[dpdk-dev] [PATCH v3 0/3] improve failsafe performance

[dpdk-dev] [PATCH v3 0/3] improve failsafe performance

[Matan Azrad]

This fail-safe series should improve some cycles in failsafe data path.
It is obvious that we cannot improve a lot because the most of data path functionality are done by the sub devices data path.
So the expectation from this series is just to improve some cycles in sub devices management done by fail-safe PMD.
I saw ~1.7% MPPS improvement for 1 core 1 queue io-forwarding mode - 2 sub-devices(mlx4 primary).

Matan Azrad (3):
  net/failsafe: fix Rx safe check compiler hint
  net/failsafe: mitigate data plan atomic operations
  net/failsafe: improve Rx sub-devices iteration

 drivers/net/failsafe/failsafe.c         |  5 ++++
 drivers/net/failsafe/failsafe_ops.c     |  1 +
 drivers/net/failsafe/failsafe_private.h |  9 +++---
 drivers/net/failsafe/failsafe_rxtx.c    | 50 ++++++++++-----------------------
 4 files changed, 26 insertions(+), 39 deletions(-)

-- 
1.8.3.1

[dpdk-dev] [PATCH v3 1/3] net/failsafe: fix Rx safe check compiler hint

[Matan Azrad]

failsafe_rx_burst function is used when there are no sub-devices or at
least one of them has been removed, on the other hand, when all the
sub-devices are present, failsafe_rx_burst_fast function is used.

So it's really expected that some of the sub-devices will be unsafe for
Rx burst in failsafe_rx_burst execution.

Remove unlikely compiler hint from fs_rx_unsafe calling.

Fixes: a46f8d584eb8 ("net/failsafe: add fail-safe PMD")
Cc: stable@dpdk.org

Signed-off-by: Matan Azrad <matan@mellanox.com>
---
 drivers/net/failsafe/failsafe_rxtx.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/failsafe/failsafe_rxtx.c b/drivers/net/failsafe/failsafe_rxtx.c
index 70157c8..178294c 100644
--- a/drivers/net/failsafe/failsafe_rxtx.c
+++ b/drivers/net/failsafe/failsafe_rxtx.c
@@ -111,7 +111,7 @@
 		if (i == priv->subs_tail)
 			i = priv->subs_head;
 		sdev = &priv->subs[i];
-		if (unlikely(fs_rx_unsafe(sdev)))
+		if (fs_rx_unsafe(sdev))
 			continue;
 		sub_rxq = ETH(sdev)->data->rx_queues[rxq->qid];
 		FS_ATOMIC_P(rxq->refcnt[sdev->sid]);
-- 
1.8.3.1

[dpdk-dev] [PATCH v3 2/3] net/failsafe: mitigate data plan atomic operations

[Matan Azrad]

Fail-safe uses atomic operations to protect sub-device close operation
calling by host thread in removal time while the removed sub-device
burst functions are still in process by application threads.

Using "set" atomic operations is a liitle bit more efficient than "add"
or "sub" atomic operations because "set" shouldn't read the value and
in fact, it does not need a special atomic mechanism in x86 platforms.

Replace "add 1" and "sub 1" atomic operations by "set 1" and "set 0"
atomic operations.

Signed-off-by: Matan Azrad <matan@mellanox.com>
---
 drivers/net/failsafe/failsafe_private.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/failsafe/failsafe_private.h b/drivers/net/failsafe/failsafe_private.h
index d81cc3c..4196ece 100644
--- a/drivers/net/failsafe/failsafe_private.h
+++ b/drivers/net/failsafe/failsafe_private.h
@@ -269,13 +269,13 @@ int failsafe_eth_lsc_event_callback(uint16_t port_id,
  * a: (rte_atomic64_t)
  */
 #define FS_ATOMIC_P(a) \
-	rte_atomic64_add(&(a), 1)
+	rte_atomic64_set(&(a), 1)
 
 /**
  * a: (rte_atomic64_t)
  */
 #define FS_ATOMIC_V(a) \
-	rte_atomic64_sub(&(a), 1)
+	rte_atomic64_set(&(a), 0)
 
 /**
  * s: (struct sub_device *)
-- 
1.8.3.1

[dpdk-dev] [PATCH v3 3/3] net/failsafe: improve Rx sub-devices iteration

[Matan Azrad]

Connecting the sub-devices each other by cyclic linked list can help to
iterate over them by Rx burst functions because there is no need to
check the sub-devices ring wraparound.

Create the aforementioned linked-list and change the Rx burst functions
iteration accordingly.

Signed-off-by: Matan Azrad <matan@mellanox.com>
---
 drivers/net/failsafe/failsafe.c         |  5 ++++
 drivers/net/failsafe/failsafe_ops.c     |  1 +
 drivers/net/failsafe/failsafe_private.h |  5 ++--
 drivers/net/failsafe/failsafe_rxtx.c    | 50 ++++++++++-----------------------
 4 files changed, 24 insertions(+), 37 deletions(-)

diff --git a/drivers/net/failsafe/failsafe.c b/drivers/net/failsafe/failsafe.c
index 6bc5aba..d230c37 100644
--- a/drivers/net/failsafe/failsafe.c
+++ b/drivers/net/failsafe/failsafe.c
@@ -55,6 +55,7 @@
 {
 	uint8_t nb_subs;
 	int ret;
+	int i;
 
 	ret = failsafe_args_count_subdevice(dev, params);
 	if (ret)
@@ -72,6 +73,10 @@
 		ERROR("Could not allocate sub_devices");
 		return -ENOMEM;
 	}
+	/* Initiate static sub devices linked list. */
+	for (i = 1; i < nb_subs; i++)
+		PRIV(dev)->subs[i - 1].next = PRIV(dev)->subs + i;
+	PRIV(dev)->subs[i - 1].next = PRIV(dev)->subs;
 	return 0;
 }
 
diff --git a/drivers/net/failsafe/failsafe_ops.c b/drivers/net/failsafe/failsafe_ops.c
index e16a590..fe957ad 100644
--- a/drivers/net/failsafe/failsafe_ops.c
+++ b/drivers/net/failsafe/failsafe_ops.c
@@ -294,6 +294,7 @@
 	rxq->info.conf = *rx_conf;
 	rxq->info.nb_desc = nb_rx_desc;
 	rxq->priv = PRIV(dev);
+	rxq->sdev = PRIV(dev)->subs;
 	dev->data->rx_queues[rx_queue_id] = rxq;
 	FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE) {
 		ret = rte_eth_rx_queue_setup(PORT_ID(sdev),
diff --git a/drivers/net/failsafe/failsafe_private.h b/drivers/net/failsafe/failsafe_private.h
index 4196ece..54b5b91 100644
--- a/drivers/net/failsafe/failsafe_private.h
+++ b/drivers/net/failsafe/failsafe_private.h
@@ -62,8 +62,8 @@
 struct rxq {
 	struct fs_priv *priv;
 	uint16_t qid;
-	/* id of last sub_device polled */
-	uint8_t last_polled;
+	/* next sub_device to poll */
+	struct sub_device *sdev;
 	unsigned int socket_id;
 	struct rte_eth_rxq_info info;
 	rte_atomic64_t refcnt[];
@@ -100,6 +100,7 @@ struct fs_stats {
 
 struct sub_device {
 	/* Exhaustive DPDK device description */
+	struct sub_device *next;
 	struct rte_devargs devargs;
 	struct rte_bus *bus;
 	struct rte_device *dev;
diff --git a/drivers/net/failsafe/failsafe_rxtx.c b/drivers/net/failsafe/failsafe_rxtx.c
index 178294c..0ebf06a 100644
--- a/drivers/net/failsafe/failsafe_rxtx.c
+++ b/drivers/net/failsafe/failsafe_rxtx.c
@@ -94,36 +94,27 @@
 		  struct rte_mbuf **rx_pkts,
 		  uint16_t nb_pkts)
 {
-	struct fs_priv *priv;
 	struct sub_device *sdev;
 	struct rxq *rxq;
 	void *sub_rxq;
 	uint16_t nb_rx;
-	uint8_t nb_polled, nb_subs;
-	uint8_t i;
 
 	rxq = queue;
-	priv = rxq->priv;
-	nb_subs = priv->subs_tail - priv->subs_head;
-	nb_polled = 0;
-	for (i = rxq->last_polled; nb_polled < nb_subs; nb_polled++) {
-		i++;
-		if (i == priv->subs_tail)
-			i = priv->subs_head;
-		sdev = &priv->subs[i];
-		if (fs_rx_unsafe(sdev))
+	sdev = rxq->sdev;
+	do {
+		if (fs_rx_unsafe(sdev)) {
+			nb_rx = 0;
 			continue;
+		}
 		sub_rxq = ETH(sdev)->data->rx_queues[rxq->qid];
 		FS_ATOMIC_P(rxq->refcnt[sdev->sid]);
 		nb_rx = ETH(sdev)->
 			rx_pkt_burst(sub_rxq, rx_pkts, nb_pkts);
 		FS_ATOMIC_V(rxq->refcnt[sdev->sid]);
-		if (nb_rx) {
-			rxq->last_polled = i;
-			return nb_rx;
-		}
-	}
-	return 0;
+		sdev = sdev->next;
+	} while (nb_rx == 0 && sdev != rxq->sdev);
+	rxq->sdev = sdev;
+	return nb_rx;
 }
 
 uint16_t
@@ -131,35 +122,24 @@
 			 struct rte_mbuf **rx_pkts,
 			 uint16_t nb_pkts)
 {
-	struct fs_priv *priv;
 	struct sub_device *sdev;
 	struct rxq *rxq;
 	void *sub_rxq;
 	uint16_t nb_rx;
-	uint8_t nb_polled, nb_subs;
-	uint8_t i;
 
 	rxq = queue;
-	priv = rxq->priv;
-	nb_subs = priv->subs_tail - priv->subs_head;
-	nb_polled = 0;
-	for (i = rxq->last_polled; nb_polled < nb_subs; nb_polled++) {
-		i++;
-		if (i == priv->subs_tail)
-			i = priv->subs_head;
-		sdev = &priv->subs[i];
+	sdev = rxq->sdev;
+	do {
 		RTE_ASSERT(!fs_rx_unsafe(sdev));
 		sub_rxq = ETH(sdev)->data->rx_queues[rxq->qid];
 		FS_ATOMIC_P(rxq->refcnt[sdev->sid]);
 		nb_rx = ETH(sdev)->
 			rx_pkt_burst(sub_rxq, rx_pkts, nb_pkts);
 		FS_ATOMIC_V(rxq->refcnt[sdev->sid]);
-		if (nb_rx) {
-			rxq->last_polled = i;
-			return nb_rx;
-		}
-	}
-	return 0;
+		sdev = sdev->next;
+	} while (nb_rx == 0 && sdev != rxq->sdev);
+	rxq->sdev = sdev;
+	return nb_rx;
 }
 
 uint16_t
-- 
1.8.3.1

Re: [dpdk-dev] [PATCH v3 3/3] net/failsafe: improve Rx sub-devices iteration

[Gaëtan Rivet]

Hi Matan,

The other commits make sense to me so no issue there.
I'm just surprised by this one so a quick question.

On Tue, Dec 19, 2017 at 05:14:29PM +0000, Matan Azrad wrote:
> Connecting the sub-devices each other by cyclic linked list can help to
> iterate over them by Rx burst functions because there is no need to
> check the sub-devices ring wraparound.
> 
> Create the aforementioned linked-list and change the Rx burst functions
> iteration accordingly.

I'm surprised that a linked-list iteration, with the usual
dereferencing, is better than doing some integer arithmetic.
Maybe the locality of the referenced data helps.

Anyway, were you able to count the cycles gained by this change? It
might be interesting to do a measure with a CPU-bound bench, such as with
a dummy device under the fail-safe (ring or such). MLX devices use a lot
of PCI bandwidth, so the bottleneck could be masked in a physical
setting.

No comments otherwise, if you are sure that this is a performance gain,
the implementation seems ok to me.

Regards,
-- 
Gaëtan Rivet
6WIND

Re: [dpdk-dev] [PATCH v3 3/3] net/failsafe: improve Rx sub-devices iteration

[Matan Azrad]

Hi Gaetan

From: Gaëtan Rivet, Friday, January 12, 2018 12:29 PM
> Hi Matan,
> 
> The other commits make sense to me so no issue there.
> I'm just surprised by this one so a quick question.
> 
> On Tue, Dec 19, 2017 at 05:14:29PM +0000, Matan Azrad wrote:
> > Connecting the sub-devices each other by cyclic linked list can help
> > to iterate over them by Rx burst functions because there is no need to
> > check the sub-devices ring wraparound.
> >
> > Create the aforementioned linked-list and change the Rx burst
> > functions iteration accordingly.
> 
> I'm surprised that a linked-list iteration, with the usual dereferencing, is
> better than doing some integer arithmetic.

This memory references are the same as the previous code because in the new code the linked list elements are still in continuous memory, so probably the addresses stay in the cache.
The removed calculations and wraparound branch probably caused to the performance gain.

> Maybe the locality of the referenced data helps.
> 
Sure.

> Anyway, were you able to count the cycles gained by this change? It might be
> interesting to do a measure with a CPU-bound bench, such as with a dummy
> device under the fail-safe (ring or such). MLX devices use a lot of PCI
> bandwidth, so the bottleneck could be masked in a physical setting.
> 
> No comments otherwise, if you are sure that this is a performance gain, the
> implementation seems ok to me.

Yes, I checked it and saw the little gain obviously.
(just run the test with and without this patch and saw the statistics).

Thanks, Have a good weekend!
> 
> Regards,
> --
> Gaëtan Rivet
> 6WIND

Re: [dpdk-dev] [PATCH v3 3/3] net/failsafe: improve Rx sub-devices iteration

[Gaëtan Rivet]

On Fri, Jan 12, 2018 at 01:29:17PM +0000, Matan Azrad wrote:
> Hi Gaetan
> 
> From: Gaëtan Rivet, Friday, January 12, 2018 12:29 PM
> > Hi Matan,
> > 
> > The other commits make sense to me so no issue there.
> > I'm just surprised by this one so a quick question.
> > 
> > On Tue, Dec 19, 2017 at 05:14:29PM +0000, Matan Azrad wrote:
> > > Connecting the sub-devices each other by cyclic linked list can help
> > > to iterate over them by Rx burst functions because there is no need to
> > > check the sub-devices ring wraparound.
> > >
> > > Create the aforementioned linked-list and change the Rx burst
> > > functions iteration accordingly.
> > 
> > I'm surprised that a linked-list iteration, with the usual dereferencing, is
> > better than doing some integer arithmetic.
> 
> This memory references are the same as the previous code because in the new code the linked list elements are still in continuous memory, so probably the addresses stay in the cache.
> The removed calculations and wraparound branch probably caused to the performance gain.
> 
> > Maybe the locality of the referenced data helps.
> > 
> Sure.

This means that the sub_device definition is critical for the datapath.
It probably goes beyond a cache-line and could be optimized.

> 
> > Anyway, were you able to count the cycles gained by this change? It might be
> > interesting to do a measure with a CPU-bound bench, such as with a dummy
> > device under the fail-safe (ring or such). MLX devices use a lot of PCI
> > bandwidth, so the bottleneck could be masked in a physical setting.
> > 
> > No comments otherwise, if you are sure that this is a performance gain, the
> > implementation seems ok to me.
> 
> Yes, I checked it and saw the little gain obviously.
> (just run the test with and without this patch and saw the statistics).

Oh I'm sure you checked, I just wanted to make sure you properly
considered the methodology.

Acked-by: Gaetan Rivet <gaetan.rivet@6wind.com>

-- 
Gaëtan Rivet
6WIND

Re: [dpdk-dev] [PATCH v3 0/3] improve failsafe performance

[Matan Azrad]

All this series patches are V1.
Sorry for V3 mistake.

> -----Original Message-----
> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Matan Azrad
> Sent: Tuesday, December 19, 2017 7:14 PM
> To: Gaetan Rivet <gaetan.rivet@6wind.com>
> Cc: dev@dpdk.org
> Subject: [dpdk-dev] [PATCH v3 0/3] improve failsafe performance
> 
> This fail-safe series should improve some cycles in failsafe data path.
> It is obvious that we cannot improve a lot because the most of data path
> functionality are done by the sub devices data path.
> So the expectation from this series is just to improve some cycles in sub
> devices management done by fail-safe PMD.
> I saw ~1.7% MPPS improvement for 1 core 1 queue io-forwarding mode - 2
> sub-devices(mlx4 primary).
> 
> Matan Azrad (3):
>   net/failsafe: fix Rx safe check compiler hint
>   net/failsafe: mitigate data plan atomic operations
>   net/failsafe: improve Rx sub-devices iteration
> 
>  drivers/net/failsafe/failsafe.c         |  5 ++++
>  drivers/net/failsafe/failsafe_ops.c     |  1 +
>  drivers/net/failsafe/failsafe_private.h |  9 +++---
>  drivers/net/failsafe/failsafe_rxtx.c    | 50 ++++++++++-----------------------
>  4 files changed, 26 insertions(+), 39 deletions(-)
> 
> --
> 1.8.3.1

Re: [dpdk-dev] [PATCH v3 0/3] improve failsafe performance

[Ferruh Yigit]

On 12/19/2017 5:14 PM, Matan Azrad wrote:
> This fail-safe series should improve some cycles in failsafe data path.
> It is obvious that we cannot improve a lot because the most of data path functionality are done by the sub devices data path.
> So the expectation from this series is just to improve some cycles in sub devices management done by fail-safe PMD.
> I saw ~1.7% MPPS improvement for 1 core 1 queue io-forwarding mode - 2 sub-devices(mlx4 primary).
> 
> Matan Azrad (3):
>   net/failsafe: fix Rx safe check compiler hint
>   net/failsafe: mitigate data plan atomic operations
>   net/failsafe: improve Rx sub-devices iteration

Hi Gaetan,

Any comment on this patchset?

Thanks,
ferruh

Re: [dpdk-dev] [PATCH v3 0/3] improve failsafe performance

[Gaëtan Rivet]

Hi Ferruh,

Only a quick question, otherwise they will probably get in rather soon.

Apologies for the delay.

On Thu, Jan 11, 2018 at 05:58:02PM +0000, Ferruh Yigit wrote:
> On 12/19/2017 5:14 PM, Matan Azrad wrote:
> > This fail-safe series should improve some cycles in failsafe data path.
> > It is obvious that we cannot improve a lot because the most of data path functionality are done by the sub devices data path.
> > So the expectation from this series is just to improve some cycles in sub devices management done by fail-safe PMD.
> > I saw ~1.7% MPPS improvement for 1 core 1 queue io-forwarding mode - 2 sub-devices(mlx4 primary).
> > 
> > Matan Azrad (3):
> >   net/failsafe: fix Rx safe check compiler hint
> >   net/failsafe: mitigate data plan atomic operations
> >   net/failsafe: improve Rx sub-devices iteration
> 
> Hi Gaetan,
> 
> Any comment on this patchset?
> 
> Thanks,
> ferruh

-- 
Gaëtan Rivet
6WIND

Re: [dpdk-dev] [PATCH v3 0/3] improve failsafe performance

[Ferruh Yigit]

On 12/19/2017 5:14 PM, Matan Azrad wrote:
> This fail-safe series should improve some cycles in failsafe data path.
> It is obvious that we cannot improve a lot because the most of data path functionality are done by the sub devices data path.
> So the expectation from this series is just to improve some cycles in sub devices management done by fail-safe PMD.
> I saw ~1.7% MPPS improvement for 1 core 1 queue io-forwarding mode - 2 sub-devices(mlx4 primary).
> 
> Matan Azrad (3):
>   net/failsafe: fix Rx safe check compiler hint
>   net/failsafe: mitigate data plan atomic operations
>   net/failsafe: improve Rx sub-devices iteration

Series applied to dpdk-next-net/master, thanks.