This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Destrictiveness of expansion and tree inlining deadlock?
- From: Jan Hubicka <jh at suse dot cz>
- To: Richard Henderson <rth at redhat dot com>, Jan Hubicka <jh at suse dot cz>,Mark Mitchell <mark at codesourcery dot com>, gcc-patches at gcc dot gnu dot org
- Date: Mon, 3 Mar 2003 19:25:31 +0100
- Subject: Re: Destrictiveness of expansion and tree inlining deadlock?
- References: <20030222102031.GB9783@kam.mff.cuni.cz> <4060000.1046041692@localhost> <20030226181316.GG27482@kam.mff.cuni.cz> <1046412993.4832.82.camel@doubledemon.codesourcery.com> <20030303155857.GH21640@kam.mff.cuni.cz> <20030303181105.GE20178@redhat.com>
> On Mon, Mar 03, 2003 at 04:58:57PM +0100, Jan Hubicka wrote:
> > * calls.c (rtx_for_function_call): Take the address as an argument
> > (expand_call): Do not modify the expression.
>
> Yes, this is ok. I should have noticed this when I
> approved the patch that you're fixing. :-(
No prob. At least I know more about trees now :)
With this patch unit-at-a-time works fine on x86-64 too. Do you see any
problem with enabling it for -O2 (or perhaps only for -O3).
At -O2 it don't do much at the present, but I hope soon it will do the
local function optimization and inlining functions called once. THat is
about 1% saving overall. With new GC scheme I measured no slowdown in
compiling binutils.
Honza
>
>
>
> r~