This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: line insn notes cleanups (haifa-sched and modulo-sched)
- From: Ian Lance Taylor <iant at google dot com>
- To: Tehila Meyzels <TEHILA at il dot ibm dot com>
- Cc: Mike Stump <mrs at apple dot com>, gcc-patches at gcc dot gnu dot org, stevenb dot gcc at gmail dot com, vmakarov at redhat dot com
- Date: 25 Jul 2006 18:16:51 -0700
- Subject: Re: line insn notes cleanups (haifa-sched and modulo-sched)
- References: <OF923E83A5.94AA0CF9-ONC22571B0.006DB23E-C22571B0.006E7DD7@il.ibm.com>
Tehila Meyzels <TEHILA@il.ibm.com> writes:
> @@ -828,38 +815,12 @@
> return false;
>
> if (!loop->single_exit)
> - {
> - if (dump_file)
> - {
> - rtx line_note = find_line_note (BB_END (loop->header));
> + {
> + return false;
> + }
>
> - fprintf (dump_file, "SMS loop many exits ");
> - if (line_note)
> - {
> - expanded_location xloc;
> - NOTE_EXPANDED_LOCATION (xloc, line_note);
> - fprintf (dump_file, " %s %d (file, line)\n",
> - xloc.file, xloc.line);
> - }
> - }
> - return false;
> - }
> -
> if (! SIMPLE_SMS_LOOP_P (loop) && ! loop_single_full_bb_p (loop))
> {
> - if (dump_file)
> - {
> - rtx line_note = find_line_note (BB_END (loop->header));
> -
> - fprintf (dump_file, "SMS loop many BBs. ");
> - if (line_note)
> - {
> - expanded_location xloc;
> - NOTE_EXPANDED_LOCATION (xloc, line_note);
> - fprintf (dump_file, " %s %d (file, line)\n",
> - xloc.file, xloc.line);
> - }
> - }
> return false;
> }
>
In these hunks, and two others later in the same file, you are
removing debugging output that doesn't have anything to do with line
notes. Is this intentional?
This patch looks OK to me. However, it is not stage 3 material. So
I'll approve it for stage 1.
But before checking it in, please run the gdb testsuite before and
after the patch, and make sure there are no new regressions. I'm
pretty sure we are OK in terms of debug info, but we don't test for it
very much in the gcc testsuite. The gdb testsuite is not
comprehensive for gcc debug info generation, but it will catch any
gross errors.
Thanks.
Ian