From: Akihiko Odaki <akihiko.odaki@gmail.com>
To: Jason Wang <jasowang@redhat.com>,
Vladislav Yaroshchuk <vladislav.yaroshchuk@jetbrains.com>
Cc: "Peter Maydell" <peter.maydell@linaro.org>,
"Gerd Hoffmann" <kraxel@redhat.com>,
"Alex Bennée" <alex.bennee@linaro.org>,
phillip.ennen@gmail.com,
"qemu Developers" <qemu-devel@nongnu.org>,
"Cameron Esfahani" <dirty@apple.com>,
"Markus Armbruster" <armbru@redhat.com>,
"Roman Bolshakov" <r.bolshakov@yadro.com>,
"Alexander Graf" <agraf@csgraf.de>,
"Phillip Tennen" <phillip@axleos.com>,
"Roman Bolshakov" <roman@roolebo.dev>,
"Howard Spoelstra" <hsp.cat7@gmail.com>,
"Alessio Dionisi" <hello@adns.io>,
"Christian Schoenebeck" <qemu_oss@crudebyte.com>,
"Eric Blake" <eblake@redhat.com>,
"Philippe Mathieu-Daudé" <f4bug@amsat.org>
Subject: Re: [PATCH v15 3/8] net/vmnet: implement shared mode (vmnet-shared)
Date: Tue, 1 Mar 2022 17:21:40 +0900 [thread overview]
Message-ID: <e5302fba-e773-c56d-388f-a3593bacd977@gmail.com> (raw)
In-Reply-To: <CAGmdLqTqcESsiyTMavKZ3U8=5yPtqdq2tunZTH-tfrzwfUMGug@mail.gmail.com>
On 2022/03/01 17:09, Vladislav Yaroshchuk wrote:
> > Not sure that only one field is enough, cause
> > we may have two states on bh execution start:
> > 1. There are packets in vmnet buffer s->packets_buf
> > that were rejected by qemu_send_async and waiting
> > to be sent. If this happens, we should complete sending
> > these waiting packets with qemu_send_async firstly,
> > and after that we should call vmnet_read to get
> > new ones and send them to QEMU;
> > 2. There are no packets in s->packets_buf to be sent to
> > qemu, we only need to get new packets from vmnet
> > with vmnet_read and send them to QEMU
>
> In case 1, you should just keep calling qemu_send_packet_async.
> Actually
> qemu_send_packet_async adds the packet to its internal queue and calls
> the callback when it is consumed.
>
>
> I'm not sure we can keep calling qemu_send_packet_async,
> because as docs from net/queue.c says:
>
> /* [...]
> * If a sent callback is provided to send(), the caller must handle a
> * zero return from the delivery handler by not sending any more packets
> * until we have invoked the callback. Only in that case will we queue
> * the packet.
> *
> * If a sent callback isn't provided, we just drop the packet to avoid
> * unbounded queueing.
> */
>
> So after we did vmnet_read and read N packets
> into temporary s->packets_buf, we begin calling
> qemu_send_packet_async. If it returns 0 - it says
> "no more packets until sent_cb called please".
> At this moment we have N packets in s->packets_buf
> and already queued K < N of them. But, packets K..N
> are not queued and keep waiting for sent_cb to be sent
> with qemu_send_packet_async.
> Thus when sent_cb called, we should finish
> our transfer of packets K..N from s->packets_buf
> to qemu calling qemu_send_packet_async.
> I meant this.
I missed the comment. The description is contradicting with the actual
code; qemu_net_queue_send_iov appends the packet to the queue whenever
it cannot send one immediately.
Jason Wang, I saw you are in the MAINTAINERS for net/. Can you tell if
calling qemu_send_packet_async is allowed after it returns 0?
Regards,
Akihiko Odaki
next prev parent reply other threads:[~2022-03-01 8:23 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-02-25 17:13 [PATCH v15 0/8] Add vmnet.framework based network backend Vladislav Yaroshchuk
2022-02-25 17:13 ` [PATCH v15 1/8] net/vmnet: add vmnet dependency and customizable option Vladislav Yaroshchuk
2022-02-25 17:13 ` [PATCH v15 2/8] net/vmnet: add vmnet backends to qapi/net Vladislav Yaroshchuk
2022-02-28 10:07 ` Markus Armbruster
2022-02-28 10:47 ` Vladislav Yaroshchuk
2022-02-25 17:13 ` [PATCH v15 3/8] net/vmnet: implement shared mode (vmnet-shared) Vladislav Yaroshchuk
2022-02-25 17:46 ` Akihiko Odaki
2022-02-26 8:37 ` Vladislav Yaroshchuk
2022-02-26 9:16 ` Akihiko Odaki
2022-02-26 11:33 ` Vladislav Yaroshchuk
2022-02-26 12:26 ` Akihiko Odaki
2022-02-28 11:59 ` Vladislav Yaroshchuk
2022-03-01 5:52 ` Akihiko Odaki
2022-03-01 8:09 ` Vladislav Yaroshchuk
2022-03-01 8:21 ` Akihiko Odaki [this message]
2022-03-03 15:43 ` Vladislav Yaroshchuk
2022-03-03 18:34 ` Akihiko Odaki
2022-03-04 4:39 ` Akihiko Odaki
2022-03-04 1:37 ` Jason Wang
2022-03-04 4:37 ` Akihiko Odaki
2022-03-07 3:59 ` Jason Wang
2022-03-07 4:25 ` Akihiko Odaki
2022-03-07 4:51 ` Jason Wang
2022-03-07 6:54 ` Akihiko Odaki
2022-02-25 17:13 ` [PATCH v15 4/8] net/vmnet: implement host mode (vmnet-host) Vladislav Yaroshchuk
2022-02-25 17:13 ` [PATCH v15 5/8] net/vmnet: implement bridged mode (vmnet-bridged) Vladislav Yaroshchuk
2022-02-25 17:50 ` Akihiko Odaki
2022-02-25 17:14 ` [PATCH v15 6/8] net/vmnet: update qemu-options.hx Vladislav Yaroshchuk
2022-02-25 17:14 ` [PATCH v15 7/8] net/vmnet: update hmp-commands.hx Vladislav Yaroshchuk
2022-02-25 17:14 ` [PATCH v15 8/8] net/vmnet: update MAINTAINERS list Vladislav Yaroshchuk
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=e5302fba-e773-c56d-388f-a3593bacd977@gmail.com \
--to=akihiko.odaki@gmail.com \
--cc=agraf@csgraf.de \
--cc=alex.bennee@linaro.org \
--cc=armbru@redhat.com \
--cc=dirty@apple.com \
--cc=eblake@redhat.com \
--cc=f4bug@amsat.org \
--cc=hello@adns.io \
--cc=hsp.cat7@gmail.com \
--cc=jasowang@redhat.com \
--cc=kraxel@redhat.com \
--cc=peter.maydell@linaro.org \
--cc=phillip.ennen@gmail.com \
--cc=phillip@axleos.com \
--cc=qemu-devel@nongnu.org \
--cc=qemu_oss@crudebyte.com \
--cc=r.bolshakov@yadro.com \
--cc=roman@roolebo.dev \
--cc=vladislav.yaroshchuk@jetbrains.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).