From morgan@physics.ucla.edu  Sat Jan 20 14:13:39 1996
Return-Path: morgan@physics.ucla.edu
Received: from MIT.EDU (SOUTH-STATION-ANNEX.MIT.EDU [18.72.1.2]) by redhat.com (8.6.12/8.6.12) with SMTP id OAA31626 for <marc@redhat.com>; Sat, 20 Jan 1996 14:13:33 -0500
Received: from physics.ucla.edu by MIT.EDU with SMTP
	id AA19432; Sat, 20 Jan 96 14:11:43 EST
Received: from gluon.physics.ucla.edu by physics.ucla.edu (5.x/SMI-SVR4)
	id AA04661; Sat, 20 Jan 1996 11:13:35 -0800
Received: by gluon.physics.ucla.edu (SMI-8.6/SMI-SVR4)
	id LAA04459; Sat, 20 Jan 1996 11:10:45 -0800
From: morgan@physics.ucla.edu (Morgan)
Message-Id: <199601201910.LAA04459@gluon.physics.ucla.edu>
Subject: Re: Semantics of "optional"
To: linux-pam@MIT.EDU (Linux PAM)
Date: Sat, 20 Jan 1996 11:10:45 -0800 (PST)
X-Mailer: ELM [version 2.4 PL25]
Content-Type: text


Theodore Ts'o wrote:
> In-Reply-To: Volker Elling's message of Tue, 16 Jan 1996 ...
> Subject: Re: Semantics of "optional"
> 
> To quote from the PAM DCE:
> 
>         (a) `required': With this flag, the module failure results in the
>             PAM framework returning the error to the caller _after_
>             executing all other modules on the stack.  For the function to
>             be able to return success to the application all `required'
>             modules have to report success.  This flag is normally set when
>             authentication by this module is a _must_.

[bits trimmed]

> In other words, "required", "optional", and "sufficient" determine how
> the PAM library should handle module success and failures.  Remember,
> the order of modules in /etc/pam.conf is important.  The PAM library
> follows these rules:
> 
> If the module has a    ... then on module	... then on the module
> control flag of ...    returning SUCCESS	returning FAILURE
> 		
> ... Required		CONTINUE		return FAILURE
> 
[bits trimmed]

Mmm? The way I read it, it looks like the module is supposed to return
FAILURE but the PAM library is supposed to only return FAILURE *after*
it has executed all of the other modules... This would seem the more
secure thing to do....

Andrew
(morgan@physics.ucla.edu)

