net/mlx5: Fix error handling when adding flow rules

If building match list or adding existing fg fails when
node is locked, function returned without unlocking it.
This happened if node version changed or adding existing fg
returned with EAGAIN after jumping to search_again_locked label.

Fixes: bd71b08 ("net/mlx5: Support multiple updates of steering rules in parallel")
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Reviewed-by: Maor Gottlieb <maorg@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

drivers/net/ethernet/mellanox/mlx5/core/fs_core.c

@@ -1759,8 +1759,11 @@ _mlx5_add_flow_rules(struct mlx5_flow_table *ft,
 
 	/* Collect all fgs which has a matching match_criteria */
 	err = build_match_list(&match_head, ft, spec);
-	if (err)
+	if (err) {
+		if (take_write)
+			up_write_ref_node(&ft->node);
 		return ERR_PTR(err);
+	}
 
 	if (!take_write)
 		up_read_ref_node(&ft->node);
@@ -1769,8 +1772,11 @@ _mlx5_add_flow_rules(struct mlx5_flow_table *ft,
 				      dest_num, version);
 	free_match_list(&match_head);
 	if (!IS_ERR(rule) ||
-	    (PTR_ERR(rule) != -ENOENT && PTR_ERR(rule) != -EAGAIN))
+	    (PTR_ERR(rule) != -ENOENT && PTR_ERR(rule) != -EAGAIN)) {
+		if (take_write)
+			up_write_ref_node(&ft->node);
 		return rule;
+	}
 
 	if (!take_write) {
 		nested_down_write_ref_node(&ft->node, FS_LOCK_GRANDPARENT);