NetBSD Problem Report #6891

Received: (qmail 14711 invoked from network); 27 Jan 1999 21:33:47 -0000
Message-Id: <199901272133.WAA25902@gnyf.t.dk>
Date: Wed, 27 Jan 1999 22:33:45 +0100 (CET)
From: "Soren S. Jorvang" <soren@t.dk>
Reply-To: soren@t.dk
To: gnats-bugs@gnats.netbsd.org
Cc: soren@t.dk
Subject: garbage memory usage fields in proc structures of kernel threads
X-Send-Pr-Version: 3.95

>Number:         6891
>Category:       kern
>Synopsis:       garbage memory usage fields in proc structures of kernel threads
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    kern-bug-people
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Jan 27 13:35:00 +0000 1999
>Closed-Date:    Thu Aug 05 05:39:00 +0000 1999
>Last-Modified:  Thu Aug 05 05:39:13 +0000 1999
>Originator:     Soren S. Jorvang
>Release:        NetBSD-current 19980127
>Organization:
Soren
>Environment:
System: NetBSD meow 1.3I i386

>Description:
Memory usage fields in the proc structures of kernel threads appear uninitialized.
I vaguely seem to recall that this was due to some unfinished work.
For now, I suggest just zeroing those fields.

>How-To-Repeat:
$ ps axu 
USER       PID %CPU %MEM   VSZ  RSS TT  STAT STARTED       TIME COMMAND
soren     1396  0.0  0.4   364  224 p1  R+   10:27PM    0:00.00 ps -axu 
root         1  0.0  0.3   332  212 ??  Ss    7:55PM    0:00.03 /sbin/init -s 
root         2  0.0 -34.3     0 -22360 ??  DL    7:55PM    0:00.11 (pagedaemon)
root         3  0.0 -34.3     0 -22360 ??  DL    7:55PM    0:00.04 (reaper)
[...]
root         0  0.0 -34.3     0 -22372 ??  DLs   7:55PM    0:00.00 (swapper)

>Fix:
Don't know.
>Release-Note:
>Audit-Trail:
State-Changed-From-To: open->analyzed 
State-Changed-By: simonb 
State-Changed-When: Sun Feb 28 02:08:26 PST 1999 
State-Changed-Why:  
From Chuck Cranor on tech-kern: 

>    i looked at this on the i386 a while back but i haven't had 
> time to figure out how to fix it.   the problem on the i386 is that 
> pmap_kenter() adds kernel mappings without bumping up the kernel 
> pmap's pm_stats.resident_count  [it can't bump it as-is because 
> it doesn't have a lock on the kernel pmap's simplelock.] 
>   
>    then, later when mappings established with pmap_kenter() are 
> removed by pmap_remove() it subtracts off resident_count irrespective 
> of if it was bumped up by pmap_enter() or not.   so, one soultion 
> would be to change these functions to always adjust resident_count. 
> another solution could be to change it so that resident_count doesn't 
> change for things entered with pmap_kenter(). 
> 
>     the thing that confuses me is you are seeing this on the mips 
> and pmap_kenter is a PMAP_NEW thing that i've been playing with. 
> if you are getting negative numbers without PMAP_NEW then there must 
> be something somewhere that is entering PTE-level mappings into the 
> pmap without bumping up resident_count? 

Adding a hack to `ps' is not the "correct" way to fix this. 
State-Changed-From-To: analyzed->closed 
State-Changed-By: thorpej 
State-Changed-When: Wed Aug 4 22:39:00 PDT 1999 
State-Changed-Why:  
This has been fixed. 
>Unformatted:

NetBSD Home
NetBSD PR Database Search

(Contact us) $NetBSD: query-full-pr,v 1.51 2026/08/10 02:28:17 riastradh Exp $
$NetBSD: gnats_config.sh,v 1.10 2026/05/13 22:00:09 riastradh Exp $
Copyright © 1994-2026 The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.