The fence made classifier_lookup() slower. Access to a size_t 'n' is safe without synchonizing, and if racing with writers matters, additional syncronization primitives are used anyway.
Signed-off-by: Jarno Rajahalme <[email protected]> --- lib/hmap.h | 1 - 1 file changed, 1 deletion(-) diff --git a/lib/hmap.h b/lib/hmap.h index 445e74f..9fb83d5 100644 --- a/lib/hmap.h +++ b/lib/hmap.h @@ -196,7 +196,6 @@ hmap_capacity(const struct hmap *hmap) static inline bool hmap_is_empty(const struct hmap *hmap) { - atomic_thread_fence(memory_order_acquire); return hmap->n == 0; } -- 1.7.10.4 _______________________________________________ dev mailing list [email protected] http://openvswitch.org/mailman/listinfo/dev
