Re: [PATCH] xfs: Properly retry failed dquot items in case of error during buffer writeback

["Darrick J. Wong" <darrick.wong@oracle.com>]

On Fri, Nov 24, 2017 at 01:03:53PM +0100, Carlos Maiolino wrote:
> Once the inode item writeback errors is already fixed, it's time to fix the same
> problem in dquot code.
> 
> Although there were no reports of users hitting this bug in dquot code (at least
> none I've seen), the bug is there and I was already planning to fix it when the
> correct approach to fix the inodes part was decided.
> 
> This patch aims to fix the same problem in dquot code, regarding failed buffers
> being unable to be resubmitted once they are flush locked.
> 
> Tested with the recently test-case sent to fstests list by Hou Tao.
> 
> Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
> ---
> 
> From RFC:
> 	- ASSERT the correct dquot flush lock in xfs_dquot_item_error()
> 
> 	As in the inode item error path, we ASSERT the inode flush lock is
> 	properly locked when settin XFS_LI_FAILED to the log item. I'm not 100%
> 	sure I've handled the dquot flush lock properly. Once dquot uses a
> 	completion queue as a flush lock, I used completion_done() helper to
> 	check if it's properly locked, but I'm not 100% sure if it is the right
> 	approach, once it uses a spin_lock/unlock irqsave/restore, and I don't
> 	know if in this code path it is ok to run such spin locks.
> 	What you guys think?
> 
> with flush lock
> 	- add an extra couple of parenthesis in xfs_qm_dqflush_done() as
> 	  suggested by Darrick.
> 
> 
>  fs/xfs/xfs_dquot.c      | 11 ++++++++---
>  fs/xfs/xfs_dquot_item.c | 38 ++++++++++++++++++++++++++++++++++++--
>  2 files changed, 44 insertions(+), 5 deletions(-)
> 
> diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c
> index d57c2db64e59..a77b5c8e4f6b 100644
> --- a/fs/xfs/xfs_dquot.c
> +++ b/fs/xfs/xfs_dquot.c
> @@ -970,14 +970,19 @@ xfs_qm_dqflush_done(
>  	 * holding the lock before removing the dquot from the AIL.
>  	 */
>  	if ((lip->li_flags & XFS_LI_IN_AIL) &&
> -	    lip->li_lsn == qip->qli_flush_lsn) {
> +	   ((lip->li_lsn == qip->qli_flush_lsn) ||
> +	    lip->li_flags & XFS_LI_FAILED)) {

/me continues to grouse about the lack of parentheses around the LI_FAILED
test...

>  
>  		/* xfs_trans_ail_delete() drops the AIL lock. */
>  		spin_lock(&ailp->xa_lock);
> -		if (lip->li_lsn == qip->qli_flush_lsn)
> +		if (lip->li_lsn == qip->qli_flush_lsn) {
>  			xfs_trans_ail_delete(ailp, lip, SHUTDOWN_CORRUPT_INCORE);
> -		else
> +		} else if (lip->li_flags & XFS_LI_FAILED) {
> +			xfs_clear_li_failed(lip);
>  			spin_unlock(&ailp->xa_lock);
> +		} else {
> +			spin_unlock(&ailp->xa_lock);
> +		}
>  	}
>  
>  	/*
> diff --git a/fs/xfs/xfs_dquot_item.c b/fs/xfs/xfs_dquot_item.c
> index 2c7a1629e064..3d73a0124988 100644
> --- a/fs/xfs/xfs_dquot_item.c
> +++ b/fs/xfs/xfs_dquot_item.c
> @@ -137,6 +137,24 @@ xfs_qm_dqunpin_wait(
>  	wait_event(dqp->q_pinwait, (atomic_read(&dqp->q_pincount) == 0));
>  }
>  
> +/*
> + * Callback used to mark a buffer with XFS_LI_FAILED when items in the buffer
> + * have been failed during writeback
> + *
> + * this informs the AIL that the dquot is already flush locked on the next push,
> + * and acquires a hold on the buffer to ensure that it isn't reclaimed before
> + * dirty data makes it to disk.
> + */
> +STATIC void
> +xfs_dquot_item_error(
> +	struct xfs_log_item	*lip,
> +	struct xfs_buf		*bp)
> +{
> +	struct xfs_dquot	*dqp = DQUOT_ITEM(lip)->qli_dquot;

Need blank line between variable definition and other code.

> +	ASSERT(!completion_done(&dqp->q_flush));
> +	xfs_set_li_failed(lip, bp);
> +}
> +
>  STATIC uint
>  xfs_qm_dquot_logitem_push(
>  	struct xfs_log_item	*lip,
> @@ -144,13 +162,28 @@ xfs_qm_dquot_logitem_push(
>  					      __acquires(&lip->li_ailp->xa_lock)
>  {
>  	struct xfs_dquot	*dqp = DQUOT_ITEM(lip)->qli_dquot;
> -	struct xfs_buf		*bp = NULL;
> +	struct xfs_buf		*bp = lip->li_buf;
>  	uint			rval = XFS_ITEM_SUCCESS;
>  	int			error;
>  
>  	if (atomic_read(&dqp->q_pincount) > 0)
>  		return XFS_ITEM_PINNED;
>  
> +	/*
> +	 * The buffer containing this item failed to be written back
> +	 * previously. Resubmit the buffer for IO
> +	 */
> +	if (lip->li_flags & XFS_LI_FAILED) {
> +		if (!xfs_buf_trylock(bp))
> +			return XFS_ITEM_LOCKED;
> +
> +		if (!xfs_buf_resubmit_failed_buffers(bp, lip, buffer_list))
> +			rval = XFS_ITEM_FLUSHING;
> +
> +		xfs_buf_unlock(bp);
> +		return rval;
> +	}
> +
>  	if (!xfs_dqlock_nowait(dqp))
>  		return XFS_ITEM_LOCKED;
>  
> @@ -242,7 +275,8 @@ static const struct xfs_item_ops xfs_dquot_item_ops = {
>  	.iop_unlock	= xfs_qm_dquot_logitem_unlock,
>  	.iop_committed	= xfs_qm_dquot_logitem_committed,
>  	.iop_push	= xfs_qm_dquot_logitem_push,
> -	.iop_committing = xfs_qm_dquot_logitem_committing
> +	.iop_committing = xfs_qm_dquot_logitem_committing,
> +	.iop_error	= xfs_dquot_item_error
>  };

Otherwise looks ok; what was the xfstest for this patch?

--D

>  
>  /*
> -- 
> 2.14.3
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html