af_iucv: wrong mapping of sent and confirmed skbs
authorUrsula Braun <ursula.braun@de.ibm.com>
Tue, 13 May 2014 12:38:02 +0000 (14:38 +0200)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 1 Jul 2014 03:11:51 +0000 (20:11 -0700)
commit f5738e2ef88070ef1372e6e718124d88e9abe4ac upstream.

When sending data through IUCV a MESSAGE COMPLETE interrupt
signals that sent data memory can be freed or reused again.
With commit f9c41a62bba3f3f7ef3541b2a025e3371bcbba97
"af_iucv: fix recvmsg by replacing skb_pull() function" the
MESSAGE COMPLETE callback iucv_callback_txdone() identifies
the wrong skb as being confirmed, which leads to data corruption.
This patch fixes the skb mapping logic in iucv_callback_txdone().

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
net/iucv/af_iucv.c

index c4b7218..1465363 100644 (file)
@@ -1829,7 +1829,7 @@ static void iucv_callback_txdone(struct iucv_path *path,
                spin_lock_irqsave(&list->lock, flags);
 
                while (list_skb != (struct sk_buff *)list) {
-                       if (msg->tag != IUCV_SKB_CB(list_skb)->tag) {
+                       if (msg->tag == IUCV_SKB_CB(list_skb)->tag) {
                                this = list_skb;
                                break;
                        }