Libertas Driver Locking Kernel

[Holger Schurig]

> Trying wireless-testing as well as the driver from 2.6.25-4
> both resulted in kernel panics from null pointers.

I'm very interested in this kernel panic because of NULL 
pointers.

Because many different firmwares from unspecified sources with 
unspecified *) changes are floating around, we must treat the 
data that the firmware sends us in (e.g. command responses) like 
data coming from userspace. That is, we must check everything 
and sanitize/discard things which are in formats unknown to us. 
Never ever should the libertas driver make a NULL pointer or 
panic because of one.


*) Maybe you can see a log of the firmware versions on the 
Marvell customer Intranet, but last year they didn't grant me 
access to the intranet. Two days ago someone in my company again 
tried to get into the Marvell Intranet. Let's see if they are 
more helpful this time compared to one year ago.