Hi Floarian, "pelzflorian (Florian Pelz)" <[email protected]> writes:
> Fredrik Salomonsson <[email protected]> writes: > >> That did fix the issue with it shutting itself off when trying to wake >> it up with the power button. However it got replaced with it directly >> going back to suspend when waking up. Only after a few suspend/wake >> cycles does it finally wake up again. > > Hello Fredrik. No idea, but just to make sure your issue really is with > the kernel, could you try > > $ sudo -i > # echo disk > /sys/power/state > > or instead of disk whatever other suspend you want from > https://docs.kernel.org/admin-guide/pm/sleep-states.html#basic-sysfs-interfaces-for-system-suspend-and-hibernation > ? Thanks for the pointer! Looks like it might be deep mem_sleep that have some issues. I tested `echo disk > /sys/power/state` but that failed. Something about it not finding the swap header or something. I'm using a swapfile so most likely that's the issue. I did however test `echo mem > /sys/power/state` with both ``` echo s2idle > /sys/power/mem_sleep ``` and ``` echo deep > /sys/power/mem_sleep ``` The former worked, the latter had the suspend/wake loop. I thought erroneously that `loginctl suspend` was using the `/sys/power/mem_sleep` but looking closely to the log I see now it always uses `deep`. And when the suspend/wake loop is broken it had actually switched to s2idle. Now that I got s2idle working, it behaves similar to what Ubuntu's suspend did. I.e. you can wake it with any input and not just the power button. So might just be that deep sleep is broken on this laptop. According to the Arch wiki [0], s2idle should give similar power saving but with a quicker wake-up time. So I'll see if I can just switch to that. The error messages are there for both s2idle and deep so might just been a red herring. [0] https://wiki.archlinux.org/title/Power_management/Suspend_and_hibernate -- s/Fred[re]+i[ck]+/Fredrik/g
