Graphical corruption with Intel HD4000 on Mint 16 Cinnamon

[Patrick Then]

With Mint 16 Cinnamon 64bit, my screen shows frequent graphical corruptions using the integrated Intel HD4000 of my i3 3110M.
It should be noted that the errors described below did not show up under Mint 15 Cinnamon - they only started after updating to Mint 16 (fresh install).

On the desktop, the corruption presents itself in form of small black rectangles (~4x8 pixels wide) appearing on screen, usually in the active window.
The program most affected is Firefox - websites frequently have the black rectangles, which only dissappear when scrolling them out of the screen.
However, it's not limited to Firefox, the screen corruption also shows up when switching between active windows - when switching, for a fraction of a second, the lower half right of the screen is showing a regular pattern of small rectangles.
Shadowrun Returns (native Linux version) is affected pretty badly, too.

Additionally, OpenGL through Wine is not working correctly anymore. In some games I get extreme corruptions (like screen flickering), others refuse to run or seem unable to use the full OpenGL feature set.

I have tried installing both the official Intel driver repo, as well as Xorg-edgers ppa, without much success. While they have altered the problem to various degree (e.g. frequency, Xorg-edgers fixed it in Firefox), most problems persist.

I suspect some connection to MDM, since one update shortly after release of Mint 16 improved the situation somewhat:
Before the update, full-screen Wine with OpenGL was impossible, now it works at least in some apps/games.
However - this is NOT a Wine issue. It's not affected in any way by the version of Wine I used (tested with several versions).

Here's some output from inxi:
Graphics:
Card: Intel 3rd Gen Core processor Graphics Controller bus-ID: 00:02.0
X.Org: 1.14.5 driver: intel Resolution: 1366x768@60.0hz
GLX Renderer: Mesa DRI Intel Ivybridge Mobile GLX Version: 3.0 Mesa 10.0.0 Direct Rendering: Yes

CPU:
Dual core Intel Core i3-3110M CPU (-HT-MCP-)

System:
Host: patrick-ThinkPad Kernel: 3.11.0-12-generic x86_64 (64 bit, gcc: 4.8.1)
Desktop: Gnome (Gtk 2.24.20) Distro: Linux Mint 16 Petra

I had submitted a bug report freedesktop.org (https://bugs.freedesktop.org/show_bug.cgi?id=72980), but nothing much came out of it.

[Patrick Then (patrickthen)]

Created attachment 91138
Output from Xorg.0.log

On my Lenovo Thinkpad Edge E330 (i3 3110M, HD4000), since upgrading to Mint 16, I got multiple issues with screen rendering:
* glitches/corruption mainly in Firefox and Thunderbird (random small black blocks in the browser window, of about 10x4 pixels, or corrupted text when scrolling)
* Various games that did run fine in Mint 15 using Wine/Playonlinux now give troubles:
DeusEx refuses to use high-res textures form NewVision even when using OpenGL renderer and correct settings, Sword of the Stars: The Pit (I know, there's a native version, but I don't own that) refuses to start at all and complains that it could not find a Direct3D device that supports XNA framework reach profile - both of which worked fine under Mint 15.
Icewind Dale is unplayable fullscreen in OpenGL mode, since the picture flickers like crazy - windowed works fine.
I doubt that it's a Wine issue, since I made sure to use the same version and settings as before.
* Shadowrun Returns (native Linux version) produces glitches on overlays in the form of, for lack of better words, an array of small boxes/rectangles and lines that appears and quickly vanishes again, sometimes flickering on and off for a short time.

I installed the xorg-edgers ppa and updated, which seems to have affected the glitches in Firefox - I now don't get the small boxes on scrolling anymore, but rather the bottom right part of the screen is diagonally filled with an array of small coloured boxes for a fraction of a second, which just shows up and vanishes again randomly from time to time.
Sorry that I can only offer this somewhat vague description, it happens too fast to capture it on a photo or screenshot...

As for my specs:
Mint 16 Cinnamon 64bit
Mesa version: 10.0.0-0ubuntu0sarvatt3 from edgers ppa
xserver-xorg-video-intel: 2:2.99.906+git20131213.f350a136-0ubuntu0sarvatt
Kernel: 3.11.0-12-generic x86_64 (64 bit)

Here's the output from inxi -Gxx:
Graphics:
Card: Intel 3rd Gen Core processor Graphics Controller bus-ID: 00:02.0
X.Org: 1.14.4.901 drivers: intel (unloaded: fbdev,vesa)
Resolution: 1366x768@60.0hz
GLX Renderer: Mesa DRI Intel Ivybridge Mobile GLX Version: 3.0 Mesa 10.0.0 Direct Rendering: Yes

Attached Xorg.0.log

[Patrick Then (patrickthen)]

To specify - I did a fresh install of Mint 16, no distro upgrade.

[Idr (idr)]

Are you using SNA for 2D acceleration?

[Patrick Then (patrickthen)]

Yes - but I think I had tried setting it to UXA, without success.
To be sure, I'll try again tonight and report back!

[Patrick Then (patrickthen)]

Ok, I tried to switch to UXA again, both by adding

/etc/X11/xorg.conf.d/20-intel.conf

and by creating

/etc/X11/xorg.conf

and putting

Section "Device"
   Identifier "Intel Graphics"
   Driver "intel"
   Option "AccelMethod" "uxa"
EndSection

into the respective files. Neither seems to have changed anything. Since I'm a noob with that stuff, is there a way to check whether I'm currently using SNA or UXA?

[Patrick Then (patrickthen)]

Ok, forget that, I have checked /var/Xorg.0.log

with the "uxa" option, I find only uxa in the log, with "sna" only sna, so I guess I had switched to uxa.
Anyway, it did not change anything about my issues.

I have downgraded mesa and xserver again, since xorg-edgers had introduced another bug I couldn't get rid of otherwise.

New output from inxi -Gx:

Graphics: Card: Intel 3rd Gen Core processor Graphics Controller bus-ID: 00:02.0
           X.Org: 1.14.5 driver: intel Resolution: 1366x768@60.0hz
           GLX Renderer: Mesa DRI Intel Ivybridge Mobile
           GLX Version: 3.0 Mesa 9.2.1 Direct Rendering: Yes

[Anders McCarthy (andersmccarthy)]

This is perhaps the same issue:

https://bugs.freedesktop.org/show_bug.cgi?id=68410

I'm running a fresh install of Mint 16 on a Lenovo X1 Carbon, and it's bad enough to be annoying. I only see it on Firefox, and usually only on a reverse scroll.

[Patrick Then (patrickthen)]

This indeed looks a lot like what I'm experiencing.
Only difference is that apart from the irregular black boxes in Firefox, I'm seeing regular patterns of boxes when switching windows (although often with Firefox active nd switching to/from it) and also in some other programs, esp. games (as the aforementioned Shadowrun Returns Linux version).

Also there's no mention of my OpenGl issues, so maybe those are unrelated?

[Paulo Roberto de Castro Filho (paulof-6)]

Happens with me too:

Apr 25 07:34:05 i7 kernel: [29521.057277] [drm:i915_hangcheck_elapsed] *ERROR* stuck on render ring
Apr 25 07:34:05 i7 kernel: [29521.057312] [drm:i915_set_reset_status] *ERROR* render ring hung flushing bo (0x85e000 ctx 0) at 0x5c
Apr 25 07:34:11 i7 kernel: [29527.082549] [drm:i915_hangcheck_elapsed] *ERROR* stuck on render ring
Apr 25 07:34:11 i7 kernel: [29527.082576] [drm:i915_set_reset_status] *ERROR* render ring hung flushing bo (0x86a000 ctx 0) at 0x5c
Apr 25 07:34:12 i7 kernel: [29527.550657] [drm:__gen6_gt_force_wake_get] *ERROR* Timed out waiting for forcewake old ack to clear.
Apr 25 07:34:18 i7 kernel: [29534.072053] [drm:i915_hangcheck_elapsed] *ERROR* stuck on render ring
Apr 25 07:34:18 i7 kernel: [29534.072079] [drm:i915_set_reset_status] *ERROR* render ring hung inside bo (0x865000 ctx 0) at 0x86522c
Apr 25 07:34:18 i7 kernel: [29534.094737] [drm:__gen6_gt_force_wake_get] *ERROR* Timed out waiting for forcewake old ack to clear.
Apr 25 07:34:24 i7 kernel: [29540.097315] [drm:i915_hangcheck_elapsed] *ERROR* stuck on render ring
Apr 25 07:34:24 i7 kernel: [29540.097343] [drm:i915_set_reset_status] *ERROR* render ring hung flushing bo (0x862000 ctx 0) at 0x5c
Apr 25 07:34:24 i7 kernel: [29540.125766] [drm:__gen6_gt_force_wake_get] *ERROR* Timed out waiting for forcewake old ack to clear.
Apr 25 07:34:30 i7 kernel: [29546.074603] [drm:i915_hangcheck_elapsed] *ERROR* stuck on render ring
Apr 25 07:34:30 i7 kernel: [29546.074628] [drm:i915_set_reset_status] *ERROR* render ring hung flushing bo (0x860000 ctx 0) at 0x5c
Apr 25 07:34:30 i7 kernel: [29546.105855] [drm:__gen6_gt_force_wake_get] *ERROR* Timed out waiting for forcewake old ack to clear.
Apr 25 07:34:36 i7 kernel: [29552.063891] [drm:i915_hangcheck_elapsed] *ERROR* stuck on render ring
Apr 25 07:34:36 i7 kernel: [29552.063917] [drm:i915_set_reset_status] *ERROR* render ring hung flushing bo (0x85e000 ctx 0) at 0x5c
Apr 25 07:34:36 i7 kernel: [29552.085157] [drm:__gen6_gt_force_wake_get] *ERROR* Timed out waiting for forcewake old ack to clear.
Apr 25 07:34:42 i7 kernel: [29558.101139] [drm:i915_hangcheck_elapsed] *ERROR* stuck on render ring
Apr 25 07:34:42 i7 kernel: [29558.101168] [drm:i915_set_reset_status] *ERROR* render ring hung flushing bo (0x85c000 ctx 0) at 0x5c
Apr 25 07:34:42 i7 kernel: [29558.289989] [drm:__gen6_gt_force_wake_get] *ERROR* Timed out waiting for forcewake old ack to clear.
Apr 25 07:34:48 i7 kernel: [29564.090438] [drm:i915_hangcheck_elapsed] *ERROR* stuck on render ring
Apr 25 07:34:48 i7 kernel: [29564.090463] [drm:i915_set_reset_status] *ERROR* render ring hung flushing bo (0x86a000 ctx 0) at 0x5c
Apr 25 07:34:48 i7 kernel: [29564.120419] [drm:__gen6_gt_force_wake_get] *ERROR* Timed out waiting for forcewake old ack to clear.
Apr 25 07:34:54 i7 kernel: [29570.079701] [drm:i915_hangcheck_elapsed] *ERROR* stuck on render ring
Apr 25 07:34:54 i7 kernel: [29570.079727] [drm:i915_set_reset_status] *ERROR* render ring hung inside bo (0x866000 ctx 0) at 0x86622c
Apr 25 07:34:54 i7 kernel: [29570.101377] [drm:__gen6_gt_force_wake_get] *ERROR* Timed out waiting for forcewake old ack to clear.
Apr 25 07:35:00 i7 kernel: [29576.0689...