Bug 85768 - [bdw] TLB issues on first render?
Summary: [bdw] TLB issues on first render?
Status: CLOSED INVALID
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: unspecified
Hardware: Other All
: medium normal
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard:
Keywords:
: 78065 (view as bug list)
Depends on:
Blocks:
 
Reported: 2014-11-02 18:16 UTC by smueller
Modified: 2017-07-24 22:50 UTC (History)
2 users (show)

See Also:
i915 platform:
i915 features:


Attachments
Error log from sysfs interface (51.45 KB, text/plain)
2014-11-02 19:24 UTC, smueller
no flags Details

Description smueller 2014-11-02 18:16:21 UTC
I see the following error in dmesg:

[14523.006746] [drm] stuck on render ring
[14523.008567] [drm] GPU HANG: ecode 0:0x0fdf00ac, in Xorg.bin [1505], reason: Ring hung, action: reset
[14523.008577] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
[14523.008582] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
[14523.008587] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
[14523.008592] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
[14523.008597] [drm] GPU crash dump saved to /sys/class/drm/card0/error

I have the error file saved, but it is big and therefore I did not add it here. I can certainly make it available though.

System behavoir:

- sometimes upon resume from suspend to RAM: it takes about 5 seconds longer to wake up, but no more obvious problems

- trying to play a video is unsuccessful: no video rendered, sound is played, X.org becomes unusable (very very slow reaction: switching to another window > 30 seconds), but the system is alive; switching to a console and killing the video player process does not help, the system is still slow; only a reboot helps; when shutting down, the screen flickers

- when logging out and sddm tries to show the logon display, screen flickers for about 30 secs before sddm is shown; no obvious problems afterwards

Software: Fedora 21 beta (not tested with earlier versions)

Hardware:
Lenovo Yoga Pro 3
Intel Broadwell CPU
No additional GPU other than the Broadwell included GPU
3200x1800 screen resolution
Comment 1 Chris Wilson 2014-11-02 19:11:43 UTC
Please compress the error state (xz works really well) and attach it.
Comment 2 smueller 2014-11-02 19:24:52 UTC
Created attachment 108807 [details]
Error log from sysfs interface
Comment 3 Chris Wilson 2014-11-02 19:36:45 UTC
It jumped into never-never land. What kernel is this?
Comment 4 smueller 2014-11-02 20:17:02 UTC
Fedora 21 beta:

3.17.1-302.fc21.x86_64
Comment 5 smueller 2014-11-02 21:08:29 UTC
Maybe to help with debugging: /var/log/Xorg.0.log contains the following warnings:

[    35.493] (II) intel: Driver for Intel(R) Integrated Graphics Chipsets:
        i810, i810-dc100, i810e, i815, i830M, 845G, 854, 852GM/855GM, 865G,
        915G, E7221 (i915), 915GM, 945G, 945GM, 945GME, Pineview GM,
        Pineview G, 965G, G35, 965Q, 946GZ, 965GM, 965GME/GLE, G33, Q35, Q33,
        GM45, 4 Series, G45/G43, Q45/Q43, G41, B43
[    35.493] (II) intel: Driver for Intel(R) HD Graphics: 2000-6000
[    35.493] (II) intel: Driver for Intel(R) Iris(TM) Graphics: 5100, 6100
[    35.493] (II) intel: Driver for Intel(R) Iris(TM) Pro Graphics: 5200, 6200, 
P6300
[    35.493] (II) modesetting: Driver for Modesetting Kernel Drivers: kms
[    35.493] (II) FBDEV: driver for framebuffer: fbdev
[    35.493] (II) VESA: driver for VESA chipsets: vesa
[    35.493] (++) using VT number 1

[    35.495] (II) intel(0): Using Kernel Mode Setting driver: i915, version 1.6.
0 20140725
[    35.497] (WW) Falling back to old probe method for modesetting
[    35.497] (WW) Falling back to old probe method for fbdev
[    35.497] (II) Loading sub module "fbdevhw"
[    35.497] (II) LoadModule: "fbdevhw"
[    35.497] (II) Loading /usr/lib64/xorg/modules/libfbdevhw.so
[    35.498] (II) Module fbdevhw: vendor="X.Org Foundation"
[    35.498]    compiled for 1.16.1, module version = 0.0.2
[    35.498]    ABI class: X.Org Video Driver, version 18.0
[    35.498] (WW) Falling back to old probe method for vesa
[    35.499] (--) intel(0): Integrated Graphics Chipset: Intel(R) HD graphics 5300
[    35.499] (--) intel(0): CPU: x86-64, sse2, sse3, ssse3, sse4.1, sse4.2, avx, avx2
[    35.499] (II) intel(0): Creating default Display subsection in Screen section
        "Default Screen Section" for depth/fbbpp 24/32
[    35.499] (==) intel(0): Depth 24, (--) framebuffer bpp 32
[    35.499] (==) intel(0): RGB weight 888
[    35.499] (==) intel(0): Default visual is TrueColor
[    35.500] (II) intel(0): Output eDP1 has no monitor section
[    35.500] (--) intel(0): Found backlight control interface intel_backlight (type 'raw') for output eDP1
[    35.500] (II) intel(0): Enabled output eDP1
[    35.500] (II) intel(0): Output HDMI1 has no monitor section
[    35.500] (II) intel(0): Enabled output HDMI1
[    35.500] (--) intel(0): Using a maximum size of 256x256 for hardware cursors
[    35.500] (II) intel(0): Output VIRTUAL1 has no monitor section
[    35.500] (II) intel(0): Enabled output VIRTUAL1
[    35.500] (--) intel(0): Output eDP1 using initial mode 3200x1800 on pipe 0
[    35.500] (==) intel(0): TearFree disabled
[    35.500] (==) intel(0): DPI set to (96, 96)
Comment 6 Rodrigo Vivi 2015-01-21 23:32:32 UTC
what video are you playing and how?
I'm running videos well on same machine.

Could you also please try latest drm-intel-nightly branch from cgit.freedesktop.org/drm-intel?
Comment 7 Rodrigo Vivi 2015-01-22 00:06:27 UTC
*** Bug 78065 has been marked as a duplicate of this bug. ***
Comment 8 smueller 2015-01-22 02:05:59 UTC
After installing some additional libraries like libva, libva-vdpau and libva-intel, I can now also play videos.

So, the playback of videos (regardless which video, regardless of using xine or vlc or mplayer) without libva does not work.
Comment 9 Rodrigo Vivi 2015-01-26 19:52:06 UTC
Intel vaapi driver allows the use of GPU to decode videos and it is the official recommendation on the stack release. Since it works well with them installed let's close this bug.
Comment 10 Chris Wilson 2015-01-26 20:46:46 UTC
That a spurious argument to ignore a kernel bug.
Comment 11 Rodrigo Vivi 2015-01-27 00:01:08 UTC
So, smueller, could you please remove libva and retest with latest drm-intel-nightly?
Also, what video do you use, how do you play and does it always fail?
Comment 12 Jesse Barnes 2015-03-20 21:23:29 UTC
Stephan, please re-open if this is still a bug you can reproduce.  There have been a number of changes since you last tested, so it's possible this was fixed.  At the very least it sounds like you have a workaround, so that's good.


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.