Re: [PATCH v2] hvm/svm: Implement Debug events

[Alexandru Stefan ISAILA <aisaila@bitdefender.com>]

On Ma, 2018-03-20 at 16:14 +0000, Andrew Cooper wrote:
> On 20/03/18 16:05, Jan Beulich wrote:
> >
> > >
> > > >
> > > > >
> > > > > On 20.03.18 at 16:53, <andrew.cooper3@citrix.com> wrote:
> > > On 20/03/18 09:40, Alexandru Isaila wrote:
> > > >
> > > > @@ -2438,16 +2439,15 @@ static bool
> > > > svm_get_pending_event(struct vcpu *v, struct x86_event *info)
> > > >      return true;
> > > >  }
> > > >
> > > > -static void svm_propagate_intr(struct vcpu *v, unsigned long
> > > > insn_len)
> > > > +static void svm_propagate_intr(unsigned long insn_len, int16_t
> > > > vector,
> > > uint8_t type)
> > >
> > > Hmm - not sure where the old unsigned long came from, but it
> > > isn't
> > > really correct.  Also, as this function no longer propagates the
> > > contents of the vmcb, it is now mis-named.
> > >
> > > Please could you delete this function and use:
> > >
> > > diff --git a/xen/include/asm-x86/hvm/hvm.h b/xen/include/asm-
> > > x86/hvm/hvm.h
> > > index 2376ed6..843dafe 100644
> > > --- a/xen/include/asm-x86/hvm/hvm.h
> > > +++ b/xen/include/asm-x86/hvm/hvm.h
> > > @@ -407,6 +407,19 @@ void hvm_migrate_pirqs(struct vcpu *v);
> > >
> > >  void hvm_inject_event(const struct x86_event *event);
> > >
> > > +static inline void hvm_inject_exception(
> > > +    unsigned int vector, unsigned int type, unsigned int
> > > insn_len)
> > > +{
> > > +    struct x86_event event = {
> > > +        .vector = vector,
> > > +        .type = type,
> > > +        .insn_len = insn_len,
> > > +        .error_code = X86_EVENT_NO_EC,
> > > +    };
> > > +
> > > +    hvm_inject_event(&event);
> > > +}
> > > +
> > >  static inline void hvm_inject_hw_exception(unsigned int vector,
> > > int
> > > errcode)
> > >  {
> > >      struct x86_event event = {
> > >
> > > as a new common helper.  (I'm not terribly happy with the name,
> > > but I
> > > can't think of a better alternative, seeing as it is needed for
> > > both
> > > software and hardware exceptions.)
> > We'll need some better name though - "exception" together with
> > an unconditional X86_EVENT_NO_EC is not really a reasonable
> > combination.
> Alternatively, keep the current name, extend the parameter list with
> an
> "int error_code" and have the new callers pass X86_EVENT_NO_EC in?
>
> ~Andrew
>
Keeping the name with an extra parameter sounds better to me.
Then hvm_inject_exception is it.

~Alex

________________________
This email was scanned by Bitdefender
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel