Environment: ----------------------------------- Platform:IVB/HSW Libdrm:(master)libdrm-2.4.56-26-gae8edc7544e566084f7b958eb93c9109b471ca30 Mesa:(master)71d4fc88d6b97d6b9a8f1a324d2dcd0c56b79f3d Xserver:(master)xorg-server-1.16.0-317-geaee6572beefca240c42791f9a3a6e547bedd410 Xf86_video_intel:(master)2.99.916-47-g2a4855a633f42d90e11f5c51342c514c2dc24307 Cairo:(master)e0c0a673ee757615bd3d22f46767535e9f53d6f5 Libva:(master)e0d25ece01e7aba819c910e98c4fb4706cdab055 Libva_intel_driver:(master)bc2e06ef0f89b264fe968fbff4f06e425385c3d8 Kernel:(drm-intel-nightly)3e5e5913a80f593aa8f67784220eac2658034c6c Bug detailed description: --------------------------------------------- Show Black screen during running OglMultithread on IVB/HSW,this issue doesn't exist on BDW/BYT-M. this problem exist both on gnome-session and Raw X. It's kernel regression(drm-intel-next-queued),show bisect result first bad commit is: commit 1ed26b0b84dab119de93723ad646229db748842d Author: Michel Thierry <michel.thierry@intel.com> AuthorDate: Fri Sep 5 14:13:16 2014 +0100 Commit: Daniel Vetter <daniel.vetter@ffwll.ch> CommitDate: Mon Sep 8 09:42:19 2014 +0200 drm/i915: Enable full PPGTT on gen7 BTW,this issue also exist on drm-intel-nightly branch. Reproduce steps: --------------------------------------------- 1. xinit& 2. ./synmark2 OglMultithread
Is there anything in dmesg? Please always attach your dmesg when reporting kernel issues.
Created attachment 106541 [details] dmesg.log
(In reply to comment #1) > Is there anything in dmesg? Please always attach your dmesg when reporting > kernel issues. OK,thanks for you remind,please see dmesg.log.
Both ddr2 and ddr3 can not be reproduced the blackscreen issue on BDW platform.
(In reply to comment #4) > Both ddr2 and ddr3 can not be reproduced the blackscreen issue on BDW > platform. Sorry,I misspelled the DRI2 and DRI3,detail as below: both DRI2 and DRI3 can not be reproduced the blackscreen issue on BDW platform.
I haven't looked at this one closely enough yet, but given the test name, the black screen may be the correct outcome. So it should prove interesting to investigate.
I've run this on a couple of ivb full-ppgtt machines and seen a slowly updating window. Do you have any more specifics on how to reproduce?
(In reply to Chris Wilson from comment #7) > I've run this on a couple of ivb full-ppgtt machines and seen a slowly > updating window. Do you have any more specifics on how to reproduce? When running case OglMultithread, the issue happened. after about 2~3s, there will be black screen, then the program will exit. it can be reproduced 100%.
Does it exit before completing the tests? i.e. what's the output from synmark? Does gdb capture an error?
(In reply to Chris Wilson from comment #9) > Does it exit before completing the tests? i.e. what's the output from > synmark? Does gdb capture an error? No. The program is still running when there is black screen. I can see that there is fps value generated. There is no error information from the output and no error for gdb capture.
Is this is with DRI3 disabled and SNA acceleration selected from xorg.conf?
I've tested successfully with both DRI2/DRI3 on ivb.
(In reply to Eero Tamminen from comment #11) > Is this is with DRI3 disabled and SNA acceleration selected from xorg.conf? Yes,We run the test case with the DRI2 configuration and SNA acceleration selected from xorg.conf.
Please retest with latest -nightly, maybe map&fenceable helped. Of course please retest with i915.enable_ppgtt=2 since this is a ppgtt issue.
We have add the i915.enable_ppgtt=2 and have a try with HSW,black screen during running OglMultithread still exists.
This bug has disappeared on the latest nightly kernel.
Closing old verified.
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.