*BSD News Article 22741


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!elroy.jpl.nasa.gov!usc!howland.reston.ans.net!vixen.cso.uiuc.edu!uxa.cso.uiuc.edu!rkb55989
From: rkb55989@uxa.cso.uiuc.edu (Rafal Boni)
Newsgroups: comp.os.386bsd.apps
Subject: Re: top-3.2 memory leak (Re: m_*bsd.c for top 3.0?)
Date: 24 Oct 1993 05:13:22 GMT
Organization: University of Illinois at Urbana
Lines: 48
Message-ID: <2ad2ti$8n@vixen.cso.uiuc.edu>
References: <29g3i2$o4q@homer.cs.mcgill.ca> <CEzt2r.1BF@luva.stgt.sub.org> <ASAMI.93Oct22220018@forgery.cs.berkeley.edu>
NNTP-Posting-Host: uxa.cso.uiuc.edu

asami@cs.berkeley.edu (Satoshi ASAMI) writes:

>(Note: orignial article was in .questions, but I moved it over because
> I thought it's more suitable here)

>In article <CEzt2r.1BF@luva.stgt.sub.org>
>        migieger@luva.stgt.sub.org (Michael Giegerich) writes:

> * It's included in top-3.2 (and top-3.3beta).
> * Important: you have to uncomment the #define PATCHED_KVM
> *            then compilation is straight forward.

>Great, I uncommented it and now top runs happily as ever.  However, I
>still have a problem (which existed in top-3.0 I got from agate's
>unsupported/from-ref directory too)---top seems to have a memory leak.
>It was about 400k when it started, and now (after running for about
>three hours), it stands at 1.4 megs.  It grows much slower than
>top-3.0, though.

>top-3.0 used to kill itself after a few hours---I haven't witnessed
>what happens to top-3.2, but my machine crashed a couple of days ago
>(with a long-running top of 4 megs or so).  I don't know if it's
>because of top, but I was away and nothing but xscreensaver and top
>was running (actively) on the machine.  This machine hasn't crashed
>for more than a week until then.

	I have seem this not only on BSD, but on a bunch of systems, 
	many of them being the university *NIX machines.... On the
	sequent symmetry that I have all my mail routed to, it grows
	much faster then on my 486 box... [But then again, the symmetry
	has gonzo memory, so it would probably take years to crash ;>]


>My system is 386DX-40, FreeBSD 1.0-Epsilon.  I have some problems with
>the disk (which I posted about a while ago, and got two replies), and
>the keyboard locks once in a while, but otherwise it's pretty stable.

>Many thanks for the great system!

>---
>                                                   o o   Satoshi Asami
>  Form follows Function.                            ^
>                                               (asami@CS.Berkeley.EDU)
>  Tel: (510)547-0336 (home)       Computer Science Div., Dept. of EECS
>       (510)642-1845 (office)     University of California at Berkeley

						--rafal