linux-kernel - [PATCH] SUNRPC/xprt: Use del_timer_sync() instead of del_singleshot_timer

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <20221028170505.44d779c3@gandalf.local.home>
Date:   Fri, 28 Oct 2022 17:05:05 -0400
From:   Steven Rostedt <rostedt@...dmis.org>
To:     LKML <linux-kernel@...r.kernel.org>
Cc:     Trond Myklebust <trond.myklebust@...merspace.com>,
        Anna Schumaker <anna@...nel.org>,
        Chuck Lever <chuck.lever@...cle.com>,
        Jeff Layton <jlayton@...nel.org>,
        "David S. Miller" <davem@...emloft.net>,
        Eric Dumazet <edumazet@...gle.com>,
        Jakub Kicinski <kuba@...nel.org>,
        Paolo Abeni <pabeni@...hat.com>, linux-nfs@...r.kernel.org,
        netdev <netdev@...r.kernel.org>
Subject: [PATCH] SUNRPC/xprt: Use del_timer_sync() instead of
 del_singleshot_timer_sync()

From: "Steven Rostedt (Google)" <rostedt@...dmis.org>

Back on June 22, 2005, it was decided to use del_singleshot_timer_sync()
because it wouldn't loop like del_timer_sync(), and since the timer that
was being removed was not about to be rearmed, it was considered an
efficiency to use del_singleshot_timer_sync() over del_timer_sync().

But on June 23, 2005, commit 55c888d6d09a0 ("timers fixes/improvements")
happened, which converted del_singleshot_timer_sync() into:

 #define del_singleshot_timer_sync(t) del_timer_sync(t)

Making the two equivalent.

Now work is being done to add a "shutdown" state to timers where a timer
must be in that state in order to be freed to prevent use-after-free bugs
caused by timers being re-armed just before being freed, the
del_singleshot_timer_sync() is now being converted into something that
will set the timer to the shutdown state. This means that once
del_singleshot_timer_sync() is called, the timer can no longer be
re-armed.

As the timer here will be re-armed, it can not use del_singleshot_timer_sync().
But as the reason it was used in the first place no longer exists, just
use del_timer_sync().

Link: https://lore.kernel.org/lkml/20221028145005.28bc324d@gandalf.local.home/

Fixes: 0f9dc2b16884b ("RPC: Clean up socket autodisconnect")
Signed-off-by: Steven Rostedt (Google) <rostedt@...dmis.org>
---
 net/sunrpc/xprt.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index 656cec208371..ab453ede54f0 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -1164,7 +1164,7 @@ xprt_request_enqueue_receive(struct rpc_task *task)
 	spin_unlock(&xprt->queue_lock);

 	/* Turn off autodisconnect */
-	del_singleshot_timer_sync(&xprt->timer);
+	del_timer_sync(&xprt->timer);
 	return 0;
 }

-- 
2.35.1