Re: [dpdk-dev] [PATCH v3 01/11] vhost: advertise support in-order feature

Re: [dpdk-dev] [PATCH v3 01/11] vhost: advertise support in-order feature

[Maxime Coquelin]

On 06/28/2018 11:52 PM, Marvin Liu wrote:
> If devices always use descriptors in the same order in which they have
> been made available. These devices can offer the VIRTIO_F_IN_ORDER
> feature. If negotiated, this knowledge allows devices to notify the use
> of a batch of buffers to virtio driver by only writing used ring index.
> 
> Vhost user device has supported this feature by default. If vhost
> dequeue zero is enabled, should disable VIRTIO_F_IN_ORDER as vhost can’t
> assure that descriptors returned from NIC are in order.
> 
> Signed-off-by: Marvin Liu <yong.liu@intel.com>
> 

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

Thanks,
Maxime

Re: [dpdk-dev] [PATCH v3 02/11] net/virtio: add in-order feature bit definition

[Maxime Coquelin]

On 06/28/2018 11:52 PM, Marvin Liu wrote:
> If VIRTIO_F_IN_ORDER has been negotiated, driver will use descriptors in
> ring order: starting from offset 0 in the table, and wrapping around at
> the end of the table. Also introduce use_inorder_[rt]x flag for
> selection of IN_ORDER [RT]x handlers.
> 
> Signed-off-by: Marvin Liu <yong.liu@intel.com>
> 

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

Thanks,
Maxime

Re: [dpdk-dev] [PATCH v3 03/11] net/virtio-user: add unsupported features mask

[Maxime Coquelin]

On 06/28/2018 11:52 PM, Marvin Liu wrote:
> This patch introduces unsupported features mask for virtio-user device.
> For virtio-user server mode, when reconnecting virtio-user will retrieve
> vhost device features as base and then unmask unsupported features.
> 
> Signed-off-by: Marvin Liu <yong.liu@intel.com>
> 

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

Thanks,
Maxime

Re: [dpdk-dev] [PATCH v3 04/11] net/virtio-user: add mrg_rxbuf and in_order vdev parameters

[Maxime Coquelin]

On 06/28/2018 11:52 PM, Marvin Liu wrote:
> Add parameters for configuring VIRTIO_NET_F_MRG_RXBUF and
> VIRTIO_F_IN_ORDER feature bits. If feature is disabled, also update
> corresponding unsupported feature bit.
> 
> Signed-off-by: Marvin Liu <yong.liu@intel.com>
> 

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

Thanks,
Maxime

Re: [dpdk-dev] [PATCH v3 05/11] net/virtio: free in-order descriptors before device start

[Maxime Coquelin]

On 06/28/2018 11:52 PM, Marvin Liu wrote:
> Add new function for freeing IN_ORDER descriptors. As descriptors will
> be allocated and freed sequentially when IN_ORDER feature was
> negotiated. There will be no need to utilize chain for freed descriptors
> management, only index update is enough.
> 
> Signed-off-by: Marvin Liu <yong.liu@intel.com>
> 


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

Thanks,
Maxime

Re: [dpdk-dev] [PATCH v3 06/11] net/virtio: extract common part for in-order functions

[Maxime Coquelin]

On 06/28/2018 11:52 PM, Marvin Liu wrote:
> IN_ORDER virtio-user Tx function support Tx checksum offloading and
> TSO which also support on normal Tx function. So extracts common part
> into separated function for reuse.
> 
> Signed-off-by: Marvin Liu <yong.liu@intel.com>


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

Thanks,
Maxime

Re: [dpdk-dev] [PATCH v3 07/11] net/virtio: support in-order Rx and Tx

[Maxime Coquelin]

On 06/28/2018 11:52 PM, Marvin Liu wrote:
> IN_ORDER Rx function depends on merge-able feature. Descriptors
> allocation and free will be done in bulk.
> 
> Virtio dequeue logic:
>      dequeue_burst_rx(burst mbufs)
>      for (each mbuf b) {
>              if (b need merge) {
>                      merge remained mbufs
>                      add merged mbuf to return mbufs list
>              } else {
>                      add mbuf to return mbufs list
>              }
>      }
>      if (last mbuf c need merge) {
>              dequeue_burst_rx(required mbufs)
>              merge last mbuf c
>      }
>      refill_avail_ring_bulk()
>      update_avail_ring()
>      return mbufs list
> 
> IN_ORDER Tx function can support offloading features. Packets which
> matched "can_push" option will be handled by simple xmit function. Those
> packets can't match "can_push" will be handled by original xmit function
> with in-order flag.
> 
> Virtio enqueue logic:
>      xmit_cleanup(used descs)
>      for (each xmit mbuf b) {
>              if (b can inorder xmit) {
>                      add mbuf b to inorder burst list
>                      continue
>              } else {
>                      xmit inorder burst list
>                      xmit mbuf b by original function
>              }
>      }
>      if (inorder burst list not empty) {
>              xmit inorder burst list
>      }
>      update_avail_ring()
> 
> Signed-off-by: Marvin Liu <yong.liu@intel.com>
> 

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

Thanks,
Maxime

Re: [dpdk-dev] [PATCH v3 08/11] net/virtio: add in-order Rx/Tx into selection

[Maxime Coquelin]

On 06/28/2018 11:52 PM, Marvin Liu wrote:
> After IN_ORDER Rx/Tx paths added, need to update Rx/Tx path selection
> logic.
> 
> Rx path select logic: If IN_ORDER is disabled will select normal Rx
> path. If IN_ORDER is enabled, Rx offload and merge-able are disabled
> will select simple Rx path. Otherwise will select IN_ORDER Rx path.
> 
> Tx path select logic: If IN_ORDER is disabled will select normal Tx
> path. If IN_ORDER is enabled and merge-able is disabled will select
> simple Tx path. Otherwise will select IN_ORDER Tx path.
> 
> Signed-off-by: Marvin Liu <yong.liu@intel.com>
> 
> diff --git a/drivers/net/virtio/virtio_ethdev.c b/drivers/net/virtio/virtio_ethdev.c
> index df50a571a..2b3d65f80 100644
> --- a/drivers/net/virtio/virtio_ethdev.c
> +++ b/drivers/net/virtio/virtio_ethdev.c
> @@ -1320,6 +1320,11 @@ set_rxtx_funcs(struct rte_eth_dev *eth_dev)
>   		PMD_INIT_LOG(INFO, "virtio: using simple Rx path on port %u",
>   			eth_dev->data->port_id);
>   		eth_dev->rx_pkt_burst = virtio_recv_pkts_vec;
> +	} else if (hw->use_inorder_rx) {
> +		PMD_INIT_LOG(INFO,
> +			"virtio: using inorder mergeable buffer Rx path on port %u",
> +			eth_dev->data->port_id);
> +		eth_dev->rx_pkt_burst = &virtio_recv_mergeable_pkts_inorder;
>   	} else if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF)) {
>   		PMD_INIT_LOG(INFO,
>   			"virtio: using mergeable buffer Rx path on port %u",
> @@ -1335,6 +1340,10 @@ set_rxtx_funcs(struct rte_eth_dev *eth_dev)
>   		PMD_INIT_LOG(INFO, "virtio: using simple Tx path on port %u",
>   			eth_dev->data->port_id);
>   		eth_dev->tx_pkt_burst = virtio_xmit_pkts_simple;
> +	} else if (hw->use_inorder_tx) {
> +		PMD_INIT_LOG(INFO, "virtio: using inorder Tx path on port %u",
> +			eth_dev->data->port_id);
> +		eth_dev->tx_pkt_burst = virtio_xmit_pkts_inorder;
>   	} else {
>   		PMD_INIT_LOG(INFO, "virtio: using standard Tx path on port %u",
>   			eth_dev->data->port_id);
> @@ -1871,24 +1880,24 @@ virtio_dev_configure(struct rte_eth_dev *dev)
>   
>   	rte_spinlock_init(&hw->state_lock);
>   
> -	hw->use_simple_rx = 1;
> -	hw->use_simple_tx = 1;
> -
>   #if defined RTE_ARCH_ARM64 || defined RTE_ARCH_ARM
>   	if (!rte_cpu_get_flag_enabled(RTE_CPUFLAG_NEON)) {
>   		hw->use_simple_rx = 0;
>   		hw->use_simple_tx = 0;
>   	}
>   #endif
> -	if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF)) {
> -		hw->use_simple_rx = 0;
> -		hw->use_simple_tx = 0;
> +	if (vtpci_with_feature(hw, VIRTIO_F_IN_ORDER)) {
> +		if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF)) {
> +			hw->use_inorder_rx = 1;
> +			hw->use_inorder_tx = 1;
> +		} else {
> +			hw->use_simple_tx = 1;
> +			if (!(rx_offloads & (DEV_RX_OFFLOAD_UDP_CKSUM |
> +					     DEV_RX_OFFLOAD_TCP_CKSUM)))
> +				hw->use_simple_rx = 1;
> +		}

It seems to be wrong.
For example if IN_ORDER hasn't been negotiated, we might want to use the 
simple path if no rx offload have been requested by the application.

It was the case before the patch if I'm not mistaken.

Also, with ARM platform, we force not to use simple path, but in case
IN_ORDER has been negotiated but not MRG_RXBUF, it gets re-enabled.

>   	}
>   
> -	if (rx_offloads & (DEV_RX_OFFLOAD_UDP_CKSUM |
> -			   DEV_RX_OFFLOAD_TCP_CKSUM))
> -		hw->use_simple_rx = 0;
> -
>   	return 0;
>   }
>   
> 

Re: [dpdk-dev] [PATCH v3 09/11] net/virtio: advertise support in-order feature

[Maxime Coquelin]

On 06/28/2018 11:52 PM, Marvin Liu wrote:
> Signed-off-by: Marvin Liu <yong.liu@intel.com>
> 

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

Thanks,
Maxime

Re: [dpdk-dev] [PATCH v3 10/11] doc: add mrg_rxbuf and in_order parameters for virtio device

[Maxime Coquelin]

On 06/28/2018 11:52 PM, Marvin Liu wrote:
> Virtio vdev parameter mrg_rxbuf and in_order can configure
> VIRTIO_NET_F_MRG_RXBUF and VIRTIO_F_IN_ORDER feature bits. These feature
> bits can effect [RT]x path selection.
> 
> Signed-off-by: Marvin Liu <yong.liu@intel.com>


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

Thanks,
Maxime

Re: [dpdk-dev] [PATCH v3 11/11] doc: add in-order funciton for virtio device

[Maxime Coquelin]

On 06/28/2018 11:52 PM, Marvin Liu wrote:
> Signed-off-by: Marvin Liu <yong.liu@intel.com>
> 

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

Thanks,
Maxime

Re: [dpdk-dev] [PATCH v3 08/11] net/virtio: add in-order Rx/Tx into selection

[Liu, Yong]

> -----Original Message-----
> From: Maxime Coquelin [mailto:maxime.coquelin@redhat.com]
> Sent: Thursday, June 28, 2018 11:16 PM
> To: Liu, Yong <yong.liu@intel.com>; Bie, Tiwei <tiwei.bie@intel.com>
> Cc: Wang, Zhihong <zhihong.wang@intel.com>; dev@dpdk.org
> Subject: Re: [PATCH v3 08/11] net/virtio: add in-order Rx/Tx into
> selection
> 
> 
> 
> On 06/28/2018 11:52 PM, Marvin Liu wrote:
> > After IN_ORDER Rx/Tx paths added, need to update Rx/Tx path selection
> > logic.
> >
> > Rx path select logic: If IN_ORDER is disabled will select normal Rx
> > path. If IN_ORDER is enabled, Rx offload and merge-able are disabled
> > will select simple Rx path. Otherwise will select IN_ORDER Rx path.
> >
> > Tx path select logic: If IN_ORDER is disabled will select normal Tx
> > path. If IN_ORDER is enabled and merge-able is disabled will select
> > simple Tx path. Otherwise will select IN_ORDER Tx path.
> >
> > Signed-off-by: Marvin Liu <yong.liu@intel.com>
> >
> > diff --git a/drivers/net/virtio/virtio_ethdev.c
> b/drivers/net/virtio/virtio_ethdev.c
> > index df50a571a..2b3d65f80 100644
> > --- a/drivers/net/virtio/virtio_ethdev.c
> > +++ b/drivers/net/virtio/virtio_ethdev.c
> > @@ -1320,6 +1320,11 @@ set_rxtx_funcs(struct rte_eth_dev *eth_dev)
> >   		PMD_INIT_LOG(INFO, "virtio: using simple Rx path on port %u",
> >   			eth_dev->data->port_id);
> >   		eth_dev->rx_pkt_burst = virtio_recv_pkts_vec;
> > +	} else if (hw->use_inorder_rx) {
> > +		PMD_INIT_LOG(INFO,
> > +			"virtio: using inorder mergeable buffer Rx path on
> port %u",
> > +			eth_dev->data->port_id);
> > +		eth_dev->rx_pkt_burst = &virtio_recv_mergeable_pkts_inorder;
> >   	} else if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF)) {
> >   		PMD_INIT_LOG(INFO,
> >   			"virtio: using mergeable buffer Rx path on port %u",
> > @@ -1335,6 +1340,10 @@ set_rxtx_funcs(struct rte_eth_dev *eth_dev)
> >   		PMD_INIT_LOG(INFO, "virtio: using simple Tx path on port %u",
> >   			eth_dev->data->port_id);
> >   		eth_dev->tx_pkt_burst = virtio_xmit_pkts_simple;
> > +	} else if (hw->use_inorder_tx) {
> > +		PMD_INIT_LOG(INFO, "virtio: using inorder Tx path on port %u",
> > +			eth_dev->data->port_id);
> > +		eth_dev->tx_pkt_burst = virtio_xmit_pkts_inorder;
> >   	} else {
> >   		PMD_INIT_LOG(INFO, "virtio: using standard Tx path on port %u",
> >   			eth_dev->data->port_id);
> > @@ -1871,24 +1880,24 @@ virtio_dev_configure(struct rte_eth_dev *dev)
> >
> >   	rte_spinlock_init(&hw->state_lock);
> >
> > -	hw->use_simple_rx = 1;
> > -	hw->use_simple_tx = 1;
> > -
> >   #if defined RTE_ARCH_ARM64 || defined RTE_ARCH_ARM
> >   	if (!rte_cpu_get_flag_enabled(RTE_CPUFLAG_NEON)) {
> >   		hw->use_simple_rx = 0;
> >   		hw->use_simple_tx = 0;
> >   	}
> >   #endif
> > -	if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF)) {
> > -		hw->use_simple_rx = 0;
> > -		hw->use_simple_tx = 0;
> > +	if (vtpci_with_feature(hw, VIRTIO_F_IN_ORDER)) {
> > +		if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF)) {
> > +			hw->use_inorder_rx = 1;
> > +			hw->use_inorder_tx = 1;
> > +		} else {
> > +			hw->use_simple_tx = 1;
> > +			if (!(rx_offloads & (DEV_RX_OFFLOAD_UDP_CKSUM |
> > +					     DEV_RX_OFFLOAD_TCP_CKSUM)))
> > +				hw->use_simple_rx = 1;
> > +		}
> 
> It seems to be wrong.
> For example if IN_ORDER hasn't been negotiated, we might want to use the
> simple path if no rx offload have been requested by the application.
> 
> It was the case before the patch if I'm not mistaken.

Maxime,
IN_ORDER is the prerequisite for selection simple rx/tx path. So when IN_ORDER + mergeable off + no rx offload will chose simple rx path.

> 
> Also, with ARM platform, we force not to use simple path, but in case
> IN_ORDER has been negotiated but not MRG_RXBUF, it gets re-enabled.

Will move ARM force action behind normal selection, thus can satisfy ARM.

Thanks,
Marvin

> 
> >   	}
> >
> > -	if (rx_offloads & (DEV_RX_OFFLOAD_UDP_CKSUM |
> > -			   DEV_RX_OFFLOAD_TCP_CKSUM))
> > -		hw->use_simple_rx = 0;
> > -
> >   	return 0;
> >   }
> >
> >

[dpdk-dev] [PATCH v3 00/11] support in-order feature

[Marvin Liu]

In latest virtio-spec, new feature bit VIRTIO_F_IN_ORDER was introduced.
When this feature has been negotiated, virtio driver will use
descriptors in ring order: starting from offset 0 in the table, and
wrapping around at the end of the table. Vhost devices will always use
descriptors in the same order in which they have been made available.
This can reduce virtio accesses to used ring.

Based on updated virtio-spec, this series realized IN_ORDER prototype
in virtio driver. Due to new [RT]x path added into selection, also add
two new parameters mrg_rx and in_order into virtio-user vdev parameters
list. This will allow user to configure feature bits thus can impact
[RT]x path selection.

Performance of virtio user with IN_ORDER feature:

    Platform: Purely
    CPU: Intel(R) Xeon(R) Platinum 8160 CPU @ 2.10GHz
    DPDK baseline: 18.05
    Setup: testpmd with vhost vdev + testpmd with virtio vdev

    +--------------+----------+----------+---------+
    |Vhost->Virtio |1 Queue   |2 Queues  |4 Queues |
    +--------------+----------+----------+---------+
    |Inorder       |12.0Mpps  |24.2Mpps  |26.0Mpps |
    |Normal        |12.1Mpps  |18.5Mpps  |18.9Mpps |
    +--------------+----------+----------+---------+
    
    +--------------+----------+----------------+---------+
    |Virtio->Vhost |1 Queue   |2 Queues        |4 Queues |
    +--------------+----------+----------------+---------+
    |Inorder       |13.8Mpps  |10.7 ~ 15.2Mpps |11.5Mpps |
    |Normal        |13.3Mpps  |9.8 ~ 14Mpps    |10.5Mpps |
    +--------------+----------+----------------+---------+
    
    +---------+----------+----------------+----------------+
    |Loopback |1 Queue   |2 Queues        |4 Queues        |
    +---------+----------+----------------+----------------+
    |Inorder  |7.4Mpps   |9.1 ~ 11.6Mpps  |10.5 ~ 11.3Mpps |
    +---------+----------+----------------+----------------+
    |Normal   |7.5Mpps   |7.7 ~ 9.0Mpps   |7.6 ~ 7.8Mpps   |
    +---------+----------+----------------+----------------+

v3:
- refine [RT]x function selection logic
- fix in-order mergable packets index error
- combine unsupport mask patch
- doc virtio in-order update
- fix checkpatch error

v2:
- merge to latest dpdk-net-virtio 
- not use in_direct for normal xmit packets
- update available ring for each descriptor
- clean up IN_ORDER xmit function
- unmask feature bits when disabled in_order or mgr_rxbuf
- extract common part between IN_ORDER and normal functions
- update performance result

Marvin Liu (11):
  vhost: advertise support in-order feature
  net/virtio: add in-order feature bit definition
  net/virtio-user: add unsupported features mask
  net/virtio-user: add mrg_rxbuf and in_order vdev parameters
  net/virtio: free in-order descriptors before device start
  net/virtio: extract common part for in-order functions
  net/virtio: support in-order Rx and Tx
  net/virtio: add in-order Rx/Tx into selection
  net/virtio: advertise support in-order feature
  doc: add mrg_rxbuf and in_order parameters for virtio device
  doc: add in-order funciton for virtio device

 doc/guides/nics/virtio.rst                    |  27 +-
 drivers/net/virtio/virtio_ethdev.c            |  29 +-
 drivers/net/virtio/virtio_ethdev.h            |   7 +
 drivers/net/virtio/virtio_pci.h               |   8 +
 drivers/net/virtio/virtio_rxtx.c              | 635 ++++++++++++++++--
 .../net/virtio/virtio_user/virtio_user_dev.c  |  26 +-
 .../net/virtio/virtio_user/virtio_user_dev.h  |   4 +-
 drivers/net/virtio/virtio_user_ethdev.c       |  45 +-
 drivers/net/virtio/virtqueue.c                |   8 +
 drivers/net/virtio/virtqueue.h                |   2 +
 lib/librte_vhost/socket.c                     |   6 +
 lib/librte_vhost/vhost.h                      |  10 +-
 12 files changed, 723 insertions(+), 84 deletions(-)

-- 
2.17.0

[dpdk-dev] [PATCH v3 01/11] vhost: advertise support in-order feature

[Marvin Liu]

If devices always use descriptors in the same order in which they have
been made available. These devices can offer the VIRTIO_F_IN_ORDER
feature. If negotiated, this knowledge allows devices to notify the use
of a batch of buffers to virtio driver by only writing used ring index.

Vhost user device has supported this feature by default. If vhost
dequeue zero is enabled, should disable VIRTIO_F_IN_ORDER as vhost can’t
assure that descriptors returned from NIC are in order.

Signed-off-by: Marvin Liu <yong.liu@intel.com>

diff --git a/lib/librte_vhost/socket.c b/lib/librte_vhost/socket.c
index 0399c37bc..d63031747 100644
--- a/lib/librte_vhost/socket.c
+++ b/lib/librte_vhost/socket.c
@@ -853,6 +853,12 @@ rte_vhost_driver_register(const char *path, uint64_t flags)
 	vsocket->supported_features = VIRTIO_NET_SUPPORTED_FEATURES;
 	vsocket->features           = VIRTIO_NET_SUPPORTED_FEATURES;
 
+	/* Dequeue zero copy can't assure descriptors returned in order */
+	if (vsocket->dequeue_zero_copy) {
+		vsocket->supported_features &= ~(1ULL << VIRTIO_F_IN_ORDER);
+		vsocket->features &= ~(1ULL << VIRTIO_F_IN_ORDER);
+	}
+
 	if (!(flags & RTE_VHOST_USER_IOMMU_SUPPORT)) {
 		vsocket->supported_features &= ~(1ULL << VIRTIO_F_IOMMU_PLATFORM);
 		vsocket->features &= ~(1ULL << VIRTIO_F_IOMMU_PLATFORM);
diff --git a/lib/librte_vhost/vhost.h b/lib/librte_vhost/vhost.h
index 786a74f64..3437b996b 100644
--- a/lib/librte_vhost/vhost.h
+++ b/lib/librte_vhost/vhost.h
@@ -191,6 +191,13 @@ struct vhost_msg {
  #define VIRTIO_F_VERSION_1 32
 #endif
 
+/*
+ * Available and used descs are in same order
+ */
+#ifndef VIRTIO_F_IN_ORDER
+#define VIRTIO_F_IN_ORDER      35
+#endif
+
 /* Features supported by this builtin vhost-user net driver. */
 #define VIRTIO_NET_SUPPORTED_FEATURES ((1ULL << VIRTIO_NET_F_MRG_RXBUF) | \
 				(1ULL << VIRTIO_F_ANY_LAYOUT) | \
@@ -214,7 +221,8 @@ struct vhost_msg {
 				(1ULL << VIRTIO_NET_F_GUEST_ECN) | \
 				(1ULL << VIRTIO_RING_F_INDIRECT_DESC) | \
 				(1ULL << VIRTIO_RING_F_EVENT_IDX) | \
-				(1ULL << VIRTIO_NET_F_MTU) | \
+				(1ULL << VIRTIO_NET_F_MTU)  | \
+				(1ULL << VIRTIO_F_IN_ORDER) | \
 				(1ULL << VIRTIO_F_IOMMU_PLATFORM))
 
 
-- 
2.17.0

[dpdk-dev] [PATCH v3 02/11] net/virtio: add in-order feature bit definition

[Marvin Liu]

If VIRTIO_F_IN_ORDER has been negotiated, driver will use descriptors in
ring order: starting from offset 0 in the table, and wrapping around at
the end of the table. Also introduce use_inorder_[rt]x flag for
selection of IN_ORDER [RT]x handlers.

Signed-off-by: Marvin Liu <yong.liu@intel.com>

diff --git a/drivers/net/virtio/virtio_pci.h b/drivers/net/virtio/virtio_pci.h
index a28ba8339..77f805df6 100644
--- a/drivers/net/virtio/virtio_pci.h
+++ b/drivers/net/virtio/virtio_pci.h
@@ -121,6 +121,12 @@ struct virtnet_ctl;
 #define VIRTIO_TRANSPORT_F_START 28
 #define VIRTIO_TRANSPORT_F_END   34
 
+/*
+ * Inorder feature indicates that all buffers are used by the device
+ * in the same order in which they have been made available.
+ */
+#define VIRTIO_F_IN_ORDER 35
+
 /* The Guest publishes the used index for which it expects an interrupt
  * at the end of the avail ring. Host should ignore the avail->flags field. */
 /* The Host publishes the avail index for which it expects a kick
@@ -233,6 +239,8 @@ struct virtio_hw {
 	uint8_t     modern;
 	uint8_t     use_simple_rx;
 	uint8_t     use_simple_tx;
+	uint8_t     use_inorder_rx;
+	uint8_t     use_inorder_tx;
 	uint16_t    port_id;
 	uint8_t     mac_addr[ETHER_ADDR_LEN];
 	uint32_t    notify_off_multiplier;
-- 
2.17.0

[dpdk-dev] [PATCH v3 03/11] net/virtio-user: add unsupported features mask

[Marvin Liu]

This patch introduces unsupported features mask for virtio-user device.
For virtio-user server mode, when reconnecting virtio-user will retrieve
vhost device features as base and then unmask unsupported features.

Signed-off-by: Marvin Liu <yong.liu@intel.com>

diff --git a/drivers/net/virtio/virtio_user/virtio_user_dev.c b/drivers/net/virtio/virtio_user/virtio_user_dev.c
index 4322527f2..0fe3b377d 100644
--- a/drivers/net/virtio/virtio_user/virtio_user_dev.c
+++ b/drivers/net/virtio/virtio_user/virtio_user_dev.c
@@ -384,6 +384,7 @@ virtio_user_dev_init(struct virtio_user_dev *dev, char *path, int queues,
 	dev->queue_pairs = 1; /* mq disabled by default */
 	dev->queue_size = queue_size;
 	dev->mac_specified = 0;
+	dev->unsupported_features = 0;
 	parse_mac(dev, mac);
 
 	if (*ifname) {
@@ -421,8 +422,10 @@ virtio_user_dev_init(struct virtio_user_dev *dev, char *path, int queues,
 
 	if (dev->mac_specified)
 		dev->device_features |= (1ull << VIRTIO_NET_F_MAC);
-	else
+	else {
 		dev->device_features &= ~(1ull << VIRTIO_NET_F_MAC);
+		dev->unsupported_features |= (1ull << VIRTIO_NET_F_MAC);
+	}
 
 	if (cq) {
 		/* device does not really need to know anything about CQ,
@@ -437,6 +440,12 @@ virtio_user_dev_init(struct virtio_user_dev *dev, char *path, int queues,
 		dev->device_features &= ~(1ull << VIRTIO_NET_F_GUEST_ANNOUNCE);
 		dev->device_features &= ~(1ull << VIRTIO_NET_F_MQ);
 		dev->device_features &= ~(1ull << VIRTIO_NET_F_CTRL_MAC_ADDR);
+		dev->unsupported_features |= (1ull << VIRTIO_NET_F_CTRL_VQ);
+		dev->unsupported_features |= (1ull << VIRTIO_NET_F_CTRL_RX);
+		dev->unsupported_features |= (1ull << VIRTIO_NET_F_CTRL_VLAN);
+		dev->unsupported_features |= (1ull << VIRTIO_NET_F_GUEST_ANNOUNCE);
+		dev->unsupported_features |= (1ull << VIRTIO_NET_F_MQ);
+		dev->unsupported_features |= (1ull << VIRTIO_NET_F_CTRL_MAC_ADDR);
 	}
 
 	/* The backend will not report this feature, we add it explicitly */
@@ -444,6 +453,7 @@ virtio_user_dev_init(struct virtio_user_dev *dev, char *path, int queues,
 		dev->device_features |= (1ull << VIRTIO_NET_F_STATUS);
 
 	dev->device_features &= VIRTIO_USER_SUPPORTED_FEATURES;
+	dev->unsupported_features |= ~VIRTIO_USER_SUPPORTED_FEATURES;
 
 	if (rte_mem_event_callback_register(VIRTIO_USER_MEM_EVENT_CLB_NAME,
 				virtio_user_mem_event_cb, dev)) {
diff --git a/drivers/net/virtio/virtio_user/virtio_user_dev.h b/drivers/net/virtio/virtio_user/virtio_user_dev.h
index d2d4cb825..c23ddfcc5 100644
--- a/drivers/net/virtio/virtio_user/virtio_user_dev.h
+++ b/drivers/net/virtio/virtio_user/virtio_user_dev.h
@@ -33,6 +33,7 @@ struct virtio_user_dev {
 				   * and will be sync with device
 				   */
 	uint64_t	device_features; /* supported features by device */
+	uint64_t	unsupported_features; /* unsupported features mask */
 	uint8_t		status;
 	uint16_t	port_id;
 	uint8_t		mac_addr[ETHER_ADDR_LEN];
diff --git a/drivers/net/virtio/virtio_user_ethdev.c b/drivers/net/virtio/virtio_user_ethdev.c
index 1c102ca72..416dbdcf8 100644
--- a/drivers/net/virtio/virtio_user_ethdev.c
+++ b/drivers/net/virtio/virtio_user_ethdev.c
@@ -30,7 +30,6 @@ virtio_user_server_reconnect(struct virtio_user_dev *dev)
 	int ret;
 	int flag;
 	int connectfd;
-	uint64_t features = dev->device_features;
 	struct rte_eth_dev *eth_dev = &rte_eth_devices[dev->port_id];
 
 	connectfd = accept(dev->listenfd, NULL, NULL);
@@ -45,15 +44,8 @@ virtio_user_server_reconnect(struct virtio_user_dev *dev)
 		return -1;
 	}
 
-	features &= ~dev->device_features;
-	/* For following bits, vhost-user doesn't really need to know */
-	features &= ~(1ull << VIRTIO_NET_F_MAC);
-	features &= ~(1ull << VIRTIO_NET_F_CTRL_VLAN);
-	features &= ~(1ull << VIRTIO_NET_F_CTRL_MAC_ADDR);
-	features &= ~(1ull << VIRTIO_NET_F_STATUS);
-	if (features)
-		PMD_INIT_LOG(ERR, "WARNING: Some features 0x%" PRIx64 " are not supported by vhost-user!",
-			     features);
+	/* umask vhost-user unsupported features */
+	dev->device_features &= ~(dev->unsupported_features);
 
 	dev->features &= dev->device_features;
 
-- 
2.17.0

[dpdk-dev] [PATCH v3 04/11] net/virtio-user: add mrg_rxbuf and in_order vdev parameters

[Marvin Liu]

Add parameters for configuring VIRTIO_NET_F_MRG_RXBUF and
VIRTIO_F_IN_ORDER feature bits. If feature is disabled, also update
corresponding unsupported feature bit.

Signed-off-by: Marvin Liu <yong.liu@intel.com>

diff --git a/drivers/net/virtio/virtio_user/virtio_user_dev.c b/drivers/net/virtio/virtio_user/virtio_user_dev.c
index 0fe3b377d..e90257f5d 100644
--- a/drivers/net/virtio/virtio_user/virtio_user_dev.c
+++ b/drivers/net/virtio/virtio_user/virtio_user_dev.c
@@ -375,7 +375,8 @@ virtio_user_dev_setup(struct virtio_user_dev *dev)
 
 int
 virtio_user_dev_init(struct virtio_user_dev *dev, char *path, int queues,
-		     int cq, int queue_size, const char *mac, char **ifname)
+		     int cq, int queue_size, const char *mac, char **ifname,
+		     int mrg_rxbuf, int in_order)
 {
 	pthread_mutex_init(&dev->mutex, NULL);
 	snprintf(dev->path, PATH_MAX, "%s", path);
@@ -420,6 +421,16 @@ virtio_user_dev_init(struct virtio_user_dev *dev, char *path, int queues,
 		dev->device_features = VIRTIO_USER_SUPPORTED_FEATURES;
 	}
 
+	if (!mrg_rxbuf) {
+		dev->device_features &= ~(1ull << VIRTIO_NET_F_MRG_RXBUF);
+		dev->unsupported_features |= (1ull << VIRTIO_NET_F_MRG_RXBUF);
+	}
+
+	if (!in_order) {
+		dev->device_features &= ~(1ull << VIRTIO_F_IN_ORDER);
+		dev->unsupported_features |= (1ull << VIRTIO_F_IN_ORDER);
+	}
+
 	if (dev->mac_specified)
 		dev->device_features |= (1ull << VIRTIO_NET_F_MAC);
 	else {
diff --git a/drivers/net/virtio/virtio_user/virtio_user_dev.h b/drivers/net/virtio/virtio_user/virtio_user_dev.h
index c23ddfcc5..d6e0e137b 100644
--- a/drivers/net/virtio/virtio_user/virtio_user_dev.h
+++ b/drivers/net/virtio/virtio_user/virtio_user_dev.h
@@ -48,7 +48,8 @@ int is_vhost_user_by_type(const char *path);
 int virtio_user_start_device(struct virtio_user_dev *dev);
 int virtio_user_stop_device(struct virtio_user_dev *dev);
 int virtio_user_dev_init(struct virtio_user_dev *dev, char *path, int queues,
-			 int cq, int queue_size, const char *mac, char **ifname);
+			 int cq, int queue_size, const char *mac, char **ifname,
+			 int mrg_rxbuf, int in_order);
 void virtio_user_dev_uninit(struct virtio_user_dev *dev);
 void virtio_user_handle_cq(struct virtio_user_dev *dev, uint16_t queue_idx);
 uint8_t virtio_user_handle_mq(struct virtio_user_dev *dev, uint16_t q_pairs);
diff --git a/drivers/net/virtio/virtio_user_ethdev.c b/drivers/net/virtio/virtio_user_ethdev.c
index 416dbdcf8..509c9b1b3 100644
--- a/drivers/net/virtio/virtio_user_ethdev.c
+++ b/drivers/net/virtio/virtio_user_ethdev.c
@@ -358,8 +358,12 @@ static const char *valid_args[] = {
 	VIRTIO_USER_ARG_QUEUE_SIZE,
 #define VIRTIO_USER_ARG_INTERFACE_NAME "iface"
 	VIRTIO_USER_ARG_INTERFACE_NAME,
-#define VIRTIO_USER_ARG_SERVER_MODE "server"
+#define VIRTIO_USER_ARG_SERVER_MODE    "server"
 	VIRTIO_USER_ARG_SERVER_MODE,
+#define VIRTIO_USER_ARG_MRG_RXBUF      "mrg_rxbuf"
+	VIRTIO_USER_ARG_MRG_RXBUF,
+#define VIRTIO_USER_ARG_IN_ORDER       "in_order"
+	VIRTIO_USER_ARG_IN_ORDER,
 	NULL
 };
 
@@ -462,6 +466,8 @@ virtio_user_pmd_probe(struct rte_vdev_device *dev)
 	uint64_t cq = VIRTIO_USER_DEF_CQ_EN;
 	uint64_t queue_size = VIRTIO_USER_DEF_Q_SZ;
 	uint64_t server_mode = VIRTIO_USER_DEF_SERVER_MODE;
+	uint64_t mrg_rxbuf = 1;
+	uint64_t in_order = 1;
 	char *path = NULL;
 	char *ifname = NULL;
 	char *mac_addr = NULL;
@@ -561,6 +567,24 @@ virtio_user_pmd_probe(struct rte_vdev_device *dev)
 		goto end;
 	}
 
+	if (rte_kvargs_count(kvlist, VIRTIO_USER_ARG_MRG_RXBUF) == 1) {
+		if (rte_kvargs_process(kvlist, VIRTIO_USER_ARG_MRG_RXBUF,
+				       &get_integer_arg, &mrg_rxbuf) < 0) {
+			PMD_INIT_LOG(ERR, "error to parse %s",
+				     VIRTIO_USER_ARG_MRG_RXBUF);
+			goto end;
+		}
+	}
+
+	if (rte_kvargs_count(kvlist, VIRTIO_USER_ARG_IN_ORDER) == 1) {
+		if (rte_kvargs_process(kvlist, VIRTIO_USER_ARG_IN_ORDER,
+				       &get_integer_arg, &in_order) < 0) {
+			PMD_INIT_LOG(ERR, "error to parse %s",
+				     VIRTIO_USER_ARG_IN_ORDER);
+			goto end;
+		}
+	}
+
 	if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
 		struct virtio_user_dev *vu_dev;
 
@@ -577,7 +601,8 @@ virtio_user_pmd_probe(struct rte_vdev_device *dev)
 		else
 			vu_dev->is_server = false;
 		if (virtio_user_dev_init(hw->virtio_user_dev, path, queues, cq,
-				 queue_size, mac_addr, &ifname) < 0) {
+				queue_size, mac_addr, &ifname, mrg_rxbuf,
+				in_order) < 0) {
 			PMD_INIT_LOG(ERR, "virtio_user_dev_init fails");
 			virtio_user_eth_dev_free(eth_dev);
 			goto end;
@@ -655,4 +680,6 @@ RTE_PMD_REGISTER_PARAM_STRING(net_virtio_user,
 	"cq=<int> "
 	"queue_size=<int> "
 	"queues=<int> "
-	"iface=<string>");
+	"iface=<string>"
+	"mrg_rxbuf=<0|1>"
+	"in_order=<0|1>");
-- 
2.17.0

[dpdk-dev] [PATCH v3 05/11] net/virtio: free in-order descriptors before device start

[Marvin Liu]

Add new function for freeing IN_ORDER descriptors. As descriptors will
be allocated and freed sequentially when IN_ORDER feature was
negotiated. There will be no need to utilize chain for freed descriptors
management, only index update is enough.

Signed-off-by: Marvin Liu <yong.liu@intel.com>

diff --git a/drivers/net/virtio/virtio_rxtx.c b/drivers/net/virtio/virtio_rxtx.c
index 92fab2174..0bca29855 100644
--- a/drivers/net/virtio/virtio_rxtx.c
+++ b/drivers/net/virtio/virtio_rxtx.c
@@ -47,6 +47,13 @@ virtio_dev_rx_queue_done(void *rxq, uint16_t offset)
 	return VIRTQUEUE_NUSED(vq) >= offset;
 }
 
+void
+vq_ring_free_inorder(struct virtqueue *vq, uint16_t desc_idx, uint16_t num)
+{
+	vq->vq_free_cnt += num;
+	vq->vq_desc_tail_idx = desc_idx & (vq->vq_nentries - 1);
+}
+
 void
 vq_ring_free_chain(struct virtqueue *vq, uint16_t desc_idx)
 {
diff --git a/drivers/net/virtio/virtqueue.c b/drivers/net/virtio/virtqueue.c
index a7d0a9cbe..56a77cc71 100644
--- a/drivers/net/virtio/virtqueue.c
+++ b/drivers/net/virtio/virtqueue.c
@@ -74,6 +74,14 @@ virtqueue_rxvq_flush(struct virtqueue *vq)
 			desc_idx = used_idx;
 			rte_pktmbuf_free(vq->sw_ring[desc_idx]);
 			vq->vq_free_cnt++;
+		} else if (hw->use_inorder_rx) {
+			desc_idx = (uint16_t)uep->id;
+			dxp = &vq->vq_descx[desc_idx];
+			if (dxp->cookie != NULL) {
+				rte_pktmbuf_free(dxp->cookie);
+				dxp->cookie = NULL;
+			}
+			vq_ring_free_inorder(vq, desc_idx, 1);
 		} else {
 			desc_idx = (uint16_t)uep->id;
 			dxp = &vq->vq_descx[desc_idx];
diff --git a/drivers/net/virtio/virtqueue.h b/drivers/net/virtio/virtqueue.h
index 14364f356..26518ed98 100644
--- a/drivers/net/virtio/virtqueue.h
+++ b/drivers/net/virtio/virtqueue.h
@@ -306,6 +306,8 @@ virtio_get_queue_type(struct virtio_hw *hw, uint16_t vtpci_queue_idx)
 #define VIRTQUEUE_NUSED(vq) ((uint16_t)((vq)->vq_ring.used->idx - (vq)->vq_used_cons_idx))
 
 void vq_ring_free_chain(struct virtqueue *vq, uint16_t desc_idx);
+void vq_ring_free_inorder(struct virtqueue *vq, uint16_t desc_idx,
+			  uint16_t num);
 
 static inline void
 vq_update_avail_idx(struct virtqueue *vq)
-- 
2.17.0

[dpdk-dev] [PATCH v3 06/11] net/virtio: extract common part for in-order functions

[Marvin Liu]

IN_ORDER virtio-user Tx function support Tx checksum offloading and
TSO which also support on normal Tx function. So extracts common part
into separated function for reuse.

Signed-off-by: Marvin Liu <yong.liu@intel.com>

diff --git a/drivers/net/virtio/virtio_rxtx.c b/drivers/net/virtio/virtio_rxtx.c
index 0bca29855..e9b1b496e 100644
--- a/drivers/net/virtio/virtio_rxtx.c
+++ b/drivers/net/virtio/virtio_rxtx.c
@@ -246,6 +246,55 @@ tx_offload_enabled(struct virtio_hw *hw)
 		(var) = (val);			\
 } while (0)
 
+static inline void
+virtqueue_xmit_offload(struct virtio_net_hdr *hdr,
+			struct rte_mbuf *cookie,
+			int offload)
+{
+	if (offload) {
+		if (cookie->ol_flags & PKT_TX_TCP_SEG)
+			cookie->ol_flags |= PKT_TX_TCP_CKSUM;
+
+		switch (cookie->ol_flags & PKT_TX_L4_MASK) {
+		case PKT_TX_UDP_CKSUM:
+			hdr->csum_start = cookie->l2_len + cookie->l3_len;
+			hdr->csum_offset = offsetof(struct udp_hdr,
+				dgram_cksum);
+			hdr->flags = VIRTIO_NET_HDR_F_NEEDS_CSUM;
+			break;
+
+		case PKT_TX_TCP_CKSUM:
+			hdr->csum_start = cookie->l2_len + cookie->l3_len;
+			hdr->csum_offset = offsetof(struct tcp_hdr, cksum);
+			hdr->flags = VIRTIO_NET_HDR_F_NEEDS_CSUM;
+			break;
+
+		default:
+			ASSIGN_UNLESS_EQUAL(hdr->csum_start, 0);
+			ASSIGN_UNLESS_EQUAL(hdr->csum_offset, 0);
+			ASSIGN_UNLESS_EQUAL(hdr->flags, 0);
+			break;
+		}
+
+		/* TCP Segmentation Offload */
+		if (cookie->ol_flags & PKT_TX_TCP_SEG) {
+			virtio_tso_fix_cksum(cookie);
+			hdr->gso_type = (cookie->ol_flags & PKT_TX_IPV6) ?
+				VIRTIO_NET_HDR_GSO_TCPV6 :
+				VIRTIO_NET_HDR_GSO_TCPV4;
+			hdr->gso_size = cookie->tso_segsz;
+			hdr->hdr_len =
+				cookie->l2_len +
+				cookie->l3_len +
+				cookie->l4_len;
+		} else {
+			ASSIGN_UNLESS_EQUAL(hdr->gso_type, 0);
+			ASSIGN_UNLESS_EQUAL(hdr->gso_size, 0);
+			ASSIGN_UNLESS_EQUAL(hdr->hdr_len, 0);
+		}
+	}
+}
+
 static inline void
 virtqueue_enqueue_xmit(struct virtnet_tx *txvq, struct rte_mbuf *cookie,
 		       uint16_t needed, int use_indirect, int can_push)
@@ -315,49 +364,7 @@ virtqueue_enqueue_xmit(struct virtnet_tx *txvq, struct rte_mbuf *cookie,
 		idx = start_dp[idx].next;
 	}
 
-	/* Checksum Offload / TSO */
-	if (offload) {
-		if (cookie->ol_flags & PKT_TX_TCP_SEG)
-			cookie->ol_flags |= PKT_TX_TCP_CKSUM;
-
-		switch (cookie->ol_flags & PKT_TX_L4_MASK) {
-		case PKT_TX_UDP_CKSUM:
-			hdr->csum_start = cookie->l2_len + cookie->l3_len;
-			hdr->csum_offset = offsetof(struct udp_hdr,
-				dgram_cksum);
-			hdr->flags = VIRTIO_NET_HDR_F_NEEDS_CSUM;
-			break;
-
-		case PKT_TX_TCP_CKSUM:
-			hdr->csum_start = cookie->l2_len + cookie->l3_len;
-			hdr->csum_offset = offsetof(struct tcp_hdr, cksum);
-			hdr->flags = VIRTIO_NET_HDR_F_NEEDS_CSUM;
-			break;
-
-		default:
-			ASSIGN_UNLESS_EQUAL(hdr->csum_start, 0);
-			ASSIGN_UNLESS_EQUAL(hdr->csum_offset, 0);
-			ASSIGN_UNLESS_EQUAL(hdr->flags, 0);
-			break;
-		}
-
-		/* TCP Segmentation Offload */
-		if (cookie->ol_flags & PKT_TX_TCP_SEG) {
-			virtio_tso_fix_cksum(cookie);
-			hdr->gso_type = (cookie->ol_flags & PKT_TX_IPV6) ?
-				VIRTIO_NET_HDR_GSO_TCPV6 :
-				VIRTIO_NET_HDR_GSO_TCPV4;
-			hdr->gso_size = cookie->tso_segsz;
-			hdr->hdr_len =
-				cookie->l2_len +
-				cookie->l3_len +
-				cookie->l4_len;
-		} else {
-			ASSIGN_UNLESS_EQUAL(hdr->gso_type, 0);
-			ASSIGN_UNLESS_EQUAL(hdr->gso_size, 0);
-			ASSIGN_UNLESS_EQUAL(hdr->hdr_len, 0);
-		}
-	}
+	virtqueue_xmit_offload(hdr, cookie, offload);
 
 	do {
 		start_dp[idx].addr  = VIRTIO_MBUF_DATA_DMA_ADDR(cookie, vq);
@@ -621,6 +628,15 @@ virtio_update_packet_stats(struct virtnet_stats *stats, struct rte_mbuf *mbuf)
 	}
 }
 
+static inline void
+virtio_rx_stats_updated(struct virtnet_rx *rxvq, struct rte_mbuf *m)
+{
+	VIRTIO_DUMP_PACKET(m, m->data_len);
+
+	rxvq->stats.bytes += m->pkt_len;
+	virtio_update_packet_stats(&rxvq->stats, m);
+}
+
 /* Optionally fill offload information in structure */
 static int
 virtio_rx_offload(struct rte_mbuf *m, struct virtio_net_hdr *hdr)
@@ -773,12 +789,9 @@ virtio_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
 			continue;
 		}
 
-		VIRTIO_DUMP_PACKET(rxm, rxm->data_len);
+		virtio_rx_stats_updated(rxvq, rxm);
 
 		rx_pkts[nb_rx++] = rxm;
-
-		rxvq->stats.bytes += rxm->pkt_len;
-		virtio_update_packet_stats(&rxvq->stats, rxm);
 	}
 
 	rxvq->stats.packets += nb_rx;
-- 
2.17.0

[dpdk-dev] [PATCH v3 07/11] net/virtio: support in-order Rx and Tx

[Marvin Liu]

IN_ORDER Rx function depends on merge-able feature. Descriptors
allocation and free will be done in bulk.

Virtio dequeue logic:
    dequeue_burst_rx(burst mbufs)
    for (each mbuf b) {
            if (b need merge) {
                    merge remained mbufs
                    add merged mbuf to return mbufs list
            } else {
                    add mbuf to return mbufs list
            }
    }
    if (last mbuf c need merge) {
            dequeue_burst_rx(required mbufs)
            merge last mbuf c
    }
    refill_avail_ring_bulk()
    update_avail_ring()
    return mbufs list

IN_ORDER Tx function can support offloading features. Packets which
matched "can_push" option will be handled by simple xmit function. Those
packets can't match "can_push" will be handled by original xmit function
with in-order flag.

Virtio enqueue logic:
    xmit_cleanup(used descs)
    for (each xmit mbuf b) {
            if (b can inorder xmit) {
                    add mbuf b to inorder burst list
                    continue
            } else {
                    xmit inorder burst list
                    xmit mbuf b by original function
            }
    }
    if (inorder burst list not empty) {
            xmit inorder burst list
    }
    update_avail_ring()

Signed-off-by: Marvin Liu <yong.liu@intel.com>

diff --git a/drivers/net/virtio/virtio_ethdev.h b/drivers/net/virtio/virtio_ethdev.h
index bb40064ea..8ab1267b0 100644
--- a/drivers/net/virtio/virtio_ethdev.h
+++ b/drivers/net/virtio/virtio_ethdev.h
@@ -83,9 +83,15 @@ uint16_t virtio_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
 uint16_t virtio_recv_mergeable_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
 		uint16_t nb_pkts);
 
+uint16_t virtio_recv_mergeable_pkts_inorder(void *rx_queue, struct rte_mbuf **rx_pkts,
+		uint16_t nb_pkts);
+
 uint16_t virtio_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
 		uint16_t nb_pkts);
 
+uint16_t virtio_xmit_pkts_inorder(void *tx_queue, struct rte_mbuf **tx_pkts,
+		uint16_t nb_pkts);
+
 uint16_t virtio_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
 		uint16_t nb_pkts);
 
diff --git a/drivers/net/virtio/virtio_rxtx.c b/drivers/net/virtio/virtio_rxtx.c
index e9b1b496e..0296b5a2f 100644
--- a/drivers/net/virtio/virtio_rxtx.c
+++ b/drivers/net/virtio/virtio_rxtx.c
@@ -122,6 +122,44 @@ virtqueue_dequeue_burst_rx(struct virtqueue *vq, struct rte_mbuf **rx_pkts,
 	return i;
 }
 
+static uint16_t
+virtqueue_dequeue_rx_inorder(struct virtqueue *vq,
+			struct rte_mbuf **rx_pkts,
+			uint32_t *len,
+			uint16_t num)
+{
+	struct vring_used_elem *uep;
+	struct rte_mbuf *cookie;
+	uint16_t used_idx = 0;
+	uint16_t i;
+
+	if (unlikely(num == 0))
+		return 0;
+
+	for (i = 0; i < num; i++) {
+		used_idx = vq->vq_used_cons_idx & (vq->vq_nentries - 1);
+		/* Desc idx same as used idx */
+		uep = &vq->vq_ring.used->ring[used_idx];
+		len[i] = uep->len;
+		cookie = (struct rte_mbuf *)vq->vq_descx[used_idx].cookie;
+
+		if (unlikely(cookie == NULL)) {
+			PMD_DRV_LOG(ERR, "vring descriptor with no mbuf cookie at %u",
+				vq->vq_used_cons_idx);
+			break;
+		}
+
+		rte_prefetch0(cookie);
+		rte_packet_prefetch(rte_pktmbuf_mtod(cookie, void *));
+		rx_pkts[i]  = cookie;
+		vq->vq_used_cons_idx++;
+		vq->vq_descx[used_idx].cookie = NULL;
+	}
+
+	vq_ring_free_inorder(vq, used_idx, i);
+	return i;
+}
+
 #ifndef DEFAULT_TX_FREE_THRESH
 #define DEFAULT_TX_FREE_THRESH 32
 #endif
@@ -150,6 +188,83 @@ virtio_xmit_cleanup(struct virtqueue *vq, uint16_t num)
 	}
 }
 
+/* Cleanup from completed inorder transmits. */
+static void
+virtio_xmit_cleanup_inorder(struct virtqueue *vq, uint16_t num)
+{
+	uint16_t i, used_idx, desc_idx, last_idx;
+	int16_t free_cnt = 0;
+	struct vq_desc_extra *dxp = NULL;
+
+	if (unlikely(num == 0))
+		return;
+
+	for (i = 0; i < num; i++) {
+		struct vring_used_elem *uep;
+
+		used_idx = vq->vq_used_cons_idx & (vq->vq_nentries - 1);
+		uep = &vq->vq_ring.used->ring[used_idx];
+		desc_idx = (uint16_t)uep->id;
+
+		dxp = &vq->vq_descx[desc_idx];
+		vq->vq_used_cons_idx++;
+
+		if (dxp->cookie != NULL) {
+			rte_pktmbuf_free(dxp->cookie);
+			dxp->cookie = NULL;
+		}
+	}
+
+	last_idx = desc_idx + dxp->ndescs - 1;
+	free_cnt = last_idx - vq->vq_desc_tail_idx;
+	if (free_cnt <= 0)
+		free_cnt += vq->vq_nentries;
+
+	vq_ring_free_inorder(vq, last_idx, free_cnt);
+}
+
+static inline int
+virtqueue_enqueue_recv_refill_inorder(struct virtqueue *vq,
+			struct rte_mbuf **cookies,
+			uint16_t num)
+{
+	struct vq_desc_extra *dxp;
+	struct virtio_hw *hw = vq->hw;
+	struct vring_desc *start_dp;
+	uint16_t head_idx, idx, i = 0;
+
+	if (unlikely(vq->vq_free_cnt == 0))
+		return -ENOSPC;
+	if (unlikely(vq->vq_free_cnt < num))
+		return -EMSGSIZE;
+
+	head_idx = vq->vq_desc_head_idx & (vq->vq_nentries - 1);
+	start_dp = vq->vq_ring.desc;
+
+	while (i < num) {
+		idx = head_idx & (vq->vq_nentries - 1);
+		dxp = &vq->vq_descx[idx];
+		dxp->cookie = (void *)cookies[i];
+		dxp->ndescs = 1;
+
+		start_dp[idx].addr =
+				VIRTIO_MBUF_ADDR(cookies[i], vq) +
+				RTE_PKTMBUF_HEADROOM - hw->vtnet_hdr_size;
+		start_dp[idx].len =
+				cookies[i]->buf_len -
+				RTE_PKTMBUF_HEADROOM +
+				hw->vtnet_hdr_size;
+		start_dp[idx].flags =  VRING_DESC_F_WRITE;
+
+		vq_update_avail_ring(vq, idx);
+		head_idx++;
+		i++;
+	}
+
+	vq->vq_desc_head_idx += num;
+	vq->vq_free_cnt = (uint16_t)(vq->vq_free_cnt - num);
+	return 0;
+}
 
 static inline int
 virtqueue_enqueue_recv_refill(struct virtqueue *vq, struct rte_mbuf *cookie)
@@ -295,9 +410,65 @@ virtqueue_xmit_offload(struct virtio_net_hdr *hdr,
 	}
 }
 
+static inline void
+virtqueue_enqueue_xmit_inorder(struct virtnet_tx *txvq,
+			struct rte_mbuf **cookies,
+			uint16_t num)
+{
+	struct vq_desc_extra *dxp;
+	struct virtqueue *vq = txvq->vq;
+	struct vring_desc *start_dp;
+	struct virtio_net_hdr *hdr;
+	uint16_t idx;
+	uint16_t head_size = vq->hw->vtnet_hdr_size;
+	int offload;
+	uint16_t i = 0;
+
+	idx = vq->vq_desc_head_idx;
+	start_dp = vq->vq_ring.desc;
+
+	offload = tx_offload_enabled(vq->hw);
+
+	while (i < num) {
+		idx = idx & (vq->vq_nentries - 1);
+		dxp = &vq->vq_descx[idx];
+		dxp->cookie = (void *)cookies[i];
+		dxp->ndescs = 1;
+
+		hdr = (struct virtio_net_hdr *)
+			rte_pktmbuf_prepend(cookies[i], head_size);
+		cookies[i]->pkt_len -= head_size;
+
+		/* if offload disabled, it is not zeroed below, do it now */
+		if (offload == 0) {
+			ASSIGN_UNLESS_EQUAL(hdr->csum_start, 0);
+			ASSIGN_UNLESS_EQUAL(hdr->csum_offset, 0);
+			ASSIGN_UNLESS_EQUAL(hdr->flags, 0);
+			ASSIGN_UNLESS_EQUAL(hdr->gso_type, 0);
+			ASSIGN_UNLESS_EQUAL(hdr->gso_size, 0);
+			ASSIGN_UNLESS_EQUAL(hdr->hdr_len, 0);
+		}
+
+		virtqueue_xmit_offload(hdr, cookies[i], offload);
+
+		start_dp[idx].addr  = VIRTIO_MBUF_DATA_DMA_ADDR(cookies[i], vq);
+		start_dp[idx].len   = cookies[i]->data_len;
+		start_dp[idx].flags = 0;
+
+		vq_update_avail_ring(vq, idx);
+
+		idx++;
+		i++;
+	};
+
+	vq->vq_free_cnt = (uint16_t)(vq->vq_free_cnt - num);
+	vq->vq_desc_head_idx = idx & (vq->vq_nentries - 1);
+}
+
 static inline void
 virtqueue_enqueue_xmit(struct virtnet_tx *txvq, struct rte_mbuf *cookie,
-		       uint16_t needed, int use_indirect, int can_push)
+			uint16_t needed, int use_indirect, int can_push,
+			int in_order)
 {
 	struct virtio_tx_region *txr = txvq->virtio_net_hdr_mz->addr;
 	struct vq_desc_extra *dxp;
@@ -310,6 +481,7 @@ virtqueue_enqueue_xmit(struct virtnet_tx *txvq, struct rte_mbuf *cookie,
 	int offload;
 
 	offload = tx_offload_enabled(vq->hw);
+
 	head_idx = vq->vq_desc_head_idx;
 	idx = head_idx;
 	dxp = &vq->vq_descx[idx];
@@ -326,6 +498,7 @@ virtqueue_enqueue_xmit(struct virtnet_tx *txvq, struct rte_mbuf *cookie,
 		 * which is wrong. Below subtract restores correct pkt size.
 		 */
 		cookie->pkt_len -= head_size;
+
 		/* if offload disabled, it is not zeroed below, do it now */
 		if (offload == 0) {
 			ASSIGN_UNLESS_EQUAL(hdr->csum_start, 0);
@@ -376,11 +549,15 @@ virtqueue_enqueue_xmit(struct virtnet_tx *txvq, struct rte_mbuf *cookie,
 	if (use_indirect)
 		idx = vq->vq_ring.desc[head_idx].next;
 
-	vq->vq_desc_head_idx = idx;
-	if (vq->vq_desc_head_idx == VQ_RING_DESC_CHAIN_END)
-		vq->vq_desc_tail_idx = idx;
 	vq->vq_free_cnt = (uint16_t)(vq->vq_free_cnt - needed);
+
+	vq->vq_desc_head_idx = idx;
 	vq_update_avail_ring(vq, head_idx);
+
+	if (!in_order) {
+		if (vq->vq_desc_head_idx == VQ_RING_DESC_CHAIN_END)
+			vq->vq_desc_tail_idx = idx;
+	}
 }
 
 void
@@ -435,7 +612,7 @@ virtio_dev_rx_queue_setup_finish(struct rte_eth_dev *dev, uint16_t queue_idx)
 	struct virtnet_rx *rxvq = &vq->rxq;
 	struct rte_mbuf *m;
 	uint16_t desc_idx;
-	int error, nbufs;
+	int error, nbufs, i;
 
 	PMD_INIT_FUNC_TRACE();
 
@@ -465,6 +642,24 @@ virtio_dev_rx_queue_setup_finish(struct rte_eth_dev *dev, uint16_t queue_idx)
 			virtio_rxq_rearm_vec(rxvq);
 			nbufs += RTE_VIRTIO_VPMD_RX_REARM_THRESH;
 		}
+	} else if (hw->use_inorder_rx) {
+		if ((!virtqueue_full(vq))) {
+			uint16_t free_cnt = vq->vq_free_cnt;
+			struct rte_mbuf *pkts[free_cnt];
+
+			if (!rte_pktmbuf_alloc_bulk(rxvq->mpool, pkts, free_cnt)) {
+				error = virtqueue_enqueue_recv_refill_inorder(vq,
+						pkts,
+						free_cnt);
+				if (unlikely(error)) {
+					for (i = 0; i < free_cnt; i++)
+						rte_pktmbuf_free(pkts[i]);
+				}
+			}
+
+			nbufs += free_cnt;
+			vq_update_avail_idx(vq);
+		}
 	} else {
 		while (!virtqueue_full(vq)) {
 			m = rte_mbuf_raw_alloc(rxvq->mpool);
@@ -574,6 +769,8 @@ virtio_dev_tx_queue_setup_finish(struct rte_eth_dev *dev,
 		for (desc_idx = mid_idx; desc_idx < vq->vq_nentries;
 		     desc_idx++)
 			vq->vq_ring.avail->ring[desc_idx] = desc_idx;
+	} else if (hw->use_inorder_tx) {
+		vq->vq_ring.desc[vq->vq_nentries - 1].next = 0;
 	}
 
 	VIRTQUEUE_DUMP(vq);
@@ -590,6 +787,19 @@ virtio_discard_rxbuf(struct virtqueue *vq, struct rte_mbuf *m)
 	 * successful since it was just dequeued.
 	 */
 	error = virtqueue_enqueue_recv_refill(vq, m);
+
+	if (unlikely(error)) {
+		RTE_LOG(ERR, PMD, "cannot requeue discarded mbuf");
+		rte_pktmbuf_free(m);
+	}
+}
+
+static void
+virtio_discard_rxbuf_inorder(struct virtqueue *vq, struct rte_mbuf *m)
+{
+	int error;
+
+	error = virtqueue_enqueue_recv_refill_inorder(vq, &m, 1);
 	if (unlikely(error)) {
 		RTE_LOG(ERR, PMD, "cannot requeue discarded mbuf");
 		rte_pktmbuf_free(m);
@@ -826,6 +1036,191 @@ virtio_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
 	return nb_rx;
 }
 
+uint16_t
+virtio_recv_mergeable_pkts_inorder(void *rx_queue,
+			struct rte_mbuf **rx_pkts,
+			uint16_t nb_pkts)
+{
+	struct virtnet_rx *rxvq = rx_queue;
+	struct virtqueue *vq = rxvq->vq;
+	struct virtio_hw *hw = vq->hw;
+	struct rte_mbuf *rxm;
+	struct rte_mbuf *prev;
+	uint16_t nb_used, num, nb_rx;
+	uint32_t len[VIRTIO_MBUF_BURST_SZ];
+	struct rte_mbuf *rcv_pkts[VIRTIO_MBUF_BURST_SZ];
+	int error;
+	uint32_t nb_enqueued;
+	uint32_t seg_num;
+	uint32_t seg_res;
+	uint32_t hdr_size;
+	int32_t i;
+	int offload;
+
+	nb_rx = 0;
+	if (unlikely(hw->started == 0))
+		return nb_rx;
+
+	nb_used = VIRTQUEUE_NUSED(vq);
+	nb_used = RTE_MIN(nb_used, nb_pkts);
+	nb_used = RTE_MIN(nb_used, VIRTIO_MBUF_BURST_SZ);
+
+	virtio_rmb();
+
+	PMD_RX_LOG(DEBUG, "used:%d", nb_used);
+
+	nb_enqueued = 0;
+	seg_num = 1;
+	seg_res = 0;
+	hdr_size = hw->vtnet_hdr_size;
+	offload = rx_offload_enabled(hw);
+
+	num = virtqueue_dequeue_rx_inorder(vq, rcv_pkts, len, nb_used);
+
+	for (i = 0; i < num; i++) {
+		struct virtio_net_hdr_mrg_rxbuf *header;
+
+		PMD_RX_LOG(DEBUG, "dequeue:%d", num);
+		PMD_RX_LOG(DEBUG, "packet len:%d", len[i]);
+
+		rxm = rcv_pkts[i];
+
+		if (unlikely(len[i] < hdr_size + ETHER_HDR_LEN)) {
+			PMD_RX_LOG(ERR, "Packet drop");
+			nb_enqueued++;
+			virtio_discard_rxbuf_inorder(vq, rxm);
+			rxvq->stats.errors++;
+			continue;
+		}
+
+		header = (struct virtio_net_hdr_mrg_rxbuf *)((char *)rxm->buf_addr +
+			RTE_PKTMBUF_HEADROOM - hdr_size);
+		seg_num = header->num_buffers;
+
+		if (seg_num == 0)
+			seg_num = 1;
+
+		rxm->data_off = RTE_PKTMBUF_HEADROOM;
+		rxm->nb_segs = seg_num;
+		rxm->ol_flags = 0;
+		rxm->vlan_tci = 0;
+		rxm->pkt_len = (uint32_t)(len[i] - hdr_size);
+		rxm->data_len = (uint16_t)(len[i] - hdr_size);
+
+		rxm->port = rxvq->port_id;
+
+		rx_pkts[nb_rx] = rxm;
+		prev = rxm;
+
+		if (offload && virtio_rx_offload(rxm, &header->hdr) < 0) {
+			virtio_discard_rxbuf_inorder(vq, rxm);
+			rxvq->stats.errors++;
+			continue;
+		}
+
+		if (hw->vlan_strip)
+			rte_vlan_strip(rx_pkts[nb_rx]);
+
+		seg_res = seg_num - 1;
+
+		/* Merge remaining segments */
+		while (seg_res != 0 && i < (num - 1)) {
+			i++;
+
+			rxm = rcv_pkts[i];
+			rxm->data_off = RTE_PKTMBUF_HEADROOM - hdr_size;
+			rxm->pkt_len = (uint32_t)(len[i]);
+			rxm->data_len = (uint16_t)(len[i]);
+
+			rx_pkts[nb_rx]->pkt_len += (uint32_t)(len[i]);
+			rx_pkts[nb_rx]->data_len += (uint16_t)(len[i]);
+
+			if (prev)
+				prev->next = rxm;
+
+			prev = rxm;
+			seg_res -= 1;
+		}
+
+		if (!seg_res) {
+			virtio_rx_stats_updated(rxvq, rx_pkts[nb_rx]);
+			nb_rx++;
+		}
+	}
+
+	/* Last packet still need merge segments */
+	while (seg_res != 0) {
+		uint16_t rcv_cnt = RTE_MIN((uint16_t)seg_res,
+					VIRTIO_MBUF_BURST_SZ);
+
+		prev = rcv_pkts[nb_rx];
+		if (likely(VIRTQUEUE_NUSED(vq) >= rcv_cnt)) {
+			num = virtqueue_dequeue_rx_inorder(vq, rcv_pkts, len,
+							   rcv_cnt);
+			uint16_t extra_idx = 0;
+
+			rcv_cnt = num;
+			while (extra_idx < rcv_cnt) {
+				rxm = rcv_pkts[extra_idx];
+				rxm->data_off = RTE_PKTMBUF_HEADROOM - hdr_size;
+				rxm->pkt_len = (uint32_t)(len[extra_idx]);
+				rxm->data_len = (uint16_t)(len[extra_idx]);
+				prev->next = rxm;
+				prev = rxm;
+				rx_pkts[nb_rx]->pkt_len += len[extra_idx];
+				rx_pkts[nb_rx]->data_len += len[extra_idx];
+				extra_idx += 1;
+			};
+			seg_res -= rcv_cnt;
+
+			if (!seg_res) {
+				virtio_rx_stats_updated(rxvq, rx_pkts[nb_rx]);
+				nb_rx++;
+			}
+		} else {
+			PMD_RX_LOG(ERR,
+					"No enough segments for packet.");
+			virtio_discard_rxbuf_inorder(vq, prev);
+			rxvq->stats.errors++;
+			break;
+		}
+	}
+
+	rxvq->stats.packets += nb_rx;
+
+	/* Allocate new mbuf for the used descriptor */
+
+	if (likely(!virtqueue_full(vq))) {
+		/* free_cnt may include mrg descs */
+		uint16_t free_cnt = vq->vq_free_cnt;
+		struct rte_mbuf *new_pkts[free_cnt];
+
+		if (!rte_pktmbuf_alloc_bulk(rxvq->mpool, new_pkts, free_cnt)) {
+			error = virtqueue_enqueue_recv_refill_inorder(vq, new_pkts,
+					free_cnt);
+			if (unlikely(error)) {
+				for (i = 0; i < free_cnt; i++)
+					rte_pktmbuf_free(new_pkts[i]);
+			}
+			nb_enqueued += free_cnt;
+		} else {
+			struct rte_eth_dev *dev = &rte_eth_devices[rxvq->port_id];
+			dev->data->rx_mbuf_alloc_failed += free_cnt;
+		}
+	}
+
+	if (likely(nb_enqueued)) {
+		vq_update_avail_idx(vq);
+
+		if (unlikely(virtqueue_kick_prepare(vq))) {
+			virtqueue_notify(vq);
+			PMD_RX_LOG(DEBUG, "Notified");
+		}
+	}
+
+	return nb_rx;
+}
+
 uint16_t
 virtio_recv_mergeable_pkts(void *rx_queue,
 			struct rte_mbuf **rx_pkts,
@@ -1073,12 +1468,124 @@ virtio_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
 		}
 
 		/* Enqueue Packet buffers */
-		virtqueue_enqueue_xmit(txvq, txm, slots, use_indirect, can_push);
+		virtqueue_enqueue_xmit(txvq, txm, slots, use_indirect,
+			can_push, 0);
+
+		txvq->stats.bytes += txm->pkt_len;
+		virtio_update_packet_stats(&txvq->stats, txm);
+	}
+
+	txvq->stats.packets += nb_tx;
+
+	if (likely(nb_tx)) {
+		vq_update_avail_idx(vq);
+
+		if (unlikely(virtqueue_kick_prepare(vq))) {
+			virtqueue_notify(vq);
+			PMD_TX_LOG(DEBUG, "Notified backend after xmit");
+		}
+	}
+
+	return nb_tx;
+}
+
+uint16_t
+virtio_xmit_pkts_inorder(void *tx_queue,
+			struct rte_mbuf **tx_pkts,
+			uint16_t nb_pkts)
+{
+	struct virtnet_tx *txvq = tx_queue;
+	struct virtqueue *vq = txvq->vq;
+	struct virtio_hw *hw = vq->hw;
+	uint16_t hdr_size = hw->vtnet_hdr_size;
+	uint16_t nb_used, nb_avail, nb_tx = 0, nb_inorder_pkts = 0;
+	struct rte_mbuf *inorder_pkts[nb_pkts];
+	int error;
+
+	if (unlikely(hw->started == 0 && tx_pkts != hw->inject_pkts))
+		return nb_tx;
+
+	if (unlikely(nb_pkts < 1))
+		return nb_pkts;
+
+	VIRTQUEUE_DUMP(vq);
+	PMD_TX_LOG(DEBUG, "%d packets to xmit", nb_pkts);
+	nb_used = VIRTQUEUE_NUSED(vq);
+
+	virtio_rmb();
+	if (likely(nb_used > vq->vq_nentries - vq->vq_free_thresh))
+		virtio_xmit_cleanup_inorder(vq, nb_used);
+
+	if (unlikely(!vq->vq_free_cnt))
+		virtio_xmit_cleanup_inorder(vq, nb_used);
+
+	nb_avail = RTE_MIN(vq->vq_free_cnt, nb_pkts);
+
+	for (nb_tx = 0; nb_tx < nb_avail; nb_tx++) {
+		struct rte_mbuf *txm = tx_pkts[nb_tx];
+		int slots, need;
+
+		/* Do VLAN tag insertion */
+		if (unlikely(txm->ol_flags & PKT_TX_VLAN_PKT)) {
+			error = rte_vlan_insert(&txm);
+			if (unlikely(error)) {
+				rte_pktmbuf_free(txm);
+				continue;
+			}
+		}
+
+		/* optimize ring usage */
+		if ((vtpci_with_feature(hw, VIRTIO_F_ANY_LAYOUT) ||
+		     vtpci_with_feature(hw, VIRTIO_F_VERSION_1)) &&
+		     rte_mbuf_refcnt_read(txm) == 1 &&
+		     RTE_MBUF_DIRECT(txm) &&
+		     txm->nb_segs == 1 &&
+		     rte_pktmbuf_headroom(txm) >= hdr_size &&
+		     rte_is_aligned(rte_pktmbuf_mtod(txm, char *),
+				__alignof__(struct virtio_net_hdr_mrg_rxbuf))) {
+			inorder_pkts[nb_inorder_pkts] = txm;
+			nb_inorder_pkts++;
+
+			txvq->stats.bytes += txm->pkt_len;
+			virtio_update_packet_stats(&txvq->stats, txm);
+			continue;
+		}
+
+		if (nb_inorder_pkts) {
+			virtqueue_enqueue_xmit_inorder(txvq, inorder_pkts,
+							nb_inorder_pkts);
+			nb_inorder_pkts = 0;
+		}
+
+		slots = txm->nb_segs + 1;
+		need = slots - vq->vq_free_cnt;
+		if (unlikely(need > 0)) {
+			nb_used = VIRTQUEUE_NUSED(vq);
+			virtio_rmb();
+			need = RTE_MIN(need, (int)nb_used);
+
+			virtio_xmit_cleanup_inorder(vq, need);
+
+			need = slots - vq->vq_free_cnt;
+
+			if (unlikely(need > 0)) {
+				PMD_TX_LOG(ERR,
+					"No free tx descriptors to transmit");
+				break;
+			}
+		}
+		/* Enqueue Packet buffers */
+		virtqueue_enqueue_xmit(txvq, txm, slots, 0, 0, 1);
 
 		txvq->stats.bytes += txm->pkt_len;
 		virtio_update_packet_stats(&txvq->stats, txm);
 	}
 
+	/* Transmit all inorder packets */
+	if (nb_inorder_pkts)
+		virtqueue_enqueue_xmit_inorder(txvq, inorder_pkts,
+						nb_inorder_pkts);
+
 	txvq->stats.packets += nb_tx;
 
 	if (likely(nb_tx)) {
@@ -1090,5 +1597,7 @@ virtio_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
 		}
 	}
 
+	VIRTQUEUE_DUMP(vq);
+
 	return nb_tx;
 }
-- 
2.17.0

[dpdk-dev] [PATCH v3 08/11] net/virtio: add in-order Rx/Tx into selection

[Marvin Liu]

After IN_ORDER Rx/Tx paths added, need to update Rx/Tx path selection
logic.

Rx path select logic: If IN_ORDER is disabled will select normal Rx
path. If IN_ORDER is enabled, Rx offload and merge-able are disabled
will select simple Rx path. Otherwise will select IN_ORDER Rx path.

Tx path select logic: If IN_ORDER is disabled will select normal Tx
path. If IN_ORDER is enabled and merge-able is disabled will select
simple Tx path. Otherwise will select IN_ORDER Tx path.

Signed-off-by: Marvin Liu <yong.liu@intel.com>

diff --git a/drivers/net/virtio/virtio_ethdev.c b/drivers/net/virtio/virtio_ethdev.c
index df50a571a..2b3d65f80 100644
--- a/drivers/net/virtio/virtio_ethdev.c
+++ b/drivers/net/virtio/virtio_ethdev.c
@@ -1320,6 +1320,11 @@ set_rxtx_funcs(struct rte_eth_dev *eth_dev)
 		PMD_INIT_LOG(INFO, "virtio: using simple Rx path on port %u",
 			eth_dev->data->port_id);
 		eth_dev->rx_pkt_burst = virtio_recv_pkts_vec;
+	} else if (hw->use_inorder_rx) {
+		PMD_INIT_LOG(INFO,
+			"virtio: using inorder mergeable buffer Rx path on port %u",
+			eth_dev->data->port_id);
+		eth_dev->rx_pkt_burst = &virtio_recv_mergeable_pkts_inorder;
 	} else if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF)) {
 		PMD_INIT_LOG(INFO,
 			"virtio: using mergeable buffer Rx path on port %u",
@@ -1335,6 +1340,10 @@ set_rxtx_funcs(struct rte_eth_dev *eth_dev)
 		PMD_INIT_LOG(INFO, "virtio: using simple Tx path on port %u",
 			eth_dev->data->port_id);
 		eth_dev->tx_pkt_burst = virtio_xmit_pkts_simple;
+	} else if (hw->use_inorder_tx) {
+		PMD_INIT_LOG(INFO, "virtio: using inorder Tx path on port %u",
+			eth_dev->data->port_id);
+		eth_dev->tx_pkt_burst = virtio_xmit_pkts_inorder;
 	} else {
 		PMD_INIT_LOG(INFO, "virtio: using standard Tx path on port %u",
 			eth_dev->data->port_id);
@@ -1871,24 +1880,24 @@ virtio_dev_configure(struct rte_eth_dev *dev)
 
 	rte_spinlock_init(&hw->state_lock);
 
-	hw->use_simple_rx = 1;
-	hw->use_simple_tx = 1;
-
 #if defined RTE_ARCH_ARM64 || defined RTE_ARCH_ARM
 	if (!rte_cpu_get_flag_enabled(RTE_CPUFLAG_NEON)) {
 		hw->use_simple_rx = 0;
 		hw->use_simple_tx = 0;
 	}
 #endif
-	if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF)) {
-		hw->use_simple_rx = 0;
-		hw->use_simple_tx = 0;
+	if (vtpci_with_feature(hw, VIRTIO_F_IN_ORDER)) {
+		if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF)) {
+			hw->use_inorder_rx = 1;
+			hw->use_inorder_tx = 1;
+		} else {
+			hw->use_simple_tx = 1;
+			if (!(rx_offloads & (DEV_RX_OFFLOAD_UDP_CKSUM |
+					     DEV_RX_OFFLOAD_TCP_CKSUM)))
+				hw->use_simple_rx = 1;
+		}
 	}
 
-	if (rx_offloads & (DEV_RX_OFFLOAD_UDP_CKSUM |
-			   DEV_RX_OFFLOAD_TCP_CKSUM))
-		hw->use_simple_rx = 0;
-
 	return 0;
 }
 
-- 
2.17.0

[dpdk-dev] [PATCH v3 09/11] net/virtio: advertise support in-order feature

[Marvin Liu]

Signed-off-by: Marvin Liu <yong.liu@intel.com>

diff --git a/drivers/net/virtio/virtio_ethdev.h b/drivers/net/virtio/virtio_ethdev.h
index 8ab1267b0..69da18d9c 100644
--- a/drivers/net/virtio/virtio_ethdev.h
+++ b/drivers/net/virtio/virtio_ethdev.h
@@ -36,6 +36,7 @@
 	 1ULL << VIRTIO_NET_F_GUEST_ANNOUNCE |	\
 	 1u << VIRTIO_RING_F_INDIRECT_DESC |    \
 	 1ULL << VIRTIO_F_VERSION_1       |	\
+	 1ULL << VIRTIO_F_IN_ORDER        |	\
 	 1ULL << VIRTIO_F_IOMMU_PLATFORM)
 
 #define VIRTIO_PMD_SUPPORTED_GUEST_FEATURES	\
diff --git a/drivers/net/virtio/virtio_user/virtio_user_dev.c b/drivers/net/virtio/virtio_user/virtio_user_dev.c
index e90257f5d..1050aed05 100644
--- a/drivers/net/virtio/virtio_user/virtio_user_dev.c
+++ b/drivers/net/virtio/virtio_user/virtio_user_dev.c
@@ -371,6 +371,7 @@ virtio_user_dev_setup(struct virtio_user_dev *dev)
 	 1ULL << VIRTIO_NET_F_GUEST_CSUM	|	\
 	 1ULL << VIRTIO_NET_F_GUEST_TSO4	|	\
 	 1ULL << VIRTIO_NET_F_GUEST_TSO6	|	\
+	 1ULL << VIRTIO_F_IN_ORDER		|   \
 	 1ULL << VIRTIO_F_VERSION_1)
 
 int
-- 
2.17.0

[dpdk-dev] [PATCH v3 10/11] doc: add mrg_rxbuf and in_order parameters for virtio device

[Marvin Liu]

Virtio vdev parameter mrg_rxbuf and in_order can configure
VIRTIO_NET_F_MRG_RXBUF and VIRTIO_F_IN_ORDER feature bits. These feature
bits can effect [RT]x path selection.

Signed-off-by: Marvin Liu <yong.liu@intel.com>

diff --git a/doc/guides/nics/virtio.rst b/doc/guides/nics/virtio.rst
index a42d1bb30..46e292c4d 100644
--- a/doc/guides/nics/virtio.rst
+++ b/doc/guides/nics/virtio.rst
@@ -331,3 +331,13 @@ The user can specify below argument in devargs.
     driver, and works as a HW vhost backend. This argument is used to specify
     a virtio device needs to work in vDPA mode.
     (Default: 0 (disabled))
+
+#. ``mrg_rxbuf``:
+
+    It is used to enable virtio device mergeable Rx buffer feature.
+    (Default: 1 (enabled))
+
+#. ``in_order``:
+
+    It is used to enable virtio device in-order feature.
+    (Default: 1 (enabled))
-- 
2.17.0

[dpdk-dev] [PATCH v3 11/11] doc: add in-order funciton for virtio device

[Marvin Liu]

Signed-off-by: Marvin Liu <yong.liu@intel.com>

diff --git a/doc/guides/nics/virtio.rst b/doc/guides/nics/virtio.rst
index 46e292c4d..1b88cec96 100644
--- a/doc/guides/nics/virtio.rst
+++ b/doc/guides/nics/virtio.rst
@@ -201,7 +201,7 @@ The packet transmission flow is:
 Virtio PMD Rx/Tx Callbacks
 --------------------------
 
-Virtio driver has 3 Rx callbacks and 2 Tx callbacks.
+Virtio driver has 4 Rx callbacks and 3 Tx callbacks.
 
 Rx callbacks:
 
@@ -215,6 +215,9 @@ Rx callbacks:
    Vector version without mergeable Rx buffer support, also fixes the available
    ring indexes and uses vector instructions to optimize performance.
 
+#. ``virtio_recv_mergeable_pkts_inorder``:
+   In-order version with mergeable Rx buffer support.
+
 Tx callbacks:
 
 #. ``virtio_xmit_pkts``:
@@ -223,13 +226,17 @@ Tx callbacks:
 #. ``virtio_xmit_pkts_simple``:
    Vector version fixes the available ring indexes to optimize performance.
 
+#. ``virtio_xmit_pkts_inorder``:
+   In-order version.
 
-By default, the non-vector callbacks are used:
+By default, the in-order callbacks are used:
 
 *   For Rx: If mergeable Rx buffers is disabled then ``virtio_recv_pkts`` is
-    used; otherwise ``virtio_recv_mergeable_pkts``.
+    used; If in-order is enabled then ``virtio_recv_mergeable_pkts_inirder``;
+    otherwise ``virtio_recv_mergeable_pkts``.
 
-*   For Tx: ``virtio_xmit_pkts``.
+*   For Tx: If in-order is enabled then ``virtio_xmit_pkts_inorder``;
+    otherwise ``virtio_xmit_pkts``.
 
 
 Vector callbacks will be used when:
@@ -242,6 +249,8 @@ Vector callbacks will be used when:
 
 *   Mergeable Rx buffers is disabled.
 
+*   In-order is enabled
+
 The corresponding callbacks are:
 
 *   For Rx: ``virtio_recv_pkts_vec``.
-- 
2.17.0

Re: [dpdk-dev] [PATCH v3 08/11] net/virtio: add in-order Rx/Tx into selection

[Maxime Coquelin]

On 06/28/2018 05:39 PM, Liu, Yong wrote:
> 
> 
>> -----Original Message-----
>> From: Maxime Coquelin [mailto:maxime.coquelin@redhat.com]
>> Sent: Thursday, June 28, 2018 11:16 PM
>> To: Liu, Yong <yong.liu@intel.com>; Bie, Tiwei <tiwei.bie@intel.com>
>> Cc: Wang, Zhihong <zhihong.wang@intel.com>; dev@dpdk.org
>> Subject: Re: [PATCH v3 08/11] net/virtio: add in-order Rx/Tx into
>> selection
>>
>>
>>
>> On 06/28/2018 11:52 PM, Marvin Liu wrote:
>>> After IN_ORDER Rx/Tx paths added, need to update Rx/Tx path selection
>>> logic.
>>>
>>> Rx path select logic: If IN_ORDER is disabled will select normal Rx
>>> path. If IN_ORDER is enabled, Rx offload and merge-able are disabled
>>> will select simple Rx path. Otherwise will select IN_ORDER Rx path.
>>>
>>> Tx path select logic: If IN_ORDER is disabled will select normal Tx
>>> path. If IN_ORDER is enabled and merge-able is disabled will select
>>> simple Tx path. Otherwise will select IN_ORDER Tx path.
>>>
>>> Signed-off-by: Marvin Liu <yong.liu@intel.com>
>>>
>>> diff --git a/drivers/net/virtio/virtio_ethdev.c
>> b/drivers/net/virtio/virtio_ethdev.c
>>> index df50a571a..2b3d65f80 100644
>>> --- a/drivers/net/virtio/virtio_ethdev.c
>>> +++ b/drivers/net/virtio/virtio_ethdev.c
>>> @@ -1320,6 +1320,11 @@ set_rxtx_funcs(struct rte_eth_dev *eth_dev)
>>>    		PMD_INIT_LOG(INFO, "virtio: using simple Rx path on port %u",
>>>    			eth_dev->data->port_id);
>>>    		eth_dev->rx_pkt_burst = virtio_recv_pkts_vec;
>>> +	} else if (hw->use_inorder_rx) {
>>> +		PMD_INIT_LOG(INFO,
>>> +			"virtio: using inorder mergeable buffer Rx path on
>> port %u",
>>> +			eth_dev->data->port_id);
>>> +		eth_dev->rx_pkt_burst = &virtio_recv_mergeable_pkts_inorder;
>>>    	} else if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF)) {
>>>    		PMD_INIT_LOG(INFO,
>>>    			"virtio: using mergeable buffer Rx path on port %u",
>>> @@ -1335,6 +1340,10 @@ set_rxtx_funcs(struct rte_eth_dev *eth_dev)
>>>    		PMD_INIT_LOG(INFO, "virtio: using simple Tx path on port %u",
>>>    			eth_dev->data->port_id);
>>>    		eth_dev->tx_pkt_burst = virtio_xmit_pkts_simple;
>>> +	} else if (hw->use_inorder_tx) {
>>> +		PMD_INIT_LOG(INFO, "virtio: using inorder Tx path on port %u",
>>> +			eth_dev->data->port_id);
>>> +		eth_dev->tx_pkt_burst = virtio_xmit_pkts_inorder;
>>>    	} else {
>>>    		PMD_INIT_LOG(INFO, "virtio: using standard Tx path on port %u",
>>>    			eth_dev->data->port_id);
>>> @@ -1871,24 +1880,24 @@ virtio_dev_configure(struct rte_eth_dev *dev)
>>>
>>>    	rte_spinlock_init(&hw->state_lock);
>>>
>>> -	hw->use_simple_rx = 1;
>>> -	hw->use_simple_tx = 1;
>>> -
>>>    #if defined RTE_ARCH_ARM64 || defined RTE_ARCH_ARM
>>>    	if (!rte_cpu_get_flag_enabled(RTE_CPUFLAG_NEON)) {
>>>    		hw->use_simple_rx = 0;
>>>    		hw->use_simple_tx = 0;
>>>    	}
>>>    #endif
>>> -	if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF)) {
>>> -		hw->use_simple_rx = 0;
>>> -		hw->use_simple_tx = 0;
>>> +	if (vtpci_with_feature(hw, VIRTIO_F_IN_ORDER)) {
>>> +		if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF)) {
>>> +			hw->use_inorder_rx = 1;
>>> +			hw->use_inorder_tx = 1;
>>> +		} else {
>>> +			hw->use_simple_tx = 1;
>>> +			if (!(rx_offloads & (DEV_RX_OFFLOAD_UDP_CKSUM |
>>> +					     DEV_RX_OFFLOAD_TCP_CKSUM)))
>>> +				hw->use_simple_rx = 1;
>>> +		}
>>
>> It seems to be wrong.
>> For example if IN_ORDER hasn't been negotiated, we might want to use the
>> simple path if no rx offload have been requested by the application.
>>
>> It was the case before the patch if I'm not mistaken.
> 
> Maxime,
> IN_ORDER is the prerequisite for selection simple rx/tx path. So when IN_ORDER + mergeable off + no rx offload will chose simple rx path.

Ok, I wonder in that case if I could just remove the simple Tx path in
my series as it is not compliant with IN_ORDER.
Tiwei, what's your take on this?

Also, I wonder if it would make sense to backport the vhost patch that 
advertize IN_ORDER features to the LTS, as no functional changes, except
that it prevents using "simple/inorder" path with dequeue zero copy,
which is a good thing.

Any thoughts?

>>
>> Also, with ARM platform, we force not to use simple path, but in case
>> IN_ORDER has been negotiated but not MRG_RXBUF, it gets re-enabled.
> 
> Will move ARM force action behind normal selection, thus can satisfy ARM.

I can change when applying.

Thanks,
Maxime

> Thanks,
> Marvin
> 
>>
>>>    	}
>>>
>>> -	if (rx_offloads & (DEV_RX_OFFLOAD_UDP_CKSUM |
>>> -			   DEV_RX_OFFLOAD_TCP_CKSUM))
>>> -		hw->use_simple_rx = 0;
>>> -
>>>    	return 0;
>>>    }
>>>
>>>

Re: [dpdk-dev] [PATCH v3 00/11] support in-order feature

[Maxime Coquelin]

Hi,

I was about to apply it, but there are quite a few checkpatch and
git-check-log.sh issues reported.

Could you please resent with this fixed?
Also, I think it would be  better to squash the doc patches in relevant
ones.

Thanks,
Maxime
On 06/28/2018 11:52 PM, Marvin Liu wrote:
> In latest virtio-spec, new feature bit VIRTIO_F_IN_ORDER was introduced.
> When this feature has been negotiated, virtio driver will use
> descriptors in ring order: starting from offset 0 in the table, and
> wrapping around at the end of the table. Vhost devices will always use
> descriptors in the same order in which they have been made available.
> This can reduce virtio accesses to used ring.
> 
> Based on updated virtio-spec, this series realized IN_ORDER prototype
> in virtio driver. Due to new [RT]x path added into selection, also add
> two new parameters mrg_rx and in_order into virtio-user vdev parameters
> list. This will allow user to configure feature bits thus can impact
> [RT]x path selection.
> 
> Performance of virtio user with IN_ORDER feature:
> 
>      Platform: Purely
>      CPU: Intel(R) Xeon(R) Platinum 8160 CPU @ 2.10GHz
>      DPDK baseline: 18.05
>      Setup: testpmd with vhost vdev + testpmd with virtio vdev
> 
>      +--------------+----------+----------+---------+
>      |Vhost->Virtio |1 Queue   |2 Queues  |4 Queues |
>      +--------------+----------+----------+---------+
>      |Inorder       |12.0Mpps  |24.2Mpps  |26.0Mpps |
>      |Normal        |12.1Mpps  |18.5Mpps  |18.9Mpps |
>      +--------------+----------+----------+---------+
>      
>      +--------------+----------+----------------+---------+
>      |Virtio->Vhost |1 Queue   |2 Queues        |4 Queues |
>      +--------------+----------+----------------+---------+
>      |Inorder       |13.8Mpps  |10.7 ~ 15.2Mpps |11.5Mpps |
>      |Normal        |13.3Mpps  |9.8 ~ 14Mpps    |10.5Mpps |
>      +--------------+----------+----------------+---------+
>      
>      +---------+----------+----------------+----------------+
>      |Loopback |1 Queue   |2 Queues        |4 Queues        |
>      +---------+----------+----------------+----------------+
>      |Inorder  |7.4Mpps   |9.1 ~ 11.6Mpps  |10.5 ~ 11.3Mpps |
>      +---------+----------+----------------+----------------+
>      |Normal   |7.5Mpps   |7.7 ~ 9.0Mpps   |7.6 ~ 7.8Mpps   |
>      +---------+----------+----------------+----------------+
> 
> v3:
> - refine [RT]x function selection logic
> - fix in-order mergable packets index error
> - combine unsupport mask patch
> - doc virtio in-order update
> - fix checkpatch error
> 
> v2:
> - merge to latest dpdk-net-virtio
> - not use in_direct for normal xmit packets
> - update available ring for each descriptor
> - clean up IN_ORDER xmit function
> - unmask feature bits when disabled in_order or mgr_rxbuf
> - extract common part between IN_ORDER and normal functions
> - update performance result
> 
> Marvin Liu (11):
>    vhost: advertise support in-order feature
>    net/virtio: add in-order feature bit definition
>    net/virtio-user: add unsupported features mask
>    net/virtio-user: add mrg_rxbuf and in_order vdev parameters
>    net/virtio: free in-order descriptors before device start
>    net/virtio: extract common part for in-order functions
>    net/virtio: support in-order Rx and Tx
>    net/virtio: add in-order Rx/Tx into selection
>    net/virtio: advertise support in-order feature
>    doc: add mrg_rxbuf and in_order parameters for virtio device
>    doc: add in-order funciton for virtio device
> 
>   doc/guides/nics/virtio.rst                    |  27 +-
>   drivers/net/virtio/virtio_ethdev.c            |  29 +-
>   drivers/net/virtio/virtio_ethdev.h            |   7 +
>   drivers/net/virtio/virtio_pci.h               |   8 +
>   drivers/net/virtio/virtio_rxtx.c              | 635 ++++++++++++++++--
>   .../net/virtio/virtio_user/virtio_user_dev.c  |  26 +-
>   .../net/virtio/virtio_user/virtio_user_dev.h  |   4 +-
>   drivers/net/virtio/virtio_user_ethdev.c       |  45 +-
>   drivers/net/virtio/virtqueue.c                |   8 +
>   drivers/net/virtio/virtqueue.h                |   2 +
>   lib/librte_vhost/socket.c                     |   6 +
>   lib/librte_vhost/vhost.h                      |  10 +-
>   12 files changed, 723 insertions(+), 84 deletions(-)
> 

Re: [dpdk-dev] [PATCH v3 08/11] net/virtio: add in-order Rx/Tx into selection

[Tiwei Bie]

On Fri, Jun 29, 2018 at 10:04:23AM +0200, Maxime Coquelin wrote:
> On 06/28/2018 05:39 PM, Liu, Yong wrote:
> > 
> > 
> > > -----Original Message-----
> > > From: Maxime Coquelin [mailto:maxime.coquelin@redhat.com]
> > > Sent: Thursday, June 28, 2018 11:16 PM
> > > To: Liu, Yong <yong.liu@intel.com>; Bie, Tiwei <tiwei.bie@intel.com>
> > > Cc: Wang, Zhihong <zhihong.wang@intel.com>; dev@dpdk.org
> > > Subject: Re: [PATCH v3 08/11] net/virtio: add in-order Rx/Tx into
> > > selection
> > > 
> > > 
> > > 
> > > On 06/28/2018 11:52 PM, Marvin Liu wrote:
> > > > After IN_ORDER Rx/Tx paths added, need to update Rx/Tx path selection
> > > > logic.
> > > > 
> > > > Rx path select logic: If IN_ORDER is disabled will select normal Rx
> > > > path. If IN_ORDER is enabled, Rx offload and merge-able are disabled
> > > > will select simple Rx path. Otherwise will select IN_ORDER Rx path.
> > > > 
> > > > Tx path select logic: If IN_ORDER is disabled will select normal Tx
> > > > path. If IN_ORDER is enabled and merge-able is disabled will select
> > > > simple Tx path. Otherwise will select IN_ORDER Tx path.
> > > > 
> > > > Signed-off-by: Marvin Liu <yong.liu@intel.com>
> > > > 
> > > > diff --git a/drivers/net/virtio/virtio_ethdev.c
> > > b/drivers/net/virtio/virtio_ethdev.c
> > > > index df50a571a..2b3d65f80 100644
> > > > --- a/drivers/net/virtio/virtio_ethdev.c
> > > > +++ b/drivers/net/virtio/virtio_ethdev.c
> > > > @@ -1320,6 +1320,11 @@ set_rxtx_funcs(struct rte_eth_dev *eth_dev)
> > > >    		PMD_INIT_LOG(INFO, "virtio: using simple Rx path on port %u",
> > > >    			eth_dev->data->port_id);
> > > >    		eth_dev->rx_pkt_burst = virtio_recv_pkts_vec;
> > > > +	} else if (hw->use_inorder_rx) {
> > > > +		PMD_INIT_LOG(INFO,
> > > > +			"virtio: using inorder mergeable buffer Rx path on
> > > port %u",
> > > > +			eth_dev->data->port_id);
> > > > +		eth_dev->rx_pkt_burst = &virtio_recv_mergeable_pkts_inorder;
> > > >    	} else if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF)) {
> > > >    		PMD_INIT_LOG(INFO,
> > > >    			"virtio: using mergeable buffer Rx path on port %u",
> > > > @@ -1335,6 +1340,10 @@ set_rxtx_funcs(struct rte_eth_dev *eth_dev)
> > > >    		PMD_INIT_LOG(INFO, "virtio: using simple Tx path on port %u",
> > > >    			eth_dev->data->port_id);
> > > >    		eth_dev->tx_pkt_burst = virtio_xmit_pkts_simple;
> > > > +	} else if (hw->use_inorder_tx) {
> > > > +		PMD_INIT_LOG(INFO, "virtio: using inorder Tx path on port %u",
> > > > +			eth_dev->data->port_id);
> > > > +		eth_dev->tx_pkt_burst = virtio_xmit_pkts_inorder;
> > > >    	} else {
> > > >    		PMD_INIT_LOG(INFO, "virtio: using standard Tx path on port %u",
> > > >    			eth_dev->data->port_id);
> > > > @@ -1871,24 +1880,24 @@ virtio_dev_configure(struct rte_eth_dev *dev)
> > > > 
> > > >    	rte_spinlock_init(&hw->state_lock);
> > > > 
> > > > -	hw->use_simple_rx = 1;
> > > > -	hw->use_simple_tx = 1;
> > > > -
> > > >    #if defined RTE_ARCH_ARM64 || defined RTE_ARCH_ARM
> > > >    	if (!rte_cpu_get_flag_enabled(RTE_CPUFLAG_NEON)) {
> > > >    		hw->use_simple_rx = 0;
> > > >    		hw->use_simple_tx = 0;
> > > >    	}
> > > >    #endif
> > > > -	if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF)) {
> > > > -		hw->use_simple_rx = 0;
> > > > -		hw->use_simple_tx = 0;
> > > > +	if (vtpci_with_feature(hw, VIRTIO_F_IN_ORDER)) {
> > > > +		if (vtpci_with_feature(hw, VIRTIO_NET_F_MRG_RXBUF)) {
> > > > +			hw->use_inorder_rx = 1;
> > > > +			hw->use_inorder_tx = 1;
> > > > +		} else {
> > > > +			hw->use_simple_tx = 1;
> > > > +			if (!(rx_offloads & (DEV_RX_OFFLOAD_UDP_CKSUM |
> > > > +					     DEV_RX_OFFLOAD_TCP_CKSUM)))
> > > > +				hw->use_simple_rx = 1;
> > > > +		}
> > > 
> > > It seems to be wrong.
> > > For example if IN_ORDER hasn't been negotiated, we might want to use the
> > > simple path if no rx offload have been requested by the application.
> > > 
> > > It was the case before the patch if I'm not mistaken.
> > 
> > Maxime,
> > IN_ORDER is the prerequisite for selection simple rx/tx path. So when IN_ORDER + mergeable off + no rx offload will chose simple rx path.
> 
> Ok, I wonder in that case if I could just remove the simple Tx path in
> my series as it is not compliant with IN_ORDER.
> Tiwei, what's your take on this?
> 
> Also, I wonder if it would make sense to backport the vhost patch that
> advertize IN_ORDER features to the LTS, as no functional changes, except
> that it prevents using "simple/inorder" path with dequeue zero copy,
> which is a good thing.
> 
> Any thoughts?

IMO it's not right to choose the current simple Tx path
when IN_ORDER is negotiated. Because in split ring, IN_ORDER
requires the driver to use the descriptors in ring order
which is not the case in the current simple Tx path, and
things will go wrong when device wants to make full use
of the IN_ORDER feature.

Best regards,
Tiwei Bie


> 
> > > 
> > > Also, with ARM platform, we force not to use simple path, but in case
> > > IN_ORDER has been negotiated but not MRG_RXBUF, it gets re-enabled.
> > 
> > Will move ARM force action behind normal selection, thus can satisfy ARM.
> 
> I can change when applying.
> 
> Thanks,
> Maxime
> 
> > Thanks,
> > Marvin
> > 
> > > 
> > > >    	}
> > > > 
> > > > -	if (rx_offloads & (DEV_RX_OFFLOAD_UDP_CKSUM |
> > > > -			   DEV_RX_OFFLOAD_TCP_CKSUM))
> > > > -		hw->use_simple_rx = 0;
> > > > -
> > > >    	return 0;
> > > >    }
> > > > 
> > > > 

Re: [dpdk-dev] [PATCH v3 00/11] support in-order feature

[Liu, Yong]

Thanks, Maxime. I have sent v4 with checkpatch and commit log warnings fixed.

Regards,
Marvin 

> -----Original Message-----
> From: Maxime Coquelin [mailto:maxime.coquelin@redhat.com]
> Sent: Friday, June 29, 2018 9:10 PM
> To: Liu, Yong <yong.liu@intel.com>; Bie, Tiwei <tiwei.bie@intel.com>
> Cc: Wang, Zhihong <zhihong.wang@intel.com>; dev@dpdk.org
> Subject: Re: [PATCH v3 00/11] support in-order feature
> 
> Hi,
> 
> I was about to apply it, but there are quite a few checkpatch and
> git-check-log.sh issues reported.
> 
> Could you please resent with this fixed?
> Also, I think it would be  better to squash the doc patches in relevant
> ones.
> 
> Thanks,
> Maxime
> On 06/28/2018 11:52 PM, Marvin Liu wrote:
> > In latest virtio-spec, new feature bit VIRTIO_F_IN_ORDER was introduced.
> > When this feature has been negotiated, virtio driver will use
> > descriptors in ring order: starting from offset 0 in the table, and
> > wrapping around at the end of the table. Vhost devices will always use
> > descriptors in the same order in which they have been made available.
> > This can reduce virtio accesses to used ring.
> >
> > Based on updated virtio-spec, this series realized IN_ORDER prototype
> > in virtio driver. Due to new [RT]x path added into selection, also add
> > two new parameters mrg_rx and in_order into virtio-user vdev parameters
> > list. This will allow user to configure feature bits thus can impact
> > [RT]x path selection.
> >
> > Performance of virtio user with IN_ORDER feature:
> >
> >      Platform: Purely
> >      CPU: Intel(R) Xeon(R) Platinum 8160 CPU @ 2.10GHz
> >      DPDK baseline: 18.05
> >      Setup: testpmd with vhost vdev + testpmd with virtio vdev
> >
> >      +--------------+----------+----------+---------+
> >      |Vhost->Virtio |1 Queue   |2 Queues  |4 Queues |
> >      +--------------+----------+----------+---------+
> >      |Inorder       |12.0Mpps  |24.2Mpps  |26.0Mpps |
> >      |Normal        |12.1Mpps  |18.5Mpps  |18.9Mpps |
> >      +--------------+----------+----------+---------+
> >
> >      +--------------+----------+----------------+---------+
> >      |Virtio->Vhost |1 Queue   |2 Queues        |4 Queues |
> >      +--------------+----------+----------------+---------+
> >      |Inorder       |13.8Mpps  |10.7 ~ 15.2Mpps |11.5Mpps |
> >      |Normal        |13.3Mpps  |9.8 ~ 14Mpps    |10.5Mpps |
> >      +--------------+----------+----------------+---------+
> >
> >      +---------+----------+----------------+----------------+
> >      |Loopback |1 Queue   |2 Queues        |4 Queues        |
> >      +---------+----------+----------------+----------------+
> >      |Inorder  |7.4Mpps   |9.1 ~ 11.6Mpps  |10.5 ~ 11.3Mpps |
> >      +---------+----------+----------------+----------------+
> >      |Normal   |7.5Mpps   |7.7 ~ 9.0Mpps   |7.6 ~ 7.8Mpps   |
> >      +---------+----------+----------------+----------------+
> >
> > v3:
> > - refine [RT]x function selection logic
> > - fix in-order mergable packets index error
> > - combine unsupport mask patch
> > - doc virtio in-order update
> > - fix checkpatch error
> >
> > v2:
> > - merge to latest dpdk-net-virtio
> > - not use in_direct for normal xmit packets
> > - update available ring for each descriptor
> > - clean up IN_ORDER xmit function
> > - unmask feature bits when disabled in_order or mgr_rxbuf
> > - extract common part between IN_ORDER and normal functions
> > - update performance result
> >
> > Marvin Liu (11):
> >    vhost: advertise support in-order feature
> >    net/virtio: add in-order feature bit definition
> >    net/virtio-user: add unsupported features mask
> >    net/virtio-user: add mrg_rxbuf and in_order vdev parameters
> >    net/virtio: free in-order descriptors before device start
> >    net/virtio: extract common part for in-order functions
> >    net/virtio: support in-order Rx and Tx
> >    net/virtio: add in-order Rx/Tx into selection
> >    net/virtio: advertise support in-order feature
> >    doc: add mrg_rxbuf and in_order parameters for virtio device
> >    doc: add in-order funciton for virtio device
> >
> >   doc/guides/nics/virtio.rst                    |  27 +-
> >   drivers/net/virtio/virtio_ethdev.c            |  29 +-
> >   drivers/net/virtio/virtio_ethdev.h            |   7 +
> >   drivers/net/virtio/virtio_pci.h               |   8 +
> >   drivers/net/virtio/virtio_rxtx.c              | 635 ++++++++++++++++--
> >   .../net/virtio/virtio_user/virtio_user_dev.c  |  26 +-
> >   .../net/virtio/virtio_user/virtio_user_dev.h  |   4 +-
> >   drivers/net/virtio/virtio_user_ethdev.c       |  45 +-
> >   drivers/net/virtio/virtqueue.c                |   8 +
> >   drivers/net/virtio/virtqueue.h                |   2 +
> >   lib/librte_vhost/socket.c                     |   6 +
> >   lib/librte_vhost/vhost.h                      |  10 +-
> >   12 files changed, 723 insertions(+), 84 deletions(-)
> >