This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Potential bug in stmt.c (expand_value_return)
- To: Joern Rennecke <amylaar at cygnus dot co dot uk>
- Subject: Re: Potential bug in stmt.c (expand_value_return)
- From: Richard Henderson <rth at cygnus dot com>
- Date: Wed, 2 Feb 2000 14:55:46 -0800
- Cc: Joern Rennecke <amylaar at pasanda dot cygnus dot co dot uk>, Denis Chertykov <denisc at overta dot ru>, gcc at gcc dot gnu dot org, gcc-patches at gcc dot gnu dot org
- References: <20000202132847.B10382@cygnus.com> <200002022230.WAA01412@phal.cygnus.co.uk>
On Wed, Feb 02, 2000 at 10:30:51PM +0000, Joern Rennecke wrote:
> I can't search through the whole compiler now, but I know that
> reload_combine needs it, and the register scavenging for out-of
> range branchs on the sh (config/sh/sh.c:gen_block_redirect).
>
> USEs are also much easier to digest for peepholes than looking into
> a global array when you hit a return insn.
Anything post flow2 can use resource.c's find_free_register.
Nothing could be easier than that.
For reload_combine ... I dunno. Probably all that code *should*
be moved after flow2. It would reduce the amount of hackery
involved in some of the guessing games.
*sigh* I guess I have to put the use back until these things
are fixed properly.
r~