MattWestB

A terrible suspicion came to me right as I was writing the previous post, so I did a quick cat /proc/cpuinfo... and it has an H3 CPU.

Everything that was written on the box and on the eBay auction was completely off.
 
It looks like this one: 
 
 
Apologies for wasting your time, I'll go and try out another image for the correct CPU!

@donluca

yes this board is NOW supported and you will have great satisfaction in using it for your purposes.

And yes ethernet will work out of the box ( and if I am not wrong, wifi too but about this let's wait @jock suggestion )

Description
This commit forces cgroup v1 usage as a workaround for docker / runc BPF issues until a proper fix is submitted
Context
Docker 20.10+ cannot run on our Asus Tinkerboard, it throws BPF related errors (Like opencontainers/runc#2959 , especially this comment). I tried a lot of armbian flavors (focal, jammy, bullseye + current, edge and legacy kernels), and even building an image with CONFIG_BPF_SYSCALL=y as suggested by this comment but it didn't fix the issue. The only thing that fixed docker was to add extraargs=systemd.unified_cgroup_hierarchy=0 to /boot/armbianEnv.txt.
Until someone else does a proper fix to make cgroup v2 working, I suggest downgrading to cgroup v1 with this extraarg.
How Has This Been Tested?
[x] Tested on Asus Tinkerboard with latest 5.15 stable kernel from apt repository. [ ] :x: NOT tested with edge and legacy kernels (Sorry I had to quickly put the machine into production and didn't have time to test other kernels) Checklist:
[x] My code follows the style guidelines of this project [x] I have performed a self-review of my own code [x] I have commented my code, particularly in hard-to-understand areas [ ] I have made corresponding changes to the documentation => Does it need a documentation change ? [x] My changes generate no new warnings [x] Any dependent changes have been merged and published in downstream modules View the full article

@jock I think i figured what the problem was .....  I have a S905L not a RK3228A processor 😂, good lord! I'm feel so stupid that i didn't open it from the beginning! 🤣 
 

please indicate box type and board version, if posible attach top and bottom images of the board.

thank you very much both @Seth @MattWestB

The right and left is VCC and GND but i dont know witch is witch.
The 2 middle is RX and TX als not know the order of them.
Start using one known GND form some shield or the power plug - to the USB-TTL GND.
Then testing  the RX from your USB-TTL cable and testing connecting is to one of the middle pins and rebooting the board. If getting some charters in the terminal (speed is 150000 baud for boot loader) its the right pin and if not testing the other middle pin.
If getting the board TX -> USB-TTL RX working is the other middle the board RX <- USB-TTL TX.
Then testing is the right or the left is GND (if both VCC and GND is on the pins you can using one multi meter for getting 3.3 Volt plus is the black GND).
 
Then having GND, RX and TX all is connected OK and you can using it as one local terminal but need changing the comport speed for Linux kernel (see first post / unbricking) and you can see and changing all in normal Linux way..

@jock
the hdmi does not work on both rounded and squared one with 5.19 kernel. date markings are exactly the same "2020/06/29" as well as board version rk3318_v1.4 on both boards. the board version was under the heatsink, i just used my phone's flashlight to see and read the board version. also out of curiousity i opened up my dead tylex v88 mini board and the external crystal is 26MHz and uses a 3R3 inductor, seems the SPxxxx wifi chips all need external crystal and inductors.
 
@MattWestB i do not know if i2c is used for communication with hdmi but both boards do behave the same, the both have hdmi output on 5.15 and both have no output on 5.19, the components around the hdmi looks exactly the same but i will take better pictures tonight when i get back home from work.
 

hello again. @jocki'll be taking pics of both boards tonight and extracting the dtb's. some vendors do give clues as to what wifi chip is placed on the board. my squared h96 max v1.4 board works fine on 2.4GHz and 5GHz and works on default distro of the 5.15 kernel build including hdmi, hdmi does not work on the 5.19 kernel though. my rounded one which i will disassemble seems to have the same board but has no 8 segment lcd. it works almost the same as the squared one with the exception of the nvram that needs replacing.
edit:
i seem to have identical boards for the rounded one and the square one, both are rk3318 v1.4 dated 2020/06/29.
differences are square and rounded respectively.
wifi chip: fx8934, sp2734c
crystal beside wifi chip: none, 37.4MHz
inductor beside wifi chip: none, 4R7
fd6551 and 8 segment: yes, no
 
square pics:https://photos.app.goo.gl/U5LEJrav5FZMh6LEA
 
rounded pics:
https://photos.app.goo.gl/dUBqpmVfWpshq4PV9
 
i think @fabiobassa can chime in on the hardware part because it seems we can put in an fd6551 and an 8 segment on the rounded h96 max box easily and make openvfd work on it too. i also think we can transplant a similar but better wifi chip on it if ever we have those on hand.

I have managed to install Armbian desktop on a RK3328 TV box folowing the instructions on the 1st post.
During the process, I appeared to have managed to 'brick' it as it would not boot/respond in any way.  I couldn't find any activity on a possible serial port.
I read some post on forcing 'maskrom mode' so decided there were 2 pins on the bottom side under the 'FORESEE' chip at bottom right of the board, so shorted them with fine pointed tweezers and powered on, removed the tweezers and got it to read the SD card again and succesfully ran Multitool again (with a correct image!).
Some pics of my TV box, a Droibox, Nano, REVO, on the PCB is marked T9-RK3328-8X4-20180718-v1.5
 
Hope it's of use to someone.
 
One question, is possible to get it to run at (upto) full speed of 1.5GHz as it is throttled to 1GHz at the moment.
 

 

 
Pins to short are at bottom left

 
 
Clock pins are below chip at bottom right...

 
 
Short these two pins and power on, then remove short and wait a little (keep fingers crossed)...

 
 

@mkultra yes, you can now unhold the packages

yup, i thought mine was an isolated incident but now i saw 2 more cases. anyway, mine is ok now since i installed the weekly build from github 22.08 v5.15 kernel cli. i set cpu speed to 1.3GHz and ram to 660MHz in rk3318-config, so far it has been stable for days, i have the same box case as @williamfj board version 1.4.

Announce:
 
Hello, I want to announce that Community Supported Configuration (CSC) board images are now built again by Armbian servers on a weekly basis!
This means that you can now download images for CSC boards (including rk3318-box) browsing from https://github.com/armbian/community
 
Images are built from trunk, GPG-signed and SHA-sum is provided.
 
I also removed the manual instructions for upgrades: Armbian 22.08 release is imminent and from that time on it will be sufficient to use apt to get kernel upgrades too! Thanks for your patience!
 
Feel free to donate if you find this useful and wish to offer support to the Armbian developers and maintainers.
 
Enjoy!

Announce:
 
Hello, I want to announce that Community Supported Configuration (CSC) board images are now built again by Armbian servers on a weekly basis!
This means that you can now download images for CSC boards (including rk322x-box) browsing from https://github.com/armbian/community
 
Images are built from trunk, GPG-signed and SHA-sum is provided.
 
Feel free to donate if you find this useful and wish to offer support to the Armbian developers and maintainers.
 
Enjoy!

Today I assembled a fully working image of OpenWRT 22.03.0  for this rk322x tvbox.

It's a patchwork/mix from multitool/debian kernel 5.16.11 and the userland from the latest OpenWRT I could find for raspberry PI 2/B (same cpu arch):

OpenWrt 22.03.0-rc6 r19590-042d558536 / LuCI openwrt-22.03 branch git-22.213.35949-d09fbe0

I've tested wireguard/vxlan/wifi, all working nice (wireless AP too, much to my surprise!).

You can turn this (rather old for 2022) TVBOX into an extremely capable router with a nice www interface (too bad for the 100Mbit LAN port...). I guess you can add some more ethernet interfaces using USB ports, but do not expect too much on the performance side.

The best way to add ethernet ports for these devices is a managed switch, so you can use VLANs (€20 for a cheap 5 ports on Amazon).

Speed of wireguard VPN is quite nice (around 70Mbps on iperf3).

If anyone is interested I can post the link for the image (flashable via multitool).

It was - as usual -  painful, but  a nice challenge.

edit:
I had to do this because there are no OpenWRT releases for rockchip/32bit armv7: arm64 builds for Rockchip 3318/3328 are officially available.

 

Thanks for the explanation. I finally unbricked it. What I figured out is that I should put a sd-card. I put it into the board, shortened contacts which MattWestB helped me to find and it booted into multitool where I choose to burn another image and now it boots into it from emmc! Thank you guys!

@DmitryS The MCC is U2300 and on the underside of the PCB you have 2 test points TP2300 and TP2301 (GND).
You can trying shutting them with one high ohm resistor for being safe (or only some metal if not playing safe and can damage the board) and putting in one USB cable from your PC in the OTG port and see if the SOC is booting in masc rom more.
 
Some of the hardware specialist here is knowing more but im 95% sure its the masc rom TPs on your board but no one can grantee it before have testing it and its all on your own risk !!
 
Its also looks you is having one unpopulated pin row for console that can being good if need doing more debug (after getting your board booting OK).

Thanks for your suggestion.

Much to my amazement network boot works pretty well: I managed to boot Alpine Linux via a modified init script that gets its rootfs from an AoE target.

I did all the debugging via tcpdump fron another system, avoiding using a serial console, and zapped the mmbclk2 partition via the usual fdisk utility.

U-boot properly started making a BOOTP/PXE/DHCP broadcast request.

I post here my findings: first the dnsmasq.conf file (tftp+dhcp all in one) - Mac address redacted, insert the one from the box you want to boot

--- dnsmasq.conf

interface=br0
bogus-priv
filterwin2k
no-resolv
localise-queries
no-negcache
no-hosts
dhcp-authoritative
listen-address=192.168.111.128
dhcp-option=3,192.168.111.1
dhcp-option=6,192.168.111.1
server=192.168.111.1
dhcp-range=interface=br0,192.168.111.35,192.168.111.40,infinite
enable-tftp
tftp-root=/var/tftp
dhcp-host=ee:58:f3:XX:XX:XX,set:d01,192.168.111.35,cazzillo,infinite
cname=set:d01,cazzillo
--- end of dnsmasq.conf
 
This assumes dns+router ad 192.168.111.1, and "cuts off" 5 IPs from 192.168.111.35 to 192.168.111.40.

U-boot requests the following file:
 
/var/tftp/pxelinux.cfg/01-ee-58-f3-XX-XX-XX

As its PXELINUX config, i.e.:

  FDT /rk322x-box.dtb
  LINUX /kernel.img
  INITRD /initramfs-rk322x
  APPEND modloop=/modloop-rk322x modules=loop,squashfs,sd-mod,usb-storage,aoe ip=dhcp
 
In the tftp root (/var/tftp) I've put the kernel, (alpine, modded by me) initrd, .dtb file, edited  "/var/tftp/pxelinux.cfg/01-ee-58-f3-XX-XX-XX" and it booted!

Pretty impressive.
 
I've not dwelled into armbian because I am not into systemd, which is the "init" of Debian derivatives.
I won't touch it with a 10-foot pole, but shouldn't be too hard to make it boot.
 

 

Great thanks @jock and  CO !!
| _ \| |/ /___ /___ // |( _ ) | __ ) _____ __ | |_) | ' / |_ \ |_ \| |/ _ \ | _ \ / _ \ \/ / | _ <| . \ ___) |__) | | (_) | | |_) | (_) > < |_| \_\_|\_\____/____/|_|\___/ |____/ \___/_/\_\ Welcome to Armbian 22.08.0-trunk Jammy with Linux 5.18.6-rockchip64 HA is up and ruing with ZHA and one Open Thread border router and looks working OK 😘
(upgraded production Ubuntu 22.04 with new kernel on eMMC without testing on SD-Card)

Dear friends, lately many many questions on this 3ad are about docker and it's problems.

I really would suggest to open a different discussion and leave here only kernel and hardware relate problems, because if we continue here with user space implementantions the 3ad becames abnormous and difficult to read.

Maybe @MattWestB and @primoitt could start soewhere else an apposit 3ad and guide other friends into solutions.

Lately if solution needs an intervent on kernel all of you will be welcome to ask here

Thanks a lot !!! 🤗

 

Hey @Bandar it was out the box installation.. it's installed from docker respository like this:

 
curl -fsSL https://download.docker.com/linux/debian/gpg | sudo apt-key add - add-apt-repository "deb [arch=armhf] https://download.docker.com/linux/debian $(lsb_release -cs) stable" apt-get install docker-ce systemctl start docker systemctl status docker curl -L https://github.com/docker/compose/releases/download/v2.3.0/docker-compose-linux-armv7 -o /usr/local/bin/docker-compose docker --version Docker version 20.10.16, build aa7e414 docker-compose --version Docker Compose version v2.3.0  

Thank you very much all.
I could install armbian; the problem was related to the HDMI issue that doesn't work on my device, and the led keep red.
The interesting part is that I bought the device as a 4gb ram (and android shows 4gb) while it only has 1.5GB when I checked by armbian!
I also took another image with a higher quality might be useful:

It is almost impossible to remove the heatsink, but as I checked carefully I could see the bga pins only.
Thanks. You solved my problem. I checked all the SD-card/USBs and Ethernet pins for the UART part, but I couldn't get any result.

@Revsearcher Also look under the green quality test sticker on the under side if some pads is being hidden there.
You have 2 pads on the upper side hear the unpopulated ram / MMC pads that very likely is forcing mask rom mode but cant being 100% sure if not tested with high impedance  resistor shutting them or oscilloscope. 
 
I think its one RL3318 box that is super stripped like my and many other RK3228A boxes and the HDMI is not working with armbian then have doing no standard layout of the PCB and need patching the RK chip for getting it working that we dont knowing for the moment. If its so you shall getting the Ethernet requesting one IP-address if waiting some minute with multi too and can connecting with SSH to it but then is normally the LED looking OK.
Some PB is using the SD-card pins for UART if no SD-card is inserted but i think its not supported for all PCBs (its using 2 pin serial = UART if not card is in and 4 pin serial then card is inserted and loading the MMC module for the SD-card).

In my experience, no I2C bus is freely routed, such as soldering point. However, if your box has an LCD display, there is a workaround. In 90% of such boxes, the display drive IC is I2C bus, so you can connect it to its corresponding pins (SCK, SDA) with another device.

Until are kernel related problems Is ok discuss here.
Or you can discuss elsewhere, test and then make requests here about missing features in kernel