[dpdk-stable] [PATCH] net/e1000: fix cannot getting Rx interrupts issue

[Luca Boccassi]

On Fri, 2017-11-03 at 17:40 +0800, Xiaoyun Li wrote:
> When using VFIO and MSIX interrupt mode, cannot get Rx interrupts.
> Because
> the interrupt vectors that are used by VFIO are not enabled. This
> patch
> fixes this issue.
> 
> Fixes: c3cd3de0ab50 ("igb: enable Rx queue interrupts for PF")
> Cc: stable at dpdk.org
> 
> Signed-off-by: Xiaoyun Li <xiaoyun.li at intel.com>
> ---
>  drivers/net/e1000/igb_ethdev.c | 16 ++++++++++++++--
>  1 file changed, 14 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/e1000/igb_ethdev.c
> b/drivers/net/e1000/igb_ethdev.c
> index 003bdf0..dc70d22 100644
> --- a/drivers/net/e1000/igb_ethdev.c
> +++ b/drivers/net/e1000/igb_ethdev.c
> @@ -5387,7 +5387,14 @@ eth_igb_rx_queue_intr_disable(struct
> rte_eth_dev *dev, uint16_t queue_id)
>  {
>  	struct e1000_hw *hw =
>  		E1000_DEV_PRIVATE_TO_HW(dev->data->dev_private);
> -	uint32_t mask = 1 << queue_id;
> +	struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev);
> +	struct rte_intr_handle *intr_handle = &pci_dev->intr_handle;
> +	uint32_t vec = E1000_MISC_VEC_ID;
> +
> +	if (rte_intr_allow_others(intr_handle))
> +		vec = E1000_RX_VEC_START;
> +
> +	uint32_t mask = 1 << (queue_id + vec);
>  
>  	E1000_WRITE_REG(hw, E1000_EIMC, mask);
>  	E1000_WRITE_FLUSH(hw);
> @@ -5402,7 +5409,12 @@ eth_igb_rx_queue_intr_enable(struct
> rte_eth_dev *dev, uint16_t queue_id)
>  		E1000_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 = &pci_dev->intr_handle;
> -	uint32_t mask = 1 << queue_id;
> +	uint32_t vec = E1000_MISC_VEC_ID;
> +
> +	if (rte_intr_allow_others(intr_handle))
> +		vec = E1000_RX_VEC_START;
> +
> +	uint32_t mask = 1 << (queue_id + vec);
>  	uint32_t regval;
>  
>  	regval = E1000_READ_REG(hw, E1000_EIMS);

Hi,

Has this fix, or a version of this fix, being accepted into dpdk/master
? I did a quick search in the git log but cannot see it.

If it has, it applies to 16.11 with a small change to get the pci_dev
in the "old way", so if you would like it in 16.11.4 and it's accepted
in mainline let me know.

-- 
Kind regards,
Luca Boccassi