Bug 215553 - Intel_backlight does not work after updating from 5.10 to 5.15
Summary: Intel_backlight does not work after updating from 5.10 to 5.15
Status: RESOLVED MOVED
Alias: None
Product: Drivers
Classification: Unclassified
Component: Video(Other) (show other bugs)
Hardware: Intel Linux
: P1 normal
Assignee: other_modules
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2022-01-31 00:01 UTC by Filippo Falezza
Modified: 2022-02-08 12:03 UTC (History)
1 user (show)

See Also:
Kernel Version: 5.15
Tree: Mainline
Regression: Yes


Attachments
dmesg of 5.15.13 (100.52 KB, application/octet-stream)
2022-01-31 16:29 UTC, Filippo Falezza
Details

Description Filippo Falezza 2022-01-31 00:01:13 UTC
As per title, after updating the kernel from to the currently latest version available in Arch repositories, the intel_backlight class has stopped working.

This is still present under /sys/class/backlight, but any change to it does not affect backlight. The only change that can be made is by setting abcklight to 0,  which turns off the screen.

To keep a usable system, I had to roll-back to 5.10.89, the last known version to work properly.

I have tried changing the pwm of the screen, which does affect the backlight, dimming it in steps, but it seems that something major has been changed.
I also tried to just roll back the linux-firmware package, with no success. From this, I concluded that the issue should be in the kernel.

Finally, I should say that I have noticed a change in the scale of intel_backlight, as maximum_backlight went from 120000 down to 512.

Thanks to anyone who will help me.


Which files should I upload to start debugging?
Comment 1 Filippo Falezza 2022-01-31 16:29:40 UTC
Created attachment 300361 [details]
dmesg of 5.15.13
Comment 2 Filippo Falezza 2022-01-31 16:30:49 UTC
I am now looking into all commits between 5.10.90 and 5.15.13 that may have affected the intel_brightness component
Comment 3 Filippo Falezza 2022-01-31 16:58:47 UTC
I also tried setting drm.debug=0x1ff, and I get the followings from the dmesg when trying to change the brightness:




>[ 1605.793461] [drm:intel_backlight_device_update_status [i915]] updating
>intel_backlight, brightness=413/512
>[ 1605.793508] i915 0000:00:02.0: [drm:intel_backlight_device_update_status
>[i915]] set backlight level = 413
>[ 1605.793770] i915 0000:00:02.0: [drm:drm_dp_dpcd_write] AUX A/DDI A/PHY A:
>0x00354 AUX <- (ret=  4) 9d 01 00 00
>[ 1605.801854] i915 0000:00:02.0:
>[drm:drm_crtc_vblank_helper_get_vblank_timestamp_internal] crtc 0 : v
>p(0,1132)@ 1605.864860 -> 1605.845740 [e 0 us, 0 rep]
>[ 1605.801863] i915 0000:00:02.0: [drm:drm_update_vblank_count] updating
>vblank count on crtc 0: current=64141, diff=0, hw=66180 hw_last=66180
>[ 1605.802007] [drm:drm_ioctl] comm="Xorg" pid=954, dev=0xe200, auth=1,
>I915_GEM_THROTTLE
>[ 1605.807081] i915 0000:00:02.0:
>[drm:drm_crtc_vblank_helper_get_vblank_timestamp_internal] crtc 0 : v
>p(0,-39)@ 1605.870088 -> 1605.870747 [e 0 us, 0 rep]
>[ 1605.807090] i915 0000:00:02.0: [drm:drm_update_vblank_count] updating
>vblank count on crtc 0: current=64141, diff=1, hw=66181 hw_last=66180
>[ 1605.807094] i915 0000:00:02.0: [drm:drm_handle_vblank] vblank event on
>64142, current 64142
>[ 1605.807098] i915 0000:00:02.0: [drm:drm_handle_vblank] vblank event on
>64142, current 64142
>[ 1605.807111] [drm:drm_atomic_state_default_clear] Clearing atomic state
>000000005cedc77d
>[ 1605.807115] [drm:drm_mode_object_put.part.0] OBJ ID: 104 (2)
>[ 1605.807117] [drm:drm_mode_object_put.part.0] OBJ ID: 110 (4)
>[ 1605.807118] [drm:drm_mode_object_put.part.0] OBJ ID: 110 (3)
>[ 1605.807119] [drm:drm_mode_object_put.part.0] OBJ ID: 109 (3)
>[ 1605.807121] [drm:drm_mode_object_put.part.0] OBJ ID: 109 (2)
>[ 1605.807122] [drm:__drm_atomic_state_free] Freeing atomic state
>000000005cedc77d
>[ 1605.807164] [drm:drm_ioctl] comm="Xorg" pid=954, dev=0xe200, auth=1,
>DRM_IOCTL_WAIT_VBLANK
>[ 1605.807169] i915 0000:00:02.0: [drm:drm_wait_vblank_ioctl] event on vblank
>count 64143, current 64142, crtc 0
>[ 1605.809057] [drm:intel_backlight_device_update_status [i915]] updating
>intel_backlight, brightness=415/512
>[ 1605.809237] i915 0000:00:02.0: [drm:intel_backlight_device_update_status
>[i915]] set backlight level = 415
>[ 1605.809503] i915 0000:00:02.0: [drm:drm_dp_dpcd_write] AUX A/DDI A/PHY A:
>0x00354 AUX <- (ret=  4) 9f 01 00 00


Might it be a n issue in writing to the AUX register?
Comment 4 Artem S. Tashkinov 2022-02-01 10:45:38 UTC
Please read this https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs and file where appropriate. Thanks.
Comment 5 The Linux kernel's regression tracker (Thorsten Leemhuis) 2022-02-08 11:56:06 UTC
Filippo, did you file the bug somewhere else? Where?
Comment 6 The Linux kernel's regression tracker (Thorsten Leemhuis) 2022-02-08 11:58:40 UTC
(In reply to Artem S. Tashkinov from comment #4)
> Please read this
> https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs and
> file where appropriate. Thanks.

Artem, when you give such advice, could you please ask people to add a link to the other report in the current report? Then everybody that stumbles on this report will know if the bug was actually filed somewhere else and where to look for further details about it.
Comment 7 Filippo Falezza 2022-02-08 12:03:59 UTC
Sorry for filing in the wrong place.

Here is the link to the bug in the freedesktop gitlab: https://gitlab.freedesktop.org/drm/intel/-/issues/5027

Note You need to log in before you can comment on or make changes to this bug.