Posts by dtech

    As I saw, there are versions of the T95M 4K equipped with S905 (gxbb) and S905X (gxl) and S905W (gxl) SoCs.

    However, if your box really contains an S905 SoC (gxbb), try a dtb starting with meson-gxbb-. For example, try meson-gxbb-p200.dtb, if you have Gigabit Ethernet, or meson-gxbb-p201.dtb if you have a 10/100 Mpbs Ethernet.

    Oh, and don't forget the toothpick method.

    At first glance, it works well. :thumbup:

    Did the driver load itself at boot, or did you start it manually with modprobe?

    In the meantime, I backported the persistent journal option to 9.2.8.10, so you can now send me a full system log.

    You can turn this on under Settings -> LibreELEC Settings -> System -> Logging.

    This setting requires a reboot, after that please send a link generated by journalctl | paste. Thanks.

    Can this firmware be updated automatically through auto-update?

    This is currently in a different branch, so please turn off auto-update for now until the final code is ready.

    If you turn it on, it will update to a version that still lacks this driver, so it would be good to avoid this.

    However, you can update the previous version with the tar on the link, so you don't have to make a new installation.

    Just download the tar and copy it to the Update folder via Samba.

    Well, it seems that the required module has not been loaded. Let's try to load it manually:

    Code
    dmesg -c > /dev/null
    modprobe 8723as
    dmesg | paste

    And thanks for the photo. ;)

    Btw, also doesn't work Power On from remote control. Can anything be done about it?

    Unfortunately, it looks like I need to dig deeper into the code, and the HD18Q image probably wouldn't be fully compatible with this 2022 box release. I feel like we can't move forward until I get a piece of this myself.

    But there is a new question: Does the box really turn off under Android when the power button is pressed on the remote, or does it just go to standby?

    P.s.: We just need to find out who is selling exactly the same box. If anyone knows accurate information about this, please let me know. Thanks.

    It seems that the SDIO bus is re-initialized, but the SDIO adapter does not appear afterwards.

    Next step: I would need an lsmod output.

    Unfortunately, I don't have an S805 device with SDIO wifi, so if I don't manage to get such a box, I can only rely on user tests.

    I tried to take the log from alexelec and it looks like what you said. It doesn't fit, does it?

    From a completely different operating system, I won't really get information about the status of the driver I modified, so it's unfortunately useless for me. However, I/O errors related to the SD card are also present here, only in a smaller volume.

    It would be worth running this in overwrite mode, before you write an image to the SD card again.

    Now some SD card related R/W error keeps repeating itself over and over again... :(

    The log should start like this:

    Code
    [    0.000000@0] Booting Linux on physical CPU 0x200
    [    0.000000@0] Linux version 3.10.108 (dtech@LibreBuilder) (gcc version 8.3.0 (GCC) ) #1 SMP PREEMPT Sat May 27 21:18:35 CEST 2023
    [    0.000000@0] CPU: ARMv7 Processor [410fc051] revision 1 (ARMv7), cr=10c5387d
    [    0.000000@0] CPU: PIPT / VIPT nonaliasing data cache, VIPT aliasing instruction cache
    [    0.000000@0] Machine: Amlogic Meson8B, model: AMLOGIC
    [    0.000000@0] physical memory start address is 0x200000
    ...

    Unfortunately, I can't help you until the essential part is invisible. I would recommend trying a different SD card first.

    I think that's behind the marketing idea: bigger is better...

    I think the manufacturer simply has a bunch of empty MXQ boxes that they haven't sold before. By the way, the "5G" label on the box is also strange because the RTL8723AS only works on 2.4GHz. So it's the usual Chinese industrial scam again. I haven't been surprised by this in a while...

    Meanwhile, I found an obvious error in the driver that I hadn't noticed before. I fixed this, please try this new snapshot version:

    Index of /snapshots/20230515_rtl8723as_67db88d/

    You should probably boot successfully with this image, but even if the wifi is working or not, I would ask for a journalctl | paste output.

    I just enabled passthrough but nothing changed. The GUI volume has always been set at 100%

    If passthrough is active and the correct output and currently used encoding are enabled, it doesn't matter what volume is set under Kodi, because then the incoming bitstream is passed directly to the output without processing.

    Edit: Okay, wait a second...

    ^ You didn't set this to 100%, did you? Because this is only the volume of the system sounds, not the master volume.

    You can adjust the master volume of Kodi using the +/- buttons on the number pad.

    Of course, with passthrough, this will have no effect, unless you only send basic PCM audio.