On Apr 30, 2014, at 3:42 PM, Jesse Gross <je...@nicira.com> wrote:

> On Tue, Apr 29, 2014 at 3:24 PM, Pritesh Kothari
> <pritesh.koth...@cisco.com> wrote:
>> diff --git a/datapath/linux/compat/include/linux/skbuff.h 
>> b/datapath/linux/compat/include/linux/skbuff.h
>> index 714c955..de9b29d 100644
>> --- a/datapath/linux/compat/include/linux/skbuff.h
>> +++ b/datapath/linux/compat/include/linux/skbuff.h
>> +#ifndef HAVE_SKB_CLEAR_HASH
>> +static inline void skb_clear_hash(struct sk_buff *skb)
>> +{
>> +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,35)
>> +       skb->rxhash = 0;
>> +#endif
>> +}
>> +#endif
> 
> Thanks - this looks mostly good to me. One thing that I noticed is
> that I think the original version of this function has a bug in it (or
> more to the point, wasn't updated).

Thanks for the review.

> 
> I think that we also need to clear skb->l4_rxhash here since otherwise
> when we retrieve the hash the next time it won't be recomputed.

Yep had it mind, but forgot to add it.

> I think we can also use HAVE_RXHASH instead of the first version check
> and may want to do something similar for the l4_rxhash, which could be
> backported separately.

Done, sent v7, also sent backport patches for lts version mentioned faq.

Regards,
Pritesh

_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev

Reply via email to