8.2.2.3 Random crashing on Le Potato

  • gadido30 is that with the new bootloader? As it’s only been 24 hours, but so far mines not crashed and it previously has by now. Obviously I need to use it for a few more days to know for sure.

  • LuRcH

    Yes,

    What I did:

    dd if=LibreELEC-S905.arm-8.2-8.2.3.1-LePotato.img of=/dev/mmcblk0 bs=4M

    wget lepotato-792M-u-boot.bin.sd.bin

    dd if=lepotato-792M-u-boot.bin.sd.bin of=/dev/mmcblk0 conv=fsync,notrunc bs=1 count=444

    dd if=lepotato-792M-u-boot.bin.sd.bin of=/dev/mmcblk0 conv=fsync,notrunc bs=512 skip=1 seek=1

    Then I boot LePotato and tried to extract a large zip file (its an addon - 550Mb) , at the end it got stack with the error log I paste.

  • kszaq

    Maybe missing this Kernel Patch?

    [stable:PATCH] arm64: Rework valid_user_regs (v3.16) - Patchwork

  • Hi,

    Mine crashed later in the evening on Monday. So doesn’t look like was the fix.

    On 12hr shifts so not have much free time, but was going to see if could cause the board to crashing with a couple of different distros such as Armbian.

  • gadido30 I applied the patch you mentioned to the kernel, please try it: kernel.img

    Simply replace kernel.img on your SD card. Please not that I have not runtime tested it!

    LuRcH You can try that one as well.

    NB:

    • you can revert to a bootloader with higher DDR clock (simply putting 8.2.3.1 tar in Update folder and "updating" will do)
    • this kernel image should work for all S905/X devices
  • Hi all,

    since yesterday I'm playing around with my new Le Potato and the LibreELEC-LePotato.arm-8.90.3 image, and I had a very few crashes either. Still too early to say anything, but two things may be the culprit in my case: The CEC from my TV, or the web interface. The web interface uses websockets, and from another project of mine I can say that websockets with python (namely: tornado) caused some crashes for me when the browser on the receiving side went down without closing the socket, e.g. when my tablet went to sleep. Could this be the case with the cases reported here?

    UPDATE: I can rule out tornado/websockets: Did not use the web interface the past day, but system crashed anyhow.

    My has hynix RAM as well - Raspberry had some problems with hynix RAM in the past: Public service announcement! - Raspberry Pi - perhaps it makes sense to figure out how they solved the issue?

    By the way: Thanks kszaq for your work!

    Kind regards

    Hauke

    Edited 2 times, last by Hauke (February 13, 2018 at 8:14 PM).

  • Know that gadido30 has still had a crash with his zip file test. But from testing purely in Kodi, mine has been stable for 5 days so far since using the kernel.img which beats the maximum of 30 or so hours previously. So might not be a perfect fix, but seems to have improved things, still testing to see how much longer it will run and if there is any more crashing.

  • Hi LuRcH,

    thanks for the hint. But it seems I'm too stupid: I just replaced kernel.img on the SD, but now the LibreELEC Logo appears, and then it says: Never touch the boot: section in extlinux.conf, your installation is now broken, will start up in 60 sec - it counts down, and then gets stuck. Do I need to run the steps in post #21? I understood post #25 that I would not need to...


    Reverted to the old kernel, and it is running again for now.

    Thanks for any pointers in the right direction!

    Figured it out myself... MD5 checksum of course was wrong. Calculated the MD5 and updated the md5-file - LibreELEC said "updateing" and is now running with the new kernel. Will report on stability!

    Thanks again!

    Hauke

    Edited 2 times, last by Hauke (February 14, 2018 at 7:25 PM).

  • Hi all,

    the kernel file seems to have done the trick - no crash in the past ~46 hours. Needed to reboot now for something else, but will keep you posted.

    Thanks!!!

    Hauke

  • /me again: Today I tried with adamg's image and with kszaq's image to copy a DVD iso file of 7.something GB size via network to the Le Potato. Both with Samba as well as with SCP the copy job crashes after ~500-700 MB, not always at the same point. I tried SD card and USB storage as target, both the same. Is there any log worth extracting for troubleshooting?