From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:48021)
	by lists.gnu.org with esmtp (Exim 4.71) (envelope-from <pl@kamp.de>)
	id 1WTEkT-0005iO-1g
	for qemu-devel@nongnu.org; Thu, 27 Mar 2014 14:10:02 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <pl@kamp.de>) id 1WTEkJ-0003tH-LO
	for qemu-devel@nongnu.org; Thu, 27 Mar 2014 14:09:52 -0400
Received: from mx-v6.kamp.de ([2a02:248:0:51::16]:40884 helo=mx01.kamp.de)
	by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from <pl@kamp.de>)
	id 1WTEkJ-0003t0-A8
	for qemu-devel@nongnu.org; Thu, 27 Mar 2014 14:09:43 -0400
Message-ID: <53346963.2050004@kamp.de>
Date: Thu, 27 Mar 2014 19:09:39 +0100
From: Peter Lieven <pl@kamp.de>
MIME-Version: 1.0
References: <1395927818-26431-1-git-send-email-pl@kamp.de>
	<53343305.7000800@redhat.com>
In-Reply-To: <53343305.7000800@redhat.com>
Content-Type: text/plain; charset=ISO-8859-15
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [PATCH] block/iscsi: speed up read for unallocated
	sectors
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: Paolo Bonzini <pbonzini@redhat.com>, qemu-devel@nongnu.org
Cc: kwolf@redhat.com, ronniesahlberg@gmail.com, stefanha@redhat.com

Am 27.03.2014 15:17, schrieb Paolo Bonzini:
> Il 27/03/2014 14:43, Peter Lieven ha scritto:
>>
>> +static void iscsi_allocationmap_set(IscsiLun *iscsilun, int64_t sector_num,
>> +                                    int nb_sectors)
>> +{
>> +    if (!iscsilun->cluster_sectors) {
>> +        return;
>> +    }
>> +    bitmap_set(iscsilun->allocationmap,
>> +               sector_num / iscsilun->cluster_sectors,
>> +               DIV_ROUND_UP(nb_sectors, iscsilun->cluster_sectors));
>> +}
>> +
>> +static void iscsi_allocationmap_clear(IscsiLun *iscsilun, int64_t sector_num,
>> +                                      int nb_sectors)
>> +{
>> +    if (!iscsilun->cluster_sectors) {
>> +        return;
>> +    }
>> +    bitmap_clear(iscsilun->allocationmap,
>> +                 sector_num / iscsilun->cluster_sectors,
>> +                 DIV_ROUND_UP(nb_sectors, iscsilun->cluster_sectors));
>> +}
>> +
>
> Here you need to round the other way, because partial clusters are not unmapped:
>
>     cluster_num = DIV_ROUND_UP(sector_num, iscsi->cluster_sectors);
>     nb_clusters = (sector_num + nb_sectors) / iscsi->cluster_sectors
>         - cluster_num;
>     bitmap_clear(iscsilun->allocationmap, cluster_num, nb_clusters);

Right, its not criticial because the allocationmap is more a hint, but its worth correcting
this. I have to check if bitmap_clear can cope with nb_clusters == 0.

thank you,
Peter

>
> Paolo