Bug 23015 - Backlight stopped working with kernels 4.14.x - Nvidia with nouveau
Summary: Backlight stopped working with kernels 4.14.x - Nvidia with nouveau
Status: RESOLVED FIXED
Alias: None
Product: Mageia
Classification: Unclassified
Component: RPM Packages (show other bugs)
Version: 6
Hardware: x86_64 Linux
Priority: Normal normal
Target Milestone: ---
Assignee: Kernel and Drivers maintainers
QA Contact:
URL: https://git.kernel.org/pub/scm/linux/...
Whiteboard:
Keywords:
Depends on: 22909
Blocks:
  Show dependency treegraph
 
Reported: 2018-05-11 15:53 CEST by Chris B
Modified: 2018-05-18 17:31 CEST (History)
2 users (show)

See Also:
Source RPM: kernel-desktop-4.14.20-1.mga6
CVE:
Status comment: fixed upstream in kernel-4.16-rc6


Attachments

Description Chris B 2018-05-11 15:53:22 CEST
While the kernels 4.9.x work as expected, on a MacBook Air 3,1 (11-inch, late 2010) with the NVIDIA GeForce 320M chipset , with the newer kernels I'm unable to control and set the screen brightness, even worse, after resuming from suspend the brightness is set to max and can't be decreased. 

It would be nice if I wouldn't have to use the older 4.9.x kernel until Mageia 7 is released.
The propr nvidia drivers are a no go on this laptop, they make the system (EFI mode) unbootable, some PCI bus ids madness.

It's a known regression, a fix already committed to kernel 4.16-rc6:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?h=v4.16-rc6&id=9e75dc61eaa9acd1bff83c3b814ac2af6dc1f64c

I've just tested, booting a recent Debian Testing live usb (xfce) with kernel 4.16, and indeed backlight is back, function keys are working.
Marja Van Waes 2018-05-12 08:54:00 CEST

Status comment: (none) => fixed upstream in kernel-4.16-rc6
URL: (none) => https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?h=v4.16-rc6&id=9e75dc61eaa9acd1bff83c3b814ac2af6dc1f64c
CC: (none) => marja11
Assignee: bugsquad => kernel

Comment 1 Thomas Backlund 2018-05-12 10:21:44 CEST
Patch added in kernel-4.14.40-1.mga6 currently building

CC: (none) => tmb
Status: NEW => ASSIGNED

Comment 2 Chris B 2018-05-12 12:37:21 CEST
Thank you, Thomas. That was fast!
When the new kernel hits updates_testing, I'll install it and report back.
Thomas Backlund 2018-05-13 10:22:25 CEST

Depends on: (none) => 22909

Comment 3 Chris B 2018-05-13 10:28:46 CEST
Installed:

- cpupower-4.14.40-1.mga6.x86_64
- kernel-desktop-4.14.40-1.mga6-1-1.mga6.x86_64
- kernel-desktop-devel-4.14.40-1.mga6-1-1.mga6.x86_64
- kernel-desktop-devel-latest-4.14.40-1.mga6.x86_64
- kernel-desktop-latest-4.14.40-1.mga6.x86_64
- kernel-userspace-headers-4.14.40-1.mga6.x86_64
- broadcom-wl dkms-broadcom
- etc. 
Installed cleanly, first reboot into the new kernel ok. Backlight is working, function keys are working, network/broadcom wifi chip working, suspend and resume from suspend working. Applications working.

After rebooting, again into the new kernel: boot process stopped. Had to hard reset the laptop.
Booted back into kernel 4.9.56: everything is normal.

Not related to this specific bug: After extensive testing - I've got the impression it's the cpupower and setting the frequency that goes crazy with the new kernel. Uninstalled the newer cpupower package, installed the older cpupower rpm. frequency was set to ondemand. Set it to performance, made sure cpupower service is running at boot time. Somehow I managed to boot the system with the new kernel again. Odd: cpupower now reports a frequency that is not supported, it throttles back and forth and sets different (not supported? only 800, 1200 and 1400 should be there?) frequencies, for example 1322, but within the range.
But ‘lscpu | grep MHz’ reports a fixed frequency, in my case with the gouvernor ‘performance’ it is 1400.
It does not do this with the older kernels.

So, this specific bug is fixed.
And meanwhile I can boot the new kernel again. Made 4 reboot tests.
Comment 4 Chris B 2018-05-13 10:59:04 CEST
I've forget something. It's possible that I missed installing a package from updates_testing. I manually selected the packages. Maybe this leads to the booting problem.

* The module in use is: acpi_cpufreq
* There are more (minor linux related) problems with this old MacBook Air 2010, but it still runs pretty fast and almost stable, overall very enjoyable with M6. But I guess I'm the only one still using this hardware, with linux that is. 
Therefore I'm unsure if it's worth to post my findings/problems at
https://bugs.mageia.org/show_bug.cgi?id=22909
Comment 5 Thomas Backlund 2018-05-18 17:31:11 CEST
An update for this issue has been pushed to the Mageia Updates repository.

https://advisories.mageia.org/MGASA-2018-0249.html

Status: ASSIGNED => RESOLVED
Resolution: (none) => FIXED


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