Posts by ghtester

    Firstly observed on latest LE12 Nightly versions, after upgrade to LE13 the issue is still there. Ocassional Kodi crashes when watching LiveTV using Tvheadend42 & DVB-T2 USB key. The crash usually happen sometimes, as soon as I go to TV/Channels/All channels menu or open the TV Guide (TV/Guide/All channels menu).

    Hopefully you can find the details in the logs:

    https://paste.libreelec.tv/simple-lark.log

    https://paste.libreelec.tv/tidy-bug.log

    https://paste.libreelec.tv/legible-dane.log

    https://paste.libreelec.tv/evolving-vervet.log

    i can confirm this is happening to me as well for last couple of years. no matter tvheadend or kodi version. I have observed, that this happens also when recording a tv show, so this looks more like a tvheadend server bug. Also from what i see, it happens to ghtester in czech republic, which is the same location as mine, so it may be the connected with h.265 codec, which is used here for dvb-t2...

    Yes in the Czech rep. it's a bit specific situation due to h.265 codec so I understand that developers from other countries (where the h.264 is still in use) don't see any issues.

    I am recording the tv programs exceptionally so I did not encounter the audio issue in this case - but when tested the behavior with bad signal level, during play of the recorded file the audio could be recovered after a temporary loss but this did not happen during live TV play of the same interval of the same program (and the audio stayed muted).

    BTW. I see that tvheadend server has more troubles with channels using fast data stream (CT1, CT2 - h.265, 1080 HD), where especially the timeshift often fails after some time.


    It's probably related to AAC audio. See here.

    Thanks for the link, unfortunately it seems it can't be fixed...

    The python script using System Tools stopped working on my RPi 4B running LE 12 nightly-20231231-dfd843f (RPi4.aarch64)

    It looks at least paths has been changed so the add-on is not backward compatible anymore and the script failed with error message:

    File "/storage/.config/shutdown.py", line 25, in <module>
    import RPi_I2C_driver
    File "/storage/LCD/RPi_I2C_driver.py", line 16, in <module>
    import smbus
    ModuleNotFoundError: No module named 'smbus'

    After updating path in the script

    from sys.path.append('/storage/.kodi/addons/virtual.system-tools/lib')

    to sys.path.append('/storage/.kodi/addons/virtual.system-tools/lib.private')


    the error message has changed to:

    File "/storage/.config/shutdown.py", line 25, in <module>
    import RPi_I2C_driver
    File "/storage/LCD/RPi_I2C_driver.py", line 17, in <module>
    import smbus
    ImportError: libi2c.so.0: cannot open shared object file: No such file or directory

    The directory content is:

    ~/.kodi/addons/virtual.system-tools/lib.private # ls -l
    total 3092
    -rw-r--r-- 1 root root 266408 Dec 22 11:44 libfuse.so
    -rw-r--r-- 1 root root 266408 Dec 22 11:44 libfuse.so.2
    -rw-r--r-- 1 root root 266408 Dec 22 11:44 libfuse.so.2.9.9
    -rw-r--r-- 1 root root 67552 Dec 22 11:44 libi2c.so
    -rw-r--r-- 1 root root 67552 Dec 22 11:44 libi2c.so.0
    -rw-r--r-- 1 root root 67552 Dec 22 11:44 libi2c.so.0.1.1
    -rw-r--r-- 1 root root 619176 Dec 22 11:44 libonig.so
    -rw-r--r-- 1 root root 619176 Dec 22 11:44 libonig.so.5
    -rw-r--r-- 1 root root 619176 Dec 22 11:44 libonig.so.5.4.0
    -rw-r--r-- 1 root root 67824 Dec 22 11:44 libulockmgr.so
    -rw-r--r-- 1 root root 67824 Dec 22 11:44 libulockmgr.so.1
    -rw-r--r-- 1 root root 67824 Dec 22 11:44 libulockmgr.so.1.0.1
    -rw-r--r-- 1 root root 68944 Dec 22 11:44 smbus.so

    My experience is (using USB to SATA & SATA HDD to boot & run LE) like described in my previous post. If the boot device is unmounted and there's SSH connection already, it's not dropped and some commands (including dmesg) still works.

    But of course it's not able to make a new SSH connection anymore (even if the boot device is remounted).

    That's not my case. I have no problem with configuration from scratch, even with fresh configuration there are some issues I had already reported here (yes I should perhaps put the reports to Kodi & Tvheadend forums as well but as the both are used in LE, I thought some day the reports could be reflected). Unfortunately as the issues are related to DVB-T2 which looks to be less used across users, the priority is low. I respect that free time of developers is limited but it's true TVH progress is very slow. Nevertheless, yet TVH 4.2 is a great product and still better product than others (I have tested) and than TVH 4.3 as well.


    I’ve been using 4.3 Alpha for quite a while and have found no issues or incompatibilities. Could you elaborate a little.

    Please look at some of my earlier posts. I was experiencing a configuration issue with adding the first DVB-T network / channel. There's also an incompatibility with Kodi - subtitles are not shown at some DVB-T channels (in fact shown 'behind the corner' while recording is OK including subtitles). I can't remember any advantage of 4.3 over 4.2 so after a long torture with subtitles issue (tried to find the root of the issue but failed and nobody here could help), I have finally reverted to 4.2.

    Now I am observing if the another annoying and old issue with randomly undetected audio streams is also related with TVH 4.3 or if it's the same with 4.2 (the next possible reason could possibly be the USB DVB-T2 tuner).

    As the issue persisted in nightly-20231219-865d980 (RPi4.aarch64) as well, I had to analyze it & find the reason.

    It looks in kernel 6.6.7 the old mapping file format is not supported anymore. I think it should not fail with Segmentation fault but...

    It was necessary to reenable kernel IR protocols (for instance ir-keytable -p lirc -p rc-5 -p jvc -p sony -p nec -p sanyo -p mce_kbd -p rc-6 -p sharp) and convert the mapping file to .toml format. Then it works again.

    Thanks as well but I believe there's no issue with the path. In fact the script works even in autostart.sh (LED can be turned on/off) except the input button.


    btn = Button(5, pull_up=False)

    the kernel log:

    export_store: invalid GPIO 5


    When the path is wrong, the script fails at import command which is not the case.

    After upgrade to nightly-20231216-63347fc (RPi4.aarch64) my IR remote stopped working.

    I am using gpio_ir_recv and found the custom mapping does not work anymore.

    Found /sys/class/rc/rc0/ with:
    Name: gpio_ir_recv
    Driver: gpio_ir_recv
    Default keymap: rc-rc6-mce
    Input device: /dev/input/event1
    LIRC device: /dev/lirc1
    Attached BPF protocols:
    Supported kernel protocols: lirc rc-5 rc-5-sz jvc sony nec sanyo mce_kbd rc-6 sharp xmp imon rc-mm
    Enabled kernel protocols: lirc rc-6
    bus: 25, vendor/product: 0001:0001, version: 0x0100
    Repeat delay: 500 ms, repeat period: 125 ms


    # ir-keytable -a /storage/.config/rc_maps.cfg -s rc0
    Segmentation fault (core dumped)