Re: [PATCH 14/14] xfs: enable non-blocking timestamp updates
Christoph Hellwig <[email protected]>
| Newsgroups | dev.linux.lists.gfs2,org.infradead.lists.linux-mtd,org.kernel.vger.io-uring,org.kernel.vger.linux-btrfs,org.kernel.vger.linux-kernel,org.kernel.vger.linux-nfs,org.kernel.vger.linux-unionfs,org.kernel.vger.linux-xfs |
|---|---|
| Message-ID | <[email protected]> |
On Sun, Nov 16, 2025 at 07:23:11PM +1100, Dave Chinner wrote:
> On Fri, Nov 14, 2025 at 07:26:17AM +0100, Christoph Hellwig wrote:
> > The lazytime path using generic_update_time can never block in XFS
> > because there is no ->dirty_inode method that could block. Allow
> > non-blocking timestamp updates for this case.
> >
> > Fixes: 66fa3cedf16a ("fs: Add async write file modification handling.")
> > Signed-off-by: Christoph Hellwig <[email protected]>
> > ---
> > fs/xfs/xfs_iops.c | 6 +++---
> > 1 file changed, 3 insertions(+), 3 deletions(-)
> >
> > diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c
> > index bd0b7e81f6ab..3d7b89ffacde 100644
> > --- a/fs/xfs/xfs_iops.c
> > +++ b/fs/xfs/xfs_iops.c
> > @@ -1195,9 +1195,6 @@ xfs_vn_update_time(
> >
> > trace_xfs_update_time(ip);
> >
> > - if (flags & S_NOWAIT)
> > - return -EAGAIN;
> > -
> > if (inode->i_sb->s_flags & SB_LAZYTIME) {
> > if (!((flags & S_VERSION) &&
> > inode_maybe_inc_iversion(inode, false)))
> > @@ -1207,6 +1204,9 @@ xfs_vn_update_time(
> > log_flags |= XFS_ILOG_CORE;
> > }
> >
> > + if (flags & S_NOWAIT)
> > + return -EAGAIN;
> > +
> > error = xfs_trans_alloc(mp, &M_RES(mp)->tr_fsyncts, 0, 0, 0, &tp);
> > if (error)
> > return error;
>
> Not sure this is correct - this can now bump iversion and then
> return -EAGAIN. That means S_VERSION likely won't be set on the
> retry, and we'll go straight through the non-blocking path to
> generic_update_time() and skip logging the iversion update....
Thanks.
I'll fix this by propagating S_NOWAIT to inode_update_timestamps.