[openib-general] [BUG]driver didn't accept ARP reply packet

[Roland Dreier]

    Shirley> I did netperf test over Mellanox 23108 4X HCA against
    Shirley> r2720, after a while the HCA stopped to accept packets.
    Shirley> I am pretty sure it's a bug in driver not IPoIB. rmmod
    Shirley> ib_ipoib didn't help, after removing ib_mthca, and
    Shirley> restarted, the driver worked again. It's easy to
    Shirley> reprocude on my 4-way intel based systems.

What is your recipe for reproducing?

    Shirley> I did saw the below messages from /var/log/messages, when
    Shirley> this happened.

    Shirley> ib_mthca 0000:04:00.0: Port change to down for port 1
    Shirley> ib_mthca 0000:04:00.0: Port change to active for port 1

This means that your link is going down and coming back up.  What do
the error counters in /sys/class/infiniband/mthca0/ports/1/counters
show before and after this happens?

    Shirley> And the HCA didn't accept ARP reply packet afterwards.

    Shirley> Now I am adding a permanent arp entry to avoid this
    Shirley> problem. It seems to work fine. Anybody else saw this
    Shirley> problem?

No, I haven't seen this happen.

 - R.