Does Ultrix 4.0 finally have a secure /dev/*mem ?

Kurt J. Lidl lidl at eng.umd.edu
Wed Aug 1 22:39:33 AEST 1990


In article <438 at e2big.mko.dec.com> francus at e2big.mko.dec.com (Yoseff Francus) writes:
>In article <1990Jul31.154912.22096 at watcgl.waterloo.edu> idallen at watcgl.waterloo.edu (Ian! D. Allen [CGL]) writes:
>>Is memory still world-readable under Ultrix 4.0?
>>-- 
>>-IAN! (Ian! D. Allen) idallen at watcgl.uwaterloo.ca idallen at watcgl.waterloo.edu
>> [129.97.128.64]  Computer Graphics Lab/University of Waterloo/Ontario/Canada
>
>No it is not. /dev/mem now has a 640 protection with ownership
>of roots and group mem. mem is a new group and has groupid 6.
>
>francus at metsny.mko.dec.com

Yes, and therein lies part of the problem.  EVERY other BSD-derived
system that I have worked on has group kmem as gid 2.  In fact, this
is one of the "standard" group/gid pairs (kmem/2) that the BSD folks
request that everyone have around in the 4.3 (or was it the Tahoe?)
release docs.  There are a few other groups in there that I wish were
standard too...

To the best of my knowledge, BSD had the idea for a group kmem first,
and as such could make reasonable requests for it.  If DEC is going to
bother to implement it, why not make it a little closer to the true
intent?

This implies to me just one more gratuitious change to a supposedly
BSD type system.  Why?  Does DEC enjoy making my life harder?  I'm
begining to think so...

(By the way, the gid of 6 is news on my system -- I thought it was
really bloodly funny to find my kernel, ps and other programs
setuid to news when I put our "standard" /etc/group file on our
first Ultrix 4.0 machine... )
--
/* Kurt J. Lidl (lidl at eng.umd.edu) | Unix is the answer, but only if you */
/* UUCP: uunet!eng.umd.edu!lidl    | phrase the question very carefully. */



More information about the Comp.unix.ultrix mailing list