System Environment: ----------------------------------------------------- Non-working platforms: Braswell Browser: Google Chrome 33.0.1750.152 Desktop environment: gnome-session Case running window: window mode by default 2015-3-24 graphics stack: Kernel: (drm-intel-nightly)26db3817223d120e791f36b6be1d52b1aece47c7 Libdrm: (master)libdrm-2.4.60-17-g8dff7a04780a0f028aa7c0218806a5ee59b1a40d Mesa: (master)43277fcd59b163a8bdd00d7875557bb71ecf48ff Xserver: (master)xorg-server-1.17.0-23-g0a78b599b34cc8b5fe6fe82f90e90234e8ab7a56 Xf86_video_intel: (master)2.99.917-203-g44452825fb862c81957f53e04f739e2d96864f91 Cairo: (master)71392653803277cec6deea9e7aa4696f8b22dd73 Bug detailed description -------------------------- This is intermittent issue, the fail rate is about 1 out 3 cycles. When issue happened, the performance values dropped to below 10 fps, while in normal status, we can get 50 fps. Regression: Yes ---------------------------- Bisect still in the progress in drm-intel-next-queued branch Good kernel commit:(drm-intel-nightly) 8b4216f91c7bf8d3459cadf9480116220bd6545e Bad kernel commit: (drm-intel-nightly)26db3817223d120e791f36b6be1d52b1aece47c7 ==Reproduce steps== ---------------------------- 1.start web server 2.run /usr/local/games/WebGL/earthscreen/earth4Video.html by google-chrome
Created attachment 114803 [details] dmesg add dmesg and xorg log for this bug
Created attachment 114804 [details] xorg log
Just to be clear: these logs are from when the slowness is apparent?
dmesg: [ 831.108882] [drm] stuck on render ring [ 831.132989] [drm] GPU HANG: ecode 8:0:0x85dffffb, in X [15360], reason: Ring hung, action: reset [ 831.133091] [drm:i915_reset_and_wakeup] resetting chip [ 831.133263] [drm:i915_context_is_banned [i915]] *ERROR* gpu hanging too fast, banning! Xorg.0.log: [ 830.730] (EE) intel(0): Failed to submit rendering commands, disabling acceleration.
last exist good commit next-queue branch 2015-02-12 a679064a7e9e8799177a64a31668a34a1bc6a4f1 since BSW will fail to boot up before 2015-03-05, I had tried next-queued branch 2015-03-06 to 2015-03-11 0306 to 0309 will block the case. 0310 not exist 0311 will meet fps decrease to 7-8 issue. The bisect is blocked.
(In reply to Chris Wilson from comment #3) > Just to be clear: these logs are from when the slowness is apparent? Yes. These logs are from when the FPS is 7-8
(In reply to Ding Heng from comment #6) > (In reply to Chris Wilson from comment #3) > > Just to be clear: these logs are from when the slowness is apparent? > > Yes. These logs are from when the FPS is 7-8 Ding Heng, can you confirm current status on this issue? Is it still occurring after updating your kernel & Mesa? If so, please attached new logs, including gpu crash dump located in /sys/class/drm/card0/error.
Presuming the GPU hang has been fixed by mesa.
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.