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.

Filesystem crash on DM8127

Hi,

We are using DM8127 in our product and develop on IPNC RDK v3.8 (Kernel 2.6.37). But we encountered the filesystem fail issue. Our device has restart function after detect anomalies error. After several restart, this issue will be occurring in the devices. Some devices will occur at mount filesystem stage. Some devices will occur elsewhere. Only re-burned the NAND flash can fix this problem. We have test bootup from Cd card, then it will be happen. If take this SD card into other device, then this SD card can't bootup system.

We want to know what kind of problem will cause this filesystem fail? Please help provide suggest to us? Thank you.

Error Message:

UBIFS error (pid 1022): ubifs_decompress: cannot decompress 2752 bytes, compressor lzo, error -22

UBIFS error (pid 1022): read_block: bad data node (block 33, inode 5182)

UBIFS error (pid 1022): do_readpage: cannot read page 33 of inode 5182, error -22

2021.UBIFS_mount_fail.zip

  • Refering to logs, you are booting from NAND not SD card. It migt occur due to Badblocks, or wrong configs of Nand drviver etc.
    below links might help you.
    www.linux-mtd.infradead.org/.../ubifs.html
    www.linux-mtd.infradead.org/.../ubifs.html
    http://elinux.org/UBIFS
  • Hi Chien,

    Check also if you have all the NAND/GPMC/UBIFS related patches in your linux kernel code base from the below branch:

    http://arago-project.org/git/projects/?p=linux-ipnc-rdk-dm81xx.git;a=shortlog;h=refs/heads/ipncrdk_psp_dm81xx_kernel

    BR
    Pavel

  • Dear Ravikiran:

    Many thanks for your Links.

    Here is our SD Card Test Log with GA3.5 kernel.

    The test procedure as following steps:

    1. Power ON, run uboot, kernel, filesystem from SD Card

    2. after all app process run up, sleep 1 sec , then reboot

    3. repeat upper steps.

    after one weeks, one of five boards crash as the attachment log.

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

    Here is another GA3.8 error message, it occurs after 10~20 times power-cut Test.

    By Albert

     1440.Error.7z

  • Hi Albert,

    Your SD card errors looks strange, because you get it after 5 days !!.
    There are few possibilities listed below. You can checkout them.
    1. SD card has gone bad. So You may have to try with a new good quality SD card.
    2. Rare but important. Issue in clocking and power supply. This can happen due to various reason including layout design. Also, if you connect some device which drag more power can cause some undeterministic behaviour. Please confirm from your end that no such things happend.
    3. Also it is possible that some of your apps with previlage permission or some drivers might corrupted your card.

    So, quick thing to tryout is Use a NEW card and observe it. Also if possible, don't run any application and experiment.

    Please share your observations
  • Dear Ravikiran:

    Depending on your suggest, we did the test as follow:

    1. Prepare 16 pcs new SD Card (ADTA and Transcend), uboot, kernel, rootfs

    2. 8 without run any app (means S99finish.sh run nothing), 8 run apps but not connect in tsp stream in.

    3. after 2 days power ON/OFF test, 3 pcs card crash.

        1 without running apps, 2 run apps

    4. And now we keep testing other 13 pcbs which enable the apps.

    We check the clock and power issue:

    a. XTAL_GND1 and XTAL_GND2 are both no connect, our H/W RD want to check these issue.

    b. Power consumption : DC12V 0.56A .

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

    The most important thing is how to make sure the roots will not crash. 

    If that is the defect of SLC nand flash, UBIFS or SD Card, how to avoid the problem, or any good idea for us to deal with it.

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

    The other question is mtd6 (data) and mtd7 (filesystem2)

    What is the different between put the data (some logs or setting parameters) in filesystem1 and mtd6? We guess the wear-leveling will cause the UBIFS mtd crash after several reboot.

    mtd7 (filesystem2) : how to use filesystem2 to repair filesystem1? or the filesystem2 is just a filesystem1 clone, and kernel do the boot switch?

    BR

    By Albert

     

    7384.20150330_ErrorSD_Test_log.7z