On Tue, 2010-02-09 at 15:29 +0300, Alexander Gordeev wrote: > Hi Inaky, > > On Mon, 08 Feb 2010 14:47:31 -0800 > Inaky Perez-Gonzalez <inaky.perez-gonzalez at intel.com> wrote: > > > > > Hello Alexander > > > > On Sun, 2010-02-07 at 03:55 -0700, Alexander Gordeev wrote: > > > > > I'm the author of madwimax (http://code.google.com/p/madwimax/). I > > > was thinking about the future of my project :) and it seems that > > > the only way forward is to integrate it into the Linux-wimax stack > > > somehow: > > > > > > * by porting drivers to the kernel > > > > > > It's good but it will effectively prevent people from using drivers > > > on some wireless routers with an old 2.4 kernel which is very > > > popular now. > > > > Geee .. 2.4 :) > > Kudos goes to Broadcom and their binary blobs :) > > > > * by allowing user-space drivers to be managed by Intel WiMAX > > > network service > > > > > > Is there any opportunity for this? AFAIK, it uses D-Bus, so adding > > > some D-Bus interface for user-space drivers would be not very hard, > > > right? > > > > Well, this code is extremely Intel specific and the long term plan > > would be to ditch it in favor of a generic service. However, that is > > currently in the "until somebody has time to work on it", as no other > > devices are coming forward with need for support. > > OK, thanks for this info! > Then the user-space port will only remain for wireless routers and > we'll move to the kernel. > > > That said, it'd be extremely difficult to modify it to work with your > > driver. How high is the mad driver in the stack level? Does it connect > > to NAPs or to NSPs? > > Well, from what I've seen so far the dongle supports only connecting to > NSPs. :( This device is still a black box. However, other devices are > coming. Can you give an outline of the HCI of the device? We need to design a new kernel API for managing NSP-based devices then. The one I have in the op_nap_scan branchs is for NAP-level devices. Thanks! -- -- Inaky
