Socket CF driver, the Zaurus, and monitor mode

[Jean Tourrilhes]

Ian Goldberg wrote :
> 
> But I'd like kismet to work, too.  :-)
[...]
>
> So I patch spectrum24-0.3.3 with the orinoco-0.11b patch from
> http://airsnort.shmoo.com/orinoco-0.11b-patch.diff and install that
> version.  No dice; I can run "iwpriv monitor eth0 2 3", but I don't
> see any packets.
> 
> --> Why?  Is the above diff only supposed to work on actual Orinoco
>     cards, not any card served by the orinoco driver?  That would
>     certainly explain it, of course.

	My bet is that the patch above was done for Lucent/Agere
firmwares, and the Symbol firmware is different and need to be done
differently.
	Either you get some Symbol documentation, or you trial and
error (based on various Spectrum driver headers).

> - I patch my 2.4.6 kernel using these patches, in order:
> 
> http://www.hpl.hp.com/personal/Jean_Tourrilhes/Linux/wireless.v12.diff
> http://www.hpl.hp.com/personal/Jean_Tourrilhes/Linux/iw_handlers.w13-5.diff
> http://www.hpl.hp.com/personal/Jean_Tourrilhes/Linux/iw_handlers.w14-5.diff

	Those patches won't do anything with regards to kismet. And
I've never tried them under 2.4.6, so it might be broken.

>   (The plan is to use the WE14 Scanning mechanism.) Also no go.  Any
>   reference to eth0 sticks a process in D state.  (I think that was the
>   problem.)

	We don't have scanning support for Symbol firmware. Same
problem as monitor mode, we need more info.

> - I build and install spectrum24-0.3.3 (no patches).  Badness: now the
>   spectrum_cs module hangs in the (initializing) state:
[...]
> One thing I forget to mention: the compile initially fails, beause
> in net/core/wirelesss.c, WE_EVENT_NETLINK is unconditionally #define'd.
> This causes compile-time errors in the event that CONFIG_RTNETLINK isn't
> defined.
> 
> I simply wrapped the definition of WE_EVENT_NETLINK in an
> "#ifdef CONFIG_RTNETLINK".  I hope that's OK.

	When setting some parameter, the new Wireless Extension will
generate an event. Obviously, something about event generation is
broken, which explain why it get stuck.
	Please try to #undef WE_EVENT_NETLINK, and also enable the
DEBUG in this file to see what's happening.

> The card is a prism2 card, right?  Only its firmware needs to be loaded
> at every powerup?
> 
> modprobe prism2_cs

	It's a Symbol card, not an Intersil card, both linux-wlan-ng
and HostAP won't be able to do anything with it (even if they pretend
they can).

> What would it take to make some wlan-ng module work with this card?

	To rewrite extensive portions of it.

	Regards,

	Jean