[Bug rtl-optimization/123070] [16 Regression] pr90178.c -m32 starts to fail since r16-5975

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123070

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2025-12-09
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(insn 37 39 38 10 (set (reg/i:SI 0 ax)
        (reg/f:SI 101 [ <retval> ])) "/app/example.cpp":12:1 -1
     (nil))

I would have assume RA would try its hardest to get r101 into EAX. But I don't
see that happening any more.
Instead of 2 bb containing a set of r101 being 0, there is only one.
For 64bit argument passing via registers it seems RA is still able to get put
that return register into EAX. This is getting out of my league now.