Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/5/84; site aicchi.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxn!ihnp4!aicchi!ignatz
From: ignatz@aicchi.UUCP (Ihnat)
Newsgroups: net.unix-wizards
Subject: Re: Xenix Crash? Trivial...
Message-ID: <526@aicchi.UUCP>
Date: Wed, 24-Jul-85 00:00:54 EDT
Article-I.D.: aicchi.526
Posted: Wed Jul 24 00:00:54 1985
Date-Received: Sat, 27-Jul-85 01:50:23 EDT
References: <518@aicchi.UUCP> <831@mulga.OZ>
Reply-To: ignatz@aicchi.UUCP (Ihnat)
Distribution: net
Organization: Analysts International Corp; Chicago Branch
Lines: 22
Keywords: Xenix 3.0, adb, crash
Summary: You can't step syscalls, but should you crash the system? Noooo...

In article <831@mulga.OZ> of net.unix-wizards, Anthony Tang responds to
my comment in <518@aicchi.UUCP> :

>>Gosh, you have to go into a program?  I just discovered tonight, while
>>'adb'ing a stripped object (DON'T ask why I would want to do it--object
>>licensees have to do many things a respectable source licensee wouldn't
>>*dream* of), that I can trash the kernel by simply trying to single-step
>>through a system call!

with:

>In the manual of XENIX 1.0 for the AT, it read  : "System calls cannot
>be single-stepped"

Mmmm-hmmmm.  No argument; they (syscalls) shouldn't be.  BUT--should any
random, non-priviliged user--malicious or not--be able to trash the system
with such a trivial mistake?  I think not...oh, well...
-- 
	Dave Ihnat
	Analysts International Corporation
	(312) 882-4673
	ihnp4!aicchi!ignatz