eMMC problem

Hello! We have a problem with three upboard's on ubuntu 18.04.4 LTS (GNU / Linux 4.15.0-37-generic x86_64, installed according to the instructions https://github.com/up-board/up-community/wiki/Ubuntu_18.04). All three sometimes freeze for exactly 10 minutes, with the following messages in the kernel logs (more detailed log attached to the post):

Sep 23 15:05:48 UP-CHT01 kernel: [280361.789559] mmcblk0: error -110 sending stop command, original cmd response 0x900, card status 0x400e00
Sep 23 15:09:32 UP-CHT01 kernel: [280586.520555] INFO: task jbd2 / mmcblk0p2-: 217 blocked for more than 120 seconds.
Sep 23 15:09:32 UP-CHT01 kernel: [280586.520564] Tainted: G OE 4.15.0-37-generic # 40 ~ upboard07-Ubuntu
...
...
...
Sep 23 15:15:48 UP-CHT01 kernel: [280961.822885] mmc0: Card stuck in programming state! mmcblk0 card_busy_detect
Sep 23 15:15:48 UP-CHT01 kernel: [280961.878891] mmc0: Tuning timeout, falling back to fixed sampling clock
Sep 23 15:15:48 UP-CHT01 kernel: [280961.879235] mmc0: cache flush error -110

Info about eMMC from dmesg:
[ 2.493106] mmc0: SDHCI controller on ACPI [80860F14:00] using ADMA
[ 2.592665] mmc0: new HS200 MMC card at address 0001
[ 2.599631] mmcblk0: mmc0:0001 DA4032 29.1 GiB
[ 2.601903] mmcblk0boot0: mmc0:0001 DA4032 partition 1 4.00 MiB
[ 2.604088] mmcblk0boot1: mmc0:0001 DA4032 partition 2 4.00 MiB
[ 2.606099] mmcblk0rpmb: mmc0:0001 DA4032 partition 3 16.0 MiB, chardev (243:0)
[ 2.612714] mmcblk0: p1 p2
[ 3.453619] EXT4-fs (mmcblk0p2): mounted filesystem with ordered data mode. Opts: (null)
[ 4.382965] EXT4-fs (mmcblk0p2): re-mounted. Opts: errors=remount-ro

What could be the cause of the problem and what should we do to solve it?

log.txt 33.6K

Comments

  • camillus
    camillus Administrator, Moderator, AAEON Posts: 122 admin

    Hi @TheCakeIsALie ,

    The kernel seems to be tainted as a result of loading proprietary module. It is tainted with flags GOE, Kindly refer to this

    It has to do with kernel loading unsigned modules, from the log it shows "3d-mod''.

    Modules without a MODULE_LICENSE or with a MODULE_LICENSE that is not recognised by insmod as GPL compatible are assumed to be proprietary.

    Since Kernel is tainted, you will have to reinstall the ubuntu 18.04 again on the upboards and ensure you load GPL compatible modules.

    Best regards,

  • TheCakeIsALie
    TheCakeIsALie New Member Posts: 2
    edited September 2021

    It has to do with kernel loading unsigned modules, from the log it shows "3d-mod''.
    Modules without a MODULE_LICENSE or with a MODULE_LICENSE that is not recognised by insmod as > GPL compatible are assumed to be proprietary.

    Thx for the reply! But 3d-mod is not a kernel module, it's my python app (doing some work with realsense camera). I think kernel is in tainted state because i've patched uvcvideo module to work with realsense and it's not the cause of the problem. More info:

    dmesg | grep -i 'taint'

    [    5.698007] videodev: loading out-of-tree module taints kernel.
    [    5.698366] videodev: module verification failed: signature and/or required key missing - tainting kernel
    

    lsmod | grep video

    uvcvideo               90112  2
    videobuf2_vmalloc      16384  1 uvcvideo
    videobuf2_memops       16384  1 videobuf2_vmalloc
    videobuf2_v4l2         24576  1 uvcvideo
    videobuf2_core         40960  2 videobuf2_v4l2,uvcvideo
    videodev              180224  5 videobuf2_core,videobuf2_v4l2,uvcvideo
    media                  40960  2 videodev,uvcvideo
    video                  45056  1 i915
    

    modinfo uvcvideo | grep "version:"

    version:        1.1.2.realsense-1.3.13
    srcversion:     0F5F28798BE263E74E073C2