Created attachment 144205 [details]
Part of kernel log.

Please disable guc submission and all unsafe cmdline options. If it occurs again, please attach the /sys/class/drm/card0/error.

Driver repeated resetting chip, after GPU Hang.
Then drawing was also repeating stop and re-motion.
This problem occurred only one time in running over 200 days totally.
At that time, I did not operate the machine, drawing only.
So, I could not reproduce yet.

[Environment]
CPU: SkyLake(core i5 6500TE)
Distribution: debian(customised)
Kernel: 4.14.98
Mesa: 18.3.3
libdrm: 2.4.89

> Chris Wilson
Sorry for my unfinished report.
I will try it.

Created attachment 144379 [details]
/sys/class/drm/card0/error

I'd disabled guc submission and unsafe cmdline options.
But, similar issue re-occurred yesterday.
I attached /sys/class/drm/card0/error.

Kernel message is following.
----
[38893.560462] [drm] GPU HANG: ecode 9:0:0x85dffffb, in mfd_draw [2656], reason: Hang on rcs0, action: reset
[38893.560470] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[38901.583889] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[38909.579903] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[38917.583892] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[38925.579889] i915 0000:00:02.0: Resetting rcs0 after gpu hang
:
:
----

(In reply to Yoshinori Gento from comment #4)
> Created attachment 144379 [details]
> /sys/class/drm/card0/error
> 
> I'd disabled guc submission and unsafe cmdline options.
> But, similar issue re-occurred yesterday.
> I attached /sys/class/drm/card0/error.
> 
> Kernel message is following.
> ----
> [38893.560462] [drm] GPU HANG: ecode 9:0:0x85dffffb, in mfd_draw [2656],
> reason: Hang on rcs0, action: reset
> [38893.560470] i915 0000:00:02.0: Resetting rcs0 after gpu hang
> [38901.583889] i915 0000:00:02.0: Resetting rcs0 after gpu hang
> [38909.579903] i915 0000:00:02.0: Resetting rcs0 after gpu hang
> [38917.583892] i915 0000:00:02.0: Resetting rcs0 after gpu hang
> [38925.579889] i915 0000:00:02.0: Resetting rcs0 after gpu hang
> :
> :
> ----

@Chris, How do you see this issue?

It's hanging in userspace (mesa), but the kernel is so old it is using known buggy dmc firmware that alone is responsible for a variety of GPU hangs all by itself.

(In reply to Chris Wilson from comment #6)
> It's hanging in userspace (mesa), but the kernel is so old it is using known
> buggy dmc firmware that alone is responsible for a variety of GPU hangs all
> by itself.

Thank you for your comment.
If you can, please tell me version that I should update kernel to.
4.14.131?, 4.19.56? or 5.1.15?
(I hope the version is LTS...)

I want to try the new one soon.

(In reply to Yoshinori Gento from comment #7)
> (In reply to Chris Wilson from comment #6)
> > It's hanging in userspace (mesa), but the kernel is so old it is using known
> > buggy dmc firmware that alone is responsible for a variety of GPU hangs all
> > by itself.
> 
> Thank you for your comment.
> If you can, please tell me version that I should update kernel to.
> 4.14.131?, 4.19.56? or 5.1.15?
> (I hope the version is LTS...)
> 
> I want to try the new one soon.

I recommend you to to verify the issue with drm-tip (https://cgit.freedesktop.org/drm-tip).

OK.
I updated kernel to 4.19 and started continuous operation.
So, if I meet same or other issue, I will verify the issue with drm-tip.

Yoshinori Gento, were you able to verify this? Thanks!

I updated Kernel to 4.19.57 and Mesa to 18.3.6 about 3 weeks ago.
After that, I don't see any problem so far.

I think the bug I met is fixed.
Thanks for all!