[PATCH rdma-next v3 01/11] RDMA/mlx5: Move device async_ctx initialization

Edward Srouji posted 11 patches 4 days, 12 hours ago
[PATCH rdma-next v3 01/11] RDMA/mlx5: Move device async_ctx initialization
Posted by Edward Srouji 4 days, 12 hours ago
From: Chiara Meiohas <cmeiohas@nvidia.com>

Move the async_ctx initialization from mlx5_mkey_cache_init() to
mlx5_ib_stage_init_init() since the async_ctx is used by both the MR
cache and DEVX.

Also add the corresponding cleanup in mlx5_ib_stage_init_cleanup() to
properly release the async_ctx resources.

Signed-off-by: Chiara Meiohas <cmeiohas@nvidia.com>
Reviewed-by: Michael Guralnik <michaelgur@nvidia.com>
Signed-off-by: Edward Srouji <edwards@nvidia.com>
---
 drivers/infiniband/hw/mlx5/main.c | 3 +++
 drivers/infiniband/hw/mlx5/mr.c   | 2 --
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index 90daa58126f4..629d09fd08bc 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -4160,6 +4160,7 @@ static const struct uapi_definition mlx5_ib_defs[] = {
 
 static void mlx5_ib_stage_init_cleanup(struct mlx5_ib_dev *dev)
 {
+	mlx5_cmd_cleanup_async_ctx(&dev->async_ctx);
 	mlx5_ib_data_direct_cleanup(dev);
 	mlx5_ib_cleanup_multiport_master(dev);
 	WARN_ON(!xa_empty(&dev->odp_mkeys));
@@ -4225,6 +4226,8 @@ static int mlx5_ib_stage_init_init(struct mlx5_ib_dev *dev)
 	if (err)
 		goto err_mp;
 
+	mlx5_cmd_init_async_ctx(mdev, &dev->async_ctx);
+
 	return 0;
 err_mp:
 	mlx5_ib_cleanup_multiport_master(dev);
diff --git a/drivers/infiniband/hw/mlx5/mr.c b/drivers/infiniband/hw/mlx5/mr.c
index 325fa04cbe8a..56687add34f7 100644
--- a/drivers/infiniband/hw/mlx5/mr.c
+++ b/drivers/infiniband/hw/mlx5/mr.c
@@ -979,7 +979,6 @@ int mlx5_mkey_cache_init(struct mlx5_ib_dev *dev)
 		return -ENOMEM;
 	}
 
-	mlx5_cmd_init_async_ctx(dev->mdev, &dev->async_ctx);
 	timer_setup(&dev->delay_timer, delay_time_func, 0);
 	mlx5_mkey_cache_debugfs_init(dev);
 	mutex_lock(&cache->rb_lock);
@@ -1041,7 +1040,6 @@ void mlx5_mkey_cache_cleanup(struct mlx5_ib_dev *dev)
 	flush_workqueue(dev->cache.wq);
 
 	mlx5_mkey_cache_debugfs_cleanup(dev);
-	mlx5_cmd_cleanup_async_ctx(&dev->async_ctx);
 
 	/* At this point all entries are disabled and have no concurrent work. */
 	mlx5r_destroy_cache_entries(dev);

-- 
2.47.1