I use HP EliteBook 745 G5 with Ryzen 5 PRO 2500U. Starting with the: commit 262485a50fd4532a8d71165190adc7a0a19bcc9e (refs/bisect/bad) Author: Anthony Koo <Anthony.Koo@amd.com> Date: Fri Oct 12 21:34:32 2018 -0400 drm/amd/display: Expand dc to use 16.16 bit backlight [Why] We want to increase precision for backlight setting. But DC interface takes 8 bit backlight level value only. [How] DMCU already takes 16 bit backlight level. Expand the DC interface to take 16.16 bit value. Max 32 bit backlight value (0xFFFF) will represent max backlight (100%) Signed-off-by: Anthony Koo <Anthony.Koo@amd.com> Reviewed-by: Tony Cheng <Tony.Cheng@amd.com> Acked-by: Leo Li <sunpeng.li@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> After resume from RAM my notebook display is all black. Before suspend: > cat /sys/class/backlight/amdgpu_bl0/{actual_brightness,brightness} 65535 255 After resume: > cat /sys/class/backlight/amdgpu_bl0/{actual_brightness,brightness} 0 255 This can be workarounded by: echo 255 > /sys/class/backlight/amdgpu_bl0/brightness assuming one can execute that without a working display. This problem still exists in the 5.0.0-rc2.
For a comparison: behavior with the commit 262485a50fd4~1 (before the regression): Before suspend: > cat /sys/class/backlight/amdgpu_bl0/{actual_brightness,brightness} 255 255 After resume: > cat /sys/class/backlight/amdgpu_bl0/{actual_brightness,brightness} 250 255 Looks like a problem with amdgpu restoring backlight after resume.
Created attachment 143141 [details] [review] drm/amd/display: Detach backlight from stream Hi Rafał, Please give the attached patch a shot. It's in our staging branch, but not in 5.0-rc2 yet.
It didn't apply cleanly on top of 5.0.0-rc2 (a trivial conflict in the dc_link.c.rej). After fixing that, compiling & testing I can confirm is solves the problem for me! Thanks! Before suspend: > cat /sys/class/backlight/amdgpu_bl0/{actual_brightness,brightness} 65535 255 After resume: > cat /sys/class/backlight/amdgpu_bl0/{actual_brightness,brightness} 64000 255 As it fixes a regression, can you submit it as a fix for the 5.0 release cycle, please?
Should be fixed with: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=35dad45d5cad3c9ca8d6a338cbf668cd7ea86469
*** Bug 109388 has been marked as a duplicate of this bug. ***
Fixed with the 5.0.0-rc3. Thanks!
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.