On Fri, Aug 23, 2019 at 11:15:18AM +0200, Vitaly Kuznetsov wrote: > Sean Christopherson <sean.j.christopherson@xxxxxxxxx> writes: > > > Move the stat.mmio_exits update into x86_emulate_instruction(). This is > > both a bug fix, e.g. the current update flows will incorrectly increment > > mmio_exits on emulation failure, and a preparatory change to set the > > stage for eliminating EMULATE_DONE and company. > > > > Signed-off-by: Sean Christopherson <sean.j.christopherson@xxxxxxxxx> > > Reviewed-by: Vitaly Kuznetsov <vkuznets@xxxxxxxxxx> > > This, however, makes me wonder why this is handled in x86-specific code > in the first place, can we just count KVM_EXIT_MMIO exits when handling > KVM_RUN? struct kvm_vcpu_stat is arch specific. At a glance, everyone is counting similar things, but often in slightly different ways. E.g. PowerPC and ARM count MMIO exits, but PowerPC counts all exits, ARM has separate counters for in-kernel vs. userspace, and x86 counts only userspace.
