[dpdk-stable] [PATCH] net/mlx5: fix resource release issue for mirror flow

[dpdk-stable] [PATCH] net/mlx5: fix resource release issue for mirror flow

[Jiawei Wang]

The mlx5 PMD allocated the resources of the sample actions, and then
moved these ones to the destination actions array. The original indices
were not cleared and the resources were referenced twice in the
flow object - as the fate actions and in the destination actions array.

This causes the failure on flow destroy because PMD tried to release the
same objects twice.

The patch clears the original indices, add the missed checking for zero
and eliminates multiple object releasing.

Fixes: 00c10c22118a ("net/mlx5: update translate function for mirroring")
Cc: stable@dpdk.org

Signed-off-by: Jiawei Wang <jiaweiw@nvidia.com>
Reviewed-by: Suanming Mou <suanmingm@nvidia.com>
---
 drivers/net/mlx5/mlx5_flow_dv.c | 25 ++++++-------------------
 1 file changed, 6 insertions(+), 19 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_flow_dv.c b/drivers/net/mlx5/mlx5_flow_dv.c
index 1818895..883fe2e 100644
--- a/drivers/net/mlx5/mlx5_flow_dv.c
+++ b/drivers/net/mlx5/mlx5_flow_dv.c
@@ -10148,24 +10148,8 @@ struct mlx5_cache_entry *
 	return &cache_resource->entry;
 error:
 	for (idx = 0; idx < resource->num_of_dest; idx++) {
-		struct mlx5_flow_sub_actions_idx *act_res =
-					&cache_resource->sample_idx[idx];
-		if (act_res->rix_hrxq &&
-		    !mlx5_hrxq_release(dev,
-				act_res->rix_hrxq))
-			act_res->rix_hrxq = 0;
-		if (act_res->rix_encap_decap &&
-			!flow_dv_encap_decap_resource_release(dev,
-				act_res->rix_encap_decap))
-			act_res->rix_encap_decap = 0;
-		if (act_res->rix_port_id_action &&
-			!flow_dv_port_id_action_resource_release(dev,
-				act_res->rix_port_id_action))
-			act_res->rix_port_id_action = 0;
-		if (act_res->rix_jump &&
-			!flow_dv_jump_tbl_resource_release(dev,
-				act_res->rix_jump))
-			act_res->rix_jump = 0;
+		flow_dv_sample_sub_actions_release(dev,
+				&cache_resource->sample_idx[idx]);
 		if (dest_attr[idx])
 			mlx5_free(dest_attr[idx]);
 	}
@@ -10533,6 +10517,7 @@ struct mlx5_cache_entry *
 				dev_flow->handle->dvh.rix_encap_decap;
 			sample_act->dr_encap_action =
 				dev_flow->dv.encap_decap->action;
+			dev_flow->handle->dvh.rix_encap_decap = 0;
 		}
 		if (sample_act->action_flags & MLX5_FLOW_ACTION_PORT_ID) {
 			normal_idx++;
@@ -10540,6 +10525,7 @@ struct mlx5_cache_entry *
 				dev_flow->handle->rix_port_id_action;
 			sample_act->dr_port_id_action =
 				dev_flow->dv.port_id_action->action;
+			dev_flow->handle->rix_port_id_action = 0;
 		}
 		if (sample_act->action_flags & MLX5_FLOW_ACTION_JUMP) {
 			normal_idx++;
@@ -12406,7 +12392,8 @@ struct mlx5_cache_entry *
 		return;
 	switch (handle->fate_action) {
 	case MLX5_FLOW_FATE_QUEUE:
-		mlx5_hrxq_release(dev, handle->rix_hrxq);
+		if (!handle->dvh.rix_sample && !handle->dvh.rix_dest_array)
+			mlx5_hrxq_release(dev, handle->rix_hrxq);
 		break;
 	case MLX5_FLOW_FATE_JUMP:
 		flow_dv_jump_tbl_resource_release(dev, handle->rix_jump);
-- 
1.8.3.1

Re: [dpdk-stable] [PATCH] net/mlx5: fix resource release issue for mirror flow

[Slava Ovsiienko]

> -----Original Message-----
> From: Jiawei Wang <jiaweiw@nvidia.com>
> Sent: Friday, April 9, 2021 15:33
> To: Matan Azrad <matan@nvidia.com>; Shahaf Shuler
> <shahafs@nvidia.com>; Slava Ovsiienko <viacheslavo@nvidia.com>;
> Jiawei(Jonny) Wang <jiaweiw@nvidia.com>
> Cc: dev@dpdk.org; stable@dpdk.org
> Subject: [PATCH] net/mlx5: fix resource release issue for mirror flow
> 
> The mlx5 PMD allocated the resources of the sample actions, and then
> moved these ones to the destination actions array. The original indices were
> not cleared and the resources were referenced twice in the flow object - as
> the fate actions and in the destination actions array.
> 
> This causes the failure on flow destroy because PMD tried to release the
> same objects twice.
> 
> The patch clears the original indices, add the missed checking for zero and
> eliminates multiple object releasing.
> 
> Fixes: 00c10c22118a ("net/mlx5: update translate function for mirroring")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Jiawei Wang <jiaweiw@nvidia.com>
> Reviewed-by: Suanming Mou <suanmingm@nvidia.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>

Re: [dpdk-stable] [dpdk-dev] [PATCH] net/mlx5: fix resource release issue for mirror flow

[Raslan Darawsheh]

Hi,

> -----Original Message-----
> From: dev <dev-bounces@dpdk.org> On Behalf Of Jiawei Wang
> Sent: Friday, April 9, 2021 3:33 PM
> To: Matan Azrad <matan@nvidia.com>; Shahaf Shuler
> <shahafs@nvidia.com>; Slava Ovsiienko <viacheslavo@nvidia.com>;
> Jiawei(Jonny) Wang <jiaweiw@nvidia.com>
> Cc: dev@dpdk.org; stable@dpdk.org
> Subject: [dpdk-dev] [PATCH] net/mlx5: fix resource release issue for mirror
> flow
> 
> The mlx5 PMD allocated the resources of the sample actions, and then
> moved these ones to the destination actions array. The original indices
> were not cleared and the resources were referenced twice in the
> flow object - as the fate actions and in the destination actions array.
> 
> This causes the failure on flow destroy because PMD tried to release the
> same objects twice.
> 
> The patch clears the original indices, add the missed checking for zero
> and eliminates multiple object releasing.
> 
> Fixes: 00c10c22118a ("net/mlx5: update translate function for mirroring")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Jiawei Wang <jiaweiw@nvidia.com>
> Reviewed-by: Suanming Mou <suanmingm@nvidia.com>

Patch applied to next-net-mlx,

Kindest regards,
Raslan Darawsheh