All of lore.kernel.org
 help / color / mirror / Atom feed
From: Aaron Conole <aconole@redhat.com>
To: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>,
	Maxime Coquelin <maxime.coquelin@redhat.com>,
	yuanhan.liu@linux.intel.com, pbonzini@redhat.com,
	qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [RFC v2 0/3] virtio-net: Add support to MTU feature
Date: Wed, 23 Nov 2016 09:02:53 -0500	[thread overview]
Message-ID: <f7tlgwal2eq.fsf@redhat.com> (raw)
In-Reply-To: <20161123062418-mutt-send-email-mst@kernel.org> (Michael S. Tsirkin's message of "Wed, 23 Nov 2016 06:26:41 +0200")

"Michael S. Tsirkin" <mst@redhat.com> writes:

> On Wed, Nov 23, 2016 at 11:42:52AM +0800, Jason Wang wrote:
>> > > > > > > >  Seems to me like an easy way to get out of sync.
>> > > > > >
>> > > > > >If we send it to the backend, that has a chance to check
>> > > > > >mtu and disconnect on error.
>> > > >
>> > > >For vhost-user backend, we can send it the MTU value with a
>> > > >vhost-user protocol feature.
>> > > >
>> > > >For tun/macvtap, how do you do without adding a new ioctl ?
>> > Have management configure same mtu on the backend and in qemu.
>> > 
>> > 
>> 
>> Then why not do same for vhost-user (instead of using two different
>> methods)?
>
> That's what I'm saying. If backend supports that, we can also
> check the mtu in some way to make sure it matches.

I'm not sure why we need a new ioctl (or an ioctl at all - netlink
supports all of this)?

ex:

08:58:34 aconole {fast-datapath-beta-rhel-7} ~/rhpkg/openvswitch$ sudo ip tuntap add dev tap0 mode tap
[sudo] password for aconole: 
08:58:40 aconole {fast-datapath-beta-rhel-7} ~/rhpkg/openvswitch$ ip l
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
...
7: tap0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000
    link/ether 46:e0:fc:83:54:1c brd ff:ff:ff:ff:ff:ff
08:58:51 aconole {fast-datapath-beta-rhel-7} ~/rhpkg/openvswitch$ sudo ip l set tap0 mtu 8000
08:58:54 aconole {fast-datapath-beta-rhel-7} ~/rhpkg/openvswitch$ ip l
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
...
7: tap0: <BROADCAST,MULTICAST> mtu 8000 qdisc noop state DOWN mode DEFAULT group default qlen 1000
    link/ether 46:e0:fc:83:54:1c brd ff:ff:ff:ff:ff:ff

So, at least with iproute2, we can already read and write using the netlink
interface for tuntap devices.  I haven't played with macvtap, but I
think it's similar support - just do a netlink query, get the configured
MTU, and advertise it.  I might be missing something though - I'm a
simple guy with simple ideas.  Maybe there's a cross-platform issue or
something?

-Aaron

  reply	other threads:[~2016-11-23 14:03 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-11-17 21:58 [Qemu-devel] [RFC v2 0/3] virtio-net: Add support to MTU feature Maxime Coquelin
2016-11-17 21:58 ` [Qemu-devel] [RFC v2 1/3] vhost-user: Add new protocol feature MTU Maxime Coquelin
2016-11-18 14:26   ` Aaron Conole
2016-11-21 12:50     ` Maxime Coquelin
2016-11-17 21:58 ` [Qemu-devel] [RFC v2 2/3] vhost-net: Add new MTU feature support Maxime Coquelin
2016-11-17 22:39   ` Michael S. Tsirkin
2016-11-21 12:51     ` Maxime Coquelin
2016-11-18 18:13   ` Aaron Conole
2016-11-17 21:58 ` [Qemu-devel] [RFC v2 3/3] virtio-net: Add " Maxime Coquelin
2016-11-17 22:38   ` Michael S. Tsirkin
2016-11-21 12:34     ` Maxime Coquelin
2016-11-21 16:48       ` Michael S. Tsirkin
2016-11-22 12:11         ` Maxime Coquelin
2016-11-22 14:18           ` Michael S. Tsirkin
2016-11-22 15:33             ` Maxime Coquelin
2016-11-17 22:34 ` [Qemu-devel] [RFC v2 0/3] virtio-net: Add support to MTU feature Michael S. Tsirkin
2016-11-18  6:42   ` John Fastabend
2016-11-18 18:15 ` Aaron Conole
2016-11-18 18:52   ` Maxime Coquelin
2016-11-18 19:21     ` Aaron Conole
2016-11-21 16:23       ` Michael S. Tsirkin
2016-11-22  4:07         ` Jason Wang
2016-11-22  7:40           ` Maxime Coquelin
2016-11-22 14:32             ` Aaron Conole
2016-11-22 14:41               ` Michael S. Tsirkin
2016-11-22 17:56                 ` Maxime Coquelin
2016-11-22 20:38                   ` Michael S. Tsirkin
2016-11-23  3:42                     ` Jason Wang
2016-11-23  4:26                       ` Michael S. Tsirkin
2016-11-23 14:02                         ` Aaron Conole [this message]
2016-11-23 17:42                           ` Michael S. Tsirkin

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=f7tlgwal2eq.fsf@redhat.com \
    --to=aconole@redhat.com \
    --cc=jasowang@redhat.com \
    --cc=maxime.coquelin@redhat.com \
    --cc=mst@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=yuanhan.liu@linux.intel.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.