[Prism54-devel] [prism54-usb] DWL-G122 and FW 2.4.6.0 differences

[raven at themaw.net]

On Fri, 10 Dec 2004, Jean-Baptiste Note wrote:

> Hello,
>> I'm still struggling with getting a log dump that I find usefull, that I
>> can compare with others. I have tried usbsnoop and ndiswrapper.
>
> Well I don't know what you mean by usefull, but you'll find here
> http://jbnote.free.fr/prism54usb/data/usbsnoop-logs/
> snoop logs from various devices, courtesy of their respective senders.

Sorry. That's just me.

atm the output from the experimental driver is enough to keep me amused.

>
> And i have put here the dumps that I had a look at and what sense i made
> of them mac-wise (you'll find more info in the code regarding the other
> fields ; i'll put them on the page as i re-learn them :/
>
> http://jbnote.free.fr/prism54usb/MgmtMeaning.html

Spotted that yesterday.

Looks like it's been updated again since.

It's play time (only wish I could spend more time on it, doh).

I'm still working through the comparison but mine looks like the first 
packet up till 0x86 and then the structure appears very sililar. It's from 
firmware 2.4.6.0.

Don't know if you saw my previous post but my driver output is at
http://themaw.net/mess-2.bz2

>
> Originally I had no knowledge of the second packet, my code only
> accomodates the first and third one, but recent devices do need a more
> in-depth knowledge. For now it seems that the magic key
>
> 04 00 01 01
>
> always introduces the MAC.

Yep. Mine as well.

The code I have downloaded from the gnu arch project seems only to know 
about the structure with the option 07 00 03 00 as it looks like the 
offset used to grab the MAC is out by 0x10.

>
>> I think there are several substructures encoded in the management frame.
>> As Feyd says they must have markers or lenghts in there some where. I
>> can't make sense of it yet.
>
> Nor can I. What i know is in the code : after the mac, there are several
> references to the available frequencies on the device, with data fields
> that are used later to generate at least the frequency change packets.
> I'm not familiar with 802.11b, so i don't really know what the meaning
> of these fields are, but the logic that i could made of them is in the
> code.
>
>> Looking at the frame from my device the offset to the MAC address is
>> different. It looks like something is missing prior to that. Everything
>> else is obviously at different offsets as well. I haven't looked further
>> as I'd like to get the MAC address encoding right first.
>
> Hum, could you perchance send me one of your usbsnoop logs so that i may
> have a look too ? (if you already did, i apologize, and i'll look for it
> in the archives...)

Yes. above.

It's just like the first frame on the recently posted packets (I haven't 
got so far as to check the frequency structures yet).

Your recently posted stuff looks quite interesting. I'll digest it for a 
while.

Ian