Ben, thanks for your comments on my comments.
On 10/15/14, 5:08 PM, Ben Pfaff wrote:
On Wed, Oct 15, 2014 at 06:14:16AM -0400, Thomas F Herbert wrote:
I am late in commenting on Ben's comments on Dave's patch but I
thought I would add my two cents.
I need to review your patches, too, obviously. I'll try to do that
soon.
Thanks Ben.
....
(This was brought to my attention some time ago by ONF issue EXT-366
"Need to create OXM field for DEI bit to set/push DEI bit for Ethernet
frames". Somehow until then I hadn't realized that the CFI bit had
ever been repurposed.)
I understand and agree. Also there may be interest in using this bit as
intended in very fast datapaths such as DPDK datapath.
For compatibility with old datapaths, userspace needs to be able to
detect whether the datapath supports non-0x8100 VLANs and fall back to
userspace processing for them if it doesn't.
What datapaths support non 0x8100 vlans? I am not sure what Dave
did, but I only tested my patch with 0x8100 and 0x88a8. I am not
sure how to test for this.
Since userspace and kernel can be arbitrarily paired, userspace must
be able to tell whether the kernel supports the new feature. This is
probably as easy as trying to "execute" a set of actions that push an
non-8100 VLAN or to "flow_add" a flow that matches on a non-8100 VLAN.
OK
--
Thomas F. Herbert
Network Implementation Engineer
Entry Point LLC
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
