Re: Regression, thermal: core: battery reading wrong after wake from S3 [Was: Bug Report according to thermal_core.c]

[fhortner@xxxxxxxx]

Currently I am running kernel 6.8.12 with reverted commit.
But yesterday I had a strange behavior.
After resuming from S3 the laptop registered a low battery, immediately 
after resume (notification in KDE Plasme). But as far as I could check, 
the battery level and stats where fine. Nevertheless, after 1 minute it 
shut off. I guess the low battery reading directly after after resume 
triggered it.

It is not directly related with the commit, because I have reverted it, 
but nevertheless, maybe it is helpful to understand the underlying 
firmware issue.

I' going to test your last patch and let you know.


Am 14.06.24 um 12:18 schrieb Rafael J. Wysocki:
> On Thu, Jun 13, 2024 at 10:04 PM fhortner@xxxxxxxx <fhortner@xxxxxxxx> wrote:
> > Am 13.06.24 um 17:14 schrieb Rafael J. Wysocki:
> >
> > Let's see if the ACPI thermal zone is the real culprit.
> >
> > The attached patch only adds the delay for thermal zone 0 which is the
> > ACPI thermal zone.  It also prints the ID and type for all of the
> > resuming thermal zones.
> >
> > Please test it (removing all of the test changes/patches tried so far)
> > and let me know what happens to the battery readings.
> >
> > Patch
> >
> > thermal-delay-resume.patch
> >
> > does not work. Output according to dmesg.txt
> >
> > Attached is a slightly modified version of the last patch I sent.
> > Please test it and let me know if it addresses the problem you are
> > seeing.
> >
> > If it helps, I think we are done with this at least for now.
> >
> > patch thermal-core-resume-prio.patch with .priority = -1 does work
> >
> > One more thing to try is the attached patch (independent of the
> > previous one) to lower the priority of the thermal PM notifier to make
> > it run always after the ACPI battery one.
> >
> > Please test this one too and let me know if it works for you.
> >
> > patch thermal-core-resume-prio.patch
> >
> > with .priority = INT_MIN does also work.
> >
> > If you need any further tests, please don't hesitate to tell me so.
> No, thank you, I think you've done enough and it is appreciated.
>
> I don't see any particular drawbacks of this approach and investing
> more time in trying to get to the bottom of the issue is probably not
> worth it.
>
> I'm going to submit the last patch as the proposed solution.
>
> > Thank you for your help!
> No problem and thank you!