[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <06761c62-93b4-8eaa-370f-f26b7c5306ca@redhat.com>
Date: Thu, 25 Aug 2022 15:03:43 +0800
From: Jason Wang <jasowang@...hat.com>
To: Guo Zhi <qtxuning1999@...u.edu.cn>, eperezma@...hat.com,
sgarzare@...hat.com, mst@...hat.com
Cc: netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
kvm@...r.kernel.org, virtualization@...ts.linux-foundation.org
Subject: Re: [RFC v2 2/7] vhost_test: batch used buffer
在 2022/8/17 21:57, Guo Zhi 写道:
> Only add to used ring when a batch of buffer have all been used. And if
> in order feature negotiated, only add the last used descriptor for a
> batch of buffer.
>
> Signed-off-by: Guo Zhi <qtxuning1999@...u.edu.cn>
> ---
> drivers/vhost/test.c | 8 +++++++-
> 1 file changed, 7 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/vhost/test.c b/drivers/vhost/test.c
> index bc8e7fb1e635..57cdb3a3edf6 100644
> --- a/drivers/vhost/test.c
> +++ b/drivers/vhost/test.c
> @@ -43,6 +43,9 @@ struct vhost_test {
> static void handle_vq(struct vhost_test *n)
> {
> struct vhost_virtqueue *vq = &n->vqs[VHOST_TEST_VQ];
> + struct vring_used_elem *heads = kmalloc(sizeof(*heads)
> + * vq->num, GFP_KERNEL);
Though it's a test device, it would be better to try avoid memory
allocation in the datapath.
And where is is freed?
Thanks
> + int batch_idx = 0;
> unsigned out, in;
> int head;
> size_t len, total_len = 0;
> @@ -84,11 +87,14 @@ static void handle_vq(struct vhost_test *n)
> vq_err(vq, "Unexpected 0 len for TX\n");
> break;
> }
> - vhost_add_used_and_signal(&n->dev, vq, head, 0);
> + heads[batch_idx].id = cpu_to_vhost32(vq, head);
> + heads[batch_idx++].len = cpu_to_vhost32(vq, len);
> total_len += len;
> if (unlikely(vhost_exceeds_weight(vq, 0, total_len)))
> break;
> }
> + if (batch_idx)
> + vhost_add_used_and_signal_n(&n->dev, vq, heads, batch_idx);
>
> mutex_unlock(&vq->mutex);
> }
Powered by blists - more mailing lists