On Thu, Oct 25, 2012 at 11:14 AM, Pravin B Shelar <pshe...@nicira.com> wrote: > Tunnel caching was added to reduce CPU utilization on TX path > by caching packet header, So performance gain is directly proportional > to number of skbs transferred. But with help of offloads skb are getting > larger. So there are less number of skbs. Therefore header caching does > not shows similar gains we seen in past. And now kernel 3.6 has removed > dst caching from networking which makes header caching even more tricky. > So this commit removes header caching from OVS tunnelling. > > Signed-off-by: Pravin B Shelar <pshe...@nicira.com>
We should probably add a NEWS item. > diff --git a/datapath/tunnel.h b/datapath/tunnel.h > index 951a6f1..f296ecb 100644 > --- a/datapath/tunnel.h > +++ b/datapath/tunnel.h > - /* > - * Updates the cached header of a packet to match the actual packet > + * Updates the header of a packet to match the actual packet > * data. Typical things that might need to be updated are length, > * checksum, etc. The IP header will have already been updated and > this > * is the final step before transmission. Returns a linked list of Is this still really updating the header now? It's really creating it from scratch. > diff --git a/datapath/vport-gre.c b/datapath/vport-gre.c > index d02d4ec..3a877f3 100644 > --- a/datapath/vport-gre.c > +++ b/datapath/vport-gre.c > @@ -196,7 +162,6 @@ static struct sk_buff *gre_update_header(const struct > vport *vport, > * packet originally had DF set. > */ > skb->local_df = 1; > - __ip_select_ident(ip_hdr(skb), dst, 0); Why is this no longer needed? _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
