Re: [PATCH net v4 0/7] net: stmmac: More selftest-related fixes

[Maxime Chevallier <maxime.chevallier@bootlin.com>]

Hi Jakub,

On 9/22/26 00:22, Jakub Kicinski wrote:
> On Thu, 17 Sep 2026 23:53:31 +0200 Maxime Chevallier wrote:
>> This is V4 of stmmac selftest fixes, addressing Sashiko's issues over
>> the MTU patch. This lead to the introduction of a new one. The
>> dev_add_pack races have been addressed, however the double-vlan issue
>> stayed there. Ovidiu is actively working on it, let's wait for his work
>> to land before fixing that.
>>
>> Nicolai, I took the liberty of keeping your review tags on affected
>> patches, let me know if this isn't ok.
>>
>> This is another round of stmmac selftest fixes, mostly about the selftests
>> themselves but a few things were discovered w.r.t MTU and buffer size
>> handling, see patch 5 anf 6.
>>
>> After this is merged, I consider the selftests to be now reliable enough
>> to run them nightly on every stmmac series that's sent, and I'll be requiring
>> clean selftests for new glue drivers.
> 
> This got to the top of the stack for me, LMK if none of the AI feedback
> is legit. You mentioned the VLAN one, not sure about the rest.
> The EEE complaints look like nit picks but also some are easily
> addressable? LMK..

I'd say the AI complaints are legit for EEE, but I have mixed feeling about
iterating more for that point only.

AI points bugs that are related to 3 main classes of problems (for EEE but
other patches as well) :

 - User reconfigures the interface during the ethtool -t run
 - The selftest outcome depends on how the link was negotiated (worth for Pause
   and EEE)
 - Selftests are sensitive to external traffic being received during the
   selftests run.

Obviously we don't want the kernel to crash because of ethtool selftests and that
has been addressed, but for now I'd say ethtool selftests failing because of
external factors is something we can live with, although we should make them
more robust.

The main thing is that we've always been subjected to these problems, and addressing
that is an entire ordeal in itself. I can followup (my goal is to have all
green runs here [1], this already proved useful for at least 2 patches in the
past 2 months that broke selftests), so I'll definitely work on the remaining
points as followup.

Maxime

[1] : https://minimaxwell.github.io/stmmac-ci/

(These are just custom tests, I want to switch to the real kselftests eventually,
the runner is currently being stress-tested to iron out the last issues before
submitting that to netdev-ci)