SH sibcalls
Richard Henderson
rth@redhat.com
Thu Nov 23 11:57:00 GMT 2000
On Thu, Nov 23, 2000 at 06:36:09PM +0000, Joern Rennecke wrote:
> However, you emit the sibcall insn as call insn, so the scheduler doesn't
> know that there is a change of control flow.
> Or will there be a BARRIER after the sibcall to tell sched2 what to do?
There will be a barrier. Moreover, SIBLING_CALL_P is set on the
call_insn. The CFG is correctly computed for sibling calls.
r~
More information about the Gcc-patches
mailing list