Johannes Berg wrote: > On Thu, 2007-08-09 at 16:52 +0200, Michael Buesch wrote: > >> There will be a problem with the PCI IDs, though. >> We can't see from the PCI ID if that's a v4 or v5 device. So >> we have to make _one_ common ssb-pci wrapper and let us probe >> from ssb, so that either b43 or b43legacy is loaded. >> I'll take a look to figure out how that works. Shouldn't be >> too hard. The probing is done through the ssb device table >> in the driver, where you can put the supported core revs. > > For the benefit of those not on IRC, these two patches should be all > that's needed (untested!) > > http://johannes.sipsolutions.net/patches/kernel/ssb-modalias.patch > http://johannes.sipsolutions.net/patches/kernel/ssb-uevent.patch > > The idea will be to strip out the PCI info from both b43 and b43legacy > and then have those two as ssb-core drivers, put the pci IDs into a > bridge module that only registers the pci/ssb bridge and then the other > modules are auto-loaded based on the ssb info.
Sounds good. I do believe we have a plan. Michael will convert bcm43xx-mac80211 into bcm43, and I'll change bcm4301 into bcm43legacy. We will have some patch conflicts over drivers/net/wireless/Kconfig and Makefile, so I'll wait for his submissions to propagate through before I push mine. That will also give me a chance to see how to use the ssb bridge modules. Larry _______________________________________________ Bcm43xx-dev mailing list [email protected] https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
