On 12/06/2010 07:48 PM, Roedel, Joerg wrote:
On Mon, Dec 06, 2010 at 11:10:46AM -0500, Avi Kivity wrote: > On 12/03/2010 06:42 PM, Joerg Roedel wrote: > > This patch implements the xsetbv intercept to the AMD part > > of KVM. This makes AVX usable in a save way for the guest on > > AVX capable AMD hardware. > > The patch is tested by using AVX in the guest and host in > > parallel and checking for data corruption. I also used the > > KVM xsave unit-tests and they all pass. > > > > That is really strange. You didn't need to do anything to get cpuid.avx > recognized. So running an older kvm on newer hardware will happily > expose avx even though it's not supported. > > We screwed up - we should have made cpuid.avx dependent on vendor support. Hmm, right. The set_supported_cpuid arch-callback should basically disable xsave on AMD for all KVM versions which do not handle the xsetbv intercept.
Please post a patch to do that, and update this patch to undo the change. We'll backport the first patch to -stable so that people running older kernels don't get a nasty surprise when they upgrade their hardware (and use -cpu host).
One more thing to watch out for. We also need to see if there aren't more mistakes like that out there.
-- error compiling committee.c: too many arguments to function -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html