[PATCH] drm: nova-drm: fix 32-bit arm build

[PATCH] drm: nova-drm: fix 32-bit arm build

[Miguel Ojeda]

In 32-bit arm, the build fails with:

    error[E0308]: mismatched types
      --> drivers/gpu/drm/nova/file.rs:42:28
       |
    42 |         getparam.set_value(value);
       |                  --------- ^^^^^ expected `u64`, found `u32`
       |                  |
       |                  arguments to this method are incorrect
       |
    note: method defined here
      --> drivers/gpu/drm/nova/uapi.rs:29:12
       |
    29 |     pub fn set_value(&self, v: u64) {
       |            ^^^^^^^^^        ------
    help: you can convert a `u32` to a `u64`
       |
    42 |         getparam.set_value(value.into());
       |                                 +++++++

The reason is that `Getparam::set_value` takes a `u64` (from the UAPI),
but `pci::Device::resource_len()` returns a `resource_size_t`, which is a
`phys_addr_t`, which may be 32- or 64-bit.

Thus add an `into()` call to support the 32-bit case, while allowing the
Clippy lint that complains in the 64-bit case where the type is the same.

Fixes: cdeaeb9dd762 ("drm: nova-drm: add initial driver skeleton")
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
---
As discussed, it may be best to have a newtype, or at least a function
to perform this -- here it is the minimal fix nevertheless.

 drivers/gpu/drm/nova/file.rs | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/nova/file.rs b/drivers/gpu/drm/nova/file.rs
index 7e59a34b830d..4fe62cf98a23 100644
--- a/drivers/gpu/drm/nova/file.rs
+++ b/drivers/gpu/drm/nova/file.rs
@@ -39,7 +39,8 @@ pub(crate) fn get_param(
             _ => return Err(EINVAL),
         };

-        getparam.set_value(value);
+        #[allow(clippy::useless_conversion)]
+        getparam.set_value(value.into());

         Ok(0)
     }

base-commit: 89be9a83ccf1f88522317ce02f854f30d6115c41
--
2.50.1

Re: [PATCH] drm: nova-drm: fix 32-bit arm build

[Danilo Krummrich]

On Thu Jul 24, 2025 at 6:54 PM CEST, Miguel Ojeda wrote:
> In 32-bit arm, the build fails with:
>
>     error[E0308]: mismatched types
>       --> drivers/gpu/drm/nova/file.rs:42:28
>        |
>     42 |         getparam.set_value(value);
>        |                  --------- ^^^^^ expected `u64`, found `u32`
>        |                  |
>        |                  arguments to this method are incorrect
>        |
>     note: method defined here
>       --> drivers/gpu/drm/nova/uapi.rs:29:12
>        |
>     29 |     pub fn set_value(&self, v: u64) {
>        |            ^^^^^^^^^        ------
>     help: you can convert a `u32` to a `u64`
>        |
>     42 |         getparam.set_value(value.into());
>        |                                 +++++++
>
> The reason is that `Getparam::set_value` takes a `u64` (from the UAPI),
> but `pci::Device::resource_len()` returns a `resource_size_t`, which is a
> `phys_addr_t`, which may be 32- or 64-bit.
>
> Thus add an `into()` call to support the 32-bit case, while allowing the
> Clippy lint that complains in the 64-bit case where the type is the same.
>
> Fixes: cdeaeb9dd762 ("drm: nova-drm: add initial driver skeleton")
> Signed-off-by: Miguel Ojeda <ojeda@kernel.org>

Thanks -- will take it through -fixes once rc-1 is out.

> ---
> As discussed, it may be best to have a newtype, or at least a function
> to perform this -- here it is the minimal fix nevertheless.

I think I will follow up with a function to perform the conversion in a single
place, but I really like the idea of a special clippy annotation to tell clippy
to not warn about unnecessary into() conversions for a specific type alias, such
as ResourceSize.

Do we agree that we want something like this? Do we even have a feature request
for this already?

Re: [PATCH] drm: nova-drm: fix 32-bit arm build

[Miguel Ojeda]

On Thu, Jul 24, 2025 at 7:05 PM Danilo Krummrich <dakr@kernel.org> wrote:
>
> I think I will follow up with a function to perform the conversion in a single
> place, but I really like the idea of a special clippy annotation to tell clippy
> to not warn about unnecessary into() conversions for a specific type alias, such
> as ResourceSize.
>
> Do we agree that we want something like this? Do we even have a feature request
> for this already?

I think we should at least ask -- done here:

    https://github.com/rust-lang/rust-clippy/issues/15337

Though, thinking about it, an explicit function may provide value
nevertheless to clearly see where this happens, and it also means that
when we see `into()` we know it cannot be a no-op.

Having said that, regardless of what we do for that lint, giving more
information to the compiler is generally a good idea, even if only for
notes/diagnostics etc.

Cheers,
Miguel

Re: [PATCH] drm: nova-drm: fix 32-bit arm build

[Miguel Ojeda]

On Thu, Jul 24, 2025 at 7:05 PM Danilo Krummrich <dakr@kernel.org> wrote:
>
> Thanks -- will take it through -fixes once rc-1 is out.

By the way, in that case we should probably:

Cc: stable@vger.kernel.org

Cheers,
Miguel

Re: [PATCH] drm: nova-drm: fix 32-bit arm build

[Christian Schrefl]

Hi Miguel,

On 24.07.25 6:54 PM, Miguel Ojeda wrote:
> In 32-bit arm, the build fails with:
> 
>     error[E0308]: mismatched types
>       --> drivers/gpu/drm/nova/file.rs:42:28
>        |
>     42 |         getparam.set_value(value);
>        |                  --------- ^^^^^ expected `u64`, found `u32`
>        |                  |
>        |                  arguments to this method are incorrect
>        |
>     note: method defined here
>       --> drivers/gpu/drm/nova/uapi.rs:29:12
>        |
>     29 |     pub fn set_value(&self, v: u64) {
>        |            ^^^^^^^^^        ------
>     help: you can convert a `u32` to a `u64`
>        |
>     42 |         getparam.set_value(value.into());
>        |                                 +++++++
> 
> The reason is that `Getparam::set_value` takes a `u64` (from the UAPI),
> but `pci::Device::resource_len()` returns a `resource_size_t`, which is a
> `phys_addr_t`, which may be 32- or 64-bit.
> 
> Thus add an `into()` call to support the 32-bit case, while allowing the
> Clippy lint that complains in the 64-bit case where the type is the same.
> 
> Fixes: cdeaeb9dd762 ("drm: nova-drm: add initial driver skeleton")
> Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
> ---
> As discussed, it may be best to have a newtype, or at least a function
> to perform this -- here it is the minimal fix nevertheless.

I agree we should at least have a specific conversion function, but for now:

Reviewed-by: Christian Schrefl <chrisi.schrefl@gmail.com>

Cheers Christian

Re: [PATCH] drm: nova-drm: fix 32-bit arm build

[Danilo Krummrich]

On 7/24/25 6:54 PM, Miguel Ojeda wrote:
> In 32-bit arm, the build fails with:
> 
>      error[E0308]: mismatched types
>        --> drivers/gpu/drm/nova/file.rs:42:28
>         |
>      42 |         getparam.set_value(value);
>         |                  --------- ^^^^^ expected `u64`, found `u32`
>         |                  |
>         |                  arguments to this method are incorrect
>         |
>      note: method defined here
>        --> drivers/gpu/drm/nova/uapi.rs:29:12
>         |
>      29 |     pub fn set_value(&self, v: u64) {
>         |            ^^^^^^^^^        ------
>      help: you can convert a `u32` to a `u64`
>         |
>      42 |         getparam.set_value(value.into());
>         |                                 +++++++
> 
> The reason is that `Getparam::set_value` takes a `u64` (from the UAPI),
> but `pci::Device::resource_len()` returns a `resource_size_t`, which is a
> `phys_addr_t`, which may be 32- or 64-bit.
> 
> Thus add an `into()` call to support the 32-bit case, while allowing the
> Clippy lint that complains in the 64-bit case where the type is the same.
> 
> Fixes: cdeaeb9dd762 ("drm: nova-drm: add initial driver skeleton")
> Signed-off-by: Miguel Ojeda <ojeda@kernel.org>

Applied to drm-misc-fixes, thanks!