[libvirt] [PATCH 1/6] kvm: Add KVM_GET_EMULATED_CPUID

[Eduardo Habkost]

On Thu, Sep 26, 2013 at 08:55:24PM +0200, Borislav Petkov wrote:
> On Thu, Sep 26, 2013 at 11:19:15AM -0300, Eduardo Habkost wrote:
> > Then we may have a problem: some CPU models already have "movbe"
> > included (e.g. Haswell), and patch 6/6 will make "-cpu Haswell" get
> > movbe enabled even if it is being emulated.
> 
> Huh? HSW has MOVBE so we won't #UD on it and MOVBE will get executed in
> hardware when executing the guest. IOW, we'll never get to the emulation
> path of piggybacking on the #UD.
> 
> > So if we really want to avoid enabling emulated features by mistake,
> > we may need a new CPU flag in addition to "enforce" to tell QEMU that
> > it is OK to enable emulated features (maybe "-cpu ...,emulate"?).
> 
> EMULATED_CPUID are off by default and only if you request them
> specifically, they get enabled.

Please point me to the code that does this, because I don't see it on
patch 6/6.

> If you start with "-cpu Haswell", MOVBE
> will be already set in the host CPUID.
> 
> Or am I missing something?

In the Haswell example, it is unlikely but possible in theory: you would
need a CPU that supported all features from Haswell except movbe. But
what will happen if you are using "-cpu n270,enforce" on a SandyBridge
host?

Also, we don't know anything about future CPUs or future features that
will end up on EMULATED_CPUID. The current code doesn't have anything to
differentiate features that were already included in the CPU definition
and ones explicitly enabled in the command-line (and I would like to
keep it that way).

And just because a feature was explicitly enabled in the command-line,
that doesn't mean the user believe it is acceptable to get it running in
emulated mode. That's why I propose a new "emulate" flag, to allow
features to be enabled in emulated mode.

> 
> > But my question still stands: suppose we had x2apic emulation
> > implemented but for some reason it was painfully slow, we wouldn't
> > want to enable it by mistake. In this case, it would end up on
> > EMULATED_CPUID and not on SUPPORTED_CPUID, right?
> 
> IMHO we want to enable emulation only when explicitly requested...
> regardless of the emulation performance.

Well, x2apic is emulated by KVM, and it is on SUPPORTED_CPUID. Ditto for
tsc-deadline. Or are you talking specifically about instruction
emulation?

-- 
Eduardo