On Thu, Apr 11, 2019 at 8:16 PM Segher Boessenkool <segher@xxxxxxxxxxxxxxxxxxx> wrote: > > On Thu, Apr 11, 2019 at 03:43:48PM -0500, William Tambe wrote: > > Wouldn't the blockage insn prevent the compiler from re-arranging > > other following instructions ? > > Yes. > > > In fact, the two emit_insn() need to be seen as one instruction with > > the compiler free to re-arrange other instructions around it. > > Then why not *make* them one (rtl) insn? > It is not obvious how to do that for "mulsidi3" that is why a define_expand that emit two instructions is used. > > Segher
