Re: x11/nvidia-driver update to 470.63.01 seems to break GK106GLM [Quadro K2100M]

On Mon, Aug 30, 2021 at 09:32:18AM -0700, Kevin Bowling wrote:
> On Sun, Aug 29, 2021 at 1:36 PM David Wolfskill <david@catwhisker.org> wrote:
> >
> > On Sun, Aug 29, 2021 at 01:32:19PM -0700, Kevin Bowling wrote:
> > > See if it is reproducible with a 13 or main kernel, there is a locking change
> > > which may interact differently on the different branches so that would
> > > be a good thing to know.  This may need nvidia help since the error
> > > seems to be deeper in the blob, they have a support forum where you
> > > can link your report to start that.
> >
> > I had encountered (very!) similar results on
> > head@main-n248984-08b9cc316a31 (from yesterday), so I believe it is thus
> > reproducible.
> 
> Can you attempt to build run it with this in your make.conf:
> WITH_DEBUG_PORTS=x11/nvidia-driver

Yes, but not until (much) later today: I use the laptop for access to
work (among many other things).

> Just as a sanity check please make sure you are building the port
> against the exact kernel sources that will be booted with the above.

I always do, for each kernel (stable/12, stable/13, and head), via

PORTS_MODULES+=x11/nvidia-driver-390

(which used to read "PORTS_MODULES+=x11/nvidia-driver") in /etc/src.conf.

> If there's a chance the shim code is involved we have a shot at doing
> something or finding a better lead.  Otherwise your best bet is to
> report it to nvidia support forum.  There is a report of a successful
> outcome on a Kepler architecture card so it seems you are encountering
> a more specific bug than HW generation and I'm not sure what to expect
> yet.

:-}

> https://forums.developer.nvidia.com/c/gpu-graphics/freebsd-solaris/147
> ...

Thanks!

Peace,
david
-- 
David H. Wolfskill                              david@catwhisker.org
Life is not intended to be a zero-sum game.

See https://www.catwhisker.org/~david/publickey.gpg for my public key.