Re: [PATCH v4 2/2] gpu: nova-core: use SizeConstants trait for u64 size constants

[Alexandre Courbot]

On Fri Apr 3, 2026 at 12:05 AM JST, Joel Fernandes wrote:
>
>
> On 4/2/2026 10:59 AM, Joel Fernandes wrote:
>> 
>> 
>> On 4/2/2026 10:55 AM, Alexandre Courbot wrote:
>>> On Thu Apr 2, 2026 at 11:33 PM JST, Joel Fernandes wrote:
>>>>
>>>>
>>>> On 4/1/2026 6:27 PM, John Hubbard wrote:
>>>>> Replace manual usize-to-u64 conversions of SZ_* constants with the
>>>>> SizeConstants trait's associated constants on u64. With the
>>>>> SizeConstants trait in scope, u64::SZ_1M replaces usize_as_u64(SZ_1M)
>>>>> and similar.
>>>>>
>>>>> This removes several now-unused imports: usize_as_u64, FromSafeCast,
>>>>> and individual SZ_* type-level constants.
>>>>>
>>>>> Reviewed-by: Eliot Courtney <ecourtney@xxxxxxxxxx>
>>>>> Signed-off-by: John Hubbard <jhubbard@xxxxxxxxxx>
>>>>
>>>> Reviewed-by: Joel Fernandes <joelagnelf@xxxxxxxxxx>
>>>>
>>>> I am very happy to see this series since I have these issues with mm
>>>> patches quite a bit, which this series will resolve. Should I update the
>>>> nova mm series to use these, or wait for it to be merged? The issue with
>>>> updating the nova mm series is, then I'll have to either:
>>>>
>>>>  a) carry the series as a dependency.
>>>>  b) wait for it to merge into drm-rust-next and rebase
>>>>
>>>> I could do a) and add a [REFERENCE] prefix to the patches, similar to how I
>>>> did for some drm buddy bindings patches. Thoughts?
>>>
>>> There is no need to carry the series into yours, just add a link to the
>>> cover letter to your dependencies and folks can pick them up.
>>>
>>> b4 also provides a dependency management mechanism, and will even fetch
>>> them automatically if the people applying your series also use it.
>> 
>> Oooh, b4 can actually track dependencies not in any tree as well? I have to
>> check that out.
>> 
>
> There is also the issue of if we sure we are going forward with this as the
> final api (I am inclined to believe it is). Trying to avoid reworking too
> much, I'd rather wait in that case for it to be confirmed as being slated
> for merge.

Yup that's probably simpler for now given that not having this series is
a mere inconvenience, as opposed to a hard blocker.