This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
[wwwdocs] Indirect inlining (PR 9079)
- From: Jan Hubicka <jh at suse dot cz>
- To: Martin Jambor <mjambor at suse dot cz>
- Cc: GCC Patches <gcc-patches at gcc dot gnu dot org>, Jan Hubicka <jh at suse dot cz>, Kenneth Zadeck <zadeck at naturalbridge dot com>, Razya Ladelsky <RAZYA at il dot ibm dot com>, Paolo Carlini <paolo dot carlini at oracle dot com>
- Date: Fri, 25 Jul 2008 14:06:29 +0200
- Subject: [wwwdocs] Indirect inlining (PR 9079)
- References: <20080715194347.569852675@virgil.suse.cz>
Hi,
while I am at it, what about this entry?
Also while reading the docs I noticed that you probably want to add code
to inlining of always inlines and functions called once to do
propagation and add the extra edges, or we would miss the hooks passed
to always inline functions, right?
(for functions called once we are doomed as we do not perform any
further inlining, but I would add it for consistency. We might want to
loop inlininf of function called once and simple function inliner until
they stabilize, but I doubt it makes real difference)
Honza
@@ -94,7 +98,11 @@
given that all the values are known at compile time and the ratio between
the new array size and the original switch branches does not exceed
the parameter <code>--param switch-conversion-max-branch-ratio</code>
- (default is eight). </li>
+ (default is eight).</li>
+ <li>Inliner now also inline indirect calls that are discovered to be known
+ at compile time thanks to previous inlining. This feature is enabled
+ at <code>-O2</code> optimization level and can be controlled via
+ <code>-findirect-inlining</code> command line option.</li>
</ul>
<h2>New Languages and Language specific improvements</h2>