Re: [PATCH v4 net 0/4] net: hns3: fix multiple bugs in reset/error paths
From: Jijie Shao
Date: Fri Nov 13 2026 - 00:28:45 EST
on 2026/11/13 3:06, Jakub Kicinski wrote:
On Wed, 11 Nov 2026 13:01:03 +0000 netdev-bot+sinfo@xxxxxxxxxx wrote:
Please do not repost the series just to address the above. Instead,Please make sure you include the answers to these questions in the
reply to this email with the missing information, so that reviewers
can take it into account. If the series needs another revision for
other reasons, please include the information in the commit messages
then.
cover letter if you repost
ok,
Patch 1: found in internal code review. Not hit at runtime - it needs
devm_kcalloc() to fail after get_vector() has succeeded.
Patch 2: found by Sashiko on v2. Error path only, not hit at runtime.
Patch 3: real customer issue. Under memory pressure the RX ring filled
up, no more RX interrupts came in and the interface went dead. No stack
trace, the interface going dead was the only symptom. Reproduced and
fixed on EVB.
Patch 4: hit on a lab EVB while reading the debugfs queue files during
reset/unload. NULL pointer dereference, machine hung.
All four were built and tested on a Huawei Kunpeng evaluation board (EVB)
with the on-board hns3 NIC. Patch 1 and 2 were not exercised on the
failure path.
Thanks,
Jijie Shao