6.12.5+: Display quickly flickers and powers off after resuming from suspend if/when external power is disconnected

Description:

After upgrading to 6.12.5 (and in 6.12.6 as well), when resuming from suspend AND when the external power is disconnected, the display brightness rises, then quickly flickers for a few seconds, and then turns off completely. If the computer is connected back to external power, the display turns back on. If the computer is restarted, the display also turns back on (even without external power) - until it has been suspended and resumed again at least once.

This does not happen with 6.12.4 and older kernels.

I have no experience bisecting kernel bugs, but looking at the commit log for 6.12.5, this looks relevant: drm/amdgpu: rework resume handling for display (v2)

Additional info:

• Kernel version: 6.12.6-arch1-1

• Computer: Lenovo IdeaPad 5 Pro 14ACN6

• CPU: AMD Ryzen 7 5800U (16) @ 4.51 GHz

• GPU: AMD Radeon Vega Series / Radeon Vega Mobile Series [Integrated]

• config and/or log files: (did not see anything relevant in dmesg or journalctl, let me know if i should be looking for something specific)

• link to upstream bug report, if any:

Steps to reproduce:

1. Reboot into 6.12.5 or 6.12.6 kernel
2. Suspend
3. Resume from suspend
4. Disconnect computer from external power

added priority3-normal severity4-low statusunconfirmed labels

assigned to @freswa, @anthraxx, @toolybird, and @gromit

Is this still an issue with the release candidate for the next stable kernel?

sudo pacman -U https://pkgbuild.com/\~gromit/linux-bisection-kernels/linux-6.12.7rc1-1-x86_64.pkg.tar.zst

Yes, same behavior with this kernel image.

Could you also try the latest mainline release?

sudo pacman -U https://pkgbuild.com/\~gromit/linux-bisection-kernels/linux-mainline-6.13rc4-1-x86_64.pkg.tar.zst

Thank you, no longer experienced! So do we consider the issue solved now?

So do we consider the issue solved now?

Kinda. Hopefully the upstream fix will find its way into a stable release. But we won't know for sure unless you and/or @gromit perform the reverse bisection to identify the fix. Probably not worth it now...

added scoperegression statuswaiting-input labels and removed statusunconfirmed label

assigned to @heftig and unassigned @freswa, @toolybird, and @gromit

assigned to @dvzrv and @tpowa

If you want you could also see if the release candidate for the next stable kernel resolves your issue:

sudo pacman -U https://pkgbuild.com/\~gromit/linux-bisection-kernels/linux-6.12.7rc1-1-x86_64.pkg.tar.zst

If it's not solved in that one or the one before we might want to reverse-bisect the issue so we can propose the fixing patch for inclusion in the stable trees. Alternatively if you find a similar sounding issue in the DRM gitlab we might be able to guess the fixing commit from the issue.

linux-6.12.7rc1-1 has the same problem with my environment.

• Computer: Lenovo IdeaPad 5 14ABA7

• CPU: AMD Ryzen 7 5825U with Radeon Graphics (16) @ 4.546GHz

• GPU: AMD ATI 05:00.0 Barcelo [Integrated]

linux-mainline-6.13rc4-1 worked.

Confirm, same issue with the linux-6.12.7rc1-1 image. Assuming that this is the correct DRM gitlab, this issue appears to be the same: https://gitlab.freedesktop.org/drm/amd/-/issues/3853

The offending commit has been identified in that thread and a patch has been proposed.

Adding the proposed fix, modified, to 6.12.7.arch1. Please let me know if this works.

Can't reproduce the issue with 6.12.7.arch1 from core-testing. Thanks!

added resolutioncompleted label

closed