Re: [PATCH v7 4/9] rust: sync: atomic: Add generic atomics

On Tue Jul 15, 2025 at 3:14 PM CEST, Boqun Feng wrote:
> On Tue, Jul 15, 2025 at 11:36:49AM +0200, Benno Lossin wrote:
>> On Mon Jul 14, 2025 at 5:32 PM CEST, Boqun Feng wrote:
>> > On Mon, Jul 14, 2025 at 05:05:40PM +0200, Benno Lossin wrote:
>> > [...]
>> >> >> >  //!
>> >> >> >  //! [`LKMM`]: srctree/tools/memory-model/
>> >> >> >  
>> >> >> > +pub mod generic;
>> >> >> 
>> >> >> Hmm, maybe just re-export the stuff? I don't think there's an advantage
>> >> >> to having the generic module be public.
>> >> >> 
>> >> >
>> >> > If `generic` is not public, then in the kernel::sync::atomic page, it
>> >> > won't should up, and there is no mentioning of struct `Atomic` either.
>> >> >
>> >> > If I made it public and also re-export the `Atomic`, there would be a
>> >> > "Re-export" section mentioning all the re-exports, so I will keep
>> >> > `generic` unless you have some tricks that I don't know.
>> >> 
>> >> Just use `#[doc(inline)]` :)
>> >> 
>> >>     https://doc.rust-lang.org/rustdoc/write-documentation/the-doc-attribute.html#inline-and-no_inline
>> >> 
>> >> > Also I feel it's a bit naturally that `AllowAtomic` and `AllowAtomicAdd`
>> >> > stay under `generic` (instead of re-export them at `atomic` mod level)
>> >> > because they are about the generic part of `Atomic`, right?
>> >> 
>> >> Why is that more natural? It only adds an extra path layer in any import
>> >> for atomics.
>> >> 
>> >
>> > Exactly, users need to go through extra steps if they want to use the
>> > "generic" part of the atomic, and I think that makes user more aware of
>> > what they are essentially doing:
>> >
>> > - If you want to use the predefined types for atomic, just
>> >
>> >   use kernel::sync::atomic::Atomic;
>> >
>> >   and just operate on an `Atomic<_>`.
>> >
>> > - If you want to bring your own type for atomic operations, you need to
>> >
>> >   use kernel::sync::atomic::generic::AllowAtomic;
>> >
>> >   (essentially you go into the "generic" part of the atomic)
>> >
>> >   and provide your own implementation for `AllowAtomic` and then you
>> >   could use it for your own type.
>> >
>> > I feel it's natural because for extra features you fetch more modules
>> > in.
>> 
>> The same would apply if you had to import `AllowAtomic` from atomic
>> directly? I don't really see the value in this.
>> 
>
> Because generic::AllowAtomic is more clear that the user is using the
> generic part of the API, or the user is extending the Atomic type with
> the generic ability. Grouping functionality with a namespace is
> basically what I want to achieve here. It's similar to why do we put
> `atomic` (and `lock`) under `sync`.

For `sync::{atomic, lock}` this makes sense, because `sync` might get
other submodules in the future (eg `once`). But for atomics, I don't see
any new modules similar to `generic` and even if, `AllowAtomic` still
makes sense in the top-level atomic module. I don't think the
namespacing argument makes sense in this case.

---
Cheers,
Benno