kernel_optimize_test/fs
Suzuki K P 87b4126f10 [PATCH] fix reiserfs bad path release panic
One of our test team hit a reiserfs_panic while running fsstress tests on
2.6.19-rc1.  The message looks like :

  REISERFS: panic(device Null superblock):
  reiserfs[5676]: assertion !(p->path_length != 1 ) failed at
  fs/reiserfs/stree.c:397:reiserfs_check_path: path not properly relsed.

The backtrace looked :

  kernel BUG in reiserfs_panic at fs/reiserfs/prints.c:361!
	.reiserfs_check_path+0x58/0x74
	.reiserfs_get_block+0x1444/0x1508
	.__block_prepare_write+0x1c8/0x558
	.block_prepare_write+0x34/0x64
	.reiserfs_prepare_write+0x118/0x1d0
	.generic_file_buffered_write+0x314/0x82c
	.__generic_file_aio_write_nolock+0x350/0x3e0
	.__generic_file_write_nolock+0x78/0xb0
	.generic_file_write+0x60/0xf0
	.reiserfs_file_write+0x198/0x2038
	.vfs_write+0xd0/0x1b4
	.sys_write+0x4c/0x8c
	syscall_exit+0x0/0x4

Upon debugging I found that the restart_transaction was not releasing
the path if the th->refcount was > 1.

/*static*/
int restart_transaction(struct reiserfs_transaction_handle *th,
                           			struct inode *inode, struct path *path)
{
	[...]

         /* we cannot restart while nested */
         if (th->t_refcount > 1) { <<- Path is not released in this case!
                 return 0;
         }

         pathrelse(path); <<- Path released here.
	[...]

This could happen in such a situation :

In reiserfs/inode.c: reiserfs_get_block() ::

      if (repeat == NO_DISK_SPACE || repeat == QUOTA_EXCEEDED) {
          /* restart the transaction to give the journal a chance to free
           ** some blocks.  releases the path, so we have to go back to
           ** research if we succeed on the second try
           */
          SB_JOURNAL(inode->i_sb)->j_next_async_flush = 1;

        -->>  retval = restart_transaction(th, inode, &path); <<--

  We are supposed to release the path, no matter we succeed or fail. But
if the th->refcount is > 1, the path is still valid. And,

          if (retval)
                   goto failure;
          repeat =
              _allocate_block(th, block, inode,
                             &allocated_block_nr, NULL, create);

If the above allocate_block fails with NO_DISK_SPACE or QUOTA_EXCEEDED,
we would have path which is not released.

         if (repeat != NO_DISK_SPACE && repeat != QUOTA_EXCEEDED) {
                   goto research;
         }
         if (repeat == QUOTA_EXCEEDED)
                   retval = -EDQUOT;
         else
                   retval = -ENOSPC;
         goto failure;
	[...]

       failure:
	[...]
         reiserfs_check_path(&path); << Panics here !

Attached here is a patch which could fix the issue.

fix reiserfs/inode.c : restart_transaction() to release the path in all
cases.

The restart_transaction() doesn't release the path when the the journal
handle has a refcount > 1.  This would trigger a reiserfs_panic() if we
encounter an -ENOSPC / -EDQUOT in reiserfs_get_block().

Signed-off-by: Suzuki K P <suzuki@in.ibm.com>
Cc: "Vladimir V. Saveliev" <vs@namesys.com>
Cc: <reiserfs-dev@namesys.com>
Cc: Jeff Mahoney <jeffm@suse.com>
Acked-by: Jan Kara <jack@suse.cz>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-07 08:39:32 -08:00
..
9p
adfs
affs
afs
autofs
autofs4
befs
bfs
cifs
coda
configfs
cramfs
debugfs
devpts
dlm
ecryptfs
efs
exportfs
ext2 [PATCH] ext2: fsid for statvfs 2006-12-07 08:39:31 -08:00
ext3 [PATCH] ext3: fsid for statvfs 2006-12-07 08:39:31 -08:00
ext4 [PATCH] ext4: fsid for statvfs 2006-12-07 08:39:31 -08:00
fat
freevxfs
fuse [PATCH] fuse: fix compile without CONFIG_BLOCK 2006-12-07 08:39:32 -08:00
gfs2
hfs
hfsplus
hostfs
hpfs
hppfs
hugetlbfs
isofs
jbd
jbd2
jffs
jffs2
jfs
lockd
minix
msdos
ncpfs
nfs
nfs_common
nfsd
nls
ntfs
ocfs2
openpromfs
partitions [PATCH] Fix check_partition routines 2006-12-07 08:39:30 -08:00
proc
qnx4
ramfs
reiserfs [PATCH] fix reiserfs bad path release panic 2006-12-07 08:39:32 -08:00
romfs
smbfs
sysfs
sysv
udf
ufs
vfat
xfs
aio.c
attr.c
bad_inode.c
binfmt_aout.c
binfmt_elf_fdpic.c
binfmt_elf.c
binfmt_em86.c
binfmt_flat.c
binfmt_misc.c
binfmt_script.c
binfmt_som.c
bio.c
block_dev.c
buffer.c
char_dev.c
compat_ioctl.c
compat.c
dcache.c
dcookies.c
direct-io.c
dnotify.c
dquot.c
drop_caches.c
eventpoll.c
exec.c
fcntl.c
fifo.c
file_table.c
file.c [PATCH] file: kill unnecessary timer in fdtable_defer 2006-12-07 08:39:32 -08:00
filesystems.c
fs-writeback.c
generic_acl.c
inode.c
inotify_user.c
inotify.c
internal.h
ioctl.c
ioprio.c
Kconfig
Kconfig.binfmt
libfs.c
locks.c
Makefile
mbcache.c
mpage.c
namei.c [PATCH] honour MNT_NOEXEC for access() 2006-12-07 08:39:30 -08:00
namespace.c
nfsctl.c
no-block.c
open.c
pipe.c
pnode.c
pnode.h
posix_acl.c
quota_v1.c
quota_v2.c
quota.c
read_write.c
read_write.h
readdir.c
select.c
seq_file.c
splice.c
stat.c
super.c
sync.c
utimes.c
xattr_acl.c
xattr.c