On Mon, Apr 27, 2026 at 06:00:46AM +0530, Ritesh Harjani wrote:
> Gautam Menghani <[email protected]> writes:
> 
> > From: Gautam Menghani <[email protected]>
> >
> > The macro kvm_trace_symbol_exit is used for providing the mappings
> > for the trap vectors and their names. Add mapping for H_FAC_UNAVAIL so that
> > trap reason is displayed as string instead of a vector number when using
> > the kvm_guest_exit tracepoint.
> >
> > Signed-off-by: Gautam Menghani <[email protected]>
> > ---
> >  arch/powerpc/kvm/trace_book3s.h | 3 ++-
> >  1 file changed, 2 insertions(+), 1 deletion(-)
> >
> > diff --git a/arch/powerpc/kvm/trace_book3s.h 
> > b/arch/powerpc/kvm/trace_book3s.h
> > index 9260ddbd557f..eafbaeb5a9db 100644
> > --- a/arch/powerpc/kvm/trace_book3s.h
> > +++ b/arch/powerpc/kvm/trace_book3s.h
> > @@ -28,6 +28,7 @@
> >     {0xea0, "H_VIRT"}, \
> >     {0xf00, "PERFMON"}, \
> >     {0xf20, "ALTIVEC"}, \
> > -   {0xf40, "VSX"}
> > +   {0xf40, "VSX"}, \
> > +   {0xf80, "H_FAC_UNAVAIL"},
> 
> So I ended up looking into other trace_symbols too and I don't think we
> should have a trailing comma here. It may cause some issues silently in
> trace-cmd or other places. I remember going through it when I was doing
> some other tracing work earlier but I don't recollect it now.
> 
> Can we just remove the trailing comma and send a v2 please?

Good catch, thanks. Will send a v2


> 
> btw, 0XF80 is correct as per "arch/powerpc/include/asm/kvm_asm.h"
> 
> #define BOOK3S_INTERRUPT_H_FAC_UNAVAIL        0xf80
> 
> -ritesh
> 

Reply via email to