Posts by chewitt

    If it works you'll see duplicate wireless network names in the connections list; one will be from the wlan0 interface and another will be from the wlan1 interface. In the long run (once you have a working USB device) it might make sense to blacklist the driver for the internal wireless card to prevent it loading at boot-time, then you will only see one set of networks.

    Suzhou SmartChip Semiconductor Co.,Ltd FCC ID Applications (2AKCE) - the phone diverts to voicemail and both voice/email messages left (in Mandarin by a native speaker) are not answered. Via third-parties we've been offered the driver in binary form for any Android kernel version we care to ask for, but nobody has sources, which is what we need. I can assure people we've been quite creative on trying to reach them - but so far no results.

    It's possible, but Amlogic boxes currently use an older 3.14 Linux kernel so the list of USB wireless devices that might work is smaller than hardware like Raspberry Pi which uses a current kernel. There is no definitive list of devices we support (too many combinations to document).

    It was a group decision in OE days to discontinue them and there are no plans to reinstate i386 builds. It's still possible to re-add support and self-build a working i386 image though, e.g. GitHub - maideii/LibreELEC.tv-Intel-i386: Just enough OS for KODI appears to have done the core mods for i386 builds with an Intel GPU and it looks reasonably up-to-date, so if you clone his branch you should be able to run "PROJECT=Generic ARCH=i386 make image" and get something that works (self-build instructions are in the wiki). If you have AMD or nVidia GPU's there are probably some missing commits, but nothing hard to track down; from memory you can crib nVidia bits from my appletv-7.0 branch.

    Code
    RTL871X: ERROR invalid mac addr:00:00:00:00:00:00, assign random MAC

    The connection manager (connman) caches configurations in a structured format keyed on the connection type and MAC of the interface so ^ that's the problem. It doesn't forget.. it's just (correctly) presented with a new randomly generated MAC on each boot.

    Code
    echo "options 8812au rtm_initmac=00:e0:4c:37:67:35" > /storage/.config/modprobe.d/8812au.conf
    or
    echo "options 8812au mac_addr=00:e0:4c:37:67:35" > /storage/.config/modprobe.d/8812au.conf

    ^ this is a complete guess based on 60 seconds of non-specific Googling, so I make no guarantees it will work. It is technically possible to change the MAC via ethtool commands, but really it should be done at driver load time which is what we're trying to achieve.

    I don't have the code knowledge to spot if this is an issue with the driver code or not, but the lack of general comments in Google search makes me suspect it's a manufacturing issue where a blank/null MAC address is set in the device. You can argue whether that's a fault as it wouldn't cause an issue in Win/macOS, but it does here. It probably indicates a cheap device where the manufacturer hasn't been assigned unique hardware ID ranges for their devices.

    We know who creates the chip. They don't respond to emails in English or Chinese. Neither Amlogic or Rockchip (where we have direct contacts) have any record of an authorised hardware manufacturer using the chip (else they would require accreditation and would have driver sources in their code archives). I'm sure sources will leak eventually, but until that happens we are out of luck.

    Make sure you follow instructions correctly. You do not copy the contents of the zip file to the SD card. You copy the contents of the folder within the zip. It's subtly different, but one results in a working "recovery" SD card that will install to eMMC (WP2 has no NAND) and the other does not.

    ^ The above is based on the composite output config I used to ship in AppleTV builds. You will need to change the TVStandard option to match your TV and store the file as /storage/.config/xorg.conf and "systemctl restart xorg" to effect the change.