2 Replies Latest reply on Jul 11, 2017 1:47 PM by robert_blair_2381936

    Interrupt Message Timing Question

    robert_blair_2381936

      The FX3 only supports 1 Interrupt message type per 125us microframe.

         

      When is a request to send an Interrupt message included in the current microframe and when is it delayed to the next microframe?

         

      And if I try to send two Interrupt messages 50 us apart, will they each be delayed to be available in their own microframe?

        • 1. Re: Interrupt Message Timing Question
          rghe

          Hi,

             

          For Host controllers and devices that operate at high-speed or super-speed, when you specify the bInterval field as 1 in the USB descriptor of an Interrupt endpoint. Host controller will poll the device every 125 us. If the device does not have any data, it will NAK the IN token from the Host. If it has data, it will send the data in response to IN token from the Host.

             

          You can refer to the following link for more details:

             

          https://msdn.microsoft.com/en-us/library/windows/hardware/ff539317(v=vs.85).aspx

          • 2. Re: Interrupt Message Timing Question
            robert_blair_2381936

            Questions are for a SuperSpeed network with Devices having INT Endpoints at the maximum polling rate.

               

            Question 1

               

            If I have two different IN endpoints for a device and both are configured for INT with 1 microframe and I send a message on each within 2 us of each other, will they generally be treated as independent messages with both being read in the same microframe?  Assumption is that they are not separated by a microframe SOF in that 2 us interval.

               

            Question 2

               

            In a system with multiple Devices with each having an INT IN Endpoint, what is the maximum number of INT messages that can be collectively sent by the Devices in any microframe?