[PATCH v3 21/45] vdpa/mlx5: use rte stdatomic API

[Tyler Retzlaff <roretzla@linux.microsoft.com>]

Replace the use of gcc builtin __atomic_xxx intrinsics with
corresponding rte_atomic_xxx optional rte stdatomic API.

Signed-off-by: Tyler Retzlaff <roretzla@linux.microsoft.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
---
 drivers/vdpa/mlx5/mlx5_vdpa.c         | 24 +++++++++---------
 drivers/vdpa/mlx5/mlx5_vdpa.h         | 14 +++++------
 drivers/vdpa/mlx5/mlx5_vdpa_cthread.c | 46 +++++++++++++++++------------------
 drivers/vdpa/mlx5/mlx5_vdpa_lm.c      |  4 ++-
 drivers/vdpa/mlx5/mlx5_vdpa_mem.c     |  4 ++-
 drivers/vdpa/mlx5/mlx5_vdpa_virtq.c   |  4 ++-
 6 files changed, 52 insertions(+), 44 deletions(-)

diff --git a/drivers/vdpa/mlx5/mlx5_vdpa.c b/drivers/vdpa/mlx5/mlx5_vdpa.c
index f900384..98c39a5 100644
--- a/drivers/vdpa/mlx5/mlx5_vdpa.c
+++ b/drivers/vdpa/mlx5/mlx5_vdpa.c
@@ -261,8 +261,8 @@
 	uint32_t timeout = 0;
 
 	/* Check and wait all close tasks done. */
-	while (__atomic_load_n(&priv->dev_close_progress,
-		__ATOMIC_RELAXED) != 0 && timeout < 1000) {
+	while (rte_atomic_load_explicit(&priv->dev_close_progress,
+		rte_memory_order_relaxed) != 0 && timeout < 1000) {
 		rte_delay_us_sleep(10000);
 		timeout++;
 	}
@@ -294,8 +294,8 @@
 			priv->last_c_thrd_idx = 0;
 		else
 			priv->last_c_thrd_idx++;
-		__atomic_store_n(&priv->dev_close_progress,
-			1, __ATOMIC_RELAXED);
+		rte_atomic_store_explicit(&priv->dev_close_progress,
+			1, rte_memory_order_relaxed);
 		if (mlx5_vdpa_task_add(priv,
 			priv->last_c_thrd_idx,
 			MLX5_VDPA_TASK_DEV_CLOSE_NOWAIT,
@@ -319,8 +319,8 @@
 	if (!priv->connected)
 		mlx5_vdpa_dev_cache_clean(priv);
 	priv->vid = 0;
-	__atomic_store_n(&priv->dev_close_progress, 0,
-		__ATOMIC_RELAXED);
+	rte_atomic_store_explicit(&priv->dev_close_progress, 0,
+		rte_memory_order_relaxed);
 	priv->state = MLX5_VDPA_STATE_PROBED;
 	DRV_LOG(INFO, "vDPA device %d was closed.", vid);
 	return ret;
@@ -664,7 +664,9 @@
 static int
 mlx5_vdpa_virtq_resource_prepare(struct mlx5_vdpa_priv *priv)
 {
-	uint32_t remaining_cnt = 0, err_cnt = 0, task_num = 0;
+	RTE_ATOMIC(uint32_t) remaining_cnt = 0;
+	RTE_ATOMIC(uint32_t) err_cnt = 0;
+	uint32_t task_num = 0;
 	uint32_t max_queues, index, thrd_idx, data[1];
 	struct mlx5_vdpa_virtq *virtq;
 
@@ -847,8 +849,8 @@
 		if (conf_thread_mng.initializer_priv == priv)
 			if (mlx5_vdpa_mult_threads_create())
 				goto error;
-		__atomic_fetch_add(&conf_thread_mng.refcnt, 1,
-			__ATOMIC_RELAXED);
+		rte_atomic_fetch_add_explicit(&conf_thread_mng.refcnt, 1,
+			rte_memory_order_relaxed);
 	}
 	if (mlx5_vdpa_create_dev_resources(priv))
 		goto error;
@@ -937,8 +939,8 @@
 	if (priv->vdev)
 		rte_vdpa_unregister_device(priv->vdev);
 	if (priv->use_c_thread)
-		if (__atomic_fetch_sub(&conf_thread_mng.refcnt,
-			1, __ATOMIC_RELAXED) == 1)
+		if (rte_atomic_fetch_sub_explicit(&conf_thread_mng.refcnt,
+			1, rte_memory_order_relaxed) == 1)
 			mlx5_vdpa_mult_threads_destroy(true);
 	rte_free(priv);
 }
diff --git a/drivers/vdpa/mlx5/mlx5_vdpa.h b/drivers/vdpa/mlx5/mlx5_vdpa.h
index 7b37c98..0cc67ed 100644
--- a/drivers/vdpa/mlx5/mlx5_vdpa.h
+++ b/drivers/vdpa/mlx5/mlx5_vdpa.h
@@ -93,8 +93,8 @@ enum mlx5_vdpa_task_type {
 struct mlx5_vdpa_task {
 	struct mlx5_vdpa_priv *priv;
 	enum mlx5_vdpa_task_type type;
-	uint32_t *remaining_cnt;
-	uint32_t *err_cnt;
+	RTE_ATOMIC(uint32_t) *remaining_cnt;
+	RTE_ATOMIC(uint32_t) *err_cnt;
 	uint32_t idx;
 } __rte_packed __rte_aligned(4);
 
@@ -107,7 +107,7 @@ struct mlx5_vdpa_c_thread {
 
 struct mlx5_vdpa_conf_thread_mng {
 	void *initializer_priv;
-	uint32_t refcnt;
+	RTE_ATOMIC(uint32_t) refcnt;
 	uint32_t max_thrds;
 	pthread_mutex_t cthrd_lock;
 	struct mlx5_vdpa_c_thread cthrd[MLX5_VDPA_MAX_C_THRD];
@@ -212,7 +212,7 @@ struct mlx5_vdpa_priv {
 	uint64_t features; /* Negotiated features. */
 	uint16_t log_max_rqt_size;
 	uint16_t last_c_thrd_idx;
-	uint16_t dev_close_progress;
+	RTE_ATOMIC(uint16_t) dev_close_progress;
 	uint16_t num_mrs; /* Number of memory regions. */
 	struct mlx5_vdpa_steer steer;
 	struct mlx5dv_var *var;
@@ -581,13 +581,13 @@ int mlx5_vdpa_dirty_bitmap_set(struct mlx5_vdpa_priv *priv, uint64_t log_base,
 mlx5_vdpa_task_add(struct mlx5_vdpa_priv *priv,
 		uint32_t thrd_idx,
 		enum mlx5_vdpa_task_type task_type,
-		uint32_t *remaining_cnt, uint32_t *err_cnt,
+		RTE_ATOMIC(uint32_t) *remaining_cnt, RTE_ATOMIC(uint32_t) *err_cnt,
 		void **task_data, uint32_t num);
 int
 mlx5_vdpa_register_mr(struct mlx5_vdpa_priv *priv, uint32_t idx);
 bool
-mlx5_vdpa_c_thread_wait_bulk_tasks_done(uint32_t *remaining_cnt,
-		uint32_t *err_cnt, uint32_t sleep_time);
+mlx5_vdpa_c_thread_wait_bulk_tasks_done(RTE_ATOMIC(uint32_t) *remaining_cnt,
+		RTE_ATOMIC(uint32_t) *err_cnt, uint32_t sleep_time);
 int
 mlx5_vdpa_virtq_setup(struct mlx5_vdpa_priv *priv, int index, bool reg_kick);
 void
diff --git a/drivers/vdpa/mlx5/mlx5_vdpa_cthread.c b/drivers/vdpa/mlx5/mlx5_vdpa_cthread.c
index 68ed841..84f611c 100644
--- a/drivers/vdpa/mlx5/mlx5_vdpa_cthread.c
+++ b/drivers/vdpa/mlx5/mlx5_vdpa_cthread.c
@@ -48,7 +48,7 @@
 mlx5_vdpa_task_add(struct mlx5_vdpa_priv *priv,
 		uint32_t thrd_idx,
 		enum mlx5_vdpa_task_type task_type,
-		uint32_t *remaining_cnt, uint32_t *err_cnt,
+		RTE_ATOMIC(uint32_t) *remaining_cnt, RTE_ATOMIC(uint32_t) *err_cnt,
 		void **task_data, uint32_t num)
 {
 	struct rte_ring *rng = conf_thread_mng.cthrd[thrd_idx].rng;
@@ -70,8 +70,8 @@
 		return -1;
 	for (i = 0 ; i < num; i++)
 		if (task[i].remaining_cnt)
-			__atomic_fetch_add(task[i].remaining_cnt, 1,
-				__ATOMIC_RELAXED);
+			rte_atomic_fetch_add_explicit(task[i].remaining_cnt, 1,
+				rte_memory_order_relaxed);
 	/* wake up conf thread. */
 	pthread_mutex_lock(&conf_thread_mng.cthrd_lock);
 	pthread_cond_signal(&conf_thread_mng.cthrd[thrd_idx].c_cond);
@@ -80,16 +80,16 @@
 }
 
 bool
-mlx5_vdpa_c_thread_wait_bulk_tasks_done(uint32_t *remaining_cnt,
-		uint32_t *err_cnt, uint32_t sleep_time)
+mlx5_vdpa_c_thread_wait_bulk_tasks_done(RTE_ATOMIC(uint32_t) *remaining_cnt,
+		RTE_ATOMIC(uint32_t) *err_cnt, uint32_t sleep_time)
 {
 	/* Check and wait all tasks done. */
-	while (__atomic_load_n(remaining_cnt,
-		__ATOMIC_RELAXED) != 0) {
+	while (rte_atomic_load_explicit(remaining_cnt,
+		rte_memory_order_relaxed) != 0) {
 		rte_delay_us_sleep(sleep_time);
 	}
-	if (__atomic_load_n(err_cnt,
-		__ATOMIC_RELAXED)) {
+	if (rte_atomic_load_explicit(err_cnt,
+		rte_memory_order_relaxed)) {
 		DRV_LOG(ERR, "Tasks done with error.");
 		return true;
 	}
@@ -137,8 +137,8 @@
 			if (ret) {
 				DRV_LOG(ERR,
 				"Failed to register mr %d.", task.idx);
-				__atomic_fetch_add(task.err_cnt, 1,
-				__ATOMIC_RELAXED);
+				rte_atomic_fetch_add_explicit(task.err_cnt, 1,
+				rte_memory_order_relaxed);
 			}
 			break;
 		case MLX5_VDPA_TASK_SETUP_VIRTQ:
@@ -149,8 +149,8 @@
 			if (ret) {
 				DRV_LOG(ERR,
 					"Failed to setup virtq %d.", task.idx);
-				__atomic_fetch_add(
-					task.err_cnt, 1, __ATOMIC_RELAXED);
+				rte_atomic_fetch_add_explicit(
+					task.err_cnt, 1, rte_memory_order_relaxed);
 			}
 			virtq->enable = 1;
 			pthread_mutex_unlock(&virtq->virtq_lock);
@@ -164,9 +164,9 @@
 				DRV_LOG(ERR,
 				"Failed to stop virtq %d.",
 				task.idx);
-				__atomic_fetch_add(
+				rte_atomic_fetch_add_explicit(
 					task.err_cnt, 1,
-					__ATOMIC_RELAXED);
+					rte_memory_order_relaxed);
 				pthread_mutex_unlock(&virtq->virtq_lock);
 				break;
 			}
@@ -176,9 +176,9 @@
 				DRV_LOG(ERR,
 		"Failed to get negotiated features virtq %d.",
 				task.idx);
-				__atomic_fetch_add(
+				rte_atomic_fetch_add_explicit(
 					task.err_cnt, 1,
-					__ATOMIC_RELAXED);
+					rte_memory_order_relaxed);
 				pthread_mutex_unlock(&virtq->virtq_lock);
 				break;
 			}
@@ -200,9 +200,9 @@
 			if (!priv->connected)
 				mlx5_vdpa_dev_cache_clean(priv);
 			priv->vid = 0;
-			__atomic_store_n(
+			rte_atomic_store_explicit(
 				&priv->dev_close_progress, 0,
-				__ATOMIC_RELAXED);
+				rte_memory_order_relaxed);
 			break;
 		case MLX5_VDPA_TASK_PREPARE_VIRTQ:
 			ret = mlx5_vdpa_virtq_single_resource_prepare(
@@ -211,9 +211,9 @@
 				DRV_LOG(ERR,
 				"Failed to prepare virtq %d.",
 				task.idx);
-				__atomic_fetch_add(
+				rte_atomic_fetch_add_explicit(
 				task.err_cnt, 1,
-				__ATOMIC_RELAXED);
+				rte_memory_order_relaxed);
 			}
 			break;
 		default:
@@ -222,8 +222,8 @@
 			break;
 		}
 		if (task.remaining_cnt)
-			__atomic_fetch_sub(task.remaining_cnt,
-			1, __ATOMIC_RELAXED);
+			rte_atomic_fetch_sub_explicit(task.remaining_cnt,
+			1, rte_memory_order_relaxed);
 	}
 	return 0;
 }
diff --git a/drivers/vdpa/mlx5/mlx5_vdpa_lm.c b/drivers/vdpa/mlx5/mlx5_vdpa_lm.c
index 0fa671f..a207734 100644
--- a/drivers/vdpa/mlx5/mlx5_vdpa_lm.c
+++ b/drivers/vdpa/mlx5/mlx5_vdpa_lm.c
@@ -92,7 +92,9 @@
 int
 mlx5_vdpa_lm_log(struct mlx5_vdpa_priv *priv)
 {
-	uint32_t remaining_cnt = 0, err_cnt = 0, task_num = 0;
+	RTE_ATOMIC(uint32_t) remaining_cnt = 0;
+	RTE_ATOMIC(uint32_t) err_cnt = 0;
+	uint32_t task_num = 0;
 	uint32_t i, thrd_idx, data[1];
 	struct mlx5_vdpa_virtq *virtq;
 	uint64_t features;
diff --git a/drivers/vdpa/mlx5/mlx5_vdpa_mem.c b/drivers/vdpa/mlx5/mlx5_vdpa_mem.c
index e333f0b..4dfe800 100644
--- a/drivers/vdpa/mlx5/mlx5_vdpa_mem.c
+++ b/drivers/vdpa/mlx5/mlx5_vdpa_mem.c
@@ -279,7 +279,9 @@
 	uint8_t mode = 0;
 	int ret = -rte_errno;
 	uint32_t i, thrd_idx, data[1];
-	uint32_t remaining_cnt = 0, err_cnt = 0, task_num = 0;
+	RTE_ATOMIC(uint32_t) remaining_cnt = 0;
+	RTE_ATOMIC(uint32_t) err_cnt = 0;
+	uint32_t task_num = 0;
 	struct rte_vhost_memory *mem = mlx5_vdpa_vhost_mem_regions_prepare
 			(priv->vid, &mode, &priv->vmem_info.size,
 			&priv->vmem_info.gcd, &priv->vmem_info.entries_num);
diff --git a/drivers/vdpa/mlx5/mlx5_vdpa_virtq.c b/drivers/vdpa/mlx5/mlx5_vdpa_virtq.c
index 607e290..093cdd0 100644
--- a/drivers/vdpa/mlx5/mlx5_vdpa_virtq.c
+++ b/drivers/vdpa/mlx5/mlx5_vdpa_virtq.c
@@ -666,7 +666,9 @@
 {
 	int ret = rte_vhost_get_negotiated_features(priv->vid, &priv->features);
 	uint16_t nr_vring = rte_vhost_get_vring_num(priv->vid);
-	uint32_t remaining_cnt = 0, err_cnt = 0, task_num = 0;
+	RTE_ATOMIC(uint32_t) remaining_cnt = 0;
+	RTE_ATOMIC(uint32_t) err_cnt = 0;
+	uint32_t task_num = 0;
 	uint32_t i, thrd_idx, data[1];
 	struct mlx5_vdpa_virtq *virtq;
 	struct rte_vhost_vring vq;
-- 
1.8.3.1