* [merged mm-hotfixes-stable] nilfs2-fix-infinite-loop-in-nilfs_mdt_get_block.patch removed from -mm tree
@ 2023-05-05 22:43 Andrew Morton
0 siblings, 0 replies; only message in thread
From: Andrew Morton @ 2023-05-05 22:43 UTC (permalink / raw)
To: mm-commits, stable, konishi.ryusuke, akpm
The quilt patch titled
Subject: nilfs2: fix infinite loop in nilfs_mdt_get_block()
has been removed from the -mm tree. Its filename was
nilfs2-fix-infinite-loop-in-nilfs_mdt_get_block.patch
This patch was dropped because it was merged into the mm-hotfixes-stable branch
of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
------------------------------------------------------
From: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Subject: nilfs2: fix infinite loop in nilfs_mdt_get_block()
Date: Mon, 1 May 2023 04:30:46 +0900
If the disk image that nilfs2 mounts is corrupted and a virtual block
address obtained by block lookup for a metadata file is invalid,
nilfs_bmap_lookup_at_level() may return the same internal return code as
-ENOENT, meaning the block does not exist in the metadata file.
This duplication of return codes confuses nilfs_mdt_get_block(), causing
it to read and create a metadata block indefinitely.
In particular, if this happens to the inode metadata file, ifile,
semaphore i_rwsem can be left held, causing task hangs in lock_mount.
Fix this issue by making nilfs_bmap_lookup_at_level() treat virtual block
address translation failures with -ENOENT as metadata corruption instead
of returning the error code.
Link: https://lkml.kernel.org/r/20230430193046.6769-1-konishi.ryusuke@gmail.com
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Tested-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Reported-by: syzbot+221d75710bde87fa0e97@syzkaller.appspotmail.com
Link: https://syzkaller.appspot.com/bug?extid=221d75710bde87fa0e97
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
fs/nilfs2/bmap.c | 16 ++++++++++++----
1 file changed, 12 insertions(+), 4 deletions(-)
--- a/fs/nilfs2/bmap.c~nilfs2-fix-infinite-loop-in-nilfs_mdt_get_block
+++ a/fs/nilfs2/bmap.c
@@ -67,20 +67,28 @@ int nilfs_bmap_lookup_at_level(struct ni
down_read(&bmap->b_sem);
ret = bmap->b_ops->bop_lookup(bmap, key, level, ptrp);
- if (ret < 0) {
- ret = nilfs_bmap_convert_error(bmap, __func__, ret);
+ if (ret < 0)
goto out;
- }
+
if (NILFS_BMAP_USE_VBN(bmap)) {
ret = nilfs_dat_translate(nilfs_bmap_get_dat(bmap), *ptrp,
&blocknr);
if (!ret)
*ptrp = blocknr;
+ else if (ret == -ENOENT) {
+ /*
+ * If there was no valid entry in DAT for the block
+ * address obtained by b_ops->bop_lookup, then pass
+ * internal code -EINVAL to nilfs_bmap_convert_error
+ * to treat it as metadata corruption.
+ */
+ ret = -EINVAL;
+ }
}
out:
up_read(&bmap->b_sem);
- return ret;
+ return nilfs_bmap_convert_error(bmap, __func__, ret);
}
int nilfs_bmap_lookup_contig(struct nilfs_bmap *bmap, __u64 key, __u64 *ptrp,
_
Patches currently in -mm which might be from konishi.ryusuke@gmail.com are
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2023-05-05 22:43 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-05-05 22:43 [merged mm-hotfixes-stable] nilfs2-fix-infinite-loop-in-nilfs_mdt_get_block.patch removed from -mm tree Andrew Morton
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox