2006-01-19 18:37:39

by Ed L. Cashin

Subject: [PATCH 2.6.15-git9] aoe [1/8]: zero packet data after skb allocation

These eight patches supercede the seven aoe patches sent on January
third for the 2.6.15-rc7 kernel. A bug existed in the Jan. 3 set of
patches where the retransmit timer for an AoE device AoE timer could
be added twice. That bug has been fixed in this set of patches.


Signed-off-by: "Ed L. Cashin" <[email protected]>

Zero the data in new socket buffers to prevent leaking information.

diff -upr 2.6.15-git9-orig/drivers/block/aoe/aoecmd.c 2.6.15-git9-aoe/drivers/block/aoe/aoecmd.c
--- 2.6.15-git9-orig/drivers/block/aoe/aoecmd.c 2006-01-18 13:22:49.000000000 -0500
+++ 2.6.15-git9-aoe/drivers/block/aoe/aoecmd.c 2006-01-18 13:48:53.000000000 -0500
@@ -28,6 +28,7 @@ new_skb(struct net_device *if_dev, ulong
skb->protocol = __constant_htons(ETH_P_AOE);
skb->priority = 0;
skb_put(skb, len);
+ memset(skb->head, 0, len);
skb->next = skb->prev = NULL;

/* tell the network layer not to perform IP checksums


--
"Ed L. Cashin" <[email protected]>

2006-01-19 22:43:04

by Greg KH

Subject: Re: [PATCH 2.6.15-git9] aoe [1/8]: zero packet data after skb allocation

On Thu, Jan 19, 2006 at 12:37:24PM -0500, Ed L. Cashin wrote:
> These eight patches supercede the seven aoe patches sent on January
> third for the 2.6.15-rc7 kernel. A bug existed in the Jan. 3 set of
> patches where the retransmit timer for an AoE device AoE timer could
> be added twice. That bug has been fixed in this set of patches.

Ok, now I got them all, sorry about that, slow email system here...

greg k-h