tipc: fix bug in multicast/broadcast message reassembly
Since commit 37e22164a8a3c39bdad45aa463b1e69a1fdf4110 ("tipc: rename and
move message reassembly function") reassembly of long broadcast messages
has been broken. This is because we test for a non-NULL return value
of the *buf parameter as criteria for succesful reassembly. However, this
parameter is left defined even after reception of the first fragment,
when reassebly is still incomplete. This leads to a kernel crash as soon
as a the first fragment of a long broadcast message is received.
We fix this with this commit, by implementing a stricter behavior of the
function and its return values.
This commit should be applied to both net and net-next.
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/net/tipc/msg.c b/net/tipc/msg.c
index 8be6e94..0a37a47 100644
--- a/net/tipc/msg.c
+++ b/net/tipc/msg.c
@@ -101,9 +101,11 @@
}
/* tipc_buf_append(): Append a buffer to the fragment list of another buffer
- * Let first buffer become head buffer
- * Returns 1 and sets *buf to headbuf if chain is complete, otherwise 0
- * Leaves headbuf pointer at NULL if failure
+ * @*headbuf: in: NULL for first frag, otherwise value returned from prev call
+ * out: set when successful non-complete reassembly, otherwise NULL
+ * @*buf: in: the buffer to append. Always defined
+ * out: head buf after sucessful complete reassembly, otherwise NULL
+ * Returns 1 when reassembly complete, otherwise 0
*/
int tipc_buf_append(struct sk_buff **headbuf, struct sk_buff **buf)
{
@@ -122,6 +124,7 @@
goto out_free;
head = *headbuf = frag;
skb_frag_list_init(head);
+ *buf = NULL;
return 0;
}
if (!head)
@@ -150,5 +153,7 @@
out_free:
pr_warn_ratelimited("Unable to build fragment list\n");
kfree_skb(*buf);
+ kfree_skb(*headbuf);
+ *buf = *headbuf = NULL;
return 0;
}