Linux NFS development
 help / color / mirror / Atom feed
From: Jeff Layton <jlayton@kernel.org>
To: Chuck Lever <chuck.lever@oracle.com>, NeilBrown <neil@brown.name>,
	 Olga Kornievskaia <okorniev@redhat.com>,
	Dai Ngo <Dai.Ngo@oracle.com>,  Tom Talpey <tom@talpey.com>,
	"J. Bruce Fields" <bfields@fieldses.org>,
	 Scott Mayhew <smayhew@redhat.com>,
	 Trond Myklebust <Trond.Myklebust@netapp.com>,
	 Andreas Gruenbacher <agruen@suse.de>,
	Mike Snitzer <snitzer@kernel.org>,
	 Rick Macklem <rmacklem@uoguelph.ca>
Cc: Chris Mason <clm@meta.com>,
	linux-nfs@vger.kernel.org,  linux-kernel@vger.kernel.org,
	Jeff Layton <jlayton@kernel.org>
Subject: [PATCH 08/10] nfsd: fix partial-write detection in nfsd_direct_write
Date: Thu, 28 May 2026 17:55:19 -0400	[thread overview]
Message-ID: <20260528-nfsd-fixes-v1-8-e78708eff77d@kernel.org> (raw)
In-Reply-To: <20260528-nfsd-fixes-v1-0-e78708eff77d@kernel.org>

From: Chris Mason <clm@meta.com>

nfsd_direct_write() walks a list of write segments and, after each
vfs_iocb_iter_write(), tries to detect a short write so the loop can
stop before placing the next segment at a wrong file offset:

    host_err = vfs_iocb_iter_write(file, kiocb, &segments[i].iter);
    if (host_err < 0)
            return host_err;
    *cnt += host_err;
    if (host_err < segments[i].iter.count)
            break;	/* partial write */

vfs_iocb_iter_write() runs the iter through ->write_iter(), which
advances the iter by the number of bytes written. By the time the
check runs, segments[i].iter.count is the residual, not the original
request length:

    before write_iter: iter.count == original_len
    after  write_iter: iter.count == original_len - host_err

The condition then reduces to host_err < original_len - host_err, so
the break fires only when less than half of the segment was written.
Any short write completing between 50% and 99% of the segment slips
through; the loop advances to the next segment with kiocb->ki_pos
only bumped by the short amount, writing the next segment's payload
at the wrong offset and over-reporting *cnt to the NFS client.

Snapshot the segment's byte count before the write and compare
host_err against that snapshot so any short write breaks the loop.

Fixes: 06c5c97293e3 ("NFSD: Implement NFSD_IO_DIRECT for NFS WRITE")
Assisted-by: kres:claude-opus-4-7
Signed-off-by: Chris Mason <clm@meta.com>
---
 fs/nfsd/vfs.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
index 980217f755b7..619f252af4d1 100644
--- a/fs/nfsd/vfs.c
+++ b/fs/nfsd/vfs.c
@@ -1380,6 +1380,7 @@ nfsd_direct_write(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	struct file *file = nf->nf_file;
 	unsigned int nsegs, i;
 	ssize_t host_err;
+	size_t expected;
 
 	nsegs = nfsd_write_dio_iters_init(nf, rqstp->rq_bvec, nvecs,
 					  kiocb, *cnt, segments);
@@ -1401,11 +1402,13 @@ nfsd_direct_write(struct svc_rqst *rqstp, struct svc_fh *fhp,
 				kiocb->ki_flags |= IOCB_DONTCACHE;
 		}
 
+		expected = iov_iter_count(&segments[i].iter);
+
 		host_err = vfs_iocb_iter_write(file, kiocb, &segments[i].iter);
 		if (host_err < 0)
 			return host_err;
 		*cnt += host_err;
-		if (host_err < segments[i].iter.count)
+		if (host_err < (ssize_t)expected)
 			break;	/* partial write */
 	}
 

-- 
2.54.0


  parent reply	other threads:[~2026-05-28 21:55 UTC|newest]

Thread overview: 40+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-05-28 21:55 [PATCH 00/10] nfsd: a pile of fixes for random bugs Jeff Layton
2026-05-28 21:55 ` [PATCH 01/10] nfsd: fix BUG_ON in nfsd4_alloc_layout_stateid on racing delegation revoke Jeff Layton
2026-05-28 23:40   ` NeilBrown
2026-05-29 14:44     ` Jeff Layton
2026-05-28 21:55 ` [PATCH 02/10] nfsd: drain callbacks and clear cl_cb_session Jeff Layton
2026-05-29 15:13   ` Chuck Lever
2026-05-29 17:31     ` Jeff Layton
2026-05-28 21:55 ` [PATCH 03/10] nfsd: serialize nfsd4_end_grace() with atomic test-and-set Jeff Layton
2026-05-29 15:38   ` Chuck Lever
2026-05-29 15:57     ` Jeff Layton
2026-05-29 16:05       ` Chuck Lever
2026-05-29 17:02         ` Jeff Layton
2026-05-28 21:55 ` [PATCH 04/10] nfsd: dedup nfs4_client_to_reclaim inserts Jeff Layton
2026-05-29 16:22   ` Chuck Lever
2026-05-28 21:55 ` [PATCH 05/10] nfsd: gate nfs3 setacl by argp->mask Jeff Layton
2026-05-28 21:55 ` [PATCH 06/10] NFSD: Enable return of an updated stable_how to NFS clients Jeff Layton
2026-05-29 10:56   ` Jeff Layton
2026-05-30  7:58   ` NFSv4.1 COMMIT of all changed areas only on flush? " Cedric Blancher
2026-05-30 10:24     ` Jeff Layton
2026-05-28 21:55 ` [PATCH 07/10] NFSD: check truncate permission under inode lock Jeff Layton
2026-05-28 21:55 ` Jeff Layton [this message]
2026-05-29 16:57   ` [PATCH 08/10] nfsd: fix partial-write detection in nfsd_direct_write Chuck Lever
2026-05-29 17:01     ` Jeff Layton
2026-05-29 17:03       ` Chuck Lever
2026-05-29 17:06         ` Jeff Layton
2026-05-29 17:09           ` Chuck Lever
2026-05-28 21:55 ` [PATCH 09/10] nfsd: cap decoded POSIX ACL count to bound sort cost Jeff Layton
2026-05-28 22:11   ` Rick Macklem
2026-05-28 23:11     ` Chuck Lever
2026-05-29  0:07       ` Chuck Lever
2026-05-29 10:48         ` Jeff Layton
2026-05-29 13:20           ` Chuck Lever
2026-05-29  7:34   ` Cedric Blancher
2026-05-29 10:50     ` Jeff Layton
2026-05-29 18:34   ` Chuck Lever
2026-05-29 18:41     ` Jeff Layton
2026-05-29 18:48       ` Chuck Lever
2026-05-29 23:04     ` Rick Macklem
2026-05-28 21:55 ` [PATCH 10/10] nfsd: validate symlink target length in NFSv4 CREATE Jeff Layton
2026-05-29 18:55   ` Chuck Lever

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=20260528-nfsd-fixes-v1-8-e78708eff77d@kernel.org \
    --to=jlayton@kernel.org \
    --cc=Dai.Ngo@oracle.com \
    --cc=Trond.Myklebust@netapp.com \
    --cc=agruen@suse.de \
    --cc=bfields@fieldses.org \
    --cc=chuck.lever@oracle.com \
    --cc=clm@meta.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-nfs@vger.kernel.org \
    --cc=neil@brown.name \
    --cc=okorniev@redhat.com \
    --cc=rmacklem@uoguelph.ca \
    --cc=smayhew@redhat.com \
    --cc=snitzer@kernel.org \
    --cc=tom@talpey.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