https://bugzilla.kernel.org/show_bug.cgi?id=216516 --- Comment #4 from Mario Limonciello (AMD) (mario.limonciello@xxxxxxx) --- > This "stays awake" behaviour is probably a bug on it's own. > (I may may open another bug report when this one is solved) This is actually the root of the problem I suspect. This might actually be fixed by b8c824a869f220c6b46df724f85794349bafbf23. > Mario, what's your opinion on this? So to further complicate things 7123d39dc24d went back to stable and is in 5.15.y and such. I think we need to better understand the failure mode and actions associated. The workaround in daf8de0874ab5b specifically helped with "aborted suspends". 7123d39dc24d fixed the behavior that regression that happened on more systems as a result of that. > Is there anything I can do to track this down further? Here's a few things that would at least help me understand the situation: 1) Does this still exist in a fully unpatched 6.0-rc6? What's the behavior? 2) What bisect points did you use to conclude that this problem on that particular commit? By chance do you have the bisect log? 3) Why don't you just use s2idle instead in BIOS setup? It's very mature on a modern kernel. Most systems with AMD APUs shipped from OEMs don't offer a toggle even, which is why I'm surprised you have one. BTW - I wasn't aware that they offered a BIOS setting for s2ram on 845. Are you patching the BIOS somehow or just changing a setting? If you change it back to s2idle does all this go away? -- You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
