Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uflorida!haven!umbc3!otter.UMBC.EDU!alex
From: alex@otter.UMBC.EDU (alex)
Newsgroups: comp.sys.att
Subject: Re: 3b1 Swap Space (sizeof)
Message-ID: <1805@umbc3.UMBC.EDU>
Date: 16 Mar 89 20:28:44 GMT
References: <914@koko.UUCP>
Sender: newspost@umbc3.UMBC.EDU
Distribution: usa
Lines: 29

In article <914@koko.UUCP>, jb@koko.UUCP (John Birchfield) writes:

> How much swap space is provided as a default on the PC7300 (3b1)?
> I have a 3b1 with 2 megs of memory (around 1.7 meg available) and
> a 67 meg hard disk.  I wrote a program to allocate memory until it
> failed and it managed to suck up around 2.3 megs before malloc failed.
> It seems to me I should be able to do a little better than that.

	Well, no actually. The 2.3 megs is limited by the process memory map,
which only knows about the (virtual) addresses between 0x80000 and 0x300000.
processes live in virtual memory, which is mapped like this:

	     0	->  80000	Kernal
	 80000	-> 300000	Process user space
	300000	-> 380000	Shared library space
	380000	-> 400000	Dynamic Kernal space

All processes live in a world like the one above, and the kernal keeps the
worlds from conflicting. Although a process doesn't have to use all of the
space allocated to it (unused pages don't exist, sortof.), Multiple copies
of the VM map exist, and thats what the swap spacve is for.

	Malloc stops at 2.3 meg because the user world reserves 100K or so 
for stack space. (2ef000 comes to mind as an upper limit, but I'm not sure)

					:alex
Alex Crain
Systems Programmer			alex@umbc3.umbc.edu
Univ Md Baltimore County		nerwin!alex@umbc3.umbc.edu