• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 98 subscribers
  • Views 1380 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.

[pandaboard es rev B2] boot: gave up waiting for rootfs. But boots perfectly on identical board

Ronny Sandig
Prodigy 20 points


Hi there!


I'm facing a strange problem here with one of my Pandaboards:

We're working with a network of 8 Pandaboards, 7 clients & 1 server. The Server is Debian Wheezy (LAMP) and is running perfectly for several months now. We had a stable system and didn't do any updates since launch. Though, yesterday I tried to boot the server image for the 1001st time and it hangs because he can't seem to find rootfs on /dev/mmcblk0p2.

I instantly tried the SD card on another identical Pandaboard where it booted without any issues.

Back on the 1st board it again stopped booting and dropped me to busybox/initramfs shell. I did

cd /dev/
ls -a

which showed me /mmcblk0p2 exists like it always did.

After typing

exit

it continued to boot normally and did so after restart and even after power-off for a few minutes.

I did some research and found solutions in case this happens again (changing bootargs) but I'm lacking the knowledge to understand the source of this issue.

Why did the image boot on identical Pandaboard and after that still fails on 1st board?

Why does it now (again) boot correctly on 1st board without altering bootargs or hardware?

Thanks in advance!

r_o

  • 0
    TI__Mastermind 29520 points

    Hello Ronny,

    Why did the image boot on identical Pandaboard and after that still fails on 1st board?

    - Your issue might be caused by a hardware problem on your first board. How did you reset your board?

    Why does it now (again) boot correctly on 1st board without altering bootargs or hardware?

    There is an known similar issue with MMC1 booting for OMAP4:

    • If the sys_boot[5:0] pins are configured for booting from MMC1, then the boot on MMC1 may not happen(board dependency) and the ROM code jumps to the next device of the boot sequence.
    • Indeed, depending on the voltage ramp, the ROM code may not wait long enough for SDMMC1_VDDSvoltage to be stabilized at 3 V before checking the voltage level supplied to OMAP.
    • In that case, the ROM code checks the SDMMC1_VDDS voltage too early, reads a bad level, and then jumps to 1.8-V configuration while the PMIC was configured to 3 V.
    • The MMC1_PBIAS cell is in a bad state and prevents the sending of commands to the MMC1 bus. The ROM code switches to the next boot sequence.

    Unfortunately, there is no workaround to overcome this issue with booting.

    I suggest you if you have such problem at boot stage, to use other power adapter for supplying the PandaBoard.

    Did you make any modifications in the OS which runs on your PandaBoard?

    I would like to notice that Texas Instruments does not provide Debian release for PandaBoard.

    Best regards,

    Yanko

  • 0 in reply to Yanko Todorov-XID
    Prodigy 20 points

    Thanks for your reply, Yanko.


    We're running an image from Robert Nelson www.rcn-ee.com with XFCE on top. For a long time we didn't make any changes in OS affecting boot process, only customizing init.d script and mysql stuff, which are run much later.

    The image (mmcblk0) contains two partitions: ~p1 ('boot', FAT16) and ~p2 ('rootfs', EXT4). The first partition is recognised correctly and starts to boot initramfs.

    Then it tries to mount mmcblk0p2 which always finished in time, but now takes too long.

    Does this mean voltage is not the issue, here?
    Is voltage relevant for different types of filesystems?

    So I increased 'rootdelay' in boot arguments (/uboot/uEnv.txt). This works out fine. As I learned, the delay could be due to SD card being worn out, maybee containing some bad sectors.

    Yanko Todorov-XID said:

    - Your issue might be caused by a hardware problem on your first board. How did you reset your board?

    What do you mean be 'resetting'? I usually shutdown the system and unplug power adapter, according to actual production environment we use this network of Pandaboards in. I rarely use the reset button.

  • 0 in reply to Ronny Sandig
    TI__Mastermind 29520 points

    Hello Ronny,

    I suggest you using a new SD card for booting your first PandaBoard. In case your problem is available yet.

    How often is occurred your issue with PandaBoard boot?

    As I already said we cannot reproduce your problem, because you use a OS which is not supported by Texas Instruments.

    I assume that your problem is caused by noises from power adapter or something similar.

    I suggest you to reset your board by the reset button, not with unplug power adapter.

    Best regards,

    Yanko