Re: Device scope wrapper for bluetooth

[Elad Efrat <elad%NetBSD.org@localhost>]

On Tue, Apr 28, 2009 at 3:04 AM, matthew green <mrg%eterna.com.au@localhost> 
wrote:
>
> i don't see why bluetooth gets its own kauth requests.  (or why usb,
> or pci, or ...)

I believe it's important to be able to indicate it's a bluetooth device. :)

(I have thoughts about desktop users using policies that grant them
special rights for bluetooth devices and other things that might make
sense in such a setup, but it's nothing concrete at the moment...)

> please create generic requests that aren't specific to any particular
> type of bus or device, beyond the high level classes we have already,
> in particular, the list in devclass_t looks to be the right set of
> "device types" to consider.  there's already a bunch of work done for
> the tty class... to me, bluetooth potentially belongs to several
> classes: bus, disk, ifnet, audiodev, tty .. at least.

The changes in question affect all bluetooth device (IIUC) and are
relevant for the bluetooth layer rather than the device's purpose.
We're at a layer where we can't tell, unless maybe if we inspect
the HCI unit structure, what the device actually is.

As for a bluetooth tty device, see my other mail (about attaching
tty devices).

Thanks,

-e.