Patchwork [Xen-devel,1/1] xen/netback: only non-freed SKB is queued into tx_queue

login
register
mail settings
Submitter annie.li@oracle.com
Date June 27, 2012, 10:46 a.m.
Message ID <1340794018-17274-1-git-send-email-annie.li@oracle.com>
Download mbox | patch
Permalink /patch/167611/
State Accepted
Delegated to: David Miller
Headers show

Comments

annie.li@oracle.com - June 27, 2012, 10:46 a.m.
From: Annie Li <Annie.li@oracle.com>

After SKB is queued into tx_queue, it will be freed if request_gop is NULL.
However, no dequeue action is called in this situation, it is likely that
tx_queue constains freed SKB. This patch should fix this issue, and it is
based on 3.5.0-rc4+.

This issue is found through code inspection, no bug is seen with it currently.
I run netperf test for several hours, and no network regression was found.

Signed-off-by: Annie Li <annie.li@oracle.com>
---
 drivers/net/xen-netback/netback.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)
David Miller - June 28, 2012, 11:55 p.m.
From: annie.li@oracle.com
Date: Wed, 27 Jun 2012 18:46:58 +0800

> From: Annie Li <Annie.li@oracle.com>
> 
> After SKB is queued into tx_queue, it will be freed if request_gop is NULL.
> However, no dequeue action is called in this situation, it is likely that
> tx_queue constains freed SKB. This patch should fix this issue, and it is
> based on 3.5.0-rc4+.
> 
> This issue is found through code inspection, no bug is seen with it currently.
> I run netperf test for several hours, and no network regression was found.
> 
> Signed-off-by: Annie Li <annie.li@oracle.com>

I lack the expertiece necessary to properly review this, so I really
need a Xen expert to look this over.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Ian Campbell - June 29, 2012, 7:23 a.m.
On Fri, 2012-06-29 at 00:55 +0100, David Miller wrote:
> From: annie.li@oracle.com
> Date: Wed, 27 Jun 2012 18:46:58 +0800
> 
> > From: Annie Li <Annie.li@oracle.com>
> > 
> > After SKB is queued into tx_queue, it will be freed if request_gop is NULL.
> > However, no dequeue action is called in this situation, it is likely that
> > tx_queue constains freed SKB. This patch should fix this issue, and it is
> > based on 3.5.0-rc4+.
> > 
> > This issue is found through code inspection, no bug is seen with it currently.
> > I run netperf test for several hours, and no network regression was found.
> > 
> > Signed-off-by: Annie Li <annie.li@oracle.com>
> 
> I lack the expertiece necessary to properly review this, so I really
> need a Xen expert to look this over.

Sorry, I put it to one side waiting for the repost to netdev and then
forgot about it...

Yes, this change looks good to me:

Acked-by: Ian Campbell <ian.campbell@citrix.com>

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
David Miller - June 29, 2012, 7:50 a.m.
From: Ian Campbell <Ian.Campbell@citrix.com>
Date: Fri, 29 Jun 2012 08:23:09 +0100

> On Fri, 2012-06-29 at 00:55 +0100, David Miller wrote:
>> From: annie.li@oracle.com
>> Date: Wed, 27 Jun 2012 18:46:58 +0800
>> 
>> > From: Annie Li <Annie.li@oracle.com>
>> > 
>> > After SKB is queued into tx_queue, it will be freed if request_gop is NULL.
>> > However, no dequeue action is called in this situation, it is likely that
>> > tx_queue constains freed SKB. This patch should fix this issue, and it is
>> > based on 3.5.0-rc4+.
>> > 
>> > This issue is found through code inspection, no bug is seen with it currently.
>> > I run netperf test for several hours, and no network regression was found.
>> > 
>> > Signed-off-by: Annie Li <annie.li@oracle.com>
>> 
>> I lack the expertiece necessary to properly review this, so I really
>> need a Xen expert to look this over.
> 
> Sorry, I put it to one side waiting for the repost to netdev and then
> forgot about it...
> 
> Yes, this change looks good to me:
> 
> Acked-by: Ian Campbell <ian.campbell@citrix.com>

Thanks, applied to net-next.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/drivers/net/xen-netback/netback.c b/drivers/net/xen-netback/netback.c
index f4a6fca..682633b 100644
--- a/drivers/net/xen-netback/netback.c
+++ b/drivers/net/xen-netback/netback.c
@@ -1363,8 +1363,6 @@  static unsigned xen_netbk_tx_build_gops(struct xen_netbk *netbk)
 					     INVALID_PENDING_IDX);
 		}
 
-		__skb_queue_tail(&netbk->tx_queue, skb);
-
 		netbk->pending_cons++;
 
 		request_gop = xen_netbk_get_requests(netbk, vif,
@@ -1376,6 +1374,8 @@  static unsigned xen_netbk_tx_build_gops(struct xen_netbk *netbk)
 		}
 		gop = request_gop;
 
+		__skb_queue_tail(&netbk->tx_queue, skb);
+
 		vif->tx.req_cons = idx;
 		xen_netbk_check_rx_xenvif(vif);