Hi,

Sorry but I'd like to withdraw this patch as the issue could be addressed by 
applications to register a pre_msg_handler.

Thanks,

Patrick

> -----Original Message-----
> From: Fu, Patrick <patrick...@intel.com>
> Sent: Thursday, July 23, 2020 1:38 PM
> To: dev@dpdk.org; maxime.coque...@redhat.com; Xia, Chenbo
> <chenbo....@intel.com>
> Cc: Fu, Patrick <patrick...@intel.com>
> Subject: [PATCH v1] vhost: fix vring disable fail in async transfer
> 
> From: Patrick Fu <patrick...@intel.com>
> 
> Async inflight packets is checked when front end virtio requests to disable
> vring. If inflight packets exists, vring disabling will fail.
> However, there is no good way for application to get notified before vring is
> disabled and properly complete inflight packets. This patch moves the inflight
> packets check from vring disabling time to the vring enabling time, so that
> applications have chances to drain out pending packets.
> 
> Fixes: 78639d54563a ("vhost: introduce async enqueue registration API")
> 
> Signed-off-by: Patrick Fu <patrick...@intel.com>
> ---
>  lib/librte_vhost/vhost_user.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/lib/librte_vhost/vhost_user.c b/lib/librte_vhost/vhost_user.c
> index 9ddeae362..36da1380a 100644
> --- a/lib/librte_vhost/vhost_user.c
> +++ b/lib/librte_vhost/vhost_user.c
> @@ -2038,9 +2038,9 @@ vhost_user_set_vring_enable(struct virtio_net
> **pdev,
>               "set queue enable: %d to qp idx: %d\n",
>               enable, index);
> 
> -     if (!enable && dev->virtqueue[index]->async_registered) {
> +     if (enable && dev->virtqueue[index]->async_registered) {
>               if (dev->virtqueue[index]->async_pkts_inflight_n) {
> -                     VHOST_LOG_CONFIG(ERR, "failed to disable vring. "
> +                     VHOST_LOG_CONFIG(ERR, "failed to enable vring. "
>                       "async inflight packets must be completed first\n");
>                       return RTE_VHOST_MSG_RESULT_ERR;
>               }
> --
> 2.18.4

Reply via email to