[8.2.3.1] LibreELEC 8.2 for S905/S905X

  • I understand that it is impossible to support all those cheap chinese devices. But SBCs are in my opinion not a compensation for those chinese boxes because they are hardware wise not ready-to-use mediaplayers. In my opinion it would be great if the list of supported mediaplayers like WeTek/Minix could carefully be increased.

    How are they not ready-to-use mediaplayers? They work far better imo, CEC power on/off actually works, they can be overclocked, all drivers are open sourced and everything works.

  • I agree that it is not a good idea to support “everything out there” including huge number of noname devices. Major and proven manufacturers should be supported of course.

  • With the current state of LE (regardless 8.x, 9.x) for S912 (H96Pro+), I do NOT see any benefits in running it on such box, in comparison to properly tuned Android 7.1.2 ROM.

    In fact - I've recently switched back to Android and it suits my needs perfectly fine.

    Just bare in mind - this is a special case (my case): H96Pro+ (S912)... ;)

    One big reason why I want LE, not Android, as my underlying OS is that Android denies rename/delete permission to my USB media drives. I only want a media appliance running KODI with full read/write permission to my media drive.

  • Hi, I can successfully install this build on some s905 boxes, but I'm having trouble getting it installed on a Mygica ATV 495 pro box that has an s905 with 2 gb of ram, I have the factory firmware and the script that forces the update is factory_update_param.aml and the image needed is in .zip format. Are there instructions on how to convert the correct build with the correct device tree into a .zip that can be loaded onto the box nand? Since I already have an sd card that I can recover the box with I'm not too worried about running libreelec from sd, but that would also help if it's possible. Thank you for any help.

  • Are there instructions on how to convert the correct build with the correct device tree into a .zip that can be loaded onto the box nand? Since I already have an sd card that I can recover the box with I'm not too worried about running libreelec from sd,

    S905: First look for device tree with your box name in it. In case there is no such image, download generic one.

    example: for a box with 2GB of RAM and 100Mbit Ethernet, you should download gxbb_p200_2G_100M.dtb

    Do not run before you can walk.

    Burn image to SD Card with correct device tree.

    [HOWTO + FAQ] Install community builds on S905/S905D/S905W/S905X/S912 device

    Run it and make sure everything works.

  • How can I fix this bug "1969" in LibreElec 8.2.3.1...it's a plugin for serials... under W10 is Kodi 17.6 it's ok....both clean installations...

    edit: LE 9 Alpha from adamg it's ok..

    Edited 2 times, last by otava5 (April 2, 2018 at 8:06 AM).

  • Before 8.2.2.2 if your device did not have MAC address burnt into the board, it would either use a random MAC address or one of default ones: co:ff:ee:00:01:9f or 00:15:18:01:81:31. This is problematic if you want to use more than one device without factory MAC in one network. 8.2.2.2 introduces generating MAC address from SoC serial number for devices that don't have factory MAC - this means that devices with proper MAC address should be unaffected (tested on all my devices), while those without MAC will get a shiny new address that will be unique and persist over reboots.

    Did anyone get this working? The 8.2.3.1 in my living room keeps getting the 00:15:18:01:81:31 MAC and none of the fixes I found perusing threads (udev, system.d, etc) work. However, I am running an older Nofan Tasi aarch64 build in my bedroom and it generates a random MAC each boot. Both boxes are identical h/w && manufacturer otherwise (and it's not a Realtek LAN adapter) - I knew about this ages ago but for some reason, it bit me in the rear today. I also booted the GDPR-2 Leia builds in Amlogic off USB and noticed it's generating the 00:15:18:01:81:31 MAC -- Anyone have a surefire method that works? I can generate a random one using dbus-uuidgen and even try to write to via fw_setenv but no luck.

    I tried everything mentioned in this thread:

    How to change duplicate MAC Address

    But the stock kszaq build and GDPR-2 build still default to that MAC. Only the aarch64 build (it's an older 8.2 from Nov 04 2017) generates a new MAC each reboot.

    Really looking for a hail mary here... I may end up building my own aarch64 build again to test with.

  • Did anyone get this working? The 8.2.3.1 in my living room keeps getting the 00:15:18:01:81:31 MAC and none of the fixes I found perusing threads (udev, system.d, etc) work. However, I am running an older Nofan Tasi aarch64 build in my bedroom and it generates a random MAC each boot. Both boxes are identical h/w && manufacturer otherwise (and it's not a Realtek LAN adapter) - I knew about this ages ago but for some reason, it bit me in the rear today. I also booted the GDPR-2 Leia builds in Amlogic off USB and noticed it's generating the 00:15:18:01:81:31 MAC -- Anyone have a surefire method that works? I can generate a random one using dbus-uuidgen and even try to write to via fw_setenv but no luck.

    I tried everything mentioned in this thread:

    How to change duplicate MAC Address

    But the stock kszaq build and GDPR-2 build still default to that MAC. Only the aarch64 build (it's an older 8.2 from Nov 04 2017) generates a new MAC each reboot.

    Really looking for a hail mary here... I may end up building my own aarch64 build again to test with.

    Mac address is generated from your CPU serial number if one is not burned to the board.

  • Thanks guys, I'll keep trying here and see what I can do.

    Poida yup, none of the methods work for me. As stated, I have 1 box that behave, 1 that doesn't and the one that behaves misbehaves when I boot an older (not that much older I think) copy of 8.90 Leia build via USB. The 8.2 build I have flashed to the unit itself that behaves works fine.

    GDPR-2 right, that's what I read -- perhaps I have two boxes with the same CPU serial #? Just weird how two boxes come up with the exact same MAC address (again, they are by the same manufacturer, had a crappy Android Lollipop on them, etc etc...)