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: Tehila Meyzels <TEHILA at il dot ibm dot com>
- To: Ian Lance Taylor <iant at google dot com>
- Cc: gcc-patches at gcc dot gnu dot org, stevenb dot gcc at gmail dot com, vmakarov at redhat dot com
- Date: Mon, 31 Jul 2006 13:37:00 +0300
- Subject: Re: line insn notes cleanups (haifa-sched and modulo-sched)
Ian Lance Taylor <iant@google.com> wrote:
>> Those hunks use the function 'find_line_note', which content is:
>> find_line_note (rtx insn)
>> {
>> if (no_line_numbers)
>> return 0;
>>
>> for (; insn; insn = PREV_INSN (insn))
>> if (NOTE_P (insn)
>> && NOTE_LINE_NUMBER (insn) >= 0)
>> break;
>>
>> return insn;
>> }
>>
>> Since it's returning the insn that is a line number note
(NOTE_LINE_NUMBER
>> (insn) >= 0), we decided to eliminate it.
>> Therefore, I think I'll keep it eliminated.
>My question was not about the code related to find_line_note. I agree
>that should go. However, this code prints "SMS loop many exits" and
>"SMS loop many BBs" regardless of whether a line note is found or not.
>There is similar code in the other hunks I mentioned. Why should
>these fprintf statements go? That seems like an unrelated change.
>Ian
You're completely right. Thanks for the comment.
We're going to replace the other printings with code uses the insn locators
instead of LINE_NUMBER_NOTES.
(need to find out how to do it).
Tehila.