[jdev] ruby xmppd, and Psi

[Eric Will]

On Thu, Sep 18, 2008 at 8:59 AM, Norman Rasmussen
<norman at rasmussen.co.za> wrote:
> So I can only assume the response values are incorrect, because the other
> values are identical.  Does your password has any non-ascii characters that
> might be being encoded as utf-8 correctly?  Can you force the server to
> generate the same nonce for both clients?  (It would require hacking at the
> DIGEST-MD5 code, but it would help validate that the response is being
> generated correctly.

The password is alnum only. Using a hard-coded nonce of
"9be91df13f8159809d392ed8dc96bdc2":

Psi:

-> TLS established
-> <?xml version="1.0"?><stream:stream
xmlns:stream="http://etherx.jabber.org/streams" version="1.0"
xmlns="jabber:client" to="malkier.net" xml:lang="en"
xmlns:xml="http://www.w3.org/XML/1998/namespace" >
<- <?xml version='1.0'?><stream:stream xmlns='jabber:client'
xmlns:stream='http://etherx.jabber.org/streams' from='malkier.net'
id='f76c54806898a90dc1f12e78796f69c9' version='1.0'>
<- <stream:features><mechanisms
xmlns='urn:ietf:params:xml:ns:xmpp-sasl'><mechanism>DIGEST-MD5</mechanism><mechanism>PLAIN</mechanism></mechanisms></stream:features>
-> <auth xmlns="urn:ietf:params:xml:ns:xmpp-sasl" mechanism="DIGEST-MD5" />
<- <challenge xmlns='urn:ietf:params:xml:ns:xmpp-sasl'>cmVhbG09bWFsa2llci5uZXQsbm9uY2U9IjliZTkxZGYxM2Y4MTU5ODA5ZDM5MmVkOGRjOTZiZGMyIixxb3A9ImF1dGgiLGNoYXJzZXQ9dXRmLTgsYWxnb3JpdGhtPW1kNS1zZXNz</challenge>
-> <response xmlns="urn:ietf:params:xml:ns:xmpp-sasl">dXNlcm5hbWU9InJha2F1ciIscmVhbG09Im1hbGtpZXIubmV0Iixub25jZT0iOWJlOTFkZjEzZjgxNTk4MDlkMzkyZWQ4ZGM5NmJkYzIiLGNub25jZT0idVBvQWVnN2J1eHJ2UFpGMnkxakpjRUxsN3NlQXFGQW1KR0phMVZZdGtVWT0iLG5jPTAwMDAwMDAxLGRpZ2VzdC11cmk9InhtcHAvbWFsa2llci5uZXQiLHFvcD1hdXRoLHJlc3BvbnNlPTMxNjBhODJhMWY4NGY0NmM2YTIwNDcxMzFlNGFmNzdlLGNoYXJzZXQ9dXRmLTg=</response>
<- <failure xmlns='urn:ietf:params:xml:ns:xmpp-sasl'><not-authorized/></failure>
<- </stream:stream>

Or, decoded:

<- challenge: realm=malkier.net,nonce="9be91df13f8159809d392ed8dc96bdc2",qop="auth",charset=utf-8,algorithm=md5-sess
-> response: username="rakaur",realm="malkier.net",nonce="9be91df13f8159809d392ed8dc96bdc2",cnonce="uPoAeg7buxrvPZF2y1jJcELl7seAqFAmJGJa1VYtkUY=",nc=00000001,digest-uri="xmpp/malkier.net",qop=auth,response=3160a82a1f84f46c6a2047131e4af77e,charset=utf-8
<- failure

Gajim (and Digsby, FYI):

-> TLS established
-> <?xml version='1.0'?><stream:stream xmlns="jabber:client"
to="malkier.net" version="1.0"
xmlns:stream="http://etherx.jabber.org/streams" >
<- <?xml version='1.0'?><stream:stream xmlns='jabber:client'
xmlns:stream='http://etherx.jabber.org/streams' from='malkier.net'
id='205d1918670c0a28dc4a8c8402e05032' version='1.0'>
<- <stream:features><mechanisms
xmlns='urn:ietf:params:xml:ns:xmpp-sasl'><mechanism>DIGEST-MD5</mechanism><mechanism>PLAIN</mechanism></mechanisms></stream:features>
-> <auth xmlns="urn:ietf:params:xml:ns:xmpp-sasl" mechanism="DIGEST-MD5" />
<- <challenge xmlns='urn:ietf:params:xml:ns:xmpp-sasl'>cmVhbG09bWFsa2llci5uZXQsbm9uY2U9IjliZTkxZGYxM2Y4MTU5ODA5ZDM5MmVkOGRjOTZiZGMyIixxb3A9ImF1dGgiLGNoYXJzZXQ9dXRmLTgsYWxnb3JpdGhtPW1kNS1zZXNz</challenge>
-> <response xmlns="urn:ietf:params:xml:ns:xmpp-sasl">Y2hhcnNldD11dGYtOCx1c2VybmFtZT0icmFrYXVyIixyZWFsbT0ibWFsa2llci5uZXQiLG5vbmNlPSI5YmU5MWRmMTNmODE1OTgwOWQzOTJlZDhkYzk2YmRjMiIsbmM9MDAwMDAwMDEsY25vbmNlPSIxZWE1ZTkwNTAyMTAxYTcwOGZlOTQ3MjMwOTM1ZWYwZTQ2MDYzZjIxM2ExMmNhMmRjIixkaWdlc3QtdXJpPSJ4bXBwL21hbGtpZXIubmV0IixyZXNwb25zZT04ZWI5YTNiNDkyNzFiNWJlZDk3Y2M2YTgzOTg4YWJhMyxxb3A9YXV0aA==</response>
<- <challenge xmlns='urn:ietf:params:xml:ns:xmpp-sasl'>cnNwYXV0aD1lYjRkYjNmMjM5N2E0NDQzY2FhNTIxYmY4ZGZjZWQyZQ==</challenge>
-> <response xmlns="urn:ietf:params:xml:ns:xmpp-sasl" />
<- <success xmlns='urn:ietf:params:xml:ns:xmpp-sasl'/>
-> SASL established

Or, decoded:

<- challenge: realm=malkier.net,nonce="9be91df13f8159809d392ed8dc96bdc2",qop="auth",charset=utf-8,algorithm=md5-sess
-> response: charset=utf-8,username="rakaur",realm="malkier.net",nonce="9be91df13f8159809d392ed8dc96bdc2",nc=00000001,cnonce="1ea5e90502101a708fe947230935ef0e46063f213a12ca2dc",digest-uri="xmpp/malkier.net",response=8eb9a3b49271b5bed97cc6a83988aba3,qop=auth
<- challenge: binary
-> response: empty
<- success

This is my (Ruby) code to generate the same response:

def h(s)
    Digest::MD5.digest(s)
end

def hh(s)
    Digest::MD5.hexdigest(s)
end

def startsasl(response)
    @jid = response['username'] + '@' + response['realm']
    a1_h = DB::User.users[@jid].password

    # Compute response and see if it matches.
    # Sorry, but there's no pretty way to do this.
    a1 = "%s:%s:%s" % [a1_h, response['nonce'], response['cnonce']]
    a2 = "AUTHENTICATE:%s" % response['digest-uri']

    myresp = "%s:%s:%s:%s:auth:%s" % [hh(a1), response['nonce'],
                                      response['nc'], response['cnonce'],
                                      hh(a2)]
    myresp = hh(myresp)

    [if myresp is equal to response, they're authorized, otherwise failure]
end

It appears as though Psi isn't base64 encoding cnonce. They're doing
what looks to be an md5 digest, and the proper encoding is a
base64-encoded md5 _hex_digest. As confusing as this (and MD5-DIGEST
in general) is, I believe the culprit to be the lack of base64
encoding on cnonce.

I have been known to be wrong, though. I'm unsure as to how Psi could
have gotten this wrong and it's never been caught, unless everyone's
using PLAIN and not a single one is using DIGEST-MD5 and they didn't
test it at all, which seems unlikely given the longevity of the
client. I just don't understand how it could be my code if it works
with (at least) two other clients using DIGEST-MD5.

--
Eric Will
EBL Engineers
National Institutes of Health
xmpp:rakaur at malkier.net