Re: [PATCH 10/10] rust: io: register: remove relative registers

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [Low] Broken intra-doc link for `WithBase` in module documentation.
- [Low] Leftover macro syntax for relative register aliases (`+ $alias_offset`) causes confusing internal macro errors.
--

--- Patch [10]: [PATCH 10/10] rust: io: register: remove relative registers ---
commit 9b1276cccc7f1b5ecb5c4e81e0c27d5b7d2f2492
Author: Gary Guo <gary@garyguo.net>

rust: io: register: remove relative registers

Relative registers can be better served by projection to subregion instead
of ad-hoc handling in register macro. Projection composes better (e.g. it
natively allows relative registers of relative registers without needing
additional support).

> diff --git a/rust/kernel/io/register.rs b/rust/kernel/io/register.rs
> index df47c28ba4f40..88ead1c1874fb 100644
> --- a/rust/kernel/io/register.rs
> +++ b/rust/kernel/io/register.rs
> @@ -191,78 +191,6 @@ fn offset(self) -> usize {
>      }
>  }
>  
> -/// Trait providing a base address to be added to the offset of a relative register to obtain
> -/// its actual offset.
[ ... ]
> -/// Trait implemented by all registers that are relative to a base.
> -pub trait WithBase {

[Severity: Low]
With WithBase removed here, does the register.rs module documentation need
to be updated? The module doc still references it in its intra-doc links:

//! items you might want to import from this module are [`WithBase`] and [`Array`].

Since kernel rustdoc compilation is strict about broken intra-doc links, could
this cause the rustdoc build to fail?

[ ... ]

> @@ -847,7 +479,7 @@ macro_rules! register {
>          $(
>              $(#[$attr:meta])* $vis:vis $name:ident ($storage:ty)
>                  $([ $size:expr $(, stride = $stride:expr)? ])?
> -                $(@ $($base:ident +)? $offset:literal)?
> +                $(@ $offset:literal)?
>                  $(=> $alias:ident $(+ $alias_offset:ident)? $([$alias_idx:expr])? )?
>              { $($fields:tt)* }
>          )*
> @@ -855,7 +487,7 @@ macro_rules! register {
>          $(
>          $crate::register!(
>              @reg [$reg_base] $(#[$attr])* $vis $name ($storage) $([$size $(, stride = $stride)?])?
> -                $(@ $($base +)? $offset)?
> +                $(@ $offset)?
>                  $(=> $alias $(+ $alias_offset)? $([$alias_idx])? )?

[Severity: Low]
Does the macro entry point still need to parse $(+ $alias_offset:ident)? and
forward $(+ $alias_offset)? now that the internal @reg macro arms handling
relative alias syntax have been removed?

If someone accidentally defines a register using the removed relative alias
syntax (e.g., => Base + Alias), could this cause a confusing internal macro
matching error instead of a clean rejection?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260721-typed_register-v1-0-452d72b60262@garyguo.net?part=10