On 2026-07-07 at 04:16 +1000, Danilo Krummrich <[email protected]> wrote...
> On Mon Jul 6, 2026 at 7:34 AM CEST, Alistair Popple wrote:
> > @@ -31,7 +31,6 @@ pub(crate) struct Nova<'bound> {
> >  
> >  /// DRM registration data, accessible from ioctl handlers via the 
> > registration guard.
> >  pub(crate) struct DrmRegData<'bound> {
> > -    #[allow(dead_code)]
> 
> Forgot to mention that in the previous patch, please use expect instead.
> 
> > @@ -12,11 +12,11 @@
> >      types::CovariantForLt, //
> >  };
> >  
> > +use crate::gpu::Chipset;
> >  use crate::gpu::Gpu;
> 
> Vertical import style.
> 
> > @@ -25,6 +25,13 @@ extern "C" {
> >   */
> >  #define NOVA_GETPARAM_VRAM_BAR_SIZE        0x1
> >  
> > +/*
> > + * NOVA_GETPARAM_GPU_CHIPSET
> > + *
> > + * Query the GPU chipset architecture/implementation.
> > + */
> > +#define NOVA_GETPARAM_GPU_CHIPSET  0x2
> 
> We could provide an enum for the chipset values and use it in the definition 
> of
> nova-core's Chipset type, which guarantees a single source of truth.
> 
>       define_chipset!({
>           TU102 = uapi::NOVA_CHIPSET_TU102,
>           TU104 = uapi::NOVA_CHIPSET_TU104,
>           // ...
>       });

The only downside I can see is it couples nova-core to nova-drm UAPI, but I
don't think that's really a problem so have done this for v2.

 - Alistair

Reply via email to