Re: Backlight control does not work no entries /sys/class/backlight

On Sat, 2025-09-27 at 09:03 +0000, Cliff mcDiarmid wrote:
> This I believe is governed by the driver 'amdgpu' which is showing as a
> loaded
> module.   There are errors though from the journald:
>
> [    1.656378] amdgpu: ATOM BIOS: 113-RENOIR-026
> [    1.658532] amdgpu 0000:04:00.0: Direct firmware load for amdgpu/
> renoir_asd.bin failed with error -2
> [    1.660242] [drm:amdgpu_device_init.cold [amdgpu]] *ERROR* early_init of
> IP
> block <psp> failed -19
> [    1.662661] amdgpu 0000:04:00.0: Direct firmware load for amdgpu/
> renoir_dmcub.bin failed with error -2
> [    1.664232] [drm:amdgpu_device_init.cold [amdgpu]] *ERROR* early_init of
> IP
> block <dm> failed -19
> [    1.665266] amdgpu 0000:04:00.0: Direct firmware load for amdgpu/
> renoir_pfp.bin failed with error -2
> [    1.665274] [drm:amdgpu_device_init.cold [amdgpu]] *ERROR* early_init of
> IP
> block <gfx_v9_0> failed -19
> [    1.666143] amdgpu 0000:04:00.0: Direct firmware load for amdgpu/
> renoir_sdma.bin failed with error -2
> [    1.666154] [drm:amdgpu_device_init.cold [amdgpu]] *ERROR* early_init of
> IP
> block <sdma_v4_0> failed -19
> [    1.675952] amdgpu 0000:04:00.0: Direct firmware load for amdgpu/
> renoir_vcn.bin failed with error -2
> [    1.677838] [drm:amdgpu_device_init.cold [amdgpu]] *ERROR* early_init of
> IP
> block <vcn_v2_0> failed -19
> [    1.680059] amdgpu 0000:04:00.0: amdgpu: Fatal error during GPU init
> [    1.681625] amdgpu 0000:04:00.0: amdgpu: amdgpu: finishing device.
>
> The module is loading but I'm assuming these errors impact the backlight
> setup?

Thanks Xi

>It's not only impacting the backlight setup, but almost all the
>functions of the GPU. You are likely running the system on the VESA or
>EFI framebuffer using LLVMpipe as the GL driver, thus most functions of
>the GPU do not work.

> So I loaded the latest linux firmware for AMD followed by the firmware from
> Tuxedo - neither has made any difference.

>What do you mean by no difference? If the error message lines are still
>there when you reboot it means something is still wrong.

As in the amdgpu driver errors are still there.   Firmware is in /lib/
firmware/amdgpu

>One possible reason is the AMDGPU driver was built as a part of vmlinux
>instead of a module. As the book already said (in the Mesa page), don't
>do that:
>Selecting CONFIG_DRM_RADEON, CONFIG_DRM_AMDGPU, or CONFIG_DRM_NOUVEAU
>as “y” is not recommended. If it is, any required firmware must be
>built as a part of the kernel image or the initramfs for the driver
>to function correctly.

Both CONFIG_DRM_RADEON, CONFIG_DRM_AMDGPU are built as modules.  Saw the
recommendation.

>If this is your case, rebuild the kernel with CONFIG_DRM_AMDGPU=m
>instead of y.

>Another possible reason is you simply installed the firmware files into
>a wrong location.

See above

The interesting thing is that the Ubuntu distro I used to build LFS has the
same firmware, drivers etc and the backlight functions!

Cliff

-- 
http://lists.linuxfromscratch.org/sympa/info/blfs-support
Unsubscribe: See the above information page