hostapd / ath9k 4-way handshake fail

Jouni Malinen j
Tue Jul 14 03:36:35 PDT 2009


On Wed, Jul 15, 2009 at 02:03:39AM +0000, Hector Akamine wrote:

> Taking a look at the debug messages, I found that hostapd and wpa_supplicant 
> create different PSKs from the same SSID and passphrase (so the PMK PTK
> calculated later are different):

Are the hostapd and wpa_supplicant running on similar CPUs and with
similar build tools/configuration? This looks like a bug somewhere and
should not really happen if everything works correctly.

> hostapd log (excerpt):
> 
> Using interface wlan1 with hwaddr 00:0d:02:3b:0a:96 and ssid 'hostapdAP-sta001'
> 
> SSID - hexdump_ascii(len=16):
>      68 6f 73 74 61 70 64 41 50 2d 73 74 61 30 30 31   hostapdAP-sta001
> PSK (ASCII passphrase) - hexdump_ascii(len=10):
>      73 74 61 30 30 31 70 61 73 73                     sta001pass      
> PSK (from passphrase) - hexdump(len=32): 18 40 0d 03 5a e6 b7 db b0 bf ca cd 
> b9 de 38 59 ba 9f d0 99 2d 35 6e a6 73 3b f0 e2 26 2e 8e 18

This PSK is incorrect.

> wpa_supplicant log (excerpt) :
> 
> Line: 9 - start of a new network block
> ssid - hexdump_ascii(len=16):
>      68 6f 73 74 61 70 64 41 50 2d 73 74 61 30 30 31   hostapdAP-sta001
> PSK (ASCII passphrase) - hexdump_ascii(len=10):
>      73 74 61 30 30 31 70 61 73 73                     sta001pass      
> PSK (from passphrase) - hexdump(len=32): 53 55 0b 81 d4 d3 a1 74 06 17 7d 3a
>  07 fe bc c9 0d 48 ea 0b 03 5a fd a2 5f d3 bd 80 f7 ce da 78

This PSK is correct.

> I expected both PSKs to be the same.
> Isn't the PSK only dependent on the passphrase and the SSID? 

Yes. I would expect that something odd happened with the hostapd build
(e.g., mismatch in build and target endianness configuration).

-- 
Jouni Malinen                                            PGP id EFC895FA



More information about the Hostap mailing list