From: "Darrick J. Wong" <djwong@kernel.org>
To: Dave Chinner <david@fromorbit.com>
Cc: linux-xfs@vger.kernel.org
Subject: Re: [PATCH 5/6] xfs: flush speculative space allocations when we run out of quota
Date: Mon, 11 Jan 2021 17:31:26 -0800 [thread overview]
Message-ID: <20210112013126.GJ1164246@magnolia> (raw)
In-Reply-To: <20210112012249.GP331610@dread.disaster.area>
On Tue, Jan 12, 2021 at 12:22:49PM +1100, Dave Chinner wrote:
> On Mon, Jan 11, 2021 at 03:23:05PM -0800, Darrick J. Wong wrote:
> > From: Darrick J. Wong <djwong@kernel.org>
> >
> > If a fs modification (creation, file write, reflink, etc.) 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_bmap_util.c | 16 ++++++++++++++++
> > fs/xfs/xfs_file.c | 2 +-
> > fs/xfs/xfs_icache.c | 9 +++++++--
> > fs/xfs/xfs_icache.h | 2 +-
> > fs/xfs/xfs_inode.c | 17 +++++++++++++++++
> > fs/xfs/xfs_ioctl.c | 2 ++
> > fs/xfs/xfs_iomap.c | 20 +++++++++++++++++++-
> > fs/xfs/xfs_reflink.c | 40 +++++++++++++++++++++++++++++++++++++---
> > fs/xfs/xfs_trace.c | 1 +
> > fs/xfs/xfs_trace.h | 40 ++++++++++++++++++++++++++++++++++++++++
> > 10 files changed, 141 insertions(+), 8 deletions(-)
> >
> >
> > diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c
> > index 7371a7f7c652..437fdc8a8fbd 100644
> > --- a/fs/xfs/xfs_bmap_util.c
> > +++ b/fs/xfs/xfs_bmap_util.c
> > @@ -761,6 +761,7 @@ xfs_alloc_file_space(
> > */
> > while (allocatesize_fsb && !error) {
> > xfs_fileoff_t s, e;
> > + bool cleared_space = false;
> >
> > /*
> > * Determine space reservations for data/realtime.
> > @@ -803,6 +804,7 @@ xfs_alloc_file_space(
> > /*
> > * Allocate and setup the transaction.
> > */
> > +retry:
> > error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, resblks,
> > resrtextents, 0, &tp);
> >
> > @@ -819,6 +821,20 @@ xfs_alloc_file_space(
> > xfs_ilock(ip, XFS_ILOCK_EXCL);
> > error = xfs_trans_reserve_quota_nblks(tp, ip, qblocks,
> > 0, quota_flag);
> > + /*
> > + * We weren't able to reserve enough quota to handle fallocate.
> > + * Flush any disk space that was being held in the hopes of
> > + * speeding up the filesystem. We hold the IOLOCK so we cannot
> > + * do a synchronous scan.
> > + */
> > + if ((error == -ENOSPC || error == -EDQUOT) && !cleared_space) {
> > + xfs_trans_cancel(tp);
> > + xfs_iunlock(ip, XFS_ILOCK_EXCL);
> > + cleared_space = xfs_inode_free_quota_blocks(ip, false);
> > + if (cleared_space)
> > + goto retry;
> > + return error;
>
> Can't say I'm a fan of repeating this everywhere. Can we move this
> into xfs_trans_reserve_quota_nblks() with a "retry" flag such that
> we do:
>
> error = xfs_trans_reserve_quota_nblks(tp, ip, qblocks, 0,
> quota_flag, &retry);
> if (error) {
> /* tp already cancelled, inode unlocked */
> return error;
> }
> if (retry) {
> /* tp already cancelled, inode unlocked */
> goto retry;
> }
Assuming you'd be interested in the same kind of change being applied to
the next patch (kick the inode reclaim from xfs_trans_reserve on ENOSPC)
then yes, that seems like a nice cleanup.
--D
> Cheers,
>
> Dave.
> --
> Dave Chinner
> david@fromorbit.com
next prev parent reply other threads:[~2021-01-12 1:32 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-01-11 23:22 [PATCHSET v2 0/6] xfs: try harder to reclaim space when we run out Darrick J. Wong
2021-01-11 23:22 ` [PATCH 1/6] xfs: hide most of the incore inode walk interface Darrick J. Wong
2021-01-13 14:34 ` Christoph Hellwig
2021-01-18 19:50 ` Darrick J. Wong
2021-01-11 23:22 ` [PATCH 2/6] xfs: refactor messy xfs_inode_free_quota_* functions Darrick J. Wong
2021-01-13 14:37 ` Christoph Hellwig
2021-01-11 23:22 ` [PATCH 3/6] xfs: don't stall cowblocks scan if we can't take locks Darrick J. Wong
2021-01-13 14:43 ` Christoph Hellwig
2021-01-14 21:54 ` Darrick J. Wong
2021-01-18 17:34 ` Christoph Hellwig
2021-01-18 19:37 ` Darrick J. Wong
2021-01-18 19:39 ` Christoph Hellwig
2021-01-18 19:44 ` Darrick J. Wong
2021-01-18 19:51 ` Christoph Hellwig
2021-01-11 23:22 ` [PATCH 4/6] xfs: xfs_inode_free_quota_blocks should scan project quota Darrick J. Wong
2021-01-13 14:45 ` Christoph Hellwig
2021-01-11 23:23 ` [PATCH 5/6] xfs: flush speculative space allocations when we run out of quota Darrick J. Wong
2021-01-12 1:22 ` Dave Chinner
2021-01-12 1:31 ` Darrick J. Wong [this message]
2021-01-12 1:40 ` Dave Chinner
2021-01-12 2:18 ` Darrick J. Wong
2021-01-14 22:10 ` Darrick J. Wong
2021-01-11 23:23 ` [PATCH 6/6] xfs: flush speculative space allocations when we run out of space Darrick J. Wong
-- strict thread matches above, loose matches on Subject: below --
2020-01-01 1:07 [PATCH v2 0/6] xfs: try harder to reclaim space when we run out Darrick J. Wong
2020-01-01 1:07 ` [PATCH 5/6] xfs: flush speculative space allocations when we run out of quota Darrick J. Wong
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20210112013126.GJ1164246@magnolia \
--to=djwong@kernel.org \
--cc=david@fromorbit.com \
--cc=linux-xfs@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).