On 2025-08-28 12:35, Paolo Abeni wrote:

On 8/28/25 10:17 AM, kernel test robot wrote:

quoted

commit: a1b445e1dcd6ee9682d77347faf3545b53354d71 ("[REGRESSION][BISECTED][PATCH] net: ipv4: fix regression in broadcast routes")
url: https://github.com/intel-lab-lkp/linux/commits/Brett-A-C-Sheffield/net-ipv4-fix-regression-in-broadcast-routes/20250825-181407
patch link: https://lore.kernel.org/all/20250822165231.4353-4-bacs@librecast.net/ (local)
patch subject: [REGRESSION][BISECTED][PATCH] net: ipv4: fix regression in broadcast routes

in testcase: trinity
version: trinity-x86_64-ba2360ed-1_20241228
with following parameters:

	runtime: 300s
	group: group-04
	nr_groups: 5



config: x86_64-randconfig-104-20250826
compiler: clang-20
test machine: qemu-system-x86_64 -enable-kvm -cpu SandyBridge -smp 2 -m 16G

(please refer to attached dmesg/kmsg for entire log/backtrace)

Since I just merged v3 of the mentioned patch and I'm wrapping the PR
for Linus, the above scared me more than a bit.

AFAICS the issue reported here is the  unconditional 'fi' dereference
spotted and fixed during code review, so no real problem after all.

Correct. Jakub spotted the error, it was fixed in a v2 5 days ago, and has since
been superceded by Oscar's patch, so nothing to worry about.

Is there a way to indicate to bots not to check superceded patches. In this case
I'd have though my v2 would have been a signal? Is there something else I should
have done?

Brett
--