linux-sg2042/fs/ext4/extents.c

4725 lines
130 KiB
C
Raw Normal View History

/*
* Copyright (c) 2003-2006, Cluster File Systems, Inc, info@clusterfs.com
* Written by Alex Tomas <alex@clusterfs.com>
*
* Architecture independence:
* Copyright (c) 2005, Bull S.A.
* Written by Pierre Peiffer <pierre.peiffer@bull.net>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*
* 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 Licens
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-
*/
/*
* Extents support for EXT4
*
* TODO:
* - ext4*_error() should be used in some situations
* - analyze all BUG()/BUG_ON(), use -EIO where appropriate
* - smart tree reduction
*/
#include <linux/fs.h>
#include <linux/time.h>
#include <linux/jbd2.h>
#include <linux/highuid.h>
#include <linux/pagemap.h>
#include <linux/quotaops.h>
#include <linux/string.h>
#include <linux/slab.h>
#include <linux/falloc.h>
#include <asm/uaccess.h>
#include <linux/fiemap.h>
#include "ext4_jbd2.h"
#include "ext4_extents.h"
#include "xattr.h"
#include <trace/events/ext4.h>
ext4: rewrite punch hole to use ext4_ext_remove_space() This commit rewrites ext4 punch hole implementation to use ext4_ext_remove_space() instead of its home gown way of doing this via ext4_ext_map_blocks(). There are several reasons for changing this. Firstly it is quite non obvious that punching hole needs to ext4_ext_map_blocks() to punch a hole, especially given that this function should map blocks, not unmap it. It also required a lot of new code in ext4_ext_map_blocks(). Secondly the design of it is not very effective. The reason is that we are trying to punch out blocks in ext4_ext_punch_hole() in opposite direction than in ext4_ext_rm_leaf() which causes the ext4_ext_rm_leaf() to iterate through the whole tree from the end to the start to find the requested extent for every extent we are going to punch out. And finally the current implementation does not use the existing code, but bring a lot of new code, which is IMO unnecessary since there already is some infrastructure we can use. Specifically ext4_ext_remove_space(). This commit changes ext4_ext_remove_space() to accept 'end' parameter so we can not only truncate to the end of file, but also remove the space in the middle of the file (punch a hole). Moreover, because the last block to punch out, might be in the middle of the extent, we have to split the extent at 'end + 1' so ext4_ext_rm_leaf() can easily either remove the whole fist part of split extent, or change its size. ext4_ext_remove_space() is then used to actually remove the space (extents) from within the hole, instead of ext4_ext_map_blocks(). Note that this also fix the issue with punch hole, where we would forget to remove empty index blocks from the extent tree, resulting in double free block error and file system corruption. This is simply because we now use different code path, where this problem does not exist. This has been tested with fsx running for several days and xfstests, plus xfstest #251 with '-o discard' run on the loop image (which converts discard requestes into punch hole to the backing file). All of it on 1K and 4K file system block size. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-03-20 11:03:19 +08:00
/*
* used by extent splitting.
*/
#define EXT4_EXT_MAY_ZEROOUT 0x1 /* safe to zeroout if split fails \
due to ENOSPC */
#define EXT4_EXT_MARK_UNINIT1 0x2 /* mark first half uninitialized */
#define EXT4_EXT_MARK_UNINIT2 0x4 /* mark second half uninitialized */
#define EXT4_EXT_DATA_VALID1 0x8 /* first half contains valid data */
#define EXT4_EXT_DATA_VALID2 0x10 /* second half contains valid data */
static __le32 ext4_extent_block_csum(struct inode *inode,
struct ext4_extent_header *eh)
{
struct ext4_inode_info *ei = EXT4_I(inode);
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
__u32 csum;
csum = ext4_chksum(sbi, ei->i_csum_seed, (__u8 *)eh,
EXT4_EXTENT_TAIL_OFFSET(eh));
return cpu_to_le32(csum);
}
static int ext4_extent_block_csum_verify(struct inode *inode,
struct ext4_extent_header *eh)
{
struct ext4_extent_tail *et;
if (!EXT4_HAS_RO_COMPAT_FEATURE(inode->i_sb,
EXT4_FEATURE_RO_COMPAT_METADATA_CSUM))
return 1;
et = find_ext4_extent_tail(eh);
if (et->et_checksum != ext4_extent_block_csum(inode, eh))
return 0;
return 1;
}
static void ext4_extent_block_csum_set(struct inode *inode,
struct ext4_extent_header *eh)
{
struct ext4_extent_tail *et;
if (!EXT4_HAS_RO_COMPAT_FEATURE(inode->i_sb,
EXT4_FEATURE_RO_COMPAT_METADATA_CSUM))
return;
et = find_ext4_extent_tail(eh);
et->et_checksum = ext4_extent_block_csum(inode, eh);
}
static int ext4_split_extent(handle_t *handle,
struct inode *inode,
struct ext4_ext_path *path,
struct ext4_map_blocks *map,
int split_flag,
int flags);
ext4: rewrite punch hole to use ext4_ext_remove_space() This commit rewrites ext4 punch hole implementation to use ext4_ext_remove_space() instead of its home gown way of doing this via ext4_ext_map_blocks(). There are several reasons for changing this. Firstly it is quite non obvious that punching hole needs to ext4_ext_map_blocks() to punch a hole, especially given that this function should map blocks, not unmap it. It also required a lot of new code in ext4_ext_map_blocks(). Secondly the design of it is not very effective. The reason is that we are trying to punch out blocks in ext4_ext_punch_hole() in opposite direction than in ext4_ext_rm_leaf() which causes the ext4_ext_rm_leaf() to iterate through the whole tree from the end to the start to find the requested extent for every extent we are going to punch out. And finally the current implementation does not use the existing code, but bring a lot of new code, which is IMO unnecessary since there already is some infrastructure we can use. Specifically ext4_ext_remove_space(). This commit changes ext4_ext_remove_space() to accept 'end' parameter so we can not only truncate to the end of file, but also remove the space in the middle of the file (punch a hole). Moreover, because the last block to punch out, might be in the middle of the extent, we have to split the extent at 'end + 1' so ext4_ext_rm_leaf() can easily either remove the whole fist part of split extent, or change its size. ext4_ext_remove_space() is then used to actually remove the space (extents) from within the hole, instead of ext4_ext_map_blocks(). Note that this also fix the issue with punch hole, where we would forget to remove empty index blocks from the extent tree, resulting in double free block error and file system corruption. This is simply because we now use different code path, where this problem does not exist. This has been tested with fsx running for several days and xfstests, plus xfstest #251 with '-o discard' run on the loop image (which converts discard requestes into punch hole to the backing file). All of it on 1K and 4K file system block size. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-03-20 11:03:19 +08:00
static int ext4_split_extent_at(handle_t *handle,
struct inode *inode,
struct ext4_ext_path *path,
ext4_lblk_t split,
int split_flag,
int flags);
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
static int ext4_find_delayed_extent(struct inode *inode,
struct extent_status *newes);
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
static int ext4_ext_truncate_extend_restart(handle_t *handle,
struct inode *inode,
int needed)
{
int err;
if (!ext4_handle_valid(handle))
return 0;
if (handle->h_buffer_credits > needed)
return 0;
err = ext4_journal_extend(handle, needed);
if (err <= 0)
return err;
err = ext4_truncate_restart_trans(handle, inode, needed);
if (err == 0)
err = -EAGAIN;
return err;
}
/*
* could return:
* - EROFS
* - ENOMEM
*/
static int ext4_ext_get_access(handle_t *handle, struct inode *inode,
struct ext4_ext_path *path)
{
if (path->p_bh) {
/* path points to block */
return ext4_journal_get_write_access(handle, path->p_bh);
}
/* path points to leaf/index in inode body */
/* we use in-core data, no need to protect them */
return 0;
}
/*
* could return:
* - EROFS
* - ENOMEM
* - EIO
*/
int __ext4_ext_dirty(const char *where, unsigned int line, handle_t *handle,
struct inode *inode, struct ext4_ext_path *path)
{
int err;
if (path->p_bh) {
ext4_extent_block_csum_set(inode, ext_block_hdr(path->p_bh));
/* path points to block */
err = __ext4_handle_dirty_metadata(where, line, handle,
inode, path->p_bh);
} else {
/* path points to leaf/index in inode body */
err = ext4_mark_inode_dirty(handle, inode);
}
return err;
}
static ext4_fsblk_t ext4_ext_find_goal(struct inode *inode,
struct ext4_ext_path *path,
ext4_lblk_t block)
{
if (path) {
int depth = path->p_depth;
struct ext4_extent *ex;
ext4: fix 32bit overflow in ext4_ext_find_goal() ext4_ext_find_goal() returns an ideal physical block number that the block allocator tries to allocate first. However, if a required file offset is smaller than the existing extent's one, ext4_ext_find_goal() returns a wrong block number because it may overflow at "block - le32_to_cpu(ex->ee_block)". This patch fixes the problem. ext4_ext_find_goal() will also return a wrong block number in case a file offset of the existing extent is too big. In this case, the ideal physical block number is fixed in ext4_mb_initialize_context(), so it's no problem. reproduce: # dd if=/dev/zero of=/mnt/mp1/tmp bs=127M count=1 oflag=sync # dd if=/dev/zero of=/mnt/mp1/file bs=512K count=1 seek=1 oflag=sync # filefrag -v /mnt/mp1/file Filesystem type is: ef53 File size of /mnt/mp1/file is 1048576 (256 blocks, blocksize 4096) ext logical physical expected length flags 0 128 67456 128 eof /mnt/mp1/file: 2 extents found # rm -rf /mnt/mp1/tmp # echo $((512*4096)) > /sys/fs/ext4/loop0/mb_stream_req # dd if=/dev/zero of=/mnt/mp1/file bs=512K count=1 oflag=sync conv=notrunc result (linux-2.6.37-rc2 + ext4 patch queue): # filefrag -v /mnt/mp1/file Filesystem type is: ef53 File size of /mnt/mp1/file is 1048576 (256 blocks, blocksize 4096) ext logical physical expected length flags 0 0 33280 128 1 128 67456 33407 128 eof /mnt/mp1/file: 2 extents found result(apply this patch): # filefrag -v /mnt/mp1/file Filesystem type is: ef53 File size of /mnt/mp1/file is 1048576 (256 blocks, blocksize 4096) ext logical physical expected length flags 0 0 66560 128 1 128 67456 66687 128 eof /mnt/mp1/file: 2 extents found Signed-off-by: Kazuya Mio <k-mio@sx.jp.nec.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-01-11 01:12:28 +08:00
/*
* Try to predict block placement assuming that we are
* filling in a file which will eventually be
* non-sparse --- i.e., in the case of libbfd writing
* an ELF object sections out-of-order but in a way
* the eventually results in a contiguous object or
* executable file, or some database extending a table
* space file. However, this is actually somewhat
* non-ideal if we are writing a sparse file such as
* qemu or KVM writing a raw image file that is going
* to stay fairly sparse, since it will end up
* fragmenting the file system's free space. Maybe we
* should have some hueristics or some way to allow
* userspace to pass a hint to file system,
* especially if the latter case turns out to be
ext4: fix 32bit overflow in ext4_ext_find_goal() ext4_ext_find_goal() returns an ideal physical block number that the block allocator tries to allocate first. However, if a required file offset is smaller than the existing extent's one, ext4_ext_find_goal() returns a wrong block number because it may overflow at "block - le32_to_cpu(ex->ee_block)". This patch fixes the problem. ext4_ext_find_goal() will also return a wrong block number in case a file offset of the existing extent is too big. In this case, the ideal physical block number is fixed in ext4_mb_initialize_context(), so it's no problem. reproduce: # dd if=/dev/zero of=/mnt/mp1/tmp bs=127M count=1 oflag=sync # dd if=/dev/zero of=/mnt/mp1/file bs=512K count=1 seek=1 oflag=sync # filefrag -v /mnt/mp1/file Filesystem type is: ef53 File size of /mnt/mp1/file is 1048576 (256 blocks, blocksize 4096) ext logical physical expected length flags 0 128 67456 128 eof /mnt/mp1/file: 2 extents found # rm -rf /mnt/mp1/tmp # echo $((512*4096)) > /sys/fs/ext4/loop0/mb_stream_req # dd if=/dev/zero of=/mnt/mp1/file bs=512K count=1 oflag=sync conv=notrunc result (linux-2.6.37-rc2 + ext4 patch queue): # filefrag -v /mnt/mp1/file Filesystem type is: ef53 File size of /mnt/mp1/file is 1048576 (256 blocks, blocksize 4096) ext logical physical expected length flags 0 0 33280 128 1 128 67456 33407 128 eof /mnt/mp1/file: 2 extents found result(apply this patch): # filefrag -v /mnt/mp1/file Filesystem type is: ef53 File size of /mnt/mp1/file is 1048576 (256 blocks, blocksize 4096) ext logical physical expected length flags 0 0 66560 128 1 128 67456 66687 128 eof /mnt/mp1/file: 2 extents found Signed-off-by: Kazuya Mio <k-mio@sx.jp.nec.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-01-11 01:12:28 +08:00
* common.
*/
ex = path[depth].p_ext;
ext4: fix 32bit overflow in ext4_ext_find_goal() ext4_ext_find_goal() returns an ideal physical block number that the block allocator tries to allocate first. However, if a required file offset is smaller than the existing extent's one, ext4_ext_find_goal() returns a wrong block number because it may overflow at "block - le32_to_cpu(ex->ee_block)". This patch fixes the problem. ext4_ext_find_goal() will also return a wrong block number in case a file offset of the existing extent is too big. In this case, the ideal physical block number is fixed in ext4_mb_initialize_context(), so it's no problem. reproduce: # dd if=/dev/zero of=/mnt/mp1/tmp bs=127M count=1 oflag=sync # dd if=/dev/zero of=/mnt/mp1/file bs=512K count=1 seek=1 oflag=sync # filefrag -v /mnt/mp1/file Filesystem type is: ef53 File size of /mnt/mp1/file is 1048576 (256 blocks, blocksize 4096) ext logical physical expected length flags 0 128 67456 128 eof /mnt/mp1/file: 2 extents found # rm -rf /mnt/mp1/tmp # echo $((512*4096)) > /sys/fs/ext4/loop0/mb_stream_req # dd if=/dev/zero of=/mnt/mp1/file bs=512K count=1 oflag=sync conv=notrunc result (linux-2.6.37-rc2 + ext4 patch queue): # filefrag -v /mnt/mp1/file Filesystem type is: ef53 File size of /mnt/mp1/file is 1048576 (256 blocks, blocksize 4096) ext logical physical expected length flags 0 0 33280 128 1 128 67456 33407 128 eof /mnt/mp1/file: 2 extents found result(apply this patch): # filefrag -v /mnt/mp1/file Filesystem type is: ef53 File size of /mnt/mp1/file is 1048576 (256 blocks, blocksize 4096) ext logical physical expected length flags 0 0 66560 128 1 128 67456 66687 128 eof /mnt/mp1/file: 2 extents found Signed-off-by: Kazuya Mio <k-mio@sx.jp.nec.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-01-11 01:12:28 +08:00
if (ex) {
ext4_fsblk_t ext_pblk = ext4_ext_pblock(ex);
ext4_lblk_t ext_block = le32_to_cpu(ex->ee_block);
if (block > ext_block)
return ext_pblk + (block - ext_block);
else
return ext_pblk - (ext_block - block);
}
/* it looks like index is empty;
* try to find starting block from index itself */
if (path[depth].p_bh)
return path[depth].p_bh->b_blocknr;
}
/* OK. use inode's group */
return ext4_inode_to_goal_block(inode);
}
/*
* Allocation for a meta data block
*/
static ext4_fsblk_t
ext4_ext_new_meta_block(handle_t *handle, struct inode *inode,
struct ext4_ext_path *path,
struct ext4_extent *ex, int *err, unsigned int flags)
{
ext4_fsblk_t goal, newblock;
goal = ext4_ext_find_goal(inode, path, le32_to_cpu(ex->ee_block));
newblock = ext4_new_meta_blocks(handle, inode, goal, flags,
NULL, err);
return newblock;
}
static inline int ext4_ext_space_block(struct inode *inode, int check)
{
int size;
size = (inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
/ sizeof(struct ext4_extent);
#ifdef AGGRESSIVE_TEST
if (!check && size > 6)
size = 6;
#endif
return size;
}
static inline int ext4_ext_space_block_idx(struct inode *inode, int check)
{
int size;
size = (inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
/ sizeof(struct ext4_extent_idx);
#ifdef AGGRESSIVE_TEST
if (!check && size > 5)
size = 5;
#endif
return size;
}
static inline int ext4_ext_space_root(struct inode *inode, int check)
{
int size;
size = sizeof(EXT4_I(inode)->i_data);
size -= sizeof(struct ext4_extent_header);
size /= sizeof(struct ext4_extent);
#ifdef AGGRESSIVE_TEST
if (!check && size > 3)
size = 3;
#endif
return size;
}
static inline int ext4_ext_space_root_idx(struct inode *inode, int check)
{
int size;
size = sizeof(EXT4_I(inode)->i_data);
size -= sizeof(struct ext4_extent_header);
size /= sizeof(struct ext4_extent_idx);
#ifdef AGGRESSIVE_TEST
if (!check && size > 4)
size = 4;
#endif
return size;
}
/*
* Calculate the number of metadata blocks needed
* to allocate @blocks
* Worse case is one block per extent
*/
int ext4_ext_calc_metadata_amount(struct inode *inode, ext4_lblk_t lblock)
{
struct ext4_inode_info *ei = EXT4_I(inode);
int idxs;
idxs = ((inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
/ sizeof(struct ext4_extent_idx));
/*
* If the new delayed allocation block is contiguous with the
* previous da block, it can share index blocks with the
* previous block, so we only need to allocate a new index
* block every idxs leaf blocks. At ldxs**2 blocks, we need
* an additional index block, and at ldxs**3 blocks, yet
* another index blocks.
*/
if (ei->i_da_metadata_calc_len &&
ei->i_da_metadata_calc_last_lblock+1 == lblock) {
int num = 0;
if ((ei->i_da_metadata_calc_len % idxs) == 0)
num++;
if ((ei->i_da_metadata_calc_len % (idxs*idxs)) == 0)
num++;
if ((ei->i_da_metadata_calc_len % (idxs*idxs*idxs)) == 0) {
num++;
ei->i_da_metadata_calc_len = 0;
} else
ei->i_da_metadata_calc_len++;
ei->i_da_metadata_calc_last_lblock++;
return num;
}
/*
* In the worst case we need a new set of index blocks at
* every level of the inode's extent tree.
*/
ei->i_da_metadata_calc_len = 1;
ei->i_da_metadata_calc_last_lblock = lblock;
return ext_depth(inode) + 1;
}
static int
ext4_ext_max_entries(struct inode *inode, int depth)
{
int max;
if (depth == ext_depth(inode)) {
if (depth == 0)
max = ext4_ext_space_root(inode, 1);
else
max = ext4_ext_space_root_idx(inode, 1);
} else {
if (depth == 0)
max = ext4_ext_space_block(inode, 1);
else
max = ext4_ext_space_block_idx(inode, 1);
}
return max;
}
static int ext4_valid_extent(struct inode *inode, struct ext4_extent *ext)
{
ext4_fsblk_t block = ext4_ext_pblock(ext);
int len = ext4_ext_get_actual_len(ext);
if (len == 0)
return 0;
return ext4_data_block_valid(EXT4_SB(inode->i_sb), block, len);
}
static int ext4_valid_extent_idx(struct inode *inode,
struct ext4_extent_idx *ext_idx)
{
ext4_fsblk_t block = ext4_idx_pblock(ext_idx);
return ext4_data_block_valid(EXT4_SB(inode->i_sb), block, 1);
}
static int ext4_valid_extent_entries(struct inode *inode,
struct ext4_extent_header *eh,
int depth)
{
unsigned short entries;
if (eh->eh_entries == 0)
return 1;
entries = le16_to_cpu(eh->eh_entries);
if (depth == 0) {
/* leaf entries */
struct ext4_extent *ext = EXT_FIRST_EXTENT(eh);
while (entries) {
if (!ext4_valid_extent(inode, ext))
return 0;
ext++;
entries--;
}
} else {
struct ext4_extent_idx *ext_idx = EXT_FIRST_INDEX(eh);
while (entries) {
if (!ext4_valid_extent_idx(inode, ext_idx))
return 0;
ext_idx++;
entries--;
}
}
return 1;
}
static int __ext4_ext_check(const char *function, unsigned int line,
struct inode *inode, struct ext4_extent_header *eh,
int depth)
{
const char *error_msg;
int max = 0;
if (unlikely(eh->eh_magic != EXT4_EXT_MAGIC)) {
error_msg = "invalid magic";
goto corrupted;
}
if (unlikely(le16_to_cpu(eh->eh_depth) != depth)) {
error_msg = "unexpected eh_depth";
goto corrupted;
}
if (unlikely(eh->eh_max == 0)) {
error_msg = "invalid eh_max";
goto corrupted;
}
max = ext4_ext_max_entries(inode, depth);
if (unlikely(le16_to_cpu(eh->eh_max) > max)) {
error_msg = "too large eh_max";
goto corrupted;
}
if (unlikely(le16_to_cpu(eh->eh_entries) > le16_to_cpu(eh->eh_max))) {
error_msg = "invalid eh_entries";
goto corrupted;
}
if (!ext4_valid_extent_entries(inode, eh, depth)) {
error_msg = "invalid extent entries";
goto corrupted;
}
/* Verify checksum on non-root extent tree nodes */
if (ext_depth(inode) != depth &&
!ext4_extent_block_csum_verify(inode, eh)) {
error_msg = "extent tree corrupted";
goto corrupted;
}
return 0;
corrupted:
ext4_error_inode(inode, function, line, 0,
"bad header/extent: %s - magic %x, "
"entries %u, max %u(%u), depth %u(%u)",
error_msg, le16_to_cpu(eh->eh_magic),
le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max),
max, le16_to_cpu(eh->eh_depth), depth);
return -EIO;
}
#define ext4_ext_check(inode, eh, depth) \
__ext4_ext_check(__func__, __LINE__, inode, eh, depth)
int ext4_ext_check_inode(struct inode *inode)
{
return ext4_ext_check(inode, ext_inode_hdr(inode), ext_depth(inode));
}
static int __ext4_ext_check_block(const char *function, unsigned int line,
struct inode *inode,
struct ext4_extent_header *eh,
int depth,
struct buffer_head *bh)
{
int ret;
if (buffer_verified(bh))
return 0;
ret = ext4_ext_check(inode, eh, depth);
if (ret)
return ret;
set_buffer_verified(bh);
return ret;
}
#define ext4_ext_check_block(inode, eh, depth, bh) \
__ext4_ext_check_block(__func__, __LINE__, inode, eh, depth, bh)
#ifdef EXT_DEBUG
static void ext4_ext_show_path(struct inode *inode, struct ext4_ext_path *path)
{
int k, l = path->p_depth;
ext_debug("path:");
for (k = 0; k <= l; k++, path++) {
if (path->p_idx) {
ext_debug(" %d->%llu", le32_to_cpu(path->p_idx->ei_block),
ext4_idx_pblock(path->p_idx));
} else if (path->p_ext) {
ext_debug(" %d:[%d]%d:%llu ",
le32_to_cpu(path->p_ext->ee_block),
ext4_ext_is_uninitialized(path->p_ext),
ext4_ext_get_actual_len(path->p_ext),
ext4_ext_pblock(path->p_ext));
} else
ext_debug(" []");
}
ext_debug("\n");
}
static void ext4_ext_show_leaf(struct inode *inode, struct ext4_ext_path *path)
{
int depth = ext_depth(inode);
struct ext4_extent_header *eh;
struct ext4_extent *ex;
int i;
if (!path)
return;
eh = path[depth].p_hdr;
ex = EXT_FIRST_EXTENT(eh);
ext_debug("Displaying leaf extents for inode %lu\n", inode->i_ino);
for (i = 0; i < le16_to_cpu(eh->eh_entries); i++, ex++) {
ext_debug("%d:[%d]%d:%llu ", le32_to_cpu(ex->ee_block),
ext4_ext_is_uninitialized(ex),
ext4_ext_get_actual_len(ex), ext4_ext_pblock(ex));
}
ext_debug("\n");
}
static void ext4_ext_show_move(struct inode *inode, struct ext4_ext_path *path,
ext4_fsblk_t newblock, int level)
{
int depth = ext_depth(inode);
struct ext4_extent *ex;
if (depth != level) {
struct ext4_extent_idx *idx;
idx = path[level].p_idx;
while (idx <= EXT_MAX_INDEX(path[level].p_hdr)) {
ext_debug("%d: move %d:%llu in new index %llu\n", level,
le32_to_cpu(idx->ei_block),
ext4_idx_pblock(idx),
newblock);
idx++;
}
return;
}
ex = path[depth].p_ext;
while (ex <= EXT_MAX_EXTENT(path[depth].p_hdr)) {
ext_debug("move %d:%llu:[%d]%d in new leaf %llu\n",
le32_to_cpu(ex->ee_block),
ext4_ext_pblock(ex),
ext4_ext_is_uninitialized(ex),
ext4_ext_get_actual_len(ex),
newblock);
ex++;
}
}
#else
#define ext4_ext_show_path(inode, path)
#define ext4_ext_show_leaf(inode, path)
#define ext4_ext_show_move(inode, path, newblock, level)
#endif
void ext4_ext_drop_refs(struct ext4_ext_path *path)
{
int depth = path->p_depth;
int i;
for (i = 0; i <= depth; i++, path++)
if (path->p_bh) {
brelse(path->p_bh);
path->p_bh = NULL;
}
}
/*
* ext4_ext_binsearch_idx:
* binary search for the closest index of the given block
* the header must be checked before calling this
*/
static void
ext4_ext_binsearch_idx(struct inode *inode,
struct ext4_ext_path *path, ext4_lblk_t block)
{
struct ext4_extent_header *eh = path->p_hdr;
struct ext4_extent_idx *r, *l, *m;
ext_debug("binsearch for %u(idx): ", block);
l = EXT_FIRST_INDEX(eh) + 1;
r = EXT_LAST_INDEX(eh);
while (l <= r) {
m = l + (r - l) / 2;
if (block < le32_to_cpu(m->ei_block))
r = m - 1;
else
l = m + 1;
ext_debug("%p(%u):%p(%u):%p(%u) ", l, le32_to_cpu(l->ei_block),
m, le32_to_cpu(m->ei_block),
r, le32_to_cpu(r->ei_block));
}
path->p_idx = l - 1;
ext_debug(" -> %u->%lld ", le32_to_cpu(path->p_idx->ei_block),
ext4_idx_pblock(path->p_idx));
#ifdef CHECK_BINSEARCH
{
struct ext4_extent_idx *chix, *ix;
int k;
chix = ix = EXT_FIRST_INDEX(eh);
for (k = 0; k < le16_to_cpu(eh->eh_entries); k++, ix++) {
if (k != 0 &&
le32_to_cpu(ix->ei_block) <= le32_to_cpu(ix[-1].ei_block)) {
printk(KERN_DEBUG "k=%d, ix=0x%p, "
"first=0x%p\n", k,
ix, EXT_FIRST_INDEX(eh));
printk(KERN_DEBUG "%u <= %u\n",
le32_to_cpu(ix->ei_block),
le32_to_cpu(ix[-1].ei_block));
}
BUG_ON(k && le32_to_cpu(ix->ei_block)
<= le32_to_cpu(ix[-1].ei_block));
if (block < le32_to_cpu(ix->ei_block))
break;
chix = ix;
}
BUG_ON(chix != path->p_idx);
}
#endif
}
/*
* ext4_ext_binsearch:
* binary search for closest extent of the given block
* the header must be checked before calling this
*/
static void
ext4_ext_binsearch(struct inode *inode,
struct ext4_ext_path *path, ext4_lblk_t block)
{
struct ext4_extent_header *eh = path->p_hdr;
struct ext4_extent *r, *l, *m;
if (eh->eh_entries == 0) {
/*
* this leaf is empty:
* we get such a leaf in split/add case
*/
return;
}
ext_debug("binsearch for %u: ", block);
l = EXT_FIRST_EXTENT(eh) + 1;
r = EXT_LAST_EXTENT(eh);
while (l <= r) {
m = l + (r - l) / 2;
if (block < le32_to_cpu(m->ee_block))
r = m - 1;
else
l = m + 1;
ext_debug("%p(%u):%p(%u):%p(%u) ", l, le32_to_cpu(l->ee_block),
m, le32_to_cpu(m->ee_block),
r, le32_to_cpu(r->ee_block));
}
path->p_ext = l - 1;
ext_debug(" -> %d:%llu:[%d]%d ",
le32_to_cpu(path->p_ext->ee_block),
ext4_ext_pblock(path->p_ext),
ext4_ext_is_uninitialized(path->p_ext),
ext4_ext_get_actual_len(path->p_ext));
#ifdef CHECK_BINSEARCH
{
struct ext4_extent *chex, *ex;
int k;
chex = ex = EXT_FIRST_EXTENT(eh);
for (k = 0; k < le16_to_cpu(eh->eh_entries); k++, ex++) {
BUG_ON(k && le32_to_cpu(ex->ee_block)
<= le32_to_cpu(ex[-1].ee_block));
if (block < le32_to_cpu(ex->ee_block))
break;
chex = ex;
}
BUG_ON(chex != path->p_ext);
}
#endif
}
int ext4_ext_tree_init(handle_t *handle, struct inode *inode)
{
struct ext4_extent_header *eh;
eh = ext_inode_hdr(inode);
eh->eh_depth = 0;
eh->eh_entries = 0;
eh->eh_magic = EXT4_EXT_MAGIC;
eh->eh_max = cpu_to_le16(ext4_ext_space_root(inode, 0));
ext4_mark_inode_dirty(handle, inode);
return 0;
}
struct ext4_ext_path *
ext4_ext_find_extent(struct inode *inode, ext4_lblk_t block,
struct ext4_ext_path *path)
{
struct ext4_extent_header *eh;
struct buffer_head *bh;
short int depth, i, ppos = 0, alloc = 0;
int ret;
eh = ext_inode_hdr(inode);
depth = ext_depth(inode);
/* account possible depth increase */
if (!path) {
path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 2),
GFP_NOFS);
if (!path)
return ERR_PTR(-ENOMEM);
alloc = 1;
}
path[0].p_hdr = eh;
path[0].p_bh = NULL;
i = depth;
/* walk through the tree */
while (i) {
ext_debug("depth %d: num %d, max %d\n",
ppos, le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max));
ext4_ext_binsearch_idx(inode, path + ppos, block);
path[ppos].p_block = ext4_idx_pblock(path[ppos].p_idx);
path[ppos].p_depth = i;
path[ppos].p_ext = NULL;
bh = sb_getblk(inode->i_sb, path[ppos].p_block);
if (unlikely(!bh)) {
ret = -ENOMEM;
goto err;
}
if (!bh_uptodate_or_lock(bh)) {
trace_ext4_ext_load_extent(inode, block,
path[ppos].p_block);
ret = bh_submit_read(bh);
if (ret < 0) {
put_bh(bh);
goto err;
}
}
eh = ext_block_hdr(bh);
ppos++;
if (unlikely(ppos > depth)) {
put_bh(bh);
EXT4_ERROR_INODE(inode,
"ppos %d > depth %d", ppos, depth);
ret = -EIO;
goto err;
}
path[ppos].p_bh = bh;
path[ppos].p_hdr = eh;
i--;
ret = ext4_ext_check_block(inode, eh, i, bh);
if (ret < 0)
goto err;
}
path[ppos].p_depth = i;
path[ppos].p_ext = NULL;
path[ppos].p_idx = NULL;
/* find extent */
ext4_ext_binsearch(inode, path + ppos, block);
/* if not an empty leaf */
if (path[ppos].p_ext)
path[ppos].p_block = ext4_ext_pblock(path[ppos].p_ext);
ext4_ext_show_path(inode, path);
return path;
err:
ext4_ext_drop_refs(path);
if (alloc)
kfree(path);
return ERR_PTR(ret);
}
/*
* ext4_ext_insert_index:
* insert new index [@logical;@ptr] into the block at @curp;
* check where to insert: before @curp or after @curp
*/
static int ext4_ext_insert_index(handle_t *handle, struct inode *inode,
struct ext4_ext_path *curp,
int logical, ext4_fsblk_t ptr)
{
struct ext4_extent_idx *ix;
int len, err;
err = ext4_ext_get_access(handle, inode, curp);
if (err)
return err;
if (unlikely(logical == le32_to_cpu(curp->p_idx->ei_block))) {
EXT4_ERROR_INODE(inode,
"logical %d == ei_block %d!",
logical, le32_to_cpu(curp->p_idx->ei_block));
return -EIO;
}
if (unlikely(le16_to_cpu(curp->p_hdr->eh_entries)
>= le16_to_cpu(curp->p_hdr->eh_max))) {
EXT4_ERROR_INODE(inode,
"eh_entries %d >= eh_max %d!",
le16_to_cpu(curp->p_hdr->eh_entries),
le16_to_cpu(curp->p_hdr->eh_max));
return -EIO;
}
if (logical > le32_to_cpu(curp->p_idx->ei_block)) {
/* insert after */
ext_debug("insert new index %d after: %llu\n", logical, ptr);
ix = curp->p_idx + 1;
} else {
/* insert before */
ext_debug("insert new index %d before: %llu\n", logical, ptr);
ix = curp->p_idx;
}
len = EXT_LAST_INDEX(curp->p_hdr) - ix + 1;
BUG_ON(len < 0);
if (len > 0) {
ext_debug("insert new index %d: "
"move %d indices from 0x%p to 0x%p\n",
logical, len, ix, ix + 1);
memmove(ix + 1, ix, len * sizeof(struct ext4_extent_idx));
}
if (unlikely(ix > EXT_MAX_INDEX(curp->p_hdr))) {
EXT4_ERROR_INODE(inode, "ix > EXT_MAX_INDEX!");
return -EIO;
}
ix->ei_block = cpu_to_le32(logical);
ext4_idx_store_pblock(ix, ptr);
le16_add_cpu(&curp->p_hdr->eh_entries, 1);
if (unlikely(ix > EXT_LAST_INDEX(curp->p_hdr))) {
EXT4_ERROR_INODE(inode, "ix > EXT_LAST_INDEX!");
return -EIO;
}
err = ext4_ext_dirty(handle, inode, curp);
ext4_std_error(inode->i_sb, err);
return err;
}
/*
* ext4_ext_split:
* inserts new subtree into the path, using free index entry
* at depth @at:
* - allocates all needed blocks (new leaf and all intermediate index blocks)
* - makes decision where to split
* - moves remaining extents and index entries (right to the split point)
* into the newly allocated blocks
* - initializes subtree
*/
static int ext4_ext_split(handle_t *handle, struct inode *inode,
unsigned int flags,
struct ext4_ext_path *path,
struct ext4_extent *newext, int at)
{
struct buffer_head *bh = NULL;
int depth = ext_depth(inode);
struct ext4_extent_header *neh;
struct ext4_extent_idx *fidx;
int i = at, k, m, a;
ext4_fsblk_t newblock, oldblock;
__le32 border;
ext4_fsblk_t *ablocks = NULL; /* array of allocated blocks */
int err = 0;
/* make decision: where to split? */
/* FIXME: now decision is simplest: at current extent */
/* if current leaf will be split, then we should use
* border from split point */
if (unlikely(path[depth].p_ext > EXT_MAX_EXTENT(path[depth].p_hdr))) {
EXT4_ERROR_INODE(inode, "p_ext > EXT_MAX_EXTENT!");
return -EIO;
}
if (path[depth].p_ext != EXT_MAX_EXTENT(path[depth].p_hdr)) {
border = path[depth].p_ext[1].ee_block;
ext_debug("leaf will be split."
" next leaf starts at %d\n",
le32_to_cpu(border));
} else {
border = newext->ee_block;
ext_debug("leaf will be added."
" next leaf starts at %d\n",
le32_to_cpu(border));
}
/*
* If error occurs, then we break processing
* and mark filesystem read-only. index won't
* be inserted and tree will be in consistent
* state. Next mount will repair buffers too.
*/
/*
* Get array to track all allocated blocks.
* We need this to handle errors and free blocks
* upon them.
*/
ablocks = kzalloc(sizeof(ext4_fsblk_t) * depth, GFP_NOFS);
if (!ablocks)
return -ENOMEM;
/* allocate all needed blocks */
ext_debug("allocate %d blocks for indexes/leaf\n", depth - at);
for (a = 0; a < depth - at; a++) {
newblock = ext4_ext_new_meta_block(handle, inode, path,
newext, &err, flags);
if (newblock == 0)
goto cleanup;
ablocks[a] = newblock;
}
/* initialize new leaf */
newblock = ablocks[--a];
if (unlikely(newblock == 0)) {
EXT4_ERROR_INODE(inode, "newblock == 0!");
err = -EIO;
goto cleanup;
}
bh = sb_getblk(inode->i_sb, newblock);
if (unlikely(!bh)) {
err = -ENOMEM;
goto cleanup;
}
lock_buffer(bh);
err = ext4_journal_get_create_access(handle, bh);
if (err)
goto cleanup;
neh = ext_block_hdr(bh);
neh->eh_entries = 0;
neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode, 0));
neh->eh_magic = EXT4_EXT_MAGIC;
neh->eh_depth = 0;
/* move remainder of path[depth] to the new leaf */
if (unlikely(path[depth].p_hdr->eh_entries !=
path[depth].p_hdr->eh_max)) {
EXT4_ERROR_INODE(inode, "eh_entries %d != eh_max %d!",
path[depth].p_hdr->eh_entries,
path[depth].p_hdr->eh_max);
err = -EIO;
goto cleanup;
}
/* start copy from next extent */
m = EXT_MAX_EXTENT(path[depth].p_hdr) - path[depth].p_ext++;
ext4_ext_show_move(inode, path, newblock, depth);
if (m) {
struct ext4_extent *ex;
ex = EXT_FIRST_EXTENT(neh);
memmove(ex, path[depth].p_ext, sizeof(struct ext4_extent) * m);
le16_add_cpu(&neh->eh_entries, m);
}
ext4_extent_block_csum_set(inode, neh);
set_buffer_uptodate(bh);
unlock_buffer(bh);
err = ext4_handle_dirty_metadata(handle, inode, bh);
if (err)
goto cleanup;
brelse(bh);
bh = NULL;
/* correct old leaf */
if (m) {
err = ext4_ext_get_access(handle, inode, path + depth);
if (err)
goto cleanup;
le16_add_cpu(&path[depth].p_hdr->eh_entries, -m);
err = ext4_ext_dirty(handle, inode, path + depth);
if (err)
goto cleanup;
}
/* create intermediate indexes */
k = depth - at - 1;
if (unlikely(k < 0)) {
EXT4_ERROR_INODE(inode, "k %d < 0!", k);
err = -EIO;
goto cleanup;
}
if (k)
ext_debug("create %d intermediate indices\n", k);
/* insert new index into current index block */
/* current depth stored in i var */
i = depth - 1;
while (k--) {
oldblock = newblock;
newblock = ablocks[--a];
bh = sb_getblk(inode->i_sb, newblock);
if (unlikely(!bh)) {
err = -ENOMEM;
goto cleanup;
}
lock_buffer(bh);
err = ext4_journal_get_create_access(handle, bh);
if (err)
goto cleanup;
neh = ext_block_hdr(bh);
neh->eh_entries = cpu_to_le16(1);
neh->eh_magic = EXT4_EXT_MAGIC;
neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode, 0));
neh->eh_depth = cpu_to_le16(depth - i);
fidx = EXT_FIRST_INDEX(neh);
fidx->ei_block = border;
ext4_idx_store_pblock(fidx, oldblock);
ext_debug("int.index at %d (block %llu): %u -> %llu\n",
i, newblock, le32_to_cpu(border), oldblock);
/* move remainder of path[i] to the new index block */
if (unlikely(EXT_MAX_INDEX(path[i].p_hdr) !=
EXT_LAST_INDEX(path[i].p_hdr))) {
EXT4_ERROR_INODE(inode,
"EXT_MAX_INDEX != EXT_LAST_INDEX ee_block %d!",
le32_to_cpu(path[i].p_ext->ee_block));
err = -EIO;
goto cleanup;
}
/* start copy indexes */
m = EXT_MAX_INDEX(path[i].p_hdr) - path[i].p_idx++;
ext_debug("cur 0x%p, last 0x%p\n", path[i].p_idx,
EXT_MAX_INDEX(path[i].p_hdr));
ext4_ext_show_move(inode, path, newblock, i);
if (m) {
memmove(++fidx, path[i].p_idx,
sizeof(struct ext4_extent_idx) * m);
le16_add_cpu(&neh->eh_entries, m);
}
ext4_extent_block_csum_set(inode, neh);
set_buffer_uptodate(bh);
unlock_buffer(bh);
err = ext4_handle_dirty_metadata(handle, inode, bh);
if (err)
goto cleanup;
brelse(bh);
bh = NULL;
/* correct old index */
if (m) {
err = ext4_ext_get_access(handle, inode, path + i);
if (err)
goto cleanup;
le16_add_cpu(&path[i].p_hdr->eh_entries, -m);
err = ext4_ext_dirty(handle, inode, path + i);
if (err)
goto cleanup;
}
i--;
}
/* insert new index */
err = ext4_ext_insert_index(handle, inode, path + at,
le32_to_cpu(border), newblock);
cleanup:
if (bh) {
if (buffer_locked(bh))
unlock_buffer(bh);
brelse(bh);
}
if (err) {
/* free all allocated blocks in error case */
for (i = 0; i < depth; i++) {
if (!ablocks[i])
continue;
ext4_free_blocks(handle, inode, NULL, ablocks[i], 1,
EXT4_FREE_BLOCKS_METADATA);
}
}
kfree(ablocks);
return err;
}
/*
* ext4_ext_grow_indepth:
* implements tree growing procedure:
* - allocates new block
* - moves top-level data (index block or leaf) into the new block
* - initializes new top-level, creating index that points to the
* just created block
*/
static int ext4_ext_grow_indepth(handle_t *handle, struct inode *inode,
unsigned int flags,
struct ext4_extent *newext)
{
struct ext4_extent_header *neh;
struct buffer_head *bh;
ext4_fsblk_t newblock;
int err = 0;
newblock = ext4_ext_new_meta_block(handle, inode, NULL,
newext, &err, flags);
if (newblock == 0)
return err;
bh = sb_getblk(inode->i_sb, newblock);
if (unlikely(!bh))
return -ENOMEM;
lock_buffer(bh);
err = ext4_journal_get_create_access(handle, bh);
if (err) {
unlock_buffer(bh);
goto out;
}
/* move top-level index/leaf into new block */
memmove(bh->b_data, EXT4_I(inode)->i_data,
sizeof(EXT4_I(inode)->i_data));
/* set size of new block */
neh = ext_block_hdr(bh);
/* old root could have indexes or leaves
* so calculate e_max right way */
if (ext_depth(inode))
neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode, 0));
else
neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode, 0));
neh->eh_magic = EXT4_EXT_MAGIC;
ext4_extent_block_csum_set(inode, neh);
set_buffer_uptodate(bh);
unlock_buffer(bh);
err = ext4_handle_dirty_metadata(handle, inode, bh);
if (err)
goto out;
/* Update top-level index: num,max,pointer */
neh = ext_inode_hdr(inode);
neh->eh_entries = cpu_to_le16(1);
ext4_idx_store_pblock(EXT_FIRST_INDEX(neh), newblock);
if (neh->eh_depth == 0) {
/* Root extent block becomes index block */
neh->eh_max = cpu_to_le16(ext4_ext_space_root_idx(inode, 0));
EXT_FIRST_INDEX(neh)->ei_block =
EXT_FIRST_EXTENT(neh)->ee_block;
}
ext_debug("new root: num %d(%d), lblock %d, ptr %llu\n",
le16_to_cpu(neh->eh_entries), le16_to_cpu(neh->eh_max),
le32_to_cpu(EXT_FIRST_INDEX(neh)->ei_block),
ext4_idx_pblock(EXT_FIRST_INDEX(neh)));
le16_add_cpu(&neh->eh_depth, 1);
ext4_mark_inode_dirty(handle, inode);
out:
brelse(bh);
return err;
}
/*
* ext4_ext_create_new_leaf:
* finds empty index and adds new leaf.
* if no free index is found, then it requests in-depth growing.
*/
static int ext4_ext_create_new_leaf(handle_t *handle, struct inode *inode,
unsigned int flags,
struct ext4_ext_path *path,
struct ext4_extent *newext)
{
struct ext4_ext_path *curp;
int depth, i, err = 0;
repeat:
i = depth = ext_depth(inode);
/* walk up to the tree and look for free index entry */
curp = path + depth;
while (i > 0 && !EXT_HAS_FREE_INDEX(curp)) {
i--;
curp--;
}
/* we use already allocated block for index block,
* so subsequent data blocks should be contiguous */
if (EXT_HAS_FREE_INDEX(curp)) {
/* if we found index with free entry, then use that
* entry: create all needed subtree and add new leaf */
err = ext4_ext_split(handle, inode, flags, path, newext, i);
if (err)
goto out;
/* refill path */
ext4_ext_drop_refs(path);
path = ext4_ext_find_extent(inode,
(ext4_lblk_t)le32_to_cpu(newext->ee_block),
path);
if (IS_ERR(path))
err = PTR_ERR(path);
} else {
/* tree is full, time to grow in depth */
err = ext4_ext_grow_indepth(handle, inode, flags, newext);
if (err)
goto out;
/* refill path */
ext4_ext_drop_refs(path);
path = ext4_ext_find_extent(inode,
(ext4_lblk_t)le32_to_cpu(newext->ee_block),
path);
if (IS_ERR(path)) {
err = PTR_ERR(path);
goto out;
}
/*
* only first (depth 0 -> 1) produces free space;
* in all other cases we have to split the grown tree
*/
depth = ext_depth(inode);
if (path[depth].p_hdr->eh_entries == path[depth].p_hdr->eh_max) {
/* now we need to split */
goto repeat;
}
}
out:
return err;
}
/*
* search the closest allocated block to the left for *logical
* and returns it at @logical + it's physical address at @phys
* if *logical is the smallest allocated block, the function
* returns 0 at @phys
* return value contains 0 (success) or error code
*/
static int ext4_ext_search_left(struct inode *inode,
struct ext4_ext_path *path,
ext4_lblk_t *logical, ext4_fsblk_t *phys)
{
struct ext4_extent_idx *ix;
struct ext4_extent *ex;
int depth, ee_len;
if (unlikely(path == NULL)) {
EXT4_ERROR_INODE(inode, "path == NULL *logical %d!", *logical);
return -EIO;
}
depth = path->p_depth;
*phys = 0;
if (depth == 0 && path->p_ext == NULL)
return 0;
/* usually extent in the path covers blocks smaller
* then *logical, but it can be that extent is the
* first one in the file */
ex = path[depth].p_ext;
ee_len = ext4_ext_get_actual_len(ex);
if (*logical < le32_to_cpu(ex->ee_block)) {
if (unlikely(EXT_FIRST_EXTENT(path[depth].p_hdr) != ex)) {
EXT4_ERROR_INODE(inode,
"EXT_FIRST_EXTENT != ex *logical %d ee_block %d!",
*logical, le32_to_cpu(ex->ee_block));
return -EIO;
}
while (--depth >= 0) {
ix = path[depth].p_idx;
if (unlikely(ix != EXT_FIRST_INDEX(path[depth].p_hdr))) {
EXT4_ERROR_INODE(inode,
"ix (%d) != EXT_FIRST_INDEX (%d) (depth %d)!",
ix != NULL ? le32_to_cpu(ix->ei_block) : 0,
EXT_FIRST_INDEX(path[depth].p_hdr) != NULL ?
le32_to_cpu(EXT_FIRST_INDEX(path[depth].p_hdr)->ei_block) : 0,
depth);
return -EIO;
}
}
return 0;
}
if (unlikely(*logical < (le32_to_cpu(ex->ee_block) + ee_len))) {
EXT4_ERROR_INODE(inode,
"logical %d < ee_block %d + ee_len %d!",
*logical, le32_to_cpu(ex->ee_block), ee_len);
return -EIO;
}
*logical = le32_to_cpu(ex->ee_block) + ee_len - 1;
*phys = ext4_ext_pblock(ex) + ee_len - 1;
return 0;
}
/*
* search the closest allocated block to the right for *logical
* and returns it at @logical + it's physical address at @phys
* if *logical is the largest allocated block, the function
* returns 0 at @phys
* return value contains 0 (success) or error code
*/
static int ext4_ext_search_right(struct inode *inode,
struct ext4_ext_path *path,
ext4_lblk_t *logical, ext4_fsblk_t *phys,
struct ext4_extent **ret_ex)
{
struct buffer_head *bh = NULL;
struct ext4_extent_header *eh;
struct ext4_extent_idx *ix;
struct ext4_extent *ex;
ext4_fsblk_t block;
int depth; /* Note, NOT eh_depth; depth from top of tree */
int ee_len;
if (unlikely(path == NULL)) {
EXT4_ERROR_INODE(inode, "path == NULL *logical %d!", *logical);
return -EIO;
}
depth = path->p_depth;
*phys = 0;
if (depth == 0 && path->p_ext == NULL)
return 0;
/* usually extent in the path covers blocks smaller
* then *logical, but it can be that extent is the
* first one in the file */
ex = path[depth].p_ext;
ee_len = ext4_ext_get_actual_len(ex);
if (*logical < le32_to_cpu(ex->ee_block)) {
if (unlikely(EXT_FIRST_EXTENT(path[depth].p_hdr) != ex)) {
EXT4_ERROR_INODE(inode,
"first_extent(path[%d].p_hdr) != ex",
depth);
return -EIO;
}
while (--depth >= 0) {
ix = path[depth].p_idx;
if (unlikely(ix != EXT_FIRST_INDEX(path[depth].p_hdr))) {
EXT4_ERROR_INODE(inode,
"ix != EXT_FIRST_INDEX *logical %d!",
*logical);
return -EIO;
}
}
goto found_extent;
}
if (unlikely(*logical < (le32_to_cpu(ex->ee_block) + ee_len))) {
EXT4_ERROR_INODE(inode,
"logical %d < ee_block %d + ee_len %d!",
*logical, le32_to_cpu(ex->ee_block), ee_len);
return -EIO;
}
if (ex != EXT_LAST_EXTENT(path[depth].p_hdr)) {
/* next allocated block in this leaf */
ex++;
goto found_extent;
}
/* go up and search for index to the right */
while (--depth >= 0) {
ix = path[depth].p_idx;
if (ix != EXT_LAST_INDEX(path[depth].p_hdr))
goto got_index;
}
/* we've gone up to the root and found no index to the right */
return 0;
got_index:
/* we've found index to the right, let's
* follow it and find the closest allocated
* block to the right */
ix++;
block = ext4_idx_pblock(ix);
while (++depth < path->p_depth) {
bh = sb_bread(inode->i_sb, block);
if (bh == NULL)
return -EIO;
eh = ext_block_hdr(bh);
/* subtract from p_depth to get proper eh_depth */
if (ext4_ext_check_block(inode, eh,
path->p_depth - depth, bh)) {
put_bh(bh);
return -EIO;
}
ix = EXT_FIRST_INDEX(eh);
block = ext4_idx_pblock(ix);
put_bh(bh);
}
bh = sb_bread(inode->i_sb, block);
if (bh == NULL)
return -EIO;
eh = ext_block_hdr(bh);
if (ext4_ext_check_block(inode, eh, path->p_depth - depth, bh)) {
put_bh(bh);
return -EIO;
}
ex = EXT_FIRST_EXTENT(eh);
found_extent:
*logical = le32_to_cpu(ex->ee_block);
*phys = ext4_ext_pblock(ex);
*ret_ex = ex;
if (bh)
put_bh(bh);
return 0;
}
/*
* ext4_ext_next_allocated_block:
ext4: Fix max file size and logical block counting of extent format file Kazuya Mio reported that he was able to hit BUG_ON(next == lblock) in ext4_ext_put_gap_in_cache() while creating a sparse file in extent format and fill the tail of file up to its end. We will hit the BUG_ON when we write the last block (2^32-1) into the sparse file. The root cause of the problem lies in the fact that we specifically set s_maxbytes so that block at s_maxbytes fit into on-disk extent format, which is 32 bit long. However, we are not storing start and end block number, but rather start block number and length in blocks. It means that in order to cover extent from 0 to EXT_MAX_BLOCK we need EXT_MAX_BLOCK+1 to fit into len (because we counting block 0 as well) - and it does not. The only way to fix it without changing the meaning of the struct ext4_extent members is, as Kazuya Mio suggested, to lower s_maxbytes by one fs block so we can cover the whole extent we can get by the on-disk extent format. Also in many places EXT_MAX_BLOCK is used as length instead of maximum logical block number as the name suggests, it is all a bit messy. So this commit renames it to EXT_MAX_BLOCKS and change its usage in some places to actually be maximum number of blocks in the extent. The bug which this commit fixes can be reproduced as follows: dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-2)) sync dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-1)) Reported-by: Kazuya Mio <k-mio@sx.jp.nec.com> Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-06-06 12:05:17 +08:00
* returns allocated block in subsequent extent or EXT_MAX_BLOCKS.
* NOTE: it considers block number from index entry as
* allocated block. Thus, index entries have to be consistent
* with leaves.
*/
static ext4_lblk_t
ext4_ext_next_allocated_block(struct ext4_ext_path *path)
{
int depth;
BUG_ON(path == NULL);
depth = path->p_depth;
if (depth == 0 && path->p_ext == NULL)
ext4: Fix max file size and logical block counting of extent format file Kazuya Mio reported that he was able to hit BUG_ON(next == lblock) in ext4_ext_put_gap_in_cache() while creating a sparse file in extent format and fill the tail of file up to its end. We will hit the BUG_ON when we write the last block (2^32-1) into the sparse file. The root cause of the problem lies in the fact that we specifically set s_maxbytes so that block at s_maxbytes fit into on-disk extent format, which is 32 bit long. However, we are not storing start and end block number, but rather start block number and length in blocks. It means that in order to cover extent from 0 to EXT_MAX_BLOCK we need EXT_MAX_BLOCK+1 to fit into len (because we counting block 0 as well) - and it does not. The only way to fix it without changing the meaning of the struct ext4_extent members is, as Kazuya Mio suggested, to lower s_maxbytes by one fs block so we can cover the whole extent we can get by the on-disk extent format. Also in many places EXT_MAX_BLOCK is used as length instead of maximum logical block number as the name suggests, it is all a bit messy. So this commit renames it to EXT_MAX_BLOCKS and change its usage in some places to actually be maximum number of blocks in the extent. The bug which this commit fixes can be reproduced as follows: dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-2)) sync dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-1)) Reported-by: Kazuya Mio <k-mio@sx.jp.nec.com> Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-06-06 12:05:17 +08:00
return EXT_MAX_BLOCKS;
while (depth >= 0) {
if (depth == path->p_depth) {
/* leaf */
if (path[depth].p_ext &&
path[depth].p_ext !=
EXT_LAST_EXTENT(path[depth].p_hdr))
return le32_to_cpu(path[depth].p_ext[1].ee_block);
} else {
/* index */
if (path[depth].p_idx !=
EXT_LAST_INDEX(path[depth].p_hdr))
return le32_to_cpu(path[depth].p_idx[1].ei_block);
}
depth--;
}
ext4: Fix max file size and logical block counting of extent format file Kazuya Mio reported that he was able to hit BUG_ON(next == lblock) in ext4_ext_put_gap_in_cache() while creating a sparse file in extent format and fill the tail of file up to its end. We will hit the BUG_ON when we write the last block (2^32-1) into the sparse file. The root cause of the problem lies in the fact that we specifically set s_maxbytes so that block at s_maxbytes fit into on-disk extent format, which is 32 bit long. However, we are not storing start and end block number, but rather start block number and length in blocks. It means that in order to cover extent from 0 to EXT_MAX_BLOCK we need EXT_MAX_BLOCK+1 to fit into len (because we counting block 0 as well) - and it does not. The only way to fix it without changing the meaning of the struct ext4_extent members is, as Kazuya Mio suggested, to lower s_maxbytes by one fs block so we can cover the whole extent we can get by the on-disk extent format. Also in many places EXT_MAX_BLOCK is used as length instead of maximum logical block number as the name suggests, it is all a bit messy. So this commit renames it to EXT_MAX_BLOCKS and change its usage in some places to actually be maximum number of blocks in the extent. The bug which this commit fixes can be reproduced as follows: dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-2)) sync dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-1)) Reported-by: Kazuya Mio <k-mio@sx.jp.nec.com> Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-06-06 12:05:17 +08:00
return EXT_MAX_BLOCKS;
}
/*
* ext4_ext_next_leaf_block:
ext4: Fix max file size and logical block counting of extent format file Kazuya Mio reported that he was able to hit BUG_ON(next == lblock) in ext4_ext_put_gap_in_cache() while creating a sparse file in extent format and fill the tail of file up to its end. We will hit the BUG_ON when we write the last block (2^32-1) into the sparse file. The root cause of the problem lies in the fact that we specifically set s_maxbytes so that block at s_maxbytes fit into on-disk extent format, which is 32 bit long. However, we are not storing start and end block number, but rather start block number and length in blocks. It means that in order to cover extent from 0 to EXT_MAX_BLOCK we need EXT_MAX_BLOCK+1 to fit into len (because we counting block 0 as well) - and it does not. The only way to fix it without changing the meaning of the struct ext4_extent members is, as Kazuya Mio suggested, to lower s_maxbytes by one fs block so we can cover the whole extent we can get by the on-disk extent format. Also in many places EXT_MAX_BLOCK is used as length instead of maximum logical block number as the name suggests, it is all a bit messy. So this commit renames it to EXT_MAX_BLOCKS and change its usage in some places to actually be maximum number of blocks in the extent. The bug which this commit fixes can be reproduced as follows: dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-2)) sync dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-1)) Reported-by: Kazuya Mio <k-mio@sx.jp.nec.com> Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-06-06 12:05:17 +08:00
* returns first allocated block from next leaf or EXT_MAX_BLOCKS
*/
static ext4_lblk_t ext4_ext_next_leaf_block(struct ext4_ext_path *path)
{
int depth;
BUG_ON(path == NULL);
depth = path->p_depth;
/* zero-tree has no leaf blocks at all */
if (depth == 0)
ext4: Fix max file size and logical block counting of extent format file Kazuya Mio reported that he was able to hit BUG_ON(next == lblock) in ext4_ext_put_gap_in_cache() while creating a sparse file in extent format and fill the tail of file up to its end. We will hit the BUG_ON when we write the last block (2^32-1) into the sparse file. The root cause of the problem lies in the fact that we specifically set s_maxbytes so that block at s_maxbytes fit into on-disk extent format, which is 32 bit long. However, we are not storing start and end block number, but rather start block number and length in blocks. It means that in order to cover extent from 0 to EXT_MAX_BLOCK we need EXT_MAX_BLOCK+1 to fit into len (because we counting block 0 as well) - and it does not. The only way to fix it without changing the meaning of the struct ext4_extent members is, as Kazuya Mio suggested, to lower s_maxbytes by one fs block so we can cover the whole extent we can get by the on-disk extent format. Also in many places EXT_MAX_BLOCK is used as length instead of maximum logical block number as the name suggests, it is all a bit messy. So this commit renames it to EXT_MAX_BLOCKS and change its usage in some places to actually be maximum number of blocks in the extent. The bug which this commit fixes can be reproduced as follows: dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-2)) sync dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-1)) Reported-by: Kazuya Mio <k-mio@sx.jp.nec.com> Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-06-06 12:05:17 +08:00
return EXT_MAX_BLOCKS;
/* go to index block */
depth--;
while (depth >= 0) {
if (path[depth].p_idx !=
EXT_LAST_INDEX(path[depth].p_hdr))
return (ext4_lblk_t)
le32_to_cpu(path[depth].p_idx[1].ei_block);
depth--;
}
ext4: Fix max file size and logical block counting of extent format file Kazuya Mio reported that he was able to hit BUG_ON(next == lblock) in ext4_ext_put_gap_in_cache() while creating a sparse file in extent format and fill the tail of file up to its end. We will hit the BUG_ON when we write the last block (2^32-1) into the sparse file. The root cause of the problem lies in the fact that we specifically set s_maxbytes so that block at s_maxbytes fit into on-disk extent format, which is 32 bit long. However, we are not storing start and end block number, but rather start block number and length in blocks. It means that in order to cover extent from 0 to EXT_MAX_BLOCK we need EXT_MAX_BLOCK+1 to fit into len (because we counting block 0 as well) - and it does not. The only way to fix it without changing the meaning of the struct ext4_extent members is, as Kazuya Mio suggested, to lower s_maxbytes by one fs block so we can cover the whole extent we can get by the on-disk extent format. Also in many places EXT_MAX_BLOCK is used as length instead of maximum logical block number as the name suggests, it is all a bit messy. So this commit renames it to EXT_MAX_BLOCKS and change its usage in some places to actually be maximum number of blocks in the extent. The bug which this commit fixes can be reproduced as follows: dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-2)) sync dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-1)) Reported-by: Kazuya Mio <k-mio@sx.jp.nec.com> Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-06-06 12:05:17 +08:00
return EXT_MAX_BLOCKS;
}
/*
* ext4_ext_correct_indexes:
* if leaf gets modified and modified extent is first in the leaf,
* then we have to correct all indexes above.
* TODO: do we need to correct tree in all cases?
*/
static int ext4_ext_correct_indexes(handle_t *handle, struct inode *inode,
struct ext4_ext_path *path)
{
struct ext4_extent_header *eh;
int depth = ext_depth(inode);
struct ext4_extent *ex;
__le32 border;
int k, err = 0;
eh = path[depth].p_hdr;
ex = path[depth].p_ext;
if (unlikely(ex == NULL || eh == NULL)) {
EXT4_ERROR_INODE(inode,
"ex %p == NULL or eh %p == NULL", ex, eh);
return -EIO;
}
if (depth == 0) {
/* there is no tree at all */
return 0;
}
if (ex != EXT_FIRST_EXTENT(eh)) {
/* we correct tree if first leaf got modified only */
return 0;
}
/*
* TODO: we need correction if border is smaller than current one
*/
k = depth - 1;
border = path[depth].p_ext->ee_block;
err = ext4_ext_get_access(handle, inode, path + k);
if (err)
return err;
path[k].p_idx->ei_block = border;
err = ext4_ext_dirty(handle, inode, path + k);
if (err)
return err;
while (k--) {
/* change all left-side indexes */
if (path[k+1].p_idx != EXT_FIRST_INDEX(path[k+1].p_hdr))
break;
err = ext4_ext_get_access(handle, inode, path + k);
if (err)
break;
path[k].p_idx->ei_block = border;
err = ext4_ext_dirty(handle, inode, path + k);
if (err)
break;
}
return err;
}
int
ext4_can_extents_be_merged(struct inode *inode, struct ext4_extent *ex1,
struct ext4_extent *ex2)
{
unsigned short ext1_ee_len, ext2_ee_len, max_len;
/*
* Make sure that both extents are initialized. We don't merge
* uninitialized extents so that we can be sure that end_io code has
* the extent that was written properly split out and conversion to
* initialized is trivial.
*/
if (ext4_ext_is_uninitialized(ex1) || ext4_ext_is_uninitialized(ex2))
return 0;
if (ext4_ext_is_uninitialized(ex1))
max_len = EXT_UNINIT_MAX_LEN;
else
max_len = EXT_INIT_MAX_LEN;
ext1_ee_len = ext4_ext_get_actual_len(ex1);
ext2_ee_len = ext4_ext_get_actual_len(ex2);
if (le32_to_cpu(ex1->ee_block) + ext1_ee_len !=
le32_to_cpu(ex2->ee_block))
return 0;
/*
* To allow future support for preallocated extents to be added
* as an RO_COMPAT feature, refuse to merge to extents if
* this can result in the top bit of ee_len being set.
*/
if (ext1_ee_len + ext2_ee_len > max_len)
return 0;
#ifdef AGGRESSIVE_TEST
if (ext1_ee_len >= 4)
return 0;
#endif
if (ext4_ext_pblock(ex1) + ext1_ee_len == ext4_ext_pblock(ex2))
return 1;
return 0;
}
/*
* This function tries to merge the "ex" extent to the next extent in the tree.
* It always tries to merge towards right. If you want to merge towards
* left, pass "ex - 1" as argument instead of "ex".
* Returns 0 if the extents (ex and ex+1) were _not_ merged and returns
* 1 if they got merged.
*/
static int ext4_ext_try_to_merge_right(struct inode *inode,
struct ext4_ext_path *path,
struct ext4_extent *ex)
{
struct ext4_extent_header *eh;
unsigned int depth, len;
int merge_done = 0;
int uninitialized = 0;
depth = ext_depth(inode);
BUG_ON(path[depth].p_hdr == NULL);
eh = path[depth].p_hdr;
while (ex < EXT_LAST_EXTENT(eh)) {
if (!ext4_can_extents_be_merged(inode, ex, ex + 1))
break;
/* merge with next extent! */
if (ext4_ext_is_uninitialized(ex))
uninitialized = 1;
ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex)
+ ext4_ext_get_actual_len(ex + 1));
if (uninitialized)
ext4_ext_mark_uninitialized(ex);
if (ex + 1 < EXT_LAST_EXTENT(eh)) {
len = (EXT_LAST_EXTENT(eh) - ex - 1)
* sizeof(struct ext4_extent);
memmove(ex + 1, ex + 2, len);
}
le16_add_cpu(&eh->eh_entries, -1);
merge_done = 1;
WARN_ON(eh->eh_entries == 0);
if (!eh->eh_entries)
EXT4_ERROR_INODE(inode, "eh->eh_entries = 0!");
}
return merge_done;
}
/*
* This function does a very simple check to see if we can collapse
* an extent tree with a single extent tree leaf block into the inode.
*/
static void ext4_ext_try_to_merge_up(handle_t *handle,
struct inode *inode,
struct ext4_ext_path *path)
{
size_t s;
unsigned max_root = ext4_ext_space_root(inode, 0);
ext4_fsblk_t blk;
if ((path[0].p_depth != 1) ||
(le16_to_cpu(path[0].p_hdr->eh_entries) != 1) ||
(le16_to_cpu(path[1].p_hdr->eh_entries) > max_root))
return;
/*
* We need to modify the block allocation bitmap and the block
* group descriptor to release the extent tree block. If we
* can't get the journal credits, give up.
*/
if (ext4_journal_extend(handle, 2))
return;
/*
* Copy the extent data up to the inode
*/
blk = ext4_idx_pblock(path[0].p_idx);
s = le16_to_cpu(path[1].p_hdr->eh_entries) *
sizeof(struct ext4_extent_idx);
s += sizeof(struct ext4_extent_header);
memcpy(path[0].p_hdr, path[1].p_hdr, s);
path[0].p_depth = 0;
path[0].p_ext = EXT_FIRST_EXTENT(path[0].p_hdr) +
(path[1].p_ext - EXT_FIRST_EXTENT(path[1].p_hdr));
path[0].p_hdr->eh_max = cpu_to_le16(max_root);
brelse(path[1].p_bh);
ext4_free_blocks(handle, inode, NULL, blk, 1,
EXT4_FREE_BLOCKS_METADATA | EXT4_FREE_BLOCKS_FORGET);
}
/*
* This function tries to merge the @ex extent to neighbours in the tree.
* return 1 if merge left else 0.
*/
static void ext4_ext_try_to_merge(handle_t *handle,
struct inode *inode,
struct ext4_ext_path *path,
struct ext4_extent *ex) {
struct ext4_extent_header *eh;
unsigned int depth;
int merge_done = 0;
depth = ext_depth(inode);
BUG_ON(path[depth].p_hdr == NULL);
eh = path[depth].p_hdr;
if (ex > EXT_FIRST_EXTENT(eh))
merge_done = ext4_ext_try_to_merge_right(inode, path, ex - 1);
if (!merge_done)
(void) ext4_ext_try_to_merge_right(inode, path, ex);
ext4_ext_try_to_merge_up(handle, inode, path);
}
/*
* check if a portion of the "newext" extent overlaps with an
* existing extent.
*
* If there is an overlap discovered, it updates the length of the newext
* such that there will be no overlap, and then returns 1.
* If there is no overlap found, it returns 0.
*/
static unsigned int ext4_ext_check_overlap(struct ext4_sb_info *sbi,
struct inode *inode,
struct ext4_extent *newext,
struct ext4_ext_path *path)
{
ext4_lblk_t b1, b2;
unsigned int depth, len1;
unsigned int ret = 0;
b1 = le32_to_cpu(newext->ee_block);
len1 = ext4_ext_get_actual_len(newext);
depth = ext_depth(inode);
if (!path[depth].p_ext)
goto out;
b2 = le32_to_cpu(path[depth].p_ext->ee_block);
b2 &= ~(sbi->s_cluster_ratio - 1);
/*
* get the next allocated block if the extent in the path
* is before the requested block(s)
*/
if (b2 < b1) {
b2 = ext4_ext_next_allocated_block(path);
ext4: Fix max file size and logical block counting of extent format file Kazuya Mio reported that he was able to hit BUG_ON(next == lblock) in ext4_ext_put_gap_in_cache() while creating a sparse file in extent format and fill the tail of file up to its end. We will hit the BUG_ON when we write the last block (2^32-1) into the sparse file. The root cause of the problem lies in the fact that we specifically set s_maxbytes so that block at s_maxbytes fit into on-disk extent format, which is 32 bit long. However, we are not storing start and end block number, but rather start block number and length in blocks. It means that in order to cover extent from 0 to EXT_MAX_BLOCK we need EXT_MAX_BLOCK+1 to fit into len (because we counting block 0 as well) - and it does not. The only way to fix it without changing the meaning of the struct ext4_extent members is, as Kazuya Mio suggested, to lower s_maxbytes by one fs block so we can cover the whole extent we can get by the on-disk extent format. Also in many places EXT_MAX_BLOCK is used as length instead of maximum logical block number as the name suggests, it is all a bit messy. So this commit renames it to EXT_MAX_BLOCKS and change its usage in some places to actually be maximum number of blocks in the extent. The bug which this commit fixes can be reproduced as follows: dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-2)) sync dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-1)) Reported-by: Kazuya Mio <k-mio@sx.jp.nec.com> Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-06-06 12:05:17 +08:00
if (b2 == EXT_MAX_BLOCKS)
goto out;
b2 &= ~(sbi->s_cluster_ratio - 1);
}
/* check for wrap through zero on extent logical start block*/
if (b1 + len1 < b1) {
ext4: Fix max file size and logical block counting of extent format file Kazuya Mio reported that he was able to hit BUG_ON(next == lblock) in ext4_ext_put_gap_in_cache() while creating a sparse file in extent format and fill the tail of file up to its end. We will hit the BUG_ON when we write the last block (2^32-1) into the sparse file. The root cause of the problem lies in the fact that we specifically set s_maxbytes so that block at s_maxbytes fit into on-disk extent format, which is 32 bit long. However, we are not storing start and end block number, but rather start block number and length in blocks. It means that in order to cover extent from 0 to EXT_MAX_BLOCK we need EXT_MAX_BLOCK+1 to fit into len (because we counting block 0 as well) - and it does not. The only way to fix it without changing the meaning of the struct ext4_extent members is, as Kazuya Mio suggested, to lower s_maxbytes by one fs block so we can cover the whole extent we can get by the on-disk extent format. Also in many places EXT_MAX_BLOCK is used as length instead of maximum logical block number as the name suggests, it is all a bit messy. So this commit renames it to EXT_MAX_BLOCKS and change its usage in some places to actually be maximum number of blocks in the extent. The bug which this commit fixes can be reproduced as follows: dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-2)) sync dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-1)) Reported-by: Kazuya Mio <k-mio@sx.jp.nec.com> Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-06-06 12:05:17 +08:00
len1 = EXT_MAX_BLOCKS - b1;
newext->ee_len = cpu_to_le16(len1);
ret = 1;
}
/* check for overlap */
if (b1 + len1 > b2) {
newext->ee_len = cpu_to_le16(b2 - b1);
ret = 1;
}
out:
return ret;
}
/*
* ext4_ext_insert_extent:
* tries to merge requsted extent into the existing extent or
* inserts requested extent as new one into the tree,
* creating new leaf in the no-space case.
*/
int ext4_ext_insert_extent(handle_t *handle, struct inode *inode,
struct ext4_ext_path *path,
struct ext4_extent *newext, int flag)
{
struct ext4_extent_header *eh;
struct ext4_extent *ex, *fex;
struct ext4_extent *nearex; /* nearest extent */
struct ext4_ext_path *npath = NULL;
int depth, len, err;
ext4_lblk_t next;
unsigned uninitialized = 0;
int flags = 0;
if (unlikely(ext4_ext_get_actual_len(newext) == 0)) {
EXT4_ERROR_INODE(inode, "ext4_ext_get_actual_len(newext) == 0");
return -EIO;
}
depth = ext_depth(inode);
ex = path[depth].p_ext;
eh = path[depth].p_hdr;
if (unlikely(path[depth].p_hdr == NULL)) {
EXT4_ERROR_INODE(inode, "path[%d].p_hdr == NULL", depth);
return -EIO;
}
/* try to insert block into found extent and return */
if (ex && !(flag & EXT4_GET_BLOCKS_PRE_IO)) {
/*
* Try to see whether we should rather test the extent on
* right from ex, or from the left of ex. This is because
* ext4_ext_find_extent() can return either extent on the
* left, or on the right from the searched position. This
* will make merging more effective.
*/
if (ex < EXT_LAST_EXTENT(eh) &&
(le32_to_cpu(ex->ee_block) +
ext4_ext_get_actual_len(ex) <
le32_to_cpu(newext->ee_block))) {
ex += 1;
goto prepend;
} else if ((ex > EXT_FIRST_EXTENT(eh)) &&
(le32_to_cpu(newext->ee_block) +
ext4_ext_get_actual_len(newext) <
le32_to_cpu(ex->ee_block)))
ex -= 1;
/* Try to append newex to the ex */
if (ext4_can_extents_be_merged(inode, ex, newext)) {
ext_debug("append [%d]%d block to %u:[%d]%d"
"(from %llu)\n",
ext4_ext_is_uninitialized(newext),
ext4_ext_get_actual_len(newext),
le32_to_cpu(ex->ee_block),
ext4_ext_is_uninitialized(ex),
ext4_ext_get_actual_len(ex),
ext4_ext_pblock(ex));
err = ext4_ext_get_access(handle, inode,
path + depth);
if (err)
return err;
/*
* ext4_can_extents_be_merged should have checked
* that either both extents are uninitialized, or
* both aren't. Thus we need to check only one of
* them here.
*/
if (ext4_ext_is_uninitialized(ex))
uninitialized = 1;
ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex)
+ ext4_ext_get_actual_len(newext));
if (uninitialized)
ext4_ext_mark_uninitialized(ex);
eh = path[depth].p_hdr;
nearex = ex;
goto merge;
}
prepend:
/* Try to prepend newex to the ex */
if (ext4_can_extents_be_merged(inode, newext, ex)) {
ext_debug("prepend %u[%d]%d block to %u:[%d]%d"
"(from %llu)\n",
le32_to_cpu(newext->ee_block),
ext4_ext_is_uninitialized(newext),
ext4_ext_get_actual_len(newext),
le32_to_cpu(ex->ee_block),
ext4_ext_is_uninitialized(ex),
ext4_ext_get_actual_len(ex),
ext4_ext_pblock(ex));
err = ext4_ext_get_access(handle, inode,
path + depth);
if (err)
return err;
/*
* ext4_can_extents_be_merged should have checked
* that either both extents are uninitialized, or
* both aren't. Thus we need to check only one of
* them here.
*/
if (ext4_ext_is_uninitialized(ex))
uninitialized = 1;
ex->ee_block = newext->ee_block;
ext4_ext_store_pblock(ex, ext4_ext_pblock(newext));
ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex)
+ ext4_ext_get_actual_len(newext));
if (uninitialized)
ext4_ext_mark_uninitialized(ex);
eh = path[depth].p_hdr;
nearex = ex;
goto merge;
}
}
depth = ext_depth(inode);
eh = path[depth].p_hdr;
if (le16_to_cpu(eh->eh_entries) < le16_to_cpu(eh->eh_max))
goto has_space;
/* probably next leaf has space for us? */
fex = EXT_LAST_EXTENT(eh);
next = EXT_MAX_BLOCKS;
if (le32_to_cpu(newext->ee_block) > le32_to_cpu(fex->ee_block))
next = ext4_ext_next_leaf_block(path);
if (next != EXT_MAX_BLOCKS) {
ext_debug("next leaf block - %u\n", next);
BUG_ON(npath != NULL);
npath = ext4_ext_find_extent(inode, next, NULL);
if (IS_ERR(npath))
return PTR_ERR(npath);
BUG_ON(npath->p_depth != path->p_depth);
eh = npath[depth].p_hdr;
if (le16_to_cpu(eh->eh_entries) < le16_to_cpu(eh->eh_max)) {
ext_debug("next leaf isn't full(%d)\n",
le16_to_cpu(eh->eh_entries));
path = npath;
goto has_space;
}
ext_debug("next leaf has no free space(%d,%d)\n",
le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max));
}
/*
* There is no free space in the found leaf.
* We're gonna add a new leaf in the tree.
*/
ext4: introduce reserved space Currently in ENOSPC condition when writing into unwritten space, or punching a hole, we might need to split the extent and grow extent tree. However since we can not allocate any new metadata blocks we'll have to zero out unwritten part of extent or punched out part of extent, or in the worst case return ENOSPC even though use actually does not allocate any space. Also in delalloc path we do reserve metadata and data blocks for the time we're going to write out, however metadata block reservation is very tricky especially since we expect that logical connectivity implies physical connectivity, however that might not be the case and hence we might end up allocating more metadata blocks than previously reserved. So in future, metadata reservation checks should be removed since we can not assure that we do not under reserve. And this is where reserved space comes into the picture. When mounting the file system we slice off a little bit of the file system space (2% or 4096 clusters, whichever is smaller) which can be then used for the cases mentioned above to prevent costly zeroout, or unexpected ENOSPC. The number of reserved clusters can be set via sysfs, however it can never be bigger than number of free clusters in the file system. Note that this patch fixes the failure of xfstest 274 as expected. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
2013-04-10 10:11:22 +08:00
if (flag & EXT4_GET_BLOCKS_METADATA_NOFAIL)
flags = EXT4_MB_USE_RESERVED;
err = ext4_ext_create_new_leaf(handle, inode, flags, path, newext);
if (err)
goto cleanup;
depth = ext_depth(inode);
eh = path[depth].p_hdr;
has_space:
nearex = path[depth].p_ext;
err = ext4_ext_get_access(handle, inode, path + depth);
if (err)
goto cleanup;
if (!nearex) {
/* there is no extent in this leaf, create first one */
ext_debug("first extent in the leaf: %u:%llu:[%d]%d\n",
le32_to_cpu(newext->ee_block),
ext4_ext_pblock(newext),
ext4_ext_is_uninitialized(newext),
ext4_ext_get_actual_len(newext));
nearex = EXT_FIRST_EXTENT(eh);
} else {
if (le32_to_cpu(newext->ee_block)
> le32_to_cpu(nearex->ee_block)) {
/* Insert after */
ext_debug("insert %u:%llu:[%d]%d before: "
"nearest %p\n",
le32_to_cpu(newext->ee_block),
ext4_ext_pblock(newext),
ext4_ext_is_uninitialized(newext),
ext4_ext_get_actual_len(newext),
nearex);
nearex++;
} else {
/* Insert before */
BUG_ON(newext->ee_block == nearex->ee_block);
ext_debug("insert %u:%llu:[%d]%d after: "
"nearest %p\n",
le32_to_cpu(newext->ee_block),
ext4_ext_pblock(newext),
ext4_ext_is_uninitialized(newext),
ext4_ext_get_actual_len(newext),
nearex);
}
len = EXT_LAST_EXTENT(eh) - nearex + 1;
if (len > 0) {
ext_debug("insert %u:%llu:[%d]%d: "
"move %d extents from 0x%p to 0x%p\n",
le32_to_cpu(newext->ee_block),
ext4_ext_pblock(newext),
ext4_ext_is_uninitialized(newext),
ext4_ext_get_actual_len(newext),
len, nearex, nearex + 1);
memmove(nearex + 1, nearex,
len * sizeof(struct ext4_extent));
}
}
le16_add_cpu(&eh->eh_entries, 1);
path[depth].p_ext = nearex;
nearex->ee_block = newext->ee_block;
ext4_ext_store_pblock(nearex, ext4_ext_pblock(newext));
nearex->ee_len = newext->ee_len;
merge:
/* try to merge extents */
if (!(flag & EXT4_GET_BLOCKS_PRE_IO))
ext4_ext_try_to_merge(handle, inode, path, nearex);
/* time to correct all indexes above */
err = ext4_ext_correct_indexes(handle, inode, path);
if (err)
goto cleanup;
err = ext4_ext_dirty(handle, inode, path + path->p_depth);
cleanup:
if (npath) {
ext4_ext_drop_refs(npath);
kfree(npath);
}
return err;
}
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
static int ext4_fill_fiemap_extents(struct inode *inode,
ext4_lblk_t block, ext4_lblk_t num,
struct fiemap_extent_info *fieinfo)
{
struct ext4_ext_path *path = NULL;
struct ext4_extent *ex;
struct extent_status es;
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
ext4_lblk_t next, next_del, start = 0, end = 0;
ext4_lblk_t last = block + num;
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
int exists, depth = 0, err = 0;
unsigned int flags = 0;
unsigned char blksize_bits = inode->i_sb->s_blocksize_bits;
ext4: Fix max file size and logical block counting of extent format file Kazuya Mio reported that he was able to hit BUG_ON(next == lblock) in ext4_ext_put_gap_in_cache() while creating a sparse file in extent format and fill the tail of file up to its end. We will hit the BUG_ON when we write the last block (2^32-1) into the sparse file. The root cause of the problem lies in the fact that we specifically set s_maxbytes so that block at s_maxbytes fit into on-disk extent format, which is 32 bit long. However, we are not storing start and end block number, but rather start block number and length in blocks. It means that in order to cover extent from 0 to EXT_MAX_BLOCK we need EXT_MAX_BLOCK+1 to fit into len (because we counting block 0 as well) - and it does not. The only way to fix it without changing the meaning of the struct ext4_extent members is, as Kazuya Mio suggested, to lower s_maxbytes by one fs block so we can cover the whole extent we can get by the on-disk extent format. Also in many places EXT_MAX_BLOCK is used as length instead of maximum logical block number as the name suggests, it is all a bit messy. So this commit renames it to EXT_MAX_BLOCKS and change its usage in some places to actually be maximum number of blocks in the extent. The bug which this commit fixes can be reproduced as follows: dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-2)) sync dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-1)) Reported-by: Kazuya Mio <k-mio@sx.jp.nec.com> Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-06-06 12:05:17 +08:00
while (block < last && block != EXT_MAX_BLOCKS) {
num = last - block;
/* find extent for this block */
ext4: Fix potential fiemap deadlock (mmap_sem vs. i_data_sem) Fix the following potential circular locking dependency between mm->mmap_sem and ei->i_data_sem: ======================================================= [ INFO: possible circular locking dependency detected ] 2.6.32-04115-gec044c5 #37 ------------------------------------------------------- ureadahead/1855 is trying to acquire lock: (&mm->mmap_sem){++++++}, at: [<ffffffff81107224>] might_fault+0x5c/0xac but task is already holding lock: (&ei->i_data_sem){++++..}, at: [<ffffffff811be1fd>] ext4_fiemap+0x11b/0x159 which lock already depends on the new lock. the existing dependency chain (in reverse order) is: -> #1 (&ei->i_data_sem){++++..}: [<ffffffff81099bfa>] __lock_acquire+0xb67/0xd0f [<ffffffff81099e7e>] lock_acquire+0xdc/0x102 [<ffffffff81516633>] down_read+0x51/0x84 [<ffffffff811a2414>] ext4_get_blocks+0x50/0x2a5 [<ffffffff811a3453>] ext4_get_block+0xab/0xef [<ffffffff81154f39>] do_mpage_readpage+0x198/0x48d [<ffffffff81155360>] mpage_readpages+0xd0/0x114 [<ffffffff811a104b>] ext4_readpages+0x1d/0x1f [<ffffffff810f8644>] __do_page_cache_readahead+0x12f/0x1bc [<ffffffff810f86f2>] ra_submit+0x21/0x25 [<ffffffff810f0cfd>] filemap_fault+0x19f/0x32c [<ffffffff81107b97>] __do_fault+0x55/0x3a2 [<ffffffff81109db0>] handle_mm_fault+0x327/0x734 [<ffffffff8151aaa9>] do_page_fault+0x292/0x2aa [<ffffffff81518205>] page_fault+0x25/0x30 [<ffffffff812a34d8>] clear_user+0x38/0x3c [<ffffffff81167e16>] padzero+0x20/0x31 [<ffffffff81168b47>] load_elf_binary+0x8bc/0x17ed [<ffffffff81130e95>] search_binary_handler+0xc2/0x259 [<ffffffff81166d64>] load_script+0x1b8/0x1cc [<ffffffff81130e95>] search_binary_handler+0xc2/0x259 [<ffffffff8113255f>] do_execve+0x1ce/0x2cf [<ffffffff81027494>] sys_execve+0x43/0x5a [<ffffffff8102918a>] stub_execve+0x6a/0xc0 -> #0 (&mm->mmap_sem){++++++}: [<ffffffff81099aa4>] __lock_acquire+0xa11/0xd0f [<ffffffff81099e7e>] lock_acquire+0xdc/0x102 [<ffffffff81107251>] might_fault+0x89/0xac [<ffffffff81139382>] fiemap_fill_next_extent+0x95/0xda [<ffffffff811bcb43>] ext4_ext_fiemap_cb+0x138/0x157 [<ffffffff811be069>] ext4_ext_walk_space+0x178/0x1f1 [<ffffffff811be21e>] ext4_fiemap+0x13c/0x159 [<ffffffff811390e6>] do_vfs_ioctl+0x348/0x4d6 [<ffffffff811392ca>] sys_ioctl+0x56/0x79 [<ffffffff81028cb2>] system_call_fastpath+0x16/0x1b other info that might help us debug this: 1 lock held by ureadahead/1855: #0: (&ei->i_data_sem){++++..}, at: [<ffffffff811be1fd>] ext4_fiemap+0x11b/0x159 stack backtrace: Pid: 1855, comm: ureadahead Not tainted 2.6.32-04115-gec044c5 #37 Call Trace: [<ffffffff81098c70>] print_circular_bug+0xa8/0xb7 [<ffffffff81099aa4>] __lock_acquire+0xa11/0xd0f [<ffffffff8102f229>] ? sched_clock+0x9/0xd [<ffffffff81099e7e>] lock_acquire+0xdc/0x102 [<ffffffff81107224>] ? might_fault+0x5c/0xac [<ffffffff81107251>] might_fault+0x89/0xac [<ffffffff81107224>] ? might_fault+0x5c/0xac [<ffffffff81124b44>] ? __kmalloc+0x13b/0x18c [<ffffffff81139382>] fiemap_fill_next_extent+0x95/0xda [<ffffffff811bcb43>] ext4_ext_fiemap_cb+0x138/0x157 [<ffffffff811bca0b>] ? ext4_ext_fiemap_cb+0x0/0x157 [<ffffffff811be069>] ext4_ext_walk_space+0x178/0x1f1 [<ffffffff811be21e>] ext4_fiemap+0x13c/0x159 [<ffffffff81107224>] ? might_fault+0x5c/0xac [<ffffffff811390e6>] do_vfs_ioctl+0x348/0x4d6 [<ffffffff8129f6d0>] ? __up_read+0x8d/0x95 [<ffffffff81517fb5>] ? retint_swapgs+0x13/0x1b [<ffffffff811392ca>] sys_ioctl+0x56/0x79 [<ffffffff81028cb2>] system_call_fastpath+0x16/0x1b Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2009-12-10 10:30:02 +08:00
down_read(&EXT4_I(inode)->i_data_sem);
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
if (path && ext_depth(inode) != depth) {
/* depth was changed. we have to realloc path */
kfree(path);
path = NULL;
}
path = ext4_ext_find_extent(inode, block, path);
if (IS_ERR(path)) {
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
up_read(&EXT4_I(inode)->i_data_sem);
err = PTR_ERR(path);
path = NULL;
break;
}
depth = ext_depth(inode);
if (unlikely(path[depth].p_hdr == NULL)) {
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
up_read(&EXT4_I(inode)->i_data_sem);
EXT4_ERROR_INODE(inode, "path[%d].p_hdr == NULL", depth);
err = -EIO;
break;
}
ex = path[depth].p_ext;
next = ext4_ext_next_allocated_block(path);
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
ext4_ext_drop_refs(path);
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
flags = 0;
exists = 0;
if (!ex) {
/* there is no extent yet, so try to allocate
* all requested space */
start = block;
end = block + num;
} else if (le32_to_cpu(ex->ee_block) > block) {
/* need to allocate space before found extent */
start = block;
end = le32_to_cpu(ex->ee_block);
if (block + num < end)
end = block + num;
} else if (block >= le32_to_cpu(ex->ee_block)
+ ext4_ext_get_actual_len(ex)) {
/* need to allocate space after found extent */
start = block;
end = block + num;
if (end >= next)
end = next;
} else if (block >= le32_to_cpu(ex->ee_block)) {
/*
* some part of requested space is covered
* by found extent
*/
start = block;
end = le32_to_cpu(ex->ee_block)
+ ext4_ext_get_actual_len(ex);
if (block + num < end)
end = block + num;
exists = 1;
} else {
BUG();
}
BUG_ON(end <= start);
if (!exists) {
es.es_lblk = start;
es.es_len = end - start;
es.es_pblk = 0;
} else {
es.es_lblk = le32_to_cpu(ex->ee_block);
es.es_len = ext4_ext_get_actual_len(ex);
es.es_pblk = ext4_ext_pblock(ex);
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
if (ext4_ext_is_uninitialized(ex))
flags |= FIEMAP_EXTENT_UNWRITTEN;
}
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
/*
* Find delayed extent and update es accordingly. We call
* it even in !exists case to find out whether es is the
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
* last existing extent or not.
*/
next_del = ext4_find_delayed_extent(inode, &es);
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
if (!exists && next_del) {
exists = 1;
flags |= FIEMAP_EXTENT_DELALLOC;
}
up_read(&EXT4_I(inode)->i_data_sem);
if (unlikely(es.es_len == 0)) {
EXT4_ERROR_INODE(inode, "es.es_len == 0");
err = -EIO;
break;
}
/*
* This is possible iff next == next_del == EXT_MAX_BLOCKS.
* we need to check next == EXT_MAX_BLOCKS because it is
* possible that an extent is with unwritten and delayed
* status due to when an extent is delayed allocated and
* is allocated by fallocate status tree will track both of
* them in a extent.
*
* So we could return a unwritten and delayed extent, and
* its block is equal to 'next'.
*/
if (next == next_del && next == EXT_MAX_BLOCKS) {
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
flags |= FIEMAP_EXTENT_LAST;
if (unlikely(next_del != EXT_MAX_BLOCKS ||
next != EXT_MAX_BLOCKS)) {
EXT4_ERROR_INODE(inode,
"next extent == %u, next "
"delalloc extent = %u",
next, next_del);
err = -EIO;
break;
}
}
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
if (exists) {
err = fiemap_fill_next_extent(fieinfo,
(__u64)es.es_lblk << blksize_bits,
(__u64)es.es_pblk << blksize_bits,
(__u64)es.es_len << blksize_bits,
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
flags);
if (err < 0)
break;
if (err == 1) {
err = 0;
break;
}
}
block = es.es_lblk + es.es_len;
}
if (path) {
ext4_ext_drop_refs(path);
kfree(path);
}
return err;
}
/*
* ext4_ext_put_gap_in_cache:
* calculate boundaries of the gap that the requested block fits into
* and cache this gap
*/
static void
ext4_ext_put_gap_in_cache(struct inode *inode, struct ext4_ext_path *path,
ext4_lblk_t block)
{
int depth = ext_depth(inode);
unsigned long len;
ext4_lblk_t lblock;
struct ext4_extent *ex;
ex = path[depth].p_ext;
if (ex == NULL) {
/*
* there is no extent yet, so gap is [0;-] and we
* don't cache it
*/
ext_debug("cache gap(whole file):");
} else if (block < le32_to_cpu(ex->ee_block)) {
lblock = block;
len = le32_to_cpu(ex->ee_block) - block;
ext_debug("cache gap(before): %u [%u:%u]",
block,
le32_to_cpu(ex->ee_block),
ext4_ext_get_actual_len(ex));
if (!ext4_find_delalloc_range(inode, lblock, lblock + len - 1))
ext4_es_insert_extent(inode, lblock, len, ~0,
EXTENT_STATUS_HOLE);
} else if (block >= le32_to_cpu(ex->ee_block)
+ ext4_ext_get_actual_len(ex)) {
ext4_lblk_t next;
lblock = le32_to_cpu(ex->ee_block)
+ ext4_ext_get_actual_len(ex);
next = ext4_ext_next_allocated_block(path);
ext_debug("cache gap(after): [%u:%u] %u",
le32_to_cpu(ex->ee_block),
ext4_ext_get_actual_len(ex),
block);
BUG_ON(next == lblock);
len = next - lblock;
if (!ext4_find_delalloc_range(inode, lblock, lblock + len - 1))
ext4_es_insert_extent(inode, lblock, len, ~0,
EXTENT_STATUS_HOLE);
} else {
lblock = len = 0;
BUG();
}
ext_debug(" -> %u:%lu\n", lblock, len);
}
/*
* ext4_ext_rm_idx:
* removes index from the index block.
*/
static int ext4_ext_rm_idx(handle_t *handle, struct inode *inode,
struct ext4_ext_path *path, int depth)
{
int err;
ext4_fsblk_t leaf;
/* free index block */
depth--;
path = path + depth;
leaf = ext4_idx_pblock(path->p_idx);
if (unlikely(path->p_hdr->eh_entries == 0)) {
EXT4_ERROR_INODE(inode, "path->p_hdr->eh_entries == 0");
return -EIO;
}
err = ext4_ext_get_access(handle, inode, path);
if (err)
return err;
if (path->p_idx != EXT_LAST_INDEX(path->p_hdr)) {
int len = EXT_LAST_INDEX(path->p_hdr) - path->p_idx;
len *= sizeof(struct ext4_extent_idx);
memmove(path->p_idx, path->p_idx + 1, len);
}
le16_add_cpu(&path->p_hdr->eh_entries, -1);
err = ext4_ext_dirty(handle, inode, path);
if (err)
return err;
ext_debug("index is empty, remove it, free block %llu\n", leaf);
trace_ext4_ext_rm_idx(inode, leaf);
ext4_free_blocks(handle, inode, NULL, leaf, 1,
EXT4_FREE_BLOCKS_METADATA | EXT4_FREE_BLOCKS_FORGET);
while (--depth >= 0) {
if (path->p_idx != EXT_FIRST_INDEX(path->p_hdr))
break;
path--;
err = ext4_ext_get_access(handle, inode, path);
if (err)
break;
path->p_idx->ei_block = (path+1)->p_idx->ei_block;
err = ext4_ext_dirty(handle, inode, path);
if (err)
break;
}
return err;
}
/*
* ext4_ext_calc_credits_for_single_extent:
* This routine returns max. credits that needed to insert an extent
* to the extent tree.
* When pass the actual path, the caller should calculate credits
* under i_data_sem.
*/
int ext4_ext_calc_credits_for_single_extent(struct inode *inode, int nrblocks,
struct ext4_ext_path *path)
{
if (path) {
int depth = ext_depth(inode);
int ret = 0;
/* probably there is space in leaf? */
if (le16_to_cpu(path[depth].p_hdr->eh_entries)
< le16_to_cpu(path[depth].p_hdr->eh_max)) {
/*
* There are some space in the leaf tree, no
* need to account for leaf block credit
*
* bitmaps and block group descriptor blocks
* and other metadata blocks still need to be
* accounted.
*/
/* 1 bitmap, 1 block group descriptor */
ret = 2 + EXT4_META_TRANS_BLOCKS(inode->i_sb);
return ret;
}
}
return ext4_chunk_trans_blocks(inode, nrblocks);
}
/*
* How many index/leaf blocks need to change/allocate to modify nrblocks?
*
* if nrblocks are fit in a single extent (chunk flag is 1), then
* in the worse case, each tree level index/leaf need to be changed
* if the tree split due to insert a new extent, then the old tree
* index/leaf need to be updated too
*
* If the nrblocks are discontiguous, they could cause
* the whole tree split more than once, but this is really rare.
*/
int ext4_ext_index_trans_blocks(struct inode *inode, int nrblocks, int chunk)
{
int index;
int depth;
/* If we are converting the inline data, only one is needed here. */
if (ext4_has_inline_data(inode))
return 1;
depth = ext_depth(inode);
if (chunk)
index = depth * 2;
else
index = depth * 3;
return index;
}
static int ext4_remove_blocks(handle_t *handle, struct inode *inode,
struct ext4_extent *ex,
ext4_fsblk_t *partial_cluster,
ext4_lblk_t from, ext4_lblk_t to)
{
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
unsigned short ee_len = ext4_ext_get_actual_len(ex);
ext4_fsblk_t pblk;
int flags = 0;
if (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode))
flags |= EXT4_FREE_BLOCKS_METADATA | EXT4_FREE_BLOCKS_FORGET;
else if (ext4_should_journal_data(inode))
flags |= EXT4_FREE_BLOCKS_FORGET;
/*
* For bigalloc file systems, we never free a partial cluster
* at the beginning of the extent. Instead, we make a note
* that we tried freeing the cluster, and check to see if we
* need to free it on a subsequent call to ext4_remove_blocks,
* or at the end of the ext4_truncate() operation.
*/
flags |= EXT4_FREE_BLOCKS_NOFREE_FIRST_CLUSTER;
trace_ext4_remove_blocks(inode, ex, from, to, *partial_cluster);
/*
* If we have a partial cluster, and it's different from the
* cluster of the last block, we need to explicitly free the
* partial cluster here.
*/
pblk = ext4_ext_pblock(ex) + ee_len - 1;
if (*partial_cluster && (EXT4_B2C(sbi, pblk) != *partial_cluster)) {
ext4_free_blocks(handle, inode, NULL,
EXT4_C2B(sbi, *partial_cluster),
sbi->s_cluster_ratio, flags);
*partial_cluster = 0;
}
#ifdef EXTENTS_STATS
{
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
spin_lock(&sbi->s_ext_stats_lock);
sbi->s_ext_blocks += ee_len;
sbi->s_ext_extents++;
if (ee_len < sbi->s_ext_min)
sbi->s_ext_min = ee_len;
if (ee_len > sbi->s_ext_max)
sbi->s_ext_max = ee_len;
if (ext_depth(inode) > sbi->s_depth_max)
sbi->s_depth_max = ext_depth(inode);
spin_unlock(&sbi->s_ext_stats_lock);
}
#endif
if (from >= le32_to_cpu(ex->ee_block)
&& to == le32_to_cpu(ex->ee_block) + ee_len - 1) {
/* tail removal */
ext4_lblk_t num;
num = le32_to_cpu(ex->ee_block) + ee_len - from;
pblk = ext4_ext_pblock(ex) + ee_len - num;
ext_debug("free last %u blocks starting %llu\n", num, pblk);
ext4_free_blocks(handle, inode, NULL, pblk, num, flags);
/*
* If the block range to be freed didn't start at the
* beginning of a cluster, and we removed the entire
* extent, save the partial cluster here, since we
* might need to delete if we determine that the
* truncate operation has removed all of the blocks in
* the cluster.
*/
if (pblk & (sbi->s_cluster_ratio - 1) &&
(ee_len == num))
*partial_cluster = EXT4_B2C(sbi, pblk);
else
*partial_cluster = 0;
} else if (from == le32_to_cpu(ex->ee_block)
&& to <= le32_to_cpu(ex->ee_block) + ee_len - 1) {
/* head removal */
ext4_lblk_t num;
ext4_fsblk_t start;
num = to - from;
start = ext4_ext_pblock(ex);
ext_debug("free first %u blocks starting %llu\n", num, start);
ext4_free_blocks(handle, inode, NULL, start, num, flags);
} else {
printk(KERN_INFO "strange request: removal(2) "
"%u-%u from %u:%u\n",
from, to, le32_to_cpu(ex->ee_block), ee_len);
}
return 0;
}
/*
* ext4_ext_rm_leaf() Removes the extents associated with the
* blocks appearing between "start" and "end", and splits the extents
* if "start" and "end" appear in the same extent
*
* @handle: The journal handle
* @inode: The files inode
* @path: The path to the leaf
* @start: The first block to remove
* @end: The last block to remove
*/
static int
ext4_ext_rm_leaf(handle_t *handle, struct inode *inode,
struct ext4_ext_path *path, ext4_fsblk_t *partial_cluster,
ext4_lblk_t start, ext4_lblk_t end)
{
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
int err = 0, correct_index = 0;
int depth = ext_depth(inode), credits;
struct ext4_extent_header *eh;
ext4_lblk_t a, b;
unsigned num;
ext4_lblk_t ex_ee_block;
unsigned short ex_ee_len;
unsigned uninitialized = 0;
struct ext4_extent *ex;
/* the header must be checked already in ext4_ext_remove_space() */
ext4: rewrite punch hole to use ext4_ext_remove_space() This commit rewrites ext4 punch hole implementation to use ext4_ext_remove_space() instead of its home gown way of doing this via ext4_ext_map_blocks(). There are several reasons for changing this. Firstly it is quite non obvious that punching hole needs to ext4_ext_map_blocks() to punch a hole, especially given that this function should map blocks, not unmap it. It also required a lot of new code in ext4_ext_map_blocks(). Secondly the design of it is not very effective. The reason is that we are trying to punch out blocks in ext4_ext_punch_hole() in opposite direction than in ext4_ext_rm_leaf() which causes the ext4_ext_rm_leaf() to iterate through the whole tree from the end to the start to find the requested extent for every extent we are going to punch out. And finally the current implementation does not use the existing code, but bring a lot of new code, which is IMO unnecessary since there already is some infrastructure we can use. Specifically ext4_ext_remove_space(). This commit changes ext4_ext_remove_space() to accept 'end' parameter so we can not only truncate to the end of file, but also remove the space in the middle of the file (punch a hole). Moreover, because the last block to punch out, might be in the middle of the extent, we have to split the extent at 'end + 1' so ext4_ext_rm_leaf() can easily either remove the whole fist part of split extent, or change its size. ext4_ext_remove_space() is then used to actually remove the space (extents) from within the hole, instead of ext4_ext_map_blocks(). Note that this also fix the issue with punch hole, where we would forget to remove empty index blocks from the extent tree, resulting in double free block error and file system corruption. This is simply because we now use different code path, where this problem does not exist. This has been tested with fsx running for several days and xfstests, plus xfstest #251 with '-o discard' run on the loop image (which converts discard requestes into punch hole to the backing file). All of it on 1K and 4K file system block size. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-03-20 11:03:19 +08:00
ext_debug("truncate since %u in leaf to %u\n", start, end);
if (!path[depth].p_hdr)
path[depth].p_hdr = ext_block_hdr(path[depth].p_bh);
eh = path[depth].p_hdr;
if (unlikely(path[depth].p_hdr == NULL)) {
EXT4_ERROR_INODE(inode, "path[%d].p_hdr == NULL", depth);
return -EIO;
}
/* find where to start removing */
ex = EXT_LAST_EXTENT(eh);
ex_ee_block = le32_to_cpu(ex->ee_block);
ex_ee_len = ext4_ext_get_actual_len(ex);
trace_ext4_ext_rm_leaf(inode, start, ex, *partial_cluster);
while (ex >= EXT_FIRST_EXTENT(eh) &&
ex_ee_block + ex_ee_len > start) {
if (ext4_ext_is_uninitialized(ex))
uninitialized = 1;
else
uninitialized = 0;
ext_debug("remove ext %u:[%d]%d\n", ex_ee_block,
uninitialized, ex_ee_len);
path[depth].p_ext = ex;
a = ex_ee_block > start ? ex_ee_block : start;
b = ex_ee_block+ex_ee_len - 1 < end ?
ex_ee_block+ex_ee_len - 1 : end;
ext_debug(" border %u:%u\n", a, b);
/* If this extent is beyond the end of the hole, skip it */
ext4: rewrite punch hole to use ext4_ext_remove_space() This commit rewrites ext4 punch hole implementation to use ext4_ext_remove_space() instead of its home gown way of doing this via ext4_ext_map_blocks(). There are several reasons for changing this. Firstly it is quite non obvious that punching hole needs to ext4_ext_map_blocks() to punch a hole, especially given that this function should map blocks, not unmap it. It also required a lot of new code in ext4_ext_map_blocks(). Secondly the design of it is not very effective. The reason is that we are trying to punch out blocks in ext4_ext_punch_hole() in opposite direction than in ext4_ext_rm_leaf() which causes the ext4_ext_rm_leaf() to iterate through the whole tree from the end to the start to find the requested extent for every extent we are going to punch out. And finally the current implementation does not use the existing code, but bring a lot of new code, which is IMO unnecessary since there already is some infrastructure we can use. Specifically ext4_ext_remove_space(). This commit changes ext4_ext_remove_space() to accept 'end' parameter so we can not only truncate to the end of file, but also remove the space in the middle of the file (punch a hole). Moreover, because the last block to punch out, might be in the middle of the extent, we have to split the extent at 'end + 1' so ext4_ext_rm_leaf() can easily either remove the whole fist part of split extent, or change its size. ext4_ext_remove_space() is then used to actually remove the space (extents) from within the hole, instead of ext4_ext_map_blocks(). Note that this also fix the issue with punch hole, where we would forget to remove empty index blocks from the extent tree, resulting in double free block error and file system corruption. This is simply because we now use different code path, where this problem does not exist. This has been tested with fsx running for several days and xfstests, plus xfstest #251 with '-o discard' run on the loop image (which converts discard requestes into punch hole to the backing file). All of it on 1K and 4K file system block size. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-03-20 11:03:19 +08:00
if (end < ex_ee_block) {
ex--;
ex_ee_block = le32_to_cpu(ex->ee_block);
ex_ee_len = ext4_ext_get_actual_len(ex);
continue;
} else if (b != ex_ee_block + ex_ee_len - 1) {
EXT4_ERROR_INODE(inode,
"can not handle truncate %u:%u "
"on extent %u:%u",
start, end, ex_ee_block,
ex_ee_block + ex_ee_len - 1);
err = -EIO;
goto out;
} else if (a != ex_ee_block) {
/* remove tail of the extent */
num = a - ex_ee_block;
} else {
/* remove whole extent: excellent! */
num = 0;
}
/*
* 3 for leaf, sb, and inode plus 2 (bmap and group
* descriptor) for each block group; assume two block
* groups plus ex_ee_len/blocks_per_block_group for
* the worst case
*/
credits = 7 + 2*(ex_ee_len/EXT4_BLOCKS_PER_GROUP(inode->i_sb));
if (ex == EXT_FIRST_EXTENT(eh)) {
correct_index = 1;
credits += (ext_depth(inode)) + 1;
}
credits += EXT4_MAXQUOTAS_TRANS_BLOCKS(inode->i_sb);
err = ext4_ext_truncate_extend_restart(handle, inode, credits);
if (err)
goto out;
err = ext4_ext_get_access(handle, inode, path + depth);
if (err)
goto out;
err = ext4_remove_blocks(handle, inode, ex, partial_cluster,
a, b);
if (err)
goto out;
if (num == 0)
/* this extent is removed; mark slot entirely unused */
ext4_ext_store_pblock(ex, 0);
ex->ee_len = cpu_to_le16(num);
/*
* Do not mark uninitialized if all the blocks in the
* extent have been removed.
*/
if (uninitialized && num)
ext4_ext_mark_uninitialized(ex);
/*
* If the extent was completely released,
* we need to remove it from the leaf
*/
if (num == 0) {
ext4: Fix max file size and logical block counting of extent format file Kazuya Mio reported that he was able to hit BUG_ON(next == lblock) in ext4_ext_put_gap_in_cache() while creating a sparse file in extent format and fill the tail of file up to its end. We will hit the BUG_ON when we write the last block (2^32-1) into the sparse file. The root cause of the problem lies in the fact that we specifically set s_maxbytes so that block at s_maxbytes fit into on-disk extent format, which is 32 bit long. However, we are not storing start and end block number, but rather start block number and length in blocks. It means that in order to cover extent from 0 to EXT_MAX_BLOCK we need EXT_MAX_BLOCK+1 to fit into len (because we counting block 0 as well) - and it does not. The only way to fix it without changing the meaning of the struct ext4_extent members is, as Kazuya Mio suggested, to lower s_maxbytes by one fs block so we can cover the whole extent we can get by the on-disk extent format. Also in many places EXT_MAX_BLOCK is used as length instead of maximum logical block number as the name suggests, it is all a bit messy. So this commit renames it to EXT_MAX_BLOCKS and change its usage in some places to actually be maximum number of blocks in the extent. The bug which this commit fixes can be reproduced as follows: dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-2)) sync dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-1)) Reported-by: Kazuya Mio <k-mio@sx.jp.nec.com> Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-06-06 12:05:17 +08:00
if (end != EXT_MAX_BLOCKS - 1) {
/*
* For hole punching, we need to scoot all the
* extents up when an extent is removed so that
* we dont have blank extents in the middle
*/
memmove(ex, ex+1, (EXT_LAST_EXTENT(eh) - ex) *
sizeof(struct ext4_extent));
/* Now get rid of the one at the end */
memset(EXT_LAST_EXTENT(eh), 0,
sizeof(struct ext4_extent));
}
le16_add_cpu(&eh->eh_entries, -1);
} else
*partial_cluster = 0;
err = ext4_ext_dirty(handle, inode, path + depth);
if (err)
goto out;
ext_debug("new extent: %u:%u:%llu\n", ex_ee_block, num,
ext4_ext_pblock(ex));
ex--;
ex_ee_block = le32_to_cpu(ex->ee_block);
ex_ee_len = ext4_ext_get_actual_len(ex);
}
if (correct_index && eh->eh_entries)
err = ext4_ext_correct_indexes(handle, inode, path);
/*
* If there is still a entry in the leaf node, check to see if
* it references the partial cluster. This is the only place
* where it could; if it doesn't, we can free the cluster.
*/
if (*partial_cluster && ex >= EXT_FIRST_EXTENT(eh) &&
(EXT4_B2C(sbi, ext4_ext_pblock(ex) + ex_ee_len - 1) !=
*partial_cluster)) {
int flags = EXT4_FREE_BLOCKS_FORGET;
if (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode))
flags |= EXT4_FREE_BLOCKS_METADATA;
ext4_free_blocks(handle, inode, NULL,
EXT4_C2B(sbi, *partial_cluster),
sbi->s_cluster_ratio, flags);
*partial_cluster = 0;
}
/* if this leaf is free, then we should
* remove it from index block above */
if (err == 0 && eh->eh_entries == 0 && path[depth].p_bh != NULL)
err = ext4_ext_rm_idx(handle, inode, path, depth);
out:
return err;
}
/*
* ext4_ext_more_to_rm:
* returns 1 if current index has to be freed (even partial)
*/
static int
ext4_ext_more_to_rm(struct ext4_ext_path *path)
{
BUG_ON(path->p_idx == NULL);
if (path->p_idx < EXT_FIRST_INDEX(path->p_hdr))
return 0;
/*
* if truncate on deeper level happened, it wasn't partial,
* so we have to consider current index for truncation
*/
if (le16_to_cpu(path->p_hdr->eh_entries) == path->p_block)
return 0;
return 1;
}
int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start,
ext4_lblk_t end)
{
struct super_block *sb = inode->i_sb;
int depth = ext_depth(inode);
struct ext4_ext_path *path = NULL;
ext4_fsblk_t partial_cluster = 0;
handle_t *handle;
int i = 0, err = 0;
ext4: rewrite punch hole to use ext4_ext_remove_space() This commit rewrites ext4 punch hole implementation to use ext4_ext_remove_space() instead of its home gown way of doing this via ext4_ext_map_blocks(). There are several reasons for changing this. Firstly it is quite non obvious that punching hole needs to ext4_ext_map_blocks() to punch a hole, especially given that this function should map blocks, not unmap it. It also required a lot of new code in ext4_ext_map_blocks(). Secondly the design of it is not very effective. The reason is that we are trying to punch out blocks in ext4_ext_punch_hole() in opposite direction than in ext4_ext_rm_leaf() which causes the ext4_ext_rm_leaf() to iterate through the whole tree from the end to the start to find the requested extent for every extent we are going to punch out. And finally the current implementation does not use the existing code, but bring a lot of new code, which is IMO unnecessary since there already is some infrastructure we can use. Specifically ext4_ext_remove_space(). This commit changes ext4_ext_remove_space() to accept 'end' parameter so we can not only truncate to the end of file, but also remove the space in the middle of the file (punch a hole). Moreover, because the last block to punch out, might be in the middle of the extent, we have to split the extent at 'end + 1' so ext4_ext_rm_leaf() can easily either remove the whole fist part of split extent, or change its size. ext4_ext_remove_space() is then used to actually remove the space (extents) from within the hole, instead of ext4_ext_map_blocks(). Note that this also fix the issue with punch hole, where we would forget to remove empty index blocks from the extent tree, resulting in double free block error and file system corruption. This is simply because we now use different code path, where this problem does not exist. This has been tested with fsx running for several days and xfstests, plus xfstest #251 with '-o discard' run on the loop image (which converts discard requestes into punch hole to the backing file). All of it on 1K and 4K file system block size. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-03-20 11:03:19 +08:00
ext_debug("truncate since %u to %u\n", start, end);
/* probably first extent we're gonna free will be last in block */
handle = ext4_journal_start(inode, EXT4_HT_TRUNCATE, depth + 1);
if (IS_ERR(handle))
return PTR_ERR(handle);
again:
trace_ext4_ext_remove_space(inode, start, depth);
ext4: rewrite punch hole to use ext4_ext_remove_space() This commit rewrites ext4 punch hole implementation to use ext4_ext_remove_space() instead of its home gown way of doing this via ext4_ext_map_blocks(). There are several reasons for changing this. Firstly it is quite non obvious that punching hole needs to ext4_ext_map_blocks() to punch a hole, especially given that this function should map blocks, not unmap it. It also required a lot of new code in ext4_ext_map_blocks(). Secondly the design of it is not very effective. The reason is that we are trying to punch out blocks in ext4_ext_punch_hole() in opposite direction than in ext4_ext_rm_leaf() which causes the ext4_ext_rm_leaf() to iterate through the whole tree from the end to the start to find the requested extent for every extent we are going to punch out. And finally the current implementation does not use the existing code, but bring a lot of new code, which is IMO unnecessary since there already is some infrastructure we can use. Specifically ext4_ext_remove_space(). This commit changes ext4_ext_remove_space() to accept 'end' parameter so we can not only truncate to the end of file, but also remove the space in the middle of the file (punch a hole). Moreover, because the last block to punch out, might be in the middle of the extent, we have to split the extent at 'end + 1' so ext4_ext_rm_leaf() can easily either remove the whole fist part of split extent, or change its size. ext4_ext_remove_space() is then used to actually remove the space (extents) from within the hole, instead of ext4_ext_map_blocks(). Note that this also fix the issue with punch hole, where we would forget to remove empty index blocks from the extent tree, resulting in double free block error and file system corruption. This is simply because we now use different code path, where this problem does not exist. This has been tested with fsx running for several days and xfstests, plus xfstest #251 with '-o discard' run on the loop image (which converts discard requestes into punch hole to the backing file). All of it on 1K and 4K file system block size. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-03-20 11:03:19 +08:00
/*
* Check if we are removing extents inside the extent tree. If that
* is the case, we are going to punch a hole inside the extent tree
* so we have to check whether we need to split the extent covering
* the last block to remove so we can easily remove the part of it
* in ext4_ext_rm_leaf().
*/
if (end < EXT_MAX_BLOCKS - 1) {
struct ext4_extent *ex;
ext4_lblk_t ee_block;
/* find extent for this block */
path = ext4_ext_find_extent(inode, end, NULL);
if (IS_ERR(path)) {
ext4_journal_stop(handle);
return PTR_ERR(path);
}
depth = ext_depth(inode);
/* Leaf not may not exist only if inode has no blocks at all */
ext4: rewrite punch hole to use ext4_ext_remove_space() This commit rewrites ext4 punch hole implementation to use ext4_ext_remove_space() instead of its home gown way of doing this via ext4_ext_map_blocks(). There are several reasons for changing this. Firstly it is quite non obvious that punching hole needs to ext4_ext_map_blocks() to punch a hole, especially given that this function should map blocks, not unmap it. It also required a lot of new code in ext4_ext_map_blocks(). Secondly the design of it is not very effective. The reason is that we are trying to punch out blocks in ext4_ext_punch_hole() in opposite direction than in ext4_ext_rm_leaf() which causes the ext4_ext_rm_leaf() to iterate through the whole tree from the end to the start to find the requested extent for every extent we are going to punch out. And finally the current implementation does not use the existing code, but bring a lot of new code, which is IMO unnecessary since there already is some infrastructure we can use. Specifically ext4_ext_remove_space(). This commit changes ext4_ext_remove_space() to accept 'end' parameter so we can not only truncate to the end of file, but also remove the space in the middle of the file (punch a hole). Moreover, because the last block to punch out, might be in the middle of the extent, we have to split the extent at 'end + 1' so ext4_ext_rm_leaf() can easily either remove the whole fist part of split extent, or change its size. ext4_ext_remove_space() is then used to actually remove the space (extents) from within the hole, instead of ext4_ext_map_blocks(). Note that this also fix the issue with punch hole, where we would forget to remove empty index blocks from the extent tree, resulting in double free block error and file system corruption. This is simply because we now use different code path, where this problem does not exist. This has been tested with fsx running for several days and xfstests, plus xfstest #251 with '-o discard' run on the loop image (which converts discard requestes into punch hole to the backing file). All of it on 1K and 4K file system block size. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-03-20 11:03:19 +08:00
ex = path[depth].p_ext;
if (!ex) {
if (depth) {
EXT4_ERROR_INODE(inode,
"path[%d].p_hdr == NULL",
depth);
err = -EIO;
}
goto out;
}
ext4: rewrite punch hole to use ext4_ext_remove_space() This commit rewrites ext4 punch hole implementation to use ext4_ext_remove_space() instead of its home gown way of doing this via ext4_ext_map_blocks(). There are several reasons for changing this. Firstly it is quite non obvious that punching hole needs to ext4_ext_map_blocks() to punch a hole, especially given that this function should map blocks, not unmap it. It also required a lot of new code in ext4_ext_map_blocks(). Secondly the design of it is not very effective. The reason is that we are trying to punch out blocks in ext4_ext_punch_hole() in opposite direction than in ext4_ext_rm_leaf() which causes the ext4_ext_rm_leaf() to iterate through the whole tree from the end to the start to find the requested extent for every extent we are going to punch out. And finally the current implementation does not use the existing code, but bring a lot of new code, which is IMO unnecessary since there already is some infrastructure we can use. Specifically ext4_ext_remove_space(). This commit changes ext4_ext_remove_space() to accept 'end' parameter so we can not only truncate to the end of file, but also remove the space in the middle of the file (punch a hole). Moreover, because the last block to punch out, might be in the middle of the extent, we have to split the extent at 'end + 1' so ext4_ext_rm_leaf() can easily either remove the whole fist part of split extent, or change its size. ext4_ext_remove_space() is then used to actually remove the space (extents) from within the hole, instead of ext4_ext_map_blocks(). Note that this also fix the issue with punch hole, where we would forget to remove empty index blocks from the extent tree, resulting in double free block error and file system corruption. This is simply because we now use different code path, where this problem does not exist. This has been tested with fsx running for several days and xfstests, plus xfstest #251 with '-o discard' run on the loop image (which converts discard requestes into punch hole to the backing file). All of it on 1K and 4K file system block size. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-03-20 11:03:19 +08:00
ee_block = le32_to_cpu(ex->ee_block);
/*
* See if the last block is inside the extent, if so split
* the extent at 'end' block so we can easily remove the
* tail of the first part of the split extent in
* ext4_ext_rm_leaf().
*/
if (end >= ee_block &&
end < ee_block + ext4_ext_get_actual_len(ex) - 1) {
int split_flag = 0;
if (ext4_ext_is_uninitialized(ex))
split_flag = EXT4_EXT_MARK_UNINIT1 |
EXT4_EXT_MARK_UNINIT2;
/*
* Split the extent in two so that 'end' is the last
ext4: introduce reserved space Currently in ENOSPC condition when writing into unwritten space, or punching a hole, we might need to split the extent and grow extent tree. However since we can not allocate any new metadata blocks we'll have to zero out unwritten part of extent or punched out part of extent, or in the worst case return ENOSPC even though use actually does not allocate any space. Also in delalloc path we do reserve metadata and data blocks for the time we're going to write out, however metadata block reservation is very tricky especially since we expect that logical connectivity implies physical connectivity, however that might not be the case and hence we might end up allocating more metadata blocks than previously reserved. So in future, metadata reservation checks should be removed since we can not assure that we do not under reserve. And this is where reserved space comes into the picture. When mounting the file system we slice off a little bit of the file system space (2% or 4096 clusters, whichever is smaller) which can be then used for the cases mentioned above to prevent costly zeroout, or unexpected ENOSPC. The number of reserved clusters can be set via sysfs, however it can never be bigger than number of free clusters in the file system. Note that this patch fixes the failure of xfstest 274 as expected. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
2013-04-10 10:11:22 +08:00
* block in the first new extent. Also we should not
* fail removing space due to ENOSPC so try to use
* reserved block if that happens.
ext4: rewrite punch hole to use ext4_ext_remove_space() This commit rewrites ext4 punch hole implementation to use ext4_ext_remove_space() instead of its home gown way of doing this via ext4_ext_map_blocks(). There are several reasons for changing this. Firstly it is quite non obvious that punching hole needs to ext4_ext_map_blocks() to punch a hole, especially given that this function should map blocks, not unmap it. It also required a lot of new code in ext4_ext_map_blocks(). Secondly the design of it is not very effective. The reason is that we are trying to punch out blocks in ext4_ext_punch_hole() in opposite direction than in ext4_ext_rm_leaf() which causes the ext4_ext_rm_leaf() to iterate through the whole tree from the end to the start to find the requested extent for every extent we are going to punch out. And finally the current implementation does not use the existing code, but bring a lot of new code, which is IMO unnecessary since there already is some infrastructure we can use. Specifically ext4_ext_remove_space(). This commit changes ext4_ext_remove_space() to accept 'end' parameter so we can not only truncate to the end of file, but also remove the space in the middle of the file (punch a hole). Moreover, because the last block to punch out, might be in the middle of the extent, we have to split the extent at 'end + 1' so ext4_ext_rm_leaf() can easily either remove the whole fist part of split extent, or change its size. ext4_ext_remove_space() is then used to actually remove the space (extents) from within the hole, instead of ext4_ext_map_blocks(). Note that this also fix the issue with punch hole, where we would forget to remove empty index blocks from the extent tree, resulting in double free block error and file system corruption. This is simply because we now use different code path, where this problem does not exist. This has been tested with fsx running for several days and xfstests, plus xfstest #251 with '-o discard' run on the loop image (which converts discard requestes into punch hole to the backing file). All of it on 1K and 4K file system block size. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-03-20 11:03:19 +08:00
*/
err = ext4_split_extent_at(handle, inode, path,
ext4: introduce reserved space Currently in ENOSPC condition when writing into unwritten space, or punching a hole, we might need to split the extent and grow extent tree. However since we can not allocate any new metadata blocks we'll have to zero out unwritten part of extent or punched out part of extent, or in the worst case return ENOSPC even though use actually does not allocate any space. Also in delalloc path we do reserve metadata and data blocks for the time we're going to write out, however metadata block reservation is very tricky especially since we expect that logical connectivity implies physical connectivity, however that might not be the case and hence we might end up allocating more metadata blocks than previously reserved. So in future, metadata reservation checks should be removed since we can not assure that we do not under reserve. And this is where reserved space comes into the picture. When mounting the file system we slice off a little bit of the file system space (2% or 4096 clusters, whichever is smaller) which can be then used for the cases mentioned above to prevent costly zeroout, or unexpected ENOSPC. The number of reserved clusters can be set via sysfs, however it can never be bigger than number of free clusters in the file system. Note that this patch fixes the failure of xfstest 274 as expected. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
2013-04-10 10:11:22 +08:00
end + 1, split_flag,
EXT4_GET_BLOCKS_PRE_IO |
EXT4_GET_BLOCKS_METADATA_NOFAIL);
ext4: rewrite punch hole to use ext4_ext_remove_space() This commit rewrites ext4 punch hole implementation to use ext4_ext_remove_space() instead of its home gown way of doing this via ext4_ext_map_blocks(). There are several reasons for changing this. Firstly it is quite non obvious that punching hole needs to ext4_ext_map_blocks() to punch a hole, especially given that this function should map blocks, not unmap it. It also required a lot of new code in ext4_ext_map_blocks(). Secondly the design of it is not very effective. The reason is that we are trying to punch out blocks in ext4_ext_punch_hole() in opposite direction than in ext4_ext_rm_leaf() which causes the ext4_ext_rm_leaf() to iterate through the whole tree from the end to the start to find the requested extent for every extent we are going to punch out. And finally the current implementation does not use the existing code, but bring a lot of new code, which is IMO unnecessary since there already is some infrastructure we can use. Specifically ext4_ext_remove_space(). This commit changes ext4_ext_remove_space() to accept 'end' parameter so we can not only truncate to the end of file, but also remove the space in the middle of the file (punch a hole). Moreover, because the last block to punch out, might be in the middle of the extent, we have to split the extent at 'end + 1' so ext4_ext_rm_leaf() can easily either remove the whole fist part of split extent, or change its size. ext4_ext_remove_space() is then used to actually remove the space (extents) from within the hole, instead of ext4_ext_map_blocks(). Note that this also fix the issue with punch hole, where we would forget to remove empty index blocks from the extent tree, resulting in double free block error and file system corruption. This is simply because we now use different code path, where this problem does not exist. This has been tested with fsx running for several days and xfstests, plus xfstest #251 with '-o discard' run on the loop image (which converts discard requestes into punch hole to the backing file). All of it on 1K and 4K file system block size. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-03-20 11:03:19 +08:00
if (err < 0)
goto out;
}
}
/*
* We start scanning from right side, freeing all the blocks
* after i_size and walking into the tree depth-wise.
*/
depth = ext_depth(inode);
if (path) {
int k = i = depth;
while (--k > 0)
path[k].p_block =
le16_to_cpu(path[k].p_hdr->eh_entries)+1;
} else {
path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 1),
GFP_NOFS);
if (path == NULL) {
ext4_journal_stop(handle);
return -ENOMEM;
}
path[0].p_depth = depth;
path[0].p_hdr = ext_inode_hdr(inode);
ext4: fix kernel BUG on large-scale rm -rf commands Commit 968dee7722: "ext4: fix hole punch failure when depth is greater than 0" introduced a regression in v3.5.1/v3.6-rc1 which caused kernel crashes when users ran run "rm -rf" on large directory hierarchy on ext4 filesystems on RAID devices: BUG: unable to handle kernel NULL pointer dereference at 0000000000000028 Process rm (pid: 18229, threadinfo ffff8801276bc000, task ffff880123631710) Call Trace: [<ffffffff81236483>] ? __ext4_handle_dirty_metadata+0x83/0x110 [<ffffffff812353d3>] ext4_ext_truncate+0x193/0x1d0 [<ffffffff8120a8cf>] ? ext4_mark_inode_dirty+0x7f/0x1f0 [<ffffffff81207e05>] ext4_truncate+0xf5/0x100 [<ffffffff8120cd51>] ext4_evict_inode+0x461/0x490 [<ffffffff811a1312>] evict+0xa2/0x1a0 [<ffffffff811a1513>] iput+0x103/0x1f0 [<ffffffff81196d84>] do_unlinkat+0x154/0x1c0 [<ffffffff8118cc3a>] ? sys_newfstatat+0x2a/0x40 [<ffffffff81197b0b>] sys_unlinkat+0x1b/0x50 [<ffffffff816135e9>] system_call_fastpath+0x16/0x1b Code: 8b 4d 20 0f b7 41 02 48 8d 04 40 48 8d 04 81 49 89 45 18 0f b7 49 02 48 83 c1 01 49 89 4d 00 e9 ae f8 ff ff 0f 1f 00 49 8b 45 28 <48> 8b 40 28 49 89 45 20 e9 85 f8 ff ff 0f 1f 80 00 00 00 RIP [<ffffffff81233164>] ext4_ext_remove_space+0xa34/0xdf0 This could be reproduced as follows: The problem in commit 968dee7722 was that caused the variable 'i' to be left uninitialized if the truncate required more space than was available in the journal. This resulted in the function ext4_ext_truncate_extend_restart() returning -EAGAIN, which caused ext4_ext_remove_space() to restart the truncate operation after starting a new jbd2 handle. Reported-by: Maciej Żenczykowski <maze@google.com> Reported-by: Marti Raudsepp <marti@juffo.org> Tested-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Cc: stable@vger.kernel.org
2012-08-17 20:54:52 +08:00
i = 0;
ext4: rewrite punch hole to use ext4_ext_remove_space() This commit rewrites ext4 punch hole implementation to use ext4_ext_remove_space() instead of its home gown way of doing this via ext4_ext_map_blocks(). There are several reasons for changing this. Firstly it is quite non obvious that punching hole needs to ext4_ext_map_blocks() to punch a hole, especially given that this function should map blocks, not unmap it. It also required a lot of new code in ext4_ext_map_blocks(). Secondly the design of it is not very effective. The reason is that we are trying to punch out blocks in ext4_ext_punch_hole() in opposite direction than in ext4_ext_rm_leaf() which causes the ext4_ext_rm_leaf() to iterate through the whole tree from the end to the start to find the requested extent for every extent we are going to punch out. And finally the current implementation does not use the existing code, but bring a lot of new code, which is IMO unnecessary since there already is some infrastructure we can use. Specifically ext4_ext_remove_space(). This commit changes ext4_ext_remove_space() to accept 'end' parameter so we can not only truncate to the end of file, but also remove the space in the middle of the file (punch a hole). Moreover, because the last block to punch out, might be in the middle of the extent, we have to split the extent at 'end + 1' so ext4_ext_rm_leaf() can easily either remove the whole fist part of split extent, or change its size. ext4_ext_remove_space() is then used to actually remove the space (extents) from within the hole, instead of ext4_ext_map_blocks(). Note that this also fix the issue with punch hole, where we would forget to remove empty index blocks from the extent tree, resulting in double free block error and file system corruption. This is simply because we now use different code path, where this problem does not exist. This has been tested with fsx running for several days and xfstests, plus xfstest #251 with '-o discard' run on the loop image (which converts discard requestes into punch hole to the backing file). All of it on 1K and 4K file system block size. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-03-20 11:03:19 +08:00
if (ext4_ext_check(inode, path[0].p_hdr, depth)) {
err = -EIO;
goto out;
}
}
err = 0;
while (i >= 0 && err == 0) {
if (i == depth) {
/* this is leaf block */
err = ext4_ext_rm_leaf(handle, inode, path,
&partial_cluster, start,
ext4: rewrite punch hole to use ext4_ext_remove_space() This commit rewrites ext4 punch hole implementation to use ext4_ext_remove_space() instead of its home gown way of doing this via ext4_ext_map_blocks(). There are several reasons for changing this. Firstly it is quite non obvious that punching hole needs to ext4_ext_map_blocks() to punch a hole, especially given that this function should map blocks, not unmap it. It also required a lot of new code in ext4_ext_map_blocks(). Secondly the design of it is not very effective. The reason is that we are trying to punch out blocks in ext4_ext_punch_hole() in opposite direction than in ext4_ext_rm_leaf() which causes the ext4_ext_rm_leaf() to iterate through the whole tree from the end to the start to find the requested extent for every extent we are going to punch out. And finally the current implementation does not use the existing code, but bring a lot of new code, which is IMO unnecessary since there already is some infrastructure we can use. Specifically ext4_ext_remove_space(). This commit changes ext4_ext_remove_space() to accept 'end' parameter so we can not only truncate to the end of file, but also remove the space in the middle of the file (punch a hole). Moreover, because the last block to punch out, might be in the middle of the extent, we have to split the extent at 'end + 1' so ext4_ext_rm_leaf() can easily either remove the whole fist part of split extent, or change its size. ext4_ext_remove_space() is then used to actually remove the space (extents) from within the hole, instead of ext4_ext_map_blocks(). Note that this also fix the issue with punch hole, where we would forget to remove empty index blocks from the extent tree, resulting in double free block error and file system corruption. This is simply because we now use different code path, where this problem does not exist. This has been tested with fsx running for several days and xfstests, plus xfstest #251 with '-o discard' run on the loop image (which converts discard requestes into punch hole to the backing file). All of it on 1K and 4K file system block size. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-03-20 11:03:19 +08:00
end);
/* root level has p_bh == NULL, brelse() eats this */
brelse(path[i].p_bh);
path[i].p_bh = NULL;
i--;
continue;
}
/* this is index block */
if (!path[i].p_hdr) {
ext_debug("initialize header\n");
path[i].p_hdr = ext_block_hdr(path[i].p_bh);
}
if (!path[i].p_idx) {
/* this level hasn't been touched yet */
path[i].p_idx = EXT_LAST_INDEX(path[i].p_hdr);
path[i].p_block = le16_to_cpu(path[i].p_hdr->eh_entries)+1;
ext_debug("init index ptr: hdr 0x%p, num %d\n",
path[i].p_hdr,
le16_to_cpu(path[i].p_hdr->eh_entries));
} else {
/* we were already here, see at next index */
path[i].p_idx--;
}
ext_debug("level %d - index, first 0x%p, cur 0x%p\n",
i, EXT_FIRST_INDEX(path[i].p_hdr),
path[i].p_idx);
if (ext4_ext_more_to_rm(path + i)) {
struct buffer_head *bh;
/* go to the next level */
ext_debug("move to level %d (block %llu)\n",
i + 1, ext4_idx_pblock(path[i].p_idx));
memset(path + i + 1, 0, sizeof(*path));
bh = sb_bread(sb, ext4_idx_pblock(path[i].p_idx));
if (!bh) {
/* should we reset i_size? */
err = -EIO;
break;
}
if (WARN_ON(i + 1 > depth)) {
err = -EIO;
break;
}
if (ext4_ext_check_block(inode, ext_block_hdr(bh),
depth - i - 1, bh)) {
err = -EIO;
break;
}
path[i + 1].p_bh = bh;
/* save actual number of indexes since this
* number is changed at the next iteration */
path[i].p_block = le16_to_cpu(path[i].p_hdr->eh_entries);
i++;
} else {
/* we finished processing this index, go up */
if (path[i].p_hdr->eh_entries == 0 && i > 0) {
/* index is empty, remove it;
* handle must be already prepared by the
* truncatei_leaf() */
err = ext4_ext_rm_idx(handle, inode, path, i);
}
/* root level has p_bh == NULL, brelse() eats this */
brelse(path[i].p_bh);
path[i].p_bh = NULL;
i--;
ext_debug("return to level %d\n", i);
}
}
trace_ext4_ext_remove_space_done(inode, start, depth, partial_cluster,
path->p_hdr->eh_entries);
/* If we still have something in the partial cluster and we have removed
* even the first extent, then we should free the blocks in the partial
* cluster as well. */
if (partial_cluster && path->p_hdr->eh_entries == 0) {
int flags = EXT4_FREE_BLOCKS_FORGET;
if (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode))
flags |= EXT4_FREE_BLOCKS_METADATA;
ext4_free_blocks(handle, inode, NULL,
EXT4_C2B(EXT4_SB(sb), partial_cluster),
EXT4_SB(sb)->s_cluster_ratio, flags);
partial_cluster = 0;
}
/* TODO: flexible tree reduction should be here */
if (path->p_hdr->eh_entries == 0) {
/*
* truncate to zero freed all the tree,
* so we need to correct eh_depth
*/
err = ext4_ext_get_access(handle, inode, path);
if (err == 0) {
ext_inode_hdr(inode)->eh_depth = 0;
ext_inode_hdr(inode)->eh_max =
cpu_to_le16(ext4_ext_space_root(inode, 0));
err = ext4_ext_dirty(handle, inode, path);
}
}
out:
ext4_ext_drop_refs(path);
kfree(path);
if (err == -EAGAIN) {
path = NULL;
goto again;
}
ext4_journal_stop(handle);
return err;
}
/*
* called at mount time
*/
void ext4_ext_init(struct super_block *sb)
{
/*
* possible initialization would be here
*/
if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS)) {
#if defined(AGGRESSIVE_TEST) || defined(CHECK_BINSEARCH) || defined(EXTENTS_STATS)
printk(KERN_INFO "EXT4-fs: file extents enabled"
#ifdef AGGRESSIVE_TEST
", aggressive tests"
#endif
#ifdef CHECK_BINSEARCH
", check binsearch"
#endif
#ifdef EXTENTS_STATS
", stats"
#endif
"\n");
#endif
#ifdef EXTENTS_STATS
spin_lock_init(&EXT4_SB(sb)->s_ext_stats_lock);
EXT4_SB(sb)->s_ext_min = 1 << 30;
EXT4_SB(sb)->s_ext_max = 0;
#endif
}
}
/*
* called at umount time
*/
void ext4_ext_release(struct super_block *sb)
{
if (!EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS))
return;
#ifdef EXTENTS_STATS
if (EXT4_SB(sb)->s_ext_blocks && EXT4_SB(sb)->s_ext_extents) {
struct ext4_sb_info *sbi = EXT4_SB(sb);
printk(KERN_ERR "EXT4-fs: %lu blocks in %lu extents (%lu ave)\n",
sbi->s_ext_blocks, sbi->s_ext_extents,
sbi->s_ext_blocks / sbi->s_ext_extents);
printk(KERN_ERR "EXT4-fs: extents: %lu min, %lu max, max depth %lu\n",
sbi->s_ext_min, sbi->s_ext_max, sbi->s_depth_max);
}
#endif
}
/* FIXME!! we need to try to merge to left or right after zero-out */
static int ext4_ext_zeroout(struct inode *inode, struct ext4_extent *ex)
{
ext4_fsblk_t ee_pblock;
unsigned int ee_len;
int ret;
ee_len = ext4_ext_get_actual_len(ex);
ee_pblock = ext4_ext_pblock(ex);
ret = sb_issue_zeroout(inode->i_sb, ee_pblock, ee_len, GFP_NOFS);
if (ret > 0)
ret = 0;
return ret;
}
/*
* ext4_split_extent_at() splits an extent at given block.
*
* @handle: the journal handle
* @inode: the file inode
* @path: the path to the extent
* @split: the logical block where the extent is splitted.
* @split_flags: indicates if the extent could be zeroout if split fails, and
* the states(init or uninit) of new extents.
* @flags: flags used to insert new extent to extent tree.
*
*
* Splits extent [a, b] into two extents [a, @split) and [@split, b], states
* of which are deterimined by split_flag.
*
* There are two cases:
* a> the extent are splitted into two extent.
* b> split is not needed, and just mark the extent.
*
* return 0 on success.
*/
static int ext4_split_extent_at(handle_t *handle,
struct inode *inode,
struct ext4_ext_path *path,
ext4_lblk_t split,
int split_flag,
int flags)
{
ext4_fsblk_t newblock;
ext4_lblk_t ee_block;
struct ext4_extent *ex, newex, orig_ex, zero_ex;
struct ext4_extent *ex2 = NULL;
unsigned int ee_len, depth;
int err = 0;
BUG_ON((split_flag & (EXT4_EXT_DATA_VALID1 | EXT4_EXT_DATA_VALID2)) ==
(EXT4_EXT_DATA_VALID1 | EXT4_EXT_DATA_VALID2));
ext_debug("ext4_split_extents_at: inode %lu, logical"
"block %llu\n", inode->i_ino, (unsigned long long)split);
ext4_ext_show_leaf(inode, path);
depth = ext_depth(inode);
ex = path[depth].p_ext;
ee_block = le32_to_cpu(ex->ee_block);
ee_len = ext4_ext_get_actual_len(ex);
newblock = split - ee_block + ext4_ext_pblock(ex);
BUG_ON(split < ee_block || split >= (ee_block + ee_len));
BUG_ON(!ext4_ext_is_uninitialized(ex) &&
split_flag & (EXT4_EXT_MAY_ZEROOUT |
EXT4_EXT_MARK_UNINIT1 |
EXT4_EXT_MARK_UNINIT2));
err = ext4_ext_get_access(handle, inode, path + depth);
if (err)
goto out;
if (split == ee_block) {
/*
* case b: block @split is the block that the extent begins with
* then we just change the state of the extent, and splitting
* is not needed.
*/
if (split_flag & EXT4_EXT_MARK_UNINIT2)
ext4_ext_mark_uninitialized(ex);
else
ext4_ext_mark_initialized(ex);
if (!(flags & EXT4_GET_BLOCKS_PRE_IO))
ext4_ext_try_to_merge(handle, inode, path, ex);
err = ext4_ext_dirty(handle, inode, path + path->p_depth);
goto out;
}
/* case a */
memcpy(&orig_ex, ex, sizeof(orig_ex));
ex->ee_len = cpu_to_le16(split - ee_block);
if (split_flag & EXT4_EXT_MARK_UNINIT1)
ext4_ext_mark_uninitialized(ex);
/*
* path may lead to new leaf, not to original leaf any more
* after ext4_ext_insert_extent() returns,
*/
err = ext4_ext_dirty(handle, inode, path + depth);
if (err)
goto fix_extent_len;
ex2 = &newex;
ex2->ee_block = cpu_to_le32(split);
ex2->ee_len = cpu_to_le16(ee_len - (split - ee_block));
ext4_ext_store_pblock(ex2, newblock);
if (split_flag & EXT4_EXT_MARK_UNINIT2)
ext4_ext_mark_uninitialized(ex2);
err = ext4_ext_insert_extent(handle, inode, path, &newex, flags);
if (err == -ENOSPC && (EXT4_EXT_MAY_ZEROOUT & split_flag)) {
if (split_flag & (EXT4_EXT_DATA_VALID1|EXT4_EXT_DATA_VALID2)) {
if (split_flag & EXT4_EXT_DATA_VALID1) {
err = ext4_ext_zeroout(inode, ex2);
zero_ex.ee_block = ex2->ee_block;
zero_ex.ee_len = cpu_to_le16(
ext4_ext_get_actual_len(ex2));
ext4_ext_store_pblock(&zero_ex,
ext4_ext_pblock(ex2));
} else {
err = ext4_ext_zeroout(inode, ex);
zero_ex.ee_block = ex->ee_block;
zero_ex.ee_len = cpu_to_le16(
ext4_ext_get_actual_len(ex));
ext4_ext_store_pblock(&zero_ex,
ext4_ext_pblock(ex));
}
} else {
err = ext4_ext_zeroout(inode, &orig_ex);
zero_ex.ee_block = orig_ex.ee_block;
zero_ex.ee_len = cpu_to_le16(
ext4_ext_get_actual_len(&orig_ex));
ext4_ext_store_pblock(&zero_ex,
ext4_ext_pblock(&orig_ex));
}
if (err)
goto fix_extent_len;
/* update the extent length and mark as initialized */
ex->ee_len = cpu_to_le16(ee_len);
ext4_ext_try_to_merge(handle, inode, path, ex);
err = ext4_ext_dirty(handle, inode, path + path->p_depth);
if (err)
goto fix_extent_len;
/* update extent status tree */
err = ext4_es_zeroout(inode, &zero_ex);
goto out;
} else if (err)
goto fix_extent_len;
out:
ext4_ext_show_leaf(inode, path);
return err;
fix_extent_len:
ex->ee_len = orig_ex.ee_len;
ext4_ext_dirty(handle, inode, path + depth);
return err;
}
/*
* ext4_split_extents() splits an extent and mark extent which is covered
* by @map as split_flags indicates
*
* It may result in splitting the extent into multiple extents (upto three)
* There are three possibilities:
* a> There is no split required
* b> Splits in two extents: Split is happening at either end of the extent
* c> Splits in three extents: Somone is splitting in middle of the extent
*
*/
static int ext4_split_extent(handle_t *handle,
struct inode *inode,
struct ext4_ext_path *path,
struct ext4_map_blocks *map,
int split_flag,
int flags)
{
ext4_lblk_t ee_block;
struct ext4_extent *ex;
unsigned int ee_len, depth;
int err = 0;
int uninitialized;
int split_flag1, flags1;
int allocated = map->m_len;
depth = ext_depth(inode);
ex = path[depth].p_ext;
ee_block = le32_to_cpu(ex->ee_block);
ee_len = ext4_ext_get_actual_len(ex);
uninitialized = ext4_ext_is_uninitialized(ex);
if (map->m_lblk + map->m_len < ee_block + ee_len) {
split_flag1 = split_flag & EXT4_EXT_MAY_ZEROOUT;
flags1 = flags | EXT4_GET_BLOCKS_PRE_IO;
if (uninitialized)
split_flag1 |= EXT4_EXT_MARK_UNINIT1 |
EXT4_EXT_MARK_UNINIT2;
if (split_flag & EXT4_EXT_DATA_VALID2)
split_flag1 |= EXT4_EXT_DATA_VALID1;
err = ext4_split_extent_at(handle, inode, path,
map->m_lblk + map->m_len, split_flag1, flags1);
if (err)
goto out;
} else {
allocated = ee_len - (map->m_lblk - ee_block);
}
/*
* Update path is required because previous ext4_split_extent_at() may
* result in split of original leaf or extent zeroout.
*/
ext4_ext_drop_refs(path);
path = ext4_ext_find_extent(inode, map->m_lblk, path);
if (IS_ERR(path))
return PTR_ERR(path);
depth = ext_depth(inode);
ex = path[depth].p_ext;
uninitialized = ext4_ext_is_uninitialized(ex);
split_flag1 = 0;
if (map->m_lblk >= ee_block) {
split_flag1 = split_flag & EXT4_EXT_DATA_VALID2;
if (uninitialized) {
split_flag1 |= EXT4_EXT_MARK_UNINIT1;
split_flag1 |= split_flag & (EXT4_EXT_MAY_ZEROOUT |
EXT4_EXT_MARK_UNINIT2);
}
err = ext4_split_extent_at(handle, inode, path,
map->m_lblk, split_flag1, flags);
if (err)
goto out;
}
ext4_ext_show_leaf(inode, path);
out:
return err ? err : allocated;
}
/*
* This function is called by ext4_ext_map_blocks() if someone tries to write
* to an uninitialized extent. It may result in splitting the uninitialized
* extent into multiple extents (up to three - one initialized and two
* uninitialized).
* There are three possibilities:
* a> There is no split required: Entire extent should be initialized
* b> Splits in two extents: Write is happening at either end of the extent
* c> Splits in three extents: Somone is writing in middle of the extent
*
* Pre-conditions:
* - The extent pointed to by 'path' is uninitialized.
* - The extent pointed to by 'path' contains a superset
* of the logical span [map->m_lblk, map->m_lblk + map->m_len).
*
* Post-conditions on success:
* - the returned value is the number of blocks beyond map->l_lblk
* that are allocated and initialized.
* It is guaranteed to be >= map->m_len.
*/
static int ext4_ext_convert_to_initialized(handle_t *handle,
struct inode *inode,
struct ext4_map_blocks *map,
ext4: introduce reserved space Currently in ENOSPC condition when writing into unwritten space, or punching a hole, we might need to split the extent and grow extent tree. However since we can not allocate any new metadata blocks we'll have to zero out unwritten part of extent or punched out part of extent, or in the worst case return ENOSPC even though use actually does not allocate any space. Also in delalloc path we do reserve metadata and data blocks for the time we're going to write out, however metadata block reservation is very tricky especially since we expect that logical connectivity implies physical connectivity, however that might not be the case and hence we might end up allocating more metadata blocks than previously reserved. So in future, metadata reservation checks should be removed since we can not assure that we do not under reserve. And this is where reserved space comes into the picture. When mounting the file system we slice off a little bit of the file system space (2% or 4096 clusters, whichever is smaller) which can be then used for the cases mentioned above to prevent costly zeroout, or unexpected ENOSPC. The number of reserved clusters can be set via sysfs, however it can never be bigger than number of free clusters in the file system. Note that this patch fixes the failure of xfstest 274 as expected. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
2013-04-10 10:11:22 +08:00
struct ext4_ext_path *path,
int flags)
{
struct ext4_sb_info *sbi;
struct ext4_extent_header *eh;
struct ext4_map_blocks split_map;
struct ext4_extent zero_ex;
struct ext4_extent *ex, *abut_ex;
ext4_lblk_t ee_block, eof_block;
unsigned int ee_len, depth, map_len = map->m_len;
int allocated = 0, max_zeroout = 0;
int err = 0;
int split_flag = 0;
ext_debug("ext4_ext_convert_to_initialized: inode %lu, logical"
"block %llu, max_blocks %u\n", inode->i_ino,
(unsigned long long)map->m_lblk, map_len);
sbi = EXT4_SB(inode->i_sb);
eof_block = (inode->i_size + inode->i_sb->s_blocksize - 1) >>
inode->i_sb->s_blocksize_bits;
if (eof_block < map->m_lblk + map_len)
eof_block = map->m_lblk + map_len;
depth = ext_depth(inode);
eh = path[depth].p_hdr;
ex = path[depth].p_ext;
ee_block = le32_to_cpu(ex->ee_block);
ee_len = ext4_ext_get_actual_len(ex);
zero_ex.ee_len = 0;
trace_ext4_ext_convert_to_initialized_enter(inode, map, ex);
/* Pre-conditions */
BUG_ON(!ext4_ext_is_uninitialized(ex));
BUG_ON(!in_range(map->m_lblk, ee_block, ee_len));
/*
* Attempt to transfer newly initialized blocks from the currently
* uninitialized extent to its neighbor. This is much cheaper
* than an insertion followed by a merge as those involve costly
* memmove() calls. Transferring to the left is the common case in
* steady state for workloads doing fallocate(FALLOC_FL_KEEP_SIZE)
* followed by append writes.
*
* Limitations of the current logic:
* - L1: we do not deal with writes covering the whole extent.
* This would require removing the extent if the transfer
* is possible.
* - L2: we only attempt to merge with an extent stored in the
* same extent tree node.
*/
if ((map->m_lblk == ee_block) &&
/* See if we can merge left */
(map_len < ee_len) && /*L1*/
(ex > EXT_FIRST_EXTENT(eh))) { /*L2*/
ext4_lblk_t prev_lblk;
ext4_fsblk_t prev_pblk, ee_pblk;
unsigned int prev_len;
abut_ex = ex - 1;
prev_lblk = le32_to_cpu(abut_ex->ee_block);
prev_len = ext4_ext_get_actual_len(abut_ex);
prev_pblk = ext4_ext_pblock(abut_ex);
ee_pblk = ext4_ext_pblock(ex);
/*
* A transfer of blocks from 'ex' to 'abut_ex' is allowed
* upon those conditions:
* - C1: abut_ex is initialized,
* - C2: abut_ex is logically abutting ex,
* - C3: abut_ex is physically abutting ex,
* - C4: abut_ex can receive the additional blocks without
* overflowing the (initialized) length limit.
*/
if ((!ext4_ext_is_uninitialized(abut_ex)) && /*C1*/
((prev_lblk + prev_len) == ee_block) && /*C2*/
((prev_pblk + prev_len) == ee_pblk) && /*C3*/
(prev_len < (EXT_INIT_MAX_LEN - map_len))) { /*C4*/
err = ext4_ext_get_access(handle, inode, path + depth);
if (err)
goto out;
trace_ext4_ext_convert_to_initialized_fastpath(inode,
map, ex, abut_ex);
/* Shift the start of ex by 'map_len' blocks */
ex->ee_block = cpu_to_le32(ee_block + map_len);
ext4_ext_store_pblock(ex, ee_pblk + map_len);
ex->ee_len = cpu_to_le16(ee_len - map_len);
ext4_ext_mark_uninitialized(ex); /* Restore the flag */
/* Extend abut_ex by 'map_len' blocks */
abut_ex->ee_len = cpu_to_le16(prev_len + map_len);
/* Result: number of initialized blocks past m_lblk */
allocated = map_len;
}
} else if (((map->m_lblk + map_len) == (ee_block + ee_len)) &&
(map_len < ee_len) && /*L1*/
ex < EXT_LAST_EXTENT(eh)) { /*L2*/
/* See if we can merge right */
ext4_lblk_t next_lblk;
ext4_fsblk_t next_pblk, ee_pblk;
unsigned int next_len;
abut_ex = ex + 1;
next_lblk = le32_to_cpu(abut_ex->ee_block);
next_len = ext4_ext_get_actual_len(abut_ex);
next_pblk = ext4_ext_pblock(abut_ex);
ee_pblk = ext4_ext_pblock(ex);
/*
* A transfer of blocks from 'ex' to 'abut_ex' is allowed
* upon those conditions:
* - C1: abut_ex is initialized,
* - C2: abut_ex is logically abutting ex,
* - C3: abut_ex is physically abutting ex,
* - C4: abut_ex can receive the additional blocks without
* overflowing the (initialized) length limit.
*/
if ((!ext4_ext_is_uninitialized(abut_ex)) && /*C1*/
((map->m_lblk + map_len) == next_lblk) && /*C2*/
((ee_pblk + ee_len) == next_pblk) && /*C3*/
(next_len < (EXT_INIT_MAX_LEN - map_len))) { /*C4*/
err = ext4_ext_get_access(handle, inode, path + depth);
if (err)
goto out;
trace_ext4_ext_convert_to_initialized_fastpath(inode,
map, ex, abut_ex);
/* Shift the start of abut_ex by 'map_len' blocks */
abut_ex->ee_block = cpu_to_le32(next_lblk - map_len);
ext4_ext_store_pblock(abut_ex, next_pblk - map_len);
ex->ee_len = cpu_to_le16(ee_len - map_len);
ext4_ext_mark_uninitialized(ex); /* Restore the flag */
/* Extend abut_ex by 'map_len' blocks */
abut_ex->ee_len = cpu_to_le16(next_len + map_len);
/* Result: number of initialized blocks past m_lblk */
allocated = map_len;
}
}
if (allocated) {
/* Mark the block containing both extents as dirty */
ext4_ext_dirty(handle, inode, path + depth);
/* Update path to point to the right extent */
path[depth].p_ext = abut_ex;
goto out;
} else
allocated = ee_len - (map->m_lblk - ee_block);
WARN_ON(map->m_lblk < ee_block);
/*
* It is safe to convert extent to initialized via explicit
* zeroout only if extent is fully insde i_size or new_size.
*/
split_flag |= ee_block + ee_len <= eof_block ? EXT4_EXT_MAY_ZEROOUT : 0;
if (EXT4_EXT_MAY_ZEROOUT & split_flag)
max_zeroout = sbi->s_extent_max_zeroout_kb >>
(inode->i_sb->s_blocksize_bits - 10);
/* If extent is less than s_max_zeroout_kb, zeroout directly */
if (max_zeroout && (ee_len <= max_zeroout)) {
err = ext4_ext_zeroout(inode, ex);
if (err)
goto out;
zero_ex.ee_block = ex->ee_block;
zero_ex.ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex));
ext4_ext_store_pblock(&zero_ex, ext4_ext_pblock(ex));
err = ext4_ext_get_access(handle, inode, path + depth);
if (err)
goto out;
ext4_ext_mark_initialized(ex);
ext4_ext_try_to_merge(handle, inode, path, ex);
err = ext4_ext_dirty(handle, inode, path + path->p_depth);
goto out;
}
/*
* four cases:
* 1. split the extent into three extents.
* 2. split the extent into two extents, zeroout the first half.
* 3. split the extent into two extents, zeroout the second half.
* 4. split the extent into two extents with out zeroout.
*/
split_map.m_lblk = map->m_lblk;
split_map.m_len = map->m_len;
if (max_zeroout && (allocated > map->m_len)) {
if (allocated <= max_zeroout) {
/* case 3 */
zero_ex.ee_block =
cpu_to_le32(map->m_lblk);
zero_ex.ee_len = cpu_to_le16(allocated);
ext4_ext_store_pblock(&zero_ex,
ext4_ext_pblock(ex) + map->m_lblk - ee_block);
err = ext4_ext_zeroout(inode, &zero_ex);
if (err)
goto out;
split_map.m_lblk = map->m_lblk;
split_map.m_len = allocated;
} else if (map->m_lblk - ee_block + map->m_len < max_zeroout) {
/* case 2 */
if (map->m_lblk != ee_block) {
zero_ex.ee_block = ex->ee_block;
zero_ex.ee_len = cpu_to_le16(map->m_lblk -
ee_block);
ext4_ext_store_pblock(&zero_ex,
ext4_ext_pblock(ex));
err = ext4_ext_zeroout(inode, &zero_ex);
if (err)
goto out;
}
split_map.m_lblk = ee_block;
split_map.m_len = map->m_lblk - ee_block + map->m_len;
allocated = map->m_len;
}
}
allocated = ext4_split_extent(handle, inode, path,
ext4: introduce reserved space Currently in ENOSPC condition when writing into unwritten space, or punching a hole, we might need to split the extent and grow extent tree. However since we can not allocate any new metadata blocks we'll have to zero out unwritten part of extent or punched out part of extent, or in the worst case return ENOSPC even though use actually does not allocate any space. Also in delalloc path we do reserve metadata and data blocks for the time we're going to write out, however metadata block reservation is very tricky especially since we expect that logical connectivity implies physical connectivity, however that might not be the case and hence we might end up allocating more metadata blocks than previously reserved. So in future, metadata reservation checks should be removed since we can not assure that we do not under reserve. And this is where reserved space comes into the picture. When mounting the file system we slice off a little bit of the file system space (2% or 4096 clusters, whichever is smaller) which can be then used for the cases mentioned above to prevent costly zeroout, or unexpected ENOSPC. The number of reserved clusters can be set via sysfs, however it can never be bigger than number of free clusters in the file system. Note that this patch fixes the failure of xfstest 274 as expected. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
2013-04-10 10:11:22 +08:00
&split_map, split_flag, flags);
if (allocated < 0)
err = allocated;
out:
/* If we have gotten a failure, don't zero out status tree */
if (!err)
err = ext4_es_zeroout(inode, &zero_ex);
return err ? err : allocated;
}
/*
* This function is called by ext4_ext_map_blocks() from
* ext4_get_blocks_dio_write() when DIO to write
* to an uninitialized extent.
*
* Writing to an uninitialized extent may result in splitting the uninitialized
* extent into multiple initialized/uninitialized extents (up to three)
* There are three possibilities:
* a> There is no split required: Entire extent should be uninitialized
* b> Splits in two extents: Write is happening at either end of the extent
* c> Splits in three extents: Somone is writing in middle of the extent
*
* One of more index blocks maybe needed if the extent tree grow after
* the uninitialized extent split. To prevent ENOSPC occur at the IO
* complete, we need to split the uninitialized extent before DIO submit
* the IO. The uninitialized extent called at this time will be split
* into three uninitialized extent(at most). After IO complete, the part
* being filled will be convert to initialized by the end_io callback function
* via ext4_convert_unwritten_extents().
*
* Returns the size of uninitialized extent to be written on success.
*/
static int ext4_split_unwritten_extents(handle_t *handle,
struct inode *inode,
struct ext4_map_blocks *map,
struct ext4_ext_path *path,
int flags)
{
ext4_lblk_t eof_block;
ext4_lblk_t ee_block;
struct ext4_extent *ex;
unsigned int ee_len;
int split_flag = 0, depth;
ext_debug("ext4_split_unwritten_extents: inode %lu, logical"
"block %llu, max_blocks %u\n", inode->i_ino,
(unsigned long long)map->m_lblk, map->m_len);
eof_block = (inode->i_size + inode->i_sb->s_blocksize - 1) >>
inode->i_sb->s_blocksize_bits;
if (eof_block < map->m_lblk + map->m_len)
eof_block = map->m_lblk + map->m_len;
/*
* It is safe to convert extent to initialized via explicit
* zeroout only if extent is fully insde i_size or new_size.
*/
depth = ext_depth(inode);
ex = path[depth].p_ext;
ee_block = le32_to_cpu(ex->ee_block);
ee_len = ext4_ext_get_actual_len(ex);
split_flag |= ee_block + ee_len <= eof_block ? EXT4_EXT_MAY_ZEROOUT : 0;
split_flag |= EXT4_EXT_MARK_UNINIT2;
if (flags & EXT4_GET_BLOCKS_CONVERT)
split_flag |= EXT4_EXT_DATA_VALID2;
flags |= EXT4_GET_BLOCKS_PRE_IO;
return ext4_split_extent(handle, inode, path, map, split_flag, flags);
}
static int ext4_convert_unwritten_extents_endio(handle_t *handle,
struct inode *inode,
struct ext4_map_blocks *map,
struct ext4_ext_path *path)
{
struct ext4_extent *ex;
ext4_lblk_t ee_block;
unsigned int ee_len;
int depth;
int err = 0;
depth = ext_depth(inode);
ex = path[depth].p_ext;
ee_block = le32_to_cpu(ex->ee_block);
ee_len = ext4_ext_get_actual_len(ex);
ext_debug("ext4_convert_unwritten_extents_endio: inode %lu, logical"
"block %llu, max_blocks %u\n", inode->i_ino,
(unsigned long long)ee_block, ee_len);
/* If extent is larger than requested it is a clear sign that we still
* have some extent state machine issues left. So extent_split is still
* required.
* TODO: Once all related issues will be fixed this situation should be
* illegal.
*/
if (ee_block != map->m_lblk || ee_len > map->m_len) {
#ifdef EXT4_DEBUG
ext4_warning("Inode (%ld) finished: extent logical block %llu,"
" len %u; IO logical block %llu, len %u\n",
inode->i_ino, (unsigned long long)ee_block, ee_len,
(unsigned long long)map->m_lblk, map->m_len);
#endif
err = ext4_split_unwritten_extents(handle, inode, map, path,
EXT4_GET_BLOCKS_CONVERT);
if (err < 0)
goto out;
ext4_ext_drop_refs(path);
path = ext4_ext_find_extent(inode, map->m_lblk, path);
if (IS_ERR(path)) {
err = PTR_ERR(path);
goto out;
}
depth = ext_depth(inode);
ex = path[depth].p_ext;
}
err = ext4_ext_get_access(handle, inode, path + depth);
if (err)
goto out;
/* first mark the extent as initialized */
ext4_ext_mark_initialized(ex);
/* note: ext4_ext_correct_indexes() isn't needed here because
* borders are not changed
*/
ext4_ext_try_to_merge(handle, inode, path, ex);
/* Mark modified extent as dirty */
err = ext4_ext_dirty(handle, inode, path + path->p_depth);
out:
ext4_ext_show_leaf(inode, path);
return err;
}
static void unmap_underlying_metadata_blocks(struct block_device *bdev,
sector_t block, int count)
{
int i;
for (i = 0; i < count; i++)
unmap_underlying_metadata(bdev, block + i);
}
/*
* Handle EOFBLOCKS_FL flag, clearing it if necessary
*/
static int check_eofblocks_fl(handle_t *handle, struct inode *inode,
ext4_lblk_t lblk,
struct ext4_ext_path *path,
unsigned int len)
{
int i, depth;
struct ext4_extent_header *eh;
struct ext4_extent *last_ex;
if (!ext4_test_inode_flag(inode, EXT4_INODE_EOFBLOCKS))
return 0;
depth = ext_depth(inode);
eh = path[depth].p_hdr;
/*
* We're going to remove EOFBLOCKS_FL entirely in future so we
* do not care for this case anymore. Simply remove the flag
* if there are no extents.
*/
if (unlikely(!eh->eh_entries))
goto out;
last_ex = EXT_LAST_EXTENT(eh);
/*
* We should clear the EOFBLOCKS_FL flag if we are writing the
* last block in the last extent in the file. We test this by
* first checking to see if the caller to
* ext4_ext_get_blocks() was interested in the last block (or
* a block beyond the last block) in the current extent. If
* this turns out to be false, we can bail out from this
* function immediately.
*/
if (lblk + len < le32_to_cpu(last_ex->ee_block) +
ext4_ext_get_actual_len(last_ex))
return 0;
/*
* If the caller does appear to be planning to write at or
* beyond the end of the current extent, we then test to see
* if the current extent is the last extent in the file, by
* checking to make sure it was reached via the rightmost node
* at each level of the tree.
*/
for (i = depth-1; i >= 0; i--)
if (path[i].p_idx != EXT_LAST_INDEX(path[i].p_hdr))
return 0;
out:
ext4_clear_inode_flag(inode, EXT4_INODE_EOFBLOCKS);
return ext4_mark_inode_dirty(handle, inode);
}
/**
* ext4_find_delalloc_range: find delayed allocated block in the given range.
*
* Return 1 if there is a delalloc block in the range, otherwise 0.
*/
int ext4_find_delalloc_range(struct inode *inode,
ext4_lblk_t lblk_start,
ext4_lblk_t lblk_end)
{
struct extent_status es;
ext4_es_find_delayed_extent_range(inode, lblk_start, lblk_end, &es);
if (es.es_len == 0)
return 0; /* there is no delay extent in this tree */
else if (es.es_lblk <= lblk_start &&
lblk_start < es.es_lblk + es.es_len)
return 1;
else if (lblk_start <= es.es_lblk && es.es_lblk <= lblk_end)
return 1;
else
return 0;
}
int ext4_find_delalloc_cluster(struct inode *inode, ext4_lblk_t lblk)
{
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
ext4_lblk_t lblk_start, lblk_end;
lblk_start = lblk & (~(sbi->s_cluster_ratio - 1));
lblk_end = lblk_start + sbi->s_cluster_ratio - 1;
return ext4_find_delalloc_range(inode, lblk_start, lblk_end);
}
/**
* Determines how many complete clusters (out of those specified by the 'map')
* are under delalloc and were reserved quota for.
* This function is called when we are writing out the blocks that were
* originally written with their allocation delayed, but then the space was
* allocated using fallocate() before the delayed allocation could be resolved.
* The cases to look for are:
* ('=' indicated delayed allocated blocks
* '-' indicates non-delayed allocated blocks)
* (a) partial clusters towards beginning and/or end outside of allocated range
* are not delalloc'ed.
* Ex:
* |----c---=|====c====|====c====|===-c----|
* |++++++ allocated ++++++|
* ==> 4 complete clusters in above example
*
* (b) partial cluster (outside of allocated range) towards either end is
* marked for delayed allocation. In this case, we will exclude that
* cluster.
* Ex:
* |----====c========|========c========|
* |++++++ allocated ++++++|
* ==> 1 complete clusters in above example
*
* Ex:
* |================c================|
* |++++++ allocated ++++++|
* ==> 0 complete clusters in above example
*
* The ext4_da_update_reserve_space will be called only if we
* determine here that there were some "entire" clusters that span
* this 'allocated' range.
* In the non-bigalloc case, this function will just end up returning num_blks
* without ever calling ext4_find_delalloc_range.
*/
static unsigned int
get_reserved_cluster_alloc(struct inode *inode, ext4_lblk_t lblk_start,
unsigned int num_blks)
{
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
ext4_lblk_t alloc_cluster_start, alloc_cluster_end;
ext4_lblk_t lblk_from, lblk_to, c_offset;
unsigned int allocated_clusters = 0;
alloc_cluster_start = EXT4_B2C(sbi, lblk_start);
alloc_cluster_end = EXT4_B2C(sbi, lblk_start + num_blks - 1);
/* max possible clusters for this allocation */
allocated_clusters = alloc_cluster_end - alloc_cluster_start + 1;
trace_ext4_get_reserved_cluster_alloc(inode, lblk_start, num_blks);
/* Check towards left side */
c_offset = lblk_start & (sbi->s_cluster_ratio - 1);
if (c_offset) {
lblk_from = lblk_start & (~(sbi->s_cluster_ratio - 1));
lblk_to = lblk_from + c_offset - 1;
if (ext4_find_delalloc_range(inode, lblk_from, lblk_to))
allocated_clusters--;
}
/* Now check towards right. */
c_offset = (lblk_start + num_blks) & (sbi->s_cluster_ratio - 1);
if (allocated_clusters && c_offset) {
lblk_from = lblk_start + num_blks;
lblk_to = lblk_from + (sbi->s_cluster_ratio - c_offset) - 1;
if (ext4_find_delalloc_range(inode, lblk_from, lblk_to))
allocated_clusters--;
}
return allocated_clusters;
}
static int
ext4_ext_handle_uninitialized_extents(handle_t *handle, struct inode *inode,
struct ext4_map_blocks *map,
struct ext4_ext_path *path, int flags,
unsigned int allocated, ext4_fsblk_t newblock)
{
int ret = 0;
int err = 0;
ext4_io_end_t *io = ext4_inode_aio(inode);
ext_debug("ext4_ext_handle_uninitialized_extents: inode %lu, logical "
"block %llu, max_blocks %u, flags %x, allocated %u\n",
inode->i_ino, (unsigned long long)map->m_lblk, map->m_len,
flags, allocated);
ext4_ext_show_leaf(inode, path);
ext4: introduce reserved space Currently in ENOSPC condition when writing into unwritten space, or punching a hole, we might need to split the extent and grow extent tree. However since we can not allocate any new metadata blocks we'll have to zero out unwritten part of extent or punched out part of extent, or in the worst case return ENOSPC even though use actually does not allocate any space. Also in delalloc path we do reserve metadata and data blocks for the time we're going to write out, however metadata block reservation is very tricky especially since we expect that logical connectivity implies physical connectivity, however that might not be the case and hence we might end up allocating more metadata blocks than previously reserved. So in future, metadata reservation checks should be removed since we can not assure that we do not under reserve. And this is where reserved space comes into the picture. When mounting the file system we slice off a little bit of the file system space (2% or 4096 clusters, whichever is smaller) which can be then used for the cases mentioned above to prevent costly zeroout, or unexpected ENOSPC. The number of reserved clusters can be set via sysfs, however it can never be bigger than number of free clusters in the file system. Note that this patch fixes the failure of xfstest 274 as expected. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
2013-04-10 10:11:22 +08:00
/*
* When writing into uninitialized space, we should not fail to
* allocate metadata blocks for the new extent block if needed.
*/
flags |= EXT4_GET_BLOCKS_METADATA_NOFAIL;
trace_ext4_ext_handle_uninitialized_extents(inode, map, flags,
allocated, newblock);
/* get_block() before submit the IO, split the extent */
if ((flags & EXT4_GET_BLOCKS_PRE_IO)) {
ret = ext4_split_unwritten_extents(handle, inode, map,
path, flags);
if (ret <= 0)
goto out;
/*
* Flag the inode(non aio case) or end_io struct (aio case)
* that this IO needs to conversion to written when IO is
* completed
*/
if (io)
ext4_set_io_unwritten_flag(inode, io);
else
ext4_set_inode_state(inode, EXT4_STATE_DIO_UNWRITTEN);
map->m_flags |= EXT4_MAP_UNWRITTEN;
if (ext4_should_dioread_nolock(inode))
map->m_flags |= EXT4_MAP_UNINIT;
goto out;
}
/* IO end_io complete, convert the filled extent to written */
if ((flags & EXT4_GET_BLOCKS_CONVERT)) {
ret = ext4_convert_unwritten_extents_endio(handle, inode, map,
path);
if (ret >= 0) {
ext4_update_inode_fsync_trans(handle, inode, 1);
err = check_eofblocks_fl(handle, inode, map->m_lblk,
path, map->m_len);
} else
err = ret;
map->m_flags |= EXT4_MAP_MAPPED;
if (allocated > map->m_len)
allocated = map->m_len;
map->m_len = allocated;
goto out2;
}
/* buffered IO case */
/*
* repeat fallocate creation request
* we already have an unwritten extent
*/
if (flags & EXT4_GET_BLOCKS_UNINIT_EXT) {
map->m_flags |= EXT4_MAP_UNWRITTEN;
goto map_out;
}
/* buffered READ or buffered write_begin() lookup */
if ((flags & EXT4_GET_BLOCKS_CREATE) == 0) {
/*
* We have blocks reserved already. We
* return allocated blocks so that delalloc
* won't do block reservation for us. But
* the buffer head will be unmapped so that
* a read from the block returns 0s.
*/
map->m_flags |= EXT4_MAP_UNWRITTEN;
goto out1;
}
/* buffered write, writepage time, convert*/
ext4: introduce reserved space Currently in ENOSPC condition when writing into unwritten space, or punching a hole, we might need to split the extent and grow extent tree. However since we can not allocate any new metadata blocks we'll have to zero out unwritten part of extent or punched out part of extent, or in the worst case return ENOSPC even though use actually does not allocate any space. Also in delalloc path we do reserve metadata and data blocks for the time we're going to write out, however metadata block reservation is very tricky especially since we expect that logical connectivity implies physical connectivity, however that might not be the case and hence we might end up allocating more metadata blocks than previously reserved. So in future, metadata reservation checks should be removed since we can not assure that we do not under reserve. And this is where reserved space comes into the picture. When mounting the file system we slice off a little bit of the file system space (2% or 4096 clusters, whichever is smaller) which can be then used for the cases mentioned above to prevent costly zeroout, or unexpected ENOSPC. The number of reserved clusters can be set via sysfs, however it can never be bigger than number of free clusters in the file system. Note that this patch fixes the failure of xfstest 274 as expected. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
2013-04-10 10:11:22 +08:00
ret = ext4_ext_convert_to_initialized(handle, inode, map, path, flags);
if (ret >= 0)
ext4_update_inode_fsync_trans(handle, inode, 1);
out:
if (ret <= 0) {
err = ret;
goto out2;
} else
allocated = ret;
map->m_flags |= EXT4_MAP_NEW;
/*
* if we allocated more blocks than requested
* we need to make sure we unmap the extra block
* allocated. The actual needed block will get
* unmapped later when we find the buffer_head marked
* new.
*/
if (allocated > map->m_len) {
unmap_underlying_metadata_blocks(inode->i_sb->s_bdev,
newblock + map->m_len,
allocated - map->m_len);
allocated = map->m_len;
}
map->m_len = allocated;
/*
* If we have done fallocate with the offset that is already
* delayed allocated, we would have block reservation
* and quota reservation done in the delayed write path.
* But fallocate would have already updated quota and block
* count for this offset. So cancel these reservation
*/
if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE) {
unsigned int reserved_clusters;
reserved_clusters = get_reserved_cluster_alloc(inode,
map->m_lblk, map->m_len);
if (reserved_clusters)
ext4_da_update_reserve_space(inode,
reserved_clusters,
0);
}
map_out:
map->m_flags |= EXT4_MAP_MAPPED;
if ((flags & EXT4_GET_BLOCKS_KEEP_SIZE) == 0) {
err = check_eofblocks_fl(handle, inode, map->m_lblk, path,
map->m_len);
if (err < 0)
goto out2;
}
out1:
if (allocated > map->m_len)
allocated = map->m_len;
ext4_ext_show_leaf(inode, path);
map->m_pblk = newblock;
map->m_len = allocated;
out2:
if (path) {
ext4_ext_drop_refs(path);
kfree(path);
}
return err ? err : allocated;
}
/*
* get_implied_cluster_alloc - check to see if the requested
* allocation (in the map structure) overlaps with a cluster already
* allocated in an extent.
* @sb The filesystem superblock structure
* @map The requested lblk->pblk mapping
* @ex The extent structure which might contain an implied
* cluster allocation
*
* This function is called by ext4_ext_map_blocks() after we failed to
* find blocks that were already in the inode's extent tree. Hence,
* we know that the beginning of the requested region cannot overlap
* the extent from the inode's extent tree. There are three cases we
* want to catch. The first is this case:
*
* |--- cluster # N--|
* |--- extent ---| |---- requested region ---|
* |==========|
*
* The second case that we need to test for is this one:
*
* |--------- cluster # N ----------------|
* |--- requested region --| |------- extent ----|
* |=======================|
*
* The third case is when the requested region lies between two extents
* within the same cluster:
* |------------- cluster # N-------------|
* |----- ex -----| |---- ex_right ----|
* |------ requested region ------|
* |================|
*
* In each of the above cases, we need to set the map->m_pblk and
* map->m_len so it corresponds to the return the extent labelled as
* "|====|" from cluster #N, since it is already in use for data in
* cluster EXT4_B2C(sbi, map->m_lblk). We will then return 1 to
* signal to ext4_ext_map_blocks() that map->m_pblk should be treated
* as a new "allocated" block region. Otherwise, we will return 0 and
* ext4_ext_map_blocks() will then allocate one or more new clusters
* by calling ext4_mb_new_blocks().
*/
static int get_implied_cluster_alloc(struct super_block *sb,
struct ext4_map_blocks *map,
struct ext4_extent *ex,
struct ext4_ext_path *path)
{
struct ext4_sb_info *sbi = EXT4_SB(sb);
ext4_lblk_t c_offset = map->m_lblk & (sbi->s_cluster_ratio-1);
ext4_lblk_t ex_cluster_start, ex_cluster_end;
ext4_lblk_t rr_cluster_start;
ext4_lblk_t ee_block = le32_to_cpu(ex->ee_block);
ext4_fsblk_t ee_start = ext4_ext_pblock(ex);
unsigned short ee_len = ext4_ext_get_actual_len(ex);
/* The extent passed in that we are trying to match */
ex_cluster_start = EXT4_B2C(sbi, ee_block);
ex_cluster_end = EXT4_B2C(sbi, ee_block + ee_len - 1);
/* The requested region passed into ext4_map_blocks() */
rr_cluster_start = EXT4_B2C(sbi, map->m_lblk);
if ((rr_cluster_start == ex_cluster_end) ||
(rr_cluster_start == ex_cluster_start)) {
if (rr_cluster_start == ex_cluster_end)
ee_start += ee_len - 1;
map->m_pblk = (ee_start & ~(sbi->s_cluster_ratio - 1)) +
c_offset;
map->m_len = min(map->m_len,
(unsigned) sbi->s_cluster_ratio - c_offset);
/*
* Check for and handle this case:
*
* |--------- cluster # N-------------|
* |------- extent ----|
* |--- requested region ---|
* |===========|
*/
if (map->m_lblk < ee_block)
map->m_len = min(map->m_len, ee_block - map->m_lblk);
/*
* Check for the case where there is already another allocated
* block to the right of 'ex' but before the end of the cluster.
*
* |------------- cluster # N-------------|
* |----- ex -----| |---- ex_right ----|
* |------ requested region ------|
* |================|
*/
if (map->m_lblk > ee_block) {
ext4_lblk_t next = ext4_ext_next_allocated_block(path);
map->m_len = min(map->m_len, next - map->m_lblk);
}
trace_ext4_get_implied_cluster_alloc_exit(sb, map, 1);
return 1;
}
trace_ext4_get_implied_cluster_alloc_exit(sb, map, 0);
return 0;
}
/*
* Block allocation/map/preallocation routine for extents based files
*
*
* Need to be called with
* down_read(&EXT4_I(inode)->i_data_sem) if not allocating file system block
* (ie, create is zero). Otherwise down_write(&EXT4_I(inode)->i_data_sem)
*
* return > 0, number of of blocks already mapped/allocated
* if create == 0 and these are pre-allocated blocks
* buffer head is unmapped
* otherwise blocks are mapped
*
* return = 0, if plain look up failed (blocks have not been allocated)
* buffer head is unmapped
*
* return < 0, error case.
*/
int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
struct ext4_map_blocks *map, int flags)
{
struct ext4_ext_path *path = NULL;
struct ext4_extent newex, *ex, *ex2;
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
ext4_fsblk_t newblock = 0;
int free_on_err = 0, err = 0, depth;
unsigned int allocated = 0, offset = 0;
unsigned int allocated_clusters = 0;
struct ext4_allocation_request ar;
ext4_io_end_t *io = ext4_inode_aio(inode);
ext4_lblk_t cluster_offset;
int set_unwritten = 0;
ext_debug("blocks %u/%u requested for inode %lu\n",
map->m_lblk, map->m_len, inode->i_ino);
trace_ext4_ext_map_blocks_enter(inode, map->m_lblk, map->m_len, flags);
/* find extent for this block */
path = ext4_ext_find_extent(inode, map->m_lblk, NULL);
if (IS_ERR(path)) {
err = PTR_ERR(path);
path = NULL;
goto out2;
}
depth = ext_depth(inode);
/*
* consistent leaf must not be empty;
* this situation is possible, though, _during_ tree modification;
* this is why assert can't be put in ext4_ext_find_extent()
*/
if (unlikely(path[depth].p_ext == NULL && depth != 0)) {
EXT4_ERROR_INODE(inode, "bad extent address "
"lblock: %lu, depth: %d pblock %lld",
(unsigned long) map->m_lblk, depth,
path[depth].p_block);
err = -EIO;
goto out2;
}
ex = path[depth].p_ext;
if (ex) {
ext4_lblk_t ee_block = le32_to_cpu(ex->ee_block);
ext4_fsblk_t ee_start = ext4_ext_pblock(ex);
unsigned short ee_len;
/*
* Uninitialized extents are treated as holes, except that
* we split out initialized portions during a write.
*/
ee_len = ext4_ext_get_actual_len(ex);
trace_ext4_ext_show_extent(inode, ee_block, ee_start, ee_len);
/* if found extent covers block, simply return it */
if (in_range(map->m_lblk, ee_block, ee_len)) {
newblock = map->m_lblk - ee_block + ee_start;
/* number of remaining blocks in the extent */
allocated = ee_len - (map->m_lblk - ee_block);
ext_debug("%u fit into %u:%d -> %llu\n", map->m_lblk,
ee_block, ee_len, newblock);
if (!ext4_ext_is_uninitialized(ex))
goto out;
allocated = ext4_ext_handle_uninitialized_extents(
handle, inode, map, path, flags,
allocated, newblock);
goto out3;
}
}
if ((sbi->s_cluster_ratio > 1) &&
ext4_find_delalloc_cluster(inode, map->m_lblk))
map->m_flags |= EXT4_MAP_FROM_CLUSTER;
/*
* requested block isn't allocated yet;
* we couldn't try to create block if create flag is zero
*/
if ((flags & EXT4_GET_BLOCKS_CREATE) == 0) {
/*
* put just found gap into cache to speed up
* subsequent requests
*/
if ((flags & EXT4_GET_BLOCKS_NO_PUT_HOLE) == 0)
ext4_ext_put_gap_in_cache(inode, path, map->m_lblk);
goto out2;
}
/*
* Okay, we need to do block allocation.
*/
map->m_flags &= ~EXT4_MAP_FROM_CLUSTER;
newex.ee_block = cpu_to_le32(map->m_lblk);
cluster_offset = map->m_lblk & (sbi->s_cluster_ratio-1);
/*
* If we are doing bigalloc, check to see if the extent returned
* by ext4_ext_find_extent() implies a cluster we can use.
*/
if (cluster_offset && ex &&
get_implied_cluster_alloc(inode->i_sb, map, ex, path)) {
ar.len = allocated = map->m_len;
newblock = map->m_pblk;
map->m_flags |= EXT4_MAP_FROM_CLUSTER;
goto got_allocated_blocks;
}
/* find neighbour allocated blocks */
ar.lleft = map->m_lblk;
err = ext4_ext_search_left(inode, path, &ar.lleft, &ar.pleft);
if (err)
goto out2;
ar.lright = map->m_lblk;
ex2 = NULL;
err = ext4_ext_search_right(inode, path, &ar.lright, &ar.pright, &ex2);
if (err)
goto out2;
/* Check if the extent after searching to the right implies a
* cluster we can use. */
if ((sbi->s_cluster_ratio > 1) && ex2 &&
get_implied_cluster_alloc(inode->i_sb, map, ex2, path)) {
ar.len = allocated = map->m_len;
newblock = map->m_pblk;
map->m_flags |= EXT4_MAP_FROM_CLUSTER;
goto got_allocated_blocks;
}
/*
* See if request is beyond maximum number of blocks we can have in
* a single extent. For an initialized extent this limit is
* EXT_INIT_MAX_LEN and for an uninitialized extent this limit is
* EXT_UNINIT_MAX_LEN.
*/
if (map->m_len > EXT_INIT_MAX_LEN &&
!(flags & EXT4_GET_BLOCKS_UNINIT_EXT))
map->m_len = EXT_INIT_MAX_LEN;
else if (map->m_len > EXT_UNINIT_MAX_LEN &&
(flags & EXT4_GET_BLOCKS_UNINIT_EXT))
map->m_len = EXT_UNINIT_MAX_LEN;
/* Check if we can really insert (m_lblk)::(m_lblk + m_len) extent */
newex.ee_len = cpu_to_le16(map->m_len);
err = ext4_ext_check_overlap(sbi, inode, &newex, path);
if (err)
allocated = ext4_ext_get_actual_len(&newex);
else
allocated = map->m_len;
/* allocate new block */
ar.inode = inode;
ar.goal = ext4_ext_find_goal(inode, path, map->m_lblk);
ar.logical = map->m_lblk;
/*
* We calculate the offset from the beginning of the cluster
* for the logical block number, since when we allocate a
* physical cluster, the physical block should start at the
* same offset from the beginning of the cluster. This is
* needed so that future calls to get_implied_cluster_alloc()
* work correctly.
*/
offset = map->m_lblk & (sbi->s_cluster_ratio - 1);
ar.len = EXT4_NUM_B2C(sbi, offset+allocated);
ar.goal -= offset;
ar.logical -= offset;
if (S_ISREG(inode->i_mode))
ar.flags = EXT4_MB_HINT_DATA;
else
/* disable in-core preallocation for non-regular files */
ar.flags = 0;
if (flags & EXT4_GET_BLOCKS_NO_NORMALIZE)
ar.flags |= EXT4_MB_HINT_NOPREALLOC;
newblock = ext4_mb_new_blocks(handle, &ar, &err);
if (!newblock)
goto out2;
ext_debug("allocate new block: goal %llu, found %llu/%u\n",
ar.goal, newblock, allocated);
free_on_err = 1;
allocated_clusters = ar.len;
ar.len = EXT4_C2B(sbi, ar.len) - offset;
if (ar.len > allocated)
ar.len = allocated;
got_allocated_blocks:
/* try to insert new extent into found leaf and return */
ext4_ext_store_pblock(&newex, newblock + offset);
newex.ee_len = cpu_to_le16(ar.len);
/* Mark uninitialized */
if (flags & EXT4_GET_BLOCKS_UNINIT_EXT){
ext4_ext_mark_uninitialized(&newex);
map->m_flags |= EXT4_MAP_UNWRITTEN;
/*
* io_end structure was created for every IO write to an
* uninitialized extent. To avoid unnecessary conversion,
* here we flag the IO that really needs the conversion.
* For non asycn direct IO case, flag the inode state
* that we need to perform conversion when IO is done.
*/
if ((flags & EXT4_GET_BLOCKS_PRE_IO))
set_unwritten = 1;
if (ext4_should_dioread_nolock(inode))
map->m_flags |= EXT4_MAP_UNINIT;
}
err = 0;
if ((flags & EXT4_GET_BLOCKS_KEEP_SIZE) == 0)
err = check_eofblocks_fl(handle, inode, map->m_lblk,
path, ar.len);
if (!err)
err = ext4_ext_insert_extent(handle, inode, path,
&newex, flags);
if (!err && set_unwritten) {
if (io)
ext4_set_io_unwritten_flag(inode, io);
else
ext4_set_inode_state(inode,
EXT4_STATE_DIO_UNWRITTEN);
}
if (err && free_on_err) {
int fb_flags = flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE ?
EXT4_FREE_BLOCKS_NO_QUOT_UPDATE : 0;
/* free data blocks we just allocated */
/* not a good idea to call discard here directly,
* but otherwise we'd need to call it every free() */
ext4_discard_preallocations(inode);
ext4_free_blocks(handle, inode, NULL, ext4_ext_pblock(&newex),
ext4_ext_get_actual_len(&newex), fb_flags);
goto out2;
}
/* previous routine could use block we allocated */
newblock = ext4_ext_pblock(&newex);
allocated = ext4_ext_get_actual_len(&newex);
if (allocated > map->m_len)
allocated = map->m_len;
map->m_flags |= EXT4_MAP_NEW;
/*
* Update reserved blocks/metadata blocks after successful
* block allocation which had been deferred till now.
*/
if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE) {
unsigned int reserved_clusters;
/*
* Check how many clusters we had reserved this allocated range
*/
reserved_clusters = get_reserved_cluster_alloc(inode,
map->m_lblk, allocated);
if (map->m_flags & EXT4_MAP_FROM_CLUSTER) {
if (reserved_clusters) {
/*
* We have clusters reserved for this range.
* But since we are not doing actual allocation
* and are simply using blocks from previously
* allocated cluster, we should release the
* reservation and not claim quota.
*/
ext4_da_update_reserve_space(inode,
reserved_clusters, 0);
}
} else {
BUG_ON(allocated_clusters < reserved_clusters);
if (reserved_clusters < allocated_clusters) {
ext4: attempt to fix race in bigalloc code path Currently, there exists a race between delayed allocated writes and the writeback when bigalloc feature is in use. The race was because we wanted to determine what blocks in a cluster are under delayed allocation and we were using buffer_delayed(bh) check for it. But, the writeback codepath clears this bit without any synchronization which resulted in a race and an ext4 warning similar to: EXT4-fs (ram1): ext4_da_update_reserve_space: ino 13, used 1 with only 0 reserved data blocks The race existed in two places. (1) between ext4_find_delalloc_range() and ext4_map_blocks() when called from writeback code path. (2) between ext4_find_delalloc_range() and ext4_da_get_block_prep() (where buffer_delayed(bh) is set. To fix (1), this patch introduces a new buffer_head state bit - BH_Da_Mapped. This bit is set under the protection of EXT4_I(inode)->i_data_sem when we have actually mapped the delayed allocated blocks during the writeout time. We can now reliably check for this bit inside ext4_find_delalloc_range() to determine whether the reservation for the blocks have already been claimed or not. To fix (2), it was necessary to set buffer_delay(bh) under the protection of i_data_sem. So, I extracted the very beginning of ext4_map_blocks into a new function - ext4_da_map_blocks() - and performed the required setting of bh_delay bit and the quota reservation under the protection of i_data_sem. These two fixes makes the checking of buffer_delay(bh) and buffer_da_mapped(bh) consistent, thus removing the race. Tested: I was able to reproduce the problem by running 'dd' and 'fsync' in parallel. Also, xfstests sometimes used to reproduce this race. After the fix both my test and xfstests were successful and no race (warning message) was observed. Google-Bug-Id: 4997027 Signed-off-by: Aditya Kali <adityakali@google.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-09-10 07:20:51 +08:00
struct ext4_inode_info *ei = EXT4_I(inode);
int reservation = allocated_clusters -
reserved_clusters;
/*
* It seems we claimed few clusters outside of
* the range of this allocation. We should give
* it back to the reservation pool. This can
* happen in the following case:
*
* * Suppose s_cluster_ratio is 4 (i.e., each
* cluster has 4 blocks. Thus, the clusters
* are [0-3],[4-7],[8-11]...
* * First comes delayed allocation write for
* logical blocks 10 & 11. Since there were no
* previous delayed allocated blocks in the
* range [8-11], we would reserve 1 cluster
* for this write.
* * Next comes write for logical blocks 3 to 8.
* In this case, we will reserve 2 clusters
* (for [0-3] and [4-7]; and not for [8-11] as
* that range has a delayed allocated blocks.
* Thus total reserved clusters now becomes 3.
* * Now, during the delayed allocation writeout
* time, we will first write blocks [3-8] and
* allocate 3 clusters for writing these
* blocks. Also, we would claim all these
* three clusters above.
* * Now when we come here to writeout the
* blocks [10-11], we would expect to claim
* the reservation of 1 cluster we had made
* (and we would claim it since there are no
* more delayed allocated blocks in the range
* [8-11]. But our reserved cluster count had
* already gone to 0.
*
* Thus, at the step 4 above when we determine
* that there are still some unwritten delayed
* allocated blocks outside of our current
* block range, we should increment the
* reserved clusters count so that when the
* remaining blocks finally gets written, we
* could claim them.
*/
ext4: attempt to fix race in bigalloc code path Currently, there exists a race between delayed allocated writes and the writeback when bigalloc feature is in use. The race was because we wanted to determine what blocks in a cluster are under delayed allocation and we were using buffer_delayed(bh) check for it. But, the writeback codepath clears this bit without any synchronization which resulted in a race and an ext4 warning similar to: EXT4-fs (ram1): ext4_da_update_reserve_space: ino 13, used 1 with only 0 reserved data blocks The race existed in two places. (1) between ext4_find_delalloc_range() and ext4_map_blocks() when called from writeback code path. (2) between ext4_find_delalloc_range() and ext4_da_get_block_prep() (where buffer_delayed(bh) is set. To fix (1), this patch introduces a new buffer_head state bit - BH_Da_Mapped. This bit is set under the protection of EXT4_I(inode)->i_data_sem when we have actually mapped the delayed allocated blocks during the writeout time. We can now reliably check for this bit inside ext4_find_delalloc_range() to determine whether the reservation for the blocks have already been claimed or not. To fix (2), it was necessary to set buffer_delay(bh) under the protection of i_data_sem. So, I extracted the very beginning of ext4_map_blocks into a new function - ext4_da_map_blocks() - and performed the required setting of bh_delay bit and the quota reservation under the protection of i_data_sem. These two fixes makes the checking of buffer_delay(bh) and buffer_da_mapped(bh) consistent, thus removing the race. Tested: I was able to reproduce the problem by running 'dd' and 'fsync' in parallel. Also, xfstests sometimes used to reproduce this race. After the fix both my test and xfstests were successful and no race (warning message) was observed. Google-Bug-Id: 4997027 Signed-off-by: Aditya Kali <adityakali@google.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-09-10 07:20:51 +08:00
dquot_reserve_block(inode,
EXT4_C2B(sbi, reservation));
spin_lock(&ei->i_block_reservation_lock);
ei->i_reserved_data_blocks += reservation;
spin_unlock(&ei->i_block_reservation_lock);
}
/*
* We will claim quota for all newly allocated blocks.
* We're updating the reserved space *after* the
* correction above so we do not accidentally free
* all the metadata reservation because we might
* actually need it later on.
*/
ext4_da_update_reserve_space(inode, allocated_clusters,
1);
}
}
/*
* Cache the extent and update transaction to commit on fdatasync only
* when it is _not_ an uninitialized extent.
*/
if ((flags & EXT4_GET_BLOCKS_UNINIT_EXT) == 0)
ext4_update_inode_fsync_trans(handle, inode, 1);
else
ext4_update_inode_fsync_trans(handle, inode, 0);
out:
if (allocated > map->m_len)
allocated = map->m_len;
ext4_ext_show_leaf(inode, path);
map->m_flags |= EXT4_MAP_MAPPED;
map->m_pblk = newblock;
map->m_len = allocated;
out2:
if (path) {
ext4_ext_drop_refs(path);
kfree(path);
}
out3:
trace_ext4_ext_map_blocks_exit(inode, map, err ? err : allocated);
return err ? err : allocated;
}
void ext4_ext_truncate(handle_t *handle, struct inode *inode)
{
struct super_block *sb = inode->i_sb;
ext4_lblk_t last_block;
int err = 0;
/*
* TODO: optimization is possible here.
* Probably we need not scan at all,
* because page truncation is enough.
*/
/* we have to know where to truncate from in crash case */
EXT4_I(inode)->i_disksize = inode->i_size;
ext4_mark_inode_dirty(handle, inode);
last_block = (inode->i_size + sb->s_blocksize - 1)
>> EXT4_BLOCK_SIZE_BITS(sb);
err = ext4_es_remove_extent(inode, last_block,
EXT_MAX_BLOCKS - last_block);
ext4: rewrite punch hole to use ext4_ext_remove_space() This commit rewrites ext4 punch hole implementation to use ext4_ext_remove_space() instead of its home gown way of doing this via ext4_ext_map_blocks(). There are several reasons for changing this. Firstly it is quite non obvious that punching hole needs to ext4_ext_map_blocks() to punch a hole, especially given that this function should map blocks, not unmap it. It also required a lot of new code in ext4_ext_map_blocks(). Secondly the design of it is not very effective. The reason is that we are trying to punch out blocks in ext4_ext_punch_hole() in opposite direction than in ext4_ext_rm_leaf() which causes the ext4_ext_rm_leaf() to iterate through the whole tree from the end to the start to find the requested extent for every extent we are going to punch out. And finally the current implementation does not use the existing code, but bring a lot of new code, which is IMO unnecessary since there already is some infrastructure we can use. Specifically ext4_ext_remove_space(). This commit changes ext4_ext_remove_space() to accept 'end' parameter so we can not only truncate to the end of file, but also remove the space in the middle of the file (punch a hole). Moreover, because the last block to punch out, might be in the middle of the extent, we have to split the extent at 'end + 1' so ext4_ext_rm_leaf() can easily either remove the whole fist part of split extent, or change its size. ext4_ext_remove_space() is then used to actually remove the space (extents) from within the hole, instead of ext4_ext_map_blocks(). Note that this also fix the issue with punch hole, where we would forget to remove empty index blocks from the extent tree, resulting in double free block error and file system corruption. This is simply because we now use different code path, where this problem does not exist. This has been tested with fsx running for several days and xfstests, plus xfstest #251 with '-o discard' run on the loop image (which converts discard requestes into punch hole to the backing file). All of it on 1K and 4K file system block size. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-03-20 11:03:19 +08:00
err = ext4_ext_remove_space(inode, last_block, EXT_MAX_BLOCKS - 1);
}
static void ext4_falloc_update_inode(struct inode *inode,
int mode, loff_t new_size, int update_ctime)
{
struct timespec now;
if (update_ctime) {
now = current_fs_time(inode->i_sb);
if (!timespec_equal(&inode->i_ctime, &now))
inode->i_ctime = now;
}
/*
* Update only when preallocation was requested beyond
* the file size.
*/
if (!(mode & FALLOC_FL_KEEP_SIZE)) {
if (new_size > i_size_read(inode))
i_size_write(inode, new_size);
if (new_size > EXT4_I(inode)->i_disksize)
ext4_update_i_disksize(inode, new_size);
} else {
/*
* Mark that we allocate beyond EOF so the subsequent truncate
* can proceed even if the new size is the same as i_size.
*/
if (new_size > i_size_read(inode))
ext4_set_inode_flag(inode, EXT4_INODE_EOFBLOCKS);
}
}
/*
* preallocate space for a file. This implements ext4's fallocate file
* operation, which gets called from sys_fallocate system call.
* For block-mapped files, posix_fallocate should fall back to the method
* of writing zeroes to the required new blocks (the same behavior which is
* expected for file systems which do not support fallocate() system call).
*/
long ext4_fallocate(struct file *file, int mode, loff_t offset, loff_t len)
{
struct inode *inode = file_inode(file);
handle_t *handle;
loff_t new_size;
unsigned int max_blocks;
int ret = 0;
int ret2 = 0;
int retries = 0;
int flags;
struct ext4_map_blocks map;
unsigned int credits, blkbits = inode->i_blkbits;
/* Return error if mode is not supported */
if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE))
return -EOPNOTSUPP;
if (mode & FALLOC_FL_PUNCH_HOLE)
return ext4_punch_hole(file, offset, len);
ret = ext4_convert_inline_data(inode);
if (ret)
return ret;
/*
* currently supporting (pre)allocate mode for extent-based
* files _only_
*/
if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
return -EOPNOTSUPP;
trace_ext4_fallocate_enter(inode, offset, len, mode);
map.m_lblk = offset >> blkbits;
/*
* We can't just convert len to max_blocks because
* If blocksize = 4096 offset = 3072 and len = 2048
*/
max_blocks = (EXT4_BLOCK_ALIGN(len + offset, blkbits) >> blkbits)
- map.m_lblk;
/*
* credits to insert 1 extent into extent tree
*/
credits = ext4_chunk_trans_blocks(inode, max_blocks);
mutex_lock(&inode->i_mutex);
ret = inode_newsize_ok(inode, (len + offset));
if (ret) {
mutex_unlock(&inode->i_mutex);
trace_ext4_fallocate_exit(inode, offset, max_blocks, ret);
return ret;
}
flags = EXT4_GET_BLOCKS_CREATE_UNINIT_EXT;
if (mode & FALLOC_FL_KEEP_SIZE)
flags |= EXT4_GET_BLOCKS_KEEP_SIZE;
/*
* Don't normalize the request if it can fit in one extent so
* that it doesn't get unnecessarily split into multiple
* extents.
*/
if (len <= EXT_UNINIT_MAX_LEN << blkbits)
flags |= EXT4_GET_BLOCKS_NO_NORMALIZE;
retry:
while (ret >= 0 && ret < max_blocks) {
map.m_lblk = map.m_lblk + ret;
map.m_len = max_blocks = max_blocks - ret;
handle = ext4_journal_start(inode, EXT4_HT_MAP_BLOCKS,
credits);
if (IS_ERR(handle)) {
ret = PTR_ERR(handle);
break;
}
ret = ext4_map_blocks(handle, inode, &map, flags);
if (ret <= 0) {
#ifdef EXT4FS_DEBUG
ext4_warning(inode->i_sb,
"inode #%lu: block %u: len %u: "
"ext4_ext_map_blocks returned %d",
inode->i_ino, map.m_lblk,
map.m_len, ret);
#endif
ext4_mark_inode_dirty(handle, inode);
ret2 = ext4_journal_stop(handle);
break;
}
if ((map.m_lblk + ret) >= (EXT4_BLOCK_ALIGN(offset + len,
blkbits) >> blkbits))
new_size = offset + len;
else
new_size = ((loff_t) map.m_lblk + ret) << blkbits;
ext4_falloc_update_inode(inode, mode, new_size,
(map.m_flags & EXT4_MAP_NEW));
ext4_mark_inode_dirty(handle, inode);
if ((file->f_flags & O_SYNC) && ret >= max_blocks)
ext4_handle_sync(handle);
ret2 = ext4_journal_stop(handle);
if (ret2)
break;
}
if (ret == -ENOSPC &&
ext4_should_retry_alloc(inode->i_sb, &retries)) {
ret = 0;
goto retry;
}
mutex_unlock(&inode->i_mutex);
trace_ext4_fallocate_exit(inode, offset, max_blocks,
ret > 0 ? ret2 : ret);
return ret > 0 ? ret2 : ret;
}
/*
* This function convert a range of blocks to written extents
* The caller of this function will pass the start offset and the size.
* all unwritten extents within this range will be converted to
* written extents.
*
* This function is called from the direct IO end io call back
* function, to convert the fallocated extents after IO is completed.
* Returns 0 on success.
*/
int ext4_convert_unwritten_extents(struct inode *inode, loff_t offset,
ssize_t len)
{
handle_t *handle;
unsigned int max_blocks;
int ret = 0;
int ret2 = 0;
struct ext4_map_blocks map;
unsigned int credits, blkbits = inode->i_blkbits;
map.m_lblk = offset >> blkbits;
/*
* We can't just convert len to max_blocks because
* If blocksize = 4096 offset = 3072 and len = 2048
*/
max_blocks = ((EXT4_BLOCK_ALIGN(len + offset, blkbits) >> blkbits) -
map.m_lblk);
/*
* credits to insert 1 extent into extent tree
*/
credits = ext4_chunk_trans_blocks(inode, max_blocks);
while (ret >= 0 && ret < max_blocks) {
map.m_lblk += ret;
map.m_len = (max_blocks -= ret);
handle = ext4_journal_start(inode, EXT4_HT_MAP_BLOCKS, credits);
if (IS_ERR(handle)) {
ret = PTR_ERR(handle);
break;
}
ret = ext4_map_blocks(handle, inode, &map,
EXT4_GET_BLOCKS_IO_CONVERT_EXT);
if (ret <= 0)
ext4_warning(inode->i_sb,
"inode #%lu: block %u: len %u: "
"ext4_ext_map_blocks returned %d",
inode->i_ino, map.m_lblk,
map.m_len, ret);
ext4_mark_inode_dirty(handle, inode);
ret2 = ext4_journal_stop(handle);
if (ret <= 0 || ret2 )
break;
}
return ret > 0 ? ret2 : ret;
}
ext4: make FIEMAP and delayed allocation play well together Fix the FIEMAP ioctl so that it returns all of the page ranges which are still subject to delayed allocation. We were missing some cases if the file was sparse. Reported by Chris Mason <chris.mason@oracle.com>: >We've had reports on btrfs that cp is giving us files full of zeros >instead of actually copying them. It was tracked down to a bug with >the btrfs fiemap implementation where it was returning holes for >delalloc ranges. > >Newer versions of cp are trusting fiemap to tell it where the holes >are, which does seem like a pretty neat trick. > >I decided to give xfs and ext4 a shot with a few tests cases too, xfs >passed with all the ones btrfs was getting wrong, and ext4 got the basic >delalloc case right. >$ mkfs.ext4 /dev/xxx >$ mount /dev/xxx /mnt >$ dd if=/dev/zero of=/mnt/foo bs=1M count=1 >$ fiemap-test foo >ext: 0 logical: [ 0.. 255] phys: 0.. 255 >flags: 0x007 tot: 256 > >Horray! But once we throw a hole in, things go bad: >$ mkfs.ext4 /dev/xxx >$ mount /dev/xxx /mnt >$ dd if=/dev/zero of=/mnt/foo bs=1M count=1 seek=1 >$ fiemap-test foo >< no output > > >We've got a delalloc extent after the hole and ext4 fiemap didn't find >it. If I run sync to kick the delalloc out: >$sync >$ fiemap-test foo >ext: 0 logical: [ 256.. 511] phys: 34048.. 34303 >flags: 0x001 tot: 256 > >fiemap-test is sitting in my /usr/local/bin, and I have no idea how it >got there. It's full of pretty comments so I know it isn't mine, but >you can grab it here: > >http://oss.oracle.com/~mason/fiemap-test.c > >xfsqa has a fiemap program too. After Fix, test results are as follows: ext: 0 logical: [ 256.. 511] phys: 0.. 255 flags: 0x007 tot: 256 ext: 0 logical: [ 256.. 511] phys: 33280.. 33535 flags: 0x001 tot: 256 $ mkfs.ext4 /dev/xxx $ mount /dev/xxx /mnt $ dd if=/dev/zero of=/mnt/foo bs=1M count=1 seek=1 $ sync $ dd if=/dev/zero of=/mnt/foo bs=1M count=1 seek=3 $ dd if=/dev/zero of=/mnt/foo bs=1M count=1 seek=5 $ fiemap-test foo ext: 0 logical: [ 256.. 511] phys: 33280.. 33535 flags: 0x000 tot: 256 ext: 1 logical: [ 768.. 1023] phys: 0.. 255 flags: 0x006 tot: 256 ext: 2 logical: [ 1280.. 1535] phys: 0.. 255 flags: 0x007 tot: 256 Tested-by: Eric Sandeen <sandeen@redhat.com> Reviewed-by: Andreas Dilger <adilger@dilger.ca> Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-02-28 06:25:47 +08:00
/*
* If newes is not existing extent (newes->ec_pblk equals zero) find
* delayed extent at start of newes and update newes accordingly and
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
* return start of the next delayed extent.
*
* If newes is existing extent (newes->ec_pblk is not equal zero)
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
* return start of next delayed extent or EXT_MAX_BLOCKS if no delayed
* extent found. Leave newes unmodified.
*/
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
static int ext4_find_delayed_extent(struct inode *inode,
struct extent_status *newes)
{
struct extent_status es;
ext4_lblk_t block, next_del;
if (newes->es_pblk == 0) {
ext4_es_find_delayed_extent_range(inode, newes->es_lblk,
newes->es_lblk + newes->es_len - 1, &es);
ext4: make FIEMAP and delayed allocation play well together Fix the FIEMAP ioctl so that it returns all of the page ranges which are still subject to delayed allocation. We were missing some cases if the file was sparse. Reported by Chris Mason <chris.mason@oracle.com>: >We've had reports on btrfs that cp is giving us files full of zeros >instead of actually copying them. It was tracked down to a bug with >the btrfs fiemap implementation where it was returning holes for >delalloc ranges. > >Newer versions of cp are trusting fiemap to tell it where the holes >are, which does seem like a pretty neat trick. > >I decided to give xfs and ext4 a shot with a few tests cases too, xfs >passed with all the ones btrfs was getting wrong, and ext4 got the basic >delalloc case right. >$ mkfs.ext4 /dev/xxx >$ mount /dev/xxx /mnt >$ dd if=/dev/zero of=/mnt/foo bs=1M count=1 >$ fiemap-test foo >ext: 0 logical: [ 0.. 255] phys: 0.. 255 >flags: 0x007 tot: 256 > >Horray! But once we throw a hole in, things go bad: >$ mkfs.ext4 /dev/xxx >$ mount /dev/xxx /mnt >$ dd if=/dev/zero of=/mnt/foo bs=1M count=1 seek=1 >$ fiemap-test foo >< no output > > >We've got a delalloc extent after the hole and ext4 fiemap didn't find >it. If I run sync to kick the delalloc out: >$sync >$ fiemap-test foo >ext: 0 logical: [ 256.. 511] phys: 34048.. 34303 >flags: 0x001 tot: 256 > >fiemap-test is sitting in my /usr/local/bin, and I have no idea how it >got there. It's full of pretty comments so I know it isn't mine, but >you can grab it here: > >http://oss.oracle.com/~mason/fiemap-test.c > >xfsqa has a fiemap program too. After Fix, test results are as follows: ext: 0 logical: [ 256.. 511] phys: 0.. 255 flags: 0x007 tot: 256 ext: 0 logical: [ 256.. 511] phys: 33280.. 33535 flags: 0x001 tot: 256 $ mkfs.ext4 /dev/xxx $ mount /dev/xxx /mnt $ dd if=/dev/zero of=/mnt/foo bs=1M count=1 seek=1 $ sync $ dd if=/dev/zero of=/mnt/foo bs=1M count=1 seek=3 $ dd if=/dev/zero of=/mnt/foo bs=1M count=1 seek=5 $ fiemap-test foo ext: 0 logical: [ 256.. 511] phys: 33280.. 33535 flags: 0x000 tot: 256 ext: 1 logical: [ 768.. 1023] phys: 0.. 255 flags: 0x006 tot: 256 ext: 2 logical: [ 1280.. 1535] phys: 0.. 255 flags: 0x007 tot: 256 Tested-by: Eric Sandeen <sandeen@redhat.com> Reviewed-by: Andreas Dilger <adilger@dilger.ca> Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-02-28 06:25:47 +08:00
/*
* No extent in extent-tree contains block @newes->es_pblk,
ext4: make FIEMAP and delayed allocation play well together Fix the FIEMAP ioctl so that it returns all of the page ranges which are still subject to delayed allocation. We were missing some cases if the file was sparse. Reported by Chris Mason <chris.mason@oracle.com>: >We've had reports on btrfs that cp is giving us files full of zeros >instead of actually copying them. It was tracked down to a bug with >the btrfs fiemap implementation where it was returning holes for >delalloc ranges. > >Newer versions of cp are trusting fiemap to tell it where the holes >are, which does seem like a pretty neat trick. > >I decided to give xfs and ext4 a shot with a few tests cases too, xfs >passed with all the ones btrfs was getting wrong, and ext4 got the basic >delalloc case right. >$ mkfs.ext4 /dev/xxx >$ mount /dev/xxx /mnt >$ dd if=/dev/zero of=/mnt/foo bs=1M count=1 >$ fiemap-test foo >ext: 0 logical: [ 0.. 255] phys: 0.. 255 >flags: 0x007 tot: 256 > >Horray! But once we throw a hole in, things go bad: >$ mkfs.ext4 /dev/xxx >$ mount /dev/xxx /mnt >$ dd if=/dev/zero of=/mnt/foo bs=1M count=1 seek=1 >$ fiemap-test foo >< no output > > >We've got a delalloc extent after the hole and ext4 fiemap didn't find >it. If I run sync to kick the delalloc out: >$sync >$ fiemap-test foo >ext: 0 logical: [ 256.. 511] phys: 34048.. 34303 >flags: 0x001 tot: 256 > >fiemap-test is sitting in my /usr/local/bin, and I have no idea how it >got there. It's full of pretty comments so I know it isn't mine, but >you can grab it here: > >http://oss.oracle.com/~mason/fiemap-test.c > >xfsqa has a fiemap program too. After Fix, test results are as follows: ext: 0 logical: [ 256.. 511] phys: 0.. 255 flags: 0x007 tot: 256 ext: 0 logical: [ 256.. 511] phys: 33280.. 33535 flags: 0x001 tot: 256 $ mkfs.ext4 /dev/xxx $ mount /dev/xxx /mnt $ dd if=/dev/zero of=/mnt/foo bs=1M count=1 seek=1 $ sync $ dd if=/dev/zero of=/mnt/foo bs=1M count=1 seek=3 $ dd if=/dev/zero of=/mnt/foo bs=1M count=1 seek=5 $ fiemap-test foo ext: 0 logical: [ 256.. 511] phys: 33280.. 33535 flags: 0x000 tot: 256 ext: 1 logical: [ 768.. 1023] phys: 0.. 255 flags: 0x006 tot: 256 ext: 2 logical: [ 1280.. 1535] phys: 0.. 255 flags: 0x007 tot: 256 Tested-by: Eric Sandeen <sandeen@redhat.com> Reviewed-by: Andreas Dilger <adilger@dilger.ca> Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-02-28 06:25:47 +08:00
* then the block may stay in 1)a hole or 2)delayed-extent.
*/
if (es.es_len == 0)
/* A hole found. */
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
return 0;
if (es.es_lblk > newes->es_lblk) {
/* A hole found. */
newes->es_len = min(es.es_lblk - newes->es_lblk,
newes->es_len);
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
return 0;
}
ext4: make FIEMAP and delayed allocation play well together Fix the FIEMAP ioctl so that it returns all of the page ranges which are still subject to delayed allocation. We were missing some cases if the file was sparse. Reported by Chris Mason <chris.mason@oracle.com>: >We've had reports on btrfs that cp is giving us files full of zeros >instead of actually copying them. It was tracked down to a bug with >the btrfs fiemap implementation where it was returning holes for >delalloc ranges. > >Newer versions of cp are trusting fiemap to tell it where the holes >are, which does seem like a pretty neat trick. > >I decided to give xfs and ext4 a shot with a few tests cases too, xfs >passed with all the ones btrfs was getting wrong, and ext4 got the basic >delalloc case right. >$ mkfs.ext4 /dev/xxx >$ mount /dev/xxx /mnt >$ dd if=/dev/zero of=/mnt/foo bs=1M count=1 >$ fiemap-test foo >ext: 0 logical: [ 0.. 255] phys: 0.. 255 >flags: 0x007 tot: 256 > >Horray! But once we throw a hole in, things go bad: >$ mkfs.ext4 /dev/xxx >$ mount /dev/xxx /mnt >$ dd if=/dev/zero of=/mnt/foo bs=1M count=1 seek=1 >$ fiemap-test foo >< no output > > >We've got a delalloc extent after the hole and ext4 fiemap didn't find >it. If I run sync to kick the delalloc out: >$sync >$ fiemap-test foo >ext: 0 logical: [ 256.. 511] phys: 34048.. 34303 >flags: 0x001 tot: 256 > >fiemap-test is sitting in my /usr/local/bin, and I have no idea how it >got there. It's full of pretty comments so I know it isn't mine, but >you can grab it here: > >http://oss.oracle.com/~mason/fiemap-test.c > >xfsqa has a fiemap program too. After Fix, test results are as follows: ext: 0 logical: [ 256.. 511] phys: 0.. 255 flags: 0x007 tot: 256 ext: 0 logical: [ 256.. 511] phys: 33280.. 33535 flags: 0x001 tot: 256 $ mkfs.ext4 /dev/xxx $ mount /dev/xxx /mnt $ dd if=/dev/zero of=/mnt/foo bs=1M count=1 seek=1 $ sync $ dd if=/dev/zero of=/mnt/foo bs=1M count=1 seek=3 $ dd if=/dev/zero of=/mnt/foo bs=1M count=1 seek=5 $ fiemap-test foo ext: 0 logical: [ 256.. 511] phys: 33280.. 33535 flags: 0x000 tot: 256 ext: 1 logical: [ 768.. 1023] phys: 0.. 255 flags: 0x006 tot: 256 ext: 2 logical: [ 1280.. 1535] phys: 0.. 255 flags: 0x007 tot: 256 Tested-by: Eric Sandeen <sandeen@redhat.com> Reviewed-by: Andreas Dilger <adilger@dilger.ca> Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-02-28 06:25:47 +08:00
newes->es_len = es.es_lblk + es.es_len - newes->es_lblk;
}
block = newes->es_lblk + newes->es_len;
ext4_es_find_delayed_extent_range(inode, block, EXT_MAX_BLOCKS, &es);
if (es.es_len == 0)
next_del = EXT_MAX_BLOCKS;
else
next_del = es.es_lblk;
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
return next_del;
}
/* fiemap flags we can handle specified here */
#define EXT4_FIEMAP_FLAGS (FIEMAP_FLAG_SYNC|FIEMAP_FLAG_XATTR)
static int ext4_xattr_fiemap(struct inode *inode,
struct fiemap_extent_info *fieinfo)
{
__u64 physical = 0;
__u64 length;
__u32 flags = FIEMAP_EXTENT_LAST;
int blockbits = inode->i_sb->s_blocksize_bits;
int error = 0;
/* in-inode? */
if (ext4_test_inode_state(inode, EXT4_STATE_XATTR)) {
struct ext4_iloc iloc;
int offset; /* offset of xattr in inode */
error = ext4_get_inode_loc(inode, &iloc);
if (error)
return error;
physical = iloc.bh->b_blocknr << blockbits;
offset = EXT4_GOOD_OLD_INODE_SIZE +
EXT4_I(inode)->i_extra_isize;
physical += offset;
length = EXT4_SB(inode->i_sb)->s_inode_size - offset;
flags |= FIEMAP_EXTENT_DATA_INLINE;
brelse(iloc.bh);
} else { /* external block */
physical = EXT4_I(inode)->i_file_acl << blockbits;
length = inode->i_sb->s_blocksize;
}
if (physical)
error = fiemap_fill_next_extent(fieinfo, 0, physical,
length, flags);
return (error < 0 ? error : 0);
}
int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
__u64 start, __u64 len)
{
ext4_lblk_t start_blk;
int error = 0;
if (ext4_has_inline_data(inode)) {
int has_inline = 1;
error = ext4_inline_data_fiemap(inode, fieinfo, &has_inline);
if (has_inline)
return error;
}
/* fallback to generic here if not in extents fmt */
if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
return generic_block_fiemap(inode, fieinfo, start, len,
ext4_get_block);
if (fiemap_check_flags(fieinfo, EXT4_FIEMAP_FLAGS))
return -EBADR;
if (fieinfo->fi_flags & FIEMAP_FLAG_XATTR) {
error = ext4_xattr_fiemap(inode, fieinfo);
} else {
ext4_lblk_t len_blks;
__u64 last_blk;
start_blk = start >> inode->i_sb->s_blocksize_bits;
last_blk = (start + len - 1) >> inode->i_sb->s_blocksize_bits;
ext4: Fix max file size and logical block counting of extent format file Kazuya Mio reported that he was able to hit BUG_ON(next == lblock) in ext4_ext_put_gap_in_cache() while creating a sparse file in extent format and fill the tail of file up to its end. We will hit the BUG_ON when we write the last block (2^32-1) into the sparse file. The root cause of the problem lies in the fact that we specifically set s_maxbytes so that block at s_maxbytes fit into on-disk extent format, which is 32 bit long. However, we are not storing start and end block number, but rather start block number and length in blocks. It means that in order to cover extent from 0 to EXT_MAX_BLOCK we need EXT_MAX_BLOCK+1 to fit into len (because we counting block 0 as well) - and it does not. The only way to fix it without changing the meaning of the struct ext4_extent members is, as Kazuya Mio suggested, to lower s_maxbytes by one fs block so we can cover the whole extent we can get by the on-disk extent format. Also in many places EXT_MAX_BLOCK is used as length instead of maximum logical block number as the name suggests, it is all a bit messy. So this commit renames it to EXT_MAX_BLOCKS and change its usage in some places to actually be maximum number of blocks in the extent. The bug which this commit fixes can be reproduced as follows: dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-2)) sync dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-1)) Reported-by: Kazuya Mio <k-mio@sx.jp.nec.com> Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-06-06 12:05:17 +08:00
if (last_blk >= EXT_MAX_BLOCKS)
last_blk = EXT_MAX_BLOCKS-1;
len_blks = ((ext4_lblk_t) last_blk) - start_blk + 1;
/*
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
* Walk the extent tree gathering extent information
* and pushing extents back to the user.
*/
ext4: prevent race while walking extent tree for fiemap Currently ext4_ext_walk_space() only takes i_data_sem for read when searching for the extent at given block with ext4_ext_find_extent(). Then it drops the lock and the extent tree can be changed at will. However later on we're searching for the 'next' extent, but the extent tree might already have changed, so the information might not be accurate. In fact we can hit BUG_ON(end <= start) if the extent got inserted into the tree after the one we found and before the block we were searching for. This has been reproduced by running xfstests 225 in loop on s390x architecture, but theoretically we could hit this on any other architecture as well, but probably not as often. Moreover the extent currently in delayed allocation might be allocated after we search the extent tree and before we search extent status tree delayed buffers resulting in those delayed buffers being completely missed, even though completely written and allocated. We fix all those problems in several steps: 1. remove unnecessary callback indirection 2. rename functions ext4_ext_walk_space -> ext4_fill_fiemap_extents ext4_ext_fiemap_cb -> ext4_find_delayed_extent 3. move fiemap_fill_next_extent() into ext4_fill_fiemap_extents() 4. hold the i_data_sem for: ext4_ext_find_extent() ext4_ext_next_allocated_block() ext4_find_delayed_extent() 5. call fiemap_fill_next_extent after releasing the i_data_sem 6. move path reinitialization into the critical section. Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-11-29 01:32:26 +08:00
error = ext4_fill_fiemap_extents(inode, start_blk,
len_blks, fieinfo);
}
return error;
}