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.

"omaplfb: OMAPLFB_Init: OMAPLFBInit failed" on am3517

Other Parts Discussed in Thread: TPS65023, AM3517

Hi,

I am testing on custom board similar with am3517evm except LCD, I substitute a 7" LCD for 4"3' LCD. (7" LCD: 800*480, CLK=33300; 4"3' LCD: 480*272,CLK=9000)

I am booting from NAND, and xloader, uboot, uImage, ubi.img are all on NAND, no MMC is attached.

Each time when I first power on the board,  "omaplfb: OMAPLFB_Init: OMAPLFBInit failed" will show on the console, and android system can't start up.

but if I reboot (just push the reboot button, not power off then power on), then the android system can start up successfully.

however, if I power off, then power on, the same case will appear, "omaplfb: OMAPLFB_Init: OMAPLFBInit failed" will come up on the console.

 

If boot from MMC, no such problem.

So, why do I have to try boot second time after first fail for the error "omaplfb: OMAPLFB_Init: OMAPLFBInit failed"?

what is the difference between second reboot and first power on?

The console log is attached. 3884.OMAPLFBInit failed.txt

 

Thanks,

Kevin

  • Hi Kevin,

    Eventhough "omaplfb: OMAPLFB_Init: OMAPLFBInit failed" error comes, it will try for multiple times, so it should success in one try...

    It is a good catch from your side...Because we have not found such problem with our setup.

    Can you do a quick try on following if your LCD is external powered ?

    - While booting from nand, switch off and switch on LCD

     

    Meanwhile I will try to consult appropriate person and let you provide more guide on that.

     

    -

    satish

     

  • Hi satish,

    My LCD is not external powered.

    I remove my LCD, and I found it is the same when the board without LCD. first power on  will show "OMAPLFBInit failed", and second reboot successfully.

     

    Thanks,

    Kevin

  • Hi Kevin,

    could you send us dmesg logs...?

    $dmesg

    it will help us to look into the  issue...

    also send us logcat logs...

    -

    satish

  • Hi Satish,

    Logs are as below,

    console: 6472.serial console.txt

    dmesg: 6607.dmesg.txt

    logcat: 7245.logcat.txt

     

    Regards,

    Kevin

  • Hi Satish,

    I compared two logs, and list the difference below, hope it will help.

     

     

    Failed Boot Log:

    ----------------------------------------------

    ......

    tps65023 1-0048: Read from reg 0x3 failed

    set_machine_constraints: failed to enable VDCDC1

    tps65023 1-0048: failed to register tps65023

    tps65023: probe of 1-0048 failed with error -5

    ......

    omapdss VENC error: can't get VDDA_DAC regulator

    omapdss CORE error: Failed to initialize venc

    ......

    omapfb omapfb: no displays

    omapfb omapfb: failed to setup omapfb

    omapfb: probe of omapfb failed with error -22

    davinci_emac_probe: using random MAC addr: f6:32:1c:03:bb:14

    ......

    omap_vout omap_vout: no displays

    omap_vout: probe of omap_vout failed with error -22

    .....

    omaplfb: OMAPLFB_Init: OMAPLFBInit failed

    omaplfb: OMAPLFB_Init: OMAPLFBInit failed

    omaplfb: OMAPLFB_Init: OMAPLFBInit failed

    warning: `zygote' uses 32-bit capabilities (legacy support in use)

    request_suspend_state: wakeup (3->0) at 17427185061 (2000-01-01 00:00:14.650787353 UTC)

    init: untracked pid 791 exited

    omaplfb: OMAPLFB_Init: OMAPLFBInit failed

    omaplfb: OMAPLFB_Init: OMAPLFBInit failed

    request_suspend_state: wakeup (0->0) at 24801116946 (2000-01-01 00:00:22.024749755 UTC)

    init: untracked pid 878 exited

    omaplfb: OMAPLFB_Init: OMAPLFBInit failed

    request_suspend_state: wakeup (0->0) at 32076507571 (2000-01-01 00:00:29.300140380 UTC)

    init: untracked pid 908 exited

    omaplfb: OMAPLFB_Init: OMAPLFBInit failed

    ==============================================

     

    Normal Boot Log:

    ----------------------------------------------

    ......

    regulator: VDCDC1: 1200 mV normal

    regulator: VDCDC2: 3300 mV normal

    regulator: VDCDC3: 1800 mV normal

    regulator: LDO1: 1800 mV normal

    regulator: LDO2: 3300 mV normal

    ......

    OMAP VENC rev 2

    ......

    omapdss DPI: Could not find exact pixel clock. Requested 33300 kHz, got 34560 kHz

    regulator_init_complete: incomplete constraints, leaving LDO2 on

    regulator_init_complete: incomplete constraints, leaving LDO1 on

    ......

    omapdss DPI error: display already enabled

    omap_vout omap_vout: 'lcd' Display already enabled

    omapdss DPI error: display already enabled

    omap_vout omap_vout: 'lcd' Display already enabled

    omap_vout omap_vout: Buffer Size = 3686400

    omap_vout omap_vout: : registered and initialized video device 1

    omap_vout omap_vout: Buffer Size = 3686400

    omap_vout omap_vout: : registered and initialized video device 2

    ==============================================

    Thanks,
    Kevin
  • Hi Kevin,

              I was going through the logs,  the reason you see omaplfb init failure error could be because of the following -

    >>>  "couldn't open framebuffer HAL (No such device)." [ from logcat ] - which probably says - there isn't enough memory for the fb.

          Can you please try to append the following to your bootargs and try (along with specifying vram size explicitly, you need to mention the memory required for the omapfb)  - 

    "vram=4M omapfb.vram=0:4M"

     

    Regards,

    Aditya

     

  • We have seen and re-produced this problem with our design.  The problem seems to be with the TPS65023 getting started in powering up the core and then having the source 5 volts drop out momentarily.  It does not seem to recover from this and start all over again and leaves the core of the AM3517 all jacked up.  Our design does not have a battery in front of it and allows for plugging and unplugging power at will.  This is much like the Logic PD am3517evm in the sense that the big slide switch just applies the 5 volts from the power wart to the PMIC.  If you take the switch and kind of bounce it on the EVM you can re-produce the problem on it.  Our solution to our design has been to put a CPU reset IC in front of the 5 volt switcher to allow for bouncing to stop for 1.1 secs before the switcher is enabled.  So far this has worked but I think the PMIC should have this built in to the IC to reset the power up sequence if the source power is dirty and needs to become stable first.