From mboxrd@z Thu Jan 1 00:00:00 1970 From: Benjamin Herrenschmidt Subject: Re: [Qemu-devel] [PATCH] virtio-pci: Fix endianness of virtio config Date: Wed, 11 Jan 2012 09:10:48 +1100 Message-ID: <1326233448.23910.74.camel@pasglop> References: <1326195311.23910.59.camel@pasglop> <5B504266-DFF3-48F8-AEA4-C947189CDB62@suse.de> <4F0CA118.30203@suse.de> <77DAA08C-9D39-491A-B8AC-A23915FE2A3D@suse.de> <1326229468.23910.69.camel@pasglop> <1326233090.23910.71.camel@pasglop> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: Andreas =?ISO-8859-1?Q?F=E4rber?= , kvm-ppc@vger.kernel.org, "qemu-devel@nongnu.org" , "kvm@vger.kernel.org" , David Gibson To: Alexander Graf Return-path: In-Reply-To: <1326233090.23910.71.camel@pasglop> Sender: kvm-ppc-owner@vger.kernel.org List-Id: kvm.vger.kernel.org On Wed, 2012-01-11 at 09:04 +1100, Benjamin Herrenschmidt wrote: > On Tue, 2012-01-10 at 22:45 +0100, Alexander Graf wrote: > > Here's the thing that I don't understand. What exactly is breaking for > > you? I tried -M pseries on a ppc box and on an x86 box and both times > > was able to see /dev/vda. > > And mount it and use it ? Here I get the capacity wrong if I don't have > my patch and can't actually boot a distro off it. Hrm actually, it might well work with Linux regardless because Linux only ever does 8 bit accesses to the virtio config space. It's SLOF that breaks because SLOF uses 2 and 4 byte accesses (which are allowed per spec as far as I can tell) but the spurrious swapping happening in that case means we lost the byte address invariance. Cheers, Ben.