From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from [140.186.70.92] (port=52311 helo=eggs.gnu.org)
	by lists.gnu.org with esmtp (Exim 4.43) id 1PxgLi-0000ev-Tf
	for qemu-devel@nongnu.org; Thu, 10 Mar 2011 08:56:20 -0500
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1PxgLg-0003xG-24
	for qemu-devel@nongnu.org; Thu, 10 Mar 2011 08:56:17 -0500
Received: from mx1.redhat.com ([209.132.183.28]:36379)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1PxgLf-0003x9-Og
	for qemu-devel@nongnu.org; Thu, 10 Mar 2011 08:56:16 -0500
Message-ID: <4D78D875.2090101@redhat.com>
Date: Thu, 10 Mar 2011 14:56:05 +0100
From: Paolo Bonzini <pbonzini@redhat.com>
MIME-Version: 1.0
References: <4D7767C0.6060609@siemens.com>
	<1299761979-15197-2-git-send-email-corentin.chary@gmail.com>
	<4D78CCDF.6090906@redhat.com> <4D78D603.5060000@us.ibm.com>
In-Reply-To: <4D78D603.5060000@us.ibm.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: [Qemu-devel] Re: [PATCH 2/2] vnc: don't mess up with iohandlers in
	the vnc thread
List-Id: qemu-devel.nongnu.org
List-Unsubscribe: <http://lists.nongnu.org/mailman/listinfo/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: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Anthony Liguori <aliguori@us.ibm.com>
Cc: kvm@vger.kernel.org, Jan Kiszka <jan.kiszka@siemens.com>, qemu-devel <qemu-devel@nongnu.org>, Anthony Liguori <aliguori@linux.vnet.ibm.com>, Corentin Chary <corentin.chary@gmail.com>, Peter Lieven <pl@dlh.net>

On 03/10/2011 02:45 PM, Anthony Liguori wrote:
> On 03/10/2011 07:06 AM, Paolo Bonzini wrote:
>> On 03/10/2011 01:59 PM, Corentin Chary wrote:
>>> Instead, we now store the data in a temporary buffer, and use a socket
>>> pair to notify the main thread that new data is available.
>>
>> You can use a bottom half for this instead of a special socket.
>> Signaling a bottom half is async-signal- and thread-safe.
>
> Bottom halves are thread safe?
>
> I don't think so.
>
> They probably should be but they aren't today.

Creating a new bottom half is not thread-safe, but scheduling one is. 
Assuming that you never use qemu_bh_schedule_idle, qemu_bh_schedule 
boils down to:

     if (bh->scheduled)
         return;
     bh->scheduled = 1;
     /* stop the currently executing CPU to execute the BH ASAP */
     qemu_notify_event();

You may have a spurious wakeup if two threads race on the same bottom 
half (including the signaling thread racing with the IO thread), but 
overall you can safely treat them as thread-safe.

Paolo