Re: Definition of Kernel BUG

Aron Griffis (agriffis@coat.com)
Wed, 10 Dec 1997 07:52:20 -0500 (EST)


On Tue, 9 Dec 1997, Royans K Tharakan wrote:

> To Begin with, is it a bug to hang the machine, even
> if I actually intended to do it. Which means that
> is it technically possible to hang a perfect bug-free
> linux machine even from root. I know one can't get
> bug free system, but I wish to know whether 2.0.32
> is buggy just because I can make it hang with root access.

Intention isn't the point. You can intentionally exploit a kernel bug to
expose it, but it is still a bug.

Basically, if you're doing something you should be able to do (even as
root) to test kernel code, and the machine hangs, you might have found a
bug. For example, copying directories while deleting them might
expose a bug in the filesystem code. (Well, it might have before
Bill Hawes came on the scene, anyway.)

However, you have permission as root to play with parts of the machine
that the kernel would normally disallow. Using root to go around kernel
protection to break the machine is not a bug; it simply means maybe you
shouldn't have root access... :-)

(I'll leave the actual bug report to someone more acquainted with the
kernel networking code to answer.)

+----------------------------------------->
| Aron Griffis (agriffis@coat.com)
| Burlington Coat Factory - Network Group
+----------------------------------------->