Thanks, it looks like a real issue to me. I sent out a patch that should fix it.
On Fri, Jun 21, 2013 at 1:19 AM, Joe Stringer <j...@wand.net.nz> wrote:
> I get a warning in gso.c after this patch:
>
> datapath/linux/gso.c: In function ‘rpl_ip_local_out’:
> datapath/linux/gso.c:134:174: warning: ‘id’ may be used uninitialized in
> this function [-Wuninitialized]
>
> Could you take a look at this?
>
>
> On Fri, Jun 21, 2013 at 3:14 PM, Pravin Shelar <pshe...@nicira.com> wrote:
>>
>> On Thu, Jun 20, 2013 at 5:50 PM, Jesse Gross <je...@nicira.com> wrote:
>> > On Thu, Jun 20, 2013 at 5:19 PM, Pravin B Shelar <pshe...@nicira.com>
>> > wrote:
>> >> diff --git a/datapath/linux/compat/skbuff-openvswitch.c
>> >> b/datapath/linux/compat/skbuff-openvswitch.c
>> >> index ef43ba9..720bc18 100644
>> >> --- a/datapath/linux/compat/skbuff-openvswitch.c
>> >> +++ b/datapath/linux/compat/skbuff-openvswitch.c
>> >> +#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,19)
>> >> +int skb_checksum_help(struct sk_buff *skb, int inward)
>> >> +#else
>> >> +int skb_checksum_help(struct sk_buff *skb)
>> >> +#endif
>> >> +{
>> >> + /* Use default features for dst device. */
>> >> + if (unlikely(skb_is_nonlinear(skb))) {
>> >> + int err;
>> >
>> > I think the above comment no longer applies since we're not pulling
>> > any device features.
>> >
>> >> diff --git a/datapath/vport-gre.c b/datapath/vport-gre.c
>> >> index dae35ac..fda6481 100644
>> >> --- a/datapath/vport-gre.c
>> >> +++ b/datapath/vport-gre.c
>> >> +static int __send(struct vport *vport, struct sk_buff *skb,
>> >> + int tunnel_hlen,
>> >> + __be32 seq, __be16 gre64_flag)
>> > [...]
>> >> - ovs_tnl_rcv(vport, skb, &tun_key);
>> >> - return 0;
>> >> + /* Push Tunnel header. */
>> >> + skb = __build_header(skb, tunnel_hlen, seq, gre64_flag);
>> >> + if (unlikely(!skb)) {
>> >> + err = 0;
>> >> + goto err_free_rt;
>> >> + }
>> >>
>> >> + df = OVS_CB(skb)->tun_key->tun_flags & TUNNEL_DONT_FRAGMENT ?
>> >> + htons(IP_DF) : 0;
>> >> +
>> >> + /*
>> >> + * Allow our local IP stack to fragment the outer packet even
>> >> + * if the DF bit is set as a last resort. We also need to
>> >> + * force selection of an IP ID here because Linux will
>> >> + * otherwise leave it at 0 if the packet originally had DF set.
>> >> + */
>> >
>> > I believe this comment was deleted/moved in the upstream version.
>> >
>> > Those are both minor though so
>> > Acked-by: Jesse Gross <je...@nicira.com>
>>
>> Thanks,
>> I pushed code with above changes and compat checksum check in
>> iptunnel_pull_header().
>> _______________________________________________
>> dev mailing list
>> dev@openvswitch.org
>> http://openvswitch.org/mailman/listinfo/dev
>
>
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
