On Mon, Jun 20, 2016 at 7:57 PM, Felipe Balbi <felipe.balbi@xxxxxxxxxxxxxxx> wrote: > > Hi, > > Muni Sekhar <munisekharrms@xxxxxxxxx> writes: >>>>>> #define USB_CTRL_GET_TIMEOUT 5000 >>>>>> >>>>>> #define USB_CTRL_SET_TIMEOUT 5000 >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> Does the above mentioned timeouts will vary based on the packet length >>>>>> of control transfers (8, 16, 32 or 64 bytes)? >>>>>> >>>>>> >>>>>> >>>>>> I do have a USB Bluetooth device and it exposes the control endpoint >>>>>> for sending HCI commands. >>>>>> >>>>>> >>>>>> >>>>>> I’m not sure what will be the good value for the >>>>>> usb_control_msg()timeout for sending HCI commands, will it be safe to >>>>>> use USB_CTRL_SET_TIMEOUT? >>>>> >>>>> why wouldn't it? If your device doesn't respond to a control message in >>>>> 5 seconds, then you should really get that FW fixed or change your >>>>> device to something that works. 5 seconds to transfer a few bytes is a >>>>> really long time. Even for low-speed pipes (1.5Mbps), you can transfer >>>>> 960 kiB in 5 seconds if you saturate the bus. How big are your HCI >>>>> commands? >>>> >>>> HCI Command Packets with up to maximum of 255 bytes of data. >>> >>> then 5s is more than enough for you ;-) >> >> What usually triggers the usb_control_msg() call to return? Is it just >> an acknowledgement from the device that the command has been received? > > No, it's completion of the entire thing. A control message is 2 or 3 > phases. It goes like below (an oversimplification of reality): > > 2-phases > ======== > > (timer starts) > -> Host sends SETUP token > -> Host sends 8-byte control > -> Peripheral sends ACK > -> Host sends STATUS token > -> Peripheral sends ACK > (timer ends) > > 3-phases > ======== > > (timer starts) > -> Host sends SETUP > -> Host sends 8-byte control > -> Peripheral sends ACK > -> Host or Peripheral sends DATA bytes > -> Host or Peripheral sends ACK > -> Host sends STATUS token > -> Peripheral sends ACK > (timer ends) > > As stated, this is an oversimplification of reality, but it's enough to > illustrate how this works. Thanks for the clarification. > > -- > balbi -- Thanks, Sekhar -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
