[PATCH] : ir256_sock_connect_cli.diff

Jean Tourrilhes (jt@bougret.hpl.hp.com)
Tue, 5 Mar 2002 14:34:31 -0800


ir256_sock_connect_cli.diff :
---------------------------
o [CRITICA] Fix socket connect to remove dangerous cli()
<Tested on SMP>

--- linux/net/irda/af_irda.d4.c Mon Mar 4 14:54:38 2002
+++ linux/net/irda/af_irda.c Mon Mar 4 17:07:45 2002
@@ -1022,28 +1022,26 @@ static int irda_connect(struct socket *s
/* Now the loop */
if (sk->state != TCP_ESTABLISHED && (flags & O_NONBLOCK))
return -EINPROGRESS;
-
- cli(); /* To avoid races on the sleep */
-
- /* A Connect Ack with Choke or timeout or failed routing will go to
- * closed. */
+
+ /* Here, there is a race condition : the state may change between
+ * our test and the sleep, via irda_connect_confirm().
+ * The way to workaround that is to sleep with a timeout, so that
+ * we don't sleep forever and check the state when waking up.
+ * 50ms is plenty good enough, because the LAP is already connected.
+ * Jean II */
while (sk->state == TCP_SYN_SENT) {
- interruptible_sleep_on(sk->sleep);
+ interruptible_sleep_on_timeout(sk->sleep, HZ/20);
if (signal_pending(current)) {
- sti();
return -ERESTARTSYS;
}
}

if (sk->state != TCP_ESTABLISHED) {
- sti();
sock->state = SS_UNCONNECTED;
return sock_error(sk); /* Always set at this point */
}

sock->state = SS_CONNECTED;
-
- sti();

/* At this point, IrLMP has assigned our source address */
self->saddr = irttp_get_saddr(self->tsap);
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/