From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43)
	id 1MldlZ-0004Av-IC
	for qemu-devel@nongnu.org; Thu, 10 Sep 2009 03:08:25 -0400
Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43)
	id 1MldlV-0004AI-5d
	for qemu-devel@nongnu.org; Thu, 10 Sep 2009 03:08:25 -0400
Received: from [199.232.76.173] (port=42179 helo=monty-python.gnu.org)
	by lists.gnu.org with esmtp (Exim 4.43) id 1MldlV-0004AF-1x
	for qemu-devel@nongnu.org; Thu, 10 Sep 2009 03:08:21 -0400
Received: from mx20.gnu.org ([199.232.41.8]:6101)
	by monty-python.gnu.org with esmtps (TLS-1.0:RSA_AES_256_CBC_SHA1:32)
	(Exim 4.60) (envelope-from <kwolf@redhat.com>) id 1MldlU-0003wB-Go
	for qemu-devel@nongnu.org; Thu, 10 Sep 2009 03:08:20 -0400
Received: from mx1.redhat.com ([209.132.183.28])
	by mx20.gnu.org with esmtp (Exim 4.60)
	(envelope-from <kwolf@redhat.com>) id 1MldlT-0003nF-Ot
	for qemu-devel@nongnu.org; Thu, 10 Sep 2009 03:08:20 -0400
Message-ID: <4AA8A59E.9090206@redhat.com>
Date: Thu, 10 Sep 2009 09:07:10 +0200
From: Kevin Wolf <kwolf@redhat.com>
MIME-Version: 1.0
References: <1252511618-19497-1-git-send-email-kwolf@redhat.com>	<1252511618-19497-2-git-send-email-kwolf@redhat.com>
	<m3ocpj67ip.fsf@neno.mitica>
In-Reply-To: <m3ocpj67ip.fsf@neno.mitica>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Subject: [Qemu-devel] Re: [PATCH v3 1/2] Add bdrv_aio_multiwrite
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.gnu.org/pipermail/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: Juan Quintela <quintela@trasno.org>
Cc: qemu-devel@nongnu.org

Am 10.09.2009 01:08, schrieb Juan Quintela:
> Kevin Wolf <kwolf@redhat.com> wrote:
> 
> Hi again
> 
> 
>> +            QEMUIOVector *qiov = qemu_mallocz(sizeof(*qiov));
>> +            qemu_iovec_init(qiov,
>> +                reqs[outidx].qiov->niov + reqs[i].qiov->niov + 1);
>> +
>> +            // Add the first request to the merged one. If the requests are
>> +            // overlapping, drop the last sectors of the first request.
>> +            size = (reqs[i].sector - reqs[outidx].sector) << 9;
>> +            qemu_iovec_concat(qiov, reqs[outidx].qiov, size);
>> +
>> +            // We might need to add some zeros between the two requests
>> +            if (reqs[i].sector > oldreq_last) {
>> +                size_t zero_bytes = (reqs[i].sector - oldreq_last) << 9;
>> +                uint8_t *buf = qemu_blockalign(bs, zero_bytes);
>> +                memset(buf, 0, zero_bytes);
>> +                qemu_iovec_add(qiov, buf, zero_bytes);
>> +                mcb->callbacks[i].free_buf = buf;
>> +            }
>> +
>> +            // Add the second request
>> +            qemu_iovec_concat(qiov, reqs[i].qiov, reqs[i].qiov->size);
>> +
>> +            reqs[outidx].nb_sectors += reqs[i].nb_sectors;
>> +            reqs[outidx].qiov = qiov;
> 
> What frees reqs[outidx].qiov previous value, or new one for that matter?

Your decision where to stop quoting the code is kind of funny. The very
next line is:

mcb->callbacks[i].free_qiov = reqs[outidx].qiov;

This saves each newly allocated qiov. It is later freed in
multiwrite_user_cb. To free the original value of qiov is the task of
the caller.

> I can't see where it is done.  As far as I can see, we are losing both
> the ones that we are overwritten and the ones for the request that got merged.

Hope you see them now. :-)

Kevin