Created attachment 136555 [details] drm.debug=0x1f kernel log during critical performance loss Arch Linux x86_64 linux-drm-tip-git 4.15rc6+1282+g5f40895798ad+724923-1 xorg-server 1.19.6-2 mesa 17.3.1-2 libdrm 2.4.89-1 xf86-video-intel 1:2.99.917+804+g708255cb-2 (w/o debug) Get the following warning in the kernel log: asynchronous wait on fence i915:[global]:3268e timed out The actual issue is the alt-tabbing away from a fullscreen Minecraft window while it is actively rendering will cause critical performance issues. It'll minimize... but slowly and without redrawing everything. Sometimes the performance effect is delayed. After that, the desktop and any VT-switching is horrendously slow (on the order of several seconds to nearly a minute), forcing me to kill the minecraft program. Even after killing it, it takes a few seconds for the graphics to respond again. After switching to a linux console, performance is fine, having no extreme CPU or RAM usage. The attachment has a drm.debug=0x1f kernel log. While it didn't seem to capture the warning in this attempt, the critical performance loss is still there. It starts at around 14:02 and lasts for about 8 minutes.
What you want to do is x11trace minecraft as it appears to flood the system with swap requests (I guess) but not drawing anything itself. But I don't understand why it would be triggering redraws if it is not visible, and not updating.
Created attachment 136560 [details] x11trace of minecraft Sample incident. I didn't even minimize it this time; it seems to have done it on its own. I used the x11trace from here: https://github.com/yuq/xtrace
Created attachment 136561 [details] x11trace of marco x11trace, not x11perf. The name just stuck in my head. :/ marco is my window manager; I'm including it just in case the culprit is merely user software.
First of all. Sorry about spam. This is mass update for our bugs. Sorry if you feel this annoying but with this trying to understand if bug still valid or not. If bug investigation still in progress, please ignore this and I apologize! If you think this is not anymore valid, please comment to the bug that can be closed. If you haven't tested with our latest pre-upstream tree(drm-tip), can you do that also to see if issue is valid there still and if you cannot see issue there, please comment to the bug.
Closing, please re-open is issue still exists.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.