[PATCH] app/testpmd: fix testpmd crash when quit with mlx5 avail_thresh enabled

[PATCH] app/testpmd: fix testpmd crash when quit with mlx5 avail_thresh enabled

[Spike Du]

When testpmd quit with mlx5 avail_thresh enabled, a rte timer handler
delays to reconfigure rx queue to re-arm this event. However at the same
time, testpmd is destroying rx queues.
It's never a valid use case for mlx5 avail_thresh. Before testpmd quit,
user should disable avail_thresh configuration to not handle the events.
This is documented in mlx5 driver guide.

To avoid the crash in such use case, check port status, if it is not
RTE_PORT_STARTED, don't process the avail_thresh event.

Fixes: 0edfc9b08316 ("app/testpmd: add Host Shaper command")

Signed-off-by: Spike Du <spiked@nvidia.com>
---
 drivers/net/mlx5/mlx5_testpmd.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/drivers/net/mlx5/mlx5_testpmd.c b/drivers/net/mlx5/mlx5_testpmd.c
index ed84583..1a9ec78 100644
--- a/drivers/net/mlx5/mlx5_testpmd.c
+++ b/drivers/net/mlx5/mlx5_testpmd.c
@@ -25,6 +25,7 @@
 
 static uint8_t host_shaper_avail_thresh_triggered[RTE_MAX_ETHPORTS];
 #define SHAPER_DISABLE_DELAY_US 100000 /* 100ms */
+extern struct rte_port *ports;
 
 /**
  * Disable the host shaper and re-arm available descriptor threshold event.
@@ -39,7 +40,15 @@
 	uint16_t port_id = port_rxq_id & 0xffff;
 	uint16_t qid = (port_rxq_id >> 16) & 0xffff;
 	struct rte_eth_rxq_info qinfo;
+	struct rte_port *port;
 
+	port = &ports[port_id];
+	if (port->port_status != RTE_PORT_STARTED) {
+		printf("%s port_status(%d) is incorrect, stop avail_thresh "
+		       "event processing.\n",
+		       __func__, port->port_status);
+		return;
+	}
 	printf("%s disable shaper\n", __func__);
 	if (rte_eth_rx_queue_info_get(port_id, qid, &qinfo)) {
 		printf("rx_queue_info_get returns error\n");
-- 
1.8.3.1

Re: [PATCH] app/testpmd: fix testpmd crash when quit with mlx5 avail_thresh enabled

[Singh, Aman Deep]

On 10/24/2022 8:44 AM, Spike Du wrote:
> When testpmd quit with mlx5 avail_thresh enabled, a rte timer handler
> delays to reconfigure rx queue to re-arm this event. However at the same
> time, testpmd is destroying rx queues.
> It's never a valid use case for mlx5 avail_thresh. Before testpmd quit,
> user should disable avail_thresh configuration to not handle the events.
> This is documented in mlx5 driver guide.
>
> To avoid the crash in such use case, check port status, if it is not
> RTE_PORT_STARTED, don't process the avail_thresh event.
>
> Fixes: 0edfc9b08316 ("app/testpmd: add Host Shaper command")
>
> Signed-off-by: Spike Du <spiked@nvidia.com>

Acked-by: Aman Singh<aman.deep.singh@intel.com>

> ---
>   drivers/net/mlx5/mlx5_testpmd.c | 9 +++++++++
>   1 file changed, 9 insertions(+)
>
> diff --git a/drivers/net/mlx5/mlx5_testpmd.c b/drivers/net/mlx5/mlx5_testpmd.c
> index ed84583..1a9ec78 100644
> --- a/drivers/net/mlx5/mlx5_testpmd.c
> +++ b/drivers/net/mlx5/mlx5_testpmd.c
> @@ -25,6 +25,7 @@
>   
>   static uint8_t host_shaper_avail_thresh_triggered[RTE_MAX_ETHPORTS];
>   #define SHAPER_DISABLE_DELAY_US 100000 /* 100ms */
> +extern struct rte_port *ports;

This might not be needed, as we have testpmd.h included.

>   
>   /**
>    * Disable the host shaper and re-arm available descriptor threshold event.
> @@ -39,7 +40,15 @@
>   	uint16_t port_id = port_rxq_id & 0xffff;
>   	uint16_t qid = (port_rxq_id >> 16) & 0xffff;
>   	struct rte_eth_rxq_info qinfo;
> +	struct rte_port *port;
>   
> +	port = &ports[port_id];
> +	if (port->port_status != RTE_PORT_STARTED) {
> +		printf("%s port_status(%d) is incorrect, stop avail_thresh "
> +		       "event processing.\n",
> +		       __func__, port->port_status);
> +		return;
> +	}
>   	printf("%s disable shaper\n", __func__);
>   	if (rte_eth_rx_queue_info_get(port_id, qid, &qinfo)) {
>   		printf("rx_queue_info_get returns error\n");

[PATCH v2] mlx5/testpmd: fix crash on quit with avail thresh enabled

[Spike Du]

When testpmd quit with mlx5 avail_thresh enabled, a rte timer handler
delays to reconfigure rx queue to re-arm this event. However at the same
time, testpmd is destroying rx queues.
It's never a valid use case for mlx5 avail_thresh. Before testpmd quit,
user should disable avail_thresh configuration to not handle the events.
This is documented in mlx5 driver guide.

To avoid the crash in such use case, check port status, if it is not
RTE_PORT_STARTED, don't process the avail_thresh event.

Fixes: f41a5092e6ae ("app/testpmd: add host shaper command")
Cc: stable@dpdk.org

Signed-off-by: Spike Du <spiked@nvidia.com>
---
 drivers/net/mlx5/mlx5_testpmd.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/net/mlx5/mlx5_testpmd.c b/drivers/net/mlx5/mlx5_testpmd.c
index ed84583..879ea28 100644
--- a/drivers/net/mlx5/mlx5_testpmd.c
+++ b/drivers/net/mlx5/mlx5_testpmd.c
@@ -39,7 +39,15 @@
 	uint16_t port_id = port_rxq_id & 0xffff;
 	uint16_t qid = (port_rxq_id >> 16) & 0xffff;
 	struct rte_eth_rxq_info qinfo;
+	struct rte_port *port;
 
+	port = &ports[port_id];
+	if (port->port_status != RTE_PORT_STARTED) {
+		printf("%s port_status(%d) is incorrect, stop avail_thresh "
+		       "event processing.\n",
+		       __func__, port->port_status);
+		return;
+	}
 	printf("%s disable shaper\n", __func__);
 	if (rte_eth_rx_queue_info_get(port_id, qid, &qinfo)) {
 		printf("rx_queue_info_get returns error\n");
-- 
1.8.3.1

RE: [PATCH v2] mlx5/testpmd: fix crash on quit with avail thresh enabled

[Matan Azrad]

From: Spike Du <spiked@nvidia.com>
> When testpmd quit with mlx5 avail_thresh enabled, a rte timer handler
> delays to reconfigure rx queue to re-arm this event. However at the same
> time, testpmd is destroying rx queues.
> It's never a valid use case for mlx5 avail_thresh. Before testpmd quit, user
> should disable avail_thresh configuration to not handle the events.
> This is documented in mlx5 driver guide.
> 
> To avoid the crash in such use case, check port status, if it is not
> RTE_PORT_STARTED, don't process the avail_thresh event.
> 
> Fixes: f41a5092e6ae ("app/testpmd: add host shaper command")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Spike Du <spiked@nvidia.com>
Acked-by: Matan Azrad <matan@nvidia.com>

RE: [PATCH v2] mlx5/testpmd: fix crash on quit with avail thresh enabled

[Raslan Darawsheh]

Hi,

> -----Original Message-----
> From: Spike Du <spiked@nvidia.com>
> Sent: Wednesday, November 2, 2022 1:44 PM
> To: Matan Azrad <matan@nvidia.com>; Slava Ovsiienko
> <viacheslavo@nvidia.com>; Ori Kam <orika@nvidia.com>; NBU-Contact-
> Thomas Monjalon (EXTERNAL) <thomas@monjalon.net>; Shahaf Shuler
> <shahafs@nvidia.com>
> Cc: dev@dpdk.org; Raslan Darawsheh <rasland@nvidia.com>;
> stable@dpdk.org
> Subject: [PATCH v2] mlx5/testpmd: fix crash on quit with avail thresh enabled
> 
> When testpmd quit with mlx5 avail_thresh enabled, a rte timer handler
> delays to reconfigure rx queue to re-arm this event. However at the same
> time, testpmd is destroying rx queues.
> It's never a valid use case for mlx5 avail_thresh. Before testpmd quit,
> user should disable avail_thresh configuration to not handle the events.
> This is documented in mlx5 driver guide.
> 
> To avoid the crash in such use case, check port status, if it is not
> RTE_PORT_STARTED, don't process the avail_thresh event.
> 
> Fixes: f41a5092e6ae ("app/testpmd: add host shaper command")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Spike Du <spiked@nvidia.com>

Patch applied to next-net-mlx,

Kindest regards,
Raslan Darawsheh