lib: use atomic thread fence recommended API

Use rte_atomic_thread_fence() instead of directly using
__atomic_thread_fence() builtin GCC intrinsic
or __rte_atomic_thread_fence() internal function.

Signed-off-by: Tyler Retzlaff <roretzla@linux.microsoft.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
Acked-by: Chengwen Feng <fengchengwen@huawei.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>

lib/distributor/rte_distributor.c

@@ -187,7 +187,7 @@ rte_distributor_return_pkt(struct rte_distributor *d,
 	}
 
 	/* Sync with distributor to acquire retptrs */
-	__atomic_thread_fence(rte_memory_order_acquire);
+	rte_atomic_thread_fence(rte_memory_order_acquire);
 	for (i = 0; i < RTE_DIST_BURST_SIZE; i++)
 		/* Switch off the return bit first */
 		buf->retptr64[i] = 0;

lib/eal/common/eal_common_trace.c

@@ -526,7 +526,7 @@ __rte_trace_point_register(rte_trace_point_t *handle, const char *name,
 
 	/* Add the trace point at tail */
 	STAILQ_INSERT_TAIL(&tp_list, tp, next);
-	__atomic_thread_fence(rte_memory_order_release);
+	rte_atomic_thread_fence(rte_memory_order_release);
 
 	/* All Good !!! */
 	return 0;

lib/eal/include/rte_mcslock.h

@@ -83,7 +83,7 @@ rte_mcslock_lock(RTE_ATOMIC(rte_mcslock_t *) *msl, rte_mcslock_t *me)
 	 * store to prev->next. Otherwise it will cause a deadlock. Need a
 	 * store-load barrier.
 	 */
-	__rte_atomic_thread_fence(rte_memory_order_acq_rel);
+	rte_atomic_thread_fence(rte_memory_order_acq_rel);
 	/* If the lock has already been acquired, it first atomically
 	 * places the node at the end of the queue and then proceeds
 	 * to spin on me->locked until the previous lock holder resets
@@ -117,7 +117,7 @@ rte_mcslock_unlock(RTE_ATOMIC(rte_mcslock_t *) *msl, RTE_ATOMIC(rte_mcslock_t *)
 		 * while-loop first. This has the potential to cause a
 		 * deadlock. Need a load barrier.
 		 */
-		__rte_atomic_thread_fence(rte_memory_order_acquire);
+		rte_atomic_thread_fence(rte_memory_order_acquire);
 		/* More nodes added to the queue by other CPUs.
 		 * Wait until the next pointer is set.
 		 */

lib/hash/rte_cuckoo_hash.c

@@ -878,7 +878,7 @@ rte_hash_cuckoo_move_insert_mw(const struct rte_hash *h,
 			/* The store to sig_current should not
 			 * move above the store to tbl_chng_cnt.
 			 */
-			__atomic_thread_fence(rte_memory_order_release);
+			rte_atomic_thread_fence(rte_memory_order_release);
 		}
 
 		/* Need to swap current/alt sig to allow later
@@ -910,7 +910,7 @@ rte_hash_cuckoo_move_insert_mw(const struct rte_hash *h,
 		/* The store to sig_current should not
 		 * move above the store to tbl_chng_cnt.
 		 */
-		__atomic_thread_fence(rte_memory_order_release);
+		rte_atomic_thread_fence(rte_memory_order_release);
 	}
 
 	curr_bkt->sig_current[curr_slot] = sig;
@@ -1403,7 +1403,7 @@ __rte_hash_lookup_with_hash_lf(const struct rte_hash *h, const void *key,
 		/* The loads of sig_current in search_one_bucket
 		 * should not move below the load from tbl_chng_cnt.
 		 */
-		__atomic_thread_fence(rte_memory_order_acquire);
+		rte_atomic_thread_fence(rte_memory_order_acquire);
 		/* Re-read the table change counter to check if the
 		 * table has changed during search. If yes, re-do
 		 * the search.
@@ -1632,7 +1632,7 @@ __rte_hash_compact_ll(const struct rte_hash *h,
 				/* The store to sig_current should
 				 * not move above the store to tbl_chng_cnt.
 				 */
-				__atomic_thread_fence(rte_memory_order_release);
+				rte_atomic_thread_fence(rte_memory_order_release);
 			}
 			last_bkt->sig_current[i] = NULL_SIGNATURE;
 			rte_atomic_store_explicit(&last_bkt->key_idx[i],
@@ -2223,7 +2223,7 @@ __bulk_lookup_lf(const struct rte_hash *h, const void **keys,
 		/* The loads of sig_current in compare_signatures
 		 * should not move below the load from tbl_chng_cnt.
 		 */
-		__atomic_thread_fence(rte_memory_order_acquire);
+		rte_atomic_thread_fence(rte_memory_order_acquire);
 		/* Re-read the table change counter to check if the
 		 * table has changed during search. If yes, re-do
 		 * the search.

lib/lpm/rte_lpm.c

@@ -1116,7 +1116,7 @@ delete_depth_big(struct __rte_lpm *i_lpm, uint32_t ip_masked,
 		 * Prevent the free of the tbl8 group from hoisting.
 		 */
 		i_lpm->lpm.tbl24[tbl24_index].valid = 0;
-		__atomic_thread_fence(__ATOMIC_RELEASE);
+		rte_atomic_thread_fence(rte_memory_order_release);
 		status = tbl8_free(i_lpm, tbl8_group_start);
 	} else if (tbl8_recycle_index > -1) {
 		/* Update tbl24 entry. */
@@ -1132,7 +1132,7 @@ delete_depth_big(struct __rte_lpm *i_lpm, uint32_t ip_masked,
 		 */
 		__atomic_store(&i_lpm->lpm.tbl24[tbl24_index], &new_tbl24_entry,
 				__ATOMIC_RELAXED);
-		__atomic_thread_fence(__ATOMIC_RELEASE);
+		rte_atomic_thread_fence(rte_memory_order_release);
 		status = tbl8_free(i_lpm, tbl8_group_start);
 	}
 #undef group_idx

lib/ring/rte_ring_c11_pvt.h

@@ -68,7 +68,7 @@ __rte_ring_move_prod_head(struct rte_ring *r, unsigned int is_sp,
 		n = max;
 
 		/* Ensure the head is read before tail */
-		__atomic_thread_fence(rte_memory_order_acquire);
+		rte_atomic_thread_fence(rte_memory_order_acquire);
 
 		/* load-acquire synchronize with store-release of ht->tail
 		 * in update_tail.
@@ -145,7 +145,7 @@ __rte_ring_move_cons_head(struct rte_ring *r, int is_sc,
 		n = max;
 
 		/* Ensure the head is read before tail */
-		__atomic_thread_fence(rte_memory_order_acquire);
+		rte_atomic_thread_fence(rte_memory_order_acquire);
 
 		/* this load-acquire synchronize with store-release of ht->tail
 		 * in update_tail.

lib/stack/rte_stack_lf_c11.h

@@ -110,7 +110,7 @@ __rte_stack_lf_pop_elems(struct rte_stack_lf_list *list,
 		 * elements are properly ordered with respect to the head
 		 * pointer read.
 		 */
-		__atomic_thread_fence(rte_memory_order_acquire);
+		rte_atomic_thread_fence(rte_memory_order_acquire);
 
 		rte_prefetch0(old_head.top);