On Monday, November 26, 2012 04:32:39 PM Greg KH wrote:
> On Mon, Nov 26, 2012 at 04:23:57PM -0800, Dmitry Torokhov wrote:
> > Hi Greg,
> >
> > For some reason it still didn't go through to our corporate mail server
> > but I see it on LKML.
>
> Good.
>
> > On Mon, Nov 26, 2012 at 04:03:04PM -0800, Greg KH wrote:
> > > On Wed, Nov 07, 2012 at 10:43:03AM -0800, George Zhang wrote:
> > > > +static inline struct vmci_handle VMCI_MAKE_HANDLE(vmci_id cid,
> > > > vmci_id rid) +{
> > > > + struct vmci_handle h;
> > > > + h.context = cid;
> > > > + h.resource = rid;
> > > > + return h;
> > > > +}
> > >
> > > You return a structure on the stack that just went away? Yeah, I know
> > > it's an inline, but come on, that's not ok.
> >
> > This is certainly OK even if it is not inline, we return the _value_,
> > not the pointer to the stacki memory. And yes, the structure is 64 bit
> > value so it is returned in registers.
>
> Even on a 32bit processor?
I thought it would, but it looks like it won't. Maybe we'll just switch it
to a macro with C99 style initializators to keep the same semantic but
avoid the question.
> Also, you already have another function that
> does this same thing, so having 2 functions in the same patch seems odd,
> right?
Yes, you can say that it is probably a bit excessive.
OK, now that we are on the same page we'll go and fix the issues.
Thanks,
Dmitry
_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization
