mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Shigeru Yoshida <syoshida@redhat.com>
To: Trond Myklebust <trondmy@kernel.org>,
	Anna Schumaker <anna@kernel.org>, Chuck Lever <cel@kernel.org>,
	Jeff Layton <jlayton@kernel.org>, NeilBrown <neil@brown.name>,
	Olga Kornievskaia <okorniev@redhat.com>,
	Dai Ngo <Dai.Ngo@oracle.com>, Tom Talpey <tom@talpey.com>,
	"David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
	Simon Horman <horms@kernel.org>, felix <fuzhen5@huawei.com>
Cc: Shigeru Yoshida <syoshida@redhat.com>,
	linux-nfs@vger.kernel.org, netdev@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: [PATCH] SUNRPC: update clnt->pipefs_sb on pipefs MOUNT/UMOUNT events
Date: Mon, 17 Aug 2026 17:00:12 +0900	[thread overview]
Message-ID: <20260817080013.564939-1-syoshida@redhat.com> (raw)

Commit bfca5fb4e97c ("SUNRPC: Fix RPC client cleaned up the freed
pipefs dentries") added a clnt->pipefs_sb field and a consistency
check in rpc_clnt_remove_pipedir() so that pipedir cleanup is skipped
when the current superblock does not match the one recorded in the
client.

However, __rpc_clnt_handle_event() never updates clnt->pipefs_sb when
it creates or removes pipedirs in response to RPC_PIPEFS_MOUNT and
RPC_PIPEFS_UMOUNT notifications.

This causes a problem for RPC clients that are registered before
rpc_pipefs is mounted (e.g., before rpc_pipefs is mounted during boot,
or after it has been unmounted). Such clients have clnt->pipefs_sb set
to NULL at registration time because rpc_get_sb_net() returns NULL.
When rpc_pipefs is mounted later, __rpc_clnt_handle_event(MOUNT)
creates the pipedir but leaves clnt->pipefs_sb as NULL. When the
client is later destroyed, rpc_clnt_remove_pipedir() sees that the
current superblock does not match clnt->pipefs_sb and skips cleanup,
leaving orphaned dentries whose RPC_I(inode)->private pointers
reference freed rpc_clnt memory. Any subsequent access to these
orphaned info files triggers a use-after-free [1].

Fix this by setting clnt->pipefs_sb in __rpc_clnt_handle_event() on
MOUNT and clearing it on UMOUNT, mirroring what rpc_setup_pipedir()
already does at client registration time.

[1]
BUG: KASAN: slab-use-after-free in rpc_info_open (net/sunrpc/rpc_pipe.c:426)
Read of size 4 at addr ffff88802ebe0800 by task cat/5079

Call Trace:
 <TASK>
 ...
 rpc_info_open (net/sunrpc/rpc_pipe.c:426)
 do_dentry_open (fs/open.c:947)
 vfs_open (fs/open.c:1052)
 path_openat (fs/namei.c:4700 fs/namei.c:4863)
 do_file_open (fs/namei.c:4892)
 do_sys_openat2 (fs/open.c:1368)
 __x64_sys_openat (fs/open.c:1385)
 do_syscall_64 (arch/x86/entry/syscall_64.c:94)
 entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:121)
 </TASK>

Allocated by task 5048:
 ...
 rpc_new_client (net/sunrpc/clnt.c:377)
 __rpc_clone_client (net/sunrpc/clnt.c:645)
 rpc_clone_client_set_auth (net/sunrpc/clnt.c:706)
 nfs_init_server_rpcclient (fs/nfs/client.c:640)
 nfs_clone_server (fs/nfs/client.c:1247)
 nfs4_clone_server (fs/nfs/nfs4proc.c:10683)
 nfs_do_submount (fs/nfs/namespace.c:288)
 ...

Freed by task 24:
 ...
 kfree (mm/slub.c:6692)
 rpc_free_client_work (net/sunrpc/clnt.c:987)
 process_scheduled_works (kernel/workqueue.c:3405)
 worker_thread (kernel/workqueue.c:3486)
 ...

Fixes: bfca5fb4e97c ("SUNRPC: Fix RPC client cleaned up the freed pipefs dentries")
Assisted-by: Claude:claude-opus-4-6
Signed-off-by: Shigeru Yoshida <syoshida@redhat.com>
---
 net/sunrpc/clnt.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index efa26899bc7d..55617d263d2f 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -179,11 +179,17 @@ static int rpc_clnt_skip_event(struct rpc_clnt *clnt, unsigned long event)
 static int __rpc_clnt_handle_event(struct rpc_clnt *clnt, unsigned long event,
 				   struct super_block *sb)
 {
+	int err;
+
 	switch (event) {
 	case RPC_PIPEFS_MOUNT:
-		return rpc_setup_pipedir_sb(sb, clnt);
+		err = rpc_setup_pipedir_sb(sb, clnt);
+		if (!err)
+			clnt->pipefs_sb = sb;
+		return err;
 	case RPC_PIPEFS_UMOUNT:
 		__rpc_clnt_remove_pipedir(clnt);
+		clnt->pipefs_sb = NULL;
 		break;
 	default:
 		printk(KERN_ERR "%s: unknown event: %ld\n", __func__, event);
-- 
2.55.0


             reply	other threads:[~2026-08-17  8:00 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-08-17  8:00 Shigeru Yoshida [this message]
2026-08-17 13:31 ` Jeff Layton

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=20260817080013.564939-1-syoshida@redhat.com \
    --to=syoshida@redhat.com \
    --cc=Dai.Ngo@oracle.com \
    --cc=anna@kernel.org \
    --cc=cel@kernel.org \
    --cc=davem@davemloft.net \
    --cc=edumazet@google.com \
    --cc=fuzhen5@huawei.com \
    --cc=horms@kernel.org \
    --cc=jlayton@kernel.org \
    --cc=kuba@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-nfs@vger.kernel.org \
    --cc=neil@brown.name \
    --cc=netdev@vger.kernel.org \
    --cc=okorniev@redhat.com \
    --cc=pabeni@redhat.com \
    --cc=tom@talpey.com \
    --cc=trondmy@kernel.org \
    /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

all inboxes | Powered by JetHome®