• State
  • Locked Locked
  • Replies 6 replies
  • Subscribers 33 subscribers
  • Views 4315 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

DP83848K: Link is Up - 100Mbps/Full - flow control off, Link is Down

ReedKacz
Intellectual 2540 points
Part Number: DP83848K

Hello Ethernet Experts!

My customer is using the DP83848KSQ/NOPB and they are facing an issue where the connection link comes up and goes down repeatedly. I have included the snip of the schematic.

Below is the error log:

 

[   35.923655] cpsw 4a100000.ethernet eth0: Link is Up - 100Mbps/Full - flow control off

[   37.933709] cpsw 4a100000.ethernet eth0: Link is Down

[   40.933598] cpsw 4a100000.ethernet eth0: Link is Up - 100Mbps/Full - flow control off

[   44.933499] cpsw 4a100000.ethernet eth0: Link is Down

[   45.933760] cpsw 4a100000.ethernet eth0: Link is Up - 100Mbps/Full - flow control off

[   49.933615] cpsw 4a100000.ethernet eth0: Link is Down

[   52.933886] cpsw 4a100000.ethernet eth0: Link is Up - 100Mbps/Full - flow control off

[   54.933599] cpsw 4a100000.ethernet eth0: Link is Down

[   57.943745] cpsw 4a100000.ethernet eth0: Link is Up - 100Mbps/Full - flow control off

[   61.943617] cpsw 4a100000.ethernet eth0: Link is Down

[   62.943775] cpsw 4a100000.ethernet eth0: Link is Up - 100Mbps/Full - flow control off

Do you have any recommendations or places to start in order to stop this behavior?

Thanks!
Reed Kaczmarek

  • 0
    TI__Mastermind 18425 points
    Hi Reed,

    What is the link partner in this case? And what is the cable length?

    Can they also provide the status of register 0x00, 0x01, 0x04, 0x05, 0x10, 0x14, 0x15, 0x17, 0x19.

    -Regards,
    Aniruddha
  • 0 in reply to Aniruddha Khadye
    TI__Intellectual 2540 points
    Hi Aniruddha,

    1. We are connecting to HP2620 switch and cable length is 1 meter.


    2. Please find below register status:

    0x00 = 3100
    0x01 = 786D
    0x04 = 01E1
    0x05 = 41E1
    0x10 = 0915
    0x14 = 00D4
    0x15 = 0000
    0x17 = 0021
    0x19 = 8001


    Thanks!
    Reed
  • 0 in reply to ReedKacz
    TI__Intellectual 2615 points

    Hi Aniruddha -

    Hope you're well.

    Just wanted to follow up and get your thoughts on the information Reed provided - this is a hot issue at the moment.

    Thanks,

    Dan

  • 0 in reply to Dan Cristy
    TI__Mastermind 18425 points
    Hi Dan, Reed,

    From Register 0x14, I see that the PHY is running in to False Carrier Sense errors. Can you check the magnetics and make sure that it matches datasheet requirement.

    Can you also check if the PHY can link to itself correctly? For this you will need a loopback cable. You can create your own loopback cable by using a Cat5e cable with RJ-45 connector on one end and open wires on the other end. Then connect Pin 1 to Pin 3 and Pin 2 to Pin 6. This will connect the TD+/- pin of the PHY to RD+/- pins.

    Lastly, clock input issues can also cause link instability. I see that clock is being split via 33ohm resistor network. For debug purposes, can you provide a clean clock directly to the PHY without splitting and see if the link gets stable.

    -Regards,
    Aniruddha
  • 0 in reply to Aniruddha Khadye
    TI__Intellectual 2540 points

    Hi Aniruddha,

    Thanks for the comments:

    1. Below are the magnetic specifications. These match the datasheet recommendations.

    2. The customer tested with the loopback cable  and read the 0x14 register its showing value as 0000.

    But when they connect the switch the it shows different values. Might be the counter is getting increased.

    3. Removed the other resistor and provided clock directly still no improvement.

    Still facing the Link Up/Down issue.

    Please let us know how to best proceed.

    Thanks,
    Reed

  • 0 in reply to ReedKacz
    TI__Mastermind 18425 points
    Hi Reed,

    I am closing this thread for now since this being debugged offline.

    -Regards,
    Aniruddha