[Bug target/117974] RISC-V: VSETVL hoisting across branch

[law at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117974

Jeffrey A. Law <law at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |law at gcc dot gnu.org

--- Comment #5 from Jeffrey A. Law <law at gcc dot gnu.org> ---
So FWIW, we'd looked at something similar in get_ref from x264, essentially
pondering if we should unroll that pixel_avg loop and our conclusion was that
with the carried dependency into the vsetvl it probably wasn't worth the
effort. 

What's starting to rattle around in my brain is the for a loop, if the count is
unknown, then we probably don't want to unroll as that's much more likely to
trigger the VL=0 problem you're trying to avoid.