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.

Executing "reboot" command (adb reboot) from terminal prompt causes cpu hang in kernel.

Other Parts Discussed in Thread: X5777ATXGABC

Hi,

 While I am running long time automation test for reboot gets  cpu hang and failed to reboot in "VAYU EVM XC5777X CPU Boards (Jacinto 6)".

When the cpu hang its failed to do 'Soft_reset' and not able to reboot, but when i press 'cpu_reset' button from the board manually, it continuous to work.

this issue is fixed or not.

If it is fixed please provide the patch.

Regards

Vincie Joe

  • Hi Vincie,

    I've moved your post to the Automotive forum in order to ensure that the proper experts can respond. Please allow for some delay in response due to the holiday season here in the United States.
  • Hi Vincie,

    Your question has been forwarded to experts. Expect delay in response due to holidays.
    Best regards
    Lucy
  • Hi Vincie

    Is this TI-EVM or custom board ? What is the J6 silicon revision?

    Regards

    Ravi

  • Hello Ravi,

     It is TI-EVM board only and the

    J6 Silicon revision is,

    OMAP X5777ATXGABC

    52ZDMZ9
    8H2 ABC G1

    Regards

    Vincie Joe

  • Hi Vincie

    Ok, its TI-EVM. What is revision of board. You can provide u-boot console log.

    There was a warm reset issue on silicon revision 1.0/1.1, that's why reboot was not working. It should work latest Rev-H EVM SR2.0 revision.

    Regards

    Ravi

  • Hello Ravi,

    Revision board is : G3

    U-boot log at hang time as follows,

    [ 11.620927] healthd: battery none chg=
    [ 11.898329] lowmemorykiller: lowmem_shrink: convert oom_adj to oom_score_adj:
    [ 11.905532] lowmemorykiller: oom_adj 0 => oom_score_adj 0
    [ 11.911068] lowmemorykiller: oom_adj 1 => oom_score_adj 58
    [ 11.916670] lowmemorykiller: oom_adj 2 => oom_score_adj 117
    [ 11.922358] lowmemorykiller: oom_adj 3 => oom_score_adj 176
    [ 11.928056] lowmemorykiller: oom_adj 9 => oom_score_adj 529
    [ 11.933690] lowmemorykiller: oom_adj 15 => oom_score_adj 1000
    [ 13.389053] acc_open
    [ 13.391267] acc_release
    [ 14.434529] mmcblk0: error -110 sending status command, retrying
    [ 14.444983] mmcblk0: timed out sending r/w cmd command, card status 0x800b00
    [ 14.452838] mmcblk0: status not valid, retrying timeout
    [ 14.561649] mtp_open
    [ 15.921662] binder: release 2225:2542 transaction 5434 in, still active
    [ 15.928619] binder: send failed reply for transaction 5434 to 2576:2576
    [ 16.796916] init: Boot Animation exit
    [ 23.600585] SysRq : Emergency Remount R/O
    [ 23.606292] EXT4-fs (mmcblk0p10): re-mounted. Opts: (null)
    [ 23.620264] EXT4-fs (mmcblk0p13): re-mounted. Opts: (null)
    [ 23.627289] Emergency Remount complete
    [ 23.726685] reboot: Restarting system with command ''

    U-Boot SPL 2014.07-00121-gfc792b4 (May 27 2015 - 21:06:00)
    DRA752-GP ES1.1

    u-boot Hangs at this position and device is not booting after this,

    When i press hard cpu reset key then it proceed rebooting.

    Regards

    Vincie Joe

  • Hello Ravi,

     Any Inputs on the issue.

    Regards

    Vincie Joe

  • Vincie

    As explained in earlier mail thread, its a known issue on ES1.1 silicon. 

    This issue has been fixed in Silicon revision 2.0.

    Regards

    Ravi

  • Hi Ravi,

    You are talking about erratum i875 Power-on-Reset (PORz) Warm Boot Hang. Am I right?

    If true, I'm wondering if Vincie can implement the workaround as described in erratum? I'm attaching it below.

    Thanks,

    Stan

    A board level workaround requires adding a 220 Ohm (+/- 5%) resistor onto the SRAM
    LDO supply (Cap_vddram_mpu1, ball K16 on the SoC package with ABC designator
    (that is, the 23 mm package)). This resistor provides a controlled discharge path for the
    charge contained within the external 1μF LDO capacitor during the reset operation. The
    workaround effectiveness assumes that the active duration of the PORz signal is a
    typical 3.4 ms or greater.
    In systems with an MCU present, a second workaround technique can be utilized. The
    MCU performs a handshake with the device following a warm PORz to ensure that the
    device is responsive after the reset. On the occasion a hang occurs, the MCU should
    assert PORz low for 200 ms. This eliminates the need for an external 220 Ohm resistor.

  • Hi Stan

    You are right, i am referring PORz warm boot issue.

    Regards
    Ravi
  • Hi Stan,

       Sorry for delay in reply

    Please share me the screen shot for board level work around to place 220 ohm  resistor onto the SRAM
    LDO .

    Regards

    Vincie Joe

  • Here you go

    Regards,

    Stan

  • Hi Stan,
    Thanks for sharing the circuit diagram,
    but my board is "OMAP X5777ATXGABC", the schematic which you shared is different.
    please share the schematic for the "OMAP X5777ATXGABC" layout, which help me to work on this.

    Regards
    Vincie Joe
  • Main goal is to have the 220ohm resistor between CAP_VDDRAM_MPU1 and GND.

    Do you have the resistor on your board? If not, is it feasible to solder one?

    Regards,

    Stan

  • Hi Stan/Ravi,

    Thank you for support ,

    My management is planed to go for the latest Rev-H EVM SR2.0 revision.

    But they need some information:

    - If we procure the product, the platform will resolve the soft boot issue? 

    - If we procure the latest Rev-H platform, your team will support us for any issue in the board and software level?

    - We can purchase the product in India itself or from other country?

    - Please share the quotation of  latest Rev-H EVM SR2.0 revision.

    Thanks 

    Vincie Joe

         

    latest Rev-H EVM SR2.0 revision

  • Hi,

    Ravi already answered on your first question about soft reset boot issue - it is solved for SR2.0.
    For the other questions - please refer to the local TI representative as this will be the fastest way to get answers.

    Best regards
    Lucy
  • Vincie

    I have verified "reboot" on J6 Rev-G EVM, with 2014.07 u-boot, kernel used was 4.4, reboot works properly. Did not see any issue.

    From u-boot logs, commit shows you are using old u-boot 2014.07. Can you check latest 2014.07 once. What is the android release & kernel version ?

    What is the frequency of issue, always occurs first time "reboot" ?

    Did check with other Rev-G boards as well ?

    Regards

    Ravi

  • Hi Ravi,

    What is the frequency of issue, always occurs first time "reboot" ?

            Issue is not occur in first reboot,When we do automation test for reboot continuously more than 10 times, then this issue is occurred.

    Did check with other Rev-G boards as well ?

           Yes, we have 3 Rev: G Board, in all 3 boards I am able to reproduce this issue.

    Regards

    Vincie Joe

  • Vincie

    Can you check with latest u-boot and kernel version.

    Regards

    Ravi

  • Hi Ravi,

     As per your advice, we ordered the platform with the  "latest Rev-H EVM SR2.0 revision", to overcome the issue.

    Thanks and Regards

    Vincie Joe

  • Hi Ravi,

    I have bought the new Jacinto 6 board, In that box i didn't find any user manual or installation guide to boot the platform.

    Please guide me how to get all source for particular jacinto 6 platform, I have shared the board details down,

    "My Board Details:

    Cpu Board: Vayu EVM xc5777x

    S/N : VA1_1703016

    REV : H

    ASSY : 516580 "

    When I tried with my old Jacinto 6 platform MLO and u-boot i am not able to enter into Fastboot mode.

    If I copy all the partition manually to the SDcard i am not able to boot it.

    It stuck in display failed error "ldc3001_read_reg: i2c transfer failed (-121)".

    I want a proper MLO, U-boot.img and the kernel and android source packages or share me the link where i could get the source for my board.



    Regards

    Vincie Joe
  • Hi Vincie,

    I will mark this thread as closed, the issue will be discussed here:
    e2e.ti.com/.../590270

    Regards,
    Yordan
  • Hi Ravi,

     I have purchased New Jacinto6EVM board with Rev-H ES:2.0 version, Still I am facing the reboot issue.

    When I have run the Reboot cmd script continuously more than 15 to 30 loop, the Board got stuck and not restarted. 

    The board got stuck after u-boot and before kernel starts.

    I have attached the screen shot were i got stuck.

    Please provide me the proper solution to resolve it.

    Regards

    Vincie Joe

  • Hi, @Vincie Joe,

    The last post is the same as the one in:

    I will mark this thread as closed. The discussion can continue in the other thread.

    Regards,

    Mariya