• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 97 subscribers
  • Views 400 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.

66AK2H14: IEEE1588 CPTS isn't time-stamping IPv6 packets

zzMcGee T
Prodigy 60 points
Part Number: 66AK2H14

I'm using XTCIEVMK2X ver4.0 board and trying to do some experiments with PTP. Things are working when configured to layer 2 or IPv4/UDP but things aren't working for IPv6/UDP. It appears that CPTS doesn't see any messages despite confirming they're received and processed correctly by higher level PTP application.

During this experiment I'm setting the registers to the following:

Pn_TS_CTL: 0x000F_0688 (I've also tried 0x000F_0FFF)

Pn_TS_CTL_LTYPE2: 0x00FF_81FF (I've also tried 0x00FF_86DD)

Pn_TS_SEQ_LTYPE: 0x001E_81FF (I've also tried 0x001E_86DD)

Any help/guidance on getting IPv6/UDP hardware-timestamping working would be appreciated.

  • 0
    Prodigy 60 points

    ^^ Correction, for Pn_TS_CTL - I've also tried 0x000F_0FFF 0x000F_07FF.

  • 0
    Prodigy 60 points

    So I saw that the 10 Gigabit User Guide has some additional information, including 2.4.5.2.8 10GbE Switch 1588 Decoder Rules. So I here's my analysis with a screenshot from wireshark as an example of what I'm sending (aligned with what's in those decoder rules):

    I notice there's a Pn_TS_CTL2 register that's mentioned. Unfortunately, there seems to be some typos regarding the documentation - Table 3-72 conflicts with Figure 3-63. I went ahead and tried setting it both ways:

    Pn_TS_CTL2: 0x0000_004F and 0x0004_FFFF

    but still not seeing any messages.

  • 0
    Prodigy 60 points

    I should also add that I've verified that TS is enabled and that the version is 1.5 (which should support Appendix E (IPv6))

  • 0
    Prodigy 60 points

    In case it is pertinent, I've also put ALE in bypass mode to take that out of the equation while troubleshooting. 

  • +1
    Prodigy 60 points

    I noticed that setting unicast enabled in the Pn_TS_CTL_LTYPE2 register bypasses the address checking (according to the decoder rules). I set that and hardware time stamps are now happening. I'm not certain why this is necessary.