On Oct 15, 2013, at 9:43 AM, Ben Pfaff <[email protected]> wrote:

> On Tue, Oct 15, 2013 at 09:19:51AM -0700, Jarno Rajahalme wrote:
>> Sets mask bits for the given field and its prerequisite fields.
>> Needed for unwildcarding the proper bits from datapath masks.
>> 
>> Signed-off-by: Jarno Rajahalme <[email protected]>
> 
> I still think that MFP_VLAN_VID doesn't have any mask prereqs,
> e.g. instead of this:
>    case MFP_ARP:
>    case MFP_IPV4:
>    case MFP_IPV6:
>    case MFP_MPLS:
>    case MFP_VLAN_VID:
>    case MFP_IP_ANY:
>        mask->dl_type = OVS_BE16_MAX;
>        /* Fall through. */
>    case MFP_NONE:
>        break;
> this:
>    case MFP_ARP:
>    case MFP_IPV4:
>    case MFP_IPV6:
>    case MFP_MPLS:
>    case MFP_IP_ANY:
>        mask->dl_type = OVS_BE16_MAX;
>        /* Fall through. */
>    case MFP_VLAN_VID:
>    case MFP_NONE:
>        break;
> 
> Otherwise:
> Acked-by: Ben Pfaff <[email protected]>

Sorry that I missed that, I was thinking about the ether type dependency on the 
wire, and did not remember that it is not visible in struct flow.

But now I wonder if the mask's VLAN_CFI bit should be set in this case, as 
mf_are_prereqs_ok() checks the CFI bit in this case. Thoughts?

  Jarno

_______________________________________________
dev mailing list
[email protected]
http://openvswitch.org/mailman/listinfo/dev

Reply via email to