* [PATCH] xfs: Fix a deadlock in xfs_log_commit_cil() code path
@ 2013-07-19 22:31 Chandra Seetharaman
2013-07-22 18:53 ` Ben Myers
0 siblings, 1 reply; 7+ messages in thread
From: Chandra Seetharaman @ 2013-07-19 22:31 UTC (permalink / raw)
To: XFS mailing list
While testing and rearranging pquota/gquota code, I stumbled
on a xfs_shutdown() during a mount. But the mount just hung.
Debugged and found that there is a deadlock involving
&log->l_cilp->xc_ctx_lock.
It is in a code path where &log->l_cilp->xc_ctx_lock is first
acquired in read mode and some levels down the same semaphore
is being acquired in write mode causing a deadlock.
This is the stack:
xfs_log_commit_cil -> acquires &log->l_cilp->xc_ctx_lock in read mode
xlog_print_tic_res
xfs_force_shutdown
xfs_log_force_umount
xlog_cil_force
xlog_cil_force_lsn
xlog_cil_push_foreground
xlog_cil_push - tries to acquire same semaphore in write mode
This patch fixes the deadlock by changing the reason code for
xfs_force_shutdown in xlog_print_tic_res() to SHUTDOWN_LOG_IO_ERROR.
SHUTDOWN_LOG_IO_ERROR is the right reason code to be set since
we are in the log path.
Thanks to Dave for suggesting this solution.
Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
---
fs/xfs/xfs_log.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c
index d852a2b..bf89eb9 100644
--- a/fs/xfs/xfs_log.c
+++ b/fs/xfs/xfs_log.c
@@ -1941,7 +1941,7 @@ xlog_print_tic_res(
xfs_alert_tag(mp, XFS_PTAG_LOGRES,
"xlog_write: reservation ran out. Need to up reservation");
- xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
+ xfs_force_shutdown(mp, SHUTDOWN_LOG_IO_ERROR);
}
/*
--
1.7.1
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCH] xfs: Fix a deadlock in xfs_log_commit_cil() code path
2013-07-19 22:31 [PATCH] xfs: Fix a deadlock in xfs_log_commit_cil() code path Chandra Seetharaman
@ 2013-07-22 18:53 ` Ben Myers
0 siblings, 0 replies; 7+ messages in thread
From: Ben Myers @ 2013-07-22 18:53 UTC (permalink / raw)
To: Chandra Seetharaman; +Cc: XFS mailing list
On Fri, Jul 19, 2013 at 05:31:38PM -0500, Chandra Seetharaman wrote:
>
> While testing and rearranging pquota/gquota code, I stumbled
> on a xfs_shutdown() during a mount. But the mount just hung.
>
> Debugged and found that there is a deadlock involving
> &log->l_cilp->xc_ctx_lock.
>
> It is in a code path where &log->l_cilp->xc_ctx_lock is first
> acquired in read mode and some levels down the same semaphore
> is being acquired in write mode causing a deadlock.
>
> This is the stack:
> xfs_log_commit_cil -> acquires &log->l_cilp->xc_ctx_lock in read mode
> xlog_print_tic_res
> xfs_force_shutdown
> xfs_log_force_umount
> xlog_cil_force
> xlog_cil_force_lsn
> xlog_cil_push_foreground
> xlog_cil_push - tries to acquire same semaphore in write mode
>
> This patch fixes the deadlock by changing the reason code for
> xfs_force_shutdown in xlog_print_tic_res() to SHUTDOWN_LOG_IO_ERROR.
>
> SHUTDOWN_LOG_IO_ERROR is the right reason code to be set since
> we are in the log path.
>
> Thanks to Dave for suggesting this solution.
>
> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
Looks fine.
Reviewed-by: Ben Myers <bpm@sgi.com>
Applied.
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] xfs: Fix a deadlock in xfs_log_commit_cil() code path
2013-07-18 2:24 ` Dave Chinner
@ 2013-07-18 19:54 ` Chandra Seetharaman
0 siblings, 0 replies; 7+ messages in thread
From: Chandra Seetharaman @ 2013-07-18 19:54 UTC (permalink / raw)
To: Dave Chinner; +Cc: XFS mailing list
On Thu, 2013-07-18 at 12:24 +1000, Dave Chinner wrote:
> On Wed, Jul 17, 2013 at 04:32:55PM -0500, Chandra Seetharaman wrote:
> > On Tue, 2013-07-16 at 10:54 +1000, Dave Chinner wrote:
> > > On Mon, Jul 15, 2013 at 05:52:34PM -0500, Chandra Seetharaman wrote:
> > > > While testing and rearranging my pquota/gquota code, I stumbled
> > > > on a xfs_shutdown() during a mount. But the mount just hung.
> > > >
> > > > I debugged and found that there is a deadlock involving
> > > > &log->l_cilp->xc_ctx_lock.
> > > >
> > > > It is in a code path where &log->l_cilp->xc_ctx_lock is first
> > > > acquired in read mode and some levels down the same semaphore
> > > > is being acquired in write mode causing a deadlock.
> > > >
> > > > This is the stack:
> > > > xfs_log_commit_cil -> acquires &log->l_cilp->xc_ctx_lock in read mode
> > > > xlog_print_tic_res
> > > > xfs_force_shutdown
> > > > xfs_log_force_umount
> > > > xlog_cil_force
> > > > xlog_cil_force_lsn
> > > > xlog_cil_push_foreground
> > > > xlog_cil_push - tries to acquire same semaphore in write mode
> > > >
> > > > This patch fixes the deadlock by not calling xfs_force_shutdown() while
> > > > holding the semaphore, instead calling it after dropping teh semaphore.
> > > >
> > > > Thanks to Dave for suggesting this solution.
> > > >
> > > > Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
> > > >
> > > > ---
> > > > fs/xfs/xfs_log.c | 6 +++---
> > > > fs/xfs/xfs_log_cil.c | 10 ++++++----
> > > > fs/xfs/xfs_log_priv.h | 2 +-
> > > > fs/xfs/xfs_trans.c | 2 +-
> > > > 4 files changed, 11 insertions(+), 9 deletions(-)
> > > >
> > > > diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c
> > > > index d852a2b..b9fa2da 100644
> > > > --- a/fs/xfs/xfs_log.c
> > > > +++ b/fs/xfs/xfs_log.c
> > > > @@ -1837,7 +1837,7 @@ xlog_state_finish_copy(
> > > > * print out info relating to regions written which consume
> > > > * the reservation
> > > > */
> > > > -void
> > > > +int
> > > > xlog_print_tic_res(
> > > > struct xfs_mount *mp,
> > > > struct xlog_ticket *ticket)
> > > > @@ -1941,7 +1941,7 @@ xlog_print_tic_res(
> > > >
> > > > xfs_alert_tag(mp, XFS_PTAG_LOGRES,
> > > > "xlog_write: reservation ran out. Need to up reservation");
> > > > - xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
> > > > + return EFSCORRUPTED;
> > >
> > > Note the "SHUTDOWN_CORRUPT_INCORE" reason given here....
> > >
> > > > diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
> > > > index 35a2299..d96022f 100644
> > > > --- a/fs/xfs/xfs_trans.c
> > > > +++ b/fs/xfs/xfs_trans.c
> > > > @@ -1547,7 +1547,7 @@ xfs_trans_commit(
> > > > xfs_trans_apply_dquot_deltas(tp);
> > > >
> > > > error = xfs_log_commit_cil(mp, tp, &commit_lsn, flags);
> > > > - if (error == ENOMEM) {
> > > > + if (error) {
> > > > xfs_force_shutdown(mp, SHUTDOWN_LOG_IO_ERROR);
> > >
> > > Which is different to the reason given here. The shutdown reason
> > > should be maintained for this particular error....
> >
> > I see.
>
> What I mean is that the code in xfs_trans_commit() should do
> something like:
>
> if (error) {
> int reason = SHUTDOWN_LOG_IO_ERROR;
> if (error == EFSCORRUPTED)
> reason = SHUTDOWN_CORRUPT_INCORE;
> xfs_force_shutdown(mp, reason);
> ....
> }
>
> >
> > Is it ok if the error reason is not propagated to the xlog_write() code
> > path ?
>
> No - if we get a transaction overflow, we need to trigger a
> shutdown. That means the error needs to be caught by the
> xlog_write() path an the filesystem shut down.
>
> Looking at it more deeply, you could probably just change the
> shutdown in xlog_print_tic_res() to use SHUTDOWN_LOG_IO_ERROR and
> the problem is solved as the shutdown won't try to force the
> log. i.e. this whole problem will go away with that one line fix...
I am confused.
In the previous response you mentioned that we have to propagate the
reason as-is in xfs_trans_commit() path. But, the new suggestion you are
making will change the behavior of all paths and they will not enter
xfs_log_force_umount().
Besides, IIUC, XFS_MOUNT_FS_SHUTDOWN is set only in
xfs_log_force_umount(), so the very first time we enter
xlog_print_tic_res(), even with SHUTDOWN_LOG_IO_ERROR we will call
xfs_log_force_umount() when can lead to the deadlock we are trying to
avoid.
>
> Cheers,
>
> Dave.
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] xfs: Fix a deadlock in xfs_log_commit_cil() code path
2013-07-17 21:32 ` Chandra Seetharaman
@ 2013-07-18 2:24 ` Dave Chinner
2013-07-18 19:54 ` Chandra Seetharaman
0 siblings, 1 reply; 7+ messages in thread
From: Dave Chinner @ 2013-07-18 2:24 UTC (permalink / raw)
To: Chandra Seetharaman; +Cc: XFS mailing list
On Wed, Jul 17, 2013 at 04:32:55PM -0500, Chandra Seetharaman wrote:
> On Tue, 2013-07-16 at 10:54 +1000, Dave Chinner wrote:
> > On Mon, Jul 15, 2013 at 05:52:34PM -0500, Chandra Seetharaman wrote:
> > > While testing and rearranging my pquota/gquota code, I stumbled
> > > on a xfs_shutdown() during a mount. But the mount just hung.
> > >
> > > I debugged and found that there is a deadlock involving
> > > &log->l_cilp->xc_ctx_lock.
> > >
> > > It is in a code path where &log->l_cilp->xc_ctx_lock is first
> > > acquired in read mode and some levels down the same semaphore
> > > is being acquired in write mode causing a deadlock.
> > >
> > > This is the stack:
> > > xfs_log_commit_cil -> acquires &log->l_cilp->xc_ctx_lock in read mode
> > > xlog_print_tic_res
> > > xfs_force_shutdown
> > > xfs_log_force_umount
> > > xlog_cil_force
> > > xlog_cil_force_lsn
> > > xlog_cil_push_foreground
> > > xlog_cil_push - tries to acquire same semaphore in write mode
> > >
> > > This patch fixes the deadlock by not calling xfs_force_shutdown() while
> > > holding the semaphore, instead calling it after dropping teh semaphore.
> > >
> > > Thanks to Dave for suggesting this solution.
> > >
> > > Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
> > >
> > > ---
> > > fs/xfs/xfs_log.c | 6 +++---
> > > fs/xfs/xfs_log_cil.c | 10 ++++++----
> > > fs/xfs/xfs_log_priv.h | 2 +-
> > > fs/xfs/xfs_trans.c | 2 +-
> > > 4 files changed, 11 insertions(+), 9 deletions(-)
> > >
> > > diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c
> > > index d852a2b..b9fa2da 100644
> > > --- a/fs/xfs/xfs_log.c
> > > +++ b/fs/xfs/xfs_log.c
> > > @@ -1837,7 +1837,7 @@ xlog_state_finish_copy(
> > > * print out info relating to regions written which consume
> > > * the reservation
> > > */
> > > -void
> > > +int
> > > xlog_print_tic_res(
> > > struct xfs_mount *mp,
> > > struct xlog_ticket *ticket)
> > > @@ -1941,7 +1941,7 @@ xlog_print_tic_res(
> > >
> > > xfs_alert_tag(mp, XFS_PTAG_LOGRES,
> > > "xlog_write: reservation ran out. Need to up reservation");
> > > - xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
> > > + return EFSCORRUPTED;
> >
> > Note the "SHUTDOWN_CORRUPT_INCORE" reason given here....
> >
> > > diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
> > > index 35a2299..d96022f 100644
> > > --- a/fs/xfs/xfs_trans.c
> > > +++ b/fs/xfs/xfs_trans.c
> > > @@ -1547,7 +1547,7 @@ xfs_trans_commit(
> > > xfs_trans_apply_dquot_deltas(tp);
> > >
> > > error = xfs_log_commit_cil(mp, tp, &commit_lsn, flags);
> > > - if (error == ENOMEM) {
> > > + if (error) {
> > > xfs_force_shutdown(mp, SHUTDOWN_LOG_IO_ERROR);
> >
> > Which is different to the reason given here. The shutdown reason
> > should be maintained for this particular error....
>
> I see.
What I mean is that the code in xfs_trans_commit() should do
something like:
if (error) {
int reason = SHUTDOWN_LOG_IO_ERROR;
if (error == EFSCORRUPTED)
reason = SHUTDOWN_CORRUPT_INCORE;
xfs_force_shutdown(mp, reason);
....
}
>
> Is it ok if the error reason is not propagated to the xlog_write() code
> path ?
No - if we get a transaction overflow, we need to trigger a
shutdown. That means the error needs to be caught by the
xlog_write() path an the filesystem shut down.
Looking at it more deeply, you could probably just change the
shutdown in xlog_print_tic_res() to use SHUTDOWN_LOG_IO_ERROR and
the problem is solved as the shutdown won't try to force the
log. i.e. this whole problem will go away with that one line fix...
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] xfs: Fix a deadlock in xfs_log_commit_cil() code path
2013-07-16 0:54 ` Dave Chinner
@ 2013-07-17 21:32 ` Chandra Seetharaman
2013-07-18 2:24 ` Dave Chinner
0 siblings, 1 reply; 7+ messages in thread
From: Chandra Seetharaman @ 2013-07-17 21:32 UTC (permalink / raw)
To: Dave Chinner; +Cc: XFS mailing list
On Tue, 2013-07-16 at 10:54 +1000, Dave Chinner wrote:
> On Mon, Jul 15, 2013 at 05:52:34PM -0500, Chandra Seetharaman wrote:
> > While testing and rearranging my pquota/gquota code, I stumbled
> > on a xfs_shutdown() during a mount. But the mount just hung.
> >
> > I debugged and found that there is a deadlock involving
> > &log->l_cilp->xc_ctx_lock.
> >
> > It is in a code path where &log->l_cilp->xc_ctx_lock is first
> > acquired in read mode and some levels down the same semaphore
> > is being acquired in write mode causing a deadlock.
> >
> > This is the stack:
> > xfs_log_commit_cil -> acquires &log->l_cilp->xc_ctx_lock in read mode
> > xlog_print_tic_res
> > xfs_force_shutdown
> > xfs_log_force_umount
> > xlog_cil_force
> > xlog_cil_force_lsn
> > xlog_cil_push_foreground
> > xlog_cil_push - tries to acquire same semaphore in write mode
> >
> > This patch fixes the deadlock by not calling xfs_force_shutdown() while
> > holding the semaphore, instead calling it after dropping teh semaphore.
> >
> > Thanks to Dave for suggesting this solution.
> >
> > Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
> >
> > ---
> > fs/xfs/xfs_log.c | 6 +++---
> > fs/xfs/xfs_log_cil.c | 10 ++++++----
> > fs/xfs/xfs_log_priv.h | 2 +-
> > fs/xfs/xfs_trans.c | 2 +-
> > 4 files changed, 11 insertions(+), 9 deletions(-)
> >
> > diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c
> > index d852a2b..b9fa2da 100644
> > --- a/fs/xfs/xfs_log.c
> > +++ b/fs/xfs/xfs_log.c
> > @@ -1837,7 +1837,7 @@ xlog_state_finish_copy(
> > * print out info relating to regions written which consume
> > * the reservation
> > */
> > -void
> > +int
> > xlog_print_tic_res(
> > struct xfs_mount *mp,
> > struct xlog_ticket *ticket)
> > @@ -1941,7 +1941,7 @@ xlog_print_tic_res(
> >
> > xfs_alert_tag(mp, XFS_PTAG_LOGRES,
> > "xlog_write: reservation ran out. Need to up reservation");
> > - xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
> > + return EFSCORRUPTED;
>
> Note the "SHUTDOWN_CORRUPT_INCORE" reason given here....
>
> > diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
> > index 35a2299..d96022f 100644
> > --- a/fs/xfs/xfs_trans.c
> > +++ b/fs/xfs/xfs_trans.c
> > @@ -1547,7 +1547,7 @@ xfs_trans_commit(
> > xfs_trans_apply_dquot_deltas(tp);
> >
> > error = xfs_log_commit_cil(mp, tp, &commit_lsn, flags);
> > - if (error == ENOMEM) {
> > + if (error) {
> > xfs_force_shutdown(mp, SHUTDOWN_LOG_IO_ERROR);
>
> Which is different to the reason given here. The shutdown reason
> should be maintained for this particular error....
I see.
Is it ok if the error reason is not propagated to the xlog_write() code
path ?
>
> Cheers,
>
> Dave.
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH] xfs: Fix a deadlock in xfs_log_commit_cil() code path
2013-07-15 22:52 Chandra Seetharaman
@ 2013-07-16 0:54 ` Dave Chinner
2013-07-17 21:32 ` Chandra Seetharaman
0 siblings, 1 reply; 7+ messages in thread
From: Dave Chinner @ 2013-07-16 0:54 UTC (permalink / raw)
To: Chandra Seetharaman; +Cc: XFS mailing list
On Mon, Jul 15, 2013 at 05:52:34PM -0500, Chandra Seetharaman wrote:
> While testing and rearranging my pquota/gquota code, I stumbled
> on a xfs_shutdown() during a mount. But the mount just hung.
>
> I debugged and found that there is a deadlock involving
> &log->l_cilp->xc_ctx_lock.
>
> It is in a code path where &log->l_cilp->xc_ctx_lock is first
> acquired in read mode and some levels down the same semaphore
> is being acquired in write mode causing a deadlock.
>
> This is the stack:
> xfs_log_commit_cil -> acquires &log->l_cilp->xc_ctx_lock in read mode
> xlog_print_tic_res
> xfs_force_shutdown
> xfs_log_force_umount
> xlog_cil_force
> xlog_cil_force_lsn
> xlog_cil_push_foreground
> xlog_cil_push - tries to acquire same semaphore in write mode
>
> This patch fixes the deadlock by not calling xfs_force_shutdown() while
> holding the semaphore, instead calling it after dropping teh semaphore.
>
> Thanks to Dave for suggesting this solution.
>
> Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
>
> ---
> fs/xfs/xfs_log.c | 6 +++---
> fs/xfs/xfs_log_cil.c | 10 ++++++----
> fs/xfs/xfs_log_priv.h | 2 +-
> fs/xfs/xfs_trans.c | 2 +-
> 4 files changed, 11 insertions(+), 9 deletions(-)
>
> diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c
> index d852a2b..b9fa2da 100644
> --- a/fs/xfs/xfs_log.c
> +++ b/fs/xfs/xfs_log.c
> @@ -1837,7 +1837,7 @@ xlog_state_finish_copy(
> * print out info relating to regions written which consume
> * the reservation
> */
> -void
> +int
> xlog_print_tic_res(
> struct xfs_mount *mp,
> struct xlog_ticket *ticket)
> @@ -1941,7 +1941,7 @@ xlog_print_tic_res(
>
> xfs_alert_tag(mp, XFS_PTAG_LOGRES,
> "xlog_write: reservation ran out. Need to up reservation");
> - xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
> + return EFSCORRUPTED;
Note the "SHUTDOWN_CORRUPT_INCORE" reason given here....
> diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
> index 35a2299..d96022f 100644
> --- a/fs/xfs/xfs_trans.c
> +++ b/fs/xfs/xfs_trans.c
> @@ -1547,7 +1547,7 @@ xfs_trans_commit(
> xfs_trans_apply_dquot_deltas(tp);
>
> error = xfs_log_commit_cil(mp, tp, &commit_lsn, flags);
> - if (error == ENOMEM) {
> + if (error) {
> xfs_force_shutdown(mp, SHUTDOWN_LOG_IO_ERROR);
Which is different to the reason given here. The shutdown reason
should be maintained for this particular error....
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 7+ messages in thread
* [PATCH] xfs: Fix a deadlock in xfs_log_commit_cil() code path
@ 2013-07-15 22:52 Chandra Seetharaman
2013-07-16 0:54 ` Dave Chinner
0 siblings, 1 reply; 7+ messages in thread
From: Chandra Seetharaman @ 2013-07-15 22:52 UTC (permalink / raw)
To: XFS mailing list
While testing and rearranging my pquota/gquota code, I stumbled
on a xfs_shutdown() during a mount. But the mount just hung.
I debugged and found that there is a deadlock involving
&log->l_cilp->xc_ctx_lock.
It is in a code path where &log->l_cilp->xc_ctx_lock is first
acquired in read mode and some levels down the same semaphore
is being acquired in write mode causing a deadlock.
This is the stack:
xfs_log_commit_cil -> acquires &log->l_cilp->xc_ctx_lock in read mode
xlog_print_tic_res
xfs_force_shutdown
xfs_log_force_umount
xlog_cil_force
xlog_cil_force_lsn
xlog_cil_push_foreground
xlog_cil_push - tries to acquire same semaphore in write mode
This patch fixes the deadlock by not calling xfs_force_shutdown() while
holding the semaphore, instead calling it after dropping teh semaphore.
Thanks to Dave for suggesting this solution.
Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
---
fs/xfs/xfs_log.c | 6 +++---
fs/xfs/xfs_log_cil.c | 10 ++++++----
fs/xfs/xfs_log_priv.h | 2 +-
fs/xfs/xfs_trans.c | 2 +-
4 files changed, 11 insertions(+), 9 deletions(-)
diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c
index d852a2b..b9fa2da 100644
--- a/fs/xfs/xfs_log.c
+++ b/fs/xfs/xfs_log.c
@@ -1837,7 +1837,7 @@ xlog_state_finish_copy(
* print out info relating to regions written which consume
* the reservation
*/
-void
+int
xlog_print_tic_res(
struct xfs_mount *mp,
struct xlog_ticket *ticket)
@@ -1941,7 +1941,7 @@ xlog_print_tic_res(
xfs_alert_tag(mp, XFS_PTAG_LOGRES,
"xlog_write: reservation ran out. Need to up reservation");
- xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
+ return EFSCORRUPTED;
}
/*
@@ -2215,7 +2215,7 @@ xlog_write(
ticket->t_curr_res -= sizeof(xlog_op_header_t);
if (ticket->t_curr_res < 0)
- xlog_print_tic_res(log->l_mp, ticket);
+ return xlog_print_tic_res(log->l_mp, ticket);
index = 0;
lv = log_vector;
diff --git a/fs/xfs/xfs_log_cil.c b/fs/xfs/xfs_log_cil.c
index 02b9cf3..93ba7bd 100644
--- a/fs/xfs/xfs_log_cil.c
+++ b/fs/xfs/xfs_log_cil.c
@@ -730,10 +730,6 @@ xfs_log_commit_cil(
/* xlog_cil_insert_items() destroys log_vector list */
xlog_cil_insert_items(log, log_vector, tp->t_ticket);
- /* check we didn't blow the reservation */
- if (tp->t_ticket->t_curr_res < 0)
- xlog_print_tic_res(log->l_mp, tp->t_ticket);
-
/* attach the transaction to the CIL if it has any busy extents */
if (!list_empty(&tp->t_busy)) {
spin_lock(&log->l_cilp->xc_cil_lock);
@@ -742,6 +738,12 @@ xfs_log_commit_cil(
spin_unlock(&log->l_cilp->xc_cil_lock);
}
+ /* check we didn't blow the reservation */
+ if (tp->t_ticket->t_curr_res < 0) {
+ up_read(&log->l_cilp->xc_ctx_lock);
+ return xlog_print_tic_res(log->l_mp, tp->t_ticket);
+ }
+
tp->t_commit_lsn = *commit_lsn;
xfs_log_done(mp, tp->t_ticket, NULL, log_flags);
xfs_trans_unreserve_and_mod_sb(tp);
diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h
index b9ea262..4f2fa6d 100644
--- a/fs/xfs/xfs_log_priv.h
+++ b/fs/xfs/xfs_log_priv.h
@@ -576,7 +576,7 @@ xlog_write_adv_cnt(void **ptr, int *len, int *off, size_t bytes)
*off += bytes;
}
-void xlog_print_tic_res(struct xfs_mount *mp, struct xlog_ticket *ticket);
+int xlog_print_tic_res(struct xfs_mount *mp, struct xlog_ticket *ticket);
int
xlog_write(
struct xlog *log,
diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
index 35a2299..d96022f 100644
--- a/fs/xfs/xfs_trans.c
+++ b/fs/xfs/xfs_trans.c
@@ -1547,7 +1547,7 @@ xfs_trans_commit(
xfs_trans_apply_dquot_deltas(tp);
error = xfs_log_commit_cil(mp, tp, &commit_lsn, flags);
- if (error == ENOMEM) {
+ if (error) {
xfs_force_shutdown(mp, SHUTDOWN_LOG_IO_ERROR);
error = XFS_ERROR(EIO);
goto out_unreserve;
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply related [flat|nested] 7+ messages in thread
end of thread, other threads:[~2013-07-22 18:53 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-07-19 22:31 [PATCH] xfs: Fix a deadlock in xfs_log_commit_cil() code path Chandra Seetharaman
2013-07-22 18:53 ` Ben Myers
-- strict thread matches above, loose matches on Subject: below --
2013-07-15 22:52 Chandra Seetharaman
2013-07-16 0:54 ` Dave Chinner
2013-07-17 21:32 ` Chandra Seetharaman
2013-07-18 2:24 ` Dave Chinner
2013-07-18 19:54 ` Chandra Seetharaman
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.