From: Stefano Garzarella <sgarzare@redhat.com>
To: Jason Wang <jasowang@redhat.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>,
netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
Stefan Hajnoczi <stefanha@redhat.com>,
"David S. Miller" <davem@davemloft.net>,
virtualization@lists.linux-foundation.org, kvm@vger.kernel.org
Subject: Re: [PATCH v4 4/5] vhost/vsock: split packets to send using multiple buffers
Date: Fri, 19 Jul 2019 11:20:11 +0200 [thread overview]
Message-ID: <20190719092011.czzju7lepn3lv3up@steredhat> (raw)
In-Reply-To: <53da84b9-184f-1377-0582-ab7cf42ebdb6@redhat.com>
On Fri, Jul 19, 2019 at 04:51:00PM +0800, Jason Wang wrote:
>
> On 2019/7/19 下午4:39, Stefano Garzarella wrote:
> > On Fri, Jul 19, 2019 at 04:21:52PM +0800, Jason Wang wrote:
> > > On 2019/7/19 下午4:08, Stefano Garzarella wrote:
> > > > On Thu, Jul 18, 2019 at 07:35:46AM -0400, Michael S. Tsirkin wrote:
> > > > > On Thu, Jul 18, 2019 at 11:37:30AM +0200, Stefano Garzarella wrote:
> > > > > > On Thu, Jul 18, 2019 at 10:13 AM Michael S. Tsirkin<mst@redhat.com> wrote:
> > > > > > > On Thu, Jul 18, 2019 at 09:50:14AM +0200, Stefano Garzarella wrote:
> > > > > > > > On Wed, Jul 17, 2019 at 4:55 PM Michael S. Tsirkin<mst@redhat.com> wrote:
> > > > > > > > > On Wed, Jul 17, 2019 at 01:30:29PM +0200, Stefano Garzarella wrote:
> > > > > > > > > > If the packets to sent to the guest are bigger than the buffer
> > > > > > > > > > available, we can split them, using multiple buffers and fixing
> > > > > > > > > > the length in the packet header.
> > > > > > > > > > This is safe since virtio-vsock supports only stream sockets.
> > > > > > > > > >
> > > > > > > > > > Signed-off-by: Stefano Garzarella<sgarzare@redhat.com>
> > > > > > > > > So how does it work right now? If an app
> > > > > > > > > does sendmsg with a 64K buffer and the other
> > > > > > > > > side publishes 4K buffers - does it just stall?
> > > > > > > > Before this series, the 64K (or bigger) user messages was split in 4K packets
> > > > > > > > (fixed in the code) and queued in an internal list for the TX worker.
> > > > > > > >
> > > > > > > > After this series, we will queue up to 64K packets and then it will be split in
> > > > > > > > the TX worker, depending on the size of the buffers available in the
> > > > > > > > vring. (The idea was to allow EWMA or a configuration of the buffers size, but
> > > > > > > > for now we postponed it)
> > > > > > > Got it. Using workers for xmit is IMHO a bad idea btw.
> > > > > > > Why is it done like this?
> > > > > > Honestly, I don't know the exact reasons for this design, but I suppose
> > > > > > that the idea was to have only one worker that uses the vring, and
> > > > > > multiple user threads that enqueue packets in the list.
> > > > > > This can simplify the code and we can put the user threads to sleep if
> > > > > > we don't have "credit" available (this means that the receiver doesn't
> > > > > > have space to receive the packet).
> > > > > I think you mean the reverse: even without credits you can copy from
> > > > > user and queue up data, then process it without waking up the user
> > > > > thread.
> > > > I checked the code better, but it doesn't seem to do that.
> > > > The .sendmsg callback of af_vsock, check if the transport has space
> > > > (virtio-vsock transport returns the credit available). If there is no
> > > > space, it put the thread to sleep on the 'sk_sleep(sk)' wait_queue.
> > > >
> > > > When the transport receives an update of credit available on the other
> > > > peer, it calls 'sk->sk_write_space(sk)' that wakes up the thread
> > > > sleeping, that will queue the new packet.
> > > >
> > > > So, in the current implementation, the TX worker doesn't check the
> > > > credit available, it only sends the packets.
> > > >
> > > > > Does it help though? It certainly adds up work outside of
> > > > > user thread context which means it's not accounted for
> > > > > correctly.
> > > > I can try to xmit the packet directly in the user thread context, to see
> > > > the improvements.
> > >
> > > It will then looks more like what virtio-net (and other networking device)
> > > did.
> > I'll try ASAP, the changes should not be too complicated... I hope :)
> >
> > >
> > > > > Maybe we want more VQs. Would help improve parallelism. The question
> > > > > would then become how to map sockets to VQs. With a simple hash
> > > > > it's easy to create collisions ...
> > > > Yes, more VQs can help but the map question is not simple to answer.
> > > > Maybe we can do an hash on the (cid, port) or do some kind of estimation
> > > > of queue utilization and try to balance.
> > > > Should the mapping be unique?
> > >
> > > It sounds to me you want some kind of fair queuing? We've already had
> > > several qdiscs that do this.
> > Thanks for pointing it out!
> >
> > > So if we use the kernel networking xmit path, all those issues could be
> > > addressed.
> > One more point to AF_VSOCK + net-stack, but we have to evaluate possible
> > drawbacks in using the net-stack. (e.g. more latency due to the complexity
> > of the net-stack?)
>
>
> Yes, we need benchmark the performance. But as we've noticed, current vsock
> implementation is not efficient, and for stream socket, the overhead should
> be minimal. The most important thing is to avoid reinventing things that has
> already existed.
Got it. I completely agree with you, and I want to avoid reinventing things
(surely in a worse way).
But the idea (suggested also by Micheal) to discover how fast can go a
new protocol separate from the networking stack, is quite attractive :)
Thanks,
Stefano
next prev parent reply other threads:[~2019-07-19 9:20 UTC|newest]
Thread overview: 69+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-07-17 11:30 [PATCH v4 0/5] vsock/virtio: optimizations to increase the throughput Stefano Garzarella
2019-07-17 11:30 ` [PATCH v4 1/5] vsock/virtio: limit the memory used per-socket Stefano Garzarella
2019-07-29 14:04 ` Michael S. Tsirkin
2019-07-29 15:36 ` Stefano Garzarella
2019-07-29 15:49 ` Michael S. Tsirkin
2019-07-29 16:19 ` Stefano Garzarella
2019-07-29 16:50 ` Stefano Garzarella
2019-07-29 19:10 ` Michael S. Tsirkin
2019-07-30 9:35 ` Stefano Garzarella
2019-07-30 20:42 ` Michael S. Tsirkin
2019-08-01 10:47 ` Stefano Garzarella
2019-08-01 13:21 ` Michael S. Tsirkin
2019-08-01 13:36 ` Stefano Garzarella
2019-09-01 8:26 ` Michael S. Tsirkin
2019-09-01 10:17 ` Michael S. Tsirkin
2019-09-02 9:57 ` Stefano Garzarella
2019-09-02 15:23 ` Michael S. Tsirkin
2019-09-03 4:39 ` Michael S. Tsirkin
2019-09-03 7:45 ` Stefano Garzarella
2019-09-03 7:52 ` Michael S. Tsirkin
2019-09-03 8:00 ` Stefano Garzarella
2019-07-29 16:01 ` Michael S. Tsirkin
2019-07-29 16:41 ` Stefano Garzarella
2019-07-29 19:33 ` Michael S. Tsirkin
2019-08-30 9:40 ` Stefano Garzarella
2019-09-01 6:56 ` Michael S. Tsirkin
2019-09-02 8:39 ` Stefan Hajnoczi
2019-09-02 8:55 ` Stefano Garzarella
2019-10-11 13:40 ` Stefano Garzarella
2019-10-11 14:11 ` Michael S. Tsirkin
2019-10-11 14:23 ` Stefano Garzarella
2019-10-14 8:17 ` Stefan Hajnoczi
2019-10-14 8:21 ` Jason Wang
2019-10-14 8:38 ` Stefano Garzarella
2019-07-17 11:30 ` [PATCH v4 2/5] vsock/virtio: reduce credit update messages Stefano Garzarella
2019-07-22 8:36 ` Stefan Hajnoczi
2019-09-03 4:38 ` Michael S. Tsirkin
2019-09-03 7:31 ` Stefano Garzarella
2019-09-03 7:38 ` Michael S. Tsirkin
2019-07-17 11:30 ` [PATCH v4 3/5] vsock/virtio: fix locking in virtio_transport_inc_tx_pkt() Stefano Garzarella
2019-07-17 14:51 ` Michael S. Tsirkin
2019-07-18 7:43 ` Stefano Garzarella
2019-07-22 8:53 ` Stefan Hajnoczi
2019-07-17 11:30 ` [PATCH v4 4/5] vhost/vsock: split packets to send using multiple buffers Stefano Garzarella
2019-07-17 14:54 ` Michael S. Tsirkin
2019-07-18 7:50 ` Stefano Garzarella
2019-07-18 8:13 ` Michael S. Tsirkin
2019-07-18 9:37 ` Stefano Garzarella
2019-07-18 11:35 ` Michael S. Tsirkin
2019-07-19 8:08 ` Stefano Garzarella
2019-07-19 8:21 ` Jason Wang
2019-07-19 8:39 ` Stefano Garzarella
2019-07-19 8:51 ` Jason Wang
2019-07-19 9:20 ` Stefano Garzarella [this message]
2019-07-22 9:06 ` Stefan Hajnoczi
2019-07-17 11:30 ` [PATCH v4 5/5] vsock/virtio: change the maximum packet size allowed Stefano Garzarella
2019-07-17 14:59 ` Michael S. Tsirkin
2019-07-18 7:52 ` Stefano Garzarella
2019-07-18 12:33 ` Michael S. Tsirkin
2019-07-19 8:29 ` Stefano Garzarella
2019-07-22 9:07 ` Stefan Hajnoczi
2019-07-22 9:08 ` [PATCH v4 0/5] vsock/virtio: optimizations to increase the throughput Stefan Hajnoczi
2019-07-22 9:14 ` Stefano Garzarella
2019-07-29 13:12 ` Stefan Hajnoczi
2019-07-29 13:59 ` Michael S. Tsirkin
2019-07-30 9:40 ` Stefano Garzarella
2019-07-30 10:03 ` Jason Wang
2019-07-30 15:38 ` Stefano Garzarella
2019-09-03 8:02 ` request for stable (was Re: [PATCH v4 0/5] vsock/virtio: optimizations to increase the throughput) 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=20190719092011.czzju7lepn3lv3up@steredhat \
--to=sgarzare@redhat.com \
--cc=davem@davemloft.net \
--cc=jasowang@redhat.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mst@redhat.com \
--cc=netdev@vger.kernel.org \
--cc=stefanha@redhat.com \
--cc=virtualization@lists.linux-foundation.org \
/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).