David Kuehling wrote: > I guess handling the IRQ pin via polling is especially a waste of > CPU-cycles.
The interrupt of atben works as a "real" interrupt. My user-space tools don't use it as such, because there's no direct way to get interrupts delivered to user space, but the kernel driver does. atusb currently only uses polling because I haven't implemented that transfer mode in the firmware yet. But I don't foresee any major technical difficulties there. Regarding the role of the user-space tools: the ones that talk directly to the hardware, be it via MMIO or libusb, only have diagnostic uses. That's all the atrf-* things under tools/ in the ben-wpan project. All "real" uses would go via the kernel driver and stack, including dirtpan. - Werner _______________________________________________ Qi Hardware Discussion List Mail to list (members only): [email protected] Subscribe or Unsubscribe: http://lists.en.qi-hardware.com/mailman/listinfo/discussion
