On Sat, Nov 17, 2012 at 11:28 AM, Ansis Atteka <aatt...@nicira.com> wrote: > diff --git a/lib/flow.c b/lib/flow.c > index 7084079..816d561 100644 > --- a/lib/flow.c > +++ b/lib/flow.c > @@ -352,6 +352,7 @@ flow_extract(struct ofpbuf *packet, uint32_t skb_priority, > } > flow->in_port = ofp_in_port; > flow->skb_priority = skb_priority; > + flow->skb_mark = 0;
I think we need to actually pass in this value so that we can initialize it based on what the kernel gave us. Otherwise, there is nothing to translate between the conversion from Netlink and actual userspace flow. We're starting to accumulate several fields in this category so we might want to group them together in a struct if it makes sense. > diff --git a/lib/odp-util.h b/lib/odp-util.h > index 5cdb204..0214a8b 100644 > --- a/lib/odp-util.h > +++ b/lib/odp-util.h > @@ -56,6 +56,7 @@ int odp_actions_from_string(const char *, const struct > simap *port_names, > * OVS_KEY_ATTR_TUN_ID 8 -- 4 12 > * OVS_KEY_ATTR_IPV4_TUNNEL 24 -- 4 28 > * OVS_KEY_ATTR_IN_PORT 4 -- 4 8 > + * OVS_KEY_ATTR_SKB_MARK 4 -- 4 8 > * OVS_KEY_ATTR_ETHERNET 12 -- 4 16 > * OVS_KEY_ATTR_ETHERTYPE 2 2 4 8 (outer VLAN ethertype) > * OVS_KEY_ATTR_8021Q 4 -- 4 8 > @@ -65,12 +66,12 @@ int odp_actions_from_string(const char *, const struct > simap *port_names, > * OVS_KEY_ATTR_ICMPV6 2 2 4 8 > * OVS_KEY_ATTR_ND 28 -- 4 32 > * ------------------------------------------------- > - * total 184 > + * total 192 > * > * We include some slack space in case the calculation isn't quite right or > we > * add another field and forget to adjust this value. > */ > -#define ODPUTIL_FLOW_KEY_BYTES 200 > +#define ODPUTIL_FLOW_KEY_BYTES 208 At this point, let's just bump this up a little more since we're starting to get close - maybe 256. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
