[RFC] net/iavf: handle iavf reset gracefully

[RFC] net/iavf: handle iavf reset gracefully

[Shiyang He]

Originally, when vf received PF-to-VF reset event, the iavf PMD did not
perform special actions, resulting in vf being offline and unavailable.
This commit handle the PF-to-VF reset event by performing all necessary
actions to bring the vf back online and available.

Signed-off-by: Shiyang He <shiyangx.he@intel.com>
---
 drivers/net/iavf/iavf.h        |  29 +++++++
 drivers/net/iavf/iavf_ethdev.c | 144 +++++++++++++++++++++++++++++++++
 drivers/net/iavf/iavf_rxtx.c   |  25 ++++++
 drivers/net/iavf/iavf_rxtx.h   |   1 +
 drivers/net/iavf/iavf_vchnl.c  |  12 +++
 5 files changed, 211 insertions(+)

diff --git a/drivers/net/iavf/iavf.h b/drivers/net/iavf/iavf.h
index aa18650ffa..79030944f7 100644
--- a/drivers/net/iavf/iavf.h
+++ b/drivers/net/iavf/iavf.h
@@ -95,6 +95,24 @@
 
 #define IAVF_L2TPV2_FLAGS_LEN	0x4000
 
+#ifndef DIV_ROUND_UP
+#define DIV_ROUND_UP(n, d) (	\
+{								\
+	const typeof(d) __d = d;	\
+	(((n) + (__d) - 1) / (__d));\
+}								\
+)
+#endif
+#ifndef DELAY
+#define DELAY(x) rte_delay_us(x)
+#endif
+#ifndef msleep
+#define msleep(x) DELAY(1000 * (x))
+#endif
+#ifndef usleep_range
+#define usleep_range(min, max) msleep(DIV_ROUND_UP(min, 1000))
+#endif
+
 struct iavf_adapter;
 struct iavf_rx_queue;
 struct iavf_tx_queue;
@@ -279,6 +297,8 @@ struct iavf_info {
 
 	uint32_t ptp_caps;
 	rte_spinlock_t phc_time_aq_lock;
+
+	bool auto_reset_enabled;
 };
 
 #define IAVF_MAX_PKT_TYPE 1024
@@ -423,6 +443,14 @@ _atomic_set_async_response_cmd(struct iavf_info *vf, enum virtchnl_ops ops)
 
 	return !ret;
 }
+
+static inline bool
+iavf_is_reset(struct iavf_hw *hw)
+{
+	return !(IAVF_READ_REG(hw, IAVF_VF_ARQLEN1) &
+		 IAVF_VF_ARQLEN1_ARQENABLE_MASK);
+}
+
 int iavf_check_api_version(struct iavf_adapter *adapter);
 int iavf_get_vf_resource(struct iavf_adapter *adapter);
 void iavf_dev_event_handler_fini(void);
@@ -498,4 +526,5 @@ int iavf_flow_unsub(struct iavf_adapter *adapter,
 		    struct iavf_fsub_conf *filter);
 int iavf_flow_sub_check(struct iavf_adapter *adapter,
 			struct iavf_fsub_conf *filter);
+int iavf_handle_hw_reset(struct rte_eth_dev *dev);
 #endif /* _IAVF_ETHDEV_H_ */
diff --git a/drivers/net/iavf/iavf_ethdev.c b/drivers/net/iavf/iavf_ethdev.c
index f6d68403ce..f421febcac 100644
--- a/drivers/net/iavf/iavf_ethdev.c
+++ b/drivers/net/iavf/iavf_ethdev.c
@@ -337,6 +337,12 @@ iavf_dev_watchdog_disable(struct iavf_adapter *adapter __rte_unused)
 #endif
 }
 
+static void
+iavf_dev_auto_reset_enable(struct iavf_adapter *adapter)
+{
+	adapter->vf.auto_reset_enabled = true;
+}
+
 static int
 iavf_set_mc_addr_list(struct rte_eth_dev *dev,
 			struct rte_ether_addr *mc_addrs,
@@ -2687,6 +2693,8 @@ iavf_dev_init(struct rte_eth_dev *eth_dev)
 	iavf_dev_watchdog_enable(adapter);
 	adapter->closed = false;
 
+	iavf_dev_auto_reset_enable(adapter);
+
 	return 0;
 
 flow_init_err:
@@ -2814,6 +2822,142 @@ iavf_dev_reset(struct rte_eth_dev *dev)
 	return iavf_dev_init(dev);
 }
 
+static bool
+iavf_is_reset_detected(struct iavf_adapter *adapter)
+{
+	struct iavf_hw *hw = &adapter->hw;
+	int i;
+
+	/* poll until we see the reset actually happen */
+	for (i = 0; i < IAVF_RESET_WAIT_CNT; i++) {
+		if (iavf_is_reset(hw))
+			return true;
+		usleep_range(5000, 10000);
+	}
+
+	return false;
+}
+
+static int
+iavf_uninit_hw(struct rte_eth_dev *dev, struct iavf_hw *hw)
+{
+	struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev);
+	struct rte_intr_handle *intr_handle = dev->intr_handle;
+	struct iavf_info *vf = IAVF_DEV_PRIVATE_TO_VF(dev->data->dev_private);
+	struct iavf_adapter *adapter = dev->data->dev_private;
+
+	iavf_reset_queues(dev);
+
+	/* Disable the interrupt for Rx */
+	rte_intr_efd_disable(intr_handle);
+	/* Rx interrupt vector mapping free */
+	rte_intr_vec_list_free(intr_handle);
+
+	adapter->stopped = 1;
+	dev->data->dev_started = 0;
+
+	adapter->closed = true;
+
+	/* free iAVF security device context all related resources */
+	iavf_security_ctx_destroy(adapter);
+
+	iavf_flow_flush(dev, NULL);
+	iavf_flow_uninit(adapter);
+
+	if (vf->promisc_unicast_enabled || vf->promisc_multicast_enabled)
+		iavf_config_promisc(adapter, false, false);
+
+	iavf_shutdown_adminq(hw);
+	if (vf->vf_res->vf_cap_flags & VIRTCHNL_VF_OFFLOAD_WB_ON_ITR) {
+		/* disable uio intr before callback unregister */
+		rte_intr_disable(intr_handle);
+
+		/* unregister callback func from eal lib */
+		rte_intr_callback_unregister(intr_handle,
+					     iavf_dev_interrupt_handler, dev);
+	}
+	iavf_disable_irq0(hw);
+
+	if (vf->vf_res->vf_cap_flags & VIRTCHNL_VF_OFFLOAD_QOS)
+		iavf_tm_conf_uninit(dev);
+
+	if (vf->vf_res->vf_cap_flags & VIRTCHNL_VF_OFFLOAD_RSS_PF) {
+		if (vf->rss_lut) {
+			rte_free(vf->rss_lut);
+			vf->rss_lut = NULL;
+		}
+		if (vf->rss_key) {
+			rte_free(vf->rss_key);
+			vf->rss_key = NULL;
+		}
+	}
+
+	rte_free(vf->vf_res);
+	vf->vsi_res = NULL;
+	vf->vf_res = NULL;
+
+	rte_free(vf->aq_resp);
+	vf->aq_resp = NULL;
+
+	if (vf->vf_reset && !rte_pci_set_bus_master(pci_dev, true))
+		vf->vf_reset = false;
+
+	/* disable watchdog */
+	iavf_dev_watchdog_disable(adapter);
+
+	return 0;
+}
+
+static int
+iavf_reset_hw(struct rte_eth_dev *dev, struct iavf_hw *hw)
+{
+	iavf_uninit_hw(dev, hw);
+
+	return iavf_dev_init(dev);
+}
+
+/*
+ * Handle hardware reset
+ */
+int
+iavf_handle_hw_reset(struct rte_eth_dev *dev)
+{
+	struct iavf_adapter *adapter = dev->data->dev_private;
+	struct iavf_hw *hw = IAVF_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+	struct iavf_info *vf = IAVF_DEV_PRIVATE_TO_VF(dev->data->dev_private);
+	int ret;
+
+	if (!adapter->vf.auto_reset_enabled)
+		return 0;
+
+	if (!iavf_is_reset_detected(adapter))
+		return 0;
+
+	if (iavf_reset_hw(dev, hw))
+		return -1;
+
+	/* configure the device */
+	ret = iavf_dev_configure(dev);
+	if (ret) {
+		PMD_DRV_LOG(ERR, "Failed to configure dev");
+		return ret;
+	}
+
+	iavf_dev_xstats_reset(dev);
+
+	/* start the device */
+	ret = iavf_dev_start(dev);
+	if (ret) {
+		PMD_DRV_LOG(ERR, "Failed to start dev");
+		return ret;
+	}
+
+	dev->data->dev_started = 1;
+	vf->vf_reset = false;
+
+	return 0;
+}
+
 static int
 iavf_dcf_cap_check_handler(__rte_unused const char *key,
 			   const char *value, __rte_unused void *opaque)
diff --git a/drivers/net/iavf/iavf_rxtx.c b/drivers/net/iavf/iavf_rxtx.c
index b1d0fbceb6..cc58564461 100644
--- a/drivers/net/iavf/iavf_rxtx.c
+++ b/drivers/net/iavf/iavf_rxtx.c
@@ -1093,6 +1093,31 @@ iavf_dev_tx_queue_release(struct rte_eth_dev *dev, uint16_t qid)
 	rte_free(q);
 }
 
+void
+iavf_reset_queues(struct rte_eth_dev *dev)
+{
+	struct iavf_rx_queue *rxq;
+	struct iavf_tx_queue *txq;
+	int i;
+
+	for (i = 0; i < dev->data->nb_tx_queues; i++) {
+		txq = dev->data->tx_queues[i];
+		if (!txq)
+			continue;
+		iavf_txq_release_mbufs_ops[txq->rel_mbufs_type].release_mbufs(txq);
+		reset_tx_queue(txq);
+		dev->data->tx_queue_state[i] = RTE_ETH_QUEUE_STATE_STOPPED;
+	}
+	for (i = 0; i < dev->data->nb_rx_queues; i++) {
+		rxq = dev->data->rx_queues[i];
+		if (!rxq)
+			continue;
+		iavf_rxq_release_mbufs_ops[rxq->rel_mbufs_type].release_mbufs(rxq);
+		reset_rx_queue(rxq);
+		dev->data->rx_queue_state[i] = RTE_ETH_QUEUE_STATE_STOPPED;
+	}
+}
+
 void
 iavf_stop_queues(struct rte_eth_dev *dev)
 {
diff --git a/drivers/net/iavf/iavf_rxtx.h b/drivers/net/iavf/iavf_rxtx.h
index 09e2127db0..622a9dbfef 100644
--- a/drivers/net/iavf/iavf_rxtx.h
+++ b/drivers/net/iavf/iavf_rxtx.h
@@ -640,6 +640,7 @@ int iavf_dev_tx_queue_start(struct rte_eth_dev *dev, uint16_t tx_queue_id);
 int iavf_dev_tx_queue_stop(struct rte_eth_dev *dev, uint16_t tx_queue_id);
 int iavf_dev_tx_done_cleanup(void *txq, uint32_t free_cnt);
 void iavf_dev_tx_queue_release(struct rte_eth_dev *dev, uint16_t qid);
+void iavf_reset_queues(struct rte_eth_dev *dev);
 void iavf_stop_queues(struct rte_eth_dev *dev);
 uint16_t iavf_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
 		       uint16_t nb_pkts);
diff --git a/drivers/net/iavf/iavf_vchnl.c b/drivers/net/iavf/iavf_vchnl.c
index ab721e082b..2a36526b10 100644
--- a/drivers/net/iavf/iavf_vchnl.c
+++ b/drivers/net/iavf/iavf_vchnl.c
@@ -31,6 +31,10 @@
 
 #define MAX_EVENT_PENDING 16
 
+static void iavf_dev_event_post(struct rte_eth_dev *dev,
+		enum rte_eth_event_type event,
+		void *param, size_t param_alloc_size);
+
 struct iavf_event_element {
 	TAILQ_ENTRY(iavf_event_element) next;
 	struct rte_eth_dev *dev;
@@ -80,6 +84,14 @@ iavf_dev_event_handle(void *param __rte_unused)
 		TAILQ_FOREACH_SAFE(pos, &pending, next, save_next) {
 			TAILQ_REMOVE(&pending, pos, next);
 			rte_eth_dev_callback_process(pos->dev, pos->event, pos->param);
+
+			switch (pos->event) {
+			case RTE_ETH_EVENT_INTR_RESET:
+				iavf_handle_hw_reset(pos->dev);
+				break;
+			default:
+				break;
+			}
 			rte_free(pos);
 		}
 	}
-- 
2.37.2