[Bug c/71785] Computed gotos are mostly optimized away

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71785

--- Comment #3 from Segher Boessenkool <segher at gcc dot gnu.org> ---
So, compgotos does its job just fine.  It has duplicated the indirect
jump to all its predecessors.  One of those is a block that just does
an add (the op++ from the testcase) and then jumps to the indirect jump.
I'd expect bb-reorder to copy the resulting block -- it's just a few
insns -- but it doesn't.  Investigating.