Re: [PATCH v16 12/26] net/nbl: do not use VFIO group bind API

["Burakov, Anatoly" <anatoly.burakov@intel.com>]

On 9/9/2026 12:02 PM, Anatoly Burakov wrote:
> The NBL driver currently uses group bind API, but it is using it only to
> get group fd and nothing else. In context of NBL driver, this is the only
> usage of VFIO API's in the driver, and it is not necessary to use it for
> what NBL driver is trying to accomplish.
> 
> Use a direct `open()` call instead, and store the group fd in common
> structure.
> 
> Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
> ---
>   drivers/net/nbl/nbl_common/nbl_userdev.c  | 24 ++++++++++++++++-------
>   drivers/net/nbl/nbl_include/nbl_include.h |  1 +
>   2 files changed, 18 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/net/nbl/nbl_common/nbl_userdev.c b/drivers/net/nbl/nbl_common/nbl_userdev.c
> index f1522772f5..b3b6548bb8 100644
> --- a/drivers/net/nbl/nbl_common/nbl_userdev.c
> +++ b/drivers/net/nbl/nbl_common/nbl_userdev.c
> @@ -387,6 +387,15 @@ nbl_userdev_mem_event_callback(enum rte_mem_event type, const void *addr, size_t
>   	}
>   }
>   
> +static int
> +nbl_open_group_fd(int iommu_group_num)
> +{
> +	char path[PATH_MAX];
> +
> +	snprintf(path, sizeof(path), DEV_VFIO_GROUP_FMT, iommu_group_num);
> +	return open(path, O_RDWR);
> +}
> +
>   static int nbl_mdev_map_device(struct nbl_adapter *adapter)
>   {
>   	const struct rte_pci_device *pci_dev = adapter->pci_dev;
> @@ -424,11 +433,16 @@ static int nbl_mdev_map_device(struct nbl_adapter *adapter)
>   	}
>   
>   	NBL_LOG(DEBUG, "nbl vfio container %d", container);
> -	vfio_group_fd = dev_vfio_container_group_bind(container, common->iommu_group_num);
> +	/*
> +	 * This assumes one device per group, as kernel will return -EBUSY on
> +	 * attempting to open the same group fd for a different device.
> +	 */
> +	vfio_group_fd = nbl_open_group_fd(common->iommu_group_num);
>   	if (vfio_group_fd < 0) {
>   		NBL_LOG(ERR, "nbl vfio group bind failed, %d", vfio_group_fd);
>   		goto free_container;
>   	}
> +	common->groupfd = vfio_group_fd;

This is indeed making an assumption that there can be only one device 
per group, while original implementation didn't make such an assumption. 
I have an implementation that addresses this issue, so a new revision 
will be sent. I'll wait for more feedback before respinning.

-- 
Thanks,
Anatoly