Garance A Drosehn wrote:
On 7/7/13 1:36 PM, A.J. Kehoe IV (Nanoman) wrote:
I commissioned Derek to come up with a solution by either updating
Steven's patch or by devising a new method. To paraphrase Derek's comments:

-----BEGIN PARAPHRASIS-----
I did some research into what other *BSDs are doing. OpenBSD and NetBSD
use the algorithm name, a comma, and then the number of rounds:

http://www.openbsd.org/cgi-bin/man.cgi?query=login.conf&sektion=5

localcipher=blowfish,6

http://netbsd.gw.com/cgi-bin/man-cgi?passwd.conf+5+NetBSD-current

localcipher=blowfish,6

To me, this isn't a good way to do it because we'd need special
rules to parse this extra field out of the previously unstructured
data.  This parsing would be algorithm dependant.

To comment only on this point, I do not think it is a significant
issue.  If OpenBSD and NetBSD are already doing this, then whatever
parsing issues are already being addressed by users on those OS's.

I think there is a significant advantage in using something that
they are already using.  Now, if they say "Wow, was this a bad
idea!", then obviously I wouldn't want to add it.  But if their
security is better with this feature, and if *they* don't have
major regrets with using it, then I think we should consider it.

I don't think the method used by OpenBSD and NetBSD is a bad idea, but there 
are a couple reasons why I prefer Derek's method:

1. Convention.  OpenBSD and NetBSD call it "blowfish", whereas FreeBSD calls it "blf", 
and others call it "bcrypt".  Modular Crypt Format, on the other hand, is pretty consistent across 
various operating systems, languages, et cetera.

2. Implementation.  Derek's method would be a relatively minimal change, 
whereas the other method would be a more invasive change with more things that 
could go wrong.

I'd certainly want to consider other ideas too.  But I don't think
we should cross this idea off the list just because it would be
too much extra effort *if* we were the only OS which used it.  I
run both FreeBSD and OpenBSD systems, and for people like me it
will be more effort if different BSD's use incompatible methods
to achieve better password security.  You won't be saving me any
effort, you'll only be adding to the effort I already have.

[admittedly that isn't much effort. :) ]

Maybe our OpenBSD and NetBSD colleagues will prefer Derek's idea too?  As Derek 
has demonstrated, it's fairly simple to implement while maintaining 
compatibility with the existing method.

--
A.J. Kehoe IV (Nanoman)     |  /"\  ASCII Ribbon Campaign
Nanoman's Company           |  \ /   - No HTML/RTF in E-mail
E-mail: [email protected]  |   X    - No proprietary attachments
WWW: http://www.nanoman.ca/ |  / \   - Respect for open standards

Attachment: smime.p7s
Description: S/MIME cryptographic signature

Reply via email to