"Chris Knipe" <[EMAIL PROTECTED]> wrote:
> I only realised after I posted.  Basically, both of them are authorised if
> they happen within a short period of time after each other...

  Ok...

> rlm_pap: login attempt by "user@realm" with password fl4shp1x13
> rlm_pap: Using password 6e8a104e79c05453663713d06b797611 for user user@realm
> authentication.
...
> rlm_pap: login attempt by "user@realm" with password fl4shp1x131
> rlm_pap: Using password  for user user@realm authentication.

  And that's probably the cause of the problem.  The passwords in the
RADIUS requests are different (the second one has an additional '1'),
but the second time, the *configured* password to check against is
empty.

  It's a bug in rlm_pap, and I'll go fix it.

  My question then is where does the empty password come from?  The
PAP module is trying to compare the user-supplied password with an
*empty*, but not non-existent password taken from some database, by
another module.

  So the fix to rlm_pap is NOT sufficient here.  We need to know which
other module you're using to get the MD5 password, and why it returns
an empty password.

  Alan DeKok.

- 
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

Reply via email to