From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([208.118.235.92]:42673)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1UFj4A-0002w6-HB
	for qemu-devel@nongnu.org; Wed, 13 Mar 2013 06:37:53 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1UFj47-0001Rc-Tm
	for qemu-devel@nongnu.org; Wed, 13 Mar 2013 06:37:50 -0400
Received: from mx1.redhat.com ([209.132.183.28]:52062)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1UFj47-0001RS-MA
	for qemu-devel@nongnu.org; Wed, 13 Mar 2013 06:37:47 -0400
Message-ID: <514056E2.2030804@redhat.com>
Date: Wed, 13 Mar 2013 11:37:22 +0100
From: Paolo Bonzini <pbonzini@redhat.com>
MIME-Version: 1.0
References: <1362624800-10682-1-git-send-email-qemulist@gmail.com>
	<1362624800-10682-2-git-send-email-qemulist@gmail.com>
	<513EEC52.2010504@redhat.com>
	<CAJnKYQ=1Gm-7W_-0dqLM3AuiNOjh5JRxGgtQA=58jjW4dbyBYg@mail.gmail.com>
In-Reply-To: <CAJnKYQ=1Gm-7W_-0dqLM3AuiNOjh5JRxGgtQA=58jjW4dbyBYg@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [PATCH v2 1/5] net: spread hub on AioContexts
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: liu ping fan <qemulist@gmail.com>
Cc: Stefan Hajnoczi <stefanha@gmail.com>, "Michael S. Tsirkin" <mst@redhat.com>, qemu-devel@nongnu.org, Anthony Liguori <anthony@codemonkey.ws>, mdroth <mdroth@linux.vnet.ibm.com>

Il 13/03/2013 03:26, liu ping fan ha scritto:
>>> >> +        qemu_net_queue_append(port->nc.peer->send_queue, &port->nc,
>>> >> +                            QEMU_NET_PACKET_FLAG_NONE, buf, len, NULL);
>>> >> +        event_notifier_set(&port->e);
>> >
>> > Why are the context and the EventNotifier a property of the port, rather
>> > than applicable to the NetClientState?
>> >
> Yes, embed context into NetClientState is more reasonable, but as for
> EventNotifier, considering about if we port tap onto context, the tap
> do not have EventNotifier.
> 

There doesn't even need to be an EventNotifier, instead you can pass the
NetClientState's AioContext to the queue and use a bottom half on the
AioContext.

Furthermore, the bottom half should be completely transparent.  Callers
can keep using qemu_net_queue_flush, qemu_net_queue_flush schedules the
bottom half (perhaps, with an optimization, it only does that if the
queue is not empty), the handler actually performs the flush.

Paolo