svn commit: r186730 - in head: lib/libusb20 sys/dev/usb2/controller sys/dev/usb2/core sys/dev/usb2/ethernet sys/dev/usb2/image sys/dev/usb2/include sys/dev/usb2/serial sys/dev/usb2/sound sys/dev/us...

[Hans Petter Selasky]

On Tuesday 06 January 2009, Stanislav Sedov wrote:
> On Tue, 6 Jan 2009 17:51:26 +0100
>
> Hans Petter Selasky <hselasky at c2i.net> mentioned:
> > On Tuesday 06 January 2009, Stanislav Sedov wrote:
> > > On Tue, 6 Jan 2009 16:13:32 +0100
> > >
> > > Hans Petter Selasky <hselasky at c2i.net> mentioned:
> > > > Try:
> > > >
> > > > usbconfig -u 0 -a 2 set_config 1
> > >
> > > This helps. Now ubsa attaches to the device. But I can't communicate
> > > with it (no echo even). I see the following messages in the log:
> > > ubsa_cfg_request:395: device request failed, err=USB_ERR_STALLED
> > > (ignored) ubsa_cfg_request:395: device request failed,
> > > err=USB_ERR_STALLED (ignored) ubsa_cfg_request:395: device request
> > > failed, err=USB_ERR_STALLED (ignored)
> > >
> > > > You can for example add a quirk for this, see "usbconfig -h | grep -i
> > > > quirk". I know this is not so easy, but if you can make a general
> > > > rule for when selecting another configuration, I will add it.
> > >
> > > Why not let the driver to set the config?
> >
> > Because:
> >
> > 1) it clashes with USB device side mode, where the config index is
> > received from Host
> >
> > 2) Different device drivers might want different configurations.
> >
> > 3) The driver for the active configuration might not be loaded.
>
> I see.
>
> > This is interesting:
> > >   bNumConfigurations = 0x0001
> >
> > The device says it only has one configuration! So setting index 1 should
> > not be allowed.
> >
> > Could you try to change the check in the ubsa2.c probe routine to check
> > for config index 0 instead?
>
> It attaches but doesn't work with the same error message as listed above.
>
> > >       iInterface = 0x0003  <retrieving string failed>
> >
> > When you cannot retrive the strings, I think your device is gone!
> > (Crashed)
>
> So at this point the device is effectively detached? Do you have ideas why
> that might happen. I don't also understand how it may work with config
> index 1 on old stack if the maximum number of configurations is 1 too...
> Looks weird.

Hi,

Some checkpoints:

1) You removed the config index 1 quirk?
2) You re-plugged the module and let it stay disconnected for some seconds?

You can reach the endpoints through ugen when config index 0 is set.

cat /dev/ugen0.2.0.1

echo "at" > /dev/ugen0.2.0.2

If you are lucky you will get something back.

--HPS