[PATCH] xfs: fix error handling in xfs_qm_log_quotaoff()

[Brian Foster]

The error handling in xfs_qm_log_quotaoff() has a couple problems. If
xfs_trans_commit() fails, we fall through to the error block and call
xfs_trans_cancel(). This is incorrect on commit failure. If
xfs_trans_reserve() fails, we jump to the error block, cancel the tp and
restore the superblock qflags to oldsbqflag. However, oldsbqflag has
been initialized to zero and not yet updated from the original flags so
we set the flags to zero.

Fix up the error handling in xfs_qm_log_quotaoff() to not restore flags
if they haven't been modified and not cancel the tp on commit failure.

Reported-by: Dan Carpenter <***@oracle.com>
Signed-off-by: Brian Foster <***@redhat.com>
---
fs/xfs/xfs_qm_syscalls.c | 35 ++++++++++++++++++++---------------
1 file changed, 20 insertions(+), 15 deletions(-)

diff --git a/fs/xfs/xfs_qm_syscalls.c b/fs/xfs/xfs_qm_syscalls.c
index 80f2d77..f2d454e 100644
--- a/fs/xfs/xfs_qm_syscalls.c
+++ b/fs/xfs/xfs_qm_syscalls.c
@@ -784,13 +784,17 @@ xfs_qm_log_quotaoff(
{
xfs_trans_t *tp;
int error;
- xfs_qoff_logitem_t *qoffi=NULL;
- uint oldsbqflag=0;
+ xfs_qoff_logitem_t *qoffi;
+ uint oldsbqflag;
+
+ *qoffstartp = NULL;

tp = xfs_trans_alloc(mp, XFS_TRANS_QM_QUOTAOFF);
error = xfs_trans_reserve(tp, &M_RES(mp)->tr_qm_quotaoff, 0, 0);
- if (error)
- goto error0;
+ if (error) {
+ xfs_trans_cancel(tp, 0);
+ return error;
+ }

qoffi = xfs_trans_get_qoff_item(tp, NULL, flags & XFS_ALL_QUOTA_ACCT);
xfs_trans_log_quotaoff_item(tp, qoffi);
@@ -809,19 +813,20 @@ xfs_qm_log_quotaoff(
*/
xfs_trans_set_sync(tp);
error = xfs_trans_commit(tp, 0);
+ if (error)
+ goto error_flags;

-error0:
- if (error) {
- xfs_trans_cancel(tp, 0);
- /*
- * No one else is modifying sb_qflags, so this is OK.
- * We still hold the quotaofflock.
- */
- spin_lock(&mp->m_sb_lock);
- mp->m_sb.sb_qflags = oldsbqflag;
- spin_unlock(&mp->m_sb_lock);
- }
*qoffstartp = qoffi;
+ return 0;
+
+error_flags:
+ /*
+ * No one else is modifying sb_qflags, so this is OK.
+ * We still hold the quotaofflock.
+ */
+ spin_lock(&mp->m_sb_lock);
+ mp->m_sb.sb_qflags = oldsbqflag;
+ spin_unlock(&mp->m_sb_lock);
return error;
}

[Dave Chinner]

Post by Brian Foster
The error handling in xfs_qm_log_quotaoff() has a couple problems. If
xfs_trans_commit() fails, we fall through to the error block and call
xfs_trans_cancel(). This is incorrect on commit failure. If
xfs_trans_reserve() fails, we jump to the error block, cancel the tp and
restore the superblock qflags to oldsbqflag. However, oldsbqflag has
been initialized to zero and not yet updated from the original flags so
we set the flags to zero.
Fix up the error handling in xfs_qm_log_quotaoff() to not restore flags
if they haven't been modified and not cancel the tp on commit failure.
---
fs/xfs/xfs_qm_syscalls.c | 35 ++++++++++++++++++++---------------
1 file changed, 20 insertions(+), 15 deletions(-)
diff --git a/fs/xfs/xfs_qm_syscalls.c b/fs/xfs/xfs_qm_syscalls.c
index 80f2d77..f2d454e 100644
--- a/fs/xfs/xfs_qm_syscalls.c
+++ b/fs/xfs/xfs_qm_syscalls.c
@@ -784,13 +784,17 @@ xfs_qm_log_quotaoff(
{
xfs_trans_t *tp;
int error;
- xfs_qoff_logitem_t *qoffi=NULL;
- uint oldsbqflag=0;
+ xfs_qoff_logitem_t *qoffi;
+ uint oldsbqflag;
+
+ *qoffstartp = NULL;
tp = xfs_trans_alloc(mp, XFS_TRANS_QM_QUOTAOFF);
error = xfs_trans_reserve(tp, &M_RES(mp)->tr_qm_quotaoff, 0, 0);
- if (error)
- goto error0;
+ if (error) {
+ xfs_trans_cancel(tp, 0);
+ return error;
+ }
qoffi = xfs_trans_get_qoff_item(tp, NULL, flags & XFS_ALL_QUOTA_ACCT);
xfs_trans_log_quotaoff_item(tp, qoffi);
@@ -809,19 +813,20 @@ xfs_qm_log_quotaoff(
*/
xfs_trans_set_sync(tp);
error = xfs_trans_commit(tp, 0);
+ if (error)
+ goto error_flags;

If the commit fails, the filesystem will be shut down and the state
of the quota flags is completely irrelevant at this point so there's
no reason to restore them. Indeed, is restoring them even the right
thing to do? The commit *may* have made it to disk, but a
subsequent error during completion handling resulted in the commit
failing....

Cheers,

Dave.

[Brian Foster]

Post by Dave Chinner

Post by Brian Foster
The error handling in xfs_qm_log_quotaoff() has a couple problems. If
xfs_trans_commit() fails, we fall through to the error block and call
xfs_trans_cancel(). This is incorrect on commit failure. If
xfs_trans_reserve() fails, we jump to the error block, cancel the tp and
restore the superblock qflags to oldsbqflag. However, oldsbqflag has
been initialized to zero and not yet updated from the original flags so
we set the flags to zero.
Fix up the error handling in xfs_qm_log_quotaoff() to not restore flags
if they haven't been modified and not cancel the tp on commit failure.
---
fs/xfs/xfs_qm_syscalls.c | 35 ++++++++++++++++++++---------------
1 file changed, 20 insertions(+), 15 deletions(-)
diff --git a/fs/xfs/xfs_qm_syscalls.c b/fs/xfs/xfs_qm_syscalls.c
index 80f2d77..f2d454e 100644
--- a/fs/xfs/xfs_qm_syscalls.c
+++ b/fs/xfs/xfs_qm_syscalls.c
@@ -784,13 +784,17 @@ xfs_qm_log_quotaoff(
{
xfs_trans_t *tp;
int error;
- xfs_qoff_logitem_t *qoffi=NULL;
- uint oldsbqflag=0;
+ xfs_qoff_logitem_t *qoffi;
+ uint oldsbqflag;
+
+ *qoffstartp = NULL;
tp = xfs_trans_alloc(mp, XFS_TRANS_QM_QUOTAOFF);
error = xfs_trans_reserve(tp, &M_RES(mp)->tr_qm_quotaoff, 0, 0);
- if (error)
- goto error0;
+ if (error) {
+ xfs_trans_cancel(tp, 0);
+ return error;
+ }
qoffi = xfs_trans_get_qoff_item(tp, NULL, flags & XFS_ALL_QUOTA_ACCT);
xfs_trans_log_quotaoff_item(tp, qoffi);
@@ -809,19 +813,20 @@ xfs_qm_log_quotaoff(
*/
xfs_trans_set_sync(tp);
error = xfs_trans_commit(tp, 0);
+ if (error)
+ goto error_flags;

If the commit fails, the filesystem will be shut down and the state
of the quota flags is completely irrelevant at this point so there's
no reason to restore them. Indeed, is restoring them even the right
thing to do? The commit *may* have made it to disk, but a
subsequent error during completion handling resulted in the commit
failing....

Good point. We don't seem to try and undo this kind of state change
anywhere else that I can see either. At most, we release references or
free memory or other such things that aren't tied to the transaction in
any way.

Brian

Post by Dave Chinner
Cheers,
Dave.
--
Dave Chinner
_______________________________________________
xfs mailing list
http://oss.sgi.com/mailman/listinfo/xfs