Devicescape's stack was not architected to handle chipsets where the MLME is provided by the hardware component. We did write a low-level driver for the Intersil Frisbee chipset, where the MLME is provided in the firmware on chip - this was achieved by translating authentication and assocation messgaes from the chip into 802.11 association frames, probe requests into scan requests, etc etc. While this worked fine, it is a bit of a hack. Interestingly there is very close to a 1-1 mapping between typical 802.11 APIs and the underlying frame exchanges. (Indeed this is part of the rationalisation for the Devicescape design, where management frames are passed untouched to userspace, rather than parsing them in the kernel and sending events).
Indeed any new netlink protocol should be designed with this layering in mind - i.e. there is 1 set of netlink messages used only with silicon that does not provide a MLME (MLME provided in software), and another set of messages used in both scenarios (MLME provided by silicon/firmware or not). Simon -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Stefan Rompf Sent: Wednesday, January 11, 2006 12:11 PM To: Jiri Benc; netdev@vger.kernel.org Subject: Devicescape stack and 'intelligent' devices Hi Jiri, to evaluate the Devicescape stack, I started porting the ipw2100 driver to it. However I do have one major problem. At least the snapshort from January 2nd does not seem to support devices that do scanning and associating controlled by firmware. Is this assumptions right, and if so, will it be changed and what help will you need to do it? It would be bad to end up with one stack for dumb devices (that by the way, do have their advantages like multiple associations controlled by the stack) and one for cards with firmware. Stefan - To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html - To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
