From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:45849) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1WLDwu-00078b-Qi for qemu-devel@nongnu.org; Wed, 05 Mar 2014 10:41:37 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1WLDwo-0007ou-Rp for qemu-devel@nongnu.org; Wed, 05 Mar 2014 10:41:36 -0500 Received: from mx1.redhat.com ([209.132.183.28]:7434) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1WLDwo-0007oh-9b for qemu-devel@nongnu.org; Wed, 05 Mar 2014 10:41:30 -0500 Date: Wed, 5 Mar 2014 16:41:20 +0100 From: Kevin Wolf Message-ID: <20140305154120.GC4483@dhcp-200-207.str.redhat.com> References: <1394031649-31915-1-git-send-email-kwolf@redhat.com> <1394031649-31915-2-git-send-email-kwolf@redhat.com> <20140305153323.GA1709@irqsave.net> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline In-Reply-To: <20140305153323.GA1709@irqsave.net> Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH 1/2] iscsi: Use bs->sg for everything else than disks List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: =?iso-8859-1?Q?Beno=EEt?= Canet Cc: pbonzini@redhat.com, pl@kamp.de, qemu-devel@nongnu.org Am 05.03.2014 um 16:33 hat Beno=EEt Canet geschrieben: > The Wednesday 05 Mar 2014 =E0 16:00:48 (+0100), Kevin Wolf wrote : > > The current iscsi block driver code makes the rather arbitrary decisi= on > > that TYPE_MEDIUM_CHANGER and TYPE_TAPE devices have bs->sg =3D 1 and = all > > other device types are disks. > >=20 > > Instead of this, check for TYPE_DISK to expose the disk interface and > > make everything else bs->sg =3D 1. In particular, this includes devic= es > > with TYPE_STORAGE_ARRAY, which is what LUN 0 of an iscsi target is. > > (See https://bugzilla.redhat.com/show_bug.cgi?id=3D1067784 for the ex= act > > scenario.) > >=20 > > Signed-off-by: Kevin Wolf > > --- > > block/iscsi.c | 9 ++++----- > > 1 file changed, 4 insertions(+), 5 deletions(-) > >=20 > > diff --git a/block/iscsi.c b/block/iscsi.c > > index 0a15f53..b490e98 100644 > > --- a/block/iscsi.c > > +++ b/block/iscsi.c > > @@ -1231,12 +1231,11 @@ static int iscsi_open(BlockDriverState *bs, Q= Dict *options, int flags, > > bs->total_sectors =3D sector_lun2qemu(iscsilun->num_blocks, iscs= ilun); > > bs->request_alignment =3D iscsilun->block_size; > > =20 > > - /* Medium changer or tape. We dont have any emulation for this s= o this must > > - * be sg ioctl compatible. We force it to be sg, otherwise qemu = will try > > - * to read from the device to guess the image format. > > + /* We don't have any emulation for devices other than disks and = CD-ROMs, so > > + * this must be sg ioctl compatible. We force it to be sg, other= wise qemu > > + * will try to read from the device to guess the image format. > > */ > > - if (iscsilun->type =3D=3D TYPE_MEDIUM_CHANGER || > > - iscsilun->type =3D=3D TYPE_TAPE) { > > + if (iscsilun->type !=3D TYPE_DISK && iscsilun->type !=3D TYPE_RO= M) { > I don't understand you are talking about type TYPE_STORAGE_ARRAY in the= commit > message but testing against TYPE_ROM and they are not the same. > Maybe a comment explaining the role of TYPE_ROM would give a clue to no= n scsi > savy persons like me. TYPE_STORAGE_ARRAY is what exposed the bug (see the Bugzilla reference). It was previously bs->sg =3D 0, but becomes now bs->sg =3D 1. The reason = for this is that the usual block layer operations only work on disks (i.e. hard disks =3D TYPE_DISK and CD-ROM drives =3D TYPE_ROM), so for all the rest we can only expose a generic SCSI device, but no real block device. Kevin