On Fri, Oct 19, 2018 at 01:01:00PM +0300, Ilya Maximets wrote: > > On Thu, Jun 28, 2018 at 02:39:32PM -0300, Flavio Leitner wrote: > >> Ian, I am afraid that the issue might be local and we most probably can > >> fix with a follow up patch later, so if others are happy with the patchset, > >> please go ahead. I can see this only in the mid of next week. > > > > The issue was indeed local. I ran few tests with the patchset and > > found no issues. I got at least +1Mpps in my simple flow and 200 > > flows test cases. > > > > Thanks again, > > fbl > > Hi Flavio, > Sorry to bother you, but my colleague tries to test HWOL with Mellanox > Connect-X5 NIC and has kind of similar issue. The performance with and > without offloading is almost equal. I'm wondering, maybe you can describe > what kind of local issue you had and how you fixed it?
Sure, the local issue was with a python script that manages Xena traffic generator. A bug was introduced when ported to Python3 which caused the test packets to be pretty much garbage. So, it was not related to OVS or the HWOL. I have plans to run more tests, just not yet. -- Flavio _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
