🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: comp.sys.next.sysadmin
Article #922

SUMMARY: Re: Two Questions From a New NeXT Sysadmin

#922
From: Dave Kinchlea
Date: Wed, 19 Aug 1992 19:12
52 lines
2637 bytes
Howdy:
As promised, the summary to my Two Questions. 

No suprises really, the -console is a process which switches ownership
to whoever is logged in at the console and lives as long as the
machine is up, or until killed (either the Wrong Way: kill, or the
Right Way: login as exit). 

Everybody said basically the same thing re: the ps process: ps is setuid; QED.
I am not sure I like this behaviour, it is (of course) not limited to
ps but is true for all setuid programs. I looked into who else does
this and my (very) limited study found:

Sun running SunOS 4.1.1 and NextStep appear to both actually change the user
field (real user?). I almost didn't notice Sun's behaviour as its ps
is setgid and (of course) the owner of the process was unaffected.

SGI running IRIX 4.0.1, Sequent running DYNIX 3.1.0 and
DYNIX/ptx 1.3.0b2, CDC running EP/IX 1.4.2, 
IBM running AIX 3.2.1: show the invocator as the owner of
the process.

[RISC/OS (don't know the version) appears to also show the invocator as
the owner but it is hard to say for sure as the only setuid program
tested was ps and it is also setgid.]

As NextStep is, I believe, Mach based and SunOS 4.1.1 is (still?) BSD
based as is DYNIX and EP/IX, this does not appear to have much to do
with historical ways of doing things. (I am sure someone will correct
me if I am wrong)

So, is the observed behaviour deliberate (ie: did the designers of
NextStep and/or Mach expect or want the owner to change with setuid
programs) or is this a consequence of something I am not taking into
consideration? (Just for the record, this is really just for the sake
of curiosity, I am not losing any sleep over this issue ;-()


If there is anybody who wants a better explanation than the one I have
provided, for either of the two questions, I can certainly send along
the replies I received.

cheerskinch

Dave Kinchlea            
email:	 	kinch@valve.heart.rri.uwo.ca
Snail-mail:	System Administrator/Research Programmer
		John P. Robarts Research Institute.
		London, Ont, Canada
A Real Live Voice: 663-5777-x4406

".... 2 to the power 759 million to 1 and falling ..." HHGTG
Technical Details
Message-ID: <KINCH.92Aug19141235@valve.heart.rri.uwo.ca>
Path: arc1-ca.pugleaf.net!sparky!uunet!utcsri!torn!newshost.uwo.ca!newshost.uwo.ca!kinch

Show Full Headers
Complete Article Headers
Newsgroups: comp.sys.next.sysadmin
Path: arc1-ca.pugleaf.net!sparky!uunet!utcsri!torn!newshost.uwo.ca!newshost.uwo.ca!kinch
From: kinch@valve.heart.rri.uwo.ca (Dave Kinchlea)
Subject: SUMMARY: Re: Two Questions From a New NeXT Sysadmin
Organization: /u/kinch/.organization
Date: Wed, 19 Aug 1992 19:12:35 GMT
Message-ID: <KINCH.92Aug19141235@valve.heart.rri.uwo.ca>
Sender: news@julian.uwo.ca (USENET News System)
Nntp-Posting-Host: valve.heart.rri.uwo.ca
Lines: 52