From mboxrd@z Thu Jan  1 00:00:00 1970
From: "Russ Cox" <rsc@plan9.bell-labs.com>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] python and lua
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Message-Id: <20010108041308.3BCC3199D7@mail.cse.psu.edu>
Date: Sun,  7 Jan 2001 23:12:50 -0500
Topicbox-Message-UUID: 476d4944-eac9-11e9-9e20-41e7f4b1d025

> What kind of situations?
> I asked this because I've never seen it.

Usually a terminal booted from local disk
can't sustain more than a few minutes of local disk swapping
before the window system or kfs swaps in a bogus page
and dies.  

Try running a few instances of the following program

	for(i in `{seq 10})
		memeat -x 10000000

and sit back.  This doesn't seem to happen when swapping
to a network file server.  Perhaps the timing is different.

I suspect something in the x86 memory management code,
although presotto says the carreras had similar problems.
I spent a day this summer trying to find anything, but
like everyone else came up empty handed.  

Instead, I moved more memory in from another machine
and turned off swapping.  On a laptop, of course,
this is less of an option, but I don't run my laptop
with swapping enabled either.  I'd much rather see 
"no physical memory" and type ^t^t k.

Russ


g% cat memeat.c
#include <u.h>
#include <libc.h>

void
main(int argc, char **argv)
{
	int xflag=0;
	int i, n;
	char *p;

	ARGBEGIN{
	case 'x':
		xflag=1;
	}ARGEND;
	n = atoi(argv[0]);
	p = malloc(n);
	memset(p, 0, n);
	i=0;
	for(;;) {
		sleep(1000+1000*(rand()%8));
		if(xflag)	memset(p, i++, n);
	}
}
g%