Subject: Re: WPA regression
To: None <current-users@netbsd.org>
From: Sam Leffler <sam@errno.com>
List: current-users
Date: 02/07/2008 14:08:39
Jukka Salmi wrote:
> Jukka Salmi --> current-users (2008-02-06 23:01:02 +0100):
>   
>> It seems that not only wpa_supplicant but also hostapd is having
>> problems on -current systems: I just upgraded my access point from
>> netbsd-4 to HEAD (4.99.52 kernel, 4.99.50 userland, latest hostapd
>> including todays fixes by Joerg) and noticed other systems are also
>> having problems authenticating. This is what syslogd logged (with the
>> station's address replaced by its operating system name):
>>
>> Feb  6 15:01:18 clam hostapd: ath0: STA [macosx] IEEE 802.11: deauthenticated due to local deauth request
>> Feb  6 15:01:18 clam hostapd: ath0: STA [macosx] IEEE 802.11: deassociated
>> Feb  6 15:01:21 clam hostapd: ath0: STA [macosx] IEEE 802.11: associated
>> Feb  6 15:01:22 clam hostapd: ath0: STA [macosx] WPA: received EAPOL-Key 2/4 Pairwise with unexpected replay counter
>> Feb  6 15:01:22 clam hostapd: ath0: STA [macosx] WPA: pairwise key handshake completed (WPA)
>> Feb  6 15:01:22 clam hostapd: ath0: STA [macosx] WPA: group key handshake completed (WPA)
>> Feb  6 15:11:16 clam hostapd: ath0: STA [macosx] WPA: received EAPOL-Key 2/2 Group with unexpected replay counter
>> Feb  6 15:11:17 clam hostapd: ath0: STA [macosx] WPA: received EAPOL-Key 2/2 Group with unexpected replay counter
>> Feb  6 15:11:17 clam hostapd: ath0: STA [macosx] WPA: group key handshake completed (WPA)
>> [...]
>> Feb  6 19:01:19 clam hostapd: ath0: STA [win] IEEE 802.11: deauthenticated due to local deauth request
>> Feb  6 19:01:19 clam hostapd: ath0: STA [win] IEEE 802.11: deassociated
>> Feb  6 19:01:19 clam hostapd: ath0: STA [win] IEEE 802.11: associated
>> Feb  6 19:01:22 clam hostapd: ath0: STA [win] IEEE 802.11: deauthenticated due to local deauth request
>> Feb  6 19:01:22 clam hostapd: ath0: STA [win] IEEE 802.11: deassociated
>> Feb  6 19:01:22 clam hostapd: ath0: STA [win] IEEE 802.11: associated
>> Feb  6 19:01:25 clam hostapd: ath0: STA [win] IEEE 802.11: deassociated
>> Feb  6 19:01:26 clam hostapd: ath0: STA [win] IEEE 802.11: associated
>> Feb  6 19:01:27 clam hostapd: ath0: STA [win] WPA: received EAPOL-Key 2/2 Group with unexpected replay counter
>> Feb  6 19:01:27 clam last message repeated 2 times
>> Feb  6 19:01:27 clam hostapd: ath0: STA [win] WPA: received EAPOL-Key 2/4 Pairwise with unexpected replay counter
>> Feb  6 19:01:27 clam last message repeated 5 times
>> Feb  6 19:01:29 clam hostapd: ath0: STA [win] IEEE 802.11: deauthenticated due to local deauth request
>> Feb  6 19:01:29 clam hostapd: ath0: STA [win] IEEE 802.11: deassociated
>> Feb  6 19:01:29 clam hostapd: ath0: STA [win] IEEE 802.11: associated
>> Feb  6 19:01:30 clam hostapd: ath0: STA [win] WPA: received EAPOL-Key 2/4 Pairwise with unexpected replay counter
>> Feb  6 19:01:31 clam hostapd: ath0: STA [win] WPA: pairwise key handshake completed (WPA)
>> Feb  6 19:01:31 clam hostapd: ath0: STA [win] WPA: group key handshake completed (WPA)
>> [...]
>> Feb  6 22:31:06 clam hostapd: ath0: STA [nbsd] WPA: received EAPOL-Key 2/4 Pairwise with unexpected replay counter
>> Feb  6 22:31:06 clam hostapd: ath0: STA [nbsd] WPA: pairwise key handshake completed (WPA)
>> Feb  6 22:31:07 clam hostapd: ath0: STA [nbsd] WPA: received EAPOL-Key 2/2 Group with unexpected replay counter
>> Feb  6 22:31:07 clam hostapd: ath0: STA [nbsd] WPA: group key handshake completed (WPA)
>>     
>
> As another data point, I don't see any of those problems when running
> -current's hostapd and wpa_supplicant (0.6.2) on netbsd-4 systems;
> therefore I suspect it's not a hostapd/wpa_supplicant bug at all...
> This makes sense since the problem [1]appeared after some changes from
> the vmlocking branch were merged, without hostapd/wpa_supplicant being
> modified.
>
>
> Regards, Jukka
>
> [1] http://mail-index.netbsd.org/current-users/2007/08/04/0009.htm

I hesitate to ask this but has anyone tried the 0.5.8 versions off the 
stable branch?  I intentionally do not track Jouni's devel branch in 
freebsd.

    Sam