From: David Marchand <david.marchand@redhat.com>
To: Maxime Coquelin <maxime.coquelin@redhat.com>
Cc: dev@dpdk.org, chenbo.xia@intel.com
Subject: Re: [PATCH 1/2] vhost: fix vduse features negotiation
Date: Wed, 5 Jul 2023 19:15:52 +0200 [thread overview]
Message-ID: <CAJFAV8y3VuPbs3otN3y-0bp7L+qfenoGy+MTtxGg3yip3Ms8Vg@mail.gmail.com> (raw)
In-Reply-To: <3cc8e20a-70d7-871c-c799-5adf37306167@redhat.com>
On Wed, Jul 5, 2023 at 7:02 PM Maxime Coquelin
<maxime.coquelin@redhat.com> wrote:
>
> On 7/5/23 15:36, David Marchand wrote:
> > On Wed, Jul 5, 2023 at 3:22 PM Maxime Coquelin
> > <maxime.coquelin@redhat.com> wrote:
> >> @@ -950,9 +954,14 @@ rte_vhost_driver_register(const char *path, uint64_t flags)
> >> * two values.
> >> */
> >> vsocket->use_builtin_virtio_net = true;
> >> - vsocket->supported_features = VIRTIO_NET_SUPPORTED_FEATURES;
> >> - vsocket->features = VIRTIO_NET_SUPPORTED_FEATURES;
> >> - vsocket->protocol_features = VHOST_USER_PROTOCOL_FEATURES;
> >> + if (vsocket->is_vduse) {
> >> + vsocket->supported_features = VDUSE_NET_SUPPORTED_FEATURES;
> >> + vsocket->features = VDUSE_NET_SUPPORTED_FEATURES;
> >> + } else {
> >> + vsocket->supported_features = VIRTIO_NET_SUPPORTED_FEATURES;
> >> + vsocket->features = VIRTIO_NET_SUPPORTED_FEATURES;
> >> + vsocket->protocol_features = VHOST_USER_PROTOCOL_FEATURES;
> >> + }
> >>
> >
> > Would it make sense to split those features in a set of features
> > shared by vhost-user and vduse, and one dedicated set for each of
> > them?
> > For example, the VIRTIO_NET_F_GUEST/HOST_* features are not bound to
> > vhost-user / vduse, are they?
>
> Most of the features are the same, but there are a few differences:
> - VIRTIO_F_RING_PACKED: this is not really Vhost or VDUSE specific. The
> issue is we just lack packed ring support in the control queue
> implementation, only used by VDUSE for now. I expect to add it in
> v23.11.
>
> - VHOST_USER_F_PROTOCOL_FEATURES: This feature is Vhost-user specific
>
> - VHOST_F_LOG_ALL: This is for live-migration, maybe it will be
> advertised by VDUSE in the future, but we miss live-migration support in
> Kernel VDUSE at the moment.
>
> - VIRTIO_NET_F_CTRL_RX: This is advertised by Vhost-user, but it does
> not do anything specific. While if we do it in VDUSE, we need to
> implement its support in CVQ.
>
> To summarize, we have some features specific to Vhost-user, but for the
> other differences, this is just some features are not yet
> implemented/tested with VDUSE.
>
> Maybe we could have something like this, which has the advantage to
> highlight the gaps we have in VDUSE (not compiled/tested):
LGTM on the principle.
Thanks Maxime.
--
David Marchand
next prev parent reply other threads:[~2023-07-05 17:16 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-07-05 13:22 [PATCH 0/2] VDUSE fixes for v23.07 Maxime Coquelin
2023-07-05 13:22 ` [PATCH 1/2] vhost: fix vduse features negotiation Maxime Coquelin
2023-07-05 13:36 ` David Marchand
2023-07-05 17:02 ` Maxime Coquelin
2023-07-05 17:15 ` David Marchand [this message]
2023-07-05 13:22 ` [PATCH 2/2] vduse: fix missing event index features Maxime Coquelin
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=CAJFAV8y3VuPbs3otN3y-0bp7L+qfenoGy+MTtxGg3yip3Ms8Vg@mail.gmail.com \
--to=david.marchand@redhat.com \
--cc=chenbo.xia@intel.com \
--cc=dev@dpdk.org \
--cc=maxime.coquelin@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.
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 NNTP newsgroup(s).