Message ID | 1440730099-29133-3-git-send-email-boqun.feng@gmail.com (mailing list archive) |
---|---|
State | Superseded, archived |
Headers | show |
On Fri, Aug 28, 2015 at 10:48:16AM +0800, Boqun Feng wrote: > Some architectures may have their special barriers for acquire, release > and fence semantics, general memory barriers(smp_mb__*_atomic()) in > __atomic_op_*() may be too strong, so arch_atomic_op_*() helpers are > introduced for architectures to provide their own version helpers to > build different variants based on _relaxed variants. > > Signed-off-by: Boqun Feng <boqun.feng@gmail.com> > --- > include/linux/atomic.h | 16 ++++++++++++++++ > 1 file changed, 16 insertions(+) > > diff --git a/include/linux/atomic.h b/include/linux/atomic.h > index 00a5763..622255b 100644 > --- a/include/linux/atomic.h > +++ b/include/linux/atomic.h > @@ -34,20 +34,33 @@ > * The idea here is to build acquire/release variants by adding explicit > * barriers on top of the relaxed variant. In the case where the relaxed > * variant is already fully ordered, no additional barriers are needed. > + * > + * Besides, if an arch has a special barrier for acquire/release, it could > + * implement its own arch_atomic_op_* and use the same framework for building > + * variants > */ > +#ifndef arch_atomic_op_acquire > #define __atomic_op_acquire(op, args...) \ > ({ \ > typeof(op##_relaxed(args)) __ret = op##_relaxed(args); \ > smp_mb__after_atomic(); \ > __ret; \ > }) > +#else > +#define __atomic_op_acquire arch_atomic_op_acquire > +#endif Not really a fan of this, its not consistent with the existing #ifndef guard style.
Hi Peter, On Fri, Aug 28, 2015 at 01:36:14PM +0200, Peter Zijlstra wrote: > On Fri, Aug 28, 2015 at 10:48:16AM +0800, Boqun Feng wrote: > > Some architectures may have their special barriers for acquire, release > > and fence semantics, general memory barriers(smp_mb__*_atomic()) in > > __atomic_op_*() may be too strong, so arch_atomic_op_*() helpers are > > introduced for architectures to provide their own version helpers to > > build different variants based on _relaxed variants. > > > > Signed-off-by: Boqun Feng <boqun.feng@gmail.com> > > --- > > include/linux/atomic.h | 16 ++++++++++++++++ > > 1 file changed, 16 insertions(+) > > > > diff --git a/include/linux/atomic.h b/include/linux/atomic.h > > index 00a5763..622255b 100644 > > --- a/include/linux/atomic.h > > +++ b/include/linux/atomic.h > > @@ -34,20 +34,33 @@ > > * The idea here is to build acquire/release variants by adding explicit > > * barriers on top of the relaxed variant. In the case where the relaxed > > * variant is already fully ordered, no additional barriers are needed. > > + * > > + * Besides, if an arch has a special barrier for acquire/release, it could > > + * implement its own arch_atomic_op_* and use the same framework for building > > + * variants > > */ > > +#ifndef arch_atomic_op_acquire > > #define __atomic_op_acquire(op, args...) \ > > ({ \ > > typeof(op##_relaxed(args)) __ret = op##_relaxed(args); \ > > smp_mb__after_atomic(); \ > > __ret; \ > > }) > > +#else > > +#define __atomic_op_acquire arch_atomic_op_acquire > > +#endif > > Not really a fan of this, its not consistent with the existing #ifndef > guard style. You suggestion is that I should: #ifndef __atomic_op_acquire #define __atomic_op_acquire(op, args...) ({ ... }) #endif ... and define powerpc specific __atomic_op_acquire in asm/atomic.h? Regards, Boqun
diff --git a/include/linux/atomic.h b/include/linux/atomic.h index 00a5763..622255b 100644 --- a/include/linux/atomic.h +++ b/include/linux/atomic.h @@ -34,20 +34,33 @@ * The idea here is to build acquire/release variants by adding explicit * barriers on top of the relaxed variant. In the case where the relaxed * variant is already fully ordered, no additional barriers are needed. + * + * Besides, if an arch has a special barrier for acquire/release, it could + * implement its own arch_atomic_op_* and use the same framework for building + * variants */ +#ifndef arch_atomic_op_acquire #define __atomic_op_acquire(op, args...) \ ({ \ typeof(op##_relaxed(args)) __ret = op##_relaxed(args); \ smp_mb__after_atomic(); \ __ret; \ }) +#else +#define __atomic_op_acquire arch_atomic_op_acquire +#endif +#ifndef arch_atomic_op_release #define __atomic_op_release(op, args...) \ ({ \ smp_mb__before_atomic(); \ op##_relaxed(args); \ }) +#else +#define __atomic_op_release arch_atomic_op_release +#endif +#ifndef arch_atomic_op_fence #define __atomic_op_fence(op, args...) \ ({ \ typeof(op##_relaxed(args)) __ret; \ @@ -56,6 +69,9 @@ smp_mb__after_atomic(); \ __ret; \ }) +#else +#define __atomic_op_fence arch_atomic_op_fence +#endif /* atomic_add_return_relaxed */ #ifndef atomic_add_return_relaxed
Some architectures may have their special barriers for acquire, release and fence semantics, general memory barriers(smp_mb__*_atomic()) in __atomic_op_*() may be too strong, so arch_atomic_op_*() helpers are introduced for architectures to provide their own version helpers to build different variants based on _relaxed variants. Signed-off-by: Boqun Feng <boqun.feng@gmail.com> --- include/linux/atomic.h | 16 ++++++++++++++++ 1 file changed, 16 insertions(+)