Re: [PATCH v3 21/33] gpu: nova-core: gsp: add the GSP_INIT request builder

[Timur Tabi <ttabi@nvidia.com>]

On Mon, 2026-09-21 at 19:42 -0700, John Hubbard wrote:
> On 9/18/26 3:26 PM, Timur Tabi wrote:
> > On Thu, 2026-09-17 at 18:07 -0700, John Hubbard wrote:
> > > 
> > 
> > > +/// Registry entries that the driver sends to GSP-RM on every boot.
> > > +///
> > > +/// `RMSecBusResetEnable` enables PCI secondary bus reset. `RMForcePcieConfigSave` makes GSP-
> > > RM
> > > +/// preserve PCI configuration registers across any PCI reset. `RMDevidCheckIgnore` lets GSP-
> > > RM
> > > +/// boot when the PCI device id is absent from its product name database.
> > > +const REGISTRY_ENTRIES: &[(&[u8], u32)] = &[
> > > +    (b"RMSecBusResetEnable\0", 1),
> > > +    (b"RMForcePcieConfigSave\0", 1),
> > > +    (b"RMDevidCheckIgnore\0", 1),
> > 
> > FYI, we may want to rethink RMDevidCheckIgnore for Nova.  For Nouveau it was a big deal, because
> > we were always lock on an older version of GSP-RM, so there's a chance that an older firmware
> > could still support a newer GPU that isn't explicitly listed as supported in GSP-RM.
> > 
> > But considering the problems we had with GB20x and r570, we may want to make RMDevidCheckIgnore
> > not the default.
> 
> Good point, but I'm not completely sure I'm making the correct choice yet:
> should I remove RMDevidCheckIgnore here, or wait for your follow-up patch?

No, don't remove it here.  We need to investigate the impact of removing it.  It would be a
shame if this patchset randomly broke some GPU that needed this key.

> > I can submit a patch (once the dust with r615 settles) that makes it a command-line option, if
> > you like.
> 
> thanks,

I'll add it to my to-do list.