Is the MIC of EAPoL-Key (2/4) affected by the .11 driver?

[Ali Bahar]

On Wed, Jun 15, 2011 at 07:31:44AM +0800, Ali Bahar wrote:
> On Tue, Jun 14, 2011 at 04:41:31PM +0200, Johannes Berg wrote:
> > On Tue, 14 Jun 2011 22:05:13 +0800, Ali Bahar wrote:
> > > On Tue, Jun 14, 2011 at 08:35:59PM +0800, Ali Bahar wrote:

> > >> I am getting a "Michael MIC failure" from the AP, and am wondering 
> 
> > I'm on my wait out, but I think you're confusing the michael MIC and 
> > the EAPOL MIC.
> 
> I know what you mean, but wireshark (the latest, development version)
> is decoding the 0x000e Reason Code as "Michael MIC failure". I did not
> look it up, and instead chose to post it as is.

The .11 Standard lists the 0x000e as "Message integrity code (MIC)
failure", so the Michael interpretation is wireshark's -- and
presumably a bug.

... Looking at the wireshark source code, that seems to be the case. I
will see if they already have an outstanding bug issued for this. 

Meanwhile, anyone with a clue as to why my AP hates me? Presuming that
the PSK has been entered properly at both ends (AP and STA), and that
the protocols (eg WPA2) match, the only remaining suspects are the
executing codes at both end -- something which I do not rule in at
all, at this point. 

My WPA2 configuration:
network={
        ssid="my"
        key_mgmt=WPA-PSK
        proto=WPA2
        pairwise=CCMP
        group=CCMP
        psk="secretphrase"
}

And for WPA1:
network={
        ssid="my"
        key_mgmt=WPA-PSK
        proto=WPA
        pairwise=TKIP
        group=TKIP
        psk="secretphrase"
}

I may try testing it against hostapd.

thanks,
ali