outstanding regressions

[Richard Henderson]

On Thu, Aug 01, 2002 at 05:22:27PM -0400, David Edelsohn wrote:
>         lfd 0,3360(9)		; load a(-10,10) into $f0
>         stfd 13,3360(9)		; store 0d0 into a(-10,10)
> 
> I know I left my causality around here somewhere...

Damn.

(insn 66 65 68 (nil) (set (mem/s/j:DF (reg/f:SI 158) [0 (* u) S8 A64])
        (reg:DF 159)) -1 (nil)
    (nil))

(insn 94 93 95 (nil) (set (reg:DF 175)
        (mem/j:DF (reg/f:SI 174) [0 a S8 A64])) -1 (nil)
    (nil))

Note the aliasing data: the inlined rtl still contains the U parameter
from the subroutine.  I'll need to nix that during inlining...


r~