[RFA:] reorg.c: fix target/7042 for main trunk
law@redhat.com
law@redhat.com
Sun Jun 16 21:51:00 GMT 2002
In message <Pine.BSF.4.30.0206161700320.96786-100000@dair.pair.com>, Hans-Peter
Nilsson writes:
> (Replying mostly to correct the subject; it's PR 7042 not 7041.
> I fixed that before committing.)
>
> On Sun, 16 Jun 2002 law@redhat.com wrote:
>
> > In message <200206161912.VAA09230@ignucius.axis.se>, Hans-Peter Nilsson
> > writes:
> > > Not to paper over a bug, but yes, it's generally the preferred
> > > solution. (Please keep in mind that the CRIS port predates
> > > prologue-and-epilogue as RTL.)
> > Really? Prologue/epilouge as RTL has been around for more than a decade
> > now.
>
> Really. Grep says there were no epilogue or prologue patterns
> in gcc 2.1.
Here's the change which added the ability to have prologues and epilogues
as RTL. I was close -- that support wil be a decade old at the end of this
month.
Mon Jun 29 12:04:57 1992 Tom Wood (wood@dg-rtp.dg.com)
* toplev.c (rest_of_compilation): Thread the prologue and epilogue
code immediately after reloading is done.
* sched.c (schedule_insns): Reposition the prologue and epilogue
notes as needed.
[ ...]
* m88k.h (FUNCTION_END_PROLOGUE, FUNCTION_BEGIN_EPILOGUE): Define.
(DELAY_SLOTS_FOR_EPILOGUE, ELIGIBLE_FOR_EPILOGUE_DELAY): Delete.
* m88k.md (prologue, epilogue, profiler, locate1, locate2): New
patterns for representing prologues and epilogues.
I don't think the CRIS port existed back in 1992, or if it did, it
certainly wasn't in the public sources.
jeff
More information about the Gcc-patches
mailing list