[Bug rtl-optimization/87678] Redundant vmovss with -fPIC

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87678

--- Comment #9 from Segher Boessenkool <segher at gcc dot gnu.org> ---
Ah.  So you want this optimisation (which is currently done by LRA) to be done
by combine as well; it's not that the resulting assembler code for this
testcase
is worse than what you'd like to see.  And what you want combine to do is just
do a substitution (and no simplification, none with known values at least).

I agree with that.  I also want the latter to be done after every split pass,
for the insns touched by that split anyway :-)