Bugzilla – Bug 1104529
Regression: Linux 4.17 cannot wake from STR after ACPICA: Events: Stop unconditionally clearing ACPI IRQs during suspend/resume
Last modified: 2022-12-23 11:24:46 UTC
Created attachment 779472 [details] 4.12 boot log Between 4.12.11 and 4.17.12-1.4 power button stopped working on STR state. You can suspend but pressing power button does not wake up system anymore.
Created attachment 779473 [details] 4.17 boot log
4.14.11 is working
4.15.0-2 is working 4.16.13-3 is working
4.16.0-1 (1b10c5b) is working 4.17.0-rc2-4 (40ccd10) is broken 4.17.0-3 is broken 4.17.13-1 is broken 4.18.0-1 is broken that's as far as kotd archives go
Created attachment 779619 [details] 4.16 boot log
Likely some ACPI changes, maybe about GPE bits, but it's hard to judge which one broke. The best would be bisection...
yes, building a couple testing kernels each day. Another thing I noticed is the broken kernels immediately wake from STR on first try and stay suspended without any way to wake up on the next.
The regression appears to be in v4.16-rc2..4860ae7f58fa ("ACPICA: Update version to 20180313") If I got the branch right it should be possible to get the exact commit soon.
Pushed to master and will merge it from there. BTW, the original commit message says: > If this patch triggers regression, the regression should be in the GPE > handlers that cannot correctly determine some spurious triggered events as > no-ops. Please report any regression related to this commit to the ACPI > component on kernel bugzilla. Lv Zheng. So is this reported?
Updated upstream bug URL
I haven't tested it yet (not a good time for rebooting my laptop), but looks like commit fa85015c0d95 ("ACPICA: Clear status of all events when entering S5"), which is tagged for stable, could fix this issue. I'll try to test it later (or, more likely, during the weekend).
Hrm, Michal has tested 4.18.0, so this should have been included. It might be not enough for his machine. Michal, care to double-check?
Ok, I missed the comment where Michal states that 4.18 is broken. I've just done a quick test in my laptop with 4.18 and it's working fine. So we must be hitting a different issue, which for me is already fixed. Sorry for the noise :-(
I think I did not test the 4.18 kernel unpatched. Only that the patch applies and the patched kernel works.
oh, right. I tested it way back and it was broken
4.18.0-1.g06ab9b3-default is broken 4.18.0-5.g494fb32-default works
Could you report this to Rafael, the upstream maintainer?
Fixed upstream