[PATCH v6] net/mlx5: fix mutex unlock in txpp cleanup

[PATCH v6] net/mlx5: fix mutex unlock in txpp cleanup

[Chengfeng Ye]

The lock sh->txpp.mutex was not correctly released on one path
of cleanup function return, potentially causing the deadlock.

Fixes: d133f4cdb7 ("net/mlx5: create clock queue for packet pacing")
Cc: stable@dpdk.org

Signed-off-by: Chengfeng Ye <cyeaa@connect.ust.hk>
---
 drivers/net/mlx5/mlx5_txpp.c | 30 ++++++++++++------------------
 1 file changed, 12 insertions(+), 18 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_txpp.c b/drivers/net/mlx5/mlx5_txpp.c
index 4f6da9f2d1..75c08c55e2 100644
--- a/drivers/net/mlx5/mlx5_txpp.c
+++ b/drivers/net/mlx5/mlx5_txpp.c
@@ -896,7 +896,6 @@ mlx5_txpp_start(struct rte_eth_dev *dev)
 	struct mlx5_priv *priv = dev->data->dev_private;
 	struct mlx5_dev_ctx_shared *sh = priv->sh;
 	int err = 0;
-	int ret;
 
 	if (!priv->config.tx_pp) {
 		/* Packet pacing is not requested for the device. */
@@ -909,14 +908,14 @@ mlx5_txpp_start(struct rte_eth_dev *dev)
 		return 0;
 	}
 	if (priv->config.tx_pp > 0) {
-		ret = rte_mbuf_dynflag_lookup
-				(RTE_MBUF_DYNFLAG_TX_TIMESTAMP_NAME, NULL);
-		if (ret < 0)
+		err = rte_mbuf_dynflag_lookup
+			(RTE_MBUF_DYNFLAG_TX_TIMESTAMP_NAME, NULL);
+		/* No flag registered means no service needed. */
+		if (err < 0)
 			return 0;
+		err = 0;
 	}
-	ret = pthread_mutex_lock(&sh->txpp.mutex);
-	MLX5_ASSERT(!ret);
-	RTE_SET_USED(ret);
+	claim_zero(pthread_mutex_lock(&sh->txpp.mutex));
 	if (sh->txpp.refcnt) {
 		priv->txpp_en = 1;
 		++sh->txpp.refcnt;
@@ -930,9 +929,7 @@ mlx5_txpp_start(struct rte_eth_dev *dev)
 			rte_errno = -err;
 		}
 	}
-	ret = pthread_mutex_unlock(&sh->txpp.mutex);
-	MLX5_ASSERT(!ret);
-	RTE_SET_USED(ret);
+	claim_zero(pthread_mutex_unlock(&sh->txpp.mutex));
 	return err;
 }
 
@@ -950,24 +947,21 @@ mlx5_txpp_stop(struct rte_eth_dev *dev)
 {
 	struct mlx5_priv *priv = dev->data->dev_private;
 	struct mlx5_dev_ctx_shared *sh = priv->sh;
-	int ret;
 
 	if (!priv->txpp_en) {
 		/* Packet pacing is already disabled for the device. */
 		return;
 	}
 	priv->txpp_en = 0;
-	ret = pthread_mutex_lock(&sh->txpp.mutex);
-	MLX5_ASSERT(!ret);
-	RTE_SET_USED(ret);
+	claim_zero(pthread_mutex_lock(&sh->txpp.mutex));
 	MLX5_ASSERT(sh->txpp.refcnt);
-	if (!sh->txpp.refcnt || --sh->txpp.refcnt)
+	if (!sh->txpp.refcnt || --sh->txpp.refcnt) {
+		claim_zero(pthread_mutex_unlock(&sh->txpp.mutex));
 		return;
+	}
 	/* No references any more, do actual destroy. */
 	mlx5_txpp_destroy(sh);
-	ret = pthread_mutex_unlock(&sh->txpp.mutex);
-	MLX5_ASSERT(!ret);
-	RTE_SET_USED(ret);
+	claim_zero(pthread_mutex_unlock(&sh->txpp.mutex));
 }
 
 /*
-- 
2.17.1

RE: [PATCH v6] net/mlx5: fix mutex unlock in txpp cleanup

[Slava Ovsiienko]

> -----Original Message-----
> From: Chengfeng Ye <cyeaa@connect.ust.hk>
> Sent: Tuesday, November 16, 2021 16:49
> To: Matan Azrad <matan@nvidia.com>; Slava Ovsiienko
> <viacheslavo@nvidia.com>; Tzahi Oved <tzahio@nvidia.com>;
> ferruh.yigit@intel.com; Raslan Darawsheh <rasland@nvidia.com>
> Cc: dev@dpdk.org; Chengfeng Ye <cyeaa@connect.ust.hk>; stable@dpdk.org
> Subject: [PATCH v6] net/mlx5: fix mutex unlock in txpp cleanup
> 
> The lock sh->txpp.mutex was not correctly released on one path of cleanup
> function return, potentially causing the deadlock.
> 
> Fixes: d133f4cdb7 ("net/mlx5: create clock queue for packet pacing")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Chengfeng Ye <cyeaa@connect.ust.hk>
Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>

Re: [PATCH v6] net/mlx5: fix mutex unlock in txpp cleanup

[Ferruh Yigit]

On 11/16/2021 3:04 PM, Slava Ovsiienko wrote:
>> -----Original Message-----
>> From: Chengfeng Ye <cyeaa@connect.ust.hk>
>> Sent: Tuesday, November 16, 2021 16:49
>> To: Matan Azrad <matan@nvidia.com>; Slava Ovsiienko
>> <viacheslavo@nvidia.com>; Tzahi Oved <tzahio@nvidia.com>;
>> ferruh.yigit@intel.com; Raslan Darawsheh <rasland@nvidia.com>
>> Cc: dev@dpdk.org; Chengfeng Ye <cyeaa@connect.ust.hk>; stable@dpdk.org
>> Subject: [PATCH v6] net/mlx5: fix mutex unlock in txpp cleanup
>>
>> The lock sh->txpp.mutex was not correctly released on one path of cleanup
>> function return, potentially causing the deadlock.
>>
>> Fixes: d133f4cdb7 ("net/mlx5: create clock queue for packet pacing")
>> Cc: stable@dpdk.org
>>
>> Signed-off-by: Chengfeng Ye <cyeaa@connect.ust.hk>
> Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
> 

Applied to dpdk-next-net/main, thanks.