hostapd/wpa_supplicant - new development release v0.6.3

[Dan Williams]

On Fri, 2008-02-22 at 19:09 -0800, Jouni Malinen wrote:
> New versions of wpa_supplicant and hostapd were just
> released and are now available from http://w1.fi/
> 
> This release is from the development branch (0.6.x). Please note that
> the 0.5.x branch continues to be the current source of stable releases.
> 
> hostapd:
> * fixed Reassociation Response callback processing when using internal
>   MLME (driver_{hostap,nl80211,test}.c)
> * updated FT support to use the latest draft, IEEE 802.11r/D9.0
> * copy optional Proxy-State attributes into RADIUS response when acting
>   as a RADIUS authentication server
> * fixed EAPOL state machine to handle a case in which no response is
>   received from the RADIUS authentication server; previous version
>   could have triggered a crash in some cases after a timeout
> * fixed EAP-SIM/AKA realm processing to allow decorated usernames to
>   be used
> * added a workaround for EAP-SIM/AKA peers that include incorrect null
>   termination in the username
> * fixed EAP-SIM/AKA protected result indication to include AT_COUNTER
>   attribute in notification messages only when using fast
>   reauthentication
> * fixed EAP-SIM Start response processing for fast reauthentication
>   case
> * added support for pending EAP processing in EAP-{PEAP,TTLS,FAST}
>   phase 2 to allow EAP-SIM and EAP-AKA to be used as the Phase 2 method
> 
> wpa_supplicant:
> * removed 'nai' and 'eappsk' network configuration variables that were
>   previously used for configuring user identity and key for EAP-PSK,
>   EAP-PAX, EAP-SAKE, and EAP-GPSK. 'identity' field is now used as the
>   replacement for 'nai' (if old configuration used a separate
>   'identity' value, that would now be configured as
>   'anonymous_identity'). 'password' field is now used as the
>   replacement for 'eappsk' (it can also be set using hexstring to
>   present random binary data)

Even though NM doesn't use these quite yet, would this mean that NM
would have to dynamically detect at runtime what version of
wpa_supplicant was installed to ensure that it wouldn't send the wrong
config options to the supplicant here?  I want to keep NM working with
both 0.5 and 0.6.

If that's the case, would you accept a patch to add a getVersion()
function to the D-Bus interface that passes back a (major, minor, micro)
tuple?

Dan


> * removed '-w' command line parameter (wait for interface to be added,
>   if needed); cleaner way of handling this functionality is to use an
>   external mechanism (e.g., hotplug scripts) that start wpa_supplicant
>   when an interface is added
> * updated FT support to use the latest draft, IEEE 802.11r/D9.0
> * added ctrl_iface monitor event (CTRL-EVENT-SCAN-RESULTS) for
>   indicating when new scan results become available
> * added new ctrl_iface command, BSS, to allow scan results to be
>   fetched without hitting the message size limits (this command
>   can be used to iterate through the scan results one BSS at the time)
> * fixed EAP-SIM not to include AT_NONCE_MT and AT_SELECTED_VERSION
>   attributes in EAP-SIM Start/Response when using fast reauthentication
> * fixed EAPOL not to end up in infinite loop when processing dynamic
>   WEP keys with IEEE 802.1X
> * fixed problems in getting NDIS events from WMI on Windows 2000
>