This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: max-stack-var-size
- From: Paul Brook <paul at nowt dot org>
- To: Steven Bosscher <s dot bosscher at student dot tudelft dot nl>
- Cc: fortran at gcc dot gnu dot org
- Date: Tue, 11 Nov 2003 23:25:22 +0000
- Subject: Re: max-stack-var-size
- References: <4195D82C2DB1D211B9910008C7C9B06F01F373D7@lr0nt3.lr.tudelft.nl> <200311101855.08231.paul@nowt.org> <200311110029.27372.s.bosscher@student.tudelft.nl>
I'm also CC'g the list. Others may have views on what the optimal behavior
is.
On Monday 10 November 2003 11:29 pm, you wrote:
> There's another thing I just found out: on PPC, if the stackframe is >16
> bits offsets, the generated code is _extremely_ poor....
> Perhaps the patch should be reconsidered.
I'd say this was an argument for the patch, abeit maybe with a smaller
default. The current code is better than before the patch in this respect.
I don't know if there is any justification for the current default limits.
Given the behavior you describe, maybe 4k (the page size on x86, dunno about
PPC) would be a better default?
At one point I started implemeting a per-stackframe limit, but it started
getting messy so I didn't bother.
Paul