Re: btsco(4) on -current

["Jared D. McNeill" <jmcneill%invisible.ca@localhost>]

Iain Hibbert wrote:
> I can't see anything obvious missing in your explanation. I've upgraded to
> a -current as of saturday kernel now and it seems to be working fine for
> me still (i386/cardbus/uhci/ubt and a Jabra headset), though the audio
> recorded seems a bit stuttery (but not if played back through the headset
> which I'm hoping is a filter problem)

FYI, I tested on 4.0/i386 with ehci/ohci and I had the same issue. The 
headset is a Motorola H350:

  http://direct.motorola.com/ENS/HKernel/H350_Headset_detail.asp

> if you use 'bthset -v' can you see that it is connected to the headset
> control channel ok? does SIGUSR1 cause ringing and can you twiddle with
> the volume settings ok?

Yes, that all works fine.

> what is there uhci attached to? (if there is any ehci there, it won't
> work as isochronous packets just get lost)

Since ubt is a full speed device, ehci hands it over to the uhci 
companion controller. For the heck of it I disabled ehci and it didn't 
make a difference.

> I just found another place that bt_lock is needed, so update
> sys/dev/bluetooth/btsco.c before you snoop, but sysutils/hcidump should
> show packets in and out of the hardware driver, can you try to see if
> anything happens there?

Sure, I can try that tomorrow -- I do remember some utility (btconfig 
maybe? it was a few weeks ago) reported no errors and that isoc packets 
were flowing in both directions.

Cheers,
Jared