diff mbox series

[09/12] xfs: flush eof/cowblocks if we can't reserve quota for chown

Message ID 161214517714.140945.1957722027452288290.stgit@magnolia (mailing list archive)
State Accepted
Headers show
Series xfs: try harder to reclaim space when we run out | expand

Commit Message

Darrick J. Wong Feb. 1, 2021, 2:06 a.m. UTC 
From: Darrick J. Wong <djwong@kernel.org>

If a file user, group, or project change is unable to reserve enough
quota to handle the modification, try clearing whatever space the
filesystem might have been hanging onto in the hopes of speeding up the
filesystem.  The flushing behavior will become particularly important
when we add deferred inode inactivation because that will increase the
amount of space that isn't actively tied to user data.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
---
 fs/xfs/xfs_trans.c |    9 +++++++++
 1 file changed, 9 insertions(+)

Comments

Christoph Hellwig Feb. 1, 2021, 12:36 p.m. UTC | #1 
On Sun, Jan 31, 2021 at 06:06:17PM -0800, Darrick J. Wong wrote:
> @@ -1175,6 +1177,13 @@ xfs_trans_alloc_ichange(
>  	if (new_udqp || new_gdqp || new_pdqp) {
>  		error = xfs_trans_reserve_quota_chown(tp, ip, new_udqp,
>  				new_gdqp, new_pdqp, force);
> +		if (!retried && (error == -EDQUOT || error == -ENOSPC)) {

One more :)

Otherwise looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>
Darrick J. Wong Feb. 1, 2021, 7:12 p.m. UTC | #2 
On Mon, Feb 01, 2021 at 12:36:37PM +0000, Christoph Hellwig wrote:
> On Sun, Jan 31, 2021 at 06:06:17PM -0800, Darrick J. Wong wrote:
> > @@ -1175,6 +1177,13 @@ xfs_trans_alloc_ichange(
> >  	if (new_udqp || new_gdqp || new_pdqp) {
> >  		error = xfs_trans_reserve_quota_chown(tp, ip, new_udqp,
> >  				new_gdqp, new_pdqp, force);
> > +		if (!retried && (error == -EDQUOT || error == -ENOSPC)) {
> 
> One more :)
> 
> Otherwise looks good:
> 
> Reviewed-by: Christoph Hellwig <hch@lst.de>

Cool, thanks for reviewing!

--D
Brian Foster Feb. 2, 2021, 3:39 p.m. UTC | #3 
On Sun, Jan 31, 2021 at 06:06:17PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <djwong@kernel.org>
> 
> If a file user, group, or project change is unable to reserve enough
> quota to handle the modification, try clearing whatever space the
> filesystem might have been hanging onto in the hopes of speeding up the
> filesystem.  The flushing behavior will become particularly important
> when we add deferred inode inactivation because that will increase the
> amount of space that isn't actively tied to user data.
> 
> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
> ---

Reviewed-by: Brian Foster <bfoster@redhat.com>

>  fs/xfs/xfs_trans.c |    9 +++++++++
>  1 file changed, 9 insertions(+)
> 
> 
> diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
> index ee3cb916c5c9..3203841ab19b 100644
> --- a/fs/xfs/xfs_trans.c
> +++ b/fs/xfs/xfs_trans.c
> @@ -1149,8 +1149,10 @@ xfs_trans_alloc_ichange(
>  	struct xfs_dquot	*new_udqp;
>  	struct xfs_dquot	*new_gdqp;
>  	struct xfs_dquot	*new_pdqp;
> +	bool			retried = false;
>  	int			error;
>  
> +retry:
>  	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ichange, 0, 0, 0, &tp);
>  	if (error)
>  		return error;
> @@ -1175,6 +1177,13 @@ xfs_trans_alloc_ichange(
>  	if (new_udqp || new_gdqp || new_pdqp) {
>  		error = xfs_trans_reserve_quota_chown(tp, ip, new_udqp,
>  				new_gdqp, new_pdqp, force);
> +		if (!retried && (error == -EDQUOT || error == -ENOSPC)) {
> +			xfs_trans_cancel(tp);
> +			xfs_blockgc_free_dquots(new_udqp, new_gdqp, new_pdqp,
> +					0);
> +			retried = true;
> +			goto retry;
> +		}
>  		if (error)
>  			goto out_cancel;
>  	}
>
diff mbox series

Patch

diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
index ee3cb916c5c9..3203841ab19b 100644
--- a/fs/xfs/xfs_trans.c
+++ b/fs/xfs/xfs_trans.c
@@ -1149,8 +1149,10 @@  xfs_trans_alloc_ichange(
 	struct xfs_dquot	*new_udqp;
 	struct xfs_dquot	*new_gdqp;
 	struct xfs_dquot	*new_pdqp;
+	bool			retried = false;
 	int			error;
 
+retry:
 	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ichange, 0, 0, 0, &tp);
 	if (error)
 		return error;
@@ -1175,6 +1177,13 @@  xfs_trans_alloc_ichange(
 	if (new_udqp || new_gdqp || new_pdqp) {
 		error = xfs_trans_reserve_quota_chown(tp, ip, new_udqp,
 				new_gdqp, new_pdqp, force);
+		if (!retried && (error == -EDQUOT || error == -ENOSPC)) {
+			xfs_trans_cancel(tp);
+			xfs_blockgc_free_dquots(new_udqp, new_gdqp, new_pdqp,
+					0);
+			retried = true;
+			goto retry;
+		}
 		if (error)
 			goto out_cancel;
 	}