public inbox for linux-nfs@vger.kernel.org
 help / color / mirror / Atom feed
From: Mike Snitzer <snitzer@kernel.org>
To: Anna Schumaker <anna@kernel.org>,
	Trond Myklebust <trondmy@hammerspace.com>,
	Chuck Lever <chuck.lever@oracle.com>,
	Jeff Layton <jlayton@kernel.org>, NeilBrown <neil@brown.name>
Cc: linux-nfs@vger.kernel.org
Subject: [for-6.16-final PATCH 8/9] nfs/localio: avoid bouncing LOCALIO if nfs_client_is_local()
Date: Sun, 13 Jul 2025 23:13:58 -0400	[thread overview]
Message-ID: <20250714031359.10192-9-snitzer@kernel.org> (raw)
In-Reply-To: <20250714031359.10192-1-snitzer@kernel.org>

From: Mike Snitzer <snitzer@hammerspace.com>

Previously nfs_local_probe() was made to disable and then attempt to
re-enable LOCALIO (via LOCALIO protocol handshake) if/when it was
called and LOCALIO already enabled.

Vague memory for _why_ this was the case is that this was useful
if/when a local NFS server were to be restarted with a local NFS
client connected to it.

But as it happens this causes an absurd amount of LOCALIO flapping
which has a side-effect of too much IO being needlessly sent to NFSD
(using RPC over the loopback network interface).  This is the
definition of "serious performance loss" (that negates the point of
having LOCALIO).

So remove this mis-optimization for re-enabling LOCALIO if/when an NFS
server is restarted (which is an extremely rare thing to do).  Will
revisit testing that scenario again but in the meantime this patch
restores the full benefit of LOCALIO.

Fixes: 56bcd0f07fdb ("nfs: implement client support for NFS_LOCALIO_PROGRAM")
Signed-off-by: Mike Snitzer <snitzer@hammerspace.com>
---
 fs/nfs/localio.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/fs/nfs/localio.c b/fs/nfs/localio.c
index a4bacd9a5052..e3ae003118cb 100644
--- a/fs/nfs/localio.c
+++ b/fs/nfs/localio.c
@@ -180,10 +180,8 @@ static void nfs_local_probe(struct nfs_client *clp)
 		return;
 	}
 
-	if (nfs_client_is_local(clp)) {
-		/* If already enabled, disable and re-enable */
-		nfs_localio_disable_client(clp);
-	}
+	if (nfs_client_is_local(clp))
+		return;
 
 	if (!nfs_uuid_begin(&clp->cl_uuid))
 		return;
@@ -241,7 +239,8 @@ __nfs_local_open_fh(struct nfs_client *clp, const struct cred *cred,
 		case -ENOMEM:
 		case -ENXIO:
 		case -ENOENT:
-			/* Revalidate localio, will disable if unsupported */
+			/* Revalidate localio */
+			nfs_localio_disable_client(clp);
 			nfs_local_probe(clp);
 		}
 	}
-- 
2.44.0


  parent reply	other threads:[~2025-07-14  3:14 UTC|newest]

Thread overview: 56+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-05-09  0:46 [PATCH 0/6 v2] nfs_localio: fixes for races and errors from older compilers NeilBrown
2025-05-09  0:46 ` [PATCH 1/6] nfs_localio: use cmpxchg() to install new nfs_file_localio NeilBrown
2025-05-09  0:46 ` [PATCH 2/6] nfs_localio: always hold nfsd net ref with nfsd_file ref NeilBrown
2025-05-09  0:46 ` [PATCH 3/6] nfs_localio: simplify interface to nfsd for getting nfsd_file NeilBrown
2025-05-09  0:46 ` [PATCH 4/6] nfs_localio: duplicate nfs_close_local_fh() NeilBrown
2025-05-09  0:46 ` [PATCH 5/6] nfs_localio: protect race between nfs_uuid_put() and nfs_close_local_fh() NeilBrown
2025-05-09  0:46 ` [PATCH 6/6] nfs_localio: change nfsd_file_put_local() to take a pointer to __rcu pointer NeilBrown
2025-05-09 11:03   ` kernel test robot
2025-07-08 14:20   ` [RFC PATCH for 6.16-rcX] Revert "nfs_localio: change nfsd_file_put_local() to take a pointer to __rcu pointer" Mike Snitzer
2025-07-14  3:13     ` [for-6.16-final PATCH 0/9] NFSD/NFS/LOCALIO: stable fixes and revert 6.16 LOCALIO changes Mike Snitzer
2025-07-14  3:13       ` [for-6.16-final PATCH 1/9] Revert "NFSD: Clean up kdoc for nfsd_open_local_fh()" Mike Snitzer
2025-07-14  3:13       ` [for-6.16-final PATCH 2/9] Revert "nfs_localio: change nfsd_file_put_local() to take a pointer to __rcu pointer" Mike Snitzer
2025-07-14  3:13       ` [for-6.16-final PATCH 3/9] Revert "nfs_localio: protect race between nfs_uuid_put() and nfs_close_local_fh()" Mike Snitzer
2025-07-14  3:13       ` [for-6.16-final PATCH 4/9] Revert "nfs_localio: duplicate nfs_close_local_fh()" Mike Snitzer
2025-07-14  3:13       ` [for-6.16-final PATCH 5/9] Revert "nfs_localio: simplify interface to nfsd for getting nfsd_file" Mike Snitzer
2025-07-14  3:13       ` [for-6.16-final PATCH 6/9] Revert "nfs_localio: always hold nfsd net ref with nfsd_file ref" Mike Snitzer
2025-07-14  3:13       ` [for-6.16-final PATCH 7/9] Revert "nfs_localio: use cmpxchg() to install new nfs_file_localio" Mike Snitzer
2025-07-14  3:13       ` Mike Snitzer [this message]
2025-07-14  4:19         ` [for-6.16-final PATCH 8/9] nfs/localio: avoid bouncing LOCALIO if nfs_client_is_local() NeilBrown
2025-07-14 14:37           ` Mike Snitzer
2025-07-14 12:23         ` Jeff Layton
2025-07-14  3:13       ` [for-6.16-final PATCH 9/9] nfs/localio: add localio_async_probe modparm Mike Snitzer
2025-07-14  4:23         ` NeilBrown
2025-07-14 12:28           ` Jeff Layton
2025-07-14 14:08             ` Mike Snitzer
2025-07-14  3:50     ` [RFC PATCH for 6.16-rcX] Revert "nfs_localio: change nfsd_file_put_local() to take a pointer to __rcu pointer" NeilBrown
2025-07-14 14:45       ` Mike Snitzer
2025-07-15 22:52     ` [PATCH 0/3] Fix localio hangs Trond Myklebust
2025-07-15 22:52       ` [PATCH 1/3] NFS/localio: nfs_close_local_fh() fix check for file closed Trond Myklebust
2025-07-15 22:52       ` [PATCH 2/3] NFS/localio: nfs_uuid_put() fix the wait for file unlink events Trond Myklebust
2025-07-15 22:52       ` [PATCH 3/3] NFS/localio: nfs_uuid_put() fix the wake up after unlinking the file Trond Myklebust
2025-07-16  1:09       ` [PATCH 1/3] NFS/localio: nfs_close_local_fh() fix check for file closed NeilBrown
2025-07-16  1:22       ` [PATCH 2/3] NFS/localio: nfs_uuid_put() fix the wait for file unlink events NeilBrown
2025-07-16  2:29         ` Trond Myklebust
2025-07-16  3:51           ` NeilBrown
2025-07-16  1:31       ` [PATCH 3/3] NFS/localio: nfs_uuid_put() fix the wake up after unlinking the file NeilBrown
2025-07-16  4:17         ` Trond Myklebust
2025-07-16  5:07           ` NeilBrown
2025-07-16 15:19             ` Trond Myklebust
2025-07-16 15:59       ` [PATCH v2 0/3] Fix localio hangs Trond Myklebust
2025-07-16 15:59         ` [PATCH v2 1/3] NFS/localio: nfs_close_local_fh() fix check for file closed Trond Myklebust
2025-07-16 15:59         ` [PATCH v2 2/3] NFS/localio: nfs_uuid_put() fix races with nfs_open/close_local_fh() Trond Myklebust
2025-07-16 15:59         ` [PATCH v2 3/3] NFS/localio: nfs_uuid_put() fix the wake up after unlinking the file Trond Myklebust
2025-07-16 22:09         ` [PATCH v2 0/3] Fix localio hangs NeilBrown
2025-07-16 23:27           ` Mike Snitzer
2025-07-18  0:18             ` NeilBrown
2025-05-09 16:01 ` [PATCH 0/6 v2] nfs_localio: fixes for races and errors from older compilers Chuck Lever
2025-05-09 21:02   ` Mike Snitzer
2025-05-10  0:16     ` Paul E. McKenney
2025-05-10  2:44       ` NeilBrown
2025-05-10  3:01   ` NeilBrown
2025-05-10 16:02     ` Chuck Lever
2025-05-10 19:57       ` Mike Snitzer
2025-05-16 15:33         ` Chuck Lever
2025-05-18 10:46           ` Pali Rohár
2025-05-19  3:49         ` NeilBrown

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=20250714031359.10192-9-snitzer@kernel.org \
    --to=snitzer@kernel.org \
    --cc=anna@kernel.org \
    --cc=chuck.lever@oracle.com \
    --cc=jlayton@kernel.org \
    --cc=linux-nfs@vger.kernel.org \
    --cc=neil@brown.name \
    --cc=trondmy@hammerspace.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