Posts by Iridium

    Full error for information:

    7:17:42.992 T:1608397728 ERROR: Traceback (most recent call last):
    File "/home/chewitt/LibreELEC.80-official/build.LibreELEC-imx6.arm-7.90.010/LibreELEC-settings-b9bda01/.install_pkg/usr/share/kodi/addons/service.libreelec.settings/oe.py", line 271, in load_url
    File "/usr/lib/python2.7/urllib2.py", line 154, in urlopen
    File "/usr/lib/python2.7/urllib2.py", line 435, in open
    File "/usr/lib/python2.7/urllib2.py", line 548, in http_response
    File "/usr/lib/python2.7/urllib2.py", line 473, in error
    File "/usr/lib/python2.7/urllib2.py", line 407, in _call_chain
    File "/usr/lib/python2.7/urllib2.py", line 556, in http_error_default
    HTTPError: HTTP Error 501: Not Implemented


    Hm... I may have been too brief. Let me try again: when Python throws errors on LibreELEC, it refers to the Python build path. LibreELEC, Python and addons are built by chewitt in his home directory, /home/chewitt... This explains messages referring to /home/chewitt/... in the log

    OK - That makes more sense. So ignore it?

    IMX6 V7.90.010
    Viewing .kodi/temp/kodi.log I noticed briefly (I think when an addon update was in progress an error occurred updating an addon /home/chewit....

    Only addons are Iplayer, Missing movies, LE Network and Le System tools.

    I lost the log, but can post more information when it occurs again.

    As it mentioned chewit I figured it was relevant.


    I see only 2 queries on boot and then nothing for some time.

    Code
    LibreELEC:~ #  journalctl --no-pager -b -0 | grep ntp:
    Jan 04 18:52:17 LibreELEC connmand[304]: ntp: time slew +1.154688 s
    Jan 04 18:52:18 LibreELEC connmand[304]: ntp: time slew -0.002825 s

    Interesting? Mine is an upgrade from V7.90.005.
    Still the same rate of updates ~16 seconds.

    Could this be a fault in the RTC?
    Anyway of disabling the update?

    IMX6 V7.90.010
    Running journalctl --no-pager -b -0 -f

    -- Logs begin at Thu 1970-01-01 01:19:51 BST. --
    Jan 04 17:57:51 Deneb connmand[547]: ntp: time slew +0.001860 s
    Jan 04 17:58:07 Deneb connmand[547]: ntp: time slew +0.001613 s
    Jan 04 17:58:23 Deneb connmand[547]: ntp: time slew +0.001042 s
    Jan 04 17:58:39 Deneb connmand[547]: ntp: time slew -0.001880 s
    Jan 04 17:58:55 Deneb connmand[547]: ntp: time slew +0.004571 s
    Jan 04 17:59:11 Deneb connmand[547]: ntp: time slew +0.001150 s
    Jan 04 17:59:27 Deneb connmand[547]: ntp: time slew -0.002975 s
    Jan 04 17:59:43 Deneb connmand[547]: ntp: time slew +0.008019 s
    Jan 04 17:59:59 Deneb connmand[547]: ntp: time slew -0.001660 s
    Jan 04 18:00:15 Deneb connmand[547]: ntp: time slew +0.000708 s

    Which seems to suggest that LE is checking the NTP server every ~15 seconds.
    Is this normal as it seems a bit excessive?

    If I understand you correctly you have a RPi1 which has a 700MHz ARM11 76JZF-S processor whereas the RPi2 has a 900 MHz 32-bit quad-core ARM Cortex-A7.

    So they are completely different processors.

    Turbo mode is just a way to get a little more "boost" - usually at start up, or to disable the cpu governor.

    RPi1 Overclocking: medium; 900 MHz ARM, 250 MHz core, 450 MHz SDRAM, 2 overvolt. Works without being too excessive.

    Read Raspberry Pi - Wikipedia for more information.

    You could try rpi-clone Here. It uses rsync which I believe is installed already on LE. It can be run standalone in the home folder as ./rpi-clone. (No need to copy to /usr/local/sbin as in the install instructions)
    I haven't tested it on LE yet but it works on Raspbian and has the added bonus that the system doesn't have to be shutdown.

    Confirmed that it is a bug and still in v7.90.010.

    To replicate, you do not need a Hifiberry-dac or clone. Just add dtoverlay=hifiberry-dac to config.txt, reboot and try and change the audio output to <alsa hifiberry> = blank screen.

    There was something exactly the same with the RPi2 and the Hifiberry-Digi, related to the LIRC, where a GPIO pin reassignment was required, but I have tried that and it still didn't work. I've tried various pin adjustments but none solve the problem.

    A "diff" between a working dtbo and the current one showed no difference.

    It's obviously a bug but I don't know where it exists. There is nothing in "dmesg" nor kodi.log. It has to be something from V7.90.006 that has caused it as it works in LE7 (stable), Xbian (stable) and moode (current).


    Remove the first line. The second line is correct.

    Code
    [ 996.977697] bcm2835-dma 20007000.dma: DMA transfer could not be terminated


    is certainly the problem. Can you post the working dmesg log with v7.90.005?

    -------
    I asked that:
    Firstly I would try with "Normal" CPU settings not "HIgh" just in case that is affecting it
    Next remove device_tree_overlay=overlays/hifiberry-dac.dtbo

    But apparently that didn't make any difference!

    Doh - spoke too soon. The add-on system-tools installs correctly on a fresh install of V7.90.009 however on an upgrade from 007 to 009 after you press "install" nothing happens. This is the same with network-tools Shader-toys and also the add-on repository. It's as if nothing in the LE repository can be installed - nothing appears in the log.

    System and network tools are still showing 8.0.102 and 8.0.100 respectively.

    Is there any way to check the correct repository is being used and if not how to update it?
    [hr]
    Doh doh doh.

    After reading another post - I figured out how to update the repository and it installs and runs!!!

    So it works, but should the repository have been updated on upgrade or is it something that happens when the add-on upgrade script runs?
    [hr]

    I would try a fresh install on a different SD card and see if it is a bug or just your configuration - you have an awful lot of stuff in your config.txt which might be affecting it.

    I expect to have a DAC clone by new year, so I should be able to debug further if required.