Re: Even more inliner cleanups

["H.J. Lu" <hjl.tools@gmail.com>]

On Mon, Apr 18, 2011 at 11:44 AM, Jan Hubicka <hubicka@ucw.cz> wrote:
> Hi,
> this patch is result of another pass through ipa-inline.c cleaning up
> confused and outdated comments.
>
> I also fixed accounting of growth caused by the small function inlining.
> We do not want to base it on program size before flattening, but afterwards.
> Just like always inlines are not (and should not) be accounted.
>
> I also broked out logic testing whether function should be inlined
> as inlined once into want_inline_function_called_once_p.
> Since it is now common to see that inlining functions called once
> makes a mistake inlining something terribly unimportant that prevents
> it to inline hot call later, I added simple logic to inline first
> hot edges and cold edges just afterwards.
>
> Finally I forgot to update recursive call inlining to be driven by frequencies
> when profile is missing.
>
> Bootstrapped/regtested x86_64-linux and comitted.
>
>        * ipa-inline.c: Fix comment typos; do not inline gt-ipa-inline.h
>        (want_inline_function_called_once_p): Break out the logic from
>        ipa_inline.
>        (edge_badness): Ensure that profile is not misupdated.
>        (lookup_recursive_calls): Prioritize by call frequencies.
>        (inline_small_functions): Move program size estimates here;
>        actually process whole queue even when unit growth has been
>        met. (to properly compute inline_failed reasons and for the
>        case unit size decrease.) Revisit comments on recursive
>        inlining.
>        (ipa_inline): Remove unit summary code; first inline hot calls
>        of functions called once, cold calls next.
>        (order, nnodes): Remove unused variables.
>        * Makefile.in (ipa-inline.o): No longer depent on ggc files.
>        (GTFILES): Remove ipa-inline.c
>        * sel-sched.c (fill_insns): Silence uninitialized var warning.
>

This caused:

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48671


-- 
H.J.