From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:34601)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbutsykin@virtuozzo.com>) id 1bhfqT-0004uQ-Hw
	for qemu-devel@nongnu.org; Wed, 07 Sep 2016 12:37:06 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <pbutsykin@virtuozzo.com>) id 1bhfqP-0002AG-Cd
	for qemu-devel@nongnu.org; Wed, 07 Sep 2016 12:37:05 -0400
References: <20160829171021.4902-1-pbutsykin@virtuozzo.com>
	<20160829171021.4902-6-pbutsykin@virtuozzo.com>
	<20160901152813.GE6355@noname.redhat.com>
	<57CEF4DE.9030901@virtuozzo.com>
	<20160906170723.GJ4667@noname.redhat.com>
From: Pavel Butsykin <pbutsykin@virtuozzo.com>
Message-ID: <57D03EA2.3020002@virtuozzo.com>
Date: Wed, 7 Sep 2016 19:21:54 +0300
MIME-Version: 1.0
In-Reply-To: <20160906170723.GJ4667@noname.redhat.com>
Content-Type: text/plain; charset="windows-1252"; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [PATCH RFC v2 05/22] block/pcache: add aio
 requests into cache
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: Kevin Wolf <kwolf@redhat.com>
Cc: qemu-block@nongnu.org, qemu-devel@nongnu.org, mreitz@redhat.com, stefanha@redhat.com, den@openvz.org, jsnow@redhat.com, eblake@redhat.com, famz@redhat.com

On 06.09.2016 20:07, Kevin Wolf wrote:
> Am 06.09.2016 um 18:54 hat Pavel Butsykin geschrieben:
>>>> +        *out_node = found;
>>>> +        return false;
>>>> +    }
>>>> +    atomic_add(&s->pcache.curr_size, new_node->cm.nb_sectors);
>>>
>>> atomic_add() implies that you have concurrent threads. I don't see any.
>>>
>> Yeah, but what about iothread? Doesn't presence of iothread lead to
>> parallel handling of requests in multiple threads?
>
> No, it doesn't. We're running under the AioContext lock.

Yes, you're right. For some reason I thought that there is such a
possibility :)

> If we were actually running in parallel threads, your locks might also
> be needed, but then they would have to be real thread locks instead of
> coroutine locks. But it doesn't happen today, so I think it's best to
> ignore.

I agree, use coroutine locks to protect the data was a bad idea.

> Kevin
>