[Bug target/63764] [5 Regression] ICE: in verify_ssa, at tree-ssa.c:939

["jakub at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63764

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
(In reply to Markus Trippelsdorf from comment #4)
> This specific ICE started with r217125.
> But it was failing before that already:
> 
> trippels@gcc2-power8 ~ % < test2.ii
> void
> fn1 ()
> {
>   __attribute__ ((altivec (vector__))) float saijplus16
>       = (__attribute__ ((altivec (vector__))) float){ 0.0, 0.0, 0.0, 0.0 };
>   ((__attribute__ ((altivec (vector__))) float) saijplus16)[0] = 0;
> }
> 
> trippels@gcc2-power8 ~ % /home/trippels/gcc_bisect/usr/local/bin/gcc
> -mcpu=power8 -O2 -c test2.ii
> test2.ii: In function âvoid fn1()â:
> test2.ii:7:1: internal compiler error: in execute_todo, at passes.c:1842
>  }
>  ^
> 0x1088b0df execute_todo
>         ../../gcc/gcc/passes.c:1842
> Please submit a full bug report,
> 
> Will bisect this later.

Yeah, 4.6 20100620 rejects that (no support for vector array subscripts), 4.8
ICEs, don't have anything in between.
That said, if something regressed because of  r217125, it better should be
analyzed what changed, because hopefully that change should affect only
-fcheck-pointer-bounds.  If not, something went very wrong.