On Thu, Mar 26, 2015 at 6:35 PM, Pravin Shelar <pshe...@nicira.com> wrote:
> On Tue, Mar 24, 2015 at 12:58 PM, Jesse Gross <je...@nicira.com> wrote:
>> On Tue, Mar 24, 2015 at 1:21 PM, Pravin Shelar <pshe...@nicira.com> wrote:
>>> On Mon, Mar 23, 2015 at 12:23 PM, Jesse Gross <je...@nicira.com> wrote:
>>>> On Mon, Mar 9, 2015 at 3:12 PM, Pravin B Shelar <pshe...@nicira.com> wrote:
>>>>> +static int skb_list_xmit(struct rtable *rt, struct sk_buff *skb, __be32
>>>>> src,
>>>>> + __be32 dst, __u8 tos, __u8 ttl, __be16 df)
>>>>> +{
>>>>> + int len = 0;
>>>>> +
>>>>> + while (skb) {
>>>>> + struct sk_buff *next = skb->next;
>>>>> +
>>>>> + if (next)
>>>>> + dst_clone(&rt->dst);
>>>>> +
>>>>> + skb->next = NULL;
>>>>> + len += iptunnel_xmit(NULL, rt, skb, src, dst, IPPROTO_TCP,
>>>>> + tos, ttl, df, false);
>>>>
>>>> I think there may be a problem in our version of ip_local_out() if it
>>>> thinks that fix_segment is set in the CB, so we need to find a way to
>>>> make sure that it is cleared.
>>>>
>>>
>>> right, I will fix STT and I will send another patch for other vport types.
>>
>> I think the existing vports should be OK because their handle offloads
>> compat code already knows how to initialize this.
>>
> I was thinking when compat handle offload is not used, but ovs calls
> compat ip_local_out(). for example on newer kernels.
I think that OVS will use compat handle_offloads in all cases where
compat ip_local_out() is used for this reason. But if you spot a bug
then we should fix it :)
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
