Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 8 Jul 1999 11:14:30 +0200
From:      Eivind Eklund <eivind@freebsd.org>
To:        Kris Kennaway <kkennawa@physics.adelaide.edu.au>
Cc:        Peter Wemm <peter@netplex.com.au>, security@freebsd.org
Subject:   Re: Improved libcrypt ready for testing
Message-ID:  <19990708111429.E46370@bitbox.follo.net>
In-Reply-To: <Pine.OSF.4.10.9907080928090.20360-100000@bragg>; from Kris Kennaway on Thu, Jul 08, 1999 at 09:35:29AM %2B0930
References:  <19990707172115.D44021@bitbox.follo.net> <Pine.OSF.4.10.9907080928090.20360-100000@bragg>

next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Jul 08, 1999 at 09:35:29AM +0930, Kris Kennaway wrote:
> On Wed, 7 Jul 1999, Eivind Eklund wrote:
> 
> > If we want to support protocol-embedded authentication data properly,
> > we need at least the ability to store several different types of
> > hashes for each user in the password file, and the ability to store
> > clear-text passwords.
> 
> Storing cleartext passwords is easy enough - just define a minimal hash
> function which base64's the plaintext, and null salt function. 
> 
> I'll have to think about how multiple password hashes could best be
> implemented - any suggestions?

For the master password file itself, I guess we could just put several
hashes in the password field, separated by commas (which I don't think
are allowed in any of the present hashes).  I don't know how to fit
multiple hashes into the databases; I've not looked too carefully at
these.  

> 
> > We should also, IMO, be switching our default password file format to
> > SRP or similar - something that allow challenges against it without
> > being the cleartext.
> 
> I have the SRP reference implementation working at home - it requires changes
> to clients, though.

Does it require changes to clients in order to be used as a normal
password hash, not to do challenges against?  I can't remember
anything about it that would force that?

Just using it as a hash format in our password file would give the
userbase time to transition, and would poise us perfectly for further
use.

> This would probably best be integrated with a PAM module talking to
> the crypt backend (such a beast exists already, but I haven't tested
> it).

It would be very, very nice if we were able to support it :)

Eivind.


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-security" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990708111429.E46370>