USB device configuration

[aseem jolly]

The device I am talking about has one configuration only. I was planning to
take bus trace or usbdump since I am not seeing it again so I don't think
that would help now.


On Thu, Oct 10, 2013 at 12:51 AM, Hans Petter Selasky <hps at bitfrost.no>wrote:

> Hi,
>
>
> On 10/10/13 09:32, aseem.jolly wrote:
>
>> It is allowed to set the configuration multiple times. Refer to USB
>>> 2.0 specification.
>>>
>> >
>
>> In what possible scenario would we want to set the configuration again? I
>> know USB 2.0 specification allows it. I have also tried to modify the code
>> and have sent SET_CONFIGURATION request to the device multiple times, and
>> device didn't report any error.
>>
>
> Some devices have multiple configurations with different interfaces, which
> the application selects depending on the use-case. Many devices only have
> one configuration.
>
>
>> Or do you mean different configurations of a device, if a device supports
>> more that one configuration?
>>
>>
>>  Is this making a difference for you?
>>>
>> I have scene a case in which we were unable to attach the umass driver to
>> the device but since USB_DEBUG was not enabled at that time so I was not
>> able to get more information out of it and being extremely intermittent in
>> nature, I am not able to reproduce it again. I am trying to figure out the
>> possible scenarios in which we might end up not attaching the driver to
>> the
>> device. I believe if we are not able to select a configuration then we
>> don't
>> attach the driver to the device.
>>
>> In usb_unconfigure func:
>>
>>          udev->curr_config_no = USB_UNCONFIG_NO;
>>          udev->curr_config_index = USB_UNCONFIG_INDEX
>>
>> In usb_probe_and_attach func:
>>          if (udev->curr_config_index == USB_UNCONFIG_INDEX) {
>>                  /* do nothing - no configuration has been set */
>>                  goto done;
>>          }
>>
>>  From the console logs, it is clear that SET_ADDRESS went successfully,
>>> have
>>>
>> noticed that we have assigned a ugen name to the device and also request
>> to
>> get the String descriptors(Product, Manufacturer etc) went successfully so
>> that make me believe that there might be an error in performing
>> SET_CONFIGURATION request. Do you have any other suggestions?
>>
>
> You can use "usbdump" to figure out exactly what went wrong with regard to
> USB requests. That might get you a clearer picture of what's going on.
>
> --HPS
>
>