Bluetooth on IBM T30 (USB device problem)

[Lee Damon]

Following up on some old USB issues with the T30 Bluethooth daughter card.
I just noticed this when I did a 'usbdevs -v -d':

Before pushing the button:

Controller /dev/usb2:
addr 1: full speed, self powered, config 1, UHCI root hub(0x0000), 
Intel(0x0000)
, rev 1.00
  uhub2
 port 1 powered
 port 2 powered

After pressing the bluetooth button, but before error message logged:

Controller /dev/usb2:
addr 1: full speed, self powered, config 1, UHCI root hub(0x0000), 
Intel(0x0000)
, rev 1.00
  uhub2
 port 1 addr 0 should never happen!
 port 2 powered

After "uhub2: device problem, disabling port 1" is posted to syslog:

Controller /dev/usb2:
addr 1: full speed, self powered, config 1, UHCI root hub(0x0000), 
Intel(0x0000), rev 1.00
  uhub2
 port 1 enabled
 port 2 powered

> Lee,
> 
> > Kernel recompiled and now no more "_mtx_assert undefined" errors. 
> > Continuing with the directions at <http://www.oook.cz/bsd/bluetooth.html>:
> > 
> > When I press the T30's bluetooth button I get the known USB problem, 
> > but never see the ubt0 device defined.  Instead of getting
> > 
> > ubt0: vendor 0x0a12 product 0x0001, rev 1.10/5.25, addr 2
> > ubt0: Interface 0 endpoints: interrupt=0x81, bulk-in=0x82, bulk-out=0x2
> > ubt0: Interface 1 (alt.config 5) endpoints: isoc-in=0x83, isoc-out=0x3;
> >       wMaxPacketSize=49; nframes=6, buffer size=294
> > 
> > I get dead silence.  Grep'ing for it in dmesg renders:
> > 
> > : || tylendel.castle.org [8] ; dmesg | grep ubt
> > Preloaded elf module "/boot/kernel/ng_ubt.ko" at 0xc06e60a8.
> 
> Of course. Because ng_ubt(4) driver *does not* know about Bluetooth
> device inside your T30. That is why you need to find out VendorID and
> Product ID for the Bluetooth device inside your T30. You can use
> ugen(4) for that. ugen(4) driver *must* be loaded *before* you press
> the Bluetooth button. Once you press the Bluetooth button ugen(4)
> driver *must* attach to it, because it is the default driver for
> the USB devices. That is *if* USB will attach the device and you
> do not get "device problem, disabling port 1" error.
> 
> Once you have ugen(4) attached to the device run
> 
> # usbdevs -v -d
> 
> This will give you VendorID/ProductID pair. Now you are ready to
> patch ng_ubt(4) driver. Just add VendorID/ProductID to the list
> of supported devices inside USB_MATCH function (in ng_ubt.c file).
> 
> Now
> 
> 1) Recompile ng_ubt(4) module
> 2) Turn off your Bluetooth device
> 3) unload ng_ubt(4) (and ugen(4) if you loaded it by hand)
> 4) load *new* ng_ubt(4) driver
> 5) Press the Bluetooth button to activate device
> 
> If everything is working then you should get the messages from the
> ng_ubt(4) driver.
> 
> > I am presuming pressing the bluetooth button is like adding a bluetooth
> > USB device but I am lacking in actual real knowledge here.  Their
> 
> right, i think it more like plugging a new USB device.
> 
> > documentation has things like "Power on the BDC by pressing Bluetooth power 
> > switch if the Bluetooth LED is off." (BDC = Bluetooth Daughter Card).
> > When I press the button the LED does not come on.  If I press and
> > hold the button the LED will eventually flash once and then I get the
> > "uhub2: device problem, disabling port 1" message on the console yet again.
> 
> Well, it seem like when LED is on then the device is plugged, but you
> still getting the "device problem, disabling port 1" message and that
> means USB stack/hub *did not* attach/activate the device and hence no
> driver was attached to it.
> 
> You *must* get the device to attach, otherwise you can not use the
> driver. For now just try to make ugen(4) attach to the device. Once
> you get the ugen(4) to recognize the device then you should move to
> ng_ubt(4).
> 
> > In searching IBM's documentation I found that they have updated firmware 
> > for the BDC, but you have to be running windows to install it. :/  
> > 
> > Kernel source was sup'd Monday evening, US Pacific time.  Bluetooth
> > bundle is ngbt-fbsd-20030305.tar.gz.
> 
> It all does not matter at this point. Again USB *did not* attach (and
> activate) the device. The firmware inside the card and other stuff does
> not event come into play.  The simple fact that you have plugged device
> into hub does not indicate that device is working.
> 
> > I went ahead and tried to run rc.bluetooth start ubt0 in debug mode
> > and got pretty much what you'd expect when the device isn't there:
> 
> [...]
> 
> > + setup_stack ubt0 hook
> > + dev=ubt0
> > + shift
> > + hook=hook
> > + shift
> > + /usr/sbin/ngctl mkpeer ubt0: hci hook drv
> > ngctl: send msg: No such file or directory
> > + exit 1
> 
> No device - no driver - no "ubt0" Netgraph node - no Bluetooth :( Sorry
> 
> In fact, just activate the Bluetooth device (LED on) and run
> 
> # usbdev -v -d
> 
> If device *was* activated you *must* see it, otherwise you will only
> see USB hubs (and possibly other devices that have been attached).
> 
> thanks,
> max
> 

nomad
 -----------                       - Lee "nomad" Damon -          \
play: nomad at castle.org    or castle!nomad                          \
work: nomad at ee.washington.edu                                       \
                                 				    /\
Seneschal, Castle PAUS.                                            /  \
                "Celebrate Diversity"                             /    \