On Sun, Mar 29, 2020 at 7:17 PM Wei Liu <[email protected]> wrote:

> On Sun, Mar 29, 2020 at 11:46:08AM +0530, Simran Singhal wrote:
> > diff --git a/xen/arch/x86/oprofile/op_model_athlon.c
> b/xen/arch/x86/oprofile/op_model_athlon.c
> > index 5c48f868ae..2bc0d04a1f 100644
> > --- a/xen/arch/x86/oprofile/op_model_athlon.c
> > +++ b/xen/arch/x86/oprofile/op_model_athlon.c
> > @@ -343,9 +343,8 @@ static int athlon_check_ctrs(unsigned int const cpu,
> >               }
> >       }
> >
> > -     ovf = handle_ibs(mode, regs);
> >       /* See op_model_ppro.c */
> > -     return ovf;
> > +     return handle_ibs(mode, regs);
>
> Hmm... ovf can potentially be set in the for loop before this hunk, but
> then immediately get overwritten by the function call. I bet the ovf = 1
> line is the reason why you didn't remove ovf out right.
>
> I think you can perhaps just remove ovf here. It would make any
> difference logically.
>
> Other changes look correct to me.
>

Thanks for your feedback.
Ah! I agree on "ovf = 1" is the reason I decided to leave ovf as it is.
I'll remove ovf now and send the new version of this patch.

Thanks
Simran


>
> Wei.
>

Reply via email to