HP-720 vs IBM-320 vs Sparc2

[Stefan Esser]

In article <1991Jun9.145014.14005 at alchemy.chem.utoronto.ca>, system at alchemy.chem.utoronto.ca (System Admin (Mike Peterson)) writes:
|> In article <1991Jun8.170028.156828 at rrz.uni-koeln.de> se at IKP.Uni-Koeln.DE, I write:
|> >
|> >Sad to say, at least the HP-PA version of HP-UX is derived from BSD!
|> 
|> Are you sure it is BSD-derived? ...
|> 					... I can't believe
|> HP would deliberately remove things from a BSD port that did not impact
|> SYS5 compliance (like 'renice', 'msgs', 'talk', ...), or would they ???
|> "Official" comments, HP? If it is BSD-based, why not still provide the
|> BSD ps (and other) commands in another directory, which can be moved to
|> the head of the $path/$PATH for users who prefer BSD options?

Since there exists a version of 'top' (a process monitor for BSD systems) 
working nice on our HP 9000/835, I wanted to have the renice feature 
normally found in 'top', but missing in the HP version. It was really quite 
simple, there is a function to access kmem, I wrote a similar one to write 
kmem, made 'top' setgid 'kmem' and it works rather nice ...
Adjtime can be simulated by a daemon process that patches a kernel 
variable (part of a HP-UX version of XNTP). 
If you look into these things, you find lots of BSD kernel structures hidden 
away and not simply accessible. 

Look at 'monitor' (in /usr/contrib/bin, not normally installed (?)).
Look at 'vmstat', and 'iostat'. BTW since I had lots of trouble with the 
machine when starts swapping and wanted to get better control over whats 
happening (HP-UX swaps out all SMALL, ACTIVE processes on our system,
when it starts swapping !!!), I took 'pstat' from our BSD Tahoe sources and 
compiled it with minor modifications. Would you ever expect this to be possible 
on a Sys.V kernel ???

Didn't you notice the way the kernel config works on the 9000/8xx (7.0)
(don't know about 3xx, 4xx) ? Its a config file quite similar to VAX BSD 4.3, 
with some braces and semicolons added. Look at the VM kernel parameters.
They look like BSD, not System V. I'm quite angry, because I miss lots of 
functionality. Its no fun to have a print spool system, that sometimes is 
capable of remote printing on BSD lpr systems, but sometimes doesn't 
send anything and at other times overflows the remote spool directory 
by sending the same file hundreds of times. 

Stefan
-- 
 Stefan Esser,  Institute of Nuclear Physics,  University of Cologne,  Germany
 se at IKP.Uni-Koeln.DE                                           [134.95.192.50]