> On Mar 3, 2015, at 10:22 AM, Ben Pfaff <b...@nicira.com> wrote: > > On Tue, Feb 24, 2015 at 03:42:05PM -0800, Jarno Rajahalme wrote: >> There is no point trying to create ukeys for non-miss upcalls, such as >> upcalls due to an explicit userspace actions. >> >> Signed-off-by: Jarno Rajahalme <jrajaha...@nicira.com> > > At first glance this was a no-brainer. When I looked closer, it became > more subtle. > > First, upcall_xlate() is only called for miss upcalls, as classified by > classify_upcall() into the MISS_UPCALL category. Most of those are in > fact DPIF_UC_MISS, but slow-pathed packets also get classified as > MISS_UPCALL. I believe that those are the only packets that this patch > affects. That's reasonable, since slow-pathed packets don't need ukeys, > but the commit message doesn't reflect the subtlety, and maybe a comment > on the new "if" statement should explain the situation also. > > Acked-by: Ben Pfaff <b...@nicira.com>
Pushed to master with a new comment and improved commit message. Thanks for the review! Jarno _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
