Posts by ehoitinga

    12.0-nightly-20230309-5426f16: Keyboard working.

    12.0-nightly-20240406-4b7642d: Keyboard not working.

    Will investigate in between versions to see were it went wrong.

    Ok, after trying 10 builds I figured out that the problem was introduced in LibreELEC-Generic.x86_64-12.0-nightly-20240311-86dba4b.img.gz.

    In build LibreELEC-Generic.x86_64-12.0-nightly-20240309-f3b540d.img.gz the keyboard is working fine.

    Just to find out what changed...

    Code
    By not connecting to Mint too I'm out of ideas.

    On my mint laptop (21.1, 6.5.0-26) the keyboard doesn't even appear in the search list of blueman app. However, with the bluetoothctl command I see it detected and subsequently deleted and it will not pair.

    I thought perhaps my keyboard was broken. But is does pair to my IPhone 11 and I can use the keyboard. So I think I can conclude that the keyboard works.

    It is a rather old keyboard but I like its backlit Fn functions. I also use "Sybu" on my Iphone but blind typing on a smartphone is somewhat difficult ;)

    With the keyboard there was also a USB 2.0 receiver included, but unfortunately I can't find it anywhere. So no possibility to test it.

    A bit odd that it stopped working when jumping from v11.3.1 to 11.95.1. Did something change in the Bluez package? It is the only thing I can think of.

    Quote

    1. Which version did you use before LE 11.95.1?

    Latest version on branch 11.0. I believe it was 11.3.1.

    Quote

    2. bluetoothctl devices is only listing known devices, please use bluetoothctl info 00:1F:20:04:44:38

    Quote

    3. Executed udev rules can be watched with udevadm monitor

    System-Tools Addon v11.80.6.1 installed.

    Switching the keyboard on:

    Code
    KERNEL[619.454889] add      /devices/pci0000:00/0000:00:14.0/usb1/1-8/1-8:1.0/bluetooth/hci0/hci0:256 (bluetooth)
    UDEV  [619.460616] add      /devices/pci0000:00/0000:00:14.0/usb1/1-8/1-8:1.0/bluetooth/hci0/hci0:256 (bluetooth)

    Switching the keyboard off:

    Code
    KERNEL[660.720462] remove   /devices/pci0000:00/0000:00:14.0/usb1/1-8/1-8:1.0/bluetooth/hci0/hci0:256 (bluetooth)
    UDEV  [660.727544] remove   /devices/pci0000:00/0000:00:14.0/usb1/1-8/1-8:1.0/bluetooth/hci0/hci0:256 (bluetooth)

    Deleting the keyboard:

    Code
    KERNEL[685.982996] remove   /devices/virtual/misc/uhid/0005:046D:B309.0001 (hid)
    UDEV  [685.990624] remove   /devices/virtual/misc/uhid/0005:046D:B309.0001 (hid)

    Switching on and Pairing (connect button) the keyboard:

    Code
    KERNEL[1192.205704] add      /devices/pci0000:00/0000:00:14.0/usb1/1-8/1-8:1.0/bluetooth/hci0/hci0:256 (bluetooth)
    UDEV  [1192.212845] add      /devices/pci0000:00/0000:00:14.0/usb1/1-8/1-8:1.0/bluetooth/hci0/hci0:256 (bluetooth)
    KERNEL[1207.450247] add      /devices/virtual/misc/uhid/0005:046D:B309.0002 (hid)
    UDEV  [1207.460254] add      /devices/virtual/misc/uhid/0005:046D:B309.0002 (hid)
    Quote

    4. After installing system System-Tools Addon evtest list and test available input devices.

    Hmm, I dont' see the keyboard here. Info on keyboard MAC address seems OK. Executed udev rules seems OK.

    Further info:

    Bluetooth wireless interface is present.

    Simple binaries that aren't compiled/linked against other libraries can often be copied from e.g. Ubuntu and used on LE. So I would try copying hid2hci to /storage/bin/ and the 70-hid2hci.rules file to /storage/.config/udev.rules.d/ and then mod the rules file (with the changes in the patch) so it runs "/storage/bin/hi2hci" (explicit path) not "hid2hci" (which relies on the binary being in $PATH).

    I copied hid2hci from my Mint laptop to /storage/bin:

    copied the 70-hid2hci.rules file to /storage/.config/udev.rules.d and adapted the content using the patch:

    Deleted the keyboard, rebooted and repaired the keyboard. No luck. Still the same problem.

    I can successfully pair it and LE says its connected but keystrokes are not arriving. When I disconnect and re connect it in LE, I get the BlueTooth error "br-connection-create-socket".

    Probably because bluetoothd is not complied with --enable-hid2hci as mentioned by mglae ?

    Thank you very much for the help so far.

    LE is including the patch in the source but bluez require --enable-hid2hci since 5.55.

    hid2hci was only included until LE 9.2.8.

    Should be easy to add again but have to test it first (currently fighting against xz 5.4.6 source archive not being available, maybe someone can upload it to the LE source mirror).

    This one?: https://sourceforge.net/projects/lzmau…ar.bz2/download

    Some update:

    LibreElec installed on Intel NUC7I5BNK.

    Code
    mediacenter:~ # bluetoothctl devices
    Device 65:A1:D3:FC:B2:93 65-A1-D3-FC-B2-93
    Device 00:1F:20:04:44:38 Logitech diNovo Edge
    Device E7:A7:BD:33:69:8E 846B213F84AC33ADE9
    mediacenter:~ #

    Looks like it is connected, however keystrokes are not arriving.

    Hi,

    After updating LibreElec to version 11.95.1 my Dinovo Edge keyboard no longer works.

    When switching on the keyboard the green Bluetooth indicator turns green for a moment
    indicating the keyboard is connected. In LibreElec --> Bluetooth it says that the keyboard
    is connected. Tried to delete it and pair it again, but no luck.

    Anyone else having problems with bluetooth keyboards?

    Thank you,

    Erik

    Hi there,

    I think the problem was on the d-link switch. I can configure the spanning tree portfast mode for each port in 'disabled', 'edge' or network. The default setting for the switch is 'network'.

    When I set eth3 (the port where LibreElec is connected to) to 'edge' my library shows up.

    It look like it's solved but it still leaves me with a big WHY? Any network specialist here who can enlighten this behavior?

    Thanks,

    Erik

    Hi There,

    I am experiencing some strange behavior of LibreElec on a NUC7I5BNK. I have several Sonos devices in my home. Two of these devices are connected to the LAN to force Sonos to use its mesh wireless network rather than the home WiFi. All other Sonos devices connect through this Sonos wireless mesh network. The Sonos devices make use of STP because one device can be connected with wireless and LAN at the same time. Below you'll find a picture which explains the connections concerned.

    My LibreElec database/library is on a MySQL server on my nas-server as well as my media files. Now here is the strange thing: when I enable STP on the DGS-1100-05-PD switch, LibreElec says "Your library is currently empty". When I disable STP on the DGS-1100-05-PD LibreElec shows my library as it should.

    I can ping the NUC7I5BNK from my laptop and when I login in to the NUC7I5BNK with ssh I can ping my nas-server. All with reasonable response times.

    My kodi.log file (STP enabled on DGS-1100-05-PD)

    In the kodi.log file I see error messages that state that kodi can not connect to the MySQL server. But why? A NAS server is not a switch and has therefore nothing to do with STP. My NAS however has 4 ethernet ports and supports link aggregation, but only one ethernet port is used to connect it to the Netgear switch and LACP is not enabled as far as I now.

    I'm a bit puzzled here... Can anyone shine a light on this.

    Thanks,

    Erik

    in continuance of my previous solution proposal to have dual audio out (in my case SPDIF + HDMI), at last I found the response from the gentleman that was kind enough to keep this alive and running "xhbl", other contributors as well.

    This is also a nice solution for those who have dual audio. Personally I don't need it because video and audio are output by HDMI from the NUC. The HDMI matrix splits the sound to Toslink (to my Sonos) and the video to HDMI (To my TV).

    Anyhow thanks for your solution.

    Hello again,

    I have received my new HDMI matrix. It did solve the problems I had concerning the loss of video signal and the screen turning pink when changing a HDMI channel. The playback (passthrough) of AC3 streams have seemed to improved a bit.

    1. The passtrough of AC3 DD5.1 streams works now more or less. No longer necessary to switch-off pass-through to get sound. However, when I start a movie with a AC3 or E-AC3 stream sometimes I have sound directly, that is, withing one or two seconds. But sometimes it takes between 20 seconds upto two minutes before the sound comes in.

      I have downloaded some testfiles from here:

      a) Amaze MP4 H.264 1080pDolby AC3 5.1 640 kbps. This file gives a Dolby Digital 5.1 input signal on my Sonos playbar but with the same symptoms described above.
      b) Countdown v1 (Lossless) MKV H.264 1080p Dolby TrueHD 5.1. This file gives a Stereo input signal on my Sonos playbar.
      c) Dolby 3D Glasses Return Trailer Flat (2D) MKV H.264 1080pDTS-HD MA 5.1. The AC3 stream gives a Dolby Digital 5.1 input signal on my Sonos playbar. The PCM 5.1 gives a Stereo input signal on my Sonos playbar. and the AC3 5.1 gives a Dolby Digital 5.1 input signal on my Sonos playbar but with the same symptoms described above.
    2. My audio output device and passthrough device keeps on switching seemingly random from "ALSA: HDA Intel PCH, MEI Panasonic-TV on Displayport #0" to "ALSA: HDA Intel PCH, MEI Panasonic-TV on HDMI #0" back an forth. It seems to happen randomly. When it is on HDMI #0 and I do a reboot it is back on Displayport #0 and then upon usage it turns back to HDMI #0 again.

    Below you will see screenshots of my audio settings and of the changing output and passthrough device:

           

           

    I am a bit puzzled here. The above symptoms can they be accredited to the buggy Intel LspCON firmware? Will this patch give any hope?

    Erik

    In the defense f the 6CAYH, it decodes flawlessly x265 10bit with ~5%/all CPU and sound is there if pass through is disabled.

    I ran a 140Mbs x265 4k UHD 8 bit and 10 bit version test files . CPU usage around 20% so the video capabilities are more than OK. And yes the sound is there but the signal is stereo. And like you, i like to watch a movie just by pressing play and not having to tinker with various settings to get sound. Good to hear that I'm not alone in this saga.

    Erik

    As weird as it sounds, I have have an 2013 Panasonic TV with an 1.4 HDMI port and could perform the LSPCON update to FW 1.66. Booted from a Windows 10 USB HDD

    To give it a go I installed Windoze 10 on the NUC and all the new Intel drivers for this NUC. After that I did a Windoze update. When I ran the HDMI firmware update utility in Windows I had no problems updating the firmware to 1.66 this time, despite all my devices, except the NUC, are HDMI 1.4a.

    The first time I tried I did not update my Windows installation an Intel drivers. That must have been the problem. Firmware is updated to 1.66 right now.

    Unfortunately this did not solve my problems with the AC3 DD5.1 streams. Still no sound. Only sound when I switch off passthrough.

    When I installed Windows I installed Kodi 17.4 to see if I had the same problems with the AC3 DD5.1 streams. Unfortunately on Kodi for Windows I had exactly the same problems, Only sound when I switch off passthrough.

    I really think that my HDMI matrix is dying. I will receive a new one today or Monday. Keep you posted with the results.

    Erik

    I'm asking, because I have seen the same (more HDMI ports) on a different (newer) Linux kernel on my NUC6CAYH.

    The strange thing is that these two 'ghost' HDMI port cannot be found as a device in the out put of alsa-info.sh but are listed in the output of aplay -l.

    Internally it's connected to DisplayPort -> LSPCON -> HDMI.

    OK. That's why I see Displayport #0 listed as available audio devices. The strange thing is that sometimes, and quite randomly, it is listed as Displayport # and sometimes as HDMI #0.

    I have no HDMI switch connected, and still see 3 HDMI devices, without the patch. I will look into it more after upgrading to 8.2. Of course you can disconnect the switch to find out, but I highly doubt it's the switch. Your alsa-info.txt only lists 3 HDMI/DP ports. As a test: try booting with 'options snd-hda-intel probe_mask=1', and there should be no HDMI.

    This was just a wild guess. I rebooted with the matrix disconnected and the bogus HDMI port still pop up.

    As weird as it sounds, I have have an 2013 Panasonic TV with an 1.4 HDMI port and could perform the LSPCON update to FW 1.66. Booted from a Windows 10 USB HDD.

    I use "Windows to Go" to create a bootable USB 3.0 external disk to update the HDMI firmware. Windows is updating right now to have the latest drivers. Now I know again why I ditched Windoze years ago. It takes a ridiculous amount of time to update. When it's updated I try the HDMI update. Will get back with the results.

    "Number of channels: 2.0"

    Do you think that might be part of the problem? Try setting the number of channels to 5.1.

    Also make sure Settings > Player > Video > Sync playback to display is OFF if you want passthrough to work.

    OK. To describe the AC3/Passthrough problems more precisely I use these three movies as a test.

       

    I did a fresh install of version 8.1.2 with exactly the same configuration as before. The movie Ice Age The Great Egg-Scapade gives always sound. The other sometimes give sound immediately or after 1 or 3 minutes or longer or not at all. I can't see a logic. When I have sound my Sonos app sees the audio signal of these three movies as Dolby Digital 5.1.

    I tried different versions of LibreElec (8.01, 8.02, 8.1.1, 8.1.2 and milhouse stable 8.2) and the symptoms above are the same with every version. I though suspect my HDMI matrix of dying. Today when I switched from the mediacenter to my STB I had no image. I had to reset my HDMI matrix to get image. Resetting my HDMI matrix in case of the above audio problems did not give any result. Symptoms stay the same.

    I ordered a new HDMI matrix which support all the new video and audio formats. I expect yo receive it end of this week or Monday. I'll keep you posted if this change anything.

    Erik