• wpa_supplicant and eduroam

    From William Unruh@2:250/1 to All on Thu Sep 19 20:58:26 2019
    I am having trouble with wpa_supplicant and am not sure where the
    trouble is. I am trying to connect to my home (UBC)(Canada) from a host here at
    TAMU (Texas). I have debugging switched on for wpa_supplicant. I am
    connecting to eduroam (a network that most universities in the world
    subscribe to so you can log into the local network with your home credentials even if you are away from home).

    My system (Mageia 6.1 up to date) tries to log on, but the connection
    fails for the first (strongest) few access points and will eventually
    (10s of seconds) connect to a much weaker access point. Looking at the
    log, the difference is that the early ones, when they come with their
    original EAP request, request method=1. The one that eventually works
    comes with method=25 (PEAP). Looking into the source code for
    wpa_supplicant eap_common/eap_defs.h I get
    EAP_TYPE_IDENTITY = 1 /* RFC 3748 */
    EAP_TYPE_PEAP = 25 /* draft-josefsson-pppext-eap-tls-eap-06.txt */

    Now, why the EAP request for Identity fails, I have no idea. NOr do I
    know if that is coming from TAMU network, or is being forwarded from UBC network. However this is very annoying.

    wpa_supplicant associates with the access point, and then requests authentication. The failed ones receive that Identity method request,
    which does not work, and wpa_supplicant then blacklists that access
    point and tries the next weaker one. It does this three times at times
    until it succeeds, leaving me attached to a very weak access point (plus
    having had to wait while it went through all those attempts).

    Anyway does anyone know what is going on here, and whether that wrong
    method =1 request would be coming from TAMU or from UBC? Or have any suggestions as to how this problem could be solved?

    (The problem with eduroam is that problems are supposed to be handled
    by the home institutions. UBC's approach seems to be to delay or ask
    irrelevant or trivial question making sure that each round takes at least a day or two,
    until the persons's trip has finished, at which time the problem is moot.


    --- MBSE BBS v1.0.7.12A (GNU/Linux-x86_64)
    * Origin: A noiseless patient Spider (2:250/1@fidonet)