hexdump

maybe build your own kernel with lima disabled (i'm not sure, but i think it might conflict with the old mali driver even if the module is not loaded) ... make sure you apply those two patches:
https://github.com/hexdump0815/linux-mainline-and-mali-generic-stable-kernel/blob/master/misc.av7/sunxi-drm-gem-cma-Export-with-handle-allocator.patch
https://github.com/hexdump0815/linux-mainline-and-mali-generic-stable-kernel/blob/master/misc.av7/sunxi-drm-sun4i-Add-GEM-allocator.patch
 
here are my own notes on getting this to work, which are verified to work up until v5.4 kernels:
https://github.com/hexdump0815/linux-mainline-and-mali-generic-stable-kernel/blob/master/readme.av7-mali-sunxi
 
good luck and best wishes - hexdump

another thing to keep in mind is that your user needs access to /dev/mali - usually granted via some udev rule and the video group (usermod -a -G video username) ... example udev rule: https://github.com/hexdump0815/imagebuilder/blob/master/files/extra-files/etc/udev/rules.d/50-mali.rules
 
good luck and best wishes - hexdump

another thing to keep in mind is that your user needs access to /dev/mali - usually granted via some udev rule and the video group (usermod -a -G video username) ... example udev rule: https://github.com/hexdump0815/imagebuilder/blob/master/files/extra-files/etc/udev/rules.d/50-mali.rules
 
good luck and best wishes - hexdump

this is just a quick note that in my experience the amount of tv boxes with fake specs has grown quite a bit in the last months and that this is something to always have in mind when getting a box for a surprisingly cheap (i.e. quite a bit cheaper than usual or most of the other offerings) price - you might be lucky and it will be a bargain or you might hit one with fake specs. some examples i saw recently: a qplus 4g ram / 32g emmc ended up to be 2g ram and 16g nand, a h6 box sold as 4g ram / 32g emmc ended up as 2g ram / 16g emmc, a x96mini 2g ram / 16g emmc ended up at only 1g ram / 16g emmc, a r39 2g ram / 16g emmc with rockchip rk3229 ends up as 1g ram / 16g emmc and an allwinner h3 cpu and so on. the fake specs are not that easy to spot: in android they even fake the storage size shown in the storage settings and with a terminal installed even the "free" command tells you most of the time that the memory amount is proper. what usualy works for storage is "cat /proc/partitions" and watching for the device itself (for instance mmcblk0) - this also quickly shows you if its emmc (=mmcblk) or nand (=nand) and for memory "dmesg | grep -i mem" (do this immediately after booting android, otherwise the memory lines from the bootup might run out of the log buffer) - both of course called in a terminal app. booting one of balbes150's armbian images usually quickly shows you the real specs of the box too.
 
good luck at not ending up with fake boxes and best wishes - hexdump
 
p.s.: one thing to keep in mind is that allwinner h6 boxes always only can use 3g ram, even if they have 4g installed - this is a limitataion of the soc ...

@mar0ni - yes this is normal behaviour i think as there is some inefficient memory copying happening between the gpu and the x server and there is no proper 2d acceleration (used for such window moves) around ... but i guess @JMCC can explain it in more detail ...
 
best wishes - hexdump

btw. i doubt that the initially mentioned box really has 2g/16g at this price - see my other posting about fake tv boxes here - i guess it will be 1g/8g maybe even nand instead of emmc if you are unlucky - all this with a hacked up android letting it look like 2g/16g so that a normal user will not recognize it ...

@martos - some things which come to my mind:
did you try my t9 dtb?  - see: most probably it will not work as your box does not seem to boot off sd card at all another thing worth trying is combined sd/usb booting - see this thread:  if all that does not work then you might have to flash it via usb (i think i have read that some tv boxes have the sd card connected to the wrong mmc port and are thus not sd card bootable at all) - see here and above and below it: good luck and best wishes - hexdump

this is just a quick note that in my experience the amount of tv boxes with fake specs has grown quite a bit in the last months and that this is something to always have in mind when getting a box for a surprisingly cheap (i.e. quite a bit cheaper than usual or most of the other offerings) price - you might be lucky and it will be a bargain or you might hit one with fake specs. some examples i saw recently: a qplus 4g ram / 32g emmc ended up to be 2g ram and 16g nand, a h6 box sold as 4g ram / 32g emmc ended up as 2g ram / 16g emmc, a x96mini 2g ram / 16g emmc ended up at only 1g ram / 16g emmc, a r39 2g ram / 16g emmc with rockchip rk3229 ends up as 1g ram / 16g emmc and an allwinner h3 cpu and so on. the fake specs are not that easy to spot: in android they even fake the storage size shown in the storage settings and with a terminal installed even the "free" command tells you most of the time that the memory amount is proper. what usualy works for storage is "cat /proc/partitions" and watching for the device itself (for instance mmcblk0) - this also quickly shows you if its emmc (=mmcblk) or nand (=nand) and for memory "dmesg | grep -i mem" (do this immediately after booting android, otherwise the memory lines from the bootup might run out of the log buffer) - both of course called in a terminal app. booting one of balbes150's armbian images usually quickly shows you the real specs of the box too.
 
good luck at not ending up with fake boxes and best wishes - hexdump
 
p.s.: one thing to keep in mind is that allwinner h6 boxes always only can use 3g ram, even if they have 4g installed - this is a limitataion of the soc ...

this is just a quick note that in my experience the amount of tv boxes with fake specs has grown quite a bit in the last months and that this is something to always have in mind when getting a box for a surprisingly cheap (i.e. quite a bit cheaper than usual or most of the other offerings) price - you might be lucky and it will be a bargain or you might hit one with fake specs. some examples i saw recently: a qplus 4g ram / 32g emmc ended up to be 2g ram and 16g nand, a h6 box sold as 4g ram / 32g emmc ended up as 2g ram / 16g emmc, a x96mini 2g ram / 16g emmc ended up at only 1g ram / 16g emmc, a r39 2g ram / 16g emmc with rockchip rk3229 ends up as 1g ram / 16g emmc and an allwinner h3 cpu and so on. the fake specs are not that easy to spot: in android they even fake the storage size shown in the storage settings and with a terminal installed even the "free" command tells you most of the time that the memory amount is proper. what usualy works for storage is "cat /proc/partitions" and watching for the device itself (for instance mmcblk0) - this also quickly shows you if its emmc (=mmcblk) or nand (=nand) and for memory "dmesg | grep -i mem" (do this immediately after booting android, otherwise the memory lines from the bootup might run out of the log buffer) - both of course called in a terminal app. booting one of balbes150's armbian images usually quickly shows you the real specs of the box too.
 
good luck at not ending up with fake boxes and best wishes - hexdump
 
p.s.: one thing to keep in mind is that allwinner h6 boxes always only can use 3g ram, even if they have 4g installed - this is a limitataion of the soc ...

this is just a quick note that in my experience the amount of tv boxes with fake specs has grown quite a bit in the last months and that this is something to always have in mind when getting a box for a surprisingly cheap (i.e. quite a bit cheaper than usual or most of the other offerings) price - you might be lucky and it will be a bargain or you might hit one with fake specs. some examples i saw recently: a qplus 4g ram / 32g emmc ended up to be 2g ram and 16g nand, a h6 box sold as 4g ram / 32g emmc ended up as 2g ram / 16g emmc, a x96mini 2g ram / 16g emmc ended up at only 1g ram / 16g emmc, a r39 2g ram / 16g emmc with rockchip rk3229 ends up as 1g ram / 16g emmc and an allwinner h3 cpu and so on. the fake specs are not that easy to spot: in android they even fake the storage size shown in the storage settings and with a terminal installed even the "free" command tells you most of the time that the memory amount is proper. what usualy works for storage is "cat /proc/partitions" and watching for the device itself (for instance mmcblk0) - this also quickly shows you if its emmc (=mmcblk) or nand (=nand) and for memory "dmesg | grep -i mem" (do this immediately after booting android, otherwise the memory lines from the bootup might run out of the log buffer) - both of course called in a terminal app. booting one of balbes150's armbian images usually quickly shows you the real specs of the box too.
 
good luck at not ending up with fake boxes and best wishes - hexdump
 
p.s.: one thing to keep in mind is that allwinner h6 boxes always only can use 3g ram, even if they have 4g installed - this is a limitataion of the soc ...

@Tarzanus - you might give this approach a try:
best wishes - hexdump

you need to dd the trust.img as well - just read the thread from the post i linked to onwards for the file and more info ...
 
good luck - hexdump

i think i got it stable finally - i built u-boot for it myself (v2020.01) and did the following:
loaderimage --pack --uboot ./u-boot-dtb.bin uboot.img 0x200000 # loaderimage on x86 dd if=uboot.img of=/dev/sdx seek=16384 the dd of trust.img as described above is required still. booted this way the box seems to run quite stable. i guess the u-boot i used before was setting a too advanced memory timing and my self build evb-rk3328 u-boot seems to have a more conservative memory timing which is ok for the box. i'll tomorrow upload a readymade dd'able u-boot here in case someone else wants to give it a try.
 
i also tried to dd the spl/tpl idbloader.img and u-boot.itb too following https://github.com/u-boot/u-boot/blob/master/doc/README.rockchip but it looks the box initally boots off the emmc and only uses the u-boot on the sd-card, so the spl/tpl on the sd card were ignored.
 
best wishes - hexdump

@jernej - success! finally ... it was the atf - i now built the atf myself (v2.2) and just emptied the axp probing function (so that it does nothing anymore and just returns an error back, i.e. no axp) and voila: it boots ... with the mainline u-boot it still resets dozens of times until it finally continues to boot, but it always does at some point in time after a few seconds or minutes (with the libdram version it boots absolutely reliable) ... so we are back to trying to improve the dram setup and timing to maybe get rid of the resets - please let me know in case you have any idea what else to maybe test ...
 
a lot of thanks for all you help and support and best wishes - hexdump
 
... resetting ... U-Boot SPL 2020.01-rc3-00007-g8d8ee47e03-dirty (Jan 10 2020 - 16:34:14 +0100) DRAM:PLL = b0003500 DRAM PHY PGSR0 = 20003d DRAM PHY DX0RSR0 = 0 DRAM PHY DX1RSR0 = 0 DRAM PHY DX2RSR0 = 0 DRAM PHY DX3RSR0 = 0 Error while initializing DRAM PHY! resetting ... U-Boot SPL 2020.01-rc3-00007-g8d8ee47e03-dirty (Jan 10 2020 - 16:34:14 +0100) DRAM:MBUS port 0 cfg0 0100000d cfg1 00640080 MBUS port 1 cfg0 06000009 cfg1 01000578 MBUS port 2 cfg0 0200000d cfg1 00600100 MBUS port 3 cfg0 01000009 cfg1 00500064 MBUS port 4 cfg0 20000209 cfg1 1388157c MBUS port 5 cfg0 00640209 cfg1 00200040 MBUS port 6 cfg0 00640209 cfg1 00200040 MBUS port 8 cfg0 01000009 cfg1 00400080 MBUS port 11 cfg0 01000009 cfg1 00640080 MBUS port 14 cfg0 04000009 cfg1 00400100 MBUS port 16 cfg0 2000060d cfg1 09600af0 MBUS port 25 cfg0 0064000d cfg1 00200040 MBUS port 26 cfg0 20000209 cfg1 1388157c MBUS port 37 cfg0 01000009 cfg1 00400080 MBUS port 38 cfg0 00640209 cfg1 00200040 MBUS port 39 cfg0 20000209 cfg1 1388157c MBUS port 40 cfg0 00640209 cfg1 00200040 2048 MiB Trying to boot from MMC1 NOTICE: BL31: v2.2(debug):v2.2-dirty NOTICE: BL31: Built : 16:17:47, Jan 10 2020 NOTICE: BL31: Detected Allwinner H6 SoC (1728) NOTICE: BL31: Found U-Boot DTB at 0xc07ad38, model: Eachlink H6 Mini INFO: ARM GICv2 driver initialized NOTICE: PMIC: Probing AXP805 <=== this is actually not doing anything in this version INFO: BL31: Platform setup done INFO: BL31: Initializing runtime services INFO: BL31: cortex_a53: CPU workaround for 855873 was applied INFO: BL31: Preparing for EL3 exit to normal world INFO: Entry point address = 0x4a000000 INFO: SPSR = 0x3c9 U-Boot 2020.01-rc3-00007-g8d8ee47e03-dirty (Jan 10 2020 - 16:34:14 +0100) Allwinner Technology CPU: Allwinner H6 (SUN50I) Model: Eachlink H6 Mini DRAM: 2 GiB MMC: mmc@4020000: 0, mmc@4022000: 1 Loading Environment from FAT... Unable to use mmc 1:1... In: serial@5000000 Out: serial@5000000 Err: serial@5000000 Net: No ethernet found. Hit any key to stop autoboot: 0 switch to partitions #0, OK mmc0 is current device Scanning mmc 0:1... Found /extlinux/extlinux.conf Retrieving file: /extlinux/extlinux.conf 1103 bytes read in 5 ms (214.8 KiB/s) 1: Armbian Retrieving file: /initrd.img-5.4.7-stb-ah6+ 11947280 bytes read in 1201 ms (9.5 MiB/s) Retrieving file: /Image-5.4.7-stb-ah6+ 21233672 bytes read in 2131 ms (9.5 MiB/s) append: root=LABEL=ROOTFS rootflags=data=writeback rw console=ttyS0,115200 console=tty0 no_console_suspend cons0 Retrieving file: /dtb-5.4.7-stb-ah6+/sun50i-h6-eachlink-h6mini.dtb 21044 bytes read in 7 ms (2.9 MiB/s) ## Flattened Device Tree blob at 4fa00000 Booting using the fdt blob at 0x4fa00000 Loading Ramdisk to 4949b000, end 49fffd10 ... OK Loading Device Tree to 0000000049492000, end 000000004949a233 ... OK Starting kernel ... Booting Linux on physical CPU 0x0000000000 [0x410fd034] Linux version 5.4.7-stb-ah6+ (root@eachlink) (gcc version 7.4.0 (Ubuntu/Linaro 7.4.0-1ubuntu1~18.04.1)) #1 SMP 0 Machine model: Eachlink H6 Mini ...  

a little update regarding my rk3318 box - at my last try it was hanging at u-boot -> see:
i finally got it booting by dd'ing the u-boot from this image: xenial-minimal-rock64-0.5.15-136-arm64.img.xz from https://github.com/ayufan-rock64/linux-build/releases/tag/0.5.15 to my sd-card and afterwards dd'ing the trust.img from:
on top of it ... not really straight forward, but works at least
 
best wishes - hexdump

hi @balbes150, i can confirm that your last image (1228) boots and runs well on a qplus h6 box ...
 
best wishes and get well into the new year - hexdump

@Redferne - try to add mem=2048M to the kernel cmdline - the eachlink h6 mini otherwise detects the memory wrong as 3gb although only 2gb are useable - thus it will crash as soon as it tries to access the part of the memory which does not really exist
 
best wishes - hexdump

@balbes150 - one other thing: try to add "mem=2048M video=HDMI-A-1:e" to your kernel cmdline - some boxes (like my eachlink h6 mini) fail to detect the memory size properly and/or fail to enable the hdmi port otherwise ...
 
best wishes - hexdump

hi @balbes150 - here is my dtb in source and binary and to make it complete also the other patches i use against the current megous 5.4 tree (just in case there is anything interesting in them for you). to my surprise the frequency scaling hack i was using with an adjusted opp table and fake vcccpu and dcdcc does not seem to work anymore in v5.4 (it only runs at fixed 900mhz again now) - i'll have to investigate why ...
 
best wishes - hexdump
emmc-fix.patch h6-ac200-eth-and-local-changes-v5.4.patch h6-ac200-eth-new-files.patch sun50i-h6-eachlink-h6mini.dts sun50i-h6-eachlink-h6mini.dtb

hi @balbes150, i think the attached one should be it (just dd it to the sd card with bs=512 skip=1 seek=1 to save the partiton table) - please let me know in case it should not work. i'll also attach my kernel config - maybe it helps you to figure out what is the problem with your generic one.
 
good luck and best wishes - hexdump
boot-h6_tv_box-aarch64.dd config-h6

its not a trivial problem (i usually get serial consoles working easily) - i tried close to all speeds etc. ... its either that the connectors are not for the serial console (but they really look like: 4 connectors, one is mass, another seems to be vcc and the other two in the middle go directly to the soc chip), there are maybe some electrical components missing to make it work properly or u-boot output is somehow scrambled (never seen something like this so far) ... these are at least my guesses ...
 
best wishes - hexdump

@oneminute - maybe give this dtb a try for booting form sd card only:
 

@jernej - good news: this patch seems to fix the emmc problem - on the eachlink i can now read and write the emmc in a stable way and on the qplus - where it was hanging on boot with emmc enabled in dtb - it does not hang anymore on boot and i can also read and write the emmc

@kissste - there is a good chance, that this patch will fix your problem too - for me it fixed the emmc problems - see
 

@jernej - bisecting in progress - 14 rounds, so it will take a while