On Thu, Aug 16, 2012 at 10:07 AM, Kevin Wolf <kw...@redhat.com> wrote: > Am 16.08.2012 10:57, schrieb Stefan Hajnoczi: >> QEMU has a policy of keeping a stable guest device ABI. When new guest >> device >> features are introduced they must not change hardware info seen by existing >> guests. This is important because operating systems or applications may >> "fingerprint" the hardware and refuse to run when the hardware changes. To >> always get the latest guest device ABI, run with x86 machine type "pc". >> >> This patch hides the new VIRTIO_BLK_F_CONFIG_WCE virtio feature bit from >> existing machine types. Only pc-1.2 and later will expose this feature >> by default. >> >> For more info on the VIRTIO_BLK_F_CONFIG_WCE feature bit, see: >> >> commit 13e3dce068773c971ff2f19d986378c55897c4a3 >> Author: Paolo Bonzini <pbonz...@redhat.com> >> Date: Thu Aug 9 16:07:19 2012 +0200 >> >> virtio-blk: support VIRTIO_BLK_F_CONFIG_WCE >> >> Also rename VIRTIO_BLK_F_WCACHE to VIRTIO_BLK_F_WCE for consistency >> with >> the spec. >> >> Signed-off-by: Paolo Bonzini <pbonz...@redhat.com> >> Signed-off-by: Kevin Wolf <kw...@redhat.com> >> >> Anthony Liguori <aligu...@us.ibm.com> reported: >> >> This broke qemu-test because it changed the pc-1.0 machine type: >> >> Setting guest RANDOM seed to 47167 >> *** Running tests *** >> Running test /tests/finger-print.sh... OK >> --- fingerprints/pc-1.0.x86_64 2011-12-18 13:08:40.000000000 -0600 >> +++ fingerprint.txt 2012-08-12 13:30:48.000000000 -0500 >> @@ -55,7 +55,7 @@ >> /sys/bus/pci/devices/0000:00:06.0/subsystem_device=0x0002 >> /sys/bus/pci/devices/0000:00:06.0/class=0x010000 >> /sys/bus/pci/devices/0000:00:06.0/revision=0x00 >> -/sys/bus/pci/devices/0000:00:06.0/virtio/host-features=0x710006d4 >> +/sys/bus/pci/devices/0000:00:06.0/virtio/host-features=0x71000ed4 >> /sys/class/dmi/id/bios_vendor=Bochs >> /sys/class/dmi/id/bios_date=01/01/2007 >> /sys/class/dmi/id/bios_version=Bochs >> Guest fingerprint changed for pc-1.0! >> >> Reported-by: Anthony Liguori <aligu...@us.ibm.com> >> Signed-off-by: Stefan Hajnoczi <stefa...@linux.vnet.ibm.com> >> --- >> Anthony, does this fix your test case? > > Looks good to me, but I think I'll wait for Anthony to check it.
Yep, we definitely need to check because I haven't run qemu-test myself. Stefan
