On Sun, Nov 29, 2020 at 2:04 AM Michael Chan [off-list ref] wrote:

On Wed, Nov 25, 2020 at 11:39 AM Marc Smith [off-list ref] wrote:

quoted

[17879.279213] bnxt_en 0000:01:09.0 s1p1: TX timeout detected,
starting reset task!
[17883.075299] bnxt_en 0000:01:09.0 s1p1: Resp cmpl intr err msg: 0x51
[17883.075302] bnxt_en 0000:01:09.0 s1p1: hwrm_ring_free type 1
failed. rc:fffffff0 err:0
[17886.957100] bnxt_en 0000:01:09.0 s1p1: Resp cmpl intr err msg: 0x51
[17886.957103] bnxt_en 0000:01:09.0 s1p1: hwrm_ring_free type 2
failed. rc:fffffff0 err:0
[17890.843023] bnxt_en 0000:01:09.0 s1p1: Resp cmpl intr err msg: 0x51
[17890.843025] bnxt_en 0000:01:09.0 s1p1: hwrm_ring_free type 2
failed. rc:fffffff0 err:0

These messages could be caused by IRQ issues or firmware issues.  It
seems that one IRQ vector that is shared by 1 TX ring (type 1) and 2
RX rings (type 2) is not getting the interrupt.  It caused the
original TX timeout and the interrupt issue persisted during the
recovery to free these rings.

quoted

firmware-version: 214.0.191.0

This firmware is about 2 years old.  I'll have someone get back to you
on how to upgrade and what version to upgrade to.

Thanks Michael. After more investigation, I'm not so sure this is a
firmware issue; we see these corresponding events on the host
(hypervisor):
AMD-Vi: Event logged [IO_PAGE_FAULT device=c8:02.0 domain=0x0001
address=0xfffffffdf8050000 flags=0x0008]

Thanks.