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]Ethernet connection failed

Other Parts Discussed in Thread: 66AK2H12

Dear Champs,

My customer failed to work Ethernet ports after successful kernel boot-up.

Could you please check their dts file and log?Desktop.zip

They used 88e1111 PHYs for 4 x 1G Ethernet ports and 88x3310 PHYs for 2 x 10G Ethernet ports.

Thanks and Best Regards,

SI.

  • Hi,

    We're working on this. Feedback will be posted directly here.

    Best Regards,
    Yordan
  • Hi Yordan,

    Is there any update?
    My customer found there was no MDIO(MDCLK) was generated. Can you guess why MDIO(MDCLK) was not generated?

    Thanks and Best Regards,
    SI.
  • Hi,

    What I've noticed from their dts file is that they don't have the mdio status set to "ok". Have they tried:

     mdio: mdio{

            compatible = "ti,davinci_mdio";

             status="okay";

            #address-cells = <1>;

            #size-cells = <0>;

            reg = <0x2090f00 0x100>;

            bus_freq = <2500000>;

            clocks = <&clkcpgmac>;

            clock-names = "fck";

            phy0: phy@0 {

             ...............................

    Best Regards,

    Yordan

  • Sorry for the delayed response on this.
    I agree with Yordan.
    Can you please check the "k2hk-evm.dts" file which has the MDIO definition.

    &mdio {
    status = "ok";
    ethphy0: ethernet-phy@0 {
    compatible = "marvell,88E1111", "ethernet-phy-ieee802.3-c22";
    reg = <0>;
    };

    ethphy1: ethernet-phy@1 {
    compatible = "marvell,88E1111", "ethernet-phy-ieee802.3-c22";
    reg = <1>;
    };
    };

    Did you ask this question for the below customer ?

    Seems similar issue.

    https://e2e.ti.com/support/dsp/c6000_multi-core_dsps/f/639/p/538863/1964638#1964638

  • Yes. It is same customer.

    Thanks for your support!!!



    Thanks and Best Regards,
    SI.

  • There were still no MDCLK output even after changing dts file as you recommended.
    Is there anything else to check?

    Thanks and Best Regards,
    SI.
  • I attached their modified dts file and their new log with new dts file.

    Could you please check this again?

    Ethernet.zip

    Thanks and Best Regards,

    SI.

  • Dear Sung-IL,
    It seems, the customer is using old kernel version (3.10) based on MCSDK.
    So ask them to try to use the latest Processor SDK package.
    I referred PSDK package and provided the code snippet for MDIO.
    Will check with internal team on this MDIO clock.
  • Hi, Sung-Il,

    Please refer to the ethernet driver section in the document, http://processors.wiki.ti.com/index.php/MCSDK_UG_Chapter_Exploring#Ethernet_Driver, and see if it helps. It mentions how to assign phy_addr. It also mentions that the u-boot env variable, has_mdio, may be needed. See if that helps.

    Rex

  • Hi, Sung-Il,

    I also checked that the mdio address is wrong in customer's dtsi file. It should be 0x02090300.

    Rex
  • FYI, the related documents to get MDIO address are:
    66AK2H12/14 data sheet - sprs866e.pdf
    Keystone Architecture NetCP - sprugz6.pdf
    Keystone Architecture GbE Switch Subsystem - sprugv9d.pdf
  • Thanks for your help, there was some progress, but still failed to ping.

    After changing MDIO address as you pointed, they found there was device presented in kernel log as below, but still fail to ping.

    [   91.875952] davinci_mdio 2090300.mdio: phy[0]: device 2090300.mdio:00, driver Marvell 88E1111

    [   91.884519] davinci_mdio 2090300.mdio: phy[1]: device 2090300.mdio:01, driver Marvell 88E1111

    [   91.893082] davinci_mdio 2090300.mdio: phy[2]: device 2090300.mdio:02, driver Marvell 88E1111

    [   91.901642] davinci_mdio 2090300.mdio: phy[4]: device 2090300.mdio:04, driver Marvell 88E1111

    INIT: Entering runlevel: 5

    Configuring network interfaces... [   97.554084] keystone-netcp 2090000.netcp: initializing cpsw version 1.3 (1) SGMII 1

    [   97.564193] keystone-netcp 2090000.netcp: Created a cpsw ale engine

    [   97.573228] davinci_mdio 2090300.mdio: resetting idled controller

    [   97.579587] keystone-netcp 2090000.netcp: phy found: id is: 0x2090300.mdio:04

    [   97.689032] pps pps0: new PPS source ptp0

    [   97.693078] cpts rftclk rate(600000000 HZ),mult(5120),shift(10),div(3)

    [   97.734783] keystone-netcp 2090000.netcp: Using Packet Accelerator Firmware version 0x03000105

    [   97.743436] keystone-netcp 2090000.netcp: pa_clk_rate(163840000 HZ),mult(25000),shift(12)

    [   98.009261] net eth0: netcp device eth0 opened

    [   98.015373] IPv6: ADDRCONF(NETDEV_UP): eth0: link is not ready

    [   98.021236] 8021q: adding VLAN 0 to HW filter on device eth0

    They still failed to ping  as below.

    root@k2hk-evm:~# ifconfig eth0 down

    [  137.428775] dma dma1chan10: out of descriptors

    [  137.433243] dma dma1chan10: out of descriptors

    [  137.437707] dma dma1chan10: out of descriptors

    [  138.828742] dma dma1chan12: patx-pdsp5-0 leaked descriptor 0

    [  138.834422] dma dma1chan12: patx-pdsp5-0 leaked descriptor 1

    ....

    [  144.813005] dma dma1chan12: patx-pdsp5-0 leaked descriptor 1023

    [  144.988758] ------------[ cut here ]------------

    [  144.993402] WARNING: at drivers/net/ethernet/ti/keystone_pa.c:822 pa_tx_dma_callback+0xf8/0xfc()

    [  145.002226] dma completion failure, status == 3

    [  145.006596] Modules linked in:

    [  145.009853] CPU: 3 PID: 1942 Comm: ifconfig Not tainted 3.10.72 #5

    [  145.016071] [<c00155cc>] (unwind_backtrace+0x0/0xf4) from [<c0011744>] (show_stack+0x10/0x14)

    [  145.024642] [<c0011744>] (show_stack+0x10/0x14) from [<c0021510>] (warn_slowpath_common+0x54/0x6c)

    ....

    [  159.908930] [<c03dc0e4>] (sock_ioctl+0x1bc/0x290) from [<c00e040c>] (do_vfs_ioctl+0x3fc/0x5c0)

    [  159.917580] [<c00e040c>] (do_vfs_ioctl+0x3fc/0x5c0) from [<c00e0608>] (SyS_ioctl+0x38/0x60)

    [  159.925970] [<c00e0608>] (SyS_ioctl+0x38/0x60) from [<c000d760>] (ret_fast_syscall+0x0/0x30)

    [  159.934449] ---[ end trace f5b1dc321997f60d ]---

    [  162.349371] dma dma1chan16: qos-bypass-0 leaked descriptor 1011

    [  162.355313] dma dma1chan16: qos-bypass-0 leaked descriptor 1012

    [  162.361264] dma dma1chan16: qos-bypass-0 leaked descriptor 1013

    [  162.367209] dma dma1chan16: qos-bypass-0 leaked descriptor 1014

    [  162.373155] dma dma1chan16: qos-bypass-0 leaked descriptor 1015

    [  162.379108] dma dma1chan16: qos-bypass-0 leaked descriptor 1016

    [  162.385046] dma dma1chan16: qos-bypass-0 leaked descriptor 1017

    [  162.390995] dma dma1chan16: qos-bypass-0 leaked descriptor 1018

    [  162.396938] dma dma1chan16: qos-bypass-0 leaked descriptor 1019

    [  162.402883] dma dma1chan16: qos-bypass-0 leaked descriptor 1020

    [  162.408835] dma dma1chan16: qos-bypass-0 leaked descriptor 1021

    [  162.414774] dma dma1chan16: qos-bypass-0 leaked descriptor 1022

    [  162.420719] dma dma1chan16: qos-bypass-0 leaked descriptor 1023

    They saw an error in packet accelerator as well, and I don't confident if they set the packet accelerator setting well.

    Could you please check attached their full log and advice on this issue?EthernetErrorLog.txt

    Thanks and Best Regards,

    SI.

  • Hi, Sung-Il,

    The serdes should have 4 lanes, but only 2 are configured. However, this does not explain why customer's eth0 not working. Customer should have at least eth0 and eth1 working if the lane2 and lane3 are not defined in serdes. Customer will need to check if their configuration is correct. Not knowing the hardware, I can't tell.

    Rex

  • Sung-Il,

    Once customer get eth0 and eth1 working, can they also try without mdio and see if eth2 and eth3 working first before enabling mdio?

    Rex