On Wed, Dec 25, 2019 at 10:24:45PM +0800, Jin Yu wrote:
> Add VHOST_USER_PROTOCOL_F_INFLIGHT_SHMFD check in
> getting inflight ring functions.
> 
> Fixes: 4d891f77ddfa ("vhost: add APIs to get inflight ring")
> Cc: sta...@dpdk.org
> 
> Signed-off-by: Jin Yu <jin...@intel.com>
> ---
>  lib/librte_vhost/vhost.c | 8 ++++++++
>  1 file changed, 8 insertions(+)
> 
> diff --git a/lib/librte_vhost/vhost.c b/lib/librte_vhost/vhost.c
> index 1cbe948f7..dfb71732b 100644
> --- a/lib/librte_vhost/vhost.c
> +++ b/lib/librte_vhost/vhost.c
> @@ -862,6 +862,10 @@ rte_vhost_get_vhost_ring_inflight(int vid, uint16_t 
> vring_idx,
>       if (unlikely(!dev))
>               return -1;
>  
> +     if (unlikely(!(dev->protocol_features &
> +         (1ULL << VHOST_USER_PROTOCOL_F_INFLIGHT_SHMFD))))
> +             return 0;

Why this is needed?
The caller needs to know whether _INFLIGHT_SHMFD is negotiated or
not anyway. Even if rte_vhost_get_vhost_ring_inflight() returns
success, the inflight info returned is still invalid.


> +
>       if (vring_idx >= VHOST_MAX_VRING)
>               return -1;
>  
> @@ -1431,6 +1435,10 @@ rte_vhost_get_vring_base_from_inflight(int vid,
>       if (dev == NULL || last_avail_idx == NULL || last_used_idx == NULL)
>               return -1;
>  
> +     if (unlikely(!(dev->protocol_features &
> +         (1ULL << VHOST_USER_PROTOCOL_F_INFLIGHT_SHMFD))))
> +             return 0;
> +
>       if (!vq_is_packed(dev))
>               return -1;
>  
> -- 
> 2.17.2
> 

Reply via email to