Re: [PATCH v1 1/3] rust: sync: atomic: Prepare AtomicOps macros for i8/i16 support

[Boqun Feng <boqun.feng@gmail.com>]

On Mon, Dec 29, 2025 at 04:36:16PM +0000, Gary Guo wrote:
> On Mon, 29 Dec 2025 19:13:11 +0800
> Boqun Feng <boqun.feng@gmail.com> wrote:
> 
> > On Sun, Dec 28, 2025 at 09:05:44PM +0900, FUJITA Tomonori wrote:
> > > Rework the internal AtomicOps macro plumbing to generate per-type
> > > implementations from a mapping list.
> > > 
> > > Capture the trait definition once and reuse it for both declaration
> > > and per-type impl expansion to reduce duplication and keep future
> > > extensions simple.
> > > 
> > > This is a preparatory refactor for enabling i8/i16 atomics cleanly.
> > > 
> > > Signed-off-by: FUJITA Tomonori <fujita.tomonori@gmail.com>  
> > 
> > Thanks! I have an idea that uses proc-macro to generate the Atomic*Ops
> > impls, e.g.
> > 
> >     #[atomic_ops(i8, i16, i32, i64)]
> >     pub trait AtomicBasicOps {
> >         #[variant(acquire)]
> >         fn read(a: &AtomicRepr<Self>) -> Self {
> > 	    unsafe { binding_call!(a.as_ptr().cast()) }
> > 	}
> >     }
> 
> Unless the proc macro is generally applicable to a wide range of subsystem
> abstractions, I would still prefer to use declarative macros.
> 

But the tt-muncher style is quite challenging for a boarder audience to
understand and change (if necessary) the declarative macros. IMO, if we
can have mod-specific proc macros (i.e. macros that are only usable in
atomic mod) then it should be fine. Thoughts?

Regards,
Boqun

> Best,
> Gary
> 
> > 
> > But I think the current solution in your patch suffices as a temporary
> > solution at least.
> > 
> > Regards,
> > Boqun