Re: [PATCH 01/10] rust: io: register: allow explicit base type specification

On Wed Jul 29, 2026 at 4:12 AM JST, John Hubbard wrote:
> On 7/27/26 1:14 AM, Alexandre Courbot wrote:
>> On Mon Jul 27, 2026 at 2:27 PM JST, Eliot Courtney wrote:
>>> On Fri Jul 24, 2026 at 10:22 PM JST, Gary Guo wrote:
>>>> On Fri Jul 24, 2026 at 7:38 AM BST, Alexandre Courbot wrote:
> ...>> It also feels natural to me to specify
>>> the space that a register is in (Region), and you can decide what metho=
d
>>> you access that region (Mmio, other, etc). The cost seems to be that yo=
u
>>> need to specify some region on every register, but afaict you could mak=
e
>>> that a Region the size of the entire BAR0 space, or a non-sized Region
>>> and just have all your accesses fallible. So it's a bit noisy but otoh
>>> it's nice to think about what space each of your registers exist in.
>>=20
>> In the case of Nova that's something we might actually want to do: as
>> their name suggest, registers are split in large blocks (PBUS, PFB,
>> etc). And we are currently trying to move every register out of the
>
> Yes. The HW ref manuals have already set up an extensive taxonomy, and
> it's probably a "feature-not-a-bug" to specify which region each access
> is in, actually.
>
> For NVIDIA GPUs and other HW, that is.

Note that if we switch to this model, the register offsets declared in
Nova will be relative to their parent block - i.e. they won't match the
offsets declared in OpenRM anymore. But that's probably not a problem
since we can always look up the registers by name.