[dpdk-dev] [PATCH v2 1/2] net/i40e: fix VFIO interrupt mapping in VF

[Wei Dai <wei.dai@intel.com>]

When a VF port is bound to VFIO-PIC, only miscellaneous interrupt
is mapped to VFIO vector 0 in i40evf_dev_init( ).
In i40evf_dev_interrupt_handle( ) and i40evf_dev_rx_queue_intr_enable( ),
if previous VFIO interrupt mapping set in i40evf_dev_init( ) is not
cleared, it will fail when PMD tries to map Rx queue interrupt to other
VFIO vectors by calling rte_intr_enable( ).
This patch clears the VFIO interrupt mappings before setting both
miscellaneous and Rx queue interrupt mappings again to avoid failure.
And remove the calling of rte_intr_enable( ) in
i40evf_dev_interrupt_handle( ) and i40evf_dev_rx_queue_intr_enable( )
as VFIO interrupt mapping can only be set in i40e_dev_start( ) and it
is not necessary to be set in above 2 functions repeatedly.

Fixes: 4b90a3ff26c5 ("i40evf: support Rx interrupt")
Cc: stable@dpdk.org

Signed-off-by: Wei Dai <wei.dai@intel.com>
---
 drivers/net/i40e/i40e_ethdev_vf.c | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)

diff --git a/drivers/net/i40e/i40e_ethdev_vf.c b/drivers/net/i40e/i40e_ethdev_vf.c
index 3b76c9e..46dac86 100644
--- a/drivers/net/i40e/i40e_ethdev_vf.c
+++ b/drivers/net/i40e/i40e_ethdev_vf.c
@@ -1433,7 +1433,6 @@ i40evf_dev_interrupt_handler(void *param)
 
 done:
 	i40evf_enable_irq0(hw);
-	rte_intr_enable(dev->intr_handle);
 }
 
 static int
@@ -1888,8 +1887,6 @@ i40evf_dev_rx_queue_intr_enable(struct rte_eth_dev *dev, uint16_t queue_id)
 
 	I40EVF_WRITE_FLUSH(hw);
 
-	rte_intr_enable(&pci_dev->intr_handle);
-
 	return 0;
 }
 
@@ -2041,7 +2038,20 @@ i40evf_dev_start(struct rte_eth_dev *dev)
 		goto err_mac;
 	}
 
+	/* When a VF port is bound to VFIO-PCI, only miscellaneous interrupt
+	 * is mapped to VFIO vector 0 in i40evf_dev_init( ).
+	 * If previous VFIO interrupt mapping set in i40evf_dev_init( ) is
+	 * not cleared, it will fail when rte_intr_enable( ) tries to map Rx
+	 * queue interrupt to other VFIO vectors.
+	 * So clear uio/vfio intr/evevnfd first to avoid failure.
+	 */
+	if (dev->data->dev_conf.intr_conf.rxq != 0) {
+		rte_intr_disable(intr_handle);
+		rte_intr_enable(intr_handle);
+	}
+
 	i40evf_enable_queues_intr(dev);
+
 	return 0;
 
 err_mac:
-- 
2.7.4