2014-07-11 08:48:36

by Hayes Wang

Subject: [PATCH net] r8152: fix r8152_csum_workaround function

The transport offset of the IPv4 packet should be fixed and wouldn't
be out of the hw limitation, so the r8152_csum_workaround() should
be used for IPv6 packets.

Signed-off-by: Hayes Wang <[email protected]>
---
drivers/net/usb/r8152.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c
index a795ecf..7bad2d3 100644
--- a/drivers/net/usb/r8152.c
+++ b/drivers/net/usb/r8152.c
@@ -1359,7 +1359,7 @@ static void r8152_csum_workaround(struct r8152 *tp, struct sk_buff *skb,
struct sk_buff_head seg_list;
struct sk_buff *segs, *nskb;

- features &= ~(NETIF_F_IP_CSUM | NETIF_F_SG | NETIF_F_TSO);
+ features &= ~(NETIF_F_SG | NETIF_F_IPV6_CSUM | NETIF_F_TSO6);
segs = skb_gso_segment(skb, features);
if (IS_ERR(segs) || !segs)
goto drop;
--
1.9.3

2014-07-11 21:52:36

by David Miller

Subject: Re: [PATCH net] r8152: fix r8152_csum_workaround function

From: Hayes Wang <[email protected]>
Date: Fri, 11 Jul 2014 16:48:27 +0800

> The transport offset of the IPv4 packet should be fixed and wouldn't
> be out of the hw limitation, so the r8152_csum_workaround() should
> be used for IPv6 packets.
>
> Signed-off-by: Hayes Wang <[email protected]>

Applied and queued up for -stable, but this patch was harder to review
than it needed to be...

> - features &= ~(NETIF_F_IP_CSUM | NETIF_F_SG | NETIF_F_TSO);
> + features &= ~(NETIF_F_SG | NETIF_F_IPV6_CSUM | NETIF_F_TSO6);

because you changed the order of the flags unnecessarily.