On Tue, Jan 04, 2011, Alexander Graf wrote about "Re: [Qemu-devel] [PATCH] Add VMX cpuid feature to qemu64": > If qemu-kvm still uses the "qemu64" type, that's plainly a bug :). It really > should use -cpu kvm64 / kvm32 as default.
When I run qemu-kvm's qemu-system-x86_64, and look at the CPU I get in the guest, I the CPU is clearly "QEMU Virtual CPU version 0.13.50" (qemu64) and not "Common KVM processor" (kvm64) the "-enable-kvm" option doesn't appear to change anything. Could this have always been a bug, which remained hidden in plain sight? :-) If I send a patch to default to kvm64, not qemu64, when KVM is being used, would that be acceptable by all parties involved (KVM people and QEMU people)? Also, kvm64 appears to have much fewer capabilities listed than qemu64. Was this list ever kept up to date with KVM's capabilities? On first glance it appars that even CPUID_EXT3_SVM is missing, so if kvm64 is used I think nested SVM will also stop working (in my patch, I can add this bit just like I plan to add CPUID_EXT_VMX). > > Well, as you can see in target-i386/cpuid.c, at least in the qemu version > > I checked out from KVM's repository, the "core2duo" and "coreduo" cpu types > > do list the CPUID_EXT_VMX - and those are the only ones that list this > > feature. > > Oops, probably my bad :). Do you want me to fix that in the patch as well? Thanks, Nadav. > If you add CPUID_EXT_VMX to the "kvm64" type and make that the default when > kvm is enabled (like we did for nested svm), all should be well, no? I don't know, I don't know how the KVM users would react to their default CPU type being changed from qemu64 to kvm64. Perhaps Avi Kivity or somebody else has an opinion about this? Thanks, Nadav. -- Nadav Har'El | Wednesday, Jan 5 2011, 29 Tevet 5771 n...@math.technion.ac.il |----------------------------------------- Phone +972-523-790466, ICQ 13349191 |Those who beat their swords into http://nadav.harel.org.il |plowshares will plow for those who don't.