*BSD News Article 6677


Return to BSD News archive

Newsgroups: comp.unix.bsd
Path: sserve!manuel.anu.edu.au!munnari.oz.au!spool.mu.edu!uunet!mcsun!sunic!ugle.unit.no!ugle.unit.no!he
From: Havard.Eidnes@runit.sintef.no
Subject: Re: [386BSD] vmstat: version for ISA architecture out?
Message-ID: <1992Oct17.185212.9864@ugle.unit.no>
Originator: he@ugle.unit.no
Keywords: memory, vmstat
Sender: news@ugle.unit.no (NetNews Administrator)
Organization: University of Trondheim, Norway
References: <1bm3ntINNbe4@frigate.doc.ic.ac.uk> <48687@shamash.cdc.com>
Date: Sat, 17 Oct 92 18:52:12 GMT
Lines: 31

In article <48687@shamash.cdc.com> pbd@runyon.cim.cdc.com (Paul Dokas) writes:
>In article <1bm3ntINNbe4@frigate.doc.ic.ac.uk>, kd@doc.ic.ac.uk (Kostis Dryllerakis) writes:
>|> The copy available from the BSD reno has to be modified
>|> and I am not the expert to do so. Not much has to be done though; if 
>|> anybody has done so I would be obliged for a copy of the modified source!
>
>I too tried to get the reno version of vmstat to work on 386BSD, but 
>unfortuately I couldn't figure out how to make it work.  The place the I 
>couldn't figure out is with names.c.  It requires several kernel symbols 
>names that I just couldn't find in 386BSD.  The offending symbols were 
>_hp_dinit for hp300, _mbdinit and _ubdinit for vax, etc.

These are structs containing information on the devices connected to the
I/O bus(es) of these systems (MB = MassBus, UB = UniBus on VAX systems). 

I remember trying to get iostat from Net2 to work (no need to use the reno
source), and I sort of got it to compile (I had to take a couple of
shortcuts and make a number of possibly false assumptions), but after
digging a while through the kernel sources I concluded that the statistics
on seek times and disk transfers were not being recorded by the 386BSD
kernel or the disk drivers (whoever's responsability that is), so I
decided that further work on "iostat" was uninteresting until someone else
fixed that problem.

>My guess is that these symbols are part of the ATT (boo hiss) code that 
>was replaced during the Great Code Purge(tm).

Not so, it's a matter of porting kernel-dependent user-space source to a
new kernel (plus making the kernel record the statistics).

- Havard