On Wed, 29 Aug 2018 at 17:10, Liran Alon <liran.alon@xxxxxxxxxx> wrote:
>
>
>
> > On 29 Aug 2018, at 8:54, Wanpeng Li <kernellwp@xxxxxxxxx> wrote:
> >
> > From: Wanpeng Li <wanpengli@xxxxxxxxxxx>
> >
> > Adds pv ipis out-of-bounds acess testing.
>
> I believe it is worth elaborating a bit more in commit message…
> At least refer to the kernel patch which fixes the issue you are testing here.
>
Will do in v2.
> >
> > Cc: Paolo Bonzini <pbonzini@xxxxxxxxxx>
> > Cc: Radim Krčmář <rkrcmar@xxxxxxxxxx>
> > Cc: Liran Alon <liran.alon@xxxxxxxxxx>
> > Signed-off-by: Wanpeng Li <wanpengli@xxxxxxxxxxx>
> > ---
> > x86/apic.c | 10 ++++++++++
> > 1 file changed, 10 insertions(+)
> >
> > diff --git a/x86/apic.c b/x86/apic.c
> > index c976e8d..460d7a5 100644
> > --- a/x86/apic.c
> > +++ b/x86/apic.c
> > @@ -566,6 +566,15 @@ static void test_apic_change_mode(void)
> > report("TMCCT should stay at zero", !apic_read(APIC_TMCCT));
> > }
> >
> > +void test_pv_ipi()
> > +{
> > + unsigned int nr = 10, ret;
>
> Define KVM_HC_SEND_IPI instead of using hard-coded “10”.
Ditto.
>
> > + unsigned long p1 = 0xFFFFFFFF, p2 = 0, p3 = 0xFFFFFFFF, p4 = 0x0;
>
> As arguments are documented in Documentation/virtual/kvm/hypercalls.txt from zero-index
> (a0,a1,a2 and etc.) it is preferred you use here same naming convention.
Ditto.
Regards,
Wanpeng Li
>
> > +
> > + asm volatile("vmcall" : "=a"(ret) :"a"(nr), "b"(p1), "c"(p2), "d"(p3), "S"(p4));
> > + report("PV IPIs testing", !ret);
> > +}
> > +
> > int main(void)
> > {
> > setup_vm();
> > @@ -581,6 +590,7 @@ int main(void)
> >
> > test_self_ipi();
> > test_physical_broadcast();
> > + test_pv_ipi();
> >
> > test_sti_nmi();
> > test_multiple_nmi();
> > --
> > 2.7.4
> >
>
> From semantics perspective:
> Reviewed-By: Liran Alon <liran.alon@xxxxxxxxxx>
>
>
