This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Fix loop hoisting bug on ia32
- To: Jakub Jelinek <jakub at redhat dot com>
- Subject: Re: [PATCH] Fix loop hoisting bug on ia32
- From: Richard Henderson <rth at redhat dot com>
- Date: Tue, 28 Nov 2000 09:50:53 -0800
- Cc: gcc-patches at gcc dot gnu dot org
- References: <20001128144155.A557@sunsite.ms.mff.cuni.cz>
On Tue, Nov 28, 2000 at 02:41:55PM +0100, Jakub Jelinek wrote:
> Is it ok to commit this and perhaps if you want later optimize as described
> above (ideas on what should be the upper bound?)?
If you want to work on the optimization, you should coordinate with
Michael Hayes, who has a rewrite pending review of loop.c to be based
off curious things called "CFGs" and "dominators".
Revolutionary, I know.
> * loop.c (load_mems): Avoid using next_label to find end_label. If
> jumping outside of the loop (other than loop end), don't hoist MEMs
> out of loop.
>
> * gcc.c-torture/execute/loop-8.c: New test.
Ok.
r~