This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Ping... Re: -fconserve-stack
- From: Jan Hubicka <hubicka at ucw dot cz>
- To: Mark Mitchell <mark at codesourcery dot com>
- Cc: Jan Hubicka <hubicka at ucw dot cz>, Jan Hubicka <jh at suse dot cz>, gcc-patches at gcc dot gnu dot org, andi at firstfloor dot org
- Date: Mon, 1 Sep 2008 12:26:12 +0200
- Subject: Re: Ping... Re: -fconserve-stack
- References: <20070909234944.GD29122@kam.mff.cuni.cz> <20070912075314.GA14746@atrey.karlin.mff.cuni.cz> <46E83CA7.4040504@codesourcery.com>
> Jan Hubicka wrote:
> >>
> >> * invoke.texi (-fconserve-stack): Document.
> >> * opts.c (decode_options): Handle conserve_stack.
> >> * common.opt (fconvserve_stack): New.
>
> >> + @item -fconserve-stack
> >> + @opindex fconserve-stack
> >> + Disable optimizations noticeably increasing stack usage. This is useful for
> >> + stack restricted environments, like kernels. Currently the option imply
> >> + setting @option{large-stack-frame} to 16 and @option{large-stack-fram-growth}
> >> + to 200.
>
> This could be written more clearly:
>
> "Attempt to minimize stack usage. The compiler will attempt to use less
> stack space, even if that makes the program smaller. This option
> implies setting the @option{large-stack-frame} parameter to 16 and the
> @option{large-stack-frame-growth} parameter to 200. For more details,
> see..."
>
> That last needs to be a cross-reference to the documentation for those
> parameters. Please adjust the TeXinfo to make that work.
>
> OK with that change.
Hi,
I was just about to ping this patch, since Andi has run into issues with
stack frame growth on kernel. I must've missed this approval, sorry for
that. Given that we are in stage 3 now, would be possible to still
re-test the patch and commit? I would like to get this in for kernel to
use.
Andi, given experience from your testcase, I guess we should change that
twofold growth to about 40% and increase large stack frame size to let
say 100 bytes?
Honza
>
> Thanks,
>
> --
> Mark Mitchell
> CodeSourcery
> mark@codesourcery.com
> (650) 331-3385 x713