On Mon, Nov 11, 2013 at 8:39 PM, William ML Leslie <[email protected]> wrote: > > Has anyone that has been following the discussion figured out what > we're supposed to do about caller-saved registers? Since they aren't > llvm registers, I can't figure out how I'm supposed to know where they > end up on the stack across a function call.
I think the current approach is to arrange for them to be aliased in such a way that the compiler has to spill them at procedure call boundaries. That's effectively what the Henderson approach does. _______________________________________________ bitc-dev mailing list [email protected] http://www.coyotos.org/mailman/listinfo/bitc-dev
