[PATCH v1 2/2] perf auxtrace: Optimize barriers with load-acquire and store-release

[Leo Yan]

Load-acquire and store-release are one-way permeable barriers, which can
be used to guarantee the memory ordering between accessing the buffer
data and the buffer's head / tail.

This patch optimizes the memory ordering with the load-acquire and
store-release barriers.

If the load-acquire is not supported by the architectures, it uses the
existed READ_ONCE() + smp_rmb() pair rather than use the fallback
definition of smp_load_acquire().  The reason is smp_rmb() is a minor
improvement than smp_mb() which is called in the fallback macro
smp_load_acquire().

In auxtrace_mmap__write_tail(), updating the tail pointer is removed; if
the store-release barrier is not supported, it rollbacks to use
smp_mb() + WRITE_ONCE() pair which is defined in the fallback macro
smp_store_release() (see include/asm/barrier.h).

Signed-off-by: Leo Yan <leo.yan@xxxxxxxxxx>
---
 tools/perf/util/auxtrace.h | 21 ++++++++++++++++-----
 1 file changed, 16 insertions(+), 5 deletions(-)

diff --git a/tools/perf/util/auxtrace.h b/tools/perf/util/auxtrace.h
index 8bed284ccc82..f18070f1993f 100644
--- a/tools/perf/util/auxtrace.h
+++ b/tools/perf/util/auxtrace.h
@@ -449,16 +449,27 @@ struct auxtrace_cache;
 static inline u64 auxtrace_mmap__read_snapshot_head(struct auxtrace_mmap *mm)
 {
 	struct perf_event_mmap_page *pc = mm->userpg;
+
+#if defined(__x86_64__) || defined(__aarch64__) || defined(__powerpc64__) || \
+    defined(__ia64__) || defined(__sparc__) && defined(__arch64__)
+	return smp_load_acquire(&pc->aux_head);
+#else
 	u64 head = READ_ONCE(pc->aux_head);
 
 	/* Ensure all reads are done after we read the head */
 	smp_rmb();
 	return head;
+#endif
 }
 
 static inline u64 auxtrace_mmap__read_head(struct auxtrace_mmap *mm)
 {
 	struct perf_event_mmap_page *pc = mm->userpg;
+
+#if defined(__x86_64__) || defined(__aarch64__) || defined(__powerpc64__) || \
+    defined(__ia64__) || defined(__sparc__) && defined(__arch64__)
+	return smp_load_acquire(&pc->aux_head);
+#else
 #if BITS_PER_LONG == 64 || !defined(HAVE_SYNC_COMPARE_AND_SWAP_SUPPORT)
 	u64 head = READ_ONCE(pc->aux_head);
 #else
@@ -468,20 +479,20 @@ static inline u64 auxtrace_mmap__read_head(struct auxtrace_mmap *mm)
 	/* Ensure all reads are done after we read the head */
 	smp_rmb();
 	return head;
+#endif
 }
 
 static inline void auxtrace_mmap__write_tail(struct auxtrace_mmap *mm, u64 tail)
 {
 	struct perf_event_mmap_page *pc = mm->userpg;
-#if BITS_PER_LONG != 64 && defined(HAVE_SYNC_COMPARE_AND_SWAP_SUPPORT)
+
+#if BITS_PER_LONG == 64 || !defined(HAVE_SYNC_COMPARE_AND_SWAP_SUPPORT)
+	smp_store_release(&pc->aux_tail, tail);
+#else
 	u64 old_tail;
-#endif
 
 	/* Ensure all reads are done before we write the tail out */
 	smp_mb();
-#if BITS_PER_LONG == 64 || !defined(HAVE_SYNC_COMPARE_AND_SWAP_SUPPORT)
-	pc->aux_tail = tail;
-#else
 	do {
 		old_tail = __sync_val_compare_and_swap(&pc->aux_tail, 0, 0);
 	} while (!__sync_bool_compare_and_swap(&pc->aux_tail, old_tail, tail));
-- 
2.25.1