Date:   Tue, 16 Nov 2021 15:33:54 +0100
From:   Stefano Garzarella <sgarzare@redhat.com>
To:     Andrey Ryabinin <arbn@yandex-team.com>
Cc:     "Michael S. Tsirkin" <mst@redhat.com>,
        Jason Wang <jasowang@redhat.com>,
        Stefan Hajnoczi <stefanha@redhat.com>, kvm@vger.kernel.org,
        virtualization@lists.linux-foundation.org, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH 2/6] vhost_net: get rid of vhost_net_flush_vq() and extra
 flush calls
Message-ID: <20211116143354.exgqcjfbmmbaahs4@steredhat>
References: <20211115153003.9140-1-arbn@yandex-team.com>
 <20211115153003.9140-2-arbn@yandex-team.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Disposition: inline
In-Reply-To: <20211115153003.9140-2-arbn@yandex-team.com>
Precedence: bulk

On Mon, Nov 15, 2021 at 06:29:59PM +0300, Andrey Ryabinin wrote:
>vhost_net_flush_vq() calls vhost_work_dev_flush() twice passing
>vhost_dev pointer obtained via 'n->poll[index].dev' and
>'n->vqs[index].vq.poll.dev'. This is actually the same pointer,
>initialized in vhost_net_open()/vhost_dev_init()/vhost_poll_init()
>
>Remove vhost_net_flush_vq() and call vhost_work_dev_flush() directly.
>Do the flushes only once instead of several flush calls in a row
>which seems rather useless.
>
>Signed-off-by: Andrey Ryabinin <arbn@yandex-team.com>
>---
> drivers/vhost/net.c   | 11 ++---------
> drivers/vhost/vhost.h |  1 +
> 2 files changed, 3 insertions(+), 9 deletions(-)
>
>diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c
>index 11221f6d11b8..b1feb5e0571e 100644
>--- a/drivers/vhost/net.c
>+++ b/drivers/vhost/net.c
>@@ -1373,16 +1373,9 @@ static void vhost_net_stop(struct vhost_net *n, struct socket **tx_sock,
> 	*rx_sock = vhost_net_stop_vq(n, &n->vqs[VHOST_NET_VQ_RX].vq);
> }
>
>-static void vhost_net_flush_vq(struct vhost_net *n, int index)
>-{
>-	vhost_work_dev_flush(n->poll[index].dev);
>-	vhost_work_dev_flush(n->vqs[index].vq.poll.dev);
>-}
>-
> static void vhost_net_flush(struct vhost_net *n)
> {
>-	vhost_net_flush_vq(n, VHOST_NET_VQ_TX);
>-	vhost_net_flush_vq(n, VHOST_NET_VQ_RX);
>+	vhost_work_dev_flush(&n->dev);
> 	if (n->vqs[VHOST_NET_VQ_TX].ubufs) {
> 		mutex_lock(&n->vqs[VHOST_NET_VQ_TX].vq.mutex);
> 		n->tx_flush = true;
>@@ -1572,7 +1565,7 @@ static long vhost_net_set_backend(struct vhost_net *n, unsigned index, int fd)
> 	}
>
> 	if (oldsock) {
>-		vhost_net_flush_vq(n, index);
>+		vhost_work_dev_flush(&n->dev);
> 		sockfd_put(oldsock);
> 	}
>
>diff --git a/drivers/vhost/vhost.h b/drivers/vhost/vhost.h
>index 638bb640d6b4..ecbaa5c6005f 100644
>--- a/drivers/vhost/vhost.h
>+++ b/drivers/vhost/vhost.h
>@@ -15,6 +15,7 @@
> #include <linux/vhost_iotlb.h>
> #include <linux/irqbypass.h>
>
>+struct vhost_dev;

Is this change needed?

Stefano