This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [RS6000] PR 40473, parameter debug location lists
- From: Alan Modra <amodra at bigpond dot net dot au>
- To: David Edelsohn <dje dot gcc at gmail dot com>
- Cc: gcc-patches <gcc-patches at gcc dot gnu dot org>
- Date: Tue, 22 Sep 2009 08:52:42 +0930
- Subject: Re: [RS6000] PR 40473, parameter debug location lists
- References: <303e1d290909210854v496770e2te0860318c8343170@mail.gmail.com>
On Mon, Sep 21, 2009 at 11:54:29AM -0400, David Edelsohn wrote:
> If the call to final() is changed to emit_insn_before_noloc(), then
> the instructions no longer are emitted and instead are generated as
> part of the function, which means they are scheduled. Your patch
> seems to make the -mno-sched-prolog flag useless.
No, we are well past scheduling. rs6000_output_function_prologue is
called from rest_of_handle_final via final_start_function.
> -mno-sched-prolog is an alias for -mno-sched-epilog. Is there a
> continuing need to allow the prologue and epilogue not to be
> scheduled?
I don't really know, but I suspect the answer is "yes". We do know
people are still using the option.. See
http://sourceware.org/bugzilla/show_bug.cgi?id=10231
--
Alan Modra
Australia Development Lab, IBM