Hi,
> >> > Index: tree-vrp.c
> >> > ===================================================================
> >> > --- tree-vrp.c (revision 173703)
> >> > +++ tree-vrp.c (working copy)
> >> > @@ -2273,7 +2273,12 @@ extract_range_from_binary_expr (value_ra
> >> > {
> >> > /* For pointer types, we are really only interested in asserting
> >> > whether the expression evaluates to non-NULL. */
> >> > - if (range_is_nonnull (&vr0) || range_is_nonnull (&vr1))
> >> > + if (flag_delete_null_pointer_checks && nowrap_type_p
> >> > (expr_type))
> >>
> >> the latter would always return true
> >>
> >> Btw, I guess you'll "miscompile" a load of code that is strictly
> >> undefined. So I'm not sure we want to do this against our users ...
> >
> > Probably not, at least unless the user explicitly asks for it -- for
> > example,
> > we could have -fdelete-null-pointer-checks=2. In fact, it might be a good
> > idea
> > to implement this flag anyway, since some current uses of
> > flag_delete_null_pointer_checks
> > can lead to "miscompilations" when user makes an error in their code and
> > would
> > probably appreciate more having their program crash.
> >
> >> Oh, and of course it's even wrong. I thing it needs &&
> >> !range_includes_zero (&vr1) (which we probably don't have). The
> >> offset may be 0 and NULL + 0
> >> is still NULL.
> >
> > actually, the result of NULL + 0 is undefined (pointer arithmetics is only
> > defined
> > for pointers to actual objects, and NULL cannot point to one).
>
> It's maybe undefined in C, but is it undefined in the middle-end? Thus,
> are you sure we never generate it from (void *)((uintptr_t)p + obfuscated_0)?
> I'm sure we simply fold that to p + obfuscated_0.
if we do, we definitely should not -- the only point of such a construction is
to bypass the pointer arithmetics restrictions,
Zdenek
