DPDK patches and discussions
 help / color / mirror / Atom feed
* [dpdk-dev] [PATCH] vdpa/mlx5: fix polling threads scheduling
@ 2021-02-08  9:28 Matan Azrad
  2021-02-08 11:17 ` Maxime Coquelin
  0 siblings, 1 reply; 4+ messages in thread
From: Matan Azrad @ 2021-02-08  9:28 UTC (permalink / raw)
  To: dev; +Cc: Maxime Coquelin

When the event mode is with 0 fixed delay, the polling-thread will never
give-up CPU.

So, when multi-polling-threads are active, the context-switch between
them will be managed by the system which may affect latency according to
the time-out decided by the system.

In order to fix multi-devices polling thread scheduling, this patch
forces rescheduling for each CQ poll iteration.

Move the polling thread to SCHED_RR mode with maximum priority to
complete the fairness.

Fixes: 6956a48cabbb ("vdpa/mlx5: set polling mode default delay to zero")

Signed-off-by: Matan Azrad <matan@nvidia.com>
---
 drivers/vdpa/mlx5/mlx5_vdpa_event.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

diff --git a/drivers/vdpa/mlx5/mlx5_vdpa_event.c b/drivers/vdpa/mlx5/mlx5_vdpa_event.c
index 0f635ff..86adc86 100644
--- a/drivers/vdpa/mlx5/mlx5_vdpa_event.c
+++ b/drivers/vdpa/mlx5/mlx5_vdpa_event.c
@@ -232,6 +232,9 @@
 	}
 	if (priv->timer_delay_us)
 		usleep(priv->timer_delay_us);
+	else
+		/* Give-up CPU to improve polling threads scheduling. */
+		pthread_yield();
 }
 
 static void *
@@ -500,6 +503,9 @@
 	rte_cpuset_t cpuset;
 	pthread_attr_t attr;
 	char name[16];
+	const struct sched_param sp = {
+		.sched_priority = sched_get_priority_max(SCHED_RR),
+	};
 
 	if (!priv->eventc)
 		/* All virtqs are in poll mode. */
@@ -520,6 +526,16 @@
 			DRV_LOG(ERR, "Failed to set thread affinity.");
 			return -1;
 		}
+		ret = pthread_attr_setschedpolicy(&attr, SCHED_RR);
+		if (ret) {
+			DRV_LOG(ERR, "Failed to set thread sched policy = RR.");
+			return -1;
+		}
+		ret = pthread_attr_setschedparam(&attr, &sp);
+		if (ret) {
+			DRV_LOG(ERR, "Failed to set thread priority.");
+			return -1;
+		}
 		ret = pthread_create(&priv->timer_tid, &attr,
 				     mlx5_vdpa_poll_handle, (void *)priv);
 		if (ret) {
-- 
1.8.3.1


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [dpdk-dev] [PATCH] vdpa/mlx5: fix polling threads scheduling
  2021-02-08  9:28 [dpdk-dev] [PATCH] vdpa/mlx5: fix polling threads scheduling Matan Azrad
@ 2021-02-08 11:17 ` Maxime Coquelin
  2021-02-09  3:15   ` Xueming(Steven) Li
  0 siblings, 1 reply; 4+ messages in thread
From: Maxime Coquelin @ 2021-02-08 11:17 UTC (permalink / raw)
  To: Matan Azrad, dev



On 2/8/21 10:28 AM, Matan Azrad wrote:
> When the event mode is with 0 fixed delay, the polling-thread will never
> give-up CPU.
> 
> So, when multi-polling-threads are active, the context-switch between
> them will be managed by the system which may affect latency according to
> the time-out decided by the system.
> 
> In order to fix multi-devices polling thread scheduling, this patch
> forces rescheduling for each CQ poll iteration.
> 
> Move the polling thread to SCHED_RR mode with maximum priority to
> complete the fairness.
> 
> Fixes: 6956a48cabbb ("vdpa/mlx5: set polling mode default delay to zero")
> 
> Signed-off-by: Matan Azrad <matan@nvidia.com>
> ---
>  drivers/vdpa/mlx5/mlx5_vdpa_event.c | 16 ++++++++++++++++
>  1 file changed, 16 insertions(+)
> 
> diff --git a/drivers/vdpa/mlx5/mlx5_vdpa_event.c b/drivers/vdpa/mlx5/mlx5_vdpa_event.c
> index 0f635ff..86adc86 100644
> --- a/drivers/vdpa/mlx5/mlx5_vdpa_event.c
> +++ b/drivers/vdpa/mlx5/mlx5_vdpa_event.c
> @@ -232,6 +232,9 @@
>  	}
>  	if (priv->timer_delay_us)
>  		usleep(priv->timer_delay_us);
> +	else
> +		/* Give-up CPU to improve polling threads scheduling. */
> +		pthread_yield();
>  }
>  
>  static void *
> @@ -500,6 +503,9 @@
>  	rte_cpuset_t cpuset;
>  	pthread_attr_t attr;
>  	char name[16];
> +	const struct sched_param sp = {
> +		.sched_priority = sched_get_priority_max(SCHED_RR),
> +	};
>  
>  	if (!priv->eventc)
>  		/* All virtqs are in poll mode. */
> @@ -520,6 +526,16 @@
>  			DRV_LOG(ERR, "Failed to set thread affinity.");
>  			return -1;
>  		}
> +		ret = pthread_attr_setschedpolicy(&attr, SCHED_RR);
> +		if (ret) {
> +			DRV_LOG(ERR, "Failed to set thread sched policy = RR.");
> +			return -1;
> +		}
> +		ret = pthread_attr_setschedparam(&attr, &sp);
> +		if (ret) {
> +			DRV_LOG(ERR, "Failed to set thread priority.");
> +			return -1;
> +		}
>  		ret = pthread_create(&priv->timer_tid, &attr,
>  				     mlx5_vdpa_poll_handle, (void *)priv);
>  		if (ret) {
> 

Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [dpdk-dev] [PATCH] vdpa/mlx5: fix polling threads scheduling
  2021-02-08 11:17 ` Maxime Coquelin
@ 2021-02-09  3:15   ` Xueming(Steven) Li
  2021-02-10 21:17     ` Thomas Monjalon
  0 siblings, 1 reply; 4+ messages in thread
From: Xueming(Steven) Li @ 2021-02-09  3:15 UTC (permalink / raw)
  To: Maxime Coquelin, Matan Azrad, dev

>-----Original Message-----
>From: dev <dev-bounces@dpdk.org> On Behalf Of Maxime Coquelin
>Sent: Monday, February 8, 2021 7:17 PM
>To: Matan Azrad <matan@nvidia.com>; dev@dpdk.org
>Subject: Re: [dpdk-dev] [PATCH] vdpa/mlx5: fix polling threads scheduling
>
>
>
>On 2/8/21 10:28 AM, Matan Azrad wrote:
>> When the event mode is with 0 fixed delay, the polling-thread will
>> never give-up CPU.
>>
>> So, when multi-polling-threads are active, the context-switch between
>> them will be managed by the system which may affect latency according
>> to the time-out decided by the system.
>>
>> In order to fix multi-devices polling thread scheduling, this patch
>> forces rescheduling for each CQ poll iteration.
>>
>> Move the polling thread to SCHED_RR mode with maximum priority to
>> complete the fairness.
>>
>> Fixes: 6956a48cabbb ("vdpa/mlx5: set polling mode default delay to
>> zero")
>>
>> Signed-off-by: Matan Azrad <matan@nvidia.com>
>> ---
>>  drivers/vdpa/mlx5/mlx5_vdpa_event.c | 16 ++++++++++++++++
>>  1 file changed, 16 insertions(+)
>>
>> diff --git a/drivers/vdpa/mlx5/mlx5_vdpa_event.c
>> b/drivers/vdpa/mlx5/mlx5_vdpa_event.c
>> index 0f635ff..86adc86 100644
>> --- a/drivers/vdpa/mlx5/mlx5_vdpa_event.c
>> +++ b/drivers/vdpa/mlx5/mlx5_vdpa_event.c
>> @@ -232,6 +232,9 @@
>>  	}
>>  	if (priv->timer_delay_us)
>>  		usleep(priv->timer_delay_us);
>> +	else
>> +		/* Give-up CPU to improve polling threads scheduling. */
>> +		pthread_yield();
>>  }
>>
>>  static void *
>> @@ -500,6 +503,9 @@
>>  	rte_cpuset_t cpuset;
>>  	pthread_attr_t attr;
>>  	char name[16];
>> +	const struct sched_param sp = {
>> +		.sched_priority = sched_get_priority_max(SCHED_RR),
>> +	};
>>
>>  	if (!priv->eventc)
>>  		/* All virtqs are in poll mode. */
>> @@ -520,6 +526,16 @@
>>  			DRV_LOG(ERR, "Failed to set thread affinity.");
>>  			return -1;
>>  		}
>> +		ret = pthread_attr_setschedpolicy(&attr, SCHED_RR);
>> +		if (ret) {
>> +			DRV_LOG(ERR, "Failed to set thread sched policy = RR.");
>> +			return -1;
>> +		}
>> +		ret = pthread_attr_setschedparam(&attr, &sp);
>> +		if (ret) {
>> +			DRV_LOG(ERR, "Failed to set thread priority.");
>> +			return -1;
>> +		}
>>  		ret = pthread_create(&priv->timer_tid, &attr,
>>  				     mlx5_vdpa_poll_handle, (void *)priv);
>>  		if (ret) {
>>
>
>Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>

Acked-by: Xueming Li <xuemingl@mellanox.com>

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [dpdk-dev] [PATCH] vdpa/mlx5: fix polling threads scheduling
  2021-02-09  3:15   ` Xueming(Steven) Li
@ 2021-02-10 21:17     ` Thomas Monjalon
  0 siblings, 0 replies; 4+ messages in thread
From: Thomas Monjalon @ 2021-02-10 21:17 UTC (permalink / raw)
  To: Matan Azrad; +Cc: Maxime Coquelin, dev, Xueming(Steven) Li

> >> When the event mode is with 0 fixed delay, the polling-thread will
> >> never give-up CPU.
> >>
> >> So, when multi-polling-threads are active, the context-switch between
> >> them will be managed by the system which may affect latency according
> >> to the time-out decided by the system.
> >>
> >> In order to fix multi-devices polling thread scheduling, this patch
> >> forces rescheduling for each CQ poll iteration.
> >>
> >> Move the polling thread to SCHED_RR mode with maximum priority to
> >> complete the fairness.
> >>
> >> Fixes: 6956a48cabbb ("vdpa/mlx5: set polling mode default delay to
> >> zero")
> >>
> >> Signed-off-by: Matan Azrad <matan@nvidia.com>
> >
> > Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
> 
> Acked-by: Xueming Li <xuemingl@mellanox.com>
converted to nvidia.com

Applied, thanks



^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2021-02-10 21:17 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-02-08  9:28 [dpdk-dev] [PATCH] vdpa/mlx5: fix polling threads scheduling Matan Azrad
2021-02-08 11:17 ` Maxime Coquelin
2021-02-09  3:15   ` Xueming(Steven) Li
2021-02-10 21:17     ` Thomas Monjalon

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).