• State
  • Locked Locked
  • Replies 2 replies
  • Subscribers 33 subscribers
  • Views 625 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

DP83640 packet based timestamp delivery and heavy receive traffic

Christian Riesch
Intellectual 710 points

Other Parts Discussed in Thread: DP83640

Hi,
the DP83640 PHYTER software development guide, paragraph 3.1.8, says: "Programmable inter-frame gap and preamble length allow the PHY to recover lost bandwidth in the case of heavy receive traffic." Could someone please explain what this means? What will happen when heavy rx traffic is present? Will timestamping data (PHY status frames) be dropped under heavy traffic? Will packets from the network be dropped to allow timestamps to be delivered? Is changing e.g. the inter-frame gap required to ensure all data and timestamps are received?
Thanks, Christian
  • 0
    TI__Mastermind 31600 points

    Christian,

    The receive traffic has priority over PHY status frames.  No receive packets will be dropped due to PHY status frames. 

    As timestamps occur, they will be placed into their respective queue.  The timestamp will then be delivered to software via PHY status frame when there is sufficient space between receive packets.

    If timestamps are placed into the queue faster than they are taken out of the queue, the queue will overflow and timestamps will be lost.  If receive traffic is heavy and timestamps accumulate faster they they can be sent via PHY status frames, the queue will overflow and timestamps will be lost. 

    Patrick

  • 0 in reply to Patrick OFarrell
    Intellectual 710 points
    Patrick,
    Thank you very much for explaining this. So in case of event timestamps, if more timestamps are placed into the queue than can be delivered via status packets, I would see an increase of the "event missed" counter, EVNTS_MISSED in the PTP_ESTS register. Is this correct?
    Best regards,
    Christian