[dpdk-dev] [PATCH v2 14/14] net/avf: enable Rx interrupt support

[Jingjing Wu <jingjing.wu@intel.com>]

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>
---
 MAINTAINERS                          |   6 ++
 doc/guides/nics/features/avf.ini     |  38 +++++++
 doc/guides/nics/features/avf_vec.ini |  38 +++++++
 doc/guides/nics/intel_vf.rst         |  16 ++-
 drivers/net/avf/avf_ethdev.c         | 190 +++++++++++++++++++++++++++++------
 5 files changed, 255 insertions(+), 33 deletions(-)
 create mode 100644 doc/guides/nics/features/avf.ini
 create mode 100644 doc/guides/nics/features/avf_vec.ini

diff --git a/MAINTAINERS b/MAINTAINERS
index f0baeb4..6692f2a 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -427,6 +427,12 @@ M: Xiao Wang <xiao.w.wang@intel.com>
 F: drivers/net/fm10k/
 F: doc/guides/nics/features/fm10k*.ini
 
+Intel avf
+M: Jingjing Wu <jingjing.wu@intel.com>
+M: Wenzhuo Lu <wenzhuo.lu@intel.com>
+F: drivers/net/avf/
+F: doc/guides/nics/features/avf*.ini
+
 Mellanox mlx4
 M: Adrien Mazarguil <adrien.mazarguil@6wind.com>
 F: drivers/net/mlx4/
diff --git a/doc/guides/nics/features/avf.ini b/doc/guides/nics/features/avf.ini
new file mode 100644
index 0000000..88d5873
--- /dev/null
+++ b/doc/guides/nics/features/avf.ini
@@ -0,0 +1,38 @@
+;
+; Supported features of the 'avf' network poll mode driver.
+;
+; Refer to default.ini for the full list of available PMD features.
+;
+[Features]
+Speed capabilities   = Y
+Link status          = Y
+Link status event    = Y
+Rx interrupt         = Y
+Queue start/stop     = Y
+MTU update           = Y
+Jumbo frame          = Y
+Scattered Rx         = Y
+TSO                  = Y
+Promiscuous mode     = Y
+Allmulticast mode    = Y
+Unicast MAC filter   = Y
+Multicast MAC filter = Y
+RSS hash             = Y
+RSS key update       = Y
+RSS reta update      = Y
+VLAN filter          = Y
+Hash filter          = Y
+CRC offload          = Y
+VLAN offload         = Y
+L3 checksum offload  = Y
+L4 checksum offload  = Y
+Packet type parsing  = Y
+Rx descriptor status = Y
+Tx descriptor status = Y
+Basic stats          = Y
+Multiprocess aware   = Y
+BSD nic_uio          = Y
+Linux UIO            = Y
+Linux VFIO           = Y
+x86-32               = Y
+x86-64               = Y
diff --git a/doc/guides/nics/features/avf_vec.ini b/doc/guides/nics/features/avf_vec.ini
new file mode 100644
index 0000000..74656ff
--- /dev/null
+++ b/doc/guides/nics/features/avf_vec.ini
@@ -0,0 +1,38 @@
+;
+; Supported features of the 'avf_vec' network poll mode driver.
+;
+; Refer to default.ini for the full list of available PMD features.
+;
+[Features]
+Speed capabilities   = Y
+Link status          = Y
+Link status event    = Y
+Rx interrupt         = Y
+Queue start/stop     = Y
+MTU update           = Y
+Jumbo frame          = Y
+Scattered Rx         = Y
+TSO                  = Y
+Promiscuous mode     = Y
+Allmulticast mode    = Y
+Unicast MAC filter   = Y
+Multicast MAC filter = Y
+RSS hash             = Y
+RSS key update       = Y
+RSS reta update      = Y
+VLAN filter          = Y
+Hash filter          = Y
+CRC offload          = Y
+VLAN offload         = P
+L3 checksum offload  = P
+L4 checksum offload  = P
+Packet type parsing  = Y
+Rx descriptor status = Y
+Tx descriptor status = Y
+Basic stats          = Y
+Multiprocess aware   = Y
+BSD nic_uio          = Y
+Linux UIO            = Y
+Linux VFIO           = Y
+x86-32               = Y
+x86-64               = Y
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 36ffa1a..7c6a1c4 100644
--- a/drivers/net/avf/avf_ethdev.c
+++ b/drivers/net/avf/avf_ethdev.c
@@ -96,9 +96,14 @@ static int avf_dev_rss_hash_conf_get(struct rte_eth_dev *dev,
 static int avf_dev_mtu_set(struct rte_eth_dev *dev, uint16_t mtu);
 static void avf_dev_set_default_mac_addr(struct rte_eth_dev *dev,
 					 struct ether_addr *mac_addr);
+static int avf_dev_rx_queue_intr_enable(struct rte_eth_dev *dev,
+					uint16_t queue_id);
+static int avf_dev_rx_queue_intr_disable(struct rte_eth_dev *dev,
+					 uint16_t queue_id);
 
 int avf_logtype_init;
 int avf_logtype_driver;
+
 static const struct rte_pci_id pci_id_avf_map[] = {
 	{ RTE_PCI_DEVICE(AVF_INTEL_VENDOR_ID, AVF_DEV_ID_ADAPTIVE_VF) },
 	{ .vendor_id = 0, /* sentinel */ },
@@ -140,6 +145,8 @@ static const struct eth_dev_ops avf_eth_dev_ops = {
 	.rx_descriptor_status       = avf_dev_rx_desc_status,
 	.tx_descriptor_status       = avf_dev_tx_desc_status,
 	.mtu_set                    = avf_dev_mtu_set,
+	.rx_queue_intr_enable       = avf_dev_rx_queue_intr_enable,
+	.rx_queue_intr_disable      = avf_dev_rx_queue_intr_disable,
 };
 
 static int
@@ -300,6 +307,88 @@ avf_init_queues(struct rte_eth_dev *dev)
 	return ret;
 }
 
+static int avf_config_rx_queues_irqs(struct rte_eth_dev *dev,
+					   struct rte_intr_handle *intr_handle)
+{
+	struct avf_adapter *adapter =
+		AVF_DEV_PRIVATE_TO_ADAPTER(dev->data->dev_private);
+	struct avf_info *vf = AVF_DEV_PRIVATE_TO_VF(adapter);
+	struct avf_hw *hw = AVF_DEV_PRIVATE_TO_HW(adapter);
+	uint16_t interval, i;
+	int vec;
+
+	if (dev->data->dev_conf.intr_conf.rxq != 0) {
+		if (rte_intr_efd_enable(intr_handle, dev->data->nb_rx_queues))
+			return -1;
+	}
+
+	if (rte_intr_dp_is_en(intr_handle) && !intr_handle->intr_vec) {
+		intr_handle->intr_vec =
+			rte_zmalloc("intr_vec",
+				    dev->data->nb_rx_queues * sizeof(int), 0);
+		if (!intr_handle->intr_vec) {
+			PMD_DRV_LOG(ERR, "Failed to allocate %d rx intr_vec",
+				    dev->data->nb_rx_queues);
+			return -1;
+		}
+	}
+
+	if (!(vf->vf_res->vf_offload_flags & VIRTCHNL_VF_OFFLOAD_WB_ON_ITR) &&
+	    dev->data->dev_conf.intr_conf.rxq == 0) {
+		/* If no WB_ON_ITR offload flags, need to set interrupt for descriptor
+		 * write back.
+		 */
+		vf->nb_msix = 1;
+		vf->msix_base = AVF_MISC_VEC_ID;
+		for (i = 0; i < dev->data->nb_rx_queues; i++)
+			vf->rxq_map[0] |= 1 << i;
+
+		PMD_DRV_LOG(DEBUG, "vector 0 are mapping to all Rx queues");
+
+		/* set ITR to max */
+		interval = avf_calc_itr_interval(AVF_QUEUE_ITR_INTERVAL_MAX);
+		AVF_WRITE_REG(hw, AVFINT_DYN_CTL01,
+			      AVFINT_DYN_CTL01_INTENA_MASK |
+			      AVFINT_DYN_CTL01_CLEARPBA_MASK |
+			      (AVF_ITR_INDEX_DEFAULT <<
+			       AVFINT_DYN_CTL01_ITR_INDX_SHIFT) |
+			      (interval << AVFINT_DYN_CTL01_INTERVAL_SHIFT));
+		AVF_WRITE_FLUSH(hw);
+	} else if (dev->data->dev_conf.intr_conf.rxq) {
+		if (!rte_intr_allow_others(intr_handle)) {
+			vf->nb_msix = 1;
+			vf->msix_base = AVF_MISC_VEC_ID;
+			for (i = 0; i < dev->data->nb_rx_queues; i++) {
+				vf->rxq_map[0] |= 1 << i;
+				intr_handle->intr_vec[i] = AVF_MISC_VEC_ID;
+			}
+			PMD_DRV_LOG(DEBUG, "vector 0 are mapping to all Rx queues");
+		} else {
+			/* If Rx interrupt is reuquired, and we can use
+			 * multi interrupts, then the vec is from 1
+			 */
+			vf->nb_msix = RTE_MIN(vf->vf_res->max_vectors,
+					      intr_handle->nb_efd);
+			vf->msix_base = AVF_RX_VEC_START;
+			vec = AVF_RX_VEC_START;
+			for (i = 0; i < dev->data->nb_rx_queues; i++) {
+				vf->rxq_map[vec] |= 1 << i;
+				intr_handle->intr_vec[i] = vec++;
+				if (vec >= vf->nb_msix)
+					vec = AVF_RX_VEC_START;
+			}
+			PMD_DRV_LOG(DEBUG, "%u vectors are mapping to %u Rx queues",
+				    vf->nb_msix, dev->data->nb_rx_queues);
+		}
+	}
+
+	if (avf_config_irq_map(adapter)) {
+		PMD_DRV_LOG(ERR, "config interrupt mapping failed");
+		return -1;
+	}
+	return 0;
+}
+
 static int
 avf_start_queues(struct rte_eth_dev *dev)
 {
@@ -339,8 +428,6 @@ avf_dev_start(struct rte_eth_dev *dev)
 	struct avf_hw *hw = AVF_DEV_PRIVATE_TO_HW(dev->data->dev_private);
 	struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev);
 	struct rte_intr_handle *intr_handle = dev->intr_handle;
-	uint16_t interval;
-	int i;
 
 	PMD_INIT_FUNC_TRACE();
 
@@ -350,8 +437,6 @@ avf_dev_start(struct rte_eth_dev *dev)
 	vf->num_queue_pairs = RTE_MAX(dev->data->nb_rx_queues,
 				      dev->data->nb_tx_queues);
 
-	/* TODO: Rx interrupt */
-
 	if (avf_init_queues(dev) != 0) {
 		PMD_DRV_LOG(ERR, "failed to do Queue init");
 		return -1;
@@ -369,29 +454,14 @@ avf_dev_start(struct rte_eth_dev *dev)
 		goto err_queue;
 	}
 
-	if (!(vf->vf_res->vf_offload_flags & VIRTCHNL_VF_OFFLOAD_WB_ON_ITR)) {
-		/* If no WB_ON_ITR offload flags, need to set interrupt for
-		 * descriptor write back.
-		 */
-		vf->nb_msix = 1;
-		vf->msix_base = AVF_MISC_VEC_ID;
-		for (i = 0; i < dev->data->nb_rx_queues; i++)
-			vf->rxq_map[0] |= 1 << i;
-
-		/* set ITR to max */
-		interval = avf_calc_itr_interval(AVF_QUEUE_ITR_INTERVAL_MAX);
-		AVF_WRITE_REG(hw, AVFINT_DYN_CTL01,
-			      AVFINT_DYN_CTL01_INTENA_MASK |
-			      AVFINT_DYN_CTL01_CLEARPBA_MASK |
-			      (AVF_ITR_INDEX_DEFAULT <<
-			       AVFINT_DYN_CTL01_ITR_INDX_SHIFT) |
-			      (interval << AVFINT_DYN_CTL01_INTERVAL_SHIFT));
-		AVF_WRITE_FLUSH(hw);
-
-		if (avf_config_irq_map(adapter)) {
-			PMD_DRV_LOG(ERR, "config interrupt mapping failed");
-			goto err_queue;
-		}
+	if (avf_config_rx_queues_irqs(dev, intr_handle) != 0) {
+		PMD_DRV_LOG(ERR, "configure irq failed");
+		goto err_queue;
+	}
+	/* re-enable intr again, because efd assign may change */
+	if (dev->data->dev_conf.intr_conf.rxq != 0) {
+		rte_intr_disable(intr_handle);
+		rte_intr_enable(intr_handle);
 	}
 
 	/* Set all mac addrs */
@@ -402,7 +472,6 @@ avf_dev_start(struct rte_eth_dev *dev)
 		goto err_mac;
 	}
 
-	/* TODO: enable interrupt for RX interrupt */
 	return 0;
 
 err_mac:
@@ -418,6 +487,8 @@ avf_dev_stop(struct rte_eth_dev *dev)
 	struct avf_adapter *adapter =
 		AVF_DEV_PRIVATE_TO_ADAPTER(dev->data->dev_private);
 	struct avf_hw *hw = AVF_DEV_PRIVATE_TO_HW(dev);
+	struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev);
+	struct rte_intr_handle *intr_handle = dev->intr_handle;
 	int ret, i;
 
 	PMD_INIT_FUNC_TRACE();
@@ -427,9 +498,13 @@ avf_dev_stop(struct rte_eth_dev *dev)
 
 	avf_stop_queues(dev);
 
-	/*TODO: Disable the interrupt for Rx*/
-
-	/* TODO: Rx interrupt vector mapping free */
+	/* Disable the interrupt for Rx */
+	rte_intr_efd_disable(intr_handle);
+	/* Rx interrupt vector mapping free */
+	if (intr_handle->intr_vec) {
+		rte_free(intr_handle->intr_vec);
+		intr_handle->intr_vec = NULL;
+	}
 
 	/* remove all mac addrs */
 	avf_add_del_all_mac_addr(adapter, FALSE);
@@ -928,6 +1003,59 @@ avf_dev_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats)
 }
 
 static int
+avf_dev_rx_queue_intr_enable(struct rte_eth_dev *dev, uint16_t queue_id)
+{
+	struct avf_adapter *adapter =
+		AVF_DEV_PRIVATE_TO_ADAPTER(dev->data->dev_private);
+	struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev);
+	struct avf_hw *hw = AVF_DEV_PRIVATE_TO_HW(adapter);
+	uint16_t msix_intr;
+
+	msix_intr = pci_dev->intr_handle.intr_vec[queue_id];
+	if (msix_intr == AVF_MISC_VEC_ID) {
+		PMD_DRV_LOG(INFO, "MISC is also enabled for control");
+		AVF_WRITE_REG(hw, AVFINT_DYN_CTL01,
+			      AVFINT_DYN_CTL01_INTENA_MASK |
+			      AVFINT_DYN_CTL01_CLEARPBA_MASK |
+			      AVFINT_DYN_CTL01_ITR_INDX_MASK);
+	} else
+		AVF_WRITE_REG(hw,
+			      AVFINT_DYN_CTLN1(msix_intr - AVF_RX_VEC_START),
+			      AVFINT_DYN_CTLN1_INTENA_MASK |
+			      AVFINT_DYN_CTLN1_CLEARPBA_MASK |
+			      AVFINT_DYN_CTLN1_ITR_INDX_MASK);
+
+	AVF_WRITE_FLUSH(hw);
+
+	rte_intr_enable(&pci_dev->intr_handle);
+
+	return 0;
+}
+
+static int
+avf_dev_rx_queue_intr_disable(struct rte_eth_dev *dev, uint16_t queue_id)
+{
+	struct avf_adapter *adapter =
+		AVF_DEV_PRIVATE_TO_ADAPTER(dev->data->dev_private);
+	struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev);
+	struct avf_hw *hw = AVF_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+	uint16_t msix_intr;
+
+	msix_intr = pci_dev->intr_handle.intr_vec[queue_id];
+	if (msix_intr == AVF_MISC_VEC_ID) {
+		PMD_DRV_LOG(ERR, "MISC is used for control, cannot disable it");
+		return -EIO;
+	}
+
+	AVF_WRITE_REG(hw,
+		      AVFINT_DYN_CTLN1(msix_intr - AVF_RX_VEC_START),
+		      0);
+
+	AVF_WRITE_FLUSH(hw);
+	return 0;
+}
+
+static int
 avf_check_vf_reset_done(struct avf_hw *hw)
 {
 	int i, reset;
-- 
2.4.11