https://bugzilla.kernel.org/show_bug.cgi?id=216516 --- Comment #21 from Mario Limonciello (AMD) (mario.limonciello@xxxxxxx) --- > --> And closing the notebook lid is what's waking the CPU! Glad it's clarified what is going on. > I did this and applied the mentioned patches. (4 commits + usleep_range > patch) > I closed the lod 22 seconds after s2idle started. Which matches exactly the > dmesg output: OK good. > 1. Stay in deepest sleep state when closing the lid. (or immediately return > to deepest sleep) When you close the lid we need to know whether the kernel leaves the suspend loop or not. If it didn't leave the suspend loop then my hypothesis is the act of going into suspend and then closing the lid causes a Linux specific race condition that prevents going into the deepest state. That's why I proposed by usleep_range patch. It will only help with power consumption if the kernel is still in the s2idle loop. > 3. Bring the s2idle power consumption further down below 2 % per hour. That last usleep_range patch is what I'm most interested in to fix this. If you run a longer suspend, I suspect it might help with going to sleep from an EC triggered event for battery life notification. -- You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
