On Thu, Aug 22, 2013 at 02:28:00PM -0400, Brian Foster wrote:
> Hi all,
>
> I hit an assert on a debug kernel while beating on some finobt work and
> eventually reproduced it on unmodified/TOT xfs/xfsprogs as of today. I
> hit it through a couple different paths, first while running fsstress on
> a CRC enabled filesystem (with otherwise default mkfs options):
>
> (These tests are running on a 4p, 4GB VM against a 100GB virtio disk,
> hosted on a single spindle desktop box).
>
> crc=1
> fsstress -z -fsymlink=1 -n99999999 -p4 -d /mnt/test
>
> XFS: Assertion failed: first <= last && last < BBTOB(bp->b_length),
Directory buffer overrun.
> [<ffffffffa031d549>] xfs_trans_log_buf+0x89/0x1b0 [xfs]
> [<ffffffffa02e7c1c>] xfs_da3_node_add+0x11c/0x210 [xfs]
> [<ffffffffa02ea703>] xfs_da3_node_split+0xc3/0x230 [xfs]
> [<ffffffffa02eaa18>] xfs_da3_split+0x1a8/0x410 [xfs]
> [<ffffffffa02f743f>] xfs_dir2_node_addname+0x47f/0xde0 [xfs]
During a split.
Easily reproduced with "seq 200000 | xargs touch" as Michael Semon
reported last week.
The fix demonstrates my concerns about modifying directory code -
the CRC changes missed a *fundamental* directory format definition,
and we've only just tripped over it....
> rm -rf /mnt/test
>
> XFS: Assertion failed: first <= last && last < BBTOB(bp->b_length),
Directory buffer overrun.
> [<ffffffffa032b549>] xfs_trans_log_buf+0x89/0x1b0 [xfs]
> [<ffffffffa02f61ff>] xfs_da3_node_unbalance+0xef/0x1d0 [xfs]
> [<ffffffffa02f98b0>] xfs_da3_join+0x240/0x290 [xfs]
> [<ffffffffa030659b>] xfs_dir2_node_removename+0x69b/0x8b0 [xfs]
During a merge. Not sure why that is happening on a v4 filesystem.
V5 filesystem, yes, due to the above bug but v4 should not be
affected.
Cheers,
Dave.
--
Dave Chinner
david@xxxxxxxxxxxxx
xfs: fix calculation of the number of node entries in a dir3 node
From: Dave Chinner <dchinner@xxxxxxxxxx>
The calculation doesn't take into account the size of the dir v3
header, so overestimates the hash entries in a node. This causes
directory buffer overruns when splitting and merging nodes.
Signed-off-by: Dave Chinner <dchinner@xxxxxxxxxx>
---
fs/xfs/xfs_da_btree.h | 11 +++++++++--
fs/xfs/xfs_dir2.c | 16 ++++++++++------
2 files changed, 19 insertions(+), 8 deletions(-)
diff --git a/fs/xfs/xfs_da_btree.h b/fs/xfs/xfs_da_btree.h
index 8cdc77b..b1f2679 100644
--- a/fs/xfs/xfs_da_btree.h
+++ b/fs/xfs/xfs_da_btree.h
@@ -133,12 +133,19 @@ extern void xfs_da3_node_hdr_to_disk(struct xfs_da_intnode *to,
struct xfs_da3_icnode_hdr *from);
static inline int
-xfs_da3_node_hdr_size(struct xfs_da_intnode *dap)
+__xfs_da3_node_hdr_size(bool v3)
{
- if (dap->hdr.info.magic == cpu_to_be16(XFS_DA3_NODE_MAGIC))
+ if (v3)
return sizeof(struct xfs_da3_node_hdr);
return sizeof(struct xfs_da_node_hdr);
}
+static inline int
+xfs_da3_node_hdr_size(struct xfs_da_intnode *dap)
+{
+ bool v3 = dap->hdr.info.magic == cpu_to_be16(XFS_DA3_NODE_MAGIC);
+
+ return __xfs_da3_node_hdr_size(v3);
+}
static inline struct xfs_da_node_entry *
xfs_da3_node_tree_p(struct xfs_da_intnode *dap)
diff --git a/fs/xfs/xfs_dir2.c b/fs/xfs/xfs_dir2.c
index d3ff96c..edf203a 100644
--- a/fs/xfs/xfs_dir2.c
+++ b/fs/xfs/xfs_dir2.c
@@ -90,6 +90,9 @@ void
xfs_dir_mount(
xfs_mount_t *mp)
{
+ int nodehdr_size;
+
+
ASSERT(xfs_sb_version_hasdirv2(&mp->m_sb));
ASSERT((1 << (mp->m_sb.sb_blocklog + mp->m_sb.sb_dirblklog)) <=
XFS_MAX_BLOCKSIZE);
@@ -98,12 +101,13 @@ xfs_dir_mount(
mp->m_dirdatablk = xfs_dir2_db_to_da(mp, XFS_DIR2_DATA_FIRSTDB(mp));
mp->m_dirleafblk = xfs_dir2_db_to_da(mp, XFS_DIR2_LEAF_FIRSTDB(mp));
mp->m_dirfreeblk = xfs_dir2_db_to_da(mp, XFS_DIR2_FREE_FIRSTDB(mp));
- mp->m_attr_node_ents =
- (mp->m_sb.sb_blocksize - (uint)sizeof(xfs_da_node_hdr_t)) /
- (uint)sizeof(xfs_da_node_entry_t);
- mp->m_dir_node_ents =
- (mp->m_dirblksize - (uint)sizeof(xfs_da_node_hdr_t)) /
- (uint)sizeof(xfs_da_node_entry_t);
+
+ nodehdr_size = __xfs_da3_node_hdr_size(xfs_sb_version_hascrc(&mp->m_sb));
+ mp->m_attr_node_ents = (mp->m_sb.sb_blocksize - nodehdr_size) /
+ (uint)sizeof(xfs_da_node_entry_t);
+ mp->m_dir_node_ents = (mp->m_dirblksize - nodehdr_size) /
+ (uint)sizeof(xfs_da_node_entry_t);
+
mp->m_dir_magicpct = (mp->m_dirblksize * 37) / 100;
if (xfs_sb_version_hasasciici(&mp->m_sb))
mp->m_dirnameops = &xfs_ascii_ci_nameops;
_______________________________________________
xfs mailing list
xfs@xxxxxxxxxxx
http://oss.sgi.com/mailman/listinfo/xfs
