lkml.org 
[lkml]   [2002]   [Jan]   [9]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
Subject[PATCH] : ir247_irlan_dynalloc-2.diff
From
ir247_irlan_dynalloc-2.diff :
---------------------------
o [CRITICA] Don't pass the same skb to two different state machines
<Following patch from Martin Diehl, modified by me>
o [CRITICA] Don't flag netdev as NETIF_F_DYNALLOC, because not kmalloc
diff -u -p -r linux/net/irda/irlan-d8/irlan_common.c linux/net/irda/irlan/irlan_common.c
--- linux/net/irda/irlan-d8/irlan_common.c Sun Sep 30 12:26:09 2001
+++ linux/net/irda/irlan/irlan_common.c Wed Jan 9 11:08:41 2002
@@ -317,8 +317,15 @@ void irlan_connect_indication(void *inst

del_timer(&self->watchdog_timer);

- irlan_do_provider_event(self, IRLAN_DATA_CONNECT_INDICATION, skb);
- irlan_do_client_event(self, IRLAN_DATA_CONNECT_INDICATION, skb);
+ /* If you want to pass the skb to *both* state machines, you will
+ * need to skb_clone() it, so that you don't free it twice.
+ * As the state machines don't need it, git rid of it here...
+ * Jean II */
+ if (skb)
+ dev_kfree_skb(skb);
+
+ irlan_do_provider_event(self, IRLAN_DATA_CONNECT_INDICATION, NULL);
+ irlan_do_client_event(self, IRLAN_DATA_CONNECT_INDICATION, NULL);

if (self->provider.access_type == ACCESS_PEER) {
/*
@@ -421,6 +428,13 @@ void irlan_disconnect_indication(void *i
break;
}

+ /* If you want to pass the skb to *both* state machines, you will
+ * need to skb_clone() it, so that you don't free it twice.
+ * As the state machines don't need it, git rid of it here...
+ * Jean II */
+ if (userdata)
+ dev_kfree_skb(userdata);
+
irlan_do_client_event(self, IRLAN_LMP_DISCONNECT, NULL);
irlan_do_provider_event(self, IRLAN_LMP_DISCONNECT, NULL);

diff -u -p -r linux/net/irda/irlan-d8/irlan_eth.c linux/net/irda/irlan/irlan_eth.c
--- linux/net/irda/irlan-d8/irlan_eth.c Wed Jan 9 11:00:53 2002
+++ linux/net/irda/irlan/irlan_eth.c Tue Jan 8 18:33:45 2002
@@ -61,7 +61,16 @@ int irlan_eth_init(struct net_device *de
dev->hard_start_xmit = irlan_eth_xmit;
dev->get_stats = irlan_eth_get_stats;
dev->set_multicast_list = irlan_eth_set_multicast_list;
- dev->features |= NETIF_F_DYNALLOC;
+
+ /* NETIF_F_DYNALLOC feature was set by irlan_eth_init() and would
+ * cause the unregister_netdev() to do asynch completion _and_
+ * kfree self->dev afterwards. Which is really bad because the
+ * netdevice was not allocated separately but is embedded in
+ * our control block and therefore gets freed with *self.
+ * The only reason why this would have been enabled is to hide
+ * some netdev refcount issues. If unregister_netdev() blocks
+ * forever, tell us about it... */
+ //dev->features |= NETIF_F_DYNALLOC;

ether_setup(dev);

-
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/
\
 
 \ /
  Last update: 2005-03-22 13:15    [from the cache]
©2003-2011 Jasper Spaans