Re: [PATCH v2 1/7] sched/fair: Provide u64 read for 32-bits arch helper

From: Tao Zhou
Date: Mon Jan 17 2022 - 18:50:20 EST


On Mon, Jan 17, 2022 at 07:42:04PM +0000, Vincent Donnefort wrote:
> [...]
>
> > > +/*
> > > + * u64_u32_load/u64_u32_store
> > > + *
> > > + * Use a copy of a u64 value to protect against data race. This is only
> > > + * applicable for 32-bits architectures.
> > > + */
> > > +#ifdef CONFIG_64BIT
> > > +# define u64_u32_load_copy(var, copy) var
> > > +# define u64_u32_store_copy(var, copy, val) (var = val)
> > > +#else
> > > +# define u64_u32_load_copy(var, copy) \
> > > +({ \
> > > + u64 __val, __val_copy; \
> > > + do { \
> > > + __val_copy = copy; \
> > > + /* \
> > > + * paired with u64_u32_store, ordering access \
> > > + * to var and copy. \
> > > + */ \
> > > + smp_rmb(); \
> > > + __val = var; \
> > > + } while (__val != __val_copy); \
> > > + __val; \
> > > +})
> > > +# define u64_u32_store_copy(var, copy, val) \
> > > +do { \
> > > + typeof(val) __val = (val); \
> > > + var = __val; \
> > > + /* \
> > > + * paired with u64_u32_load, ordering access to var and \
> > > + * copy. \
> > > + */ \
> > > + smp_wmb(); \
> > > + copy = __val; \
> > > +} while (0)
> >
> > Code stay there some time from me. Just from my crude review;
> > The above macro need a variable to load @var temporarily for
> > later store; that means the @copy value is from @var not @val.
> >
> > # define u64_u32_store_copy(var, copy, val) \
> > do { \
> > typeof(val) __val = (val), __var = (var); \
> > var = __val; \
> > /* \
> > * paired with u64_u32_load, ordering access to var and \
> > * copy. \
> > */ \
> > smp_wmb(); \
> > copy = __var; \
> > } while (0)
>
>
> Hi Tao,
>
> __var would then contain the previous value of @var, wouldn't it? We need
> both @var and @copy to be equal to @val.

Sorry for the noise, I'm wrong here.

>
> >
> >
> >
> > Thanks,
> > Tao