> From: Jérôme LELEU
> Sent: Thursday, January 22, 2015 6:49 AM
> 
> Yes indeed, you should upgrade to close the vulnerability if you use LDAP
> authentication.

You know, if you're going to announce a "holy crap upgrade now" security issue, 
it would be nice to get a little advance notice that it's coming 8-/.

I don't quite understand this vulnerability. According to the announcement 
(http://seclists.org/oss-sec/2015/q1/205), it says "CAS Server 3.5.2 allows 
remote attackers to bypass LDAP authentication via crafted wildcards".

Then under the description it says "A valid username and password required." It 
further says "The login will be sucessfully only if the ldap bind search return 
one unique member."

If you need to know a valid username and the correct password for that 
username, how exactly are you "bypassing" authentication? It sounds like if you 
specify a wildcard that matches one and exactly one identity in your directory, 
*and* you supply the correct password for that identity, you successfully 
authenticate? Again, I don't understand how that can be considered to bypass 
authentication? It looks like the only ramification is that you can 
successfully authenticate with a string that isn't exactly the username, and 
that string is then presumably provided to the application you are trying to 
authenticate to? So instead of the application thinking the user "henson" 
logged in, it would think the user "hens*" logged in? Presumably undesirable, 
with potentially unknown ramifications depending on the application, but still 
not bypassing authentication.

Also, I can't seem to reproduce it on my deployment. The LDAP wildcard "henso*" 
matches one and exactly one entry in my directory. If I type "henso*" and my 
correct password into the CAS login form, it tells me it is invalid.

If I try the example in the announcement:

curl -k -L -d "username=henso%2A&password=XXXXXXXXX" 
https://auth.csupomona.edu/cas/v1/tickets

All I get in return is the CAS login page.

Is this vulnerability dependent on how you have LDAP configured? I am using the 
FastBindLdapAuthenticationHandler mechanism. I don't believe there is any way 
for this vulnerability to apply to my configuration, as attempting to directly 
bind with the provided wildcard will always fail. Perhaps the vulnerability is 
only applicable to people using the BindLdapAuthenticationHandler, which would 
perform a wildcard search and find an entry which it would then try to bind as?

Please clarify the issues surrounding this vulnerability so users can respond 
appropriately. My initial impression is that if you are using the 
FastBindLdapAuthenticationHandler you are not affected, so perhaps instead of 
announcing "You must upgrade if you use LDAP authentication" you should 
announce "You should upgrade if you are using the BindLdapAuthenticationHandler 
for LDAP authentication"? I also don't think the CVE should have a title that 
it bypasses authentication, as you're hardly bypassing authentication if you 
are required to know the username and password for the account 8-/. More 
accurately, it seems you can simply misrepresent your username to an 
application.

Thanks…

--
Paul B. Henson  |  (909) 979-6361  |  http://www.cpp.edu/~henson/
Operating Systems and Network Analyst  |  [email protected]
California State Polytechnic University  |  Pomona CA 91768



-- 
You are currently subscribed to [email protected] as: 
[email protected]
To unsubscribe, change settings or access archives, see 
http://www.ja-sig.org/wiki/display/JSG/cas-user

Reply via email to