[Bug rtl-optimization/125727] [16/17 Regression] ICE on valid code at -O{s,2,3} with "-fno-tree-forwprop -fno-tree-pre -fno-tree-vrp -fno-tree-dce" on x86_64-linux-gnu: in simplify_subreg, at simplify-rtx.cc:8465 since r16-7261

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125727

--- Comment #5 from Drea Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Jeffrey A. Law from comment #4)
> So the combine change just makes it more likely that we try 4 insn
> combinations.  So I suspect it's just exposing a problem elsewhere.

Right, we are starting to play `whack a mole` with the bad handling of the
clobber inside combine as we improve other parts of combine or other areas
before combine.