On Fri, Aug 12, 2005 at 10:35:33PM +0200, Jan Hubicka wrote: > Unforutnately the actual size we want to limit is somewhat variable (for > kernel it is pretty small, in usual case few hounderd K is probably good > choice).
I disagree. If recusion is involved, even a few K may be unacceptable. That can mean the difference between being able to recurse 500k times and being able to recurse only 100k times. I would think that, in the extreme, we don't want *any* stack expansion, but in the norm a percentage expansion. r~