On 31.8.2020 18.15, Alan Stern wrote:
> On Mon, Aug 31, 2020 at 02:46:49PM +0300, Mathias Nyman wrote:
>> Userspace drivers that use a SetConfiguration() request to "lightweight"
>> reset a already configured usb device might cause data toggles to get out
>> of sync between the device and host, and the device becomes unusable.
>>
>> The xHCI host requires endpoints to be dropped and added back to reset the
>> toggle. USB core avoids these otherwise extra steps if the current active
>> configuration is the same as the new requested configuration.
>
> You should mention usb_reset_configuration() here. After all, that's
> where most of the changes in this patch occur.
>
>>
>> A SetConfiguration() request will reset the device side data toggles.
>> Make sure usb core drops and adds back the endpoints in this case.
>>
>> To avoid code duplication split the current usb_disable_device() function
>> and reuse the endpoint specific part.
>>
>> Cc: stable <stable@xxxxxxxxxxxxxxx>
>> Tested-by: Martin Thierer <mthierer@xxxxxxxxx>
>> Signed-off-by: Mathias Nyman <mathias.nyman@xxxxxxxxxxxxxxx>
>> ---
>
>> @@ -1589,8 +1579,12 @@ int usb_reset_configuration(struct usb_device *dev)
>> USB_REQ_SET_CONFIGURATION, 0,
>> config->desc.bConfigurationValue, 0,
>> NULL, 0, USB_CTRL_SET_TIMEOUT);
>> - if (retval < 0)
>> - goto reset_old_alts;
>> + if (retval < 0) {
>> + retval = usb_hcd_alloc_bandwidth(dev, NULL, NULL, NULL);
>> + usb_enable_lpm(dev);
>> + mutex_unlock(hcd->bandwidth_mutex);
>> + return retval;
>
> That's not right; we want to return the original error code. Not 0,
> which usb_hcd_alloc_bandwidth() will probably give us. Just remove
> the "retval =" from the call above.
>
Thanks, somehow I didn't notice that.
-Mathias
