Bug 23296

Summary:	NVIDIA GPU "fallen off the bus" - upgrading to 4.14.50 broke the GPU
Product:	Mageia	Reporter:	Jérôme Hénin <heninj>
Component:	RPM Packages	Assignee:	Kernel and Drivers maintainers <kernel>
Status:	RESOLVED FIXED	QA Contact:
Severity:	normal
Priority:	Normal	CC:	fri, marja11, tmb
Version:	6
Target Milestone:	---
Hardware:	All
OS:	Linux
Whiteboard:
Source RPM:	kernel-4.14.50-2.mga6.src.rpm	CVE:
Status comment:

Description Jérôme Hénin 2018-07-12 10:56:57 CEST

Description of problem:
On an Optimus laptop (ASUS) with an MX150 discrete GPU, upgrading to 4.14.50 broke the GPU (whether I'm using Bumblebee or mageia-prime). I'm using the "pcie_port_pm=off" boot option for all kernel versions.

Trying to run anything with bumblebee gives:
> $ primusrun glxspheres64 
> primus: fatal: Bumblebee daemon reported: error: Could not load GPU driver

And dmesg gives:
> [   50.081664] nvidia-nvlink: Nvlink Core is being initialized, major device number 244
> [   50.475307] nvidia 0000:01:00.0: enabling device (0000 -> 0003)
> [   50.625646] NVRM: The NVIDIA GPU 0000:01:00.0
>                NVRM: (PCI ID: 10de:1d10) installed in this system has
>                NVRM: fallen off the bus and is not responding to commands.
> [   50.675784] nvidia: probe of 0000:01:00.0 failed with error -1
> [   50.675799] NVRM: The NVIDIA probe routine failed for 1 device(s).
> [   50.675800] NVRM: None of the NVIDIA graphics adapters were initialized!
> [   50.675885] nvidia-nvlink: Unregistered the Nvlink Core, major device number 244


Version-Release number of selected component (if applicable):
4.14.50-desktop-2.mga6

How reproducible:
All the time.

Steps to Reproduce:
1. Install bumblebee on said hardware
2. Start any program with bumblebee, or run "modprobe nvidia-current"

Marja Van Waes 2018-07-12 15:34:27 CEST

CC: (none) => marja11
Assignee: bugsquad => kernel

Marja Van Waes 2018-07-12 15:35:15 CEST

Summary: NVIDIA GPU "fallen off the bus" => NVIDIA GPU "fallen off the bus" - upgrading to 4.14.50 broke the GPU

Comment 1 Morgan Leijström 2018-07-13 09:23:24 CEST

FYI, there is now kernel 4.14.55-desktop in updates testing, i am running it OK now on a Nvidia-only system.

CC: (none) => fri

Comment 2 Jérôme Hénin 2018-07-13 10:10:46 CEST

Thanks for the hint Morgan. I've just tried 4.14.55, it has the same problem as 4.14.50 on my setup.

Comment 3 Thomas Backlund 2018-07-18 00:48:49 CEST

There is now a 4.14.56 in testing:
https://bugs.mageia.org/show_bug.cgi?id=23315

And a new nvidia-current:
https://bugs.mageia.org/show_bug.cgi?id=23316

Does any of them solve the issue for you ?

CC: (none) => tmb

Comment 4 Jérôme Hénin 2018-07-18 10:32:35 CEST

Problem is solved with the new kernel. Thank you Thomas.

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

Comment 5 Jérôme Hénin 2018-07-18 21:26:27 CEST

I'm sorry, I must have missed something (not sure how). The problem is still there with 4.14.56.

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

Comment 6 Jérôme Hénin 2019-03-24 19:13:09 CET

Problem is fixed with current kernel and nvidia driver.

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