Posts by dtech

    Hmm, the remote.conf file on my Nexbx A1 clone has no key-code 28. Maybe that's why OK button does not work. This remote also worked with H96 Pro+. File is attached below and it has comments in the file but I am not sure how accurate are the comments.

    That would be better, but it's a completely different code config:
    factory_code = 0xfe010001
    ^ The sample above, which I linked from GitHub, also had this code. If this were good, then the remote control should work by default, because this is included in my image by default.

    However, what you need should be:
    factory_code = 0xdf000001

    I just found another one with the right code, which already has 28 for OK button (attached). I think this might be good.

    The problem is that the remote.conf what I found has no explanations, so I can't bind the function button to the IR code.
    In the modifications to the IR code tables I made earlier, I commented everything, so for example the OK button code is 28:

    LibreELEC-AML/projects/S912/devices/Generic_S912/filesystem/etc/amremote/remote.conf at master · dtechsrv/LibreELEC-AML
    'Just enough OS' for Kodi for some Amlogic TV boxes - dtechsrv/LibreELEC-AML
    github.com

    You can find keyboard codes for Linux here, most of them are the same on Kodi:

    linux/include/uapi/linux/input-event-codes.h at master · torvalds/linux
    Linux kernel source tree. Contribute to torvalds/linux development by creating an account on GitHub.
    github.com

    However, if this is the only button that doesn't work, try this first:
    0x06 28

    But, if your remote.conf contains any comments or explanations, you can send it and I can help you make a working one based on them.

    I got rather distracted with some recent discoveries on the hardware decoders. I need to circle back and cherry-pick the older device tree and then do some fixups for changes; some of the kernel bindings have evolved since LE11 days.

    If I have some free time, I will try to collect my fixes from that time slot, because the network stack fixes of the S905Ds (similar to the GXMs) were never included in the official images, but they were implemented incorrectly. This affected most internal tuner boxes and the Vero4+ as well. In addition, OSMC's support for the latter has ended recently, so it is possible that the number of LE users may start to increase on these boxes.

    Thanks for this, but my original query was pertaining to the non working lan on the box.

    Sorry, I didn't mean to off-topic. As far as I remember, the T95Z Plus has the same ZTE chip instead of Realtek GbE.

    Regarding this, I sent a DTB patch to chewitt a few years ago (including LED GPIOs and VFD display), and it was included, sometime around the LE11 test versions. However, since then this DTB has completely disappeared from the current AMLGX images.

    The Ethernet patch was based on this:

    T95Z Plus (ethernet working for kernel 5.3.0)
    I tried to install Armbian_19.11.3_Aml-g12_buster_legacy_5.3.0_20191126.img.xz to my T95Z Plus box (S912 with 3GB RAM) but got only wifi running. Any ethernet…
    forum.armbian.com

    I've since checked, and the Ethernet chip of T95Z Plus is exactly the same as yours:

    No progress due to other priorities over the last week. Read: https://wiki.libreelec.tv/hardware/amlogic#install2internal

    I don't understand why, but this script was never available under the name install2internal, and it's even referred to by that name in the official documentation, but in fact its original name was always installtointernal.

    I still have this script in my legacy build, and I've even received feedback from users who couldn't run the install2internal script. Of course, if it has a different name, how the hell would it run?! I don't know who started spreading this typo, but it's already driving me crazy...

    installtointernal -> This is the correct name, it's always been this!

    On my legacy amlogic images, only the amremote format is supported, so you need a compatible remote.conf file.

    I only found one "in theory" match for yours, so I don't know if it will work.

    Please download the attached file, extract its content, and place it here: /storage/.config/, then restart your box.

    Don't be confused by the file name, the remote control of the A1 and the A95X are identical.

    the 45 degree rotation on the chip makes Amlogic unlikely

    There is a HD18Q marking on the PCB, so I'm 99% sure that there is an Amlogic S805 SoC under the heatsink.

    It looks very similar to the Eny EM6Q-MXQ PCB, which also had a 45 degree heatsink, although there is no battery for the RTC.

    The faulty lwfinger driver was defined here: https://github.com/dtechsrv/Libre…88EU/package.mk . Perhaps you could update this, as it appears to be your repo after all? ;)

    It may perform better on some devices, while it may not on others. Unfortunately, this is a double-edged sword, and since I can't test it, I wouldn't want to modify it, because it's an inherited driver from the official LE repo. In addition, I already indicated that I have never used this driver at all.

    ... I have never used the RTL8188EU made by lwfinger in my images.

    My LE-AML repo is freely forkable, so anyone can modify it. However, I don't want to put more devices and/or drivers in my repo that I can't test. (There are already several devices in 3rdParty that I don't own.)

    I've discovered that the WiFi driver we've been using for these builds (lwfinger/rtl8188eu) is causing serious instability with this media box. When streaming over WiFi, the media player is prone to lock-up / freeze. Further inspection of the kernel crash logs revealed that it was this wifi module causing the crashes...

    There is a separate RTL8188EU-aml package for Amlogic devices, so I have never used the RTL8188EU made by lwfinger in my images.

    For example, for the options file for S805 project looks like this: https://github.com/dtechsrv/Libre…05/options#L118

    However, this has been a topic here before (#337), but since it is quite difficult to trace back where was published the source code you used, I don't know what you guys actually included earlier.

    Updated zip containing LibreELEC-AML 9.2.8.16 and inputstream adaptive addon (for youtube)

    Feel free to use my WeTek_Play binary repository from my server, you just need to change the end of the file in your projects/Amlogic_MX2/options source:

    Code
      # Addon Server Url
        ADDON_SERVER_URL="http://libreelec.dtech.hu/packages"
    
      # set the addon dirs
        ADDON_PATH="$ADDON_VERSION/WeTek_Play/$TARGET_ARCH"
        ADDON_URL="$ADDON_SERVER_URL/$ADDON_PATH"

    The inputstream.adaptive addon that you attached to the update file is already there... ;)

    I have a tv box that I migrated from Android to LE OS, whose internal memory nand is 128 GB.

    I don't really believe that any manufacturer has made an S805 box with such a high capacity NAND or eMMC. The kernel recognizes the real size of the flash memory, so either Android is lying or you ran into an excessive marketing ad when you bought the box.

    Regarding the compile errors, I suspect that you are trying to compile on a system that is too new, and the tools delivered with the actual distro version are not always backwards compatible. I'm still using Debian 10 to compile my LE 9.2 images.

    I git clone it and use make, but apparently I f up. If it's much more complex like this, i'll just use it without internet, but perhaps there was an easy way. cheers!

    Compiling a kernel module is a bit more complicated than running a simple make.

    The package created for the build of the module used under LE-AML 9.2 can be found here:

    LibreELEC-AML/packages/amlogic/linux-drivers/RTL8723AS-aml/package.mk at master · dtechsrv/LibreELEC-AML
    'Just enough OS' for Kodi for some Amlogic TV boxes - dtechsrv/LibreELEC-AML
    github.com

    However, ilmich uses mainline kernel (6.x) in his builds where this module is probably no longer usable.

    I have not noticed any boot time change between USB boot (multiboot) and SDcard only boot.

    Unfortunately, I don't know what boot scripts you actually use, since you said you didn't use the toothpick method again. The boot order may also be different, so the information is unfortunately not relevant.

    By the way, it might be device specific, but for me it takes 10 seconds for the USB ports to be scanned, which is a lot for me.

    Normally, once boot, I always keep it on or on standby mode & use remote to wake-up.

    In this build, power-off have only 3 option (power off, custom shutdown & restart), don't have 4th option for suspend (sleep). I had to override in setting - power saving - Force suspend instead of poweroff .

    For many devices, standby mode is not the default or is not available at all. This is also a uboot-specific property defined by the manufacturer. (By the way, this "Force suspend instead of poweroff" is only available in my builds, because it's my own solution.)

    And, if someone boots from an internal NAND/eMMC, they always have to wait to scan the USB unnecessarily. And, since many people can't even use the suspend option, every start will take forever. This is not good, so we need an intermediate solution that is acceptable to everyone, and I am currently working on this.

    The original 3-script solution is unnecessarily overcomplicated. I'm currently testing the single-script solution and I introduced a small modification, so the USB boot is inactive by default, but it can be simply activated by an fw_setenv enableusbboot 1 command.

    I installed docket but how to install portainer? It's not in add-on. Needs to use CLI?

    It is recommended to use the CLI, but there are many places describing how to use it, so I don't want to write complex documentation for this.

    Code
    mkdir /storage/.config/dockers
    docker run \
     --name=portainer \
     --restart unless-stopped \
     -v /var/run/docker.sock:/var/run/docker.sock \
     -v /storage/.config/dockers/portainer/data:/data \
     -p 9000:9000 \
     -d portainer/portainer-ce

    From your attachment, unzipped 3 files in root folder ( aml_autoscript file replaced ) of USB & booted straight from USB flash-drive. I already had multiboot activated from previous installed so I do NOT have to use toothpick first time.

    Thanks for the test, maybe I will restore these boot scripts on the 3rdParty k200 and n200 images.

    Edit:

    In the meantime, I tested it and now it works for me too, but whether I boot from USB (or not), it causes +10 seconds in the boot time, which I think is too much. With the current script, the time elapsed until the LibreELEC logo appeared on the M8S+ from power-on was 14 seconds, with the restored multiboot scripts it was 24 seconds.

    Basically, I don't know how many people would use USB boot, but I'm strongly thinking about reworking these bootscripts because I don't want to restore them in their current state.

    ... it looks like the same "search for boot files" stuff used everywhere.

    I'd still appreciate some links, because many of the developers building S8X2 on the old BSP kernel have simply deleted their GitHub repos. (Demetris, balbes150, etc.)

    By the way, if these are really the affected files, then I indicate that I removed them because none of my boxes were able to boot from USB (omitted u-boot function?), despite the above scripts. For this reason, I deleted them, and only the aml_autoscript remained, without the USB boot option, because the original 3-script solution increased the start of the boxes by about 10 seconds, and this was unacceptable to me.

    MakaaiD Please try these boot scripts with my image: https://libreelec.dtech.hu/.upload/s8x2-old-bootscripts.zip
    Of course, the toothpick method will be necessary again to load them.