init_special_inode(i, i->i_mode, 0); ++ break; ++ } ++ default: ++ ERROR("Unknown inode type %d in squashfs_iget!\n", ++ inodeb->inode_type); ++ goto failed_read1; ++ } ++ ++ insert_inode_hash(i); ++ return i; ++ ++failed_read: ++ ERROR("Unable to read inode [%llx:%x]\n", block, offset); ++ ++failed_read1: ++ return NULL; ++} ++ ++ ++static int read_fragment_index_table(struct super_block *s) ++{ ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ ++ /* Allocate fragment index table */ ++ if (!(msblk->fragment_index = kmalloc(SQUASHFS_FRAGMENT_INDEX_BYTES ++ (sblk->fragments), GFP_KERNEL))) { ++ ERROR("Failed to allocate uid/gid table\n"); ++ return 0; ++ } ++ ++ if (SQUASHFS_FRAGMENT_INDEX_BYTES(sblk->fragments) && ++ !squashfs_read_data(s, (char *) ++ msblk->fragment_index, ++ sblk->fragment_table_start, ++ SQUASHFS_FRAGMENT_INDEX_BYTES ++ (sblk->fragments) | ++ SQUASHFS_COMPRESSED_BIT_BLOCK, NULL)) { ++ ERROR("unable to read fragment index table\n"); ++ return 0; ++ } ++ ++ if (msblk->swap) { ++ int i; ++ long long fragment; ++ ++ for (i = 0; i < SQUASHFS_FRAGMENT_INDEXES(sblk->fragments); ++ i++) { ++ SQUASHFS_SWAP_FRAGMENT_INDEXES((&fragment), ++ &msblk->fragment_index[i], 1); ++ msblk->fragment_index[i] = fragment; ++ } ++ } ++ ++ return 1; ++} ++ ++ ++static int supported_squashfs_filesystem(struct squashfs_sb_info *msblk, int silent) ++{ ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ ++ msblk->iget = squashfs_iget; ++ msblk->read_blocklist = read_blocklist; ++ msblk->read_fragment_index_table = read_fragment_index_table; ++ ++ if (sblk->s_major == 1) { ++ if (!squashfs_1_0_supported(msblk)) { ++ SERROR("Major/Minor mismatch, Squashfs 1.0 filesystems " ++ "are unsupported\n"); ++ SERROR("Please recompile with " ++ "Squashfs 1.0 support enabled\n"); ++ return 0; ++ } ++ } else if (sblk->s_major == 2) { ++ if (!squashfs_2_0_supported(msblk)) { ++ SERROR("Major/Minor mismatch, Squashfs 2.0 filesystems " ++ "are unsupported\n"); ++ SERROR("Please recompile with " ++ "Squashfs 2.0 support enabled\n"); ++ return 0; ++ } ++ } else if(sblk->s_major != SQUASHFS_MAJOR || sblk->s_minor > ++ SQUASHFS_MINOR) { ++ SERROR("Major/Minor mismatch, trying to mount newer %d.%d " ++ "filesystem\n", sblk->s_major, sblk->s_minor); ++ SERROR("Please update your kernel\n"); ++ return 0; ++ } ++ ++ return 1; ++} ++ ++ ++static int squashfs_fill_super(struct super_block *s, void *data, int silent) ++{ ++ struct squashfs_sb_info *msblk; ++ struct squashfs_super_block *sblk; ++ int i; ++ char b[BDEVNAME_SIZE]; ++ struct inode *root; ++ ++ TRACE("Entered squashfs_read_superblock\n"); ++ ++ if (!(s->s_fs_info = kmalloc(sizeof(struct squashfs_sb_info), ++ GFP_KERNEL))) { ++ ERROR("Failed to allocate superblock\n"); ++ goto failure; ++ } ++ memset(s->s_fs_info, 0, sizeof(struct squashfs_sb_info)); ++ msblk = s->s_fs_info; ++ sblk = &msblk->sblk; ++ ++ msblk->devblksize = sb_min_blocksize(s, BLOCK_SIZE); ++ msblk->devblksize_log2 = ffz(~msblk->devblksize); ++ ++ init_MUTEX(&msblk->read_data_mutex); ++ init_MUTEX(&msblk->read_page_mutex); ++ init_MUTEX(&msblk->block_cache_mutex); ++ init_MUTEX(&msblk->fragment_mutex); ++ init_MUTEX(&msblk->meta_index_mutex); ++ ++ init_waitqueue_head(&msblk->waitq); ++ init_waitqueue_head(&msblk->fragment_wait_queue); ++ ++ if (!squashfs_read_data(s, (char *) sblk, SQUASHFS_START, ++ sizeof(struct squashfs_super_block) | ++ SQUASHFS_COMPRESSED_BIT_BLOCK, NULL)) { ++ SERROR("unable to read superblock\n"); ++ goto failed_mount; ++ } ++ ++ /* Check it is a SQUASHFS superblock */ ++ msblk->swap = 0; ++ if ((s->s_magic = sblk->s_magic) != SQUASHFS_MAGIC) { ++ if (sblk->s_magic == SQUASHFS_MAGIC_SWAP) { ++ struct squashfs_super_block ssblk; ++ ++ WARNING("Mounting a different endian SQUASHFS " ++ "filesystem on %s\n", bdevname(s->s_bdev, b)); ++ ++ SQUASHFS_SWAP_SUPER_BLOCK(&ssblk, sblk); ++ memcpy(sblk, &ssblk, sizeof(struct squashfs_super_block)); ++ msblk->swap = 1; ++ } else { ++ SERROR("Can't find a SQUASHFS superblock on %s\n", ++ bdevname(s->s_bdev, b)); ++ goto failed_mount; ++ } ++ } ++ ++ /* Check the MAJOR & MINOR versions */ ++ if(!supported_squashfs_filesystem(msblk, silent)) ++ goto failed_mount; ++ ++ TRACE("Found valid superblock on %s\n", bdevname(s->s_bdev, b)); ++ TRACE("Inodes are %scompressed\n", ++ SQUASHFS_UNCOMPRESSED_INODES ++ (sblk->flags) ? "un" : ""); ++ TRACE("Data is %scompressed\n", ++ SQUASHFS_UNCOMPRESSED_DATA(sblk->flags) ++ ? "un" : ""); ++ TRACE("Check data is %s present in the filesystem\n", ++ SQUASHFS_CHECK_DATA(sblk->flags) ? ++ "" : "not"); ++ TRACE("Filesystem size %lld bytes\n", sblk->bytes_used); ++ TRACE("Block size %d\n", sblk->block_size); ++ TRACE("Number of inodes %d\n", sblk->inodes); ++ if (sblk->s_major > 1) ++ TRACE("Number of fragments %d\n", sblk->fragments); ++ TRACE("Number of uids %d\n", sblk->no_uids); ++ TRACE("Number of gids %d\n", sblk->no_guids); ++ TRACE("sblk->inode_table_start %llx\n", sblk->inode_table_start); ++ TRACE("sblk->directory_table_start %llx\n", sblk->directory_table_start); ++ if (sblk->s_major > 1) ++ TRACE("sblk->fragment_table_start %llx\n", ++ sblk->fragment_table_start); ++ TRACE("sblk->uid_start %llx\n", sblk->uid_start); ++ ++ s->s_flags |= MS_RDONLY; ++ s->s_op = &squashfs_ops; ++ ++ /* Init inode_table block pointer array */ ++ if (!(msblk->block_cache = kmalloc(sizeof(struct squashfs_cache) * ++ SQUASHFS_CACHED_BLKS, GFP_KERNEL))) { ++ ERROR("Failed to allocate block cache\n"); ++ goto failed_mount; ++ } ++ ++ for (i = 0; i < SQUASHFS_CACHED_BLKS; i++) ++ msblk->block_cache[i].block = SQUASHFS_INVALID_BLK; ++ ++ msblk->next_cache = 0; ++ ++ /* Allocate read_data block */ ++ msblk->read_size = (sblk->block_size < SQUASHFS_METADATA_SIZE) ? ++ SQUASHFS_METADATA_SIZE : ++ sblk->block_size; ++ ++ if (!(msblk->read_data = kmalloc(msblk->read_size, GFP_KERNEL))) { ++ ERROR("Failed to allocate read_data block\n"); ++ goto failed_mount; ++ } ++ ++ /* Allocate read_page block */ ++ if (!(msblk->read_page = kmalloc(sblk->block_size, GFP_KERNEL))) { ++ ERROR("Failed to allocate read_page block\n"); ++ goto failed_mount; ++ } ++ ++ /* Allocate uid and gid tables */ ++ if (!(msblk->uid = kmalloc((sblk->no_uids + sblk->no_guids) * ++ sizeof(unsigned int), GFP_KERNEL))) { ++ ERROR("Failed to allocate uid/gid table\n"); ++ goto failed_mount; ++ } ++ msblk->guid = msblk->uid + sblk->no_uids; ++ ++ if (msblk->swap) { ++ unsigned int suid[sblk->no_uids + sblk->no_guids]; ++ ++ if (!squashfs_read_data(s, (char *) &suid, sblk->uid_start, ++ ((sblk->no_uids + sblk->no_guids) * ++ sizeof(unsigned int)) | ++ SQUASHFS_COMPRESSED_BIT_BLOCK, NULL)) { ++ ERROR("unable to read uid/gid table\n"); ++ goto failed_mount; ++ } ++ ++ SQUASHFS_SWAP_DATA(msblk->uid, suid, (sblk->no_uids + ++ sblk->no_guids), (sizeof(unsigned int) * 8)); ++ } else ++ if (!squashfs_read_data(s, (char *) msblk->uid, sblk->uid_start, ++ ((sblk->no_uids + sblk->no_guids) * ++ sizeof(unsigned int)) | ++ SQUASHFS_COMPRESSED_BIT_BLOCK, NULL)) { ++ ERROR("unable to read uid/gid table\n"); ++ goto failed_mount; ++ } ++ ++ ++ if (sblk->s_major == 1 && squashfs_1_0_supported(msblk)) ++ goto allocate_root; ++ ++ if (!(msblk->fragment = kmalloc(sizeof(struct squashfs_fragment_cache) * ++ SQUASHFS_CACHED_FRAGMENTS, GFP_KERNEL))) { ++ ERROR("Failed to allocate fragment block cache\n"); ++ goto failed_mount; ++ } ++ ++ for (i = 0; i < SQUASHFS_CACHED_FRAGMENTS; i++) { ++ msblk->fragment[i].locked = 0; ++ msblk->fragment[i].block = SQUASHFS_INVALID_BLK; ++ msblk->fragment[i].data = NULL; ++ } ++ ++ msblk->next_fragment = 0; ++ ++ /* Allocate fragment index table */ ++ if (msblk->read_fragment_index_table(s) == 0) ++ goto failed_mount; ++ ++allocate_root: ++ if ((root = (msblk->iget)(s, sblk->root_inode)) == NULL) ++ goto failed_mount; ++ ++ if ((s->s_root = d_alloc_root(root)) == NULL) { ++ ERROR("Root inode create failed\n"); ++ iput(root); ++ goto failed_mount; ++ } ++ ++ TRACE("Leaving squashfs_read_super\n"); ++ return 0; ++ ++failed_mount: ++ kfree(msblk->fragment_index); ++ kfree(msblk->fragment); ++ kfree(msblk->uid); ++ kfree(msblk->read_page); ++ kfree(msblk->read_data); ++ kfree(msblk->block_cache); ++ kfree(msblk->fragment_index_2); ++ kfree(s->s_fs_info); ++ s->s_fs_info = NULL; ++ return -EINVAL; ++ ++failure: ++ return -ENOMEM; ++} ++ ++ ++static int squashfs_statfs(struct dentry *dentry, struct kstatfs *buf) ++{ ++ struct squashfs_sb_info *msblk = dentry->d_inode->i_sb->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ ++ TRACE("Entered squashfs_statfs\n"); ++ ++ buf->f_type = SQUASHFS_MAGIC; ++ buf->f_bsize = sblk->block_size; ++ buf->f_blocks = ((sblk->bytes_used - 1) >> sblk->block_log) + 1; ++ buf->f_bfree = buf->f_bavail = 0; ++ buf->f_files = sblk->inodes; ++ buf->f_ffree = 0; ++ buf->f_namelen = SQUASHFS_NAME_LEN; ++ ++ return 0; ++} ++ ++ ++static int squashfs_symlink_readpage(struct file *file, struct page *page) ++{ ++ struct inode *inode = page->mapping->host; ++ int index = page->index << PAGE_CACHE_SHIFT, length, bytes; ++ long long block = SQUASHFS_I(inode)->start_block; ++ int offset = SQUASHFS_I(inode)->offset; ++ void *pageaddr = kmap(page); ++ ++ TRACE("Entered squashfs_symlink_readpage, page index %ld, start block " ++ "%llx, offset %x\n", page->index, ++ SQUASHFS_I(inode)->start_block, ++ SQUASHFS_I(inode)->offset); ++ ++ for (length = 0; length < index; length += bytes) { ++ if (!(bytes = squashfs_get_cached_block(inode->i_sb, NULL, ++ block, offset, PAGE_CACHE_SIZE, &block, ++ &offset))) { ++ ERROR("Unable to read symbolic link [%llx:%x]\n", block, ++ offset); ++ goto skip_read; ++ } ++ } ++ ++ if (length != index) { ++ ERROR("(squashfs_symlink_readpage) length != index\n"); ++ bytes = 0; ++ goto skip_read; ++ } ++ ++ bytes = (i_size_read(inode) - length) > PAGE_CACHE_SIZE ? PAGE_CACHE_SIZE : ++ i_size_read(inode) - length; ++ ++ if (!(bytes = squashfs_get_cached_block(inode->i_sb, pageaddr, block, ++ offset, bytes, &block, &offset))) ++ ERROR("Unable to read symbolic link [%llx:%x]\n", block, offset); ++ ++skip_read: ++ memset(pageaddr + bytes, 0, PAGE_CACHE_SIZE - bytes); ++ kunmap(page); ++ SetPageUptodate(page); ++ unlock_page(page); ++ ++ return 0; ++} ++ ++ ++struct meta_index *locate_meta_index(struct inode *inode, int index, int offset) ++{ ++ struct meta_index *meta = NULL; ++ struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info; ++ int i; ++ ++ down(&msblk->meta_index_mutex); ++ ++ TRACE("locate_meta_index: index %d, offset %d\n", index, offset); ++ ++ if(msblk->meta_index == NULL) ++ goto not_allocated; ++ ++ for (i = 0; i < SQUASHFS_META_NUMBER; i ++) ++ if (msblk->meta_index[i].inode_number == inode->i_ino && ++ msblk->meta_index[i].offset >= offset && ++ msblk->meta_index[i].offset <= index && ++ msblk->meta_index[i].locked == 0) { ++ TRACE("locate_meta_index: entry %d, offset %d\n", i, ++ msblk->meta_index[i].offset); ++ meta = &msblk->meta_index[i]; ++ offset = meta->offset; ++ } ++ ++ if (meta) ++ meta->locked = 1; ++ ++not_allocated: ++ up(&msblk->meta_index_mutex); ++ ++ return meta; ++} ++ ++ ++struct meta_index *empty_meta_index(struct inode *inode, int offset, int skip) ++{ ++ struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info; ++ struct meta_index *meta = NULL; ++ int i; ++ ++ down(&msblk->meta_index_mutex); ++ ++ TRACE("empty_meta_index: offset %d, skip %d\n", offset, skip); ++ ++ if(msblk->meta_index == NULL) { ++ if (!(msblk->meta_index = kmalloc(sizeof(struct meta_index) * ++ SQUASHFS_META_NUMBER, GFP_KERNEL))) { ++ ERROR("Failed to allocate meta_index\n"); ++ goto failed; ++ } ++ for(i = 0; i < SQUASHFS_META_NUMBER; i++) { ++ msblk->meta_index[i].inode_number = 0; ++ msblk->meta_index[i].locked = 0; ++ } ++ msblk->next_meta_index = 0; ++ } ++ ++ for(i = SQUASHFS_META_NUMBER; i && ++ msblk->meta_index[msblk->next_meta_index].locked; i --) ++ msblk->next_meta_index = (msblk->next_meta_index + 1) % ++ SQUASHFS_META_NUMBER; ++ ++ if(i == 0) { ++ TRACE("empty_meta_index: failed!\n"); ++ goto failed; ++ } ++ ++ TRACE("empty_meta_index: returned meta entry %d, %p\n", ++ msblk->next_meta_index, ++ &msblk->meta_index[msblk->next_meta_index]); ++ ++ meta = &msblk->meta_index[msblk->next_meta_index]; ++ msblk->next_meta_index = (msblk->next_meta_index + 1) % ++ SQUASHFS_META_NUMBER; ++ ++ meta->inode_number = inode->i_ino; ++ meta->offset = offset; ++ meta->skip = skip; ++ meta->entries = 0; ++ meta->locked = 1; ++ ++failed: ++ up(&msblk->meta_index_mutex); ++ return meta; ++} ++ ++ ++void release_meta_index(struct inode *inode, struct meta_index *meta) ++{ ++ meta->locked = 0; ++} ++ ++ ++static int read_block_index(struct super_block *s, int blocks, char *block_list, ++ long long *start_block, int *offset) ++{ ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ unsigned int *block_listp; ++ int block = 0; ++ ++ if (msblk->swap) { ++ char sblock_list[blocks << 2]; ++ ++ if (!squashfs_get_cached_block(s, sblock_list, *start_block, ++ *offset, blocks << 2, start_block, offset)) { ++ ERROR("Unable to read block list [%llx:%x]\n", ++ *start_block, *offset); ++ goto failure; ++ } ++ SQUASHFS_SWAP_INTS(((unsigned int *)block_list), ++ ((unsigned int *)sblock_list), blocks); ++ } else ++ if (!squashfs_get_cached_block(s, block_list, *start_block, ++ *offset, blocks << 2, start_block, offset)) { ++ ERROR("Unable to read block list [%llx:%x]\n", ++ *start_block, *offset); ++ goto failure; ++ } ++ ++ for (block_listp = (unsigned int *) block_list; blocks; ++ block_listp++, blocks --) ++ block += SQUASHFS_COMPRESSED_SIZE_BLOCK(*block_listp); ++ ++ return block; ++ ++failure: ++ return -1; ++} ++ ++ ++#define SIZE 256 ++ ++static inline int calculate_skip(int blocks) { ++ int skip = (blocks - 1) / ((SQUASHFS_SLOTS * SQUASHFS_META_ENTRIES + 1) * SQUASHFS_META_INDEXES); ++ return skip >= 7 ? 7 : skip + 1; ++} ++ ++ ++static int get_meta_index(struct inode *inode, int index, ++ long long *index_block, int *index_offset, ++ long long *data_block, char *block_list) ++{ ++ struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ int skip = calculate_skip(i_size_read(inode) >> sblk->block_log); ++ int offset = 0; ++ struct meta_index *meta; ++ struct meta_entry *meta_entry; ++ long long cur_index_block = SQUASHFS_I(inode)->u.s1.block_list_start; ++ int cur_offset = SQUASHFS_I(inode)->offset; ++ long long cur_data_block = SQUASHFS_I(inode)->start_block; ++ int i; ++ ++ index /= SQUASHFS_META_INDEXES * skip; ++ ++ while ( offset < index ) { ++ meta = locate_meta_index(inode, index, offset + 1); ++ ++ if (meta == NULL) { ++ if ((meta = empty_meta_index(inode, offset + 1, ++ skip)) == NULL) ++ goto all_done; ++ } else { ++ offset = index < meta->offset + meta->entries ? index : ++ meta->offset + meta->entries - 1; ++ meta_entry = &meta->meta_entry[offset - meta->offset]; ++ cur_index_block = meta_entry->index_block + sblk->inode_table_start; ++ cur_offset = meta_entry->offset; ++ cur_data_block = meta_entry->data_block; ++ TRACE("get_meta_index: offset %d, meta->offset %d, " ++ "meta->entries %d\n", offset, meta->offset, ++ meta->entries); ++ TRACE("get_meta_index: index_block 0x%llx, offset 0x%x" ++ " data_block 0x%llx\n", cur_index_block, ++ cur_offset, cur_data_block); ++ } ++ ++ for (i = meta->offset + meta->entries; i <= index && ++ i < meta->offset + SQUASHFS_META_ENTRIES; i++) { ++ int blocks = skip * SQUASHFS_META_INDEXES; ++ ++ while (blocks) { ++ int block = blocks > (SIZE >> 2) ? (SIZE >> 2) : ++ blocks; ++ int res = read_block_index(inode->i_sb, block, ++ block_list, &cur_index_block, ++ &cur_offset); ++ ++ if (res == -1) ++ goto failed; ++ ++ cur_data_block += res; ++ blocks -= block; ++ } ++ ++ meta_entry = &meta->meta_entry[i - meta->offset]; ++ meta_entry->index_block = cur_index_block - sblk->inode_table_start; ++ meta_entry->offset = cur_offset; ++ meta_entry->data_block = cur_data_block; ++ meta->entries ++; ++ offset ++; ++ } ++ ++ TRACE("get_meta_index: meta->offset %d, meta->entries %d\n", ++ meta->offset, meta->entries); ++ ++ release_meta_index(inode, meta); ++ } ++ ++all_done: ++ *index_block = cur_index_block; ++ *index_offset = cur_offset; ++ *data_block = cur_data_block; ++ ++ return offset * SQUASHFS_META_INDEXES * skip; ++ ++failed: ++ release_meta_index(inode, meta); ++ return -1; ++} ++ ++ ++static long long read_blocklist(struct inode *inode, int index, ++ int readahead_blks, char *block_list, ++ unsigned short **block_p, unsigned int *bsize) ++{ ++ long long block_ptr; ++ int offset; ++ long long block; ++ int res = get_meta_index(inode, index, &block_ptr, &offset, &block, ++ block_list); ++ ++ TRACE("read_blocklist: res %d, index %d, block_ptr 0x%llx, offset" ++ " 0x%x, block 0x%llx\n", res, index, block_ptr, offset, ++ block); ++ ++ if(res == -1) ++ goto failure; ++ ++ index -= res; ++ ++ while ( index ) { ++ int blocks = index > (SIZE >> 2) ? (SIZE >> 2) : index; ++ int res = read_block_index(inode->i_sb, blocks, block_list, ++ &block_ptr, &offset); ++ if (res == -1) ++ goto failure; ++ block += res; ++ index -= blocks; ++ } ++ ++ if (read_block_index(inode->i_sb, 1, block_list, ++ &block_ptr, &offset) == -1) ++ goto failure; ++ *bsize = *((unsigned int *) block_list); ++ ++ return block; ++ ++failure: ++ return 0; ++} ++ ++ ++static int squashfs_readpage(struct file *file, struct page *page) ++{ ++ struct inode *inode = page->mapping->host; ++ struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ unsigned char block_list[SIZE]; ++ long long block; ++ unsigned int bsize, i = 0, bytes = 0, byte_offset = 0; ++ int index = page->index >> (sblk->block_log - PAGE_CACHE_SHIFT); ++ void *pageaddr; ++ struct squashfs_fragment_cache *fragment = NULL; ++ char *data_ptr = msblk->read_page; ++ ++ int mask = (1 << (sblk->block_log - PAGE_CACHE_SHIFT)) - 1; ++ int start_index = page->index & ~mask; ++ int end_index = start_index | mask; ++ ++ TRACE("Entered squashfs_readpage, page index %lx, start block %llx\n", ++ page->index, ++ SQUASHFS_I(inode)->start_block); ++ ++ if (page->index >= ((i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> ++ PAGE_CACHE_SHIFT)) ++ goto skip_read; ++ ++ if (SQUASHFS_I(inode)->u.s1.fragment_start_block == SQUASHFS_INVALID_BLK ++ || index < (i_size_read(inode) >> ++ sblk->block_log)) { ++ if ((block = (msblk->read_blocklist)(inode, index, 1, ++ block_list, NULL, &bsize)) == 0) ++ goto skip_read; ++ ++ down(&msblk->read_page_mutex); ++ ++ if (!(bytes = squashfs_read_data(inode->i_sb, msblk->read_page, ++ block, bsize, NULL))) { ++ ERROR("Unable to read page, block %llx, size %x\n", block, ++ bsize); ++ up(&msblk->read_page_mutex); ++ goto skip_read; ++ } ++ } else { ++ if ((fragment = get_cached_fragment(inode->i_sb, ++ SQUASHFS_I(inode)-> ++ u.s1.fragment_start_block, ++ SQUASHFS_I(inode)->u.s1.fragment_size)) ++ == NULL) { ++ ERROR("Unable to read page, block %llx, size %x\n", ++ SQUASHFS_I(inode)-> ++ u.s1.fragment_start_block, ++ (int) SQUASHFS_I(inode)-> ++ u.s1.fragment_size); ++ goto skip_read; ++ } ++ bytes = SQUASHFS_I(inode)->u.s1.fragment_offset + ++ (i_size_read(inode) & (sblk->block_size ++ - 1)); ++ byte_offset = SQUASHFS_I(inode)->u.s1.fragment_offset; ++ data_ptr = fragment->data; ++ } ++ ++ for (i = start_index; i <= end_index && byte_offset < bytes; ++ i++, byte_offset += PAGE_CACHE_SIZE) { ++ struct page *push_page; ++ int available_bytes = (bytes - byte_offset) > PAGE_CACHE_SIZE ? ++ PAGE_CACHE_SIZE : bytes - byte_offset; ++ ++ TRACE("bytes %d, i %d, byte_offset %d, available_bytes %d\n", ++ bytes, i, byte_offset, available_bytes); ++ ++ if (i == page->index) { ++ pageaddr = kmap_atomic(page, KM_USER0); ++ memcpy(pageaddr, data_ptr + byte_offset, ++ available_bytes); ++ memset(pageaddr + available_bytes, 0, ++ PAGE_CACHE_SIZE - available_bytes); ++ kunmap_atomic(pageaddr, KM_USER0); ++ flush_dcache_page(page); ++ SetPageUptodate(page); ++ unlock_page(page); ++ } else if ((push_page = ++ grab_cache_page_nowait(page->mapping, i))) { ++ pageaddr = kmap_atomic(push_page, KM_USER0); ++ ++ memcpy(pageaddr, data_ptr + byte_offset, ++ available_bytes); ++ memset(pageaddr + available_bytes, 0, ++ PAGE_CACHE_SIZE - available_bytes); ++ kunmap_atomic(pageaddr, KM_USER0); ++ flush_dcache_page(push_page); ++ SetPageUptodate(push_page); ++ unlock_page(push_page); ++ page_cache_release(push_page); ++ } ++ } ++ ++ if (SQUASHFS_I(inode)->u.s1.fragment_start_block == SQUASHFS_INVALID_BLK ++ || index < (i_size_read(inode) >> ++ sblk->block_log)) ++ up(&msblk->read_page_mutex); ++ else ++ release_cached_fragment(msblk, fragment); ++ ++ return 0; ++ ++skip_read: ++ pageaddr = kmap_atomic(page, KM_USER0); ++ memset(pageaddr + bytes, 0, PAGE_CACHE_SIZE - bytes); ++ kunmap_atomic(pageaddr, KM_USER0); ++ flush_dcache_page(page); ++ SetPageUptodate(page); ++ unlock_page(page); ++ ++ return 0; ++} ++ ++ ++static int squashfs_readpage4K(struct file *file, struct page *page) ++{ ++ struct inode *inode = page->mapping->host; ++ struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ unsigned char block_list[SIZE]; ++ long long block; ++ unsigned int bsize, bytes = 0; ++ void *pageaddr; ++ ++ TRACE("Entered squashfs_readpage4K, page index %lx, start block %llx\n", ++ page->index, ++ SQUASHFS_I(inode)->start_block); ++ ++ if (page->index >= ((i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> ++ PAGE_CACHE_SHIFT)) { ++ pageaddr = kmap_atomic(page, KM_USER0); ++ goto skip_read; ++ } ++ ++ if (SQUASHFS_I(inode)->u.s1.fragment_start_block == SQUASHFS_INVALID_BLK ++ || page->index < (i_size_read(inode) >> ++ sblk->block_log)) { ++ block = (msblk->read_blocklist)(inode, page->index, 1, ++ block_list, NULL, &bsize); ++ ++ down(&msblk->read_page_mutex); ++ bytes = squashfs_read_data(inode->i_sb, msblk->read_page, block, ++ bsize, NULL); ++ pageaddr = kmap_atomic(page, KM_USER0); ++ if (bytes) ++ memcpy(pageaddr, msblk->read_page, bytes); ++ else ++ ERROR("Unable to read page, block %llx, size %x\n", ++ block, bsize); ++ up(&msblk->read_page_mutex); ++ } else { ++ struct squashfs_fragment_cache *fragment = ++ get_cached_fragment(inode->i_sb, ++ SQUASHFS_I(inode)-> ++ u.s1.fragment_start_block, ++ SQUASHFS_I(inode)-> u.s1.fragment_size); ++ pageaddr = kmap_atomic(page, KM_USER0); ++ if (fragment) { ++ bytes = i_size_read(inode) & (sblk->block_size - 1); ++ memcpy(pageaddr, fragment->data + SQUASHFS_I(inode)-> ++ u.s1.fragment_offset, bytes); ++ release_cached_fragment(msblk, fragment); ++ } else ++ ERROR("Unable to read page, block %llx, size %x\n", ++ SQUASHFS_I(inode)-> ++ u.s1.fragment_start_block, (int) ++ SQUASHFS_I(inode)-> u.s1.fragment_size); ++ } ++ ++skip_read: ++ memset(pageaddr + bytes, 0, PAGE_CACHE_SIZE - bytes); ++ kunmap_atomic(pageaddr, KM_USER0); ++ flush_dcache_page(page); ++ SetPageUptodate(page); ++ unlock_page(page); ++ ++ return 0; ++} ++ ++ ++static int get_dir_index_using_offset(struct super_block *s, long long ++ *next_block, unsigned int *next_offset, ++ long long index_start, ++ unsigned int index_offset, int i_count, ++ long long f_pos) ++{ ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ int i, length = 0; ++ struct squashfs_dir_index index; ++ ++ TRACE("Entered get_dir_index_using_offset, i_count %d, f_pos %d\n", ++ i_count, (unsigned int) f_pos); ++ ++ f_pos =- 3; ++ if (f_pos == 0) ++ goto finish; ++ ++ for (i = 0; i < i_count; i++) { ++ if (msblk->swap) { ++ struct squashfs_dir_index sindex; ++ squashfs_get_cached_block(s, (char *) &sindex, ++ index_start, index_offset, ++ sizeof(sindex), &index_start, ++ &index_offset); ++ SQUASHFS_SWAP_DIR_INDEX(&index, &sindex); ++ } else ++ squashfs_get_cached_block(s, (char *) &index, ++ index_start, index_offset, ++ sizeof(index), &index_start, ++ &index_offset); ++ ++ if (index.index > f_pos) ++ break; ++ ++ squashfs_get_cached_block(s, NULL, index_start, index_offset, ++ index.size + 1, &index_start, ++ &index_offset); ++ ++ length = index.index; ++ *next_block = index.start_block + sblk->directory_table_start; ++ } ++ ++ *next_offset = (length + *next_offset) % SQUASHFS_METADATA_SIZE; ++ ++finish: ++ return length + 3; ++} ++ ++ ++static int get_dir_index_using_name(struct super_block *s, long long ++ *next_block, unsigned int *next_offset, ++ long long index_start, ++ unsigned int index_offset, int i_count, ++ const char *name, int size) ++{ ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ int i, length = 0; ++ char buffer[sizeof(struct squashfs_dir_index) + SQUASHFS_NAME_LEN + 1]; ++ struct squashfs_dir_index *index = (struct squashfs_dir_index *) buffer; ++ char str[SQUASHFS_NAME_LEN + 1]; ++ ++ TRACE("Entered get_dir_index_using_name, i_count %d\n", i_count); ++ ++ strncpy(str, name, size); ++ str[size] = '\0'; ++ ++ for (i = 0; i < i_count; i++) { ++ if (msblk->swap) { ++ struct squashfs_dir_index sindex; ++ squashfs_get_cached_block(s, (char *) &sindex, ++ index_start, index_offset, ++ sizeof(sindex), &index_start, ++ &index_offset); ++ SQUASHFS_SWAP_DIR_INDEX(index, &sindex); ++ } else ++ squashfs_get_cached_block(s, (char *) index, ++ index_start, index_offset, ++ sizeof(struct squashfs_dir_index), ++ &index_start, &index_offset); ++ ++ squashfs_get_cached_block(s, index->name, index_start, ++ index_offset, index->size + 1, ++ &index_start, &index_offset); ++ ++ index->name[index->size + 1] = '\0'; ++ ++ if (strcmp(index->name, str) > 0) ++ break; ++ ++ length = index->index; ++ *next_block = index->start_block + sblk->directory_table_start; ++ } ++ ++ *next_offset = (length + *next_offset) % SQUASHFS_METADATA_SIZE; ++ return length + 3; ++} ++ ++ ++static int squashfs_readdir(struct file *file, void *dirent, filldir_t filldir) ++{ ++ struct inode *i = file->f_dentry->d_inode; ++ struct squashfs_sb_info *msblk = i->i_sb->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ long long next_block = SQUASHFS_I(i)->start_block + ++ sblk->directory_table_start; ++ int next_offset = SQUASHFS_I(i)->offset, length = 0, dirs_read = 0, ++ dir_count; ++ struct squashfs_dir_header dirh; ++ char buffer[sizeof(struct squashfs_dir_entry) + SQUASHFS_NAME_LEN + 1]; ++ struct squashfs_dir_entry *dire = (struct squashfs_dir_entry *) buffer; ++ ++ TRACE("Entered squashfs_readdir [%llx:%x]\n", next_block, next_offset); ++ ++ while(file->f_pos < 3) { ++ char *name; ++ int size, i_ino; ++ ++ if(file->f_pos == 0) { ++ name = "."; ++ size = 1; ++ i_ino = i->i_ino; ++ } else { ++ name = ".."; ++ size = 2; ++ i_ino = SQUASHFS_I(i)->u.s2.parent_inode; ++ } ++ TRACE("Calling filldir(%x, %s, %d, %d, %d, %d)\n", ++ (unsigned int) dirent, name, size, (int) ++ file->f_pos, i_ino, ++ squashfs_filetype_table[1]); ++ ++ if (filldir(dirent, name, size, ++ file->f_pos, i_ino, ++ squashfs_filetype_table[1]) < 0) { ++ TRACE("Filldir returned less than 0\n"); ++ goto finish; ++ } ++ file->f_pos += size; ++ dirs_read++; ++ } ++ ++ length = get_dir_index_using_offset(i->i_sb, &next_block, &next_offset, ++ SQUASHFS_I(i)->u.s2.directory_index_start, ++ SQUASHFS_I(i)->u.s2.directory_index_offset, ++ SQUASHFS_I(i)->u.s2.directory_index_count, ++ file->f_pos); ++ ++ while (length < i_size_read(i)) { ++ /* read directory header */ ++ if (msblk->swap) { ++ struct squashfs_dir_header sdirh; ++ ++ if (!squashfs_get_cached_block(i->i_sb, (char *) &sdirh, ++ next_block, next_offset, sizeof(sdirh), ++ &next_block, &next_offset)) ++ goto failed_read; ++ ++ length += sizeof(sdirh); ++ SQUASHFS_SWAP_DIR_HEADER(&dirh, &sdirh); ++ } else { ++ if (!squashfs_get_cached_block(i->i_sb, (char *) &dirh, ++ next_block, next_offset, sizeof(dirh), ++ &next_block, &next_offset)) ++ goto failed_read; ++ ++ length += sizeof(dirh); ++ } ++ ++ dir_count = dirh.count + 1; ++ while (dir_count--) { ++ if (msblk->swap) { ++ struct squashfs_dir_entry sdire; ++ if (!squashfs_get_cached_block(i->i_sb, (char *) ++ &sdire, next_block, next_offset, ++ sizeof(sdire), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ length += sizeof(sdire); ++ SQUASHFS_SWAP_DIR_ENTRY(dire, &sdire); ++ } else { ++ if (!squashfs_get_cached_block(i->i_sb, (char *) ++ dire, next_block, next_offset, ++ sizeof(*dire), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ length += sizeof(*dire); ++ } ++ ++ if (!squashfs_get_cached_block(i->i_sb, dire->name, ++ next_block, next_offset, ++ dire->size + 1, &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ length += dire->size + 1; ++ ++ if (file->f_pos >= length) ++ continue; ++ ++ dire->name[dire->size + 1] = '\0'; ++ ++ TRACE("Calling filldir(%x, %s, %d, %d, %x:%x, %d, %d)\n", ++ (unsigned int) dirent, dire->name, ++ dire->size + 1, (int) file->f_pos, ++ dirh.start_block, dire->offset, ++ dirh.inode_number + dire->inode_number, ++ squashfs_filetype_table[dire->type]); ++ ++ if (filldir(dirent, dire->name, dire->size + 1, ++ file->f_pos, ++ dirh.inode_number + dire->inode_number, ++ squashfs_filetype_table[dire->type]) ++ < 0) { ++ TRACE("Filldir returned less than 0\n"); ++ goto finish; ++ } ++ file->f_pos = length; ++ dirs_read++; ++ } ++ } ++ ++finish: ++ return dirs_read; ++ ++failed_read: ++ ERROR("Unable to read directory block [%llx:%x]\n", next_block, ++ next_offset); ++ return 0; ++} ++ ++ ++static struct dentry *squashfs_lookup(struct inode *i, struct dentry *dentry, ++ struct nameidata *nd) ++{ ++ const unsigned char *name = dentry->d_name.name; ++ int len = dentry->d_name.len; ++ struct inode *inode = NULL; ++ struct squashfs_sb_info *msblk = i->i_sb->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ long long next_block = SQUASHFS_I(i)->start_block + ++ sblk->directory_table_start; ++ int next_offset = SQUASHFS_I(i)->offset, length = 0, ++ dir_count; ++ struct squashfs_dir_header dirh; ++ char buffer[sizeof(struct squashfs_dir_entry) + SQUASHFS_NAME_LEN]; ++ struct squashfs_dir_entry *dire = (struct squashfs_dir_entry *) buffer; ++ ++ TRACE("Entered squashfs_lookup [%llx:%x]\n", next_block, next_offset); ++ ++ if (len > SQUASHFS_NAME_LEN) ++ goto exit_loop; ++ ++ length = get_dir_index_using_name(i->i_sb, &next_block, &next_offset, ++ SQUASHFS_I(i)->u.s2.directory_index_start, ++ SQUASHFS_I(i)->u.s2.directory_index_offset, ++ SQUASHFS_I(i)->u.s2.directory_index_count, name, ++ len); ++ ++ while (length < i_size_read(i)) { ++ /* read directory header */ ++ if (msblk->swap) { ++ struct squashfs_dir_header sdirh; ++ if (!squashfs_get_cached_block(i->i_sb, (char *) &sdirh, ++ next_block, next_offset, sizeof(sdirh), ++ &next_block, &next_offset)) ++ goto failed_read; ++ ++ length += sizeof(sdirh); ++ SQUASHFS_SWAP_DIR_HEADER(&dirh, &sdirh); ++ } else { ++ if (!squashfs_get_cached_block(i->i_sb, (char *) &dirh, ++ next_block, next_offset, sizeof(dirh), ++ &next_block, &next_offset)) ++ goto failed_read; ++ ++ length += sizeof(dirh); ++ } ++ ++ dir_count = dirh.count + 1; ++ while (dir_count--) { ++ if (msblk->swap) { ++ struct squashfs_dir_entry sdire; ++ if (!squashfs_get_cached_block(i->i_sb, (char *) ++ &sdire, next_block,next_offset, ++ sizeof(sdire), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ length += sizeof(sdire); ++ SQUASHFS_SWAP_DIR_ENTRY(dire, &sdire); ++ } else { ++ if (!squashfs_get_cached_block(i->i_sb, (char *) ++ dire, next_block,next_offset, ++ sizeof(*dire), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ length += sizeof(*dire); ++ } ++ ++ if (!squashfs_get_cached_block(i->i_sb, dire->name, ++ next_block, next_offset, dire->size + 1, ++ &next_block, &next_offset)) ++ goto failed_read; ++ ++ length += dire->size + 1; ++ ++ if (name[0] < dire->name[0]) ++ goto exit_loop; ++ ++ if ((len == dire->size + 1) && !strncmp(name, ++ dire->name, len)) { ++ squashfs_inode_t ino = ++ SQUASHFS_MKINODE(dirh.start_block, ++ dire->offset); ++ ++ TRACE("calling squashfs_iget for directory " ++ "entry %s, inode %x:%x, %d\n", name, ++ dirh.start_block, dire->offset, ++ dirh.inode_number + dire->inode_number); ++ ++ inode = (msblk->iget)(i->i_sb, ino); ++ ++ goto exit_loop; ++ } ++ } ++ } ++ ++exit_loop: ++ d_add(dentry, inode); ++ return ERR_PTR(0); ++ ++failed_read: ++ ERROR("Unable to read directory block [%llx:%x]\n", next_block, ++ next_offset); ++ goto exit_loop; ++} ++ ++ ++static void squashfs_put_super(struct super_block *s) ++{ ++ int i; ++ ++ if (s->s_fs_info) { ++ struct squashfs_sb_info *sbi = s->s_fs_info; ++ if (sbi->block_cache) ++ for (i = 0; i < SQUASHFS_CACHED_BLKS; i++) ++ if (sbi->block_cache[i].block != ++ SQUASHFS_INVALID_BLK) ++ kfree(sbi->block_cache[i].data); ++ if (sbi->fragment) ++ for (i = 0; i < SQUASHFS_CACHED_FRAGMENTS; i++) ++ SQUASHFS_FREE(sbi->fragment[i].data); ++ kfree(sbi->fragment); ++ kfree(sbi->block_cache); ++ kfree(sbi->read_data); ++ kfree(sbi->read_page); ++ kfree(sbi->uid); ++ kfree(sbi->fragment_index); ++ kfree(sbi->fragment_index_2); ++ kfree(sbi->meta_index); ++ kfree(s->s_fs_info); ++ s->s_fs_info = NULL; ++ } ++} ++ ++ ++static int squashfs_get_sb(struct file_system_type *fs_type, ++ int flags, const char *dev_name, void *data, ++ struct vfsmount *mnt) ++{ ++ return get_sb_bdev(fs_type, flags, dev_name, data, squashfs_fill_super, mnt); ++} ++ ++ ++static int __init init_squashfs_fs(void) ++{ ++ int err = init_inodecache(); ++ if (err) ++ goto out; ++ ++ printk(KERN_INFO "squashfs: version 3.0 (2006/03/15) " ++ "Phillip Lougher\n"); ++ ++ if (!(stream.workspace = vmalloc(zlib_inflate_workspacesize()))) { ++ ERROR("Failed to allocate zlib workspace\n"); ++ destroy_inodecache(); ++ err = -ENOMEM; ++ goto out; ++ } ++ ++ if ((err = register_filesystem(&squashfs_fs_type))) { ++ vfree(stream.workspace); ++ destroy_inodecache(); ++ } ++ ++out: ++ return err; ++} ++ ++ ++static void __exit exit_squashfs_fs(void) ++{ ++ vfree(stream.workspace); ++ unregister_filesystem(&squashfs_fs_type); ++ destroy_inodecache(); ++} ++ ++ ++static struct kmem_cache * squashfs_inode_cachep; ++ ++ ++static struct inode *squashfs_alloc_inode(struct super_block *sb) ++{ ++ struct squashfs_inode_info *ei; ++ ei = kmem_cache_alloc(squashfs_inode_cachep, GFP_KERNEL); ++ if (!ei) ++ return NULL; ++ return &ei->vfs_inode; ++} ++ ++ ++static void squashfs_destroy_inode(struct inode *inode) ++{ ++ kmem_cache_free(squashfs_inode_cachep, SQUASHFS_I(inode)); ++} ++ ++ ++static void init_once(void * foo, struct kmem_cache * cachep, unsigned long flags) ++{ ++ struct squashfs_inode_info *ei = foo; ++ ++ inode_init_once(&ei->vfs_inode); ++} ++ ++ ++static int __init init_inodecache(void) ++{ ++ squashfs_inode_cachep = kmem_cache_create("squashfs_inode_cache", ++ sizeof(struct squashfs_inode_info), ++ 0, SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT, ++ init_once); ++ if (squashfs_inode_cachep == NULL) ++ return -ENOMEM; ++ return 0; ++} ++ ++ ++static void destroy_inodecache(void) ++{ ++ kmem_cache_destroy(squashfs_inode_cachep); ++} ++ ++ ++module_init(init_squashfs_fs); ++module_exit(exit_squashfs_fs); ++MODULE_DESCRIPTION("squashfs, a compressed read-only filesystem"); ++MODULE_AUTHOR("Phillip Lougher <phillip@lougher.org.uk>"); ++MODULE_LICENSE("GPL"); +Index: linux-2.6.23/fs/squashfs/Makefile +=================================================================== +--- /dev/null 1970-01-01 00:00:00.000000000 +0000 ++++ linux-2.6.23/fs/squashfs/Makefile 2007-10-10 13:52:14.000000000 +0800 +@@ -0,0 +1,7 @@ ++# ++# Makefile for the linux squashfs routines. ++# ++ ++obj-$(CONFIG_SQUASHFS) += squashfs.o ++squashfs-y += inode.o ++squashfs-y += squashfs2_0.o +Index: linux-2.6.23/fs/squashfs/squashfs2_0.c +=================================================================== +--- /dev/null 1970-01-01 00:00:00.000000000 +0000 ++++ linux-2.6.23/fs/squashfs/squashfs2_0.c 2007-10-10 13:52:14.000000000 +0800 +@@ -0,0 +1,758 @@ ++/* ++ * Squashfs - a compressed read only filesystem for Linux ++ * ++ * Copyright (c) 2002, 2003, 2004, 2005, 2006 ++ * Phillip Lougher <phillip@lougher.org.uk> ++ * ++ * This program 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. ++ * ++ * This program 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 this program; if not, write to the Free Software ++ * Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. ++ * ++ * squashfs2_0.c ++ */ ++ ++#include <linux/types.h> ++#include <linux/squashfs_fs.h> ++#include <linux/module.h> ++#include <linux/errno.h> ++#include <linux/slab.h> ++#include <linux/fs.h> ++#include <linux/smp_lock.h> ++#include <linux/slab.h> ++#include <linux/squashfs_fs_sb.h> ++#include <linux/squashfs_fs_i.h> ++#include <linux/buffer_head.h> ++#include <linux/vfs.h> ++#include <linux/init.h> ++#include <linux/dcache.h> ++#include <linux/wait.h> ++#include <linux/zlib.h> ++#include <linux/blkdev.h> ++#include <linux/vmalloc.h> ++#include <asm/uaccess.h> ++#include <asm/semaphore.h> ++ ++#include "squashfs.h" ++static int squashfs_readdir_2(struct file *file, void *dirent, filldir_t filldir); ++static struct dentry *squashfs_lookup_2(struct inode *, struct dentry *, ++ struct nameidata *); ++ ++static struct file_operations squashfs_dir_ops_2 = { ++ .read = generic_read_dir, ++ .readdir = squashfs_readdir_2 ++}; ++ ++static struct inode_operations squashfs_dir_inode_ops_2 = { ++ .lookup = squashfs_lookup_2 ++}; ++ ++static unsigned char squashfs_filetype_table[] = { ++ DT_UNKNOWN, DT_DIR, DT_REG, DT_LNK, DT_BLK, DT_CHR, DT_FIFO, DT_SOCK ++}; ++ ++static int read_fragment_index_table_2(struct super_block *s) ++{ ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ ++ if (!(msblk->fragment_index_2 = kmalloc(SQUASHFS_FRAGMENT_INDEX_BYTES_2 ++ (sblk->fragments), GFP_KERNEL))) { ++ ERROR("Failed to allocate uid/gid table\n"); ++ return 0; ++ } ++ ++ if (SQUASHFS_FRAGMENT_INDEX_BYTES_2(sblk->fragments) && ++ !squashfs_read_data(s, (char *) ++ msblk->fragment_index_2, ++ sblk->fragment_table_start, ++ SQUASHFS_FRAGMENT_INDEX_BYTES_2 ++ (sblk->fragments) | ++ SQUASHFS_COMPRESSED_BIT_BLOCK, NULL)) { ++ ERROR("unable to read fragment index table\n"); ++ return 0; ++ } ++ ++ if (msblk->swap) { ++ int i; ++ unsigned int fragment; ++ ++ for (i = 0; i < SQUASHFS_FRAGMENT_INDEXES_2(sblk->fragments); ++ i++) { ++ SQUASHFS_SWAP_FRAGMENT_INDEXES_2((&fragment), ++ &msblk->fragment_index_2[i], 1); ++ msblk->fragment_index_2[i] = fragment; ++ } ++ } ++ ++ return 1; ++} ++ ++ ++static int get_fragment_location_2(struct super_block *s, unsigned int fragment, ++ long long *fragment_start_block, ++ unsigned int *fragment_size) ++{ ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ long long start_block = ++ msblk->fragment_index_2[SQUASHFS_FRAGMENT_INDEX_2(fragment)]; ++ int offset = SQUASHFS_FRAGMENT_INDEX_OFFSET_2(fragment); ++ struct squashfs_fragment_entry_2 fragment_entry; ++ ++ if (msblk->swap) { ++ struct squashfs_fragment_entry_2 sfragment_entry; ++ ++ if (!squashfs_get_cached_block(s, (char *) &sfragment_entry, ++ start_block, offset, ++ sizeof(sfragment_entry), &start_block, ++ &offset)) ++ goto out; ++ SQUASHFS_SWAP_FRAGMENT_ENTRY_2(&fragment_entry, &sfragment_entry); ++ } else ++ if (!squashfs_get_cached_block(s, (char *) &fragment_entry, ++ start_block, offset, ++ sizeof(fragment_entry), &start_block, ++ &offset)) ++ goto out; ++ ++ *fragment_start_block = fragment_entry.start_block; ++ *fragment_size = fragment_entry.size; ++ ++ return 1; ++ ++out: ++ return 0; ++} ++ ++ ++static struct inode *squashfs_new_inode(struct super_block *s, ++ struct squashfs_base_inode_header_2 *inodeb, unsigned int ino) ++{ ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ struct inode *i = new_inode(s); ++ ++ if (i) { ++ i->i_ino = ino; ++ i->i_mtime.tv_sec = sblk->mkfs_time; ++ i->i_atime.tv_sec = sblk->mkfs_time; ++ i->i_ctime.tv_sec = sblk->mkfs_time; ++ i->i_uid = msblk->uid[inodeb->uid]; ++ i->i_mode = inodeb->mode; ++ i->i_nlink = 1; ++ i->i_size = 0; ++ if (inodeb->guid == SQUASHFS_GUIDS) ++ i->i_gid = i->i_uid; ++ else ++ i->i_gid = msblk->guid[inodeb->guid]; ++ } ++ ++ return i; ++} ++ ++ ++static struct inode *squashfs_iget_2(struct super_block *s, squashfs_inode_t inode) ++{ ++ struct inode *i; ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ unsigned int block = SQUASHFS_INODE_BLK(inode) + ++ sblk->inode_table_start; ++ unsigned int offset = SQUASHFS_INODE_OFFSET(inode); ++ unsigned int ino = SQUASHFS_MK_VFS_INODE(block ++ - sblk->inode_table_start, offset); ++ long long next_block; ++ unsigned int next_offset; ++ union squashfs_inode_header_2 id, sid; ++ struct squashfs_base_inode_header_2 *inodeb = &id.base, ++ *sinodeb = &sid.base; ++ ++ TRACE("Entered squashfs_iget\n"); ++ ++ if (msblk->swap) { ++ if (!squashfs_get_cached_block(s, (char *) sinodeb, block, ++ offset, sizeof(*sinodeb), &next_block, ++ &next_offset)) ++ goto failed_read; ++ SQUASHFS_SWAP_BASE_INODE_HEADER_2(inodeb, sinodeb, ++ sizeof(*sinodeb)); ++ } else ++ if (!squashfs_get_cached_block(s, (char *) inodeb, block, ++ offset, sizeof(*inodeb), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ switch(inodeb->inode_type) { ++ case SQUASHFS_FILE_TYPE: { ++ struct squashfs_reg_inode_header_2 *inodep = &id.reg; ++ struct squashfs_reg_inode_header_2 *sinodep = &sid.reg; ++ long long frag_blk; ++ unsigned int frag_size; ++ ++ if (msblk->swap) { ++ if (!squashfs_get_cached_block(s, (char *) ++ sinodep, block, offset, ++ sizeof(*sinodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ SQUASHFS_SWAP_REG_INODE_HEADER_2(inodep, sinodep); ++ } else ++ if (!squashfs_get_cached_block(s, (char *) ++ inodep, block, offset, ++ sizeof(*inodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ frag_blk = SQUASHFS_INVALID_BLK; ++ if (inodep->fragment != SQUASHFS_INVALID_FRAG && ++ !get_fragment_location_2(s, ++ inodep->fragment, &frag_blk, &frag_size)) ++ goto failed_read; ++ ++ if((i = squashfs_new_inode(s, inodeb, ino)) == NULL) ++ goto failed_read1; ++ ++ i->i_size = inodep->file_size; ++ i->i_fop = &generic_ro_fops; ++ i->i_mode |= S_IFREG; ++ i->i_mtime.tv_sec = inodep->mtime; ++ i->i_atime.tv_sec = inodep->mtime; ++ i->i_ctime.tv_sec = inodep->mtime; ++ i->i_blocks = ((i->i_size - 1) >> 9) + 1; ++ i->i_blksize = PAGE_CACHE_SIZE; ++ SQUASHFS_I(i)->u.s1.fragment_start_block = frag_blk; ++ SQUASHFS_I(i)->u.s1.fragment_size = frag_size; ++ SQUASHFS_I(i)->u.s1.fragment_offset = inodep->offset; ++ SQUASHFS_I(i)->start_block = inodep->start_block; ++ SQUASHFS_I(i)->u.s1.block_list_start = next_block; ++ SQUASHFS_I(i)->offset = next_offset; ++ if (sblk->block_size > 4096) ++ i->i_data.a_ops = &squashfs_aops; ++ else ++ i->i_data.a_ops = &squashfs_aops_4K; ++ ++ TRACE("File inode %x:%x, start_block %x, " ++ "block_list_start %llx, offset %x\n", ++ SQUASHFS_INODE_BLK(inode), offset, ++ inodep->start_block, next_block, ++ next_offset); ++ break; ++ } ++ case SQUASHFS_DIR_TYPE: { ++ struct squashfs_dir_inode_header_2 *inodep = &id.dir; ++ struct squashfs_dir_inode_header_2 *sinodep = &sid.dir; ++ ++ if (msblk->swap) { ++ if (!squashfs_get_cached_block(s, (char *) ++ sinodep, block, offset, ++ sizeof(*sinodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ SQUASHFS_SWAP_DIR_INODE_HEADER_2(inodep, sinodep); ++ } else ++ if (!squashfs_get_cached_block(s, (char *) ++ inodep, block, offset, ++ sizeof(*inodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ if((i = squashfs_new_inode(s, inodeb, ino)) == NULL) ++ goto failed_read1; ++ ++ i->i_size = inodep->file_size; ++ i->i_op = &squashfs_dir_inode_ops_2; ++ i->i_fop = &squashfs_dir_ops_2; ++ i->i_mode |= S_IFDIR; ++ i->i_mtime.tv_sec = inodep->mtime; ++ i->i_atime.tv_sec = inodep->mtime; ++ i->i_ctime.tv_sec = inodep->mtime; ++ SQUASHFS_I(i)->start_block = inodep->start_block; ++ SQUASHFS_I(i)->offset = inodep->offset; ++ SQUASHFS_I(i)->u.s2.directory_index_count = 0; ++ SQUASHFS_I(i)->u.s2.parent_inode = 0; ++ ++ TRACE("Directory inode %x:%x, start_block %x, offset " ++ "%x\n", SQUASHFS_INODE_BLK(inode), ++ offset, inodep->start_block, ++ inodep->offset); ++ break; ++ } ++ case SQUASHFS_LDIR_TYPE: { ++ struct squashfs_ldir_inode_header_2 *inodep = &id.ldir; ++ struct squashfs_ldir_inode_header_2 *sinodep = &sid.ldir; ++ ++ if (msblk->swap) { ++ if (!squashfs_get_cached_block(s, (char *) ++ sinodep, block, offset, ++ sizeof(*sinodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ SQUASHFS_SWAP_LDIR_INODE_HEADER_2(inodep, ++ sinodep); ++ } else ++ if (!squashfs_get_cached_block(s, (char *) ++ inodep, block, offset, ++ sizeof(*inodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ if((i = squashfs_new_inode(s, inodeb, ino)) == NULL) ++ goto failed_read1; ++ ++ i->i_size = inodep->file_size; ++ i->i_op = &squashfs_dir_inode_ops_2; ++ i->i_fop = &squashfs_dir_ops_2; ++ i->i_mode |= S_IFDIR; ++ i->i_mtime.tv_sec = inodep->mtime; ++ i->i_atime.tv_sec = inodep->mtime; ++ i->i_ctime.tv_sec = inodep->mtime; ++ SQUASHFS_I(i)->start_block = inodep->start_block; ++ SQUASHFS_I(i)->offset = inodep->offset; ++ SQUASHFS_I(i)->u.s2.directory_index_start = next_block; ++ SQUASHFS_I(i)->u.s2.directory_index_offset = ++ next_offset; ++ SQUASHFS_I(i)->u.s2.directory_index_count = ++ inodep->i_count; ++ SQUASHFS_I(i)->u.s2.parent_inode = 0; ++ ++ TRACE("Long directory inode %x:%x, start_block %x, " ++ "offset %x\n", ++ SQUASHFS_INODE_BLK(inode), offset, ++ inodep->start_block, inodep->offset); ++ break; ++ } ++ case SQUASHFS_SYMLINK_TYPE: { ++ struct squashfs_symlink_inode_header_2 *inodep = ++ &id.symlink; ++ struct squashfs_symlink_inode_header_2 *sinodep = ++ &sid.symlink; ++ ++ if (msblk->swap) { ++ if (!squashfs_get_cached_block(s, (char *) ++ sinodep, block, offset, ++ sizeof(*sinodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ SQUASHFS_SWAP_SYMLINK_INODE_HEADER_2(inodep, ++ sinodep); ++ } else ++ if (!squashfs_get_cached_block(s, (char *) ++ inodep, block, offset, ++ sizeof(*inodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ if((i = squashfs_new_inode(s, inodeb, ino)) == NULL) ++ goto failed_read1; ++ ++ i->i_size = inodep->symlink_size; ++ i->i_op = &page_symlink_inode_operations; ++ i->i_data.a_ops = &squashfs_symlink_aops; ++ i->i_mode |= S_IFLNK; ++ SQUASHFS_I(i)->start_block = next_block; ++ SQUASHFS_I(i)->offset = next_offset; ++ ++ TRACE("Symbolic link inode %x:%x, start_block %llx, " ++ "offset %x\n", ++ SQUASHFS_INODE_BLK(inode), offset, ++ next_block, next_offset); ++ break; ++ } ++ case SQUASHFS_BLKDEV_TYPE: ++ case SQUASHFS_CHRDEV_TYPE: { ++ struct squashfs_dev_inode_header_2 *inodep = &id.dev; ++ struct squashfs_dev_inode_header_2 *sinodep = &sid.dev; ++ ++ if (msblk->swap) { ++ if (!squashfs_get_cached_block(s, (char *) ++ sinodep, block, offset, ++ sizeof(*sinodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ SQUASHFS_SWAP_DEV_INODE_HEADER_2(inodep, sinodep); ++ } else ++ if (!squashfs_get_cached_block(s, (char *) ++ inodep, block, offset, ++ sizeof(*inodep), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ if ((i = squashfs_new_inode(s, inodeb, ino)) == NULL) ++ goto failed_read1; ++ ++ i->i_mode |= (inodeb->inode_type == ++ SQUASHFS_CHRDEV_TYPE) ? S_IFCHR : ++ S_IFBLK; ++ init_special_inode(i, i->i_mode, ++ old_decode_dev(inodep->rdev)); ++ ++ TRACE("Device inode %x:%x, rdev %x\n", ++ SQUASHFS_INODE_BLK(inode), offset, ++ inodep->rdev); ++ break; ++ } ++ case SQUASHFS_FIFO_TYPE: ++ case SQUASHFS_SOCKET_TYPE: { ++ if ((i = squashfs_new_inode(s, inodeb, ino)) == NULL) ++ goto failed_read1; ++ ++ i->i_mode |= (inodeb->inode_type == SQUASHFS_FIFO_TYPE) ++ ? S_IFIFO : S_IFSOCK; ++ init_special_inode(i, i->i_mode, 0); ++ break; ++ } ++ default: ++ ERROR("Unknown inode type %d in squashfs_iget!\n", ++ inodeb->inode_type); ++ goto failed_read1; ++ } ++ ++ insert_inode_hash(i); ++ return i; ++ ++failed_read: ++ ERROR("Unable to read inode [%x:%x]\n", block, offset); ++ ++failed_read1: ++ return NULL; ++} ++ ++ ++static int get_dir_index_using_offset(struct super_block *s, long long ++ *next_block, unsigned int *next_offset, ++ long long index_start, ++ unsigned int index_offset, int i_count, ++ long long f_pos) ++{ ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ int i, length = 0; ++ struct squashfs_dir_index_2 index; ++ ++ TRACE("Entered get_dir_index_using_offset, i_count %d, f_pos %d\n", ++ i_count, (unsigned int) f_pos); ++ ++ if (f_pos == 0) ++ goto finish; ++ ++ for (i = 0; i < i_count; i++) { ++ if (msblk->swap) { ++ struct squashfs_dir_index_2 sindex; ++ squashfs_get_cached_block(s, (char *) &sindex, ++ index_start, index_offset, ++ sizeof(sindex), &index_start, ++ &index_offset); ++ SQUASHFS_SWAP_DIR_INDEX_2(&index, &sindex); ++ } else ++ squashfs_get_cached_block(s, (char *) &index, ++ index_start, index_offset, ++ sizeof(index), &index_start, ++ &index_offset); ++ ++ if (index.index > f_pos) ++ break; ++ ++ squashfs_get_cached_block(s, NULL, index_start, index_offset, ++ index.size + 1, &index_start, ++ &index_offset); ++ ++ length = index.index; ++ *next_block = index.start_block + sblk->directory_table_start; ++ } ++ ++ *next_offset = (length + *next_offset) % SQUASHFS_METADATA_SIZE; ++ ++finish: ++ return length; ++} ++ ++ ++static int get_dir_index_using_name(struct super_block *s, long long ++ *next_block, unsigned int *next_offset, ++ long long index_start, ++ unsigned int index_offset, int i_count, ++ const char *name, int size) ++{ ++ struct squashfs_sb_info *msblk = s->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ int i, length = 0; ++ char buffer[sizeof(struct squashfs_dir_index_2) + SQUASHFS_NAME_LEN + 1]; ++ struct squashfs_dir_index_2 *index = (struct squashfs_dir_index_2 *) buffer; ++ char str[SQUASHFS_NAME_LEN + 1]; ++ ++ TRACE("Entered get_dir_index_using_name, i_count %d\n", i_count); ++ ++ strncpy(str, name, size); ++ str[size] = '\0'; ++ ++ for (i = 0; i < i_count; i++) { ++ if (msblk->swap) { ++ struct squashfs_dir_index_2 sindex; ++ squashfs_get_cached_block(s, (char *) &sindex, ++ index_start, index_offset, ++ sizeof(sindex), &index_start, ++ &index_offset); ++ SQUASHFS_SWAP_DIR_INDEX_2(index, &sindex); ++ } else ++ squashfs_get_cached_block(s, (char *) index, ++ index_start, index_offset, ++ sizeof(struct squashfs_dir_index_2), ++ &index_start, &index_offset); ++ ++ squashfs_get_cached_block(s, index->name, index_start, ++ index_offset, index->size + 1, ++ &index_start, &index_offset); ++ ++ index->name[index->size + 1] = '\0'; ++ ++ if (strcmp(index->name, str) > 0) ++ break; ++ ++ length = index->index; ++ *next_block = index->start_block + sblk->directory_table_start; ++ } ++ ++ *next_offset = (length + *next_offset) % SQUASHFS_METADATA_SIZE; ++ return length; ++} ++ ++ ++static int squashfs_readdir_2(struct file *file, void *dirent, filldir_t filldir) ++{ ++ struct inode *i = file->f_dentry->d_inode; ++ struct squashfs_sb_info *msblk = i->i_sb->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ long long next_block = SQUASHFS_I(i)->start_block + ++ sblk->directory_table_start; ++ int next_offset = SQUASHFS_I(i)->offset, length = 0, dirs_read = 0, ++ dir_count; ++ struct squashfs_dir_header_2 dirh; ++ char buffer[sizeof(struct squashfs_dir_entry_2) + SQUASHFS_NAME_LEN + 1]; ++ struct squashfs_dir_entry_2 *dire = (struct squashfs_dir_entry_2 *) buffer; ++ ++ TRACE("Entered squashfs_readdir_2 [%llx:%x]\n", next_block, next_offset); ++ ++ length = get_dir_index_using_offset(i->i_sb, &next_block, &next_offset, ++ SQUASHFS_I(i)->u.s2.directory_index_start, ++ SQUASHFS_I(i)->u.s2.directory_index_offset, ++ SQUASHFS_I(i)->u.s2.directory_index_count, ++ file->f_pos); ++ ++ while (length < i_size_read(i)) { ++ /* read directory header */ ++ if (msblk->swap) { ++ struct squashfs_dir_header_2 sdirh; ++ ++ if (!squashfs_get_cached_block(i->i_sb, (char *) &sdirh, ++ next_block, next_offset, sizeof(sdirh), ++ &next_block, &next_offset)) ++ goto failed_read; ++ ++ length += sizeof(sdirh); ++ SQUASHFS_SWAP_DIR_HEADER_2(&dirh, &sdirh); ++ } else { ++ if (!squashfs_get_cached_block(i->i_sb, (char *) &dirh, ++ next_block, next_offset, sizeof(dirh), ++ &next_block, &next_offset)) ++ goto failed_read; ++ ++ length += sizeof(dirh); ++ } ++ ++ dir_count = dirh.count + 1; ++ while (dir_count--) { ++ if (msblk->swap) { ++ struct squashfs_dir_entry_2 sdire; ++ if (!squashfs_get_cached_block(i->i_sb, (char *) ++ &sdire, next_block, next_offset, ++ sizeof(sdire), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ length += sizeof(sdire); ++ SQUASHFS_SWAP_DIR_ENTRY_2(dire, &sdire); ++ } else { ++ if (!squashfs_get_cached_block(i->i_sb, (char *) ++ dire, next_block, next_offset, ++ sizeof(*dire), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ length += sizeof(*dire); ++ } ++ ++ if (!squashfs_get_cached_block(i->i_sb, dire->name, ++ next_block, next_offset, ++ dire->size + 1, &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ length += dire->size + 1; ++ ++ if (file->f_pos >= length) ++ continue; ++ ++ dire->name[dire->size + 1] = '\0'; ++ ++ TRACE("Calling filldir(%x, %s, %d, %d, %x:%x, %d)\n", ++ (unsigned int) dirent, dire->name, ++ dire->size + 1, (int) file->f_pos, ++ dirh.start_block, dire->offset, ++ squashfs_filetype_table[dire->type]); ++ ++ if (filldir(dirent, dire->name, dire->size + 1, ++ file->f_pos, SQUASHFS_MK_VFS_INODE( ++ dirh.start_block, dire->offset), ++ squashfs_filetype_table[dire->type]) ++ < 0) { ++ TRACE("Filldir returned less than 0\n"); ++ goto finish; ++ } ++ file->f_pos = length; ++ dirs_read++; ++ } ++ } ++ ++finish: ++ return dirs_read; ++ ++failed_read: ++ ERROR("Unable to read directory block [%llx:%x]\n", next_block, ++ next_offset); ++ return 0; ++} ++ ++ ++static struct dentry *squashfs_lookup_2(struct inode *i, struct dentry *dentry, ++ struct nameidata *nd) ++{ ++ const unsigned char *name = dentry->d_name.name; ++ int len = dentry->d_name.len; ++ struct inode *inode = NULL; ++ struct squashfs_sb_info *msblk = i->i_sb->s_fs_info; ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ long long next_block = SQUASHFS_I(i)->start_block + ++ sblk->directory_table_start; ++ int next_offset = SQUASHFS_I(i)->offset, length = 0, ++ dir_count; ++ struct squashfs_dir_header_2 dirh; ++ char buffer[sizeof(struct squashfs_dir_entry_2) + SQUASHFS_NAME_LEN]; ++ struct squashfs_dir_entry_2 *dire = (struct squashfs_dir_entry_2 *) buffer; ++ int sorted = sblk->s_major == 2 && sblk->s_minor >= 1; ++ ++ TRACE("Entered squashfs_lookup [%llx:%x]\n", next_block, next_offset); ++ ++ if (len > SQUASHFS_NAME_LEN) ++ goto exit_loop; ++ ++ length = get_dir_index_using_name(i->i_sb, &next_block, &next_offset, ++ SQUASHFS_I(i)->u.s2.directory_index_start, ++ SQUASHFS_I(i)->u.s2.directory_index_offset, ++ SQUASHFS_I(i)->u.s2.directory_index_count, name, ++ len); ++ ++ while (length < i_size_read(i)) { ++ /* read directory header */ ++ if (msblk->swap) { ++ struct squashfs_dir_header_2 sdirh; ++ if (!squashfs_get_cached_block(i->i_sb, (char *) &sdirh, ++ next_block, next_offset, sizeof(sdirh), ++ &next_block, &next_offset)) ++ goto failed_read; ++ ++ length += sizeof(sdirh); ++ SQUASHFS_SWAP_DIR_HEADER_2(&dirh, &sdirh); ++ } else { ++ if (!squashfs_get_cached_block(i->i_sb, (char *) &dirh, ++ next_block, next_offset, sizeof(dirh), ++ &next_block, &next_offset)) ++ goto failed_read; ++ ++ length += sizeof(dirh); ++ } ++ ++ dir_count = dirh.count + 1; ++ while (dir_count--) { ++ if (msblk->swap) { ++ struct squashfs_dir_entry_2 sdire; ++ if (!squashfs_get_cached_block(i->i_sb, (char *) ++ &sdire, next_block,next_offset, ++ sizeof(sdire), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ length += sizeof(sdire); ++ SQUASHFS_SWAP_DIR_ENTRY_2(dire, &sdire); ++ } else { ++ if (!squashfs_get_cached_block(i->i_sb, (char *) ++ dire, next_block,next_offset, ++ sizeof(*dire), &next_block, ++ &next_offset)) ++ goto failed_read; ++ ++ length += sizeof(*dire); ++ } ++ ++ if (!squashfs_get_cached_block(i->i_sb, dire->name, ++ next_block, next_offset, dire->size + 1, ++ &next_block, &next_offset)) ++ goto failed_read; ++ ++ length += dire->size + 1; ++ ++ if (sorted && name[0] < dire->name[0]) ++ goto exit_loop; ++ ++ if ((len == dire->size + 1) && !strncmp(name, ++ dire->name, len)) { ++ squashfs_inode_t ino = ++ SQUASHFS_MKINODE(dirh.start_block, ++ dire->offset); ++ ++ TRACE("calling squashfs_iget for directory " ++ "entry %s, inode %x:%x, %lld\n", name, ++ dirh.start_block, dire->offset, ino); ++ ++ inode = (msblk->iget)(i->i_sb, ino); ++ ++ goto exit_loop; ++ } ++ } ++ } ++ ++exit_loop: ++ d_add(dentry, inode); ++ return ERR_PTR(0); ++ ++failed_read: ++ ERROR("Unable to read directory block [%llx:%x]\n", next_block, ++ next_offset); ++ goto exit_loop; ++} ++ ++ ++int squashfs_2_0_supported(struct squashfs_sb_info *msblk) ++{ ++ struct squashfs_super_block *sblk = &msblk->sblk; ++ ++ msblk->iget = squashfs_iget_2; ++ msblk->read_fragment_index_table = read_fragment_index_table_2; ++ ++ sblk->bytes_used = sblk->bytes_used_2; ++ sblk->uid_start = sblk->uid_start_2; ++ sblk->guid_start = sblk->guid_start_2; ++ sblk->inode_table_start = sblk->inode_table_start_2; ++ sblk->directory_table_start = sblk->directory_table_start_2; ++ sblk->fragment_table_start = sblk->fragment_table_start_2; ++ ++ return 1; ++} +Index: linux-2.6.23/fs/squashfs/squashfs.h +=================================================================== +--- /dev/null 1970-01-01 00:00:00.000000000 +0000 ++++ linux-2.6.23/fs/squashfs/squashfs.h 2007-10-10 13:52:14.000000000 +0800 +@@ -0,0 +1,86 @@ ++/* ++ * Squashfs - a compressed read only filesystem for Linux ++ * ++ * Copyright (c) 2002, 2003, 2004, 2005, 2006 ++ * Phillip Lougher <phillip@lougher.org.uk> ++ * ++ * This program 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. ++ * ++ * This program 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 this program; if not, write to the Free Software ++ * Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. ++ * ++ * squashfs.h ++ */ ++ ++#ifdef CONFIG_SQUASHFS_1_0_COMPATIBILITY ++#undef CONFIG_SQUASHFS_1_0_COMPATIBILITY ++#endif ++ ++#ifdef SQUASHFS_TRACE ++#define TRACE(s, args...) printk(KERN_NOTICE "SQUASHFS: "s, ## args) ++#else ++#define TRACE(s, args...) {} ++#endif ++ ++#define ERROR(s, args...) printk(KERN_ERR "SQUASHFS error: "s, ## args) ++ ++#define SERROR(s, args...) do { \ ++ if (!silent) \ ++ printk(KERN_ERR "SQUASHFS error: "s, ## args);\ ++ } while(0) ++ ++#define WARNING(s, args...) printk(KERN_WARNING "SQUASHFS: "s, ## args) ++ ++static inline struct squashfs_inode_info *SQUASHFS_I(struct inode *inode) ++{ ++ return list_entry(inode, struct squashfs_inode_info, vfs_inode); ++} ++ ++#if defined(CONFIG_SQUASHFS_1_0_COMPATIBILITY ) || defined(CONFIG_SQUASHFS_2_0_COMPATIBILITY) ++#define SQSH_EXTERN ++extern unsigned int squashfs_read_data(struct super_block *s, char *buffer, ++ long long index, unsigned int length, ++ long long *next_index); ++extern int squashfs_get_cached_block(struct super_block *s, char *buffer, ++ long long block, unsigned int offset, ++ int length, long long *next_block, ++ unsigned int *next_offset); ++extern void release_cached_fragment(struct squashfs_sb_info *msblk, struct ++ squashfs_fragment_cache *fragment); ++extern struct squashfs_fragment_cache *get_cached_fragment(struct super_block ++ *s, long long start_block, ++ int length); ++extern struct address_space_operations squashfs_symlink_aops; ++extern struct address_space_operations squashfs_aops; ++extern struct address_space_operations squashfs_aops_4K; ++extern struct inode_operations squashfs_dir_inode_ops; ++#else ++#define SQSH_EXTERN static ++#endif ++ ++#ifdef CONFIG_SQUASHFS_1_0_COMPATIBILITY ++extern int squashfs_1_0_supported(struct squashfs_sb_info *msblk); ++#else ++static inline int squashfs_1_0_supported(struct squashfs_sb_info *msblk) ++{ ++ return 0; ++} ++#endif ++ ++#ifdef CONFIG_SQUASHFS_2_0_COMPATIBILITY ++extern int squashfs_2_0_supported(struct squashfs_sb_info *msblk); ++#else ++static inline int squashfs_2_0_supported(struct squashfs_sb_info *msblk) ++{ ++ return 0; ++} ++#endif +Index: linux-2.6.23/include/linux/magic.h +=================================================================== +--- linux-2.6.23.orig/include/linux/magic.h 2007-10-10 13:52:12.000000000 +0800 ++++ linux-2.6.23/include/linux/magic.h 2007-10-10 13:52:14.000000000 +0800 +@@ -35,6 +35,9 @@ + #define REISER2FS_SUPER_MAGIC_STRING "ReIsEr2Fs" + #define REISER2FS_JR_SUPER_MAGIC_STRING "ReIsEr3Fs" + ++#define SQUASHFS_MAGIC 0x73717368 ++#define SQUASHFS_MAGIC_SWAP 0x68737173 ++ + #define SMB_SUPER_MAGIC 0x517B + #define USBDEVICE_SUPER_MAGIC 0x9fa2 + +Index: linux-2.6.23/include/linux/squashfs_fs.h +=================================================================== +--- /dev/null 1970-01-01 00:00:00.000000000 +0000 ++++ linux-2.6.23/include/linux/squashfs_fs.h 2007-10-10 13:52:14.000000000 +0800 +@@ -0,0 +1,911 @@ ++#ifndef SQUASHFS_FS ++#define SQUASHFS_FS ++ ++/* ++ * Squashfs ++ * ++ * Copyright (c) 2002, 2003, 2004, 2005, 2006 ++ * Phillip Lougher <phillip@lougher.org.uk> ++ * ++ * This program 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. ++ * ++ * This program 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 this program; if not, write to the Free Software ++ * Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. ++ * ++ * squashfs_fs.h ++ */ ++ ++#ifndef CONFIG_SQUASHFS_2_0_COMPATIBILITY ++#define CONFIG_SQUASHFS_2_0_COMPATIBILITY ++#endif ++ ++#ifdef CONFIG_SQUASHFS_VMALLOC ++#define SQUASHFS_ALLOC(a) vmalloc(a) ++#define SQUASHFS_FREE(a) vfree(a) ++#else ++#define SQUASHFS_ALLOC(a) kmalloc(a, GFP_KERNEL) ++#define SQUASHFS_FREE(a) kfree(a) ++#endif ++#define SQUASHFS_CACHED_FRAGMENTS CONFIG_SQUASHFS_FRAGMENT_CACHE_SIZE ++#define SQUASHFS_MAJOR 3 ++#define SQUASHFS_MINOR 0 ++#define SQUASHFS_START 0 ++ ++/* size of metadata (inode and directory) blocks */ ++#define SQUASHFS_METADATA_SIZE 8192 ++#define SQUASHFS_METADATA_LOG 13 ++ ++/* default size of data blocks */ ++#define SQUASHFS_FILE_SIZE 65536 ++#define SQUASHFS_FILE_LOG 16 ++ ++#define SQUASHFS_FILE_MAX_SIZE 65536 ++ ++/* Max number of uids and gids */ ++#define SQUASHFS_UIDS 256 ++#define SQUASHFS_GUIDS 255 ++ ++/* Max length of filename (not 255) */ ++#define SQUASHFS_NAME_LEN 256 ++ ++#define SQUASHFS_INVALID ((long long) 0xffffffffffff) ++#define SQUASHFS_INVALID_FRAG ((unsigned int) 0xffffffff) ++#define SQUASHFS_INVALID_BLK ((long long) -1) ++#define SQUASHFS_USED_BLK ((long long) -2) ++ ++/* Filesystem flags */ ++#define SQUASHFS_NOI 0 ++#define SQUASHFS_NOD 1 ++#define SQUASHFS_CHECK 2 ++#define SQUASHFS_NOF 3 ++#define SQUASHFS_NO_FRAG 4 ++#define SQUASHFS_ALWAYS_FRAG 5 ++#define SQUASHFS_DUPLICATE 6 ++ ++#define SQUASHFS_BIT(flag, bit) ((flag >> bit) & 1) ++ ++#define SQUASHFS_UNCOMPRESSED_INODES(flags) SQUASHFS_BIT(flags, \ ++ SQUASHFS_NOI) ++ ++#define SQUASHFS_UNCOMPRESSED_DATA(flags) SQUASHFS_BIT(flags, \ ++ SQUASHFS_NOD) ++ ++#define SQUASHFS_UNCOMPRESSED_FRAGMENTS(flags) SQUASHFS_BIT(flags, \ ++ SQUASHFS_NOF) ++ ++#define SQUASHFS_NO_FRAGMENTS(flags) SQUASHFS_BIT(flags, \ ++ SQUASHFS_NO_FRAG) ++ ++#define SQUASHFS_ALWAYS_FRAGMENTS(flags) SQUASHFS_BIT(flags, \ ++ SQUASHFS_ALWAYS_FRAG) ++ ++#define SQUASHFS_DUPLICATES(flags) SQUASHFS_BIT(flags, \ ++ SQUASHFS_DUPLICATE) ++ ++#define SQUASHFS_CHECK_DATA(flags) SQUASHFS_BIT(flags, \ ++ SQUASHFS_CHECK) ++ ++#define SQUASHFS_MKFLAGS(noi, nod, check_data, nof, no_frag, always_frag, \ ++ duplicate_checking) (noi | (nod << 1) | (check_data << 2) \ ++ | (nof << 3) | (no_frag << 4) | (always_frag << 5) | \ ++ (duplicate_checking << 6)) ++ ++/* Max number of types and file types */ ++#define SQUASHFS_DIR_TYPE 1 ++#define SQUASHFS_FILE_TYPE 2 ++#define SQUASHFS_SYMLINK_TYPE 3 ++#define SQUASHFS_BLKDEV_TYPE 4 ++#define SQUASHFS_CHRDEV_TYPE 5 ++#define SQUASHFS_FIFO_TYPE 6 ++#define SQUASHFS_SOCKET_TYPE 7 ++#define SQUASHFS_LDIR_TYPE 8 ++#define SQUASHFS_LREG_TYPE 9 ++ ++/* 1.0 filesystem type definitions */ ++#define SQUASHFS_TYPES 5 ++#define SQUASHFS_IPC_TYPE 0 ++ ++/* Flag whether block is compressed or uncompressed, bit is set if block is ++ * uncompressed */ ++#define SQUASHFS_COMPRESSED_BIT (1 << 15) ++ ++#define SQUASHFS_COMPRESSED_SIZE(B) (((B) & ~SQUASHFS_COMPRESSED_BIT) ? \ ++ (B) & ~SQUASHFS_COMPRESSED_BIT : SQUASHFS_COMPRESSED_BIT) ++ ++#define SQUASHFS_COMPRESSED(B) (!((B) & SQUASHFS_COMPRESSED_BIT)) ++ ++#define SQUASHFS_COMPRESSED_BIT_BLOCK (1 << 24) ++ ++#define SQUASHFS_COMPRESSED_SIZE_BLOCK(B) (((B) & \ ++ ~SQUASHFS_COMPRESSED_BIT_BLOCK) ? (B) & \ ++ ~SQUASHFS_COMPRESSED_BIT_BLOCK : SQUASHFS_COMPRESSED_BIT_BLOCK) ++ ++#define SQUASHFS_COMPRESSED_BLOCK(B) (!((B) & SQUASHFS_COMPRESSED_BIT_BLOCK)) ++ ++/* ++ * Inode number ops. Inodes consist of a compressed block number, and an ++ * uncompressed offset within that block ++ */ ++#define SQUASHFS_INODE_BLK(a) ((unsigned int) ((a) >> 16)) ++ ++#define SQUASHFS_INODE_OFFSET(a) ((unsigned int) ((a) & 0xffff)) ++ ++#define SQUASHFS_MKINODE(A, B) ((squashfs_inode_t)(((squashfs_inode_t) (A)\ ++ << 16) + (B))) ++ ++/* Compute 32 bit VFS inode number from squashfs inode number */ ++#define SQUASHFS_MK_VFS_INODE(a, b) ((unsigned int) (((a) << 8) + \ ++ ((b) >> 2) + 1)) ++/* XXX */ ++ ++/* Translate between VFS mode and squashfs mode */ ++#define SQUASHFS_MODE(a) ((a) & 0xfff) ++ ++/* fragment and fragment table defines */ ++#define SQUASHFS_FRAGMENT_BYTES(A) (A * sizeof(struct squashfs_fragment_entry)) ++ ++#define SQUASHFS_FRAGMENT_INDEX(A) (SQUASHFS_FRAGMENT_BYTES(A) / \ ++ SQUASHFS_METADATA_SIZE) ++ ++#define SQUASHFS_FRAGMENT_INDEX_OFFSET(A) (SQUASHFS_FRAGMENT_BYTES(A) % \ ++ SQUASHFS_METADATA_SIZE) ++ ++#define SQUASHFS_FRAGMENT_INDEXES(A) ((SQUASHFS_FRAGMENT_BYTES(A) + \ ++ SQUASHFS_METADATA_SIZE - 1) / \ ++ SQUASHFS_METADATA_SIZE) ++ ++#define SQUASHFS_FRAGMENT_INDEX_BYTES(A) (SQUASHFS_FRAGMENT_INDEXES(A) *\ ++ sizeof(long long)) ++ ++/* cached data constants for filesystem */ ++#define SQUASHFS_CACHED_BLKS 8 ++ ++#define SQUASHFS_MAX_FILE_SIZE_LOG 64 ++ ++#define SQUASHFS_MAX_FILE_SIZE ((long long) 1 << \ ++ (SQUASHFS_MAX_FILE_SIZE_LOG - 2)) ++ ++#define SQUASHFS_MARKER_BYTE 0xff ++ ++/* meta index cache */ ++#define SQUASHFS_META_INDEXES (SQUASHFS_METADATA_SIZE / sizeof(unsigned int)) ++#define SQUASHFS_META_ENTRIES 31 ++#define SQUASHFS_META_NUMBER 8 ++#define SQUASHFS_SLOTS 4 ++ ++#include <linux/magic.h> ++ ++struct meta_entry { ++ long long data_block; ++ unsigned int index_block; ++ unsigned short offset; ++ unsigned short pad; ++}; ++ ++struct meta_index { ++ unsigned int inode_number; ++ unsigned int offset; ++ unsigned short entries; ++ unsigned short skip; ++ unsigned short locked; ++ unsigned short pad; ++ struct meta_entry meta_entry[SQUASHFS_META_ENTRIES]; ++}; ++ ++ ++/* ++ * definitions for structures on disk ++ */ ++ ++typedef long long squashfs_block_t; ++typedef long long squashfs_inode_t; ++ ++struct squashfs_super_block { ++ unsigned int s_magic; ++ unsigned int inodes; ++ unsigned int bytes_used_2; ++ unsigned int uid_start_2; ++ unsigned int guid_start_2; ++ unsigned int inode_table_start_2; ++ unsigned int directory_table_start_2; ++ unsigned int s_major:16; ++ unsigned int s_minor:16; ++ unsigned int block_size_1:16; ++ unsigned int block_log:16; ++ unsigned int flags:8; ++ unsigned int no_uids:8; ++ unsigned int no_guids:8; ++ unsigned int mkfs_time /* time of filesystem creation */; ++ squashfs_inode_t root_inode; ++ unsigned int block_size; ++ unsigned int fragments; ++ unsigned int fragment_table_start_2; ++ long long bytes_used; ++ long long uid_start; ++ long long guid_start; ++ long long inode_table_start; ++ long long directory_table_start; ++ long long fragment_table_start; ++ long long unused; ++} __attribute__ ((packed)); ++ ++struct squashfs_dir_index { ++ unsigned int index; ++ unsigned int start_block; ++ unsigned char size; ++ unsigned char name[0]; ++} __attribute__ ((packed)); ++ ++#define SQUASHFS_BASE_INODE_HEADER \ ++ unsigned int inode_type:4; \ ++ unsigned int mode:12; \ ++ unsigned int uid:8; \ ++ unsigned int guid:8; \ ++ unsigned int mtime; \ ++ unsigned int inode_number; ++ ++struct squashfs_base_inode_header { ++ SQUASHFS_BASE_INODE_HEADER; ++} __attribute__ ((packed)); ++ ++struct squashfs_ipc_inode_header { ++ SQUASHFS_BASE_INODE_HEADER; ++ unsigned int nlink; ++} __attribute__ ((packed)); ++ ++struct squashfs_dev_inode_header { ++ SQUASHFS_BASE_INODE_HEADER; ++ unsigned int nlink; ++ unsigned short rdev; ++} __attribute__ ((packed)); ++ ++struct squashfs_symlink_inode_header { ++ SQUASHFS_BASE_INODE_HEADER; ++ unsigned int nlink; ++ unsigned short symlink_size; ++ char symlink[0]; ++} __attribute__ ((packed)); ++ ++struct squashfs_reg_inode_header { ++ SQUASHFS_BASE_INODE_HEADER; ++ squashfs_block_t start_block; ++ unsigned int fragment; ++ unsigned int offset; ++ unsigned int file_size; ++ unsigned short block_list[0]; ++} __attribute__ ((packed)); ++ ++struct squashfs_lreg_inode_header { ++ SQUASHFS_BASE_INODE_HEADER; ++ unsigned int nlink; ++ squashfs_block_t start_block; ++ unsigned int fragment; ++ unsigned int offset; ++ long long file_size; ++ unsigned short block_list[0]; ++} __attribute__ ((packed)); ++ ++struct squashfs_dir_inode_header { ++ SQUASHFS_BASE_INODE_HEADER; ++ unsigned int nlink; ++ unsigned int file_size:19; ++ unsigned int offset:13; ++ unsigned int start_block; ++ unsigned int parent_inode; ++} __attribute__ ((packed)); ++ ++struct squashfs_ldir_inode_header { ++ SQUASHFS_BASE_INODE_HEADER; ++ unsigned int nlink; ++ unsigned int file_size:27; ++ unsigned int offset:13; ++ unsigned int start_block; ++ unsigned int i_count:16; ++ unsigned int parent_inode; ++ struct squashfs_dir_index index[0]; ++} __attribute__ ((packed)); ++ ++union squashfs_inode_header { ++ struct squashfs_base_inode_header base; ++ struct squashfs_dev_inode_header dev; ++ struct squashfs_symlink_inode_header symlink; ++ struct squashfs_reg_inode_header reg; ++ struct squashfs_lreg_inode_header lreg; ++ struct squashfs_dir_inode_header dir; ++ struct squashfs_ldir_inode_header ldir; ++ struct squashfs_ipc_inode_header ipc; ++}; ++ ++struct squashfs_dir_entry { ++ unsigned int offset:13; ++ unsigned int type:3; ++ unsigned int size:8; ++ int inode_number:16; ++ char name[0]; ++} __attribute__ ((packed)); ++ ++struct squashfs_dir_header { ++ unsigned int count:8; ++ unsigned int start_block; ++ unsigned int inode_number; ++} __attribute__ ((packed)); ++ ++struct squashfs_fragment_entry { ++ long long start_block; ++ unsigned int size; ++ unsigned int unused; ++} __attribute__ ((packed)); ++ ++extern int squashfs_uncompress_block(void *d, int dstlen, void *s, int srclen); ++extern int squashfs_uncompress_init(void); ++extern int squashfs_uncompress_exit(void); ++ ++/* ++ * macros to convert each packed bitfield structure from little endian to big ++ * endian and vice versa. These are needed when creating or using a filesystem ++ * on a machine with different byte ordering to the target architecture. ++ * ++ */ ++ ++#define SQUASHFS_SWAP_START \ ++ int bits;\ ++ int b_pos;\ ++ unsigned long long val;\ ++ unsigned char *s;\ ++ unsigned char *d; ++ ++#define SQUASHFS_SWAP_SUPER_BLOCK(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_MEMSET(s, d, sizeof(struct squashfs_super_block));\ ++ SQUASHFS_SWAP((s)->s_magic, d, 0, 32);\ ++ SQUASHFS_SWAP((s)->inodes, d, 32, 32);\ ++ SQUASHFS_SWAP((s)->bytes_used_2, d, 64, 32);\ ++ SQUASHFS_SWAP((s)->uid_start_2, d, 96, 32);\ ++ SQUASHFS_SWAP((s)->guid_start_2, d, 128, 32);\ ++ SQUASHFS_SWAP((s)->inode_table_start_2, d, 160, 32);\ ++ SQUASHFS_SWAP((s)->directory_table_start_2, d, 192, 32);\ ++ SQUASHFS_SWAP((s)->s_major, d, 224, 16);\ ++ SQUASHFS_SWAP((s)->s_minor, d, 240, 16);\ ++ SQUASHFS_SWAP((s)->block_size_1, d, 256, 16);\ ++ SQUASHFS_SWAP((s)->block_log, d, 272, 16);\ ++ SQUASHFS_SWAP((s)->flags, d, 288, 8);\ ++ SQUASHFS_SWAP((s)->no_uids, d, 296, 8);\ ++ SQUASHFS_SWAP((s)->no_guids, d, 304, 8);\ ++ SQUASHFS_SWAP((s)->mkfs_time, d, 312, 32);\ ++ SQUASHFS_SWAP((s)->root_inode, d, 344, 64);\ ++ SQUASHFS_SWAP((s)->block_size, d, 408, 32);\ ++ SQUASHFS_SWAP((s)->fragments, d, 440, 32);\ ++ SQUASHFS_SWAP((s)->fragment_table_start_2, d, 472, 32);\ ++ SQUASHFS_SWAP((s)->bytes_used, d, 504, 64);\ ++ SQUASHFS_SWAP((s)->uid_start, d, 568, 64);\ ++ SQUASHFS_SWAP((s)->guid_start, d, 632, 64);\ ++ SQUASHFS_SWAP((s)->inode_table_start, d, 696, 64);\ ++ SQUASHFS_SWAP((s)->directory_table_start, d, 760, 64);\ ++ SQUASHFS_SWAP((s)->fragment_table_start, d, 824, 64);\ ++ SQUASHFS_SWAP((s)->unused, d, 888, 64);\ ++} ++ ++#define SQUASHFS_SWAP_BASE_INODE_CORE(s, d, n)\ ++ SQUASHFS_MEMSET(s, d, n);\ ++ SQUASHFS_SWAP((s)->inode_type, d, 0, 4);\ ++ SQUASHFS_SWAP((s)->mode, d, 4, 12);\ ++ SQUASHFS_SWAP((s)->uid, d, 16, 8);\ ++ SQUASHFS_SWAP((s)->guid, d, 24, 8);\ ++ SQUASHFS_SWAP((s)->mtime, d, 32, 32);\ ++ SQUASHFS_SWAP((s)->inode_number, d, 64, 32); ++ ++#define SQUASHFS_SWAP_BASE_INODE_HEADER(s, d, n) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE(s, d, n)\ ++} ++ ++#define SQUASHFS_SWAP_IPC_INODE_HEADER(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE(s, d, \ ++ sizeof(struct squashfs_ipc_inode_header))\ ++ SQUASHFS_SWAP((s)->nlink, d, 96, 32);\ ++} ++ ++#define SQUASHFS_SWAP_DEV_INODE_HEADER(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE(s, d, \ ++ sizeof(struct squashfs_dev_inode_header)); \ ++ SQUASHFS_SWAP((s)->nlink, d, 96, 32);\ ++ SQUASHFS_SWAP((s)->rdev, d, 128, 16);\ ++} ++ ++#define SQUASHFS_SWAP_SYMLINK_INODE_HEADER(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE(s, d, \ ++ sizeof(struct squashfs_symlink_inode_header));\ ++ SQUASHFS_SWAP((s)->nlink, d, 96, 32);\ ++ SQUASHFS_SWAP((s)->symlink_size, d, 128, 16);\ ++} ++ ++#define SQUASHFS_SWAP_REG_INODE_HEADER(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE(s, d, \ ++ sizeof(struct squashfs_reg_inode_header));\ ++ SQUASHFS_SWAP((s)->start_block, d, 96, 64);\ ++ SQUASHFS_SWAP((s)->fragment, d, 160, 32);\ ++ SQUASHFS_SWAP((s)->offset, d, 192, 32);\ ++ SQUASHFS_SWAP((s)->file_size, d, 224, 32);\ ++} ++ ++#define SQUASHFS_SWAP_LREG_INODE_HEADER(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE(s, d, \ ++ sizeof(struct squashfs_lreg_inode_header));\ ++ SQUASHFS_SWAP((s)->nlink, d, 96, 32);\ ++ SQUASHFS_SWAP((s)->start_block, d, 128, 64);\ ++ SQUASHFS_SWAP((s)->fragment, d, 192, 32);\ ++ SQUASHFS_SWAP((s)->offset, d, 224, 32);\ ++ SQUASHFS_SWAP((s)->file_size, d, 256, 64);\ ++} ++ ++#define SQUASHFS_SWAP_DIR_INODE_HEADER(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE(s, d, \ ++ sizeof(struct squashfs_dir_inode_header));\ ++ SQUASHFS_SWAP((s)->nlink, d, 96, 32);\ ++ SQUASHFS_SWAP((s)->file_size, d, 128, 19);\ ++ SQUASHFS_SWAP((s)->offset, d, 147, 13);\ ++ SQUASHFS_SWAP((s)->start_block, d, 160, 32);\ ++ SQUASHFS_SWAP((s)->parent_inode, d, 192, 32);\ ++} ++ ++#define SQUASHFS_SWAP_LDIR_INODE_HEADER(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE(s, d, \ ++ sizeof(struct squashfs_ldir_inode_header));\ ++ SQUASHFS_SWAP((s)->nlink, d, 96, 32);\ ++ SQUASHFS_SWAP((s)->file_size, d, 128, 27);\ ++ SQUASHFS_SWAP((s)->offset, d, 155, 13);\ ++ SQUASHFS_SWAP((s)->start_block, d, 168, 32);\ ++ SQUASHFS_SWAP((s)->i_count, d, 200, 16);\ ++ SQUASHFS_SWAP((s)->parent_inode, d, 216, 32);\ ++} ++ ++#define SQUASHFS_SWAP_DIR_INDEX(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_MEMSET(s, d, sizeof(struct squashfs_dir_index));\ ++ SQUASHFS_SWAP((s)->index, d, 0, 32);\ ++ SQUASHFS_SWAP((s)->start_block, d, 32, 32);\ ++ SQUASHFS_SWAP((s)->size, d, 64, 8);\ ++} ++ ++#define SQUASHFS_SWAP_DIR_HEADER(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_MEMSET(s, d, sizeof(struct squashfs_dir_header));\ ++ SQUASHFS_SWAP((s)->count, d, 0, 8);\ ++ SQUASHFS_SWAP((s)->start_block, d, 8, 32);\ ++ SQUASHFS_SWAP((s)->inode_number, d, 40, 32);\ ++} ++ ++#define SQUASHFS_SWAP_DIR_ENTRY(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_MEMSET(s, d, sizeof(struct squashfs_dir_entry));\ ++ SQUASHFS_SWAP((s)->offset, d, 0, 13);\ ++ SQUASHFS_SWAP((s)->type, d, 13, 3);\ ++ SQUASHFS_SWAP((s)->size, d, 16, 8);\ ++ SQUASHFS_SWAP((s)->inode_number, d, 24, 16);\ ++} ++ ++#define SQUASHFS_SWAP_FRAGMENT_ENTRY(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_MEMSET(s, d, sizeof(struct squashfs_fragment_entry));\ ++ SQUASHFS_SWAP((s)->start_block, d, 0, 64);\ ++ SQUASHFS_SWAP((s)->size, d, 64, 32);\ ++} ++ ++#define SQUASHFS_SWAP_SHORTS(s, d, n) {\ ++ int entry;\ ++ int bit_position;\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_MEMSET(s, d, n * 2);\ ++ for(entry = 0, bit_position = 0; entry < n; entry++, bit_position += \ ++ 16)\ ++ SQUASHFS_SWAP(s[entry], d, bit_position, 16);\ ++} ++ ++#define SQUASHFS_SWAP_INTS(s, d, n) {\ ++ int entry;\ ++ int bit_position;\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_MEMSET(s, d, n * 4);\ ++ for(entry = 0, bit_position = 0; entry < n; entry++, bit_position += \ ++ 32)\ ++ SQUASHFS_SWAP(s[entry], d, bit_position, 32);\ ++} ++ ++#define SQUASHFS_SWAP_LONG_LONGS(s, d, n) {\ ++ int entry;\ ++ int bit_position;\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_MEMSET(s, d, n * 8);\ ++ for(entry = 0, bit_position = 0; entry < n; entry++, bit_position += \ ++ 64)\ ++ SQUASHFS_SWAP(s[entry], d, bit_position, 64);\ ++} ++ ++#define SQUASHFS_SWAP_DATA(s, d, n, bits) {\ ++ int entry;\ ++ int bit_position;\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_MEMSET(s, d, n * bits / 8);\ ++ for(entry = 0, bit_position = 0; entry < n; entry++, bit_position += \ ++ bits)\ ++ SQUASHFS_SWAP(s[entry], d, bit_position, bits);\ ++} ++ ++#define SQUASHFS_SWAP_FRAGMENT_INDEXES(s, d, n) SQUASHFS_SWAP_LONG_LONGS(s, d, n) ++ ++#ifdef CONFIG_SQUASHFS_1_0_COMPATIBILITY ++ ++struct squashfs_base_inode_header_1 { ++ unsigned int inode_type:4; ++ unsigned int mode:12; /* protection */ ++ unsigned int uid:4; /* index into uid table */ ++ unsigned int guid:4; /* index into guid table */ ++} __attribute__ ((packed)); ++ ++struct squashfs_ipc_inode_header_1 { ++ unsigned int inode_type:4; ++ unsigned int mode:12; /* protection */ ++ unsigned int uid:4; /* index into uid table */ ++ unsigned int guid:4; /* index into guid table */ ++ unsigned int type:4; ++ unsigned int offset:4; ++} __attribute__ ((packed)); ++ ++struct squashfs_dev_inode_header_1 { ++ unsigned int inode_type:4; ++ unsigned int mode:12; /* protection */ ++ unsigned int uid:4; /* index into uid table */ ++ unsigned int guid:4; /* index into guid table */ ++ unsigned short rdev; ++} __attribute__ ((packed)); ++ ++struct squashfs_symlink_inode_header_1 { ++ unsigned int inode_type:4; ++ unsigned int mode:12; /* protection */ ++ unsigned int uid:4; /* index into uid table */ ++ unsigned int guid:4; /* index into guid table */ ++ unsigned short symlink_size; ++ char symlink[0]; ++} __attribute__ ((packed)); ++ ++struct squashfs_reg_inode_header_1 { ++ unsigned int inode_type:4; ++ unsigned int mode:12; /* protection */ ++ unsigned int uid:4; /* index into uid table */ ++ unsigned int guid:4; /* index into guid table */ ++ unsigned int mtime; ++ unsigned int start_block; ++ unsigned int file_size:32; ++ unsigned short block_list[0]; ++} __attribute__ ((packed)); ++ ++struct squashfs_dir_inode_header_1 { ++ unsigned int inode_type:4; ++ unsigned int mode:12; /* protection */ ++ unsigned int uid:4; /* index into uid table */ ++ unsigned int guid:4; /* index into guid table */ ++ unsigned int file_size:19; ++ unsigned int offset:13; ++ unsigned int mtime; ++ unsigned int start_block:24; ++} __attribute__ ((packed)); ++ ++#define SQUASHFS_SWAP_BASE_INODE_CORE_1(s, d, n) \ ++ SQUASHFS_MEMSET(s, d, n);\ ++ SQUASHFS_SWAP((s)->inode_type, d, 0, 4);\ ++ SQUASHFS_SWAP((s)->mode, d, 4, 12);\ ++ SQUASHFS_SWAP((s)->uid, d, 16, 4);\ ++ SQUASHFS_SWAP((s)->guid, d, 20, 4); ++ ++#define SQUASHFS_SWAP_BASE_INODE_HEADER_1(s, d, n) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE_1(s, d, n)\ ++} ++ ++#define SQUASHFS_SWAP_IPC_INODE_HEADER_1(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE_1(s, d, \ ++ sizeof(struct squashfs_ipc_inode_header_1));\ ++ SQUASHFS_SWAP((s)->type, d, 24, 4);\ ++ SQUASHFS_SWAP((s)->offset, d, 28, 4);\ ++} ++ ++#define SQUASHFS_SWAP_DEV_INODE_HEADER_1(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE_1(s, d, \ ++ sizeof(struct squashfs_dev_inode_header_1));\ ++ SQUASHFS_SWAP((s)->rdev, d, 24, 16);\ ++} ++ ++#define SQUASHFS_SWAP_SYMLINK_INODE_HEADER_1(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE_1(s, d, \ ++ sizeof(struct squashfs_symlink_inode_header_1));\ ++ SQUASHFS_SWAP((s)->symlink_size, d, 24, 16);\ ++} ++ ++#define SQUASHFS_SWAP_REG_INODE_HEADER_1(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE_1(s, d, \ ++ sizeof(struct squashfs_reg_inode_header_1));\ ++ SQUASHFS_SWAP((s)->mtime, d, 24, 32);\ ++ SQUASHFS_SWAP((s)->start_block, d, 56, 32);\ ++ SQUASHFS_SWAP((s)->file_size, d, 88, 32);\ ++} ++ ++#define SQUASHFS_SWAP_DIR_INODE_HEADER_1(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE_1(s, d, \ ++ sizeof(struct squashfs_dir_inode_header_1));\ ++ SQUASHFS_SWAP((s)->file_size, d, 24, 19);\ ++ SQUASHFS_SWAP((s)->offset, d, 43, 13);\ ++ SQUASHFS_SWAP((s)->mtime, d, 56, 32);\ ++ SQUASHFS_SWAP((s)->start_block, d, 88, 24);\ ++} ++ ++#endif ++ ++#ifdef CONFIG_SQUASHFS_2_0_COMPATIBILITY ++ ++struct squashfs_dir_index_2 { ++ unsigned int index:27; ++ unsigned int start_block:29; ++ unsigned char size; ++ unsigned char name[0]; ++} __attribute__ ((packed)); ++ ++struct squashfs_base_inode_header_2 { ++ unsigned int inode_type:4; ++ unsigned int mode:12; /* protection */ ++ unsigned int uid:8; /* index into uid table */ ++ unsigned int guid:8; /* index into guid table */ ++} __attribute__ ((packed)); ++ ++struct squashfs_ipc_inode_header_2 { ++ unsigned int inode_type:4; ++ unsigned int mode:12; /* protection */ ++ unsigned int uid:8; /* index into uid table */ ++ unsigned int guid:8; /* index into guid table */ ++} __attribute__ ((packed)); ++ ++struct squashfs_dev_inode_header_2 { ++ unsigned int inode_type:4; ++ unsigned int mode:12; /* protection */ ++ unsigned int uid:8; /* index into uid table */ ++ unsigned int guid:8; /* index into guid table */ ++ unsigned short rdev; ++} __attribute__ ((packed)); ++ ++struct squashfs_symlink_inode_header_2 { ++ unsigned int inode_type:4; ++ unsigned int mode:12; /* protection */ ++ unsigned int uid:8; /* index into uid table */ ++ unsigned int guid:8; /* index into guid table */ ++ unsigned short symlink_size; ++ char symlink[0]; ++} __attribute__ ((packed)); ++ ++struct squashfs_reg_inode_header_2 { ++ unsigned int inode_type:4; ++ unsigned int mode:12; /* protection */ ++ unsigned int uid:8; /* index into uid table */ ++ unsigned int guid:8; /* index into guid table */ ++ unsigned int mtime; ++ unsigned int start_block; ++ unsigned int fragment; ++ unsigned int offset; ++ unsigned int file_size:32; ++ unsigned short block_list[0]; ++} __attribute__ ((packed)); ++ ++struct squashfs_dir_inode_header_2 { ++ unsigned int inode_type:4; ++ unsigned int mode:12; /* protection */ ++ unsigned int uid:8; /* index into uid table */ ++ unsigned int guid:8; /* index into guid table */ ++ unsigned int file_size:19; ++ unsigned int offset:13; ++ unsigned int mtime; ++ unsigned int start_block:24; ++} __attribute__ ((packed)); ++ ++struct squashfs_ldir_inode_header_2 { ++ unsigned int inode_type:4; ++ unsigned int mode:12; /* protection */ ++ unsigned int uid:8; /* index into uid table */ ++ unsigned int guid:8; /* index into guid table */ ++ unsigned int file_size:27; ++ unsigned int offset:13; ++ unsigned int mtime; ++ unsigned int start_block:24; ++ unsigned int i_count:16; ++ struct squashfs_dir_index_2 index[0]; ++} __attribute__ ((packed)); ++ ++union squashfs_inode_header_2 { ++ struct squashfs_base_inode_header_2 base; ++ struct squashfs_dev_inode_header_2 dev; ++ struct squashfs_symlink_inode_header_2 symlink; ++ struct squashfs_reg_inode_header_2 reg; ++ struct squashfs_dir_inode_header_2 dir; ++ struct squashfs_ldir_inode_header_2 ldir; ++ struct squashfs_ipc_inode_header_2 ipc; ++}; ++ ++struct squashfs_dir_header_2 { ++ unsigned int count:8; ++ unsigned int start_block:24; ++} __attribute__ ((packed)); ++ ++struct squashfs_dir_entry_2 { ++ unsigned int offset:13; ++ unsigned int type:3; ++ unsigned int size:8; ++ char name[0]; ++} __attribute__ ((packed)); ++ ++struct squashfs_fragment_entry_2 { ++ unsigned int start_block; ++ unsigned int size; ++} __attribute__ ((packed)); ++ ++#define SQUASHFS_SWAP_BASE_INODE_CORE_2(s, d, n)\ ++ SQUASHFS_MEMSET(s, d, n);\ ++ SQUASHFS_SWAP((s)->inode_type, d, 0, 4);\ ++ SQUASHFS_SWAP((s)->mode, d, 4, 12);\ ++ SQUASHFS_SWAP((s)->uid, d, 16, 8);\ ++ SQUASHFS_SWAP((s)->guid, d, 24, 8);\ ++ ++#define SQUASHFS_SWAP_BASE_INODE_HEADER_2(s, d, n) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE_2(s, d, n)\ ++} ++ ++#define SQUASHFS_SWAP_IPC_INODE_HEADER_2(s, d) \ ++ SQUASHFS_SWAP_BASE_INODE_HEADER_2(s, d, sizeof(struct squashfs_ipc_inode_header_2)) ++ ++#define SQUASHFS_SWAP_DEV_INODE_HEADER_2(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE_2(s, d, \ ++ sizeof(struct squashfs_dev_inode_header_2)); \ ++ SQUASHFS_SWAP((s)->rdev, d, 32, 16);\ ++} ++ ++#define SQUASHFS_SWAP_SYMLINK_INODE_HEADER_2(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE_2(s, d, \ ++ sizeof(struct squashfs_symlink_inode_header_2));\ ++ SQUASHFS_SWAP((s)->symlink_size, d, 32, 16);\ ++} ++ ++#define SQUASHFS_SWAP_REG_INODE_HEADER_2(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE_2(s, d, \ ++ sizeof(struct squashfs_reg_inode_header_2));\ ++ SQUASHFS_SWAP((s)->mtime, d, 32, 32);\ ++ SQUASHFS_SWAP((s)->start_block, d, 64, 32);\ ++ SQUASHFS_SWAP((s)->fragment, d, 96, 32);\ ++ SQUASHFS_SWAP((s)->offset, d, 128, 32);\ ++ SQUASHFS_SWAP((s)->file_size, d, 160, 32);\ ++} ++ ++#define SQUASHFS_SWAP_DIR_INODE_HEADER_2(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE_2(s, d, \ ++ sizeof(struct squashfs_dir_inode_header_2));\ ++ SQUASHFS_SWAP((s)->file_size, d, 32, 19);\ ++ SQUASHFS_SWAP((s)->offset, d, 51, 13);\ ++ SQUASHFS_SWAP((s)->mtime, d, 64, 32);\ ++ SQUASHFS_SWAP((s)->start_block, d, 96, 24);\ ++} ++ ++#define SQUASHFS_SWAP_LDIR_INODE_HEADER_2(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_SWAP_BASE_INODE_CORE_2(s, d, \ ++ sizeof(struct squashfs_ldir_inode_header_2));\ ++ SQUASHFS_SWAP((s)->file_size, d, 32, 27);\ ++ SQUASHFS_SWAP((s)->offset, d, 59, 13);\ ++ SQUASHFS_SWAP((s)->mtime, d, 72, 32);\ ++ SQUASHFS_SWAP((s)->start_block, d, 104, 24);\ ++ SQUASHFS_SWAP((s)->i_count, d, 128, 16);\ ++} ++ ++#define SQUASHFS_SWAP_DIR_INDEX_2(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_MEMSET(s, d, sizeof(struct squashfs_dir_index_2));\ ++ SQUASHFS_SWAP((s)->index, d, 0, 27);\ ++ SQUASHFS_SWAP((s)->start_block, d, 27, 29);\ ++ SQUASHFS_SWAP((s)->size, d, 56, 8);\ ++} ++#define SQUASHFS_SWAP_DIR_HEADER_2(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_MEMSET(s, d, sizeof(struct squashfs_dir_header_2));\ ++ SQUASHFS_SWAP((s)->count, d, 0, 8);\ ++ SQUASHFS_SWAP((s)->start_block, d, 8, 24);\ ++} ++ ++#define SQUASHFS_SWAP_DIR_ENTRY_2(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_MEMSET(s, d, sizeof(struct squashfs_dir_entry_2));\ ++ SQUASHFS_SWAP((s)->offset, d, 0, 13);\ ++ SQUASHFS_SWAP((s)->type, d, 13, 3);\ ++ SQUASHFS_SWAP((s)->size, d, 16, 8);\ ++} ++ ++#define SQUASHFS_SWAP_FRAGMENT_ENTRY_2(s, d) {\ ++ SQUASHFS_SWAP_START\ ++ SQUASHFS_MEMSET(s, d, sizeof(struct squashfs_fragment_entry_2));\ ++ SQUASHFS_SWAP((s)->start_block, d, 0, 32);\ ++ SQUASHFS_SWAP((s)->size, d, 32, 32);\ ++} ++ ++#define SQUASHFS_SWAP_FRAGMENT_INDEXES_2(s, d, n) SQUASHFS_SWAP_INTS(s, d, n) ++ ++/* fragment and fragment table defines */ ++#define SQUASHFS_FRAGMENT_BYTES_2(A) (A * sizeof(struct squashfs_fragment_entry_2)) ++ ++#define SQUASHFS_FRAGMENT_INDEX_2(A) (SQUASHFS_FRAGMENT_BYTES_2(A) / \ ++ SQUASHFS_METADATA_SIZE) ++ ++#define SQUASHFS_FRAGMENT_INDEX_OFFSET_2(A) (SQUASHFS_FRAGMENT_BYTES_2(A) % \ ++ SQUASHFS_METADATA_SIZE) ++ ++#define SQUASHFS_FRAGMENT_INDEXES_2(A) ((SQUASHFS_FRAGMENT_BYTES_2(A) + \ ++ SQUASHFS_METADATA_SIZE - 1) / \ ++ SQUASHFS_METADATA_SIZE) ++ ++#define SQUASHFS_FRAGMENT_INDEX_BYTES_2(A) (SQUASHFS_FRAGMENT_INDEXES_2(A) *\ ++ sizeof(int)) ++ ++#endif ++ ++#ifdef __KERNEL__ ++ ++/* ++ * macros used to swap each structure entry, taking into account ++ * bitfields and different bitfield placing conventions on differing ++ * architectures ++ */ ++ ++#include <asm/byteorder.h> ++ ++#ifdef __BIG_ENDIAN ++ /* convert from little endian to big endian */ ++#define SQUASHFS_SWAP(value, p, pos, tbits) _SQUASHFS_SWAP(value, p, pos, \ ++ tbits, b_pos) ++#else ++ /* convert from big endian to little endian */ ++#define SQUASHFS_SWAP(value, p, pos, tbits) _SQUASHFS_SWAP(value, p, pos, \ ++ tbits, 64 - tbits - b_pos) ++#endif ++ ++#define _SQUASHFS_SWAP(value, p, pos, tbits, SHIFT) {\ ++ b_pos = pos % 8;\ ++ val = 0;\ ++ s = (unsigned char *)p + (pos / 8);\ ++ d = ((unsigned char *) &val) + 7;\ ++ for(bits = 0; bits < (tbits + b_pos); bits += 8) \ ++ *d-- = *s++;\ ++ value = (val >> (SHIFT))/* & ((1 << tbits) - 1)*/;\ ++} ++ ++#define SQUASHFS_MEMSET(s, d, n) memset(s, 0, n); ++ ++#endif ++#endif +Index: linux-2.6.23/include/linux/squashfs_fs_i.h +=================================================================== +--- /dev/null 1970-01-01 00:00:00.000000000 +0000 ++++ linux-2.6.23/include/linux/squashfs_fs_i.h 2007-10-10 13:52:14.000000000 +0800 +@@ -0,0 +1,45 @@ ++#ifndef SQUASHFS_FS_I ++#define SQUASHFS_FS_I ++/* ++ * Squashfs ++ * ++ * Copyright (c) 2002, 2003, 2004, 2005, 2006 ++ * Phillip Lougher <phillip@lougher.org.uk> ++ * ++ * This program 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. ++ * ++ * This program 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 this program; if not, write to the Free Software ++ * Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. ++ * ++ * squashfs_fs_i.h ++ */ ++ ++struct squashfs_inode_info { ++ long long start_block; ++ unsigned int offset; ++ union { ++ struct { ++ long long fragment_start_block; ++ unsigned int fragment_size; ++ unsigned int fragment_offset; ++ long long block_list_start; ++ } s1; ++ struct { ++ long long directory_index_start; ++ unsigned int directory_index_offset; ++ unsigned int directory_index_count; ++ unsigned int parent_inode; ++ } s2; ++ } u; ++ struct inode vfs_inode; ++}; ++#endif +Index: linux-2.6.23/include/linux/squashfs_fs_sb.h +=================================================================== +--- /dev/null 1970-01-01 00:00:00.000000000 +0000 ++++ linux-2.6.23/include/linux/squashfs_fs_sb.h 2007-10-10 13:52:14.000000000 +0800 +@@ -0,0 +1,74 @@ ++#ifndef SQUASHFS_FS_SB ++#define SQUASHFS_FS_SB ++/* ++ * Squashfs ++ * ++ * Copyright (c) 2002, 2003, 2004, 2005, 2006 ++ * Phillip Lougher <phillip@lougher.org.uk> ++ * ++ * This program 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. ++ * ++ * This program 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 this program; if not, write to the Free Software ++ * Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. ++ * ++ * squashfs_fs_sb.h ++ */ ++ ++#include <linux/squashfs_fs.h> ++ ++struct squashfs_cache { ++ long long block; ++ int length; ++ long long next_index; ++ char *data; ++}; ++ ++struct squashfs_fragment_cache { ++ long long block; ++ int length; ++ unsigned int locked; ++ char *data; ++}; ++ ++struct squashfs_sb_info { ++ struct squashfs_super_block sblk; ++ int devblksize; ++ int devblksize_log2; ++ int swap; ++ struct squashfs_cache *block_cache; ++ struct squashfs_fragment_cache *fragment; ++ int next_cache; ++ int next_fragment; ++ int next_meta_index; ++ unsigned int *uid; ++ unsigned int *guid; ++ long long *fragment_index; ++ unsigned int *fragment_index_2; ++ unsigned int read_size; ++ char *read_data; ++ char *read_page; ++ struct semaphore read_data_mutex; ++ struct semaphore read_page_mutex; ++ struct semaphore block_cache_mutex; ++ struct semaphore fragment_mutex; ++ struct semaphore meta_index_mutex; ++ wait_queue_head_t waitq; ++ wait_queue_head_t fragment_wait_queue; ++ struct meta_index *meta_index; ++ struct inode *(*iget)(struct super_block *s, squashfs_inode_t \ ++ inode); ++ long long (*read_blocklist)(struct inode *inode, int \ ++ index, int readahead_blks, char *block_list, \ ++ unsigned short **block_p, unsigned int *bsize); ++ int (*read_fragment_index_table)(struct super_block *s); ++}; ++#endif +Index: linux-2.6.23/init/do_mounts_rd.c +=================================================================== +--- linux-2.6.23.orig/init/do_mounts_rd.c 2007-10-10 13:52:12.000000000 +0800 ++++ linux-2.6.23/init/do_mounts_rd.c 2007-10-10 13:52:14.000000000 +0800 +@@ -5,6 +5,7 @@ + #include <linux/ext2_fs.h> + #include <linux/romfs_fs.h> + #include <linux/cramfs_fs.h> ++#include <linux/squashfs_fs.h> + #include <linux/initrd.h> + #include <linux/string.h> + +@@ -39,6 +40,7 @@ + * numbers could not be found. + * + * We currently check for the following magic numbers: ++ * squashfs + * minix + * ext2 + * romfs +@@ -53,6 +55,7 @@ + struct ext2_super_block *ext2sb; + struct romfs_super_block *romfsb; + struct cramfs_super *cramfsb; ++ struct squashfs_super_block *squashfsb; + int nblocks = -1; + unsigned char *buf; + +@@ -64,6 +67,7 @@ + ext2sb = (struct ext2_super_block *) buf; + romfsb = (struct romfs_super_block *) buf; + cramfsb = (struct cramfs_super *) buf; ++ squashfsb = (struct squashfs_super_block *) buf; + memset(buf, 0xe5, size); + + /* +@@ -101,6 +105,15 @@ + goto done; + } + ++ /* squashfs is at block zero too */ ++ if (squashfsb->s_magic == SQUASHFS_MAGIC) { ++ printk(KERN_NOTICE ++ "RAMDISK: squashfs filesystem found at block %d\n", ++ start_block); ++ nblocks = (squashfsb->bytes_used+BLOCK_SIZE-1)>>BLOCK_SIZE_BITS; ++ goto done; ++ } ++ + /* + * Read block 1 to test for minix and ext2 superblock + */ |