Arch Linux

Maybe related: https://bugs.archlinux.org/task/69565

The following assumes the base-devel group and git are installed and I recommend enabling https://wiki.archlinux.org/index.php/Makepkg#Parallel_compilation to reduce build times.

$ git clone git://git.archlinux.org/svntogit/packages.git --single-branch --branch "packages/linux"
$ cd packages/trunk
$ makepkg -odd
$ cd src/archlinux-linux/
$ git checkout 05f6d2aa7e2f2cdd137ee600785704139e6dd3b7 #checkout v5.10.12, the tag is not available so using commit instead
$ cd ../..
$ makepkg -rsi # build 5.10.12 check that does not have the issue

$ cd packages/trunk/src/archlinux-linux/
$ git checkout 0c245c5fe93f0e9769de4a8b31f129b2759bf802 #checkout v5.10.13, again by commit
cd ../..
$ makepkg -rsif # build 5.10.13 check that does have the issue

$ cd packages/trunk/src/archlinux-linux/
$ git bisect start
$ git bisect bad
$ git bisect good 05f6d2aa7e2f2cdd137ee600785704139e6dd3b7
$ cd ../..
$ makepkg -rsif # build the commit git bisect need to check

$ cd packages/trunk/src/archlinux-linux/
$ git bisect $result #replace $result with good or bad
$ cd ../..
$ makepkg -rsif # repeat last four commands until git bisect finds the causal commit

@Nesser That's definitely the same problem, I searched for Linux bugs so didn't see your issue open

@loqs I don't understand how building and packaging the kernel myself (?) will solve the issue? Can you explain more?

@ali_molaei to identify which commit in [1] introduced the issue, so you know who to contact and can provide upstream with the exact cause of the issue, it also rules out the issue being caused by a change in in the kernel config or build system. See [2] for more information on bisection and [3] for reporting kernel bugs and [4] for i915 bugs in particular. You could also try linux-mainline available from [5] to see if the issue has already been fixed upstream.

[1] https://cdn.kernel.org/pub/linux/kernel/v5.x/ChangeLog-5.10.13
[2] https://wiki.archlinux.org/index.php/Bisecting_bugs_with_Git
[3] https://www.kernel.org/doc/html/latest/admin-guide/reporting-issues.html
[4] https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs
[5] https://wiki.archlinux.org/index.php/Unofficial_user_repositories#miffe

No glitches in Firefox after update to Linux 5.10.13-arch1-2. Glitches remains in Minetest, nothing of value from running MT from console.

Well, even glxgears show corrupted frames

Linux 5.10.14-arch1-1: still having the issue.

There is nothing actionable in the bug report.

I was busy, sorry, I will try the commands you said to see if I can identify the problem

Something quicker to test would be linux-mainline from [1]. See if the issue is already fixed in the release candidate of the next release.

[1] https://wiki.archlinux.org/index.php/Unofficial_user_repositories#miffe

Tried with `linux-mainline 5.11rc7-1` (Linux b590 5.11.0-rc7-1-mainline #1 SMP PREEMPT Sun, 07 Feb 2021 22:39:24 +0000 x86_64 GNU/Linux)

The problem is not solved. I will try the steps you mentioned before, btw, here is 2 AM right now, I put the commands to clone the kernel (to be downloaded) and let you know about any progress

--double comment--

After about 10 times kernel compilation with my not strong machine (each compile took about 2 hours), Finally, I found the commit that causes the problem.

```
520d05a77b2866eb4cb9e548e1d8c8abcfe60ec5 is the first bad commit
commit 520d05a77b2866eb4cb9e548e1d8c8abcfe60ec5
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date: Sun Jan 17 09:30:15 2021 +0000

drm/i915/gt: Clear CACHE_MODE prior to clearing residuals

commit ef99a60ffd9b918354e038bc5e61f007ff7e901d upstream.

Since we do a bare context switch with no restore, the clear residual
kernel runs on dirty state, and we must be careful to avoid executing
with bad state from context registers inherited from a malicious client.

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2955
Fixes: 09aa9e45863e ("drm/i915/gt: Restore clear-residual mitigations for Ivybridge, Baytrail")
Testcase: igt/gem_ctx_isolation # ivb,vlv
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Reviewed-by: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Link:20210117093015.29143-1-chris@chris-wilson.co.uk"> https://patchwork.freedesktop.org/patch/msgid/20210117093015.29143-1-chris@chris-wilson.co.uk
(cherry picked from commit ace44e13e577c2ae59980e9a6ff5ca253b1cf831)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drivers/gpu/drm/i915/gt/gen7_renderclear.c | 12 ++++++++++++
1 file changed, 12 insertions(+)
```

I don't know if I should report it upstream or not doe.

This issue: https://gitlab.freedesktop.org/drm/intel/-/issues/3071 was on i915 bug tracker so I commented a link to this issue there too.

I think is useless now, but you may want to use ccache for speeding the compiling of the same package

https://wiki.archlinux.org/index.php/Ccache

@ali_molaei thank you for the bisect and upstream report.
Does the proposed fix [1] work?

[1] https://patchwork.freedesktop.org/patch/420436/?series=86947&rev=1

@logs: No it didn't work, https://gitlab.freedesktop.org/drm/intel/-/issues/3071#note_796697

Just updated Linux to version 5.10.15-arch1-1, glxgears and Firefox still shows artifacts. Maybe the patch will be in a future kernel release.

@logs: No it didn't work, https://gitlab.freedesktop.org/drm/intel/-/issues/3071#note_796697

Linux 5.10.16-arch1-1
Artifacts in Firefox and glxgears.

Please try building the attached source archive it contains an additional two patches from  FS#69406 

Hey @loqs sorry for being late, I compiled with these patches too, I still have the problems

Linux: 5.11.1-arch1-1

Effects: Firefox, Minetest and glxgears still shows artifacts.

I was experiencing an issue that I believe is the same as this one. It was resolved for me since upgrading from linux-5.11.2-arch1-1 to linux 5.11.4-arch1-1 (yeah, I skipped 5.11.3).

I also believe I am seeing this same i915 related issue but it is quite infrequent so difficult to say what kernel version it started and I have don't know what initiates it. I guess I first saw graphics glitches about kernel 5.10.16 and have seen glitches on most versions since (although not yet on 5.11.6 over last 2 days). Once they start they will remain until I reboot. Sometimes I get colored "snow" lines the appear on animations, sometimes I get rectangular/triangular artifacts and/or flashing. Here is a small video of an odd one I had on 5.11.5 although I have not seen another one like this before: https://vimeo.com/522537489.

@jantman I still have the problem even in 5.11.6 and 5.12.rc1

@bulletmark I watched your video, it seems to be another problem though, The best way I found to detect this issue is by running the command: glxgears

I think the problem occurs from all 5.11.x to 5.12, but I'm not so sure about 5.12. The only reason I know about the bug's existence on 5.12 is because of the comments. :) I'm on 5.11.8 and 3D rendering is not the greatest it's ever been.

For me the problem is present in 5.12.rc8 too. I am using drm in a intel hd 2500 gpu.
For me it just looks like bits of the geometry is missing.

Latest update: https://gitlab.freedesktop.org/drm/intel/-/issues/3071#note_895128

Kernel: linux 5.12.1-arch1-1
glxgears: still artifacts

Latest update: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/drivers/gpu/drm/i915/gt/gen7_renderclear.c?id=023dfa9602f561952c0e19d74f66614a56d7e57a

https://gitlab.freedesktop.org/drm/intel/-/issues/3081#note_928867

This should solve the issue, but I think it's in the kernel that comes after v5.13-rc2

BTW, maybe a patch can solve the issue for now on arch?

Should be fixed in linux 5.12.6.arch1-1

It's fixed in 5.12.6.arch1-1, I think this issue can be closed. Thank you to all the people who helped with this :D