This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Updated version of -Winline warnings patch
- From: Jan Hubicka <jh at suse dot cz>
- To: Gábor Lóki <loki at inf dot u-szeged dot hu>
- Cc: gcc-patches at gcc dot gnu dot org, jh at suse dot cz
- Date: Thu, 8 Jan 2004 15:58:54 +0100
- Subject: Re: Updated version of -Winline warnings patch
- References: <3FFD5E2F.6000200@inf.u-szeged.hu>
> >I've managed to forget to commit the patch for proper -Winline
> >warning. Since
> >this is a regression relative to older GCC versions that used to
> >produce
> >usefull messages, I've updated the patch and will commit it tomorrow.
> >I will
> >also prepare tree-ssa branch version.
> >
> >Bootstrapped/regtested i686-pc-gnu-linux
>
> According to the results measured with CSiBE this patch is responsible
> for about 0.5% code size increase for all targets but i386.
> After I looked through the patch I had found the problematic part.
>
> As you see below, the "node1->callers->inline_call" condition was
> converted into "node1->callers->inline_failed", not into a
> "!node1->callers->inline_failed".
>
> Is this intentional or is it a typo?!
It is typo. Thanks for noticing it! (perhaps adding some function
abstraction here would help)
I tested it on Gerald's testcase and obviously it has no sites to be
inlined once. Does fixing it fix the code size growth?
Honza