[Fwd: Arla 0.20 and linux 2.2-pre9, misc problems, probably n ot kernel specific]

Mon Jan 25 19:19:52 CET 1999

> -----Original Message-----
> From: Derek Atkins [mailto:warlord at MIT.EDU]
> Sent: Monday, January 25, 1999 9:13 AM
> To: Nathan Neulinger
> Cc: kenh at cmf.nrl.navy.mil
> Subject: Re: [Fwd: Arla 0.20 and linux 2.2-pre9, misc 
> problems, probably
> not kernel specific]
> 
> 
> Linux-AFS replaces the 'initgroups' system call, does Arla does the
> same?  I don't know how the Linux-2.2 version of AFS (which is
> Transarc 3.5) will behave.  I doubt it is a kernel change.

I figured it did something like that. 

I verified that a simple initgroups() test fails on 2.2 and works on 2.0.35
(well, behaves the way I want anyway).

Is your replacement for initgroups() something you can release, given that
it was not intrinsic to transarc's code? Arla already preloads getcwd.so,
should be easy enough to add another one to do initgroups().

> Check Arla on 2.0.35 and see if it works there.
> 
> -derek
> 
> Nathan Neulinger <nneul at umr.edu> writes:
> 
> > 
> > This is a multi-part message in MIME format.
> > --------------9E78052BF545843A7AB710B7
> > Content-Type: text/plain; charset=us-ascii
> > Content-Transfer-Encoding: 7bit
> > 
> > Hey guys. 
> > 
> > I just sent this to the arla list. Got any suggestions on 
> how to clean
> > this up in krb5 (or elsewhere) in some way that is not just a really
> > ugly hack?
> > 
> > This will no doubt affect use with transarc's client as well.
> > 
> > How is this handled on other architectures? Or is initgroups() just
> > broken everywhere?
> > 
> > -- Nathan
> > 
> > ------------------------------------------------------------
> > Nathan Neulinger                       EMail:  nneul at umr.edu
> > University of Missouri - Rolla         Phone: (573) 341-4841
> > Computing Services                       Fax: (573) 341-4216
> > --------------9E78052BF545843A7AB710B7
> > Content-Type: message/rfc822
> > Content-Transfer-Encoding: 7bit
> > Content-Disposition: inline
> > 
> > Received: from umr.edu (hermes.cc.umr.edu [131.151.1.68]) 
> by umr-mail01.cc.umr.edu with SMTP (Microsoft Exchange 
> Internet Mail Service Version 5.5.2232.9)
> > 	id DQ2R7SPM; Sun, 24 Jan 1999 15:23:56 -0600
> > Received: from sundance.stacken.kth.se 
> (sundance.stacken.kth.se [130.237.234.41]) via ESMTP by 
> hermes.cc.umr.edu (8.8.7/R.4.20) id PAA28277; Sun, 24 Jan 
> 1999 15:23:43 -0600 (CST)
> > Received: (from majordom at localhost)
> > 	by sundance.stacken.kth.se (8.8.8/8.8.8) id WAA12481
> > 	for arla-drinkers-list; Sun, 24 Jan 1999 22:12:25 +0100 (MET)
> > Received: from umr.edu (hermes.cc.umr.edu [131.151.1.68])
> > 	by sundance.stacken.kth.se (8.8.8/8.8.8) with ESMTP id WAA12477;
> > 	Sun, 24 Jan 1999 22:12:20 +0100 (MET)
> > Received: from umr-mail01.cc.umr.edu (umr-mail01.cc.umr.edu 
> [131.151.37.121]) via ESMTP by hermes.cc.umr.edu 
> (8.8.7/R.4.20) id PAA24905; Sun, 24 Jan 1999 15:12:18 -0600 (CST)
> > Received: by umr-mail01.cc.umr.edu with Internet Mail 
> Service (5.5.2232.9)
> > 	id <DQ2R7S3Q>; Sun, 24 Jan 1999 15:12:27 -0600
> > Message-ID: 
> <9DA8D24B915BD1118911006094516EAF019C7E77 at umr-mail02.cc.umr.edu>
> > From: "Neulinger, Nathan R." <nneul at umr.edu>
> > To: "'Magnus Ahltorp'" <map at stacken.kth.se>
> > Cc: arla-drinkers at stacken.kth.se
> > Subject: RE: Arla 0.20 and linux 2.2-pre9, misc problems, 
> probably not ker
> > 	nel  	specific
> > Date: Sun, 24 Jan 1999 15:12:17 -0600
> > MIME-Version: 1.0
> > X-Mailer: Internet Mail Service (5.5.2232.9)
> > Content-Type: text/plain;
> > 	charset="ISO-8859-1"
> > Sender: owner-arla-drinkers at stacken.kth.se
> > Precedence: bulk
> > 
> > > -----Original Message-----
> > > From: Magnus Ahltorp [mailto:map at stacken.kth.se]
> > > Sent: Sunday, January 24, 1999 2:58 PM
> > > To: Neulinger, Nathan R.
> > > Cc: arla-drinkers at stacken.kth.se
> > > Subject: Re: Arla 0.20 and linux 2.2-pre9, misc problems, 
> probably not
> > > kernel specific
> > > 
> > > 
> > > > It seems to be working ok for me. 
> > > 
> > > Great.
> > > 
> > > > I do notice one problem. It appears that Arla treats PAG's 
> > > differently than
> > > > transarc's AFS. If I do an su to root while logged in, with 
> > > a token, I no
> > > > longer have the token, or the pag, in the su'd session:
> > > 
> > > That is because your su throws away the secondary groups:
> > > 
> > > > infinity(49)>id
> > > > uid=5879(nneul) gid=5000(afsuser) 
> groups=33536,32512,5000(afsuser)
> > > > infinity(50)>su - 
> > > > Password: 
> > > > [root at infinity /root]# id
> > > > uid=0(root) gid=0(root)
> > > > groups=0(root),1(bin),2(daemon),3(sys),4(adm),6(disk),10(wheel)
> > > 
> > > Here, the 33536,32512 has been thrown away, and therefore you are
> > > assumed to be in the default PAG.
> > 
> > Right, but it didn't do that before. It's possible that it 
> is a change in
> > the kernel from
> > 2.0.35 to 2.2.
> > 
> > I have another machine running 2.0.35 with transarc's code, 
> and the same
> > su/ksu. Both perform correctly on that machine. 
> > 
> > > 
> > > > Yes, I am using a mixture of transarc excutables and AFS, 
> > > but only to
> > > > demonstrate the problem. Might be good if arla included a 
> > > tokens executable.
> > > > Probably is easy enough to write, in fact, I believe I have 
> > > the code lying
> > > > around somewhere to do it.
> > > 
> > > If you use the kth-krb kerberos distribution, your klist 
> is able to do
> > > this (klist -T).
> > 
> > Nope, running krb5 with hornstein's patches.
> > 
> > -- Nathan
> > 
> > --------------9E78052BF545843A7AB710B7--
> > 
> 
> -- 
>        Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
>        Member, MIT Student Information Processing Board  (SIPB)
>        URL: http://web.mit.edu/warlord/      PP-ASEL      N1NWH
>        warlord at MIT.EDU                        PGP key available
> 