Loop strip mining to decreases safepoint poll frequency

[Roland Westrelin]

> I would like to keep UseCLS to pretend to be "independent". E.g. UseCLS enables
> safepoints in counted loops, and LSM>1 additionally strip mines the loops.
> Regardless what the actual implementation is doing, the saner interface seems
> better if done like that.

So what I suggested in my previous email is ok, right?

> Do we foresee loop strip mining to be useful for something else besides
> safepoint poll improvements? If so, untying their respective options make even
> more sense.

I don't think we do at this point. The advantage of only using
LoopStripMiningIter in c2 itself is that the logic that checks both
UseCountedLoopSafepoints and LoopStripMiningIter is not duplicated every
time c2 has to check what kind of safepoint elimination needs to be
performed. To be fair there are only a few such checks which also means
it should be straightforward to revisit that logic if needed.

Roland.