https://bugzilla.kernel.org/show_bug.cgi?id=216516 --- Comment #16 from Mario Limonciello (AMD) (mario.limonciello@xxxxxxx) --- > And indeed, power usage went down to 2 % per hour! :-) > So should I open a bug for s2idle in openSUSE instead!? Eventually; yes. But I think you need to further root cause why this is happening in openSUSE before opening a bug asking them to change a policy or userspace software. Narrow down which wakeup trigger is causing the spurious events, and confirm that in Ubuntu 22.10 that wakeup trigger is disabled. > But I couldn't get the s2idle power usage significantly below 2 % per hour. When the power consumption is at this level in your test environment, can you please double check with amd_pmc.dyndbg enabled matches a high percentage of the time you were suspended? > Tomorrow I'm planning to do the debugging steps described by Mario in comment > #14. > So maybe with that the consumption can be brought further down for s2idle. Yeah let's see if that series of patches changes anything. Like I suggested on the otherwise unpatched configuration you should also correlate the number you get for time (in microseconds) you were in the deepest state against the amount of time you were in suspend. > But I also still the idea of using S3 (especially because of the very low > power consumption). > So I'd still like to see a fix for the commit initially mentioned. > Or maybe not a fix, but a kernel command line option to disable that > behavior. s2idle should be less power consumption than S3 on AMD SoCs and also have a much faster suspend and wakeup process. I suspect what is happening here is that HP's EC is issuing wakeups too frequently and causing a wasted power consumption. Let's see the information from my other debugging steps and proposed patch if they help. For your request I sent up https://patchwork.freedesktop.org/patch/504971/ for discussion. You can feel free to leave a Tested-by: comment there if that works for you. -- You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
