[Bug rtl-optimization/84660] Combine doing wrong optimization for 64 bits with SHIFT_COUNT_TRUNCATED target

[wilson at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84660

--- Comment #5 from Jim Wilson <wilson at gcc dot gnu.org> ---
This is what the original patch I wrote looked like.  But then I tried to
optimize it a bit to eliminate the redundant nonzero_bits call, and that is
where I screwed up.  Then I got overwhelmed by work and haven't had a chance to
get back to it yet.

I can't self approve patches to combine, so it needs to be submitted to
gcc-patches for review.

It probably should be tested on at least one other target that isn't
SHIFT_COUNT_TRUNCATED.