On Wed, 19 Jun 2013, Greg KH wrote: > On Wed, Jun 19, 2013 at 03:33:51PM -0500, Thomas Pugliese wrote: > > Hi, > > I am attempting to debug a problem where the hwa_hc module occasionally > > fails to start correctly when an HWA device is plugged in. An HWA device > > consists of two USB interfaces each with its own driver: the radio control > > interface (hwa_rc.ko), and the host controller interface (hwa_hc.ko). > > Both of these modules depend on a common subcomponent (uwb.ko) but they do > > not depend directly on each other as far as depmod is concerned. > > Why don't we just build them both together, as they aren't ever used > separately, right? Would that solve your issue? > > thanks, > > greg k-h > I'm not exactly sure what the implications of combining the modules are so I don't know if that would fix it or not. It is true that an HWA will always have an associated radio controller interface but the inverse is not always true. A UWB radio controller can be paired with any number of different protocol interfaces that run on top of UWB. I think HWA_RC is a bit of a misnomer for the radio control module. It should really be something like uwb_usb_rc since it is basically a USB interface to the UWB bandwidth reservation functionality. The best approach is probably to further decouple the HC and RC interfaces instead of tying them closer together. They are meant to be independent and there are only a few areas where the HC driver uses the RC handle. That functionality could be moved to userspace since the HWA already needs input from userspace to start the WUSB channel. Instead of the HWA_HC querying the RC directly for its channel characteristics, userspace could query the RC and send that info to the HC device. Doing that will probably break the ABI though. Is that allowed? Thanks, Tom -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
