[SOLVED] Laptop backlight dims to zero on boot with kernel 5.14.0

Daedra · 09-02-2021, 01:31 PM

I am going to assume this is some how caused by kernel 5.14.0 since the problem only started occurring after that upgrade. The laptop I have is an nvidia/AMD vega graphics hybrid, the problem is during boot once amdgpudrmfb takes over the backlight of the laptop drops to 0 and is barely visible. As a workaround for now I added this to rc.local

Code:

echo 255 > /sys/class/backlight/amdgpu_bl0/brightness

This will bring the backlight back up to max, but this was not necessary before 5.14.0. Any ideas on how to fix this are appreciated. Here is some dmesg output. I didn't find anything useful in it though

Code:

[    6.324280] [drm] amdgpu kernel modesetting enabled.
[    6.330234] amdgpu: Virtual CRAT table created for CPU
[    6.331078] amdgpu: Topology: Add CPU node
[    6.331926] fb0: switching to amdgpudrmfb from EFI VGA
[    6.332815] amdgpu 0000:06:00.0: vgaarb: deactivate vga console
[    6.332839] amdgpu 0000:06:00.0: enabling device (0006 -> 0007)
[    6.332884] amdgpu 0000:06:00.0: amdgpu: Trusted Memory Zone (TMZ) feature enabled
[    6.334177] amdgpu 0000:06:00.0: amdgpu: Fetched VBIOS from VFCT
[    6.334180] amdgpu: ATOM BIOS: 113-RENOIR-026
[    6.334623] amdgpu 0000:06:00.0: amdgpu: VRAM: 512M 0x000000F400000000 - 0x000000F41FFFFFFF (512M used)
[    6.334626] amdgpu 0000:06:00.0: amdgpu: GART: 1024M 0x0000000000000000 - 0x000000003FFFFFFF
[    6.334628] amdgpu 0000:06:00.0: amdgpu: AGP: 267419648M 0x000000F800000000 - 0x0000FFFFFFFFFFFF
[    6.334670] [drm] amdgpu: 512M of VRAM memory ready
[    6.334672] [drm] amdgpu: 3072M of GTT memory ready.
[    6.335443] amdgpu 0000:06:00.0: amdgpu: PSP runtime database doesn't exist
[    7.040261] amdgpu 0000:06:00.0: amdgpu: RAS: optional ras ta ucode is not available
[    7.047797] amdgpu 0000:06:00.0: amdgpu: RAP: optional rap ta ucode is not available
[    7.047808] amdgpu 0000:06:00.0: amdgpu: SECUREDISPLAY: securedisplay ta ucode is not available
[    7.048940] amdgpu 0000:06:00.0: amdgpu: SMU is initialized successfully!
[    7.159638] kfd kfd: amdgpu: Allocated 3969056 bytes on gart
[    7.173434] amdgpu: HMM registered 512MB device memory
[    7.173493] amdgpu: SRAT table not found
[    7.173498] amdgpu: Virtual CRAT table created for GPU
[    7.173819] amdgpu: Topology: Add dGPU node [0x1636:0x1002]
[    7.173834] kfd kfd: amdgpu: added device 1002:1636
[    7.173845] amdgpu 0000:06:00.0: amdgpu: SE 1, SH per SE 2, CU per SH 18, active_cu_number 26
[    7.176896] fbcon: amdgpu (fb0) is primary device
[    7.218403] amdgpu 0000:06:00.0: [drm] fb0: amdgpu frame buffer device
[    7.223883] amdgpu 0000:06:00.0: amdgpu: ring gfx uses VM inv eng 0 on hub 0
[    7.223907] amdgpu 0000:06:00.0: amdgpu: ring comp_1.0.0 uses VM inv eng 1 on hub 0
[    7.223929] amdgpu 0000:06:00.0: amdgpu: ring comp_1.1.0 uses VM inv eng 4 on hub 0
[    7.223946] amdgpu 0000:06:00.0: amdgpu: ring comp_1.2.0 uses VM inv eng 5 on hub 0
[    7.223963] amdgpu 0000:06:00.0: amdgpu: ring comp_1.3.0 uses VM inv eng 6 on hub 0
[    7.223980] amdgpu 0000:06:00.0: amdgpu: ring comp_1.0.1 uses VM inv eng 7 on hub 0
[    7.223997] amdgpu 0000:06:00.0: amdgpu: ring comp_1.1.1 uses VM inv eng 8 on hub 0
[    7.224014] amdgpu 0000:06:00.0: amdgpu: ring comp_1.2.1 uses VM inv eng 9 on hub 0
[    7.224031] amdgpu 0000:06:00.0: amdgpu: ring comp_1.3.1 uses VM inv eng 10 on hub 0
[    7.224048] amdgpu 0000:06:00.0: amdgpu: ring kiq_2.1.0 uses VM inv eng 11 on hub 0
[    7.224065] amdgpu 0000:06:00.0: amdgpu: ring sdma0 uses VM inv eng 0 on hub 1
[    7.224089] amdgpu 0000:06:00.0: amdgpu: ring vcn_dec uses VM inv eng 1 on hub 1
[    7.224106] amdgpu 0000:06:00.0: amdgpu: ring vcn_enc0 uses VM inv eng 4 on hub 1
[    7.224123] amdgpu 0000:06:00.0: amdgpu: ring vcn_enc1 uses VM inv eng 5 on hub 1
[    7.224140] amdgpu 0000:06:00.0: amdgpu: ring jpeg_dec uses VM inv eng 6 on hub 1
[    7.225750] [drm] Initialized amdgpu 3.42.0 20150101 for 0000:06:00.0 on minor 1

Zexuo · 09-02-2021, 02:35 PM

This started happening on my AMD laptop (with AMD GPU) with the 5.14.0 update as well. It returns to normal brightness after resuming from suspend. Funny enough, xbacklight does not recognize any outputs with the backlight property on this machine, but I can control the brightness with xrander --brightness. The xrander command uses the dimmed level as the baseline and won't restore full brightness until the machine resumes from suspend.

This problem did not manifest on my Intel laptop.

Daedra · 09-02-2021, 02:49 PM

Thanks for the report, glad to know I am not the only one. I may start checking the kernel mailing lists to see if I can find anything.

GazL · 09-02-2021, 03:37 PM

For my intel backlight I use this /etc/udev/rules.d/ file

Code:

# 81-backlight.rules:
#
#    Set the initial backlight level.
#    When present, use the 'acpi_video0' control in preference to
#    the vendor 'intel_backlight' control.
#

SUBSYSTEM!="backlight", GOTO="backlight_end"
ACTION!="add", GOTO="backlight_end"

KERNEL=="intel_backlight", TEST!="/sys/class/backlight/acpi_video0/brightness", ATTR{brightness}="364"

KERNEL=="acpi_video0", ATTR{brightness}="39"

LABEL="backlight_end"

You could modify it to suit your amd hardware.

Still doesn't explain why yours is being set to 0 all of a sudden, but it's a little less crude than a rc.local hack.

Daedra · 09-02-2021, 03:39 PM

Quote:

Originally Posted by GazL

For my intel backlight I use this /etc/udev/rules.d/ file

Code:

# 81-backlight.rules:
#
#    Set the initial backlight level.
#    When present, use the 'acpi_video0' control in preference to
#    the vendor 'intel_backlight' control.
#

SUBSYSTEM!="backlight", GOTO="backlight_end"
ACTION!="add", GOTO="backlight_end"

KERNEL=="intel_backlight", TEST!="/sys/class/backlight/acpi_video0/brightness", ATTR{brightness}="364"

KERNEL=="acpi_video0", ATTR{brightness}="39"

LABEL="backlight_end"

You could modify it to suit your amd hardware.

Still doesn't explain why yours is being set to 0 all of a sudden, but it's a little less crude than a rc.local hack.

Thanks, I'll modify it and report back.

marav · 09-02-2021, 04:18 PM

Here is mine

Code:

$ cat /etc/udev/rules.d/30-amdgpu-bl.rules
KERNEL=="card0", SUBSYSTEM=="drm", DRIVERS=="amdgpu", ATTR{device/backlight/amdgpu_bl0/brightness}="255"

I had found the trick here :
https://wiki.archlinux.org/title/AMDGPU

marav · 09-02-2021, 05:15 PM

Another workaround that works well :

in elilo.conf

Code:

image=vmlinuz
...
  append="root=/dev/nvme0n1p3 vga=normal ro resume=/dev/nvme0n1p4 quiet amdgpu.backlight=1"

Daedra · 09-02-2021, 05:21 PM

Thanks guys, all three of those suggestions worked. amdgpu.backlight=1 works best since there is no dimming whatsoever with that route.

marav · 09-02-2021, 05:41 PM

Quote:

Originally Posted by Daedra

Thanks guys, all three of those suggestions worked. amdgpu.backlight=1 works best since there is no dimming whatsoever with that route.

erf ...

There's an unwanted behaviour with amdgpu.backlight=1

Function keys ( fn+F6 & fn+F7 ) that normally decrease/increase the brightness, on my Dell laptop, don't work anymore ...

Daedra · 09-02-2021, 05:52 PM

Just tested on my laptop. They fn+f11/f12 keys are working in mate,kde, and xfce. In non X they don't function, but I had never tried them out side of X before so that may be normal.

marav · 09-02-2021, 06:07 PM

Quote:

Originally Posted by Daedra

Just tested on my laptop. They fn+f11/f12 keys are working in mate,kde, and xfce. In non X they don't function, but I had never tried them out side of X before so that may be normal.

In KDE, no matter if my brightness is 0% or 100%, the brightness is still full

But :

Code:

root:~/ # cat /sys/class/drm/card0/device/backlight/amdgpu_bl0/{actual_brightness,brightness,max_brightness}
1
1
255

when I decrese to 0%
And

Code:

255
255
255

When I increase to 100%

But nothing change

EDIT: I'll keep my .rules file as long as it works under X/Wayland.

Daedra · 09-03-2021, 03:57 PM

Well spoke too soon. Like you, changing brightness with keyboarding was not working on battery, and the laptop was not dimming on idle. Went back to the udev rule route. Still all fixed though

glats · 09-19-2021, 12:58 AM

I have the same problem with my thinkpad t14. Only my card is a single Renoir. Is this due to some change in the amdgpu driver or kernel?

Daedra · 09-19-2021, 04:18 AM

Quote:

Originally Posted by glats

I have the same problem with my thinkpad t14. Only my card is a single Renoir. Is this due to some change in the amdgpu driver or kernel?

For me it started with kernel 5.14. Not sure what commit may have caused this, but the uvdev rule posted above fixes the issue for me.

thunderweasel · 09-30-2021, 08:55 AM

Quote:

Originally Posted by Daedra

For me it started with kernel 5.14. Not sure what commit may have caused this, but the uvdev rule posted above fixes the issue for me.

I'm running Void on an Ideapad Flex 5 14ARE05, but experiencing the same thing. Issue introduced with the recent upgrade from the 5.13.9 kernel to the 5.14.8 kernel. amdgpu.backlight=1 kernel param in GRUB CMD restores full screen brightness, including in the pre-boot environment and my disk encryption password prompt, however, it breaks the functionality of the screen backlight keyboard function keys (using acpilight; the brightness - key simply turns off the display, and the brightness + key restores it to 100% full brightness with no steps inbetween; my Fn keys are mapped to exec:xbacklight -inc 10 and exec:xbacklight -dec 10; using JWM, no DE). For the time being I am using the udev rule ITT as a workaround. Screen backlight keyboard function keys work properly with correct stepping, and I get full screen brightness once the kernel has loaded, but it it is still set to its lowest brightness level during my disk encryption password prompt and the first couple lines of the boot sequence. Annoying but tolerable. On the bright side (pun semi-intended), battery life on this kernel is noticeably better.