Re: [ovs-dev] [PATCH] odp-util.c: Fix dp_hash execution with slowpath actions.

On Mon, Jul 06, 2020 at 04:06:48PM -0700, Han Zhou wrote: > On Mon, Jul 6, 2020 at 1:42 PM Ben Pfaff wrote: > > > > On Tue, Jun 16, 2020 at 03:51:06PM -0700, Han Zhou wrote: > > > Thanks Ben. I think it is better to be backported to at least 2.12, > 2.11. I > > > am not sure about older branches.

Re: [ovs-dev] [PATCH] odp-util.c: Fix dp_hash execution with slowpath actions.

On Mon, Jul 6, 2020 at 1:42 PM Ben Pfaff wrote: > > On Tue, Jun 16, 2020 at 03:51:06PM -0700, Han Zhou wrote: > > Thanks Ben. I think it is better to be backported to at least 2.12, 2.11. I > > am not sure about older branches. > > I guess 2.13, then, as well. Yes, sorry I forgot to mention 2.13.

Re: [ovs-dev] [PATCH] odp-util.c: Fix dp_hash execution with slowpath actions.

On Tue, Jun 16, 2020 at 03:51:06PM -0700, Han Zhou wrote: > Thanks Ben. I think it is better to be backported to at least 2.12, 2.11. I > am not sure about older branches. I guess 2.13, then, as well. Would you mind posting backport patches? I see that there are patch rejects, although they are

Re: [ovs-dev] [PATCH] odp-util.c: Fix dp_hash execution with slowpath actions.

On Tue, Jun 16, 2020 at 2:40 PM Ben Pfaff wrote: > > On Fri, May 15, 2020 at 12:17:47AM -0700, Han Zhou wrote: > > When dp_hash is executed with slowpath actions, it results in endless > > recirc loop in kernel datapath, and finally drops the packet, with > > kernel logs: > > > > openvswitch: ovs-

Re: [ovs-dev] [PATCH] odp-util.c: Fix dp_hash execution with slowpath actions.

On Thu, Jun 11, 2020 at 03:15:13PM +0200, Ilya Maximets wrote: > I understand that this patch fixes this particular case, however I still > think it's dangerous to pass the hash calculated in userspace to kernel > since it might cause mismatch for the later packets in case where the > flow doesn't

Re: [ovs-dev] [PATCH] odp-util.c: Fix dp_hash execution with slowpath actions.

On Fri, May 15, 2020 at 12:17:47AM -0700, Han Zhou wrote: > When dp_hash is executed with slowpath actions, it results in endless > recirc loop in kernel datapath, and finally drops the packet, with > kernel logs: > > openvswitch: ovs-system: deferred action limit reached, drop recirc action > >

Re: [ovs-dev] [PATCH] odp-util.c: Fix dp_hash execution with slowpath actions.

On Thu, Jun 11, 2020 at 6:15 AM Ilya Maximets wrote: > > On 5/15/20 8:55 PM, Han Zhou wrote: > > > > > > On Fri, May 15, 2020 at 12:18 AM Han Zhou > wrote: > >> > >> When dp_hash is executed with slowpath actions, it results in endless > >> recirc loop in kernel datapath, and finally drops the pac

Re: [ovs-dev] [PATCH] odp-util.c: Fix dp_hash execution with slowpath actions.

On 5/15/20 8:55 PM, Han Zhou wrote: > > > On Fri, May 15, 2020 at 12:18 AM Han Zhou > wrote: >> >> When dp_hash is executed with slowpath actions, it results in endless >> recirc loop in kernel datapath, and finally drops the packet, with >> kernel logs: >> >> openvswitch:

Re: [ovs-dev] [PATCH] odp-util.c: Fix dp_hash execution with slowpath actions.

Thanks Han for the fix: I verified it in the internal ecmp environment and Ping works fine with no more drop recirc action error messages as per example above. Maybe it would be good to show if R3 has a VIF too in the diagram and tweak the commit message a bit stating VIFs under R3 can ping R3 in

Re: [ovs-dev] [PATCH] odp-util.c: Fix dp_hash execution with slowpath actions.

On Fri, May 15, 2020 at 12:18 AM Han Zhou wrote: > > When dp_hash is executed with slowpath actions, it results in endless > recirc loop in kernel datapath, and finally drops the packet, with > kernel logs: > > openvswitch: ovs-system: deferred action limit reached, drop recirc action > > The root

[ovs-dev] [PATCH] odp-util.c: Fix dp_hash execution with slowpath actions.

When dp_hash is executed with slowpath actions, it results in endless recirc loop in kernel datapath, and finally drops the packet, with kernel logs: openvswitch: ovs-system: deferred action limit reached, drop recirc action The root cause is that the dp_hash value calculated by slowpath is not p