From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([208.118.235.92]:51836)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <paolo.bonzini@gmail.com>) id 1TDzPz-0003cO-0B
	for qemu-devel@nongnu.org; Tue, 18 Sep 2012 11:09:04 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <paolo.bonzini@gmail.com>) id 1TDzPy-0002lb-7D
	for qemu-devel@nongnu.org; Tue, 18 Sep 2012 11:08:54 -0400
Received: from mail-pb0-f45.google.com ([209.85.160.45]:64687)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <paolo.bonzini@gmail.com>) id 1TDzPy-0002lP-0b
	for qemu-devel@nongnu.org; Tue, 18 Sep 2012 11:08:54 -0400
Received: by pbbrp12 with SMTP id rp12so188750pbb.4
	for <qemu-devel@nongnu.org>; Tue, 18 Sep 2012 08:08:53 -0700 (PDT)
Sender: Paolo Bonzini <paolo.bonzini@gmail.com>
Message-ID: <50588E7F.2060908@redhat.com>
Date: Tue, 18 Sep 2012 17:08:47 +0200
From: Paolo Bonzini <pbonzini@redhat.com>
MIME-Version: 1.0
References: <1347968442-8860-1-git-send-email-kwolf@redhat.com>
	<1347968442-8860-16-git-send-email-kwolf@redhat.com>
	<505888CC.8070207@redhat.com> <50588D22.6080703@redhat.com>
	<50588DB0.1000207@redhat.com>
In-Reply-To: <50588DB0.1000207@redhat.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [RFC PATCH 15/16] qcow2: Cancel COW when
	overwritten
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>
Cc: Kevin Wolf <kwolf@redhat.com>, qemu-devel@nongnu.org

Il 18/09/2012 17:05, Paolo Bonzini ha scritto:
>>>>>> >>> >> +    qemu_co_mutex_unlock(&s->lock);
>>>>>> >>> >> +    qemu_co_rwlock_wrlock(&m->l2_writeback_lock);
>>>> >> > 
>>>> >> > Can anybody else take the lock as reader again at this point?  If not, I
>>>> >> > wonder if this is more clear if you write it as a CoQueue.
>> > This isn't "let one writer complete, then wake up n readers" (which
>> > would indeed be represented more naturally as CoQueue), but rather "let
>> > n readers complete, then wake up one writer".
> So would it be representable as a list of children (matching the parent
> pointer) and then unlocking is
> 
>   remove self from children list
>   if the children list is empty, wake parent->co
> 
> ?

Of course this only work if the list of children cannot become non-empty
anymore after the writer restarts.

Paolo