diff mbox

[dpdk-dev,RFC,8/9] net/avf: enable ops to check queue info and status

Message ID 1508488012-82704-9-git-send-email-jingjing.wu@intel.com (mailing list archive)
State Changes Requested, archived
Delegated to: Ferruh Yigit
Headers show

Checks

Context Check Description
ci/checkpatch warning coding style issues
ci/Intel-compilation fail Compilation issues

Commit Message

Jingjing Wu Oct. 20, 2017, 8:26 a.m. UTC
- rxq_info_get
 - txq_info_get
 - rx_queue_count
 - rx_descriptor_done
 - rx_descriptor_status
 - tx_descriptor_status

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>
---
 doc/guides/nics/intel_vf.rst |  16 ++++-
 drivers/net/avf/avf_ethdev.c |   8 +++
 drivers/net/avf/avf_rxtx.c   | 145 +++++++++++++++++++++++++++++++++++++++++++
 drivers/net/avf/avf_rxtx.h   |   8 +++
 4 files changed, 175 insertions(+), 2 deletions(-)

Comments

Ferruh Yigit Nov. 22, 2017, 12:09 a.m. UTC | #1
On 10/20/2017 1:26 AM, Jingjing Wu wrote:
>  - rxq_info_get
>  - txq_info_get
>  - rx_queue_count
>  - rx_descriptor_done
>  - rx_descriptor_status
>  - tx_descriptor_status

+ some documentation.

<...>

> @@ -28,8 +28,8 @@
>      (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
>      OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
>  
> -I40E/IXGBE/IGB Virtual Function Driver

AVF doesn't cover ixgbe/igb virtual functions right? Perhaps it can be good to
keep information for them.

<...>

> @@ -132,7 +132,15 @@ static const struct eth_dev_ops avf_eth_dev_ops = {
>  	.reta_query                 = avf_dev_rss_reta_query,
>  	.rss_hash_update            = avf_dev_rss_hash_update,
>  	.rss_hash_conf_get          = avf_dev_rss_hash_conf_get,
> +	.rxq_info_get               = avf_dev_rxq_info_get,
> +	.txq_info_get               = avf_dev_txq_info_get,
> +	.rx_queue_count             = avf_dev_rxq_count,
> +	.rx_descriptor_done         = avf_dev_rx_desc_done,

If you implemented "rx_descriptor_status" no need to implement
"rx_descriptor_done", one covers other.

cc'ed Olivier if I am missing anything.

<...>
Jingjing Wu Nov. 22, 2017, 8:23 a.m. UTC | #2
> -----Original Message-----

> From: Yigit, Ferruh

> Sent: Wednesday, November 22, 2017 8:09 AM

> To: Wu, Jingjing <jingjing.wu@intel.com>; dev@dpdk.org

> Cc: Lu, Wenzhuo <wenzhuo.lu@intel.com>; Olivier MATZ

> <olivier.matz@6wind.com>

> Subject: Re: [dpdk-dev] [RFC 8/9] net/avf: enable ops to check queue info and

> status

> 

> On 10/20/2017 1:26 AM, Jingjing Wu wrote:

> >  - rxq_info_get

> >  - txq_info_get

> >  - rx_queue_count

> >  - rx_descriptor_done

> >  - rx_descriptor_status

> >  - tx_descriptor_status

> 

> + some documentation.

> 

> <...>

> 

> > @@ -28,8 +28,8 @@

> >      (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF

> THE USE

> >      OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH

> DAMAGE.

> >

> > -I40E/IXGBE/IGB Virtual Function Driver

> 

> AVF doesn't cover ixgbe/igb virtual functions right? Perhaps it can be good to

> keep information for them.

> 

Just rename "I40E/IXGBE/IGB Virtual Function Driver" to be " Intel Virtual Function Driver" which doesn't indicate AVF.

And i40e, ixgbe, igb NICs are described below in this doc.

> <...>

> 

> > @@ -132,7 +132,15 @@ static const struct eth_dev_ops avf_eth_dev_ops = {

> >  	.reta_query                 = avf_dev_rss_reta_query,

> >  	.rss_hash_update            = avf_dev_rss_hash_update,

> >  	.rss_hash_conf_get          = avf_dev_rss_hash_conf_get,

> > +	.rxq_info_get               = avf_dev_rxq_info_get,

> > +	.txq_info_get               = avf_dev_txq_info_get,

> > +	.rx_queue_count             = avf_dev_rxq_count,

> > +	.rx_descriptor_done         = avf_dev_rx_desc_done,

> 

> If you implemented "rx_descriptor_status" no need to implement

> "rx_descriptor_done", one covers other.

> 


Thanks, will change.

> cc'ed Olivier if I am missing anything.

> 

> <...>
diff mbox

Patch

diff --git a/doc/guides/nics/intel_vf.rst b/doc/guides/nics/intel_vf.rst
index 1e83bf6..3adb684 100644
--- a/doc/guides/nics/intel_vf.rst
+++ b/doc/guides/nics/intel_vf.rst
@@ -28,8 +28,8 @@ 
     (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
     OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
-I40E/IXGBE/IGB Virtual Function Driver
-======================================
+Intel Virtual Function Driver
+=============================
 
 Supported Intel® Ethernet Controllers (see the *DPDK Release Notes* for details)
 support the following modes of operation in a virtualized environment:
@@ -93,6 +93,18 @@  and the Physical Function operates on the global resources on behalf of the Virt
 For this out-of-band communication, an SR-IOV enabled NIC provides a memory buffer for each Virtual Function,
 which is called a "Mailbox".
 
+Intel® Ethernet Adaptive Virtual Function
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+Adaptive Virtual Function (AVF) is a SR-IOV Virtual Function with the same device id (8086:1889) on different Intel Ethernet Controller.
+AVF Driver is VF driver which supports for all future Intel devices without requiring a VM update. And since this happens to be an adaptive VF driver,
+every new drop of the VF driver would add more and more advanced features that can be turned on in the VM if the underlying HW device supports those
+advanced features based on a device agnostic way without ever compromising on the base functionality. AVF provides generic hardware interface and
+interface between AVF driver and a compliant PF driver is specified.
+
+Intel products starting Ethernet Controller 710 Series to support Adaptive Virtual Function.
+
+The way to generate Virtual Function is like normal, and the resource of VF assignment depends on the NIC Infrastructure.
+
 The PCIE host-interface of Intel Ethernet Switch FM10000 Series VF infrastructure
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
diff --git a/drivers/net/avf/avf_ethdev.c b/drivers/net/avf/avf_ethdev.c
index d3946d6..550bead 100644
--- a/drivers/net/avf/avf_ethdev.c
+++ b/drivers/net/avf/avf_ethdev.c
@@ -132,7 +132,15 @@  static const struct eth_dev_ops avf_eth_dev_ops = {
 	.reta_query                 = avf_dev_rss_reta_query,
 	.rss_hash_update            = avf_dev_rss_hash_update,
 	.rss_hash_conf_get          = avf_dev_rss_hash_conf_get,
+	.rxq_info_get               = avf_dev_rxq_info_get,
+	.txq_info_get               = avf_dev_txq_info_get,
+	.rx_queue_count             = avf_dev_rxq_count,
+	.rx_descriptor_done         = avf_dev_rx_desc_done,
+	.rx_descriptor_status       = avf_dev_rx_desc_status,
+	.tx_descriptor_status       = avf_dev_tx_desc_status,
 	.mtu_set                    = avf_dev_mtu_set,
+	/* TODO: Get statistics/xstatistics */
+	/* TODO: Rx interrupt */
 };
 
 static int
diff --git a/drivers/net/avf/avf_rxtx.c b/drivers/net/avf/avf_rxtx.c
index 95992fc..b3fe550 100644
--- a/drivers/net/avf/avf_rxtx.c
+++ b/drivers/net/avf/avf_rxtx.c
@@ -1384,4 +1384,149 @@  avf_set_tx_function(struct rte_eth_dev *dev)
 {
 	dev->tx_pkt_burst = avf_xmit_pkts;
 	dev->tx_pkt_prepare = avf_prep_pkts;
+}
+void
+avf_dev_rxq_info_get(struct rte_eth_dev *dev, uint16_t queue_id,
+		 struct rte_eth_rxq_info *qinfo)
+{
+	struct avf_rx_queue *rxq;
+
+	rxq = dev->data->rx_queues[queue_id];
+
+	qinfo->mp = rxq->mp;
+	qinfo->scattered_rx = dev->data->scattered_rx;
+	qinfo->nb_desc = rxq->nb_rx_desc;
+
+	qinfo->conf.rx_free_thresh = rxq->rx_free_thresh;
+	qinfo->conf.rx_drop_en = TRUE;
+	qinfo->conf.rx_deferred_start = rxq->rx_deferred_start;
+}
+
+void
+avf_dev_txq_info_get(struct rte_eth_dev *dev, uint16_t queue_id,
+		 struct rte_eth_txq_info *qinfo)
+{
+	struct avf_tx_queue *txq;
+
+	txq = dev->data->tx_queues[queue_id];
+
+	qinfo->nb_desc = txq->nb_tx_desc;
+
+	qinfo->conf.tx_free_thresh = txq->free_thresh;
+	qinfo->conf.tx_rs_thresh = txq->rs_thresh;
+	qinfo->conf.txq_flags = txq->txq_flags;
+	qinfo->conf.tx_deferred_start = txq->tx_deferred_start;
+}
+
+/* Get the number of used descriptors of a rx queue */
+uint32_t
+avf_dev_rxq_count(struct rte_eth_dev *dev, uint16_t queue_id)
+{
+#define AVF_RXQ_SCAN_INTERVAL 4
+	volatile union avf_rx_desc *rxdp;
+	struct avf_rx_queue *rxq;
+	uint16_t desc = 0;
+
+	rxq = dev->data->rx_queues[queue_id];
+	rxdp = &(rxq->rx_ring[rxq->rx_tail]);
+	while ((desc < rxq->nb_rx_desc) &&
+		((rte_le_to_cpu_64(rxdp->wb.qword1.status_error_len) &
+		AVF_RXD_QW1_STATUS_MASK) >> AVF_RXD_QW1_STATUS_SHIFT) &
+				(1 << AVF_RX_DESC_STATUS_DD_SHIFT)) {
+		/* Check the DD bit of a rx descriptor of each 4 in a group,
+		 * to avoid checking too frequently and downgrading performance
+		 * too much.
+		 */
+		desc += AVF_RXQ_SCAN_INTERVAL;
+		rxdp += AVF_RXQ_SCAN_INTERVAL;
+		if (rxq->rx_tail + desc >= rxq->nb_rx_desc)
+			rxdp = &(rxq->rx_ring[rxq->rx_tail +
+					desc - rxq->nb_rx_desc]);
+	}
+
+	return desc;
+}
+
+int
+avf_dev_rx_desc_done(void *rx_queue, uint16_t offset)
+{
+	volatile union avf_rx_desc *rxdp;
+	struct avf_rx_queue *rxq = rx_queue;
+	uint16_t desc;
+	int ret;
+
+	if (unlikely(offset >= rxq->nb_rx_desc)) {
+		PMD_DRV_LOG(ERR, "Invalid RX descriptor id %u", offset);
+		return 0;
+	}
+
+	desc = rxq->rx_tail + offset;
+	if (desc >= rxq->nb_rx_desc)
+		desc -= rxq->nb_rx_desc;
+
+	rxdp = &(rxq->rx_ring[desc]);
+
+	ret = !!(((rte_le_to_cpu_64(rxdp->wb.qword1.status_error_len) &
+		  AVF_RXD_QW1_STATUS_MASK) >> AVF_RXD_QW1_STATUS_SHIFT) &
+		 (1 << AVF_RX_DESC_STATUS_DD_SHIFT));
+
+	return ret;
+}
+
+int
+avf_dev_rx_desc_status(void *rx_queue, uint16_t offset)
+{
+	struct avf_rx_queue *rxq = rx_queue;
+	volatile uint64_t *status;
+	uint64_t mask;
+	uint32_t desc;
+
+	if (unlikely(offset >= rxq->nb_rx_desc))
+		return -EINVAL;
+
+	if (offset >= rxq->nb_rx_desc - rxq->nb_rx_hold)
+		return RTE_ETH_RX_DESC_UNAVAIL;
+
+	desc = rxq->rx_tail + offset;
+	if (desc >= rxq->nb_rx_desc)
+		desc -= rxq->nb_rx_desc;
+
+	status = &rxq->rx_ring[desc].wb.qword1.status_error_len;
+	mask = rte_le_to_cpu_64((1ULL << AVF_RX_DESC_STATUS_DD_SHIFT)
+		<< AVF_RXD_QW1_STATUS_SHIFT);
+	if (*status & mask)
+		return RTE_ETH_RX_DESC_DONE;
+
+	return RTE_ETH_RX_DESC_AVAIL;
+}
+
+int
+avf_dev_tx_desc_status(void *tx_queue, uint16_t offset)
+{
+	struct avf_tx_queue *txq = tx_queue;
+	volatile uint64_t *status;
+	uint64_t mask, expect;
+	uint32_t desc;
+
+	if (unlikely(offset >= txq->nb_tx_desc))
+		return -EINVAL;
+
+	desc = txq->tx_tail + offset;
+	/* go to next desc that has the RS bit */
+	desc = ((desc + txq->rs_thresh - 1) / txq->rs_thresh) *
+		txq->rs_thresh;
+	if (desc >= txq->nb_tx_desc) {
+		desc -= txq->nb_tx_desc;
+		if (desc >= txq->nb_tx_desc)
+			desc -= txq->nb_tx_desc;
+	}
+
+	status = &txq->tx_ring[desc].cmd_type_offset_bsz;
+	mask = rte_le_to_cpu_64(AVF_TXD_QW1_DTYPE_MASK);
+	expect = rte_cpu_to_le_64(
+		 AVF_TX_DESC_DTYPE_DESC_DONE << AVF_TXD_QW1_DTYPE_SHIFT);
+	if ((*status & mask) == expect)
+		return RTE_ETH_TX_DESC_DONE;
+
+	return RTE_ETH_TX_DESC_FULL;
 }
\ No newline at end of file
diff --git a/drivers/net/avf/avf_rxtx.h b/drivers/net/avf/avf_rxtx.h
index de98ce3..c52bd5f 100644
--- a/drivers/net/avf/avf_rxtx.h
+++ b/drivers/net/avf/avf_rxtx.h
@@ -180,6 +180,14 @@  uint16_t avf_prep_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
 		       uint16_t nb_pkts);
 void avf_set_rx_function(struct rte_eth_dev *dev);
 void avf_set_tx_function(struct rte_eth_dev *dev);
+void avf_dev_rxq_info_get(struct rte_eth_dev *dev, uint16_t queue_id,
+			  struct rte_eth_rxq_info *qinfo);
+void avf_dev_txq_info_get(struct rte_eth_dev *dev, uint16_t queue_id,
+			  struct rte_eth_txq_info *qinfo);
+uint32_t avf_dev_rxq_count(struct rte_eth_dev *dev, uint16_t queue_id);
+int avf_dev_rx_desc_done(void *rx_queue, uint16_t offset);
+int avf_dev_rx_desc_status(void *rx_queue, uint16_t offset);
+int avf_dev_tx_desc_status(void *tx_queue, uint16_t offset);
 
 static inline
 void avf_dump_rx_descriptor(struct avf_rx_queue *rxq,