Skip to contentSkip to conversation

Add missing socket check in can/bcm release.

PATCH Dave Jones 2 participants 4 messages Apr 20, 2011

Subsystems (3)
lore
Share & tools
lore

Dave Jones

author
#0 lore raw

We can get here with a NULL socket argument passed from userspace, so we need to handle it accordingly.

net/can/bcm.c +6 -1
index 57b1aed..8a6a05e 100644
@@ -1427,9 +1427,14 @@ static int bcm_init(struct sock *sk)
 static int bcm_release(struct socket *sock)
 {
 	struct sock *sk = sock->sk;
-	struct bcm_sock *bo = bcm_sk(sk);
+	struct bcm_sock *bo;
 	struct bcm_op *op, *next;
 
+	if (sk == NULL)
+		return 0;
+
+	bo = bcm_sk(sk);
+
 	/* remove bcm_ops, timer, rx_unregister(), etc. */
 
 	unregister_netdevice_notifier(&bo->notifier);
author
#2 lore raw
Quoted 9 lines · David Miller

From: Dave Jones <[email protected]>
Date: Tue, 19 Apr 2011 23:30:01 -0400

Quoted 4 lines

We can get here with a NULL socket argument passed from userspace, so we need to handle it accordingly.

Applied and queued up for -stable, thanks Dave.

Out of curiousity, while I was asleep it occured to me.. is it ever valid for a ->release to get passed a NULL socket->sk ? I'm wondering if we can't do this check a layer up in sock_release, in case future protocols reintroduce the same bug.

Quoted 1 line

From a quick look, almost every protocol has this check in its ->release.

Though it seems some do something different instead of using socket->sk, so it would be a pointless check for some of the lesser used ones.

thoughts?

Dave

maintainer
#3 lore raw

From: Dave Jones <[email protected]>
Date: Wed, 20 Apr 2011 12:03:50 -0400

Quoted 7 lines · David Miller
Quoted 6 lines

From: Dave Jones <[email protected]>
Date: Tue, 19 Apr 2011 23:30:01 -0400

Quoted 3 lines

We can get here with a NULL socket argument passed from userspace, so we need to handle it accordingly.

Quoted 6 lines
Quoted 3 lines
Quoted 1 line

Applied and queued up for -stable, thanks Dave.

Out of curiousity, while I was asleep it occured to me.. is it ever valid for a ->release to get passed a NULL socket->sk ?

Yes, it happens all the time.

If accept() fails mid-stream, we'll have an 'sk' that hasn't been hooked up to ->socket yet, but we still have to release the 'sk' in the error handling.

See also commit c100c8f4c3c6f2a407bdbaaad2c4f1062e6a473a, which fixes a bug triggered via the same code path.