Re: [patch 2/3] KVM: x86: introduce ioctl to allow frequency hypercalls

[Paolo Bonzini]

On 23/02/2017 17:48, Radim Krcmar wrote:
> Doing it manually doesn't sound useful outside of testing ...
> is DPDK actually being used "dynamically"?
> (I thought that the setup is decided when the host boots.)

Or at the very least when the guest boots.  I agree that KVM_ENABLE_CAP
is enough.  However, changing KVM_CPUID_FEATURES should be handled by
userspace (KVM doesn't know about 0x400000xx leaves at all).