From: Nick Piggin <npiggin@kernel.dk>
To: linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org
Cc: Andrew Morton <akpm@linux-foundation.org>
Subject: [patch 5/8] fs: ext2 inode sync fix
Date: Sat, 18 Dec 2010 12:46:39 +1100 [thread overview]
Message-ID: <20101218015117.436300300@kernel.dk> (raw)
In-Reply-To: 20101218014634.943276411@kernel.dk
[-- Attachment #1: ext2-sync-fixes.patch --]
[-- Type: text/plain, Size: 11623 bytes --]
There is a big fuckup with inode metadata writeback (I suspect in a lot
of filesystems, but I've only dared to look at a couple as yet).
ext2 relies on ->write_inode being called from sync_inode_metadata in
fsync in order to sync the inode. However I_DIRTY_SYNC gets cleared
after a call to this guy, and it doesn't actually write back and wait
on the inode block unless it is called for sync. This means that write_inode
from background writeback can kill the inode dirty bits without the data
getting to disk. Fsync will subsequently miss it.
The fix is for ->write_inode to dirty the buffers/cache, and then ->fsync to
write back the dirty data. In the full filesystem sync case, buffercache
writeback in the generic code will write back the dirty data. Other
filesystems could use ->sync_fs for this.
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Index: linux-2.6/fs/ext2/inode.c
===================================================================
--- linux-2.6.orig/fs/ext2/inode.c 2010-12-11 20:57:04.000000000 +1100
+++ linux-2.6/fs/ext2/inode.c 2010-12-16 18:33:25.000000000 +1100
@@ -1211,7 +1211,7 @@ static int ext2_setsize(struct inode *in
return 0;
}
-static struct ext2_inode *ext2_get_inode(struct super_block *sb, ino_t ino,
+struct ext2_inode *ext2_get_inode(struct super_block *sb, ino_t ino,
struct buffer_head **p)
{
struct buffer_head * bh;
@@ -1505,16 +1505,8 @@ static int __ext2_write_inode(struct ino
} else for (n = 0; n < EXT2_N_BLOCKS; n++)
raw_inode->i_block[n] = ei->i_data[n];
mark_buffer_dirty(bh);
- if (do_sync) {
- sync_dirty_buffer(bh);
- if (buffer_req(bh) && !buffer_uptodate(bh)) {
- printk ("IO error syncing ext2 inode [%s:%08lx]\n",
- sb->s_id, (unsigned long) ino);
- err = -EIO;
- }
- }
- ei->i_state &= ~EXT2_STATE_NEW;
brelse (bh);
+ ei->i_state &= ~EXT2_STATE_NEW;
return err;
}
Index: linux-2.6/fs/ext2/file.c
===================================================================
--- linux-2.6.orig/fs/ext2/file.c 2010-12-16 18:28:58.000000000 +1100
+++ linux-2.6/fs/ext2/file.c 2010-12-16 18:33:25.000000000 +1100
@@ -21,6 +21,7 @@
#include <linux/time.h>
#include <linux/pagemap.h>
#include <linux/quotaops.h>
+#include <linux/buffer_head.h>
#include "ext2.h"
#include "xattr.h"
#include "acl.h"
@@ -43,16 +44,33 @@ static int ext2_release_file (struct ino
int ext2_fsync(struct file *file, int datasync)
{
int ret;
- struct super_block *sb = file->f_mapping->host->i_sb;
- struct address_space *mapping = sb->s_bdev->bd_inode->i_mapping;
+ struct inode *inode = file->f_mapping->host;
+ ino_t ino = inode->i_ino;
+ struct super_block *sb = inode->i_sb;
+ struct address_space *sb_mapping = sb->s_bdev->bd_inode->i_mapping;
+ struct buffer_head *bh;
+ struct ext2_inode *raw_inode;
ret = generic_file_fsync(file, datasync);
- if (ret == -EIO || test_and_clear_bit(AS_EIO, &mapping->flags)) {
+ if (ret == -EIO || test_and_clear_bit(AS_EIO, &sb_mapping->flags)) {
/* We don't really know where the IO error happened... */
ext2_error(sb, __func__,
"detected IO error when writing metadata buffers");
+ return -EIO;
+ }
+
+ raw_inode = ext2_get_inode(sb, ino, &bh);
+ if (IS_ERR(raw_inode))
+ return -EIO;
+
+ sync_dirty_buffer(bh);
+ if (buffer_req(bh) && !buffer_uptodate(bh)) {
+ printk ("IO error syncing ext2 inode [%s:%08lx]\n",
+ sb->s_id, (unsigned long) ino);
ret = -EIO;
}
+ brelse (bh);
+
return ret;
}
Index: linux-2.6/fs/ext2/ext2.h
===================================================================
--- linux-2.6.orig/fs/ext2/ext2.h 2010-12-11 20:57:04.000000000 +1100
+++ linux-2.6/fs/ext2/ext2.h 2010-12-16 18:33:25.000000000 +1100
@@ -124,6 +124,8 @@ extern int ext2_get_block(struct inode *
extern int ext2_setattr (struct dentry *, struct iattr *);
extern void ext2_set_inode_flags(struct inode *inode);
extern void ext2_get_inode_flags(struct ext2_inode_info *);
+extern struct ext2_inode *ext2_get_inode(struct super_block *sb, ino_t ino,
+ struct buffer_head **p);
extern int ext2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
u64 start, u64 len);
Index: linux-2.6/fs/adfs/inode.c
===================================================================
--- linux-2.6.orig/fs/adfs/inode.c 2010-12-11 20:57:04.000000000 +1100
+++ linux-2.6/fs/adfs/inode.c 2010-12-16 18:33:25.000000000 +1100
@@ -383,7 +383,7 @@ int adfs_write_inode(struct inode *inode
obj.attr = ADFS_I(inode)->attr;
obj.size = inode->i_size;
- ret = adfs_dir_update(sb, &obj, wbc->sync_mode == WB_SYNC_ALL);
+ ret = adfs_dir_update(sb, &obj, 1 /* XXX: fix fsync and use 'wbc->sync_mode == WB_SYNC_ALL' */);
unlock_kernel();
return ret;
}
Index: linux-2.6/fs/affs/file.c
===================================================================
--- linux-2.6.orig/fs/affs/file.c 2010-12-11 20:57:04.000000000 +1100
+++ linux-2.6/fs/affs/file.c 2010-12-16 18:33:25.000000000 +1100
@@ -931,6 +931,7 @@ int affs_file_fsync(struct file *filp, i
int ret, err;
ret = write_inode_now(inode, 0);
+ /* XXX: could just sync the buffer been dirtied by write_inode */
err = sync_blockdev(inode->i_sb->s_bdev);
if (!ret)
ret = err;
Index: linux-2.6/fs/bfs/inode.c
===================================================================
--- linux-2.6.orig/fs/bfs/inode.c 2010-12-16 18:29:02.000000000 +1100
+++ linux-2.6/fs/bfs/inode.c 2010-12-16 18:33:25.000000000 +1100
@@ -151,7 +151,7 @@ static int bfs_write_inode(struct inode
di->i_eoffset = cpu_to_le32(i_sblock * BFS_BSIZE + inode->i_size - 1);
mark_buffer_dirty(bh);
- if (wbc->sync_mode == WB_SYNC_ALL) {
+ if (1 /* XXX: fix fsync and use wbc->sync_mode == WB_SYNC_ALL */ ) {
sync_dirty_buffer(bh);
if (buffer_req(bh) && !buffer_uptodate(bh))
err = -EIO;
Index: linux-2.6/fs/exofs/inode.c
===================================================================
--- linux-2.6.orig/fs/exofs/inode.c 2010-12-11 20:57:04.000000000 +1100
+++ linux-2.6/fs/exofs/inode.c 2010-12-16 18:33:25.000000000 +1100
@@ -1273,7 +1273,7 @@ static int exofs_update_inode(struct ino
int exofs_write_inode(struct inode *inode, struct writeback_control *wbc)
{
- return exofs_update_inode(inode, wbc->sync_mode == WB_SYNC_ALL);
+ return exofs_update_inode(inode, 1 /* XXX: fix fsync and use wbc->sync_mode == WB_SYNC_ALL */ );
}
/*
Index: linux-2.6/fs/ext4/inode.c
===================================================================
--- linux-2.6.orig/fs/ext4/inode.c 2010-12-16 18:30:20.000000000 +1100
+++ linux-2.6/fs/ext4/inode.c 2010-12-16 18:33:25.000000000 +1100
@@ -5243,7 +5243,7 @@ int ext4_write_inode(struct inode *inode
err = __ext4_get_inode_loc(inode, &iloc, 0);
if (err)
return err;
- if (wbc->sync_mode == WB_SYNC_ALL)
+ if (1 /* XXX: fix fxync and use wbc->sync_mode == WB_SYNC_ALL */)
sync_dirty_buffer(iloc.bh);
if (buffer_req(iloc.bh) && !buffer_uptodate(iloc.bh)) {
EXT4_ERROR_INODE_BLOCK(inode, iloc.bh->b_blocknr,
Index: linux-2.6/fs/fat/inode.c
===================================================================
--- linux-2.6.orig/fs/fat/inode.c 2010-12-16 18:29:02.000000000 +1100
+++ linux-2.6/fs/fat/inode.c 2010-12-16 18:33:25.000000000 +1100
@@ -645,7 +645,7 @@ static int __fat_write_inode(struct inod
spin_unlock(&sbi->inode_hash_lock);
mark_buffer_dirty(bh);
err = 0;
- if (wait)
+ if (1 /* XXX: fix fsync and use wait */)
err = sync_dirty_buffer(bh);
brelse(bh);
return err;
Index: linux-2.6/fs/jfs/inode.c
===================================================================
--- linux-2.6.orig/fs/jfs/inode.c 2010-12-11 20:57:04.000000000 +1100
+++ linux-2.6/fs/jfs/inode.c 2010-12-16 18:33:25.000000000 +1100
@@ -123,7 +123,7 @@ int jfs_commit_inode(struct inode *inode
int jfs_write_inode(struct inode *inode, struct writeback_control *wbc)
{
- int wait = wbc->sync_mode == WB_SYNC_ALL;
+ int wait = 1; /* XXX fix fsync and use wbc->sync_mode == WB_SYNC_ALL; */
if (test_cflag(COMMIT_Nolink, inode))
return 0;
Index: linux-2.6/fs/minix/inode.c
===================================================================
--- linux-2.6.orig/fs/minix/inode.c 2010-12-16 18:29:02.000000000 +1100
+++ linux-2.6/fs/minix/inode.c 2010-12-16 18:33:25.000000000 +1100
@@ -576,7 +576,7 @@ static int minix_write_inode(struct inod
bh = V2_minix_update_inode(inode);
if (!bh)
return -EIO;
- if (wbc->sync_mode == WB_SYNC_ALL && buffer_dirty(bh)) {
+ if (1 /* XXX: fix fsync and use wbc->sync_mode == WB_SYNC_ALL */ && buffer_dirty(bh)) {
sync_dirty_buffer(bh);
if (buffer_req(bh) && !buffer_uptodate(bh)) {
printk("IO error syncing minix inode [%s:%08lx]\n",
Index: linux-2.6/fs/omfs/inode.c
===================================================================
--- linux-2.6.orig/fs/omfs/inode.c 2010-12-11 20:57:04.000000000 +1100
+++ linux-2.6/fs/omfs/inode.c 2010-12-16 18:33:25.000000000 +1100
@@ -169,7 +169,7 @@ static int __omfs_write_inode(struct ino
static int omfs_write_inode(struct inode *inode, struct writeback_control *wbc)
{
- return __omfs_write_inode(inode, wbc->sync_mode == WB_SYNC_ALL);
+ return __omfs_write_inode(inode, 1 /* XXX: fix fsync and use wbc->sync_mode == WB_SYNC_ALL */);
}
int omfs_sync_inode(struct inode *inode)
Index: linux-2.6/fs/reiserfs/inode.c
===================================================================
--- linux-2.6.orig/fs/reiserfs/inode.c 2010-12-11 20:57:04.000000000 +1100
+++ linux-2.6/fs/reiserfs/inode.c 2010-12-16 18:33:25.000000000 +1100
@@ -1635,6 +1635,8 @@ int reiserfs_write_inode(struct inode *i
** these cases are just when the system needs ram, not when the
** inode needs to reach disk for safety, and they can safely be
** ignored because the altered inode has already been logged.
+ ** XXX: is this really OK? The caller clears the inode dirty bit, so
+ ** a subsequent sync for integrity might never reach here.
*/
if (wbc->sync_mode == WB_SYNC_ALL && !(current->flags & PF_MEMALLOC)) {
reiserfs_write_lock(inode->i_sb);
Index: linux-2.6/fs/sysv/inode.c
===================================================================
--- linux-2.6.orig/fs/sysv/inode.c 2010-12-16 18:29:02.000000000 +1100
+++ linux-2.6/fs/sysv/inode.c 2010-12-16 18:33:25.000000000 +1100
@@ -286,7 +286,7 @@ static int __sysv_write_inode(struct ino
write3byte(sbi, (u8 *)&si->i_data[block],
&raw_inode->i_data[3*block]);
mark_buffer_dirty(bh);
- if (wait) {
+ if (1 /* XXX: fix fsync and use wait */) {
sync_dirty_buffer(bh);
if (buffer_req(bh) && !buffer_uptodate(bh)) {
printk ("IO error syncing sysv inode [%s:%08x]\n",
Index: linux-2.6/fs/udf/inode.c
===================================================================
--- linux-2.6.orig/fs/udf/inode.c 2010-12-11 20:57:04.000000000 +1100
+++ linux-2.6/fs/udf/inode.c 2010-12-16 18:33:25.000000000 +1100
@@ -1598,7 +1598,7 @@ static int udf_update_inode(struct inode
/* write the data blocks */
mark_buffer_dirty(bh);
- if (do_sync) {
+ if (1 /* XXX fix fsync and use do_sync */) {
sync_dirty_buffer(bh);
if (buffer_write_io_error(bh)) {
printk(KERN_WARNING "IO error syncing udf inode "
Index: linux-2.6/fs/ufs/inode.c
===================================================================
--- linux-2.6.orig/fs/ufs/inode.c 2010-12-11 20:57:04.000000000 +1100
+++ linux-2.6/fs/ufs/inode.c 2010-12-16 18:33:25.000000000 +1100
@@ -889,7 +889,7 @@ static int ufs_update_inode(struct inode
}
mark_buffer_dirty(bh);
- if (do_sync)
+ if (1 /* XXX: fix fsync and use do_sync */)
sync_dirty_buffer(bh);
brelse (bh);
next prev parent reply other threads:[~2010-12-18 1:46 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-12-18 1:46 [patch 0/8] Inode data integrity patches Nick Piggin
2010-12-18 1:46 ` [patch 1/8] fs: mark_inode_dirty barrier fix Nick Piggin
2010-12-18 1:46 ` [patch 2/8] fs: simple fsync race fix Nick Piggin
2010-12-18 1:46 ` [patch 3/8] fs: introduce inode writeback helpers Nick Piggin
2010-12-18 1:46 ` [patch 4/8] fs: preserve inode dirty bits on failed metadata writeback Nick Piggin
2010-12-18 1:46 ` Nick Piggin [this message]
2011-01-07 19:08 ` [patch 5/8] fs: ext2 inode sync fix Ted Ts'o
2010-12-18 1:46 ` [patch 6/8] fs: fsync optimisations Nick Piggin
2010-12-18 1:46 ` [patch 7/8] fs: fix or note I_DIRTY handling bugs in filesystems Nick Piggin
2010-12-29 15:01 ` Christoph Hellwig
2011-01-03 15:03 ` Steven Whitehouse
2011-01-03 16:58 ` Christoph Hellwig
2011-01-04 7:12 ` Nick Piggin
2011-01-04 14:22 ` Steven Whitehouse
2011-01-04 6:04 ` Nick Piggin
2011-01-04 6:39 ` Christoph Hellwig
2011-01-04 7:52 ` Nick Piggin
2011-01-04 9:13 ` Christoph Hellwig
2011-01-04 9:28 ` Nick Piggin
2010-12-18 1:46 ` [patch 8/8] fs: add i_op->sync_inode Nick Piggin
2010-12-29 15:12 ` Christoph Hellwig
2011-01-04 6:27 ` Nick Piggin
2011-01-04 6:57 ` Christoph Hellwig
2011-01-04 8:03 ` Nick Piggin
2011-01-04 8:31 ` Nick Piggin
2011-01-04 9:25 ` Christoph Hellwig
2011-01-04 9:52 ` Nick Piggin
2011-01-06 20:49 ` Christoph Hellwig
2011-01-07 4:48 ` Nick Piggin
2011-01-07 7:25 ` Christoph Hellwig
2011-01-11 3:44 ` Nick Piggin
2011-01-04 9:25 ` Christoph Hellwig
2011-01-04 9:49 ` Nick Piggin
2011-01-06 20:45 ` Christoph Hellwig
2011-01-07 4:47 ` Nick Piggin
2011-01-07 7:24 ` Christoph Hellwig
2011-01-07 7:29 ` Christoph Hellwig
2011-01-07 13:10 ` Christoph Hellwig
2011-01-07 18:30 ` Ted Ts'o
2011-01-07 18:32 ` Christoph Hellwig
2011-01-07 19:06 ` Ted Ts'o
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20101218015117.436300300@kernel.dk \
--to=npiggin@kernel.dk \
--cc=akpm@linux-foundation.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).