From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([140.186.70.92]:53407)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <avi@redhat.com>) id 1RmoHJ-0000AK-1S
	for qemu-devel@nongnu.org; Mon, 16 Jan 2012 10:15:21 -0500
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <avi@redhat.com>) id 1RmoHH-0007WD-UU
	for qemu-devel@nongnu.org; Mon, 16 Jan 2012 10:15:21 -0500
Message-ID: <4F143F03.5020001@redhat.com>
Date: Mon, 16 Jan 2012 17:15:15 +0200
From: Avi Kivity <avi@redhat.com>
MIME-Version: 1.0
References: <1325894809-17322-1-git-send-email-andreas.faerber@web.de>
	<1326424168-15705-1-git-send-email-andreas.faerber@web.de>
	<1326424168-15705-4-git-send-email-andreas.faerber@web.de>
	<4F129A1A.9070709@redhat.com> <4F143D83.3050502@web.de>
In-Reply-To: <4F143D83.3050502@web.de>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Subject: Re: [Qemu-devel] [PATCH v3 3/8] prep_pci: Update I/O to
	MemoryRegion ops
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: =?UTF-8?B?QW5kcmVhcyBGw6RyYmVy?= <andreas.faerber@web.de>
Cc: Alexander Graf <agraf@suse.de>, =?UTF-8?B?QmVub8OudCBDYW5ldA==?= <benoit.canet@gmail.com>, qemu-ppc@nongnu.org, qemu-devel@nongnu.org, "Michael S. Tsirkin" <mst@redhat.com>

On 01/16/2012 05:08 PM, Andreas F=C3=A4rber wrote:
> Am 15.01.2012 10:19, schrieb Avi Kivity:
> > On 01/13/2012 05:09 AM, Andreas F=C3=A4rber wrote:
> >> Convert to new-style read/write callbacks.
> >>
> >> =20
> >> -static uint32_t PPC_PCIIO_readl (void *opaque, target_phys_addr_t a=
ddr)
> >> +static uint64_t ppc_pci_io_read(void *opaque, target_phys_addr_t ad=
dr,
> >> +                                unsigned int size)
> >>  {
> >>      PREPPCIState *s =3D opaque;
> >> -    uint32_t val;
> >> -    val =3D pci_data_read(s->bus, PPC_PCIIO_config(addr), 4);
> >> -    return val;
> >> +    switch (size) {
> >> +    case 1:
> >> +    case 2:
> >> +    case 4:
> >> +        return pci_data_read(s->bus, PPC_PCIIO_config(addr), size);
> >> +    default:
> >> +        abort();
> >> +    }
> >>  }
> >=20
> > Huh? just call pci_data_read() unconditionally.
>
> Just so that I understand, is that because PReP is 32-bit ppc? In the
> above mechanical conversion, size 8 would abort.

The memory core never issues size 8 transactions, since the code is
unprepared for it.  When we will support it, you'll have to explicitly
declare it with .impl.max_access_size =3D 8 or something.

> BTW did we agree on an indentation style for switch?

What you wrote conforms to the de facto standard.

--=20
error compiling committee.c: too many arguments to function