On Wed, Jan 03, 2018 at 11:45:55AM -0500, Nathan Sidwell wrote:
> On 01/03/2018 11:31 AM, Marek Polacek wrote:
> > Here we are crashing because cxx_fold_indirect_ref got a POINTER_PLUS_EXPR
> > with offset > signed HOST_WIDE_INT and we tried to convert it to sHWI.
> >
> > The matching code in fold_indirect_ref_1 uses uHWIs so I've followed suit.
> > But that code now also uses poly_uint64 and I'm not sure if any of the
> > constexpr.c code should use it, too. But this patch fixes the ICE.
>
> This doesn't look right to me, but it doesn't help that the test case
> invokes UB.
Hmm, like I said, I just followed fold_indirect_ref_1 so I was pretty confident
that this is not a totally wrong thing to do ;).
> > +typedef int V __attribute__ ((__vector_size__ (16)));
> > +V a;
> > +
> > +int
> > +main ()
> > +{
> > + reinterpret_cast <int *> (&a)[-1] += 1;
> > +}
>
> one could make this code well formed with (I think)
>
> typedef int V __attribute__ ((__vector_size__ (16)));
> V a[2];
>
> int main ()
> {
> return reinterpret_cast <int *> (&a[1])[-1];
> }
>
> That should access the final element of the a[0] vector.
Unfortunately, this doesn't trigger the ICE.
Marek
