Linux NFS development
 help / color / mirror / Atom feed
From: Trond Myklebust <trondmy@hammerspace.com>
To: "schumaker.anna@gmail.com" <schumaker.anna@gmail.com>
Cc: "linux-nfs@vger.kernel.org" <linux-nfs@vger.kernel.org>
Subject: Re: [PATCH 12/19] NFSv4: Fix up delegated attributes in nfs_setattr
Date: Fri, 14 Jun 2024 19:59:23 +0000	[thread overview]
Message-ID: <bddef727ec63731061a7c80f477f8dc112085b0a.camel@hammerspace.com> (raw)
In-Reply-To: <CAFX2Jfk7u37+AOX_o1ZRf-QX_abSDXpKaEpHt=iOvL5Bq6opTQ@mail.gmail.com>

On Fri, 2024-06-14 at 12:32 -0400, Anna Schumaker wrote:
> Hi Trond,
> 
> On Thu, Jun 13, 2024 at 12:18 AM <trondmy@gmail.com> wrote:
> > 
> > From: Trond Myklebust <trond.myklebust@primarydata.com>
> > 
> > nfs_setattr calls nfs_update_inode() directly, so we have to reset
> > the
> > m/ctime there.
> > 
> > Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
> > Signed-off-by: Lance Shelton <lance.shelton@hammerspace.com>
> > Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
> > ---
> >  fs/nfs/inode.c | 43 +++++++++++++++++++++++++++++++++++++++++++
> >  1 file changed, 43 insertions(+)
> 
> After applying this patch I see a handful of new failures on
> xfstests:
> generic/075, generic/086, generic/112, generic/332, generic/346,
> generic/647, and generic/729. I see the first five on NFS v4.2, but
> 647 and 729 both fail on v4.1 in addition to v4.2.

Thanks Anna!

Yes, I think that is a consequence of the changes that were made in
commit cc7f2dae63bc ("NFS: Don't store NFS_INO_REVAL_FORCED"). I can
just remove that "if (!(cache_validity & NFS_INO_REVAL_FORCED))"
altogether with that change applied.

Version 2 will be forthcoming.

> 
> I hope this helps!
> Anna
> 
> > 
> > diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
> > index 91c0aeaf6c1e..e03c512c8535 100644
> > --- a/fs/nfs/inode.c
> > +++ b/fs/nfs/inode.c
> > @@ -605,6 +605,46 @@ nfs_fhget(struct super_block *sb, struct
> > nfs_fh *fh, struct nfs_fattr *fattr)
> >  }
> >  EXPORT_SYMBOL_GPL(nfs_fhget);
> > 
> > +static void
> > +nfs_fattr_fixup_delegated(struct inode *inode, struct nfs_fattr
> > *fattr)
> > +{
> > +       unsigned long cache_validity = NFS_I(inode)-
> > >cache_validity;
> > +
> > +       if (!nfs_have_read_or_write_delegation(inode))
> > +               return;
> > +
> > +       if (!(cache_validity & NFS_INO_REVAL_FORCED))
> > +               cache_validity &= ~(NFS_INO_INVALID_ATIME
> > +                               | NFS_INO_INVALID_CHANGE
> > +                               | NFS_INO_INVALID_CTIME
> > +                               | NFS_INO_INVALID_MTIME
> > +                               | NFS_INO_INVALID_SIZE);
> > +
> > +       if (!(cache_validity & NFS_INO_INVALID_SIZE))
> > +               fattr->valid &= ~(NFS_ATTR_FATTR_PRESIZE
> > +                               | NFS_ATTR_FATTR_SIZE);
> > +
> > +       if (!(cache_validity & NFS_INO_INVALID_CHANGE))
> > +               fattr->valid &= ~(NFS_ATTR_FATTR_PRECHANGE
> > +                               | NFS_ATTR_FATTR_CHANGE);
> > +
> > +       if (nfs_have_delegated_mtime(inode)) {
> > +               if (!(cache_validity & NFS_INO_INVALID_CTIME))
> > +                       fattr->valid &= ~(NFS_ATTR_FATTR_PRECTIME
> > +                                       | NFS_ATTR_FATTR_CTIME);
> > +
> > +               if (!(cache_validity & NFS_INO_INVALID_MTIME))
> > +                       fattr->valid &= ~(NFS_ATTR_FATTR_PREMTIME
> > +                                       | NFS_ATTR_FATTR_MTIME);
> > +
> > +               if (!(cache_validity & NFS_INO_INVALID_ATIME))
> > +                       fattr->valid &= ~NFS_ATTR_FATTR_ATIME;
> > +       } else if (nfs_have_delegated_atime(inode)) {
> > +               if (!(cache_validity & NFS_INO_INVALID_ATIME))
> > +                       fattr->valid &= ~NFS_ATTR_FATTR_ATIME;
> > +       }
> > +}
> > +
> >  void nfs_update_delegated_atime(struct inode *inode)
> >  {
> >         spin_lock(&inode->i_lock);
> > @@ -2163,6 +2203,9 @@ static int nfs_update_inode(struct inode
> > *inode, struct nfs_fattr *fattr)
> >          */
> >         nfsi->read_cache_jiffies = fattr->time_start;
> > 
> > +       /* Fix up any delegated attributes in the struct nfs_fattr
> > */
> > +       nfs_fattr_fixup_delegated(inode, fattr);
> > +
> >         save_cache_validity = nfsi->cache_validity;
> >         nfsi->cache_validity &= ~(NFS_INO_INVALID_ATTR
> >                         | NFS_INO_INVALID_ATIME
> > --
> > 2.45.2
> > 
> > 

-- 
Trond Myklebust
Linux NFS client maintainer, Hammerspace
trond.myklebust@hammerspace.com



  reply	other threads:[~2024-06-14 19:59 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-06-13  4:11 [PATCH 00/19] OPEN optimisations and Attribute delegations trondmy
2024-06-13  4:11 ` [PATCH 01/19] NFSv4: Clean up open delegation return structure trondmy
2024-06-13  4:11   ` [PATCH 02/19] NFSv4: Refactor nfs4_opendata_check_deleg() trondmy
2024-06-13  4:11     ` [PATCH 03/19] NFSv4: Add new attribute delegation definitions trondmy
2024-06-13  4:11       ` [PATCH 04/19] NFSv4: Plumb in XDR support for the new delegation-only setattr op trondmy
2024-06-13  4:11         ` [PATCH 05/19] NFSv4: Add CB_GETATTR support for delegated attributes trondmy
2024-06-13  4:11           ` [PATCH 06/19] NFSv4: Add a flags argument to the 'have_delegation' callback trondmy
2024-06-13  4:11             ` [PATCH 07/19] NFSv4: Add support for delegated atime and mtime attributes trondmy
2024-06-13  4:11               ` [PATCH 08/19] NFSv4: Add recovery of attribute delegations trondmy
2024-06-13  4:11                 ` [PATCH 09/19] NFSv4: Add a capability for delegated attributes trondmy
2024-06-13  4:11                   ` [PATCH 10/19] NFSv4: Enable attribute delegations trondmy
2024-06-13  4:11                     ` [PATCH 11/19] NFSv4: Delegreturn must set m/atime when they are delegated trondmy
2024-06-13  4:11                       ` [PATCH 12/19] NFSv4: Fix up delegated attributes in nfs_setattr trondmy
2024-06-13  4:11                         ` [PATCH 13/19] NFSv4: Don't request atime/mtime/size if they are delegated to us trondmy
2024-06-13  4:11                           ` [PATCH 14/19] NFSv4: Add support for the FATTR4_OPEN_ARGUMENTS attribute trondmy
2024-06-13  4:11                             ` [PATCH 15/19] NFSv4: Detect support for OPEN4_SHARE_ACCESS_WANT_OPEN_XOR_DELEGATION trondmy
2024-06-13  4:11                               ` [PATCH 16/19] NFSv4: Add support for OPEN4_RESULT_NO_OPEN_STATEID trondmy
2024-06-13  4:11                                 ` [PATCH 17/19] NFSv4: Ask for a delegation or an open stateid in OPEN trondmy
2024-06-13  4:11                                   ` [PATCH 18/19] NFS: Add a generic callback to return the delegation trondmy
2024-06-13  4:11                                     ` [PATCH 19/19] Return the delegation when deleting the sillyrenamed file trondmy
2024-06-14 16:32                         ` [PATCH 12/19] NFSv4: Fix up delegated attributes in nfs_setattr Anna Schumaker
2024-06-14 19:59                           ` Trond Myklebust [this message]
2024-06-15  0:25                             ` Trond Myklebust
2024-06-13 20:26               ` [PATCH 07/19] NFSv4: Add support for delegated atime and mtime attributes Anna Schumaker
2024-06-14 12:34 ` [PATCH 00/19] OPEN optimisations and Attribute delegations Jeff Layton
2024-06-15  6:27   ` Christoph Hellwig
2024-06-17  1:39     ` Trond Myklebust
2024-06-17  5:35       ` hch
2024-06-15  6:25 ` Christoph Hellwig
2024-06-17  1:28   ` Trond Myklebust
2024-06-17  5:37     ` hch

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=bddef727ec63731061a7c80f477f8dc112085b0a.camel@hammerspace.com \
    --to=trondmy@hammerspace.com \
    --cc=linux-nfs@vger.kernel.org \
    --cc=schumaker.anna@gmail.com \
    /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