[PATCH] drm/nouveau: expose GSP-RM logging buffers via debugfs

[David Airlie]

On Tue, Feb 6, 2024 at 1:55 PM Timur Tabi <ttabi at nvidia.com> wrote:
>
> On Tue, 2024-02-06 at 10:35 +1000, David Airlie wrote:
> > >
> > > An end-user can capture the logs using the following commands:
> > >
> > >     cp /sys/kernel/debug/nouveau/loginit loginit
> > >     cp /sys/kernel/debug/nouveau/logrm logrm
> > >     cp /sys/kernel/debug/nouveau/logintr logintr
> > >     cp /sys/kernel/debug/nouveau/logpmu logpmu
> >
> > If we have 2 GPUs won't this conflict on driver load?
> >
> > Do we need to at least make subdirs or if two early in boot to have
> > ids, use the pci path?
>
> I knew I was forgetting something.  I will fix this to use per-GPU paths.  Any
> suggestion as to what to use to differentiate?  PCI ID?  Is there some other
> identifier?

I'm not sure if we've registered dri minor devices at that point to
get card numbers, so probably would have to come from the pci id (and
we'd have to fix arm later).

Dave.