• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 98 subscribers
  • Views 967 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.

PHY issues in uboot of Custom K2L board

striker qian
Intellectual 755 points

Hi ,expert

I am now trying to get uboot working on our custom k2l board. Here's issue:

By running via CCS I got uboot up and entered main loop. However it seems network is not up though no error message appears during the starting process.

We are using a different PHY(88E1543M) with that of EVM board(88E1514). In baremetal mode I got NDK running well with 88E1543M without changing any single line.

After debugging uboot for some time I found the return code of  keystone2_emac_initialize() is fine but keystone2_eth_open_close() is not good(they are both in board_eth_init() func , board_k2l.c). after dig a little bit further I found keystone_get_link_status() function is not good either.

Do you have any ideas of how to change uboot code on a different Phy chip?

Thanks,

striker

  • 0
    TI__Guru** 116100 points
    Hi,

    After debugging uboot for some time I found the return code of keystone2_emac_initialize() is fine but keystone2_eth_open_close() is not good(they are both in board_eth_init() func , board_k2l.c). after dig a little bit further I found keystone_get_link_status() function is not good either.


    Please check the following driver file.
    u-boot/drivers/net/keystone_net.c

    1) Undefine the "CONFIG_ETH_PHY_MARVEL_88E1111" in "u-boot/include/configs/ks2_evm.h" file.

    2) Enable the debug in EMAC driver file "keystone_net.c"
    emac_dbg = 1 ;
  • 0 in reply to Titusrathinaraj Stalin
    Intellectual 755 points

    Stalin,

    I am very glad to tell you that  I got network working  after undefine "CONFIG_ETH_PHY_MARVEL_88E1111". But there's additional code modified as below, otherwise it will still fail:

    In func gen_get_link_speed(),Keystone_net.c , the code still calls the keystone2_eth_phy_read() function which is not right I suppose. So I copied the original gen_get_link_speed()  as well as davinci_eth_phy_write() and davinci_eth_phy_read() from davinci_emac.c . To avoid naming confusion I added static keyword in all three functions. And that solved my issue.

    Thanks a lot for your help,

    striker

  • 0 in reply to striker qian
    TI__Guru** 116100 points
    Hi Striker,
    Sounds good.
    I'm glad that you able to solve the problem.
    Thanks for your update with your add-on solution.
    I hope it could help other customers too.
  • 0 in reply to Titusrathinaraj Stalin
    Intellectual 755 points

    Thanks Stalin,

    Now I face the simiar issue in the kernel . 

    here's error log:

    [15159.653404] mdio_bus 26200f00.mdio: cannot get PHY at address 0
    [15159.659352] mdio_bus 26200f00.mdio: cannot get PHY at address 1
     
    after my  digging into the code I found that get_phy_id returns EIO(-5 ) error.
    I am not quite aware of porting linux,so your suggestions will be very helpful . Do you have any ideas about it?
    Thanks,
    striker
     
     
  • 0 in reply to striker qian
    TI__Guru** 116100 points
    Have you tried to modify the "k2l-net.dtsi" file "arch/arm/boot/dts/" for ethernet?
  • 0 in reply to Titusrathinaraj Stalin
    Intellectual 755 points

    Yes, can you give me more hint about the device tree file. I am not sure which part I should modify because I am using 88e1545 but I can't find this type in marvell.c .

    Besides, what's usual reason why I can't read the phy id from address 0 ?

    Thanks,

    striker