powerpc: Make value-returning atomics fully ordered
authorBoqun Feng <boqun.feng@gmail.com>
Mon, 2 Nov 2015 01:30:31 +0000 (09:30 +0800)
committerBen Hutchings <ben@decadent.org.uk>
Sat, 13 Feb 2016 10:34:05 +0000 (10:34 +0000)
commit 49e9cf3f0c04bf76ffa59242254110309554861d upstream.

According to memory-barriers.txt:

> Any atomic operation that modifies some state in memory and returns
> information about the state (old or new) implies an SMP-conditional
> general memory barrier (smp_mb()) on each side of the actual
> operation ...

Which mean these operations should be fully ordered. However on PPC,
PPC_ATOMIC_ENTRY_BARRIER is the barrier before the actual operation,
which is currently "lwsync" if SMP=y. The leading "lwsync" can not
guarantee fully ordered atomics, according to Paul Mckenney:


To fix this, we define PPC_ATOMIC_ENTRY_BARRIER as "sync" to guarantee
the fully-ordered semantics.

This also makes futex atomics fully ordered, which can avoid possible
memory ordering problems if userspace code relies on futex system call
for fully ordered semantics.

Fixes: b97021f85517 ("powerpc: Fix atomic_xxx_return barrier semantics")
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>

index e682a71..c508686 100644 (file)
@@ -44,7 +44,7 @@ static inline void isync(void)
        MAKE_LWSYNC_SECTION_ENTRY(97, __lwsync_fixup);
 #define PPC_ACQUIRE_BARRIER     "\n" stringify_in_c(__PPC_ACQUIRE_BARRIER)
 #define PPC_RELEASE_BARRIER     stringify_in_c(LWSYNC) "\n"
-#define PPC_ATOMIC_ENTRY_BARRIER "\n" stringify_in_c(LWSYNC) "\n"
+#define PPC_ATOMIC_ENTRY_BARRIER "\n" stringify_in_c(sync) "\n"
 #define PPC_ATOMIC_EXIT_BARRIER         "\n" stringify_in_c(sync) "\n"