From: Laurent Vivier <lvivier@redhat.com>
To: Stefano Brivio <sbrivio@redhat.com>, passt-dev@passt.top
Cc: David Gibson <david@gibson.dropbear.id.au>
Subject: Re: [RFC PATCH 1/3] vu_common: Stick to size of input buffer in vu_send_single()
Date: Tue, 4 Nov 2025 16:01:07 +0100 [thread overview]
Message-ID: <7e3386c8-e0cc-4dc2-9be1-c938ef757896@redhat.com> (raw)
In-Reply-To: <20251103101629.1412331-2-sbrivio@redhat.com>
On 11/3/25 11:16, Stefano Brivio wrote:
> ...instead of copying, from the input buffer, the amount of available
> bytes in the buffer provided via vhost-user.
>
> The existing behaviour should be harmless due to the fact that we
> don't request overly large buffers from vhost-user, but it doesn't
> look correct to me.
I don't understand where is the problem.
I think the existing behaviour is correct because the iov array is padded to size in
vu_collect() by:
if (iov->iov_len > size - current_size)
iov->iov_len = size - current_size;
total can be lesser than size but in this case we exit.
So total should be equal to input size + sizeof(struct virtio_net_hdr_mrg_rxbuf).
And in iov_from_buf() we copy the content of the buffer according the available room in
the iovec and the size of the input buffer.
Thanks,
Laurent>
> Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
> ---
> vu_common.c | 22 +++++++++++-----------
> 1 file changed, 11 insertions(+), 11 deletions(-)
>
> diff --git a/vu_common.c b/vu_common.c
> index b13b7c3..21cfb2a 100644
> --- a/vu_common.c
> +++ b/vu_common.c
> @@ -238,7 +238,7 @@ void vu_kick_cb(struct vu_dev *vdev, union epoll_ref ref,
> * vu_send_single() - Send a buffer to the front-end using the RX virtqueue
> * @c: execution context
> * @buf: address of the buffer
> - * @size: size of the buffer
> + * @size: size of the input buffer
> *
> * Return: number of bytes sent, -1 if there is an error
> */
> @@ -248,7 +248,7 @@ int vu_send_single(const struct ctx *c, const void *buf, size_t size)
> struct vu_virtq *vq = &vdev->vq[VHOST_USER_RX_QUEUE];
> struct vu_virtq_element elem[VIRTQUEUE_MAX_SIZE];
> struct iovec in_sg[VIRTQUEUE_MAX_SIZE];
> - size_t total;
> + size_t vu_buf_size;
> int elem_cnt;
> int i;
>
> @@ -261,21 +261,20 @@ int vu_send_single(const struct ctx *c, const void *buf, size_t size)
>
> vu_init_elem(elem, in_sg, VIRTQUEUE_MAX_SIZE);
>
> - size += sizeof(struct virtio_net_hdr_mrg_rxbuf);
> - elem_cnt = vu_collect(vdev, vq, elem, VIRTQUEUE_MAX_SIZE, size, &total);
> - if (total < size) {
> + elem_cnt = vu_collect(vdev, vq, elem, VIRTQUEUE_MAX_SIZE,
> + size + sizeof(struct virtio_net_hdr_mrg_rxbuf),
> + &vu_buf_size);
> + if (vu_buf_size < size + sizeof(struct virtio_net_hdr_mrg_rxbuf)) {
> debug("vu_send_single: no space to send the data "
> - "elem_cnt %d size %zd", elem_cnt, total);
> + "elem_cnt %d size %zd", elem_cnt, size);
> goto err;
> }
>
> vu_set_vnethdr(vdev, in_sg[0].iov_base, elem_cnt);
>
> - total -= sizeof(struct virtio_net_hdr_mrg_rxbuf);
> -
> /* copy data from the buffer to the iovec */
> iov_from_buf(in_sg, elem_cnt, sizeof(struct virtio_net_hdr_mrg_rxbuf),
> - buf, total);
> + buf, size);
>
> if (*c->pcap) {
> pcap_iov(in_sg, elem_cnt,
> @@ -284,9 +283,10 @@ int vu_send_single(const struct ctx *c, const void *buf, size_t size)
>
> vu_flush(vdev, vq, elem, elem_cnt);
>
> - trace("vhost-user sent %zu", total);
> + trace("vhost-user sent %zu",
> + vu_buf_size - sizeof(struct virtio_net_hdr_mrg_rxbuf));
>
> - return total;
> + return size;
> err:
> for (i = 0; i < elem_cnt; i++)
> vu_queue_detach_element(vq);
next prev parent reply other threads:[~2025-11-04 15:01 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-11-03 10:16 [RFC PATCH 0/3] Fill exceeding size of vhost-user buffers explicitly Stefano Brivio
2025-11-03 10:16 ` [RFC PATCH 1/3] vu_common: Stick to size of input buffer in vu_send_single() Stefano Brivio
2025-11-04 15:01 ` Laurent Vivier [this message]
2025-11-04 16:09 ` Stefano Brivio
2025-11-04 16:24 ` Laurent Vivier
2025-11-03 10:16 ` [RFC PATCH 2/3] iov: Fix coding style of basic (non-IOV_TAIL) parts Stefano Brivio
2025-11-04 15:14 ` Laurent Vivier
2025-11-05 3:58 ` David Gibson
2025-11-03 10:16 ` [RFC PATCH 3/3] iov, vu_common: Make iov_from_buf() fill destination iov entirely Stefano Brivio
2025-11-04 15:11 ` Laurent Vivier
2025-11-05 4:01 ` David Gibson
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=7e3386c8-e0cc-4dc2-9be1-c938ef757896@redhat.com \
--to=lvivier@redhat.com \
--cc=david@gibson.dropbear.id.au \
--cc=passt-dev@passt.top \
--cc=sbrivio@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://passt.top/passt
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for IMAP folder(s).