Re: [PATCH 27/33] locking/atomic: powerpc: move to ARCH_ATOMIC

[Mark Rutland]

On Tue, May 11, 2021 at 10:16:21AM +0100, Mark Rutland wrote:
> On Mon, May 10, 2021 at 10:37:47AM +0100, Mark Rutland wrote:
> > We'd like all architectures to convert to ARCH_ATOMIC, as once all
> > architectures are converted it will be possible to make significant
> > cleanups to the atomics headers, and this will make it much easier to
> > generically enable atomic functionality (e.g. debug logic in the
> > instrumented wrappers).
> > 
> > As a step towards that, this patch migrates powerpc to ARCH_ATOMIC. The
> > arch code provides arch_{atomic,atomic64,xchg,cmpxchg}*(), and common
> > code wraps these with optional instrumentation to provide the regular
> > functions.
> > 
> > Signed-off-by: Mark Rutland <mark.rutland@xxxxxxx>
> > Cc: Benjamin Herrenschmidt <benh@xxxxxxxxxxxxxxxxxxx>
> > Cc: Boqun Feng <boqun.feng@xxxxxxxxx>
> > Cc: Michael Ellerman <mpe@xxxxxxxxxxxxxx>
> > Cc: Paul Mackerras <paulus@xxxxxxxxx>
> > Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> > Cc: Will Deacon <will@xxxxxxxxxx>
> > ---
> >  arch/powerpc/Kconfig               |   1 +
> >  arch/powerpc/include/asm/atomic.h  | 140 +++++++++++++++++++------------------
> >  arch/powerpc/include/asm/cmpxchg.h |  30 ++++----
> >  3 files changed, 89 insertions(+), 82 deletions(-)
> 
> The kbuild test robot spotted a couple of bits I'd got wrong; I've noted
> those below (and both are now fixed in my kernel.org branch).
> 
> >  static __always_inline bool
> > -atomic_try_cmpxchg_lock(atomic_t *v, int *old, int new)
> > +arch_atomic_try_cmpxchg_lock(atomic_t *v, int *old, int new)
> 
> Since this isn't part of the core atomic API, and is used directly by
> powerpc's spinlock implementation, this should have stayed as-is (or we
> should use the `arch_` prefix consitently and update the spinlock code).
> 
> I've dropped the `arch_` prefix for now.

On second thought (and having hit a similar issue on m68k), I've added
the `arch_` prefix and updated the caller, to more clearly separate the
common atomic API and the arch-specific bits.

Thanks,
Mark.