From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:56645) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1adZ6B-0005Rx-77 for qemu-devel@nongnu.org; Wed, 09 Mar 2016 03:04:04 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1adZ67-0008UF-Vx for qemu-devel@nongnu.org; Wed, 09 Mar 2016 03:04:03 -0500 Received: from mx1.redhat.com ([209.132.183.28]:52430) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1adZ67-0008UA-QY for qemu-devel@nongnu.org; Wed, 09 Mar 2016 03:03:59 -0500 Received: from int-mx13.intmail.prod.int.phx2.redhat.com (int-mx13.intmail.prod.int.phx2.redhat.com [10.5.11.26]) by mx1.redhat.com (Postfix) with ESMTPS id 767E290E47 for ; Wed, 9 Mar 2016 08:03:59 +0000 (UTC) Message-ID: <1457510637.676.15.camel@redhat.com> From: Gerd Hoffmann Date: Wed, 09 Mar 2016 09:03:57 +0100 In-Reply-To: <20160308081533.57e0fa8d@t450s.home> References: <1457080913-30018-1-git-send-email-kraxel@redhat.com> <20160307144151.5d17dfb4@t450s.home> <1457424290.22567.24.camel@redhat.com> <20160308081533.57e0fa8d@t450s.home> Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Mime-Version: 1.0 Subject: Re: [Qemu-devel] [PATCH v2] vfio/igd: handle q35 machine type List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Alex Williamson Cc: qemu-devel@nongnu.org Hi, > > q35 already has a isa bridge @ 1f.0, and it's not a dummy device but > > emulates the ich9 lpc. Just overwriting the identity of that device > > isn't a good idea I think. >=20 > It's not a good idea in practice or in theory? It's exactly what we're > proposing to do for the host bridge. We don't change the pci ids for the host bridge. It's still a i440fx/q35 host bridge, only the subsystem ID is changed (although I'm sometimes wondering why), and the gfx registers added of course. > Do we know for certain that any of > the emulated ich9 registers do not align with newer chipsets? See drivers/mfd/lpc_ich.c (lpc_chipset_info array). Lots of different versions exist for GPIO wiring. Which probably isn't that much of a problem as I think we don't emulate gpio. For iTCO there are tree different versions too, and we *do* emulate that. > It seems > quite limiting of igd assignment on q35 VMs if we're going to remove > collateral device modification. Should we only claim "legacy" igd > support on 440FX and support only universal passthrough on q35? Thanks, I'll go test this a bit more. Possibly we can drop the whole lpc tweaking. At least when looking at the linux driver code it doesn't seem to be very important. cheers, Gerd