Intel i915 random freeze

Hi,

The last 2 weeks I have had several random freeze. According to what I investigated the problem could be related this bug: https://github.com/intel/gvt-linux/issues/137

I am not totally sure, here I leave some messages from journalctl

feb 07 16:29:56 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:29:56 kernel: [drm:gen8_reset_engines] *ERROR* rcs0 reset request timed out: {request: 00000001, RESET_CTL: 00000001}
feb 07 16:29:56 kernel: i915 0000:00:02.0: Resetting chip for hang on rcs0
feb 07 16:29:56 kernel: [drm:gen8_reset_engines] *ERROR* rcs0 reset request timed out {request: 00000001, RESET_CTL: 00000001}
feb 07 16:29:56 kernel: [drm:gen8_reset_engines] *ERROR* rcs0 reset request timed out: {request: 00000001, RESET_CTL: 00000001}
feb 07 16:29:59 kernel: Asynchronous wait on fence i915:gnome-shell[980]:8e3ee timed out (hint:intel_atomic_commit_ready+0x0/0x60)
feb 07 16:30:02 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[...]
feb 07 16:30:10 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:12 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:14 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:16 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:18 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:20 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:22 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:24 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:26 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:28 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:30 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:32 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:34 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:36 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:38 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:40 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:42 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:44 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:46 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:48 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:50 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:52 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:54 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:56 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:30:58 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:31:00 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:31:02 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:31:04 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
feb 07 16:31:06 kernel: i915 0000:00:02.0: GPU recovery timed out, cancelling all in-flight rendering.
feb 07 16:31:06 kernel: i915 0000:00:02.0: Resetting chip for hang on rcs0

This error has been reproduced with both Wayland and Xorg

My host system:
System: Dell Latitude 7400
SO: Clear Linux 32260
Linux Kernel: 5.4.18-902.native
DE: GNOME 3.34.3
CPU: Intel i7-8665U (8) @ 4.800GHz
GPU: Intel UHD Graphics 620

According to this discussion: https://bugzilla.redhat.com/show_bug.cgi?id=1780800, the workaround is back to kernel 5.3.

How could I do this in Clear Linux?
Is there any other workaround?
What other useful information can I add?

1 Like

I would recommend installing the lts kernel, which is several releases behind. That removes the need to compile anything and you will get updates for it.

1 Like

Thanks,

If I install this:

swupd bundle-add kernel-lts

Will this kernel version automatically take on boot up? Or should I do something additional?
And if I want to go back to the current kernel version, how can I do it?

1 Like

Confirm, I’ve the same trouble. This should to be fixed.

I’ve also experienced freezes.

Could you also point us to how to compile the latest kernel please as my laptop is pretty new and would benefit from a newer kernel.

Thanks,
Carolos

This is very strange,I have this issue before only when I running Brave Browser but after the latest kernel-native upgrade 5.4.18 all is ok now.

I have the same latest kernel-native 5.4.18-902 but I got system crash of video yesterday (while firefox surfing). I can’t use lts kernel (by other causes)

I have Intel UHD Graphics 605
but it still see that output in kernel like that:

user@host # dmesg -k | grep drm
...
[drm:mipi_exec_gpio] *ERROR* GPIO index 1 request failed (-2)
[drm:mipi_exec_gpio] *ERROR* GPIO index 5 request failed (-2)
[drm:mipi_exec_gpio] *ERROR* GPIO index 0 request failed (-2)
[drm:mipi_exec_gpio] *ERROR* GPIO index 5 request failed (-2)
[drm:mipi_exec_gpio] *ERROR* GPIO index 5 request failed (-2)
...

Today I updated to Clear Linux 32300, and now I have the 5.5.2-903.native kernel.

For now I have not had the issue back :smiley:

Has anyone experienced the issue with this upgrade?

Use clr-boot-manager list-kernels and clr-boot-manager set-kernel to select the kernel to boot.

1 Like

We generally provide the latest kernel (we’re already shipping 5.5.3), so while you can compile your own, it may not bring you much benefit. Nonetheless:
https://docs.01.org/clearlinux/latest/guides/kernel/kernel-development.html

1 Like

Thanks!

I upgraded this morning to 32300 (kernel 5.5) and my issue has disappeared.

Before 32300 I could consistently reproduce the issue and now everything looks good.

1 Like

I still have the same issue with kernel 5.5.5-911.

i915: Resetting rcs0 for stopped heartbeat on rcs0
i915: GPU HANG: ecode 8:1:0x86dffffd, in Xwayland [875], stopped heartbeat on rcs0

I use firefox

I change to kernel 4.19.105-115.lts2018