FS#54175 - [linux] checkerboard artifact over all graphics after installing kernel 4.11.2
Attached to Project:
Arch Linux
Opened by steelwing (steelwing) - Wednesday, 24 May 2017, 21:02 GMT
Last edited by Eli Schwartz (eschwartz) - Wednesday, 13 December 2017, 13:04 GMT
Opened by steelwing (steelwing) - Wednesday, 24 May 2017, 21:02 GMT
Last edited by Eli Schwartz (eschwartz) - Wednesday, 13 December 2017, 13:04 GMT
|
Details
Following installation of kernel 4.11.2, I rebooted my
machine. Thereafter, anything using OpenGL (or higher level
rendering from my video card) was overlaid with a black
checkerboard pattern. Screenshot of glxgears with the
artifact is attached.
Only errors shown in Xorg.log are: [ 4699.256] (EE) Failed to load module "fbdev" (module does not exist, 0) [ 4699.257] (EE) Failed to load module "vesa" (module does not exist, 0) No errors found in system journal related to amdgpu. Packages affected: * linux 4.11.2 * linux-headers 4.11.2 Steps to reproduce: 1. Use amdgpu as your video driver. 2. Install linux 4.11.2 and linux-headers 4.11.2 from core. 3. Reboot. 4. You should see the video artifacts shown in the attachment. |
This task depends upon
Closed by Eli Schwartz (eschwartz)
Wednesday, 13 December 2017, 13:04 GMT
Reason for closing: Fixed
Additional comments about closing: linux 4.14.4-1
Wednesday, 13 December 2017, 13:04 GMT
Reason for closing: Fixed
Additional comments about closing: linux 4.14.4-1
Comment by steelwing (steelwing) -
Wednesday, 24 May 2017, 21:07 GMT
Comment by
Britt Yazel (brittyazel) -
Wednesday, 24 May 2017, 23:43 GMT
Comment by steelwing (steelwing) -
Thursday, 25 May 2017, 01:15 GMT
Comment by Martin Doege (Morn) -
Sunday, 30 July 2017, 15:25 GMT
Comment by Martin Doege (Morn) -
Monday, 04 December 2017, 00:40 GMT
Comment by steelwing (steelwing) -
Monday, 04 December 2017, 09:41 GMT
Comment by
Eli Schwartz (eschwartz) - Monday,
04 December 2017, 13:13 GMT
Comment by steelwing (steelwing) -
Wednesday, 06 December 2017, 09:58 GMT
Comment by Martin Doege (Morn) -
Wednesday, 06 December 2017, 16:09 GMT
Comment by steelwing (steelwing) -
Wednesday, 06 December 2017, 16:31 GMT
Comment by
Britt Yazel (brittyazel) -
Wednesday, 06 December 2017, 17:01 GMT
Comment by loqs (loqs) - Wednesday,
06 December 2017, 17:29 GMT
Comment by steelwing (steelwing) -
Wednesday, 06 December 2017, 17:33 GMT
Comment by steelwing (steelwing) -
Wednesday, 13 December 2017, 09:46 GMT
Downgrading to the last working kernel (in my case, 4.10.13) fixed
the issue, fyi.
I had this as well on my work computer using AMGPU on a GCN 1.0
card, reverting back to radeon fixed the issue. AMDGPU on my r9
290 on my personal computer did not result in this issue.
My video card is R7 240.
This checkerboard pattern bug still persists with the current
kernel 4.12.3. I am also running an R7 240-based card.
On kernel 4.13.12 this issue seems to be solved now. So it looks
like this bug only affected the 4.11 and 4.12 kernels.
Will test with 4.13.x at next opportunity.
That should be quite easy as it's been in [core] since September
28. :)
The kernel installs, but refuses to boot, and I haven't figured
out why yet. I get the line where my drive passes its check, and
then the boot process freezes, like it can't find the kernel where
it expects to. That's a different problem from my original report,
though.
Keep in mind 4.13 and 4.14 need new kernel parameters to boot at
all on GCN 1.0/1.1 cards:
https://bbs.archlinux.org/viewtopic.php?id=232403
Also see the Wiki's AMDGPU entry for details.
.....Having read all that, I echo one commenter from the forum
post: Who checks the wiki on every kernel upgrade? I checked the
front page of the site for major announcements regarding new
kernel versions, but didn't see anything there. I never would have
known about a "fine-print" kind of thing like this without you
telling me about it.
That's why this support is considered experimental for our cards.
If you enable experimental features, it's up to you to keep up on
the changes and fine print
@steelwing after you applied the parameters was the original issue
still present under 4.14?
@loqs I'll let you know once I have applied the parameters. I
won't be able to try it until later today.
Finally confirmed as fixed. System is working on kernel 4.14.4.