[usb2] Garmin Edge 705 disconnects immediately

[Hans Petter Selasky]

On Sunday 19 July 2009 14:14:23 Michael Gmelin wrote:
> Hans Petter Selasky wrote:
> > On Saturday 18 July 2009 20:46:22 Michael Gmelin wrote:
> >> Hello,
> >>
> >> I'm trying to access a Garmin Edge 705 bike computer, which should show
> >> up as a mass storage device.
> >>
> >> kldload usb2_controller_uhci|ehci
> >>
> >> Jul 18 19:37:55 ufo kernel: ugen2.2: <Garmin International> at usbus2
> >> Jul 18 19:37:55 ufo kernel: ugen2.2: <Garmin International> at usbus2
> >> (disconnected)
> >>
> >> kldload usb2_storage_mass
> >> kldload usb2_controller_uhci
> >> Jul 18 19:38:06 ufo kernel: ugen2.2: <Garmin International> at usbus2
> >> Jul 18 19:38:06 ufo kernel: umass0: <Garmin International Edge 705,
> >> class 0/0, rev 1.10/5.09, addr 2> on usbus2
> >> Jul 18 19:38:06 ufo kernel: umass0:  SCSI over Bulk-Only; quirks =
> >> 0x0000 Jul 18 19:38:08 ufo kernel: umass0:1:0:-1: Attached to scbus1
> >> Jul 18 19:38:08 ufo kernel: umass0: at ushub2, port 1, addr 2
> >> (disconnected) Jul 18 19:38:08 ufo kernel: ugen2.2: <Garmin
> >> International> at usbus2 (disconnected)
> >>
> >> Tried various quirks without any success for usb2_storage_mass (afaik it
> >> should be SCSI | BBB). Also tried all possible orders of attaching the
> >> device/
> >>
> >> Setting various debugging sysctls didn't give any useful hints (didn't
> >> see anything extraordinary).
> >> usbconfig doesn't show anything because the device disappears
> >> immediately. I would assume that this is not a problem with _storage but
> >> the device expecting some initialization which doesn't happen (or
> >> happens too often). Also tried various usb2_quirk settings, all without
> >> success.
> >>
> >> VendorId: 0x091e
> >> ProductId: 0x2271
> >>
> >> Maybe someone can give me a hint where to start debugging this.
> >
> > You can try:
> >
> > sysctl hw.usb.debug=15
> >
> > I suspect that the USB stack in your device has some bugs. Sounds like
> > the bike-computer manufacturer are not the ones to spend the most money
> > on USB.
> >
> > --HPS
>
> I attached you the log output which clearly indicates trouble. Maybe
> some more info on the device: It has in-built memory + a memory card,
> which should come up as separate devices. It boots up and shows that it
> is connected to a computer, after a couple of seconds (once detach
> failed) it will fallback to charging mode. I think the device could also
> used as a serial device to gather GPS input, so maybe that's part of the
> problem.
>
> Considering the price of the device I wish they would have spent more
> money on usb. On the other hand the GPS sensor is fantastic and the
> device overall is really awesome (it seems to mount ok under linux and I
> couldn't find anyone who had to do any patching for that, the linux
> kernel also no quirks for it, so I hope it's just a minor flaw in the
> protocol that makes usb2 fail because it aims to be standard compliant :).
>

Could you redo the trace without umass loaded?

--HPS