Bug#1005311: forcing dependency is not enough, at least for optirun

[Aniol Martí]

On Mon, 14 Feb 2022 20:31:33 +0100 Gianluigi Tiesi <sherpya at netfarm.it> 
wrote:
> optirun glxinfo
> 
> optirun glxinfo
> [ 1566.529702] [ERROR]Cannot access secondary GPU - error: [XORG] (EE) Unable to locate/open config directory: 
> "/etc/bumblebee/xorg.conf.d"
> 
> [ 1566.529735] [ERROR]Aborting because fallback start is disabled.
> 
> 
> in xorg log:
> 
> [  1566.527] (II) NVIDIA dlloader X Driver  470.103.01  Thu Jan  6 12:18:33 UTC 2022
> [  1566.527] (II) NVIDIA Unified Driver for all Supported NVIDIA GPUs
> [  1566.529] (EE) No devices detected.
> [  1566.529] (EE)
> Fatal server error:
> [  1566.529] (EE) no screens found(EE)
> [  1566.529] (EE)
> 
> perhaps nvidia module loads:
> 
> [ 1566.540788] nvidia 0000:01:00.0: enabling device (0000 -> 0003)
> [ 1566.540886] nvidia 0000:01:00.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=none:owns=none
> [ 1566.587576] NVRM: loading NVIDIA UNIX x86_64 Kernel Module  470.103.01  Thu Jan  6 12:10:04 UTC 2022
> 

I solved it manually setting the BusID in /etc/bumblebee/xorg.conf.nvidia.

To find the BusID run `lspci | egrep 'VGA|3D'`. You will get something 
like "03:00.0". Then, add BusID "PCI:03:00:0" to 
/etc/bumblebee/xorg.conf.nvidia (note that you need to change the dot to 
a colon).

After doing this primusrun worked again.