Re: [PATCH] KVM: SVM: Add tracepoint for cr_interception

[Haiwei Li]

On 20/9/4 20:01, Vitaly Kuznetsov wrote:
> Haiwei Li <lihaiwei.kernel@xxxxxxxxx> writes:
>
> > From: Haiwei Li <lihaiwei@xxxxxxxxxxx>
> >
> > Add trace_kvm_cr_write and trace_kvm_cr_read for svm.
> >
> > Signed-off-by: Haiwei Li <lihaiwei@xxxxxxxxxxx>
> > ---
> >    arch/x86/kvm/svm/svm.c | 2 ++
> >    1 file changed, 2 insertions(+)
> >
> > diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
> > index 03dd7bac8034..2c6dea48ba62 100644
> > --- a/arch/x86/kvm/svm/svm.c
> > +++ b/arch/x86/kvm/svm/svm.c
> > @@ -2261,6 +2261,7 @@ static int cr_interception(struct vcpu_svm *svm)
>
> There are two special cases when we go to emulate_on_interception() and
> these won't be logged but I don't think this is a must.
>
> >    	if (cr >= 16) { /* mov to cr */
> >    		cr -= 16;
> >    		val = kvm_register_read(&svm->vcpu, reg);
> > +		trace_kvm_cr_write(cr, val);
> >    		switch (cr) {
> >    		case 0:
> >    			if (!check_selective_cr0_intercepted(svm, val))
> > @@ -2306,6 +2307,7 @@ static int cr_interception(struct vcpu_svm *svm)
> >    			return 1;
> >    		}
> >    		kvm_register_write(&svm->vcpu, reg, val);
> > +		trace_kvm_cr_read(cr, val);
>
> The 'default:' case above does 'return 1;' so we won't get the trace but
> I understand you put trace_kvm_cr_read() here so you can log the
> returned 'val', #UD should be clearly visible.
>
> >    	}
> >    	return kvm_complete_insn_gp(&svm->vcpu, err);
> >    }
> > --
> > 2.18.4
>
> Reviewed-by: Vitaly Kuznetsov <vkuznets@xxxxxxxxxx>

Thanks a lot.

>