From: Trond Myklebust <trondmy@kernel.org>
To: Michael Stoler <michael.stoler@vastdata.com>, linux-nfs@vger.kernel.org
Cc: Dan Aloni <dan.aloni@vastdata.com>
Subject: Re: Spurious -EEXIST from NFSv3
Date: Wed, 19 Nov 2025 08:48:50 -0500 [thread overview]
Message-ID: <53bc46637bdc4b267a318c74fb4c97cb382f29d1.camel@kernel.org> (raw)
In-Reply-To: <CAGztG2BdRW8fy9iF5u0iJmMoXrc2G0NQTt8jwk12Q=Q+e0FaLQ@mail.gmail.com>
On Wed, 2025-11-19 at 12:52 +0200, Michael Stoler wrote:
>
>
> You don't often get email from michael.stoler@vastdata.com.
> Learn why this is important
>
>
>
>
>
>
> I’m having an issue with an NFS driver based on Linux 5.15.147.
> The function nfs_verifier_is_delegated() spuriously returns true for
> NFSv3 file inside nfs_do_lookup_revalidate(). This causes the
> d_revalidate method to return true for a removed file, which in turn
> leads to an -EEXIST error during exclusive creation of a non-existent
> file.
>
> It appears that the root cause is an initialization races or an
> uninitialized d_time value. The attached patch resolves the issue,
> but is there a more graceful or proper solution?
>
That patch is incorrect. The verifier must be initialised for all
visible dentries, irrespective of the NFS version. Let me see if I can
come up with something.
--
Trond Myklebust
Linux NFS client maintainer, Hammerspace
trondmy@kernel.org, trond.myklebust@hammerspace.com
next prev parent reply other threads:[~2025-11-19 13:48 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-11-19 10:52 Spurious -EEXIST from NFSv3 Michael Stoler
2025-11-19 13:48 ` Trond Myklebust [this message]
2025-11-19 13:58 ` [PATCH 0/3] Fix verifier initialisation races for dentries Trond Myklebust
2025-11-19 13:58 ` [PATCH 1/3] NFS: Initialise verifiers for visible dentries in readdir and lookup Trond Myklebust
2025-11-19 13:58 ` [PATCH 2/3] NFS: Initialise verifiers for visible dentries in nfs_atomic_open() Trond Myklebust
2025-11-19 13:58 ` [PATCH 3/3] NFS: Initialise verifiers for visible dentries in _nfs4_open_and_get_state Trond Myklebust
2025-11-19 16:14 ` Anna Schumaker
2025-11-19 16:50 ` Trond Myklebust
2025-11-20 10:47 ` kernel test robot
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=53bc46637bdc4b267a318c74fb4c97cb382f29d1.camel@kernel.org \
--to=trondmy@kernel.org \
--cc=dan.aloni@vastdata.com \
--cc=linux-nfs@vger.kernel.org \
--cc=michael.stoler@vastdata.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.