Re: [PATCH v1 1/2] rust: sync: atomic: Add perfromance-optimal Flag type for atomic booleans

[Boqun Feng <boqun@kernel.org>]

On Wed, Jan 28, 2026 at 09:19:14AM -0800, Boqun Feng wrote:
[...]
> > > +    #[inline(always)]
> > > +    pub fn get_mut(&mut self) -> &mut bool {
> > > +        &mut self.0.get_mut().bool_field
> > > +    }
> > > +
> > > +    /// Loads the value from the atomic flag.
> > > +    #[inline(always)]
> > > +    pub fn load<Ordering: ordering::AcquireOrRelaxed>(&self, o: Ordering) -> bool {
> > > +        self.0.load(o).bool_field
> 
> For load(), xchg() and cmpxchg(), I think we should not use
> `.bool_field`, because we know that `padding` is always zero, but
> compilers don't. Using `.bool_field` will make the compiler generate i32
> to i8 or a bit mask instruction to get the boolean. We need to implement
> `From<Flag>` for `bool` and use `.into()` here.
> 

Scratch this for a moment, I need to do some more investigation.

Regards,
Boqun

> Regards,
> Boqun
> 
> 
[..]