On Sun, Jun 16, 2024 at 07:16:30AM -0700, Boqun Feng wrote:
> On Sun, Jun 16, 2024 at 05:51:07AM -0400, Kent Overstreet wrote:
> > On Sat, Jun 15, 2024 at 03:12:33PM -0700, Boqun Feng wrote:
> > > What's the issue of having AtomicI32 and AtomicI64 first then? We don't
> > > need to do 1 or 2 until the real users show up.
> > >
> > > And I'd like also to point out that there are a few more trait bound
> > > designs needed for Atomic<T>, for example, Atomic<u32> and Atomic<i32>
> > > have different sets of API (no inc_unless_negative() for u32).
> > >
> > > Don't make me wrong, I have no doubt we can handle this in the type
> > > system, but given the design work need, won't it make sense that we take
> > > baby steps on this? We can first introduce AtomicI32 and AtomicI64 which
> > > already have real users, and then if there are some values of generic
> > > atomics, we introduce them and have proper discussion on design.
> > >
> > > To me, it's perfectly fine that Atomic{I32,I64} co-exist with Atomic<T>.
> > > What's the downside? A bit specific example would help me understand
> > > the real concern here.
> >
> > Err, what?
> >
> > Of course we want generic atomics, and we need that for properly
> > supporting cmpxchg.
> >
>
> Nope. Note this series only introduces the atomic types (atomic_ C
> APIs), but cmpxchg C APIs (no atomic_ prefix) are probably presented via
> a different API, where we need to make it easier to interact with normal
> types, and we may use generic there.
>
Or it could be a generic function instead of generic type like:
pub unsafe fn cmpxchg<T>(ptr: * mut T, old: T, new T) -> T
the "unsafe" part is due to `ptr` may not be a valid pointer or this may
make normal accesses data race.
Regards,
Boqun
> > Bogun, you've got all the rust guys pushing for doing this with
> > generics, I'm not sure why you're being stubborn here?
>
> Hmm? Have you seen the email I replied to John, a broader Rust community
> seems doesn't appreciate the idea of generic atomics.
>
> Regards,
> Boqun
