Memory overcommitment and guessing about stack size (was Re: [Python-Dev] stack check on Unix: any suggestions?)

Peter Funk pf@artcom-gmbh.de
Wed, 30 Aug 2000 10:39:39 +0200 (MEST)


--ELM967624779-23128-0_
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

Hi,

Any attempts to *reliable* predict the amount of virtual memory (stack+heap)
available to a process are *DOOMED TO FAIL* by principle on any unixoid
System.

Some of you might have missed all those repeated threads about virtual memory
allocation and the overcommitment strategy in the various Linux groups.  

M.-A. Lemburg:
> Skip Montanaro wrote:
> > 
> >     MAL> The C program can suck memory in large chunks and consume great
> >     MAL> amounts of stack, it just doesn't dump core... (don't know what I'm
> >     MAL> doing wrong here).
> > 
> > Are you overwriting all that memory you malloc with random junk?  If not,
> > the stack and the heap may have collided but not corrupted each other.
> 
> Not random junk, but all 1s:
[...]

For anyone interested in more details, I attach an email written by
Linus Thorvalds in the thread 'Re: Linux is 'creating' memory ?!'
on 'comp.os.linux.developmen.apps' on Mar 20th 1995, since I was
unable to locate this article on Deja (you know).

Regards, Peter
-- 
Peter Funk, Oldenburger Str.86, D-27777 Ganderkesee, Germany, Fax:+49 4222950260
office: +49 421 20419-0 (ArtCom GmbH, Grazer Str.8, D-28359 Bremen)

--ELM967624779-23128-0_
Content-Type: text/plain; charset=ISO-8859-1
Content-Disposition: attachment; filename=lt
Content-Description: News-Posting by Linus Torvalds in 1995
Content-Transfer-Encoding: 7bit