which ioctl in driver takes WPA-TKIP passphrase?

[hong zhang]

Dan, Bryan,
   
  Beside key, should I also enable countermeasurement inside driver?
   
  ---henry


Dan Williams <dcbw at redhat.com> wrote:
  On Tue, 2007-02-20 at 22:06 -0500, Bryan Kadzban wrote:
> Dan Williams wrote:
> > You don't push the passphrase to the driver, you hash the passphrase 
> > with wpa_passphrase to get the key and you push the key down to the 
> > driver using SIOCSIWENCODEEXT. Drivers certainly shouldn't be
> > accepting passphrases unless they are smoking really, really unholy
> > crack.
> 
> Is even the key accepted by the driver? I thought the client needed to
> generate a key from the passphrase, then do a 4-way handshake with the
> AP using that key, to get a temporal key, and then finally program the
> temporal key into the driver. (Then it has to do a group handshake to
> get the group key as well, although WPA2 does this during the second
> half of the 4-way handshake. Either way, the client needs to program
> the group temporal key into the driver as well.)
> 
> And in any case, if you don't do the 4-way handshake after associating,
> then the AP will either disconnect you, or not let you pass any traffic.

Right; drivers have ioctls that let wpa_supplicant (or any other
supplicant) send the key to the driver after it's completed the 4 way
handshake, otherwise the driver wouldn't know how to decrypt the packets
after the handshake had taken place.

Dan



-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.shmoo.com/pipermail/hostap/attachments/20070221/63f0b33a/attachment.htm