CSC Armbian for RK322x TV box boards

[ArkhanLK]

1 hour ago, jock said:

@ArkhanLK It's hard to say if it is an hardware failute or not. Surely if reinstalling the original android does not work, most probably there is an hardware fault.

Multitool also should always boot without issues, if it fails to boot with similar error messages, maybe your board is faulty.

 

Did you try the cpu-stability overlay in /boot/armbianEnv.txt?

Hey Jock! Thanks for answering so fast!

 

Here is what happened: android was installed (I do not remember if I used it, but I believe I did at least tested...), anyways, the only different thing I did was to install armbian through multitool using "restore" instead of "install", could that have been the issue?

Then, after that, the board does not boot at all. If I try to boot it pressing the reset button with the multitool sd card I get that screen. The sd card is 100% working as I have installed and backed up and restored multiple version on multiple boxes.
I have not tried anything with the rkdev tool ever, but I could try something. Regardless, I can't even revert back to android since I can't get multiool working. That's why I believe it has to be a hardware failure...

 

Also: can, somehow, an action from multitool (restore, backup, install...) break a box/emmc? It doesn't mess with clocks or voltages, does it? Like, I suppose a energy shortage midway process could be dangerous... But, assuming everything to be normal, can multitool actions themselves break/brick the box?

 

As always, I appreciate your guys work and efforts, thank you so much! 

[jock]

6 minutes ago, ArkhanLK said:

Also: can, somehow, an action from multitool (restore, backup, install...) break a box/emmc? It doesn't mess with clocks or voltages, does it? Like, I suppose a energy shortage midway process could be dangerous... But, assuming everything to be normal, can multitool actions themselves break/brick the box?

Normally no, multitool does not mess with clocks and voltages, moreover it is designed to run the board at maximum 800 MHz for extreme precaution.

Using restore instead of burn is right the same thing; restore will just pick from backups folder, burn picks from images folder.

 

You may erase eMMC completely and see if Armbian runs from sdcard: in such case maybe the eMMC is the first candidate as being faulty.

[ArkhanLK]

4 minutes ago, jock said:

Normally no, multitool does not mess with clocks and voltages, moreover it is designed to run the board at maximum 800 MHz for extreme precaution.

Using restore instead of burn is right the same thing; restore will just pick from backups folder, burn picks from images folder.

I see, thanks!

 

4 minutes ago, jock said:

You may erase eMMC completely and see if Armbian runs from sdcard: in such case maybe the eMMC is the first candidate as being faulty.

Well, yeah. But for that I'd need multitool to open, haha

Currently, I either have no boot or boot into that "error like" screen I posted the images for. It doesn't go anywhere, no terminal, no multitool, no nothing.

Just that and only that. That's why I am strongly inclined to believe something on the board (hardware) is faulty.

[jock]

2 hours ago, ArkhanLK said:

Currently, I either have no boot or boot into that "error like" screen I posted the images for. It doesn't go anywhere, no terminal, no multitool, no nothing.

Just that and only that. That's why I am strongly inclined to believe something on the board (hardware) is faulty.

Ok, i read again your last message and I see that the last thing you did was (most probably) install back Android via multitool and you get that, am I right?

Debugging via serial will tell something more, but you may also try mwronglyaskrom mode and rkdeveloptool to erase the mmc and boot from sdcard.

 

That's bad multitool is not booting with such error, but is it possible you installed back the wrong Android image (maybe coming from another box)? If so, it could be that DDR memory is wrongly initialized and makes the system unstable; even the multitool can't cope with that and for that problem you need to erase mmc.

[ArkhanLK]

9 hours ago, jock said:

Ok, i read again your last message and I see that the last thing you did was (most probably) install back Android via multitool and you get that, am I right?

Well, no, not at all. Here is all steps on what happened:

1. Bought the box with android 7.1 and it booted OK;
2. Inserted SD card with multitool and ran "restore" instead of "install" with an Armbian version that I already had "perfected" on another exactly equal box;
   • I have installed (thoruhg "install" menu option in multitool) this same version of Armbian on another number of this exact same box, multiple times, and nothing similar has ever happened! 
3. Tried to boot the box, nothing happened;
4. Tried to boot the box with the multitool SD card in and got that error screen;
5. Nothing could change the results of steps 3 and 4 so far.

Steps 3 and 4 happened on multiple attempts. I was never able to get back to android or start multitool again. The box is, therefore, bricked.

 

Could this have happened because of a faulty multitool install? Could multitool have damaged something to make it "bricked"? Or is it more likely that the box is somehow faulty?

I believe on the latter but have not tried investigating. I never used the rkdevelopt tool, but I will try to use it later!

[jock]

@ArkhanLKI guess for "Install" you intend "Burn image on flash", since there is no such install option in multitool.

Besides that, it could be that the board has a slightly different memory and it does not like the default settings.

If I'm right, the issue is the same 1T/2T Command Rate problem.

 

The latest version of the multitool is able to autodetect this setting from the existing Android, and also allows to manually change the setting from the menu.

 

The only issue is that, at the moment, you are stuck with the problem and the only things you can do are:

- clean the eMMC using maskrom mode and rkdeveloptool or

- use the serial and clean the first sectors of the eMMC using u-boot

 

[fabiobassa]

Quote

that I already had "perfected" on another exactly equal box......

and this assemption is soooo far away from the truth that you can imagine. Imagine those big factory that assemblies boxes, one factory.. many reseller !
And in a certain moment a employee says " hey boss.. we finished ddr 3 ram and we have only ddr 2 now"
And the boss answers " doesn't matter take the ddr 2 put in washing machine, erase old name and put on chip the right ddr  3 name and assembly 1000 other fuc**ng box!!  "
This taht could sems just a lie or a joke I believe happenns very most often than we can even imagine
Do you know how many box i have ABSOLUTELY  egual in name, brand shape, even names of chips inside.. and the uart log tells me the REAL SITUATION?

Plenty !!! And me and @jock bumping head in the wall because one box working and the other NO ?!!!!

So only solution is nice uart log or short the emmc/nand pins so you are sure you booting in maskrom and program a generic android even older that puts vesry piece in the right place of internal memory so you can start form scratch again
Those box are quite unbrickable, and NO multitool and other tools offered with great generosity from jock cannot " destroy" boxes
 

[gianlucaf]

On 9/18/2021 at 12:19 PM, gianlucaf said:

Ciao, sto provando ad installare Docker su un Alfawise A8, voglio provare a farci girare Home Assistant. Le varie procedure trovate qui e qui mi portano sempre ad un errore simile nel momento in cui do install.

Sapete se è possbile installarlo e come, accetto ogni tipo di suggerimento, sono abbastanza neofita di linux. Grazie

 

Hi, I'm trying to install Docker on an Alfawise A8, I want to try to run Home Assistant on it. The various procedures found here and here always lead me to a similar error when I do install.

You know if it is possible to install it and how, I accept all kinds of suggestions, I am quite new to linux. Thanks

 

root@rk322x-box:~# sudo apt-get install docker-ce docker-ce-cli containerd.io
Reading package lists... Done
Building dependency tree
Reading state information... Done
Package docker-ce is not available, but is referred to by another package.
This may mean that the package is missing, has been obsoleted, or
is only available from another source
However the following packages replace it:
  docker-ce-cli:arm64

E: Package 'docker-ce' has no installation candidate
E: Unable to locate package docker-ce-cli
E: Unable to locate package containerd.io
E: Couldn't find any package by glob 'containerd.io'
E: Couldn't find any package by regex 'containerd.io'

 

 

I find a docker repository, i downloaded the .deb files and everything worked! stupendous! it is also very fast.
I could not activate the wifi, it would be convenient for me to free up a lan cable, is it possible to do this? And how? thx

[dotbg]

Any chance to boot from USB on eMCP board?

[jock]

3 hours ago, dotbg said:

Any chance to boot from USB on eMCP board?

 

You can install Armbian on eMCP using latest multitool: it is still experimental and you have to overwrite the existing Android.

 

Once Armbian is installed, the bootloader will search for bootable devices in this order: sdcard -> USB -> eMCP

 

 

[dotbg]

It's possible to install u-boot only ( rkdeveloptool method ) and boot from External USB drive? 

[jock]

3 hours ago, dotbg said:

It's possible to install u-boot only ( rkdeveloptool method ) and boot from External USB drive? 

Indeed, you need to install the first megabyte from any armbian image onto the eMMC starting from sector 0.

That's all you need.

 

The suggested experimental way with the multitool is better because it patches the installed bootloader, otherwise your board may brick and you need to rescue using the maskrom mode, which is difficult on eMCP.

[Willy Moto]

Hi, it's me again.

 

Have installed Armbian Buster via stepnand method, and running it for 2 weeks now.

I observed something & would like to check:

 

I found that the Load Average seems to be higher than other SoCs even the box is not running much loading.

Here is the spec:

 

RK3229 :  1GB RAM + 8 GB NAND storage

OS:           Armbian Buster 21.08.02  ( via apt-get upgrade )

                 Legacy Kernel 4.4.194  (for NAND memory model)

 

Software installed & running

                 UFW  (for allow a few ports to access Pi-Hole)

                 Pi Hole  (installed via armbian-config, 3rd party softy menu)

                 DuckDNS update curl script  (crontab, only run once each hour)

 

 

===========

 

Using htop,   I found that the Load Average almost always stays at 1 1 1.  It stays like that even there is not much DNS traffic going on...

As RK3229 has 4 cores, this translates to 1/4 = 25% system load at all time.

 

However,  the CPU utilization is lower & mostly below 10%.    I don't understand this discrepancy (25% system load vs 8~10% cpu utilization).   

For comparison,  I am running Armbian on another Cubieboard 2 compatible TV box  (1 GB RAM, CPU SoC is Allwinner A20 dual core A7 @ 1 Ghz).

 

The load average in that TV box is very often 0.5 or below (remember it's only dual core system, running almost same software configuration except it's using latest mainline stable Kernel 5.10.y).

 

 

Question:

Is there something odd about the higher than usual Load Average value  ( 1 1 1 ) in RK3229 (Legacy Kernel + NAND),  even when not much task is going on?

 

 

[fabiobassa]

@Willy Moto
I have several board rk322x running pihole and it is a very interesting task to do with thsese boards. What I don't understand is :
 

Quote

   UFW  (for allow a few ports to access Pi-Hole)

What you mean ?? What is the topology of  your network? I don' undestand the necessity of opening ports on the box via ufw to access dns requestes .
 the pihole is running on the same lan subnet of your work/home and you do requests from lan by setting the dns of your internal topology to point at pihole address instead of canonical google dns or opendns etc etc.
And even if you want to do personal requestes from outside ( wan) you should open ports on the border router, not on box itself
And finally.... if you real want to open port , use iptables and iptables-persistent and not user space 3rd  programs ( ufw)
 

Quote

I am running Armbian on another Cubieboard 2 compatible TV box

This is another long story ... often done on this 3ad. Is not possible compare raspberry or amlogic or allwinner among themselves or among rockchip.  Is a bit comparing pears and appples
Is fundamental understand different behaviours due to different topology even if the substrate is always arm architecture

To obtain more complex infos, unplug lan , stop pihole and again htop to check if there is offending tasks

 

[Willy Moto]

19 hours ago, fabiobassa said:

@Willy Moto
And finally.... if you real want to open port , use iptables and iptables-persistent and not user space 3rd  programs ( ufw)
 

 

I understand, but I wanted to make a shortcut & use UFW.   Pi Hole documentation has the exact info for UFW rules.

To make it simple, just want to protect the Armbian box & not allow anyone else to interfere & hack in.

 

 

Quote

This is another long story ... often done on this 3ad. Is not possible compare raspberry or amlogic or allwinner among themselves or among rockchip.  Is a bit comparing pears and appples
Is fundamental understand different behaviours due to different topology even if the substrate is always arm architecture

To obtain more complex infos, unplug lan , stop pihole and again htop to check if there is offending tasks

 

 

I understand.   I am currently playing with another rockchip TV box lately (RK3328), but I thought not to bring this in here so as not to mix things up.

I will observe a little longer but I am thinking if the older Kernel or NAND device driver could be the culprit?  Because all other Armbian boxes I've worked on were all on latest Kernel 5.x, on SD-card / EMMC.   They didn't have this high Load Average problem (ie. the discrepancy between Load Average vs CPU utilization ).

 

I would try to gather more information and post back.    Maybe I should run  Mainline Kernel build (5.10.y Kernel on a USB drive) to compare.

[jock]

@Willy Moto Yes, your observation is right: load average is broken on legacy kernel. For some reason, it idles with a value of 4.00, where instead mainline kernel correctly idles with a value of 0.00

[gianlucaf]

sorry, how can i activate the wifi module?

[Luis]

Disculpen pero en la última actualización Buster Xfce  4.4 no me aparecen USB que pongo probé con otros USB pero nada no aparecen algo estoy haciendo mal ?

[tommy]

Thank you for make a great project for this RK322x board.

 

I try to build a retro gaming box but I stuck by a obstracle. I successfully compile retroarch (retro gaming frontend) but the performance is very good many gaming platforms but not enough for some , snes, PS one, psp because video fps is below 60fps in 480p resolution.

 

Libreelec has a good performance but it lack control system.

 

Lakka(Libreelec modify for retro gaming) maybe good enough but I could compile or find a working binary image.

 

Android has a good performance too but bad interface.

 

Could someone guide me to make this retro gaming box successful?

[dotbg]

Is there any limits to the maximum usable size of RAM?

[fabiobassa]

@gianlucaf

Quote

sorry, how can i activate the wifi module?

uhmmmmm     .... how can i drive my car? uhmmm what car have you ? fuel, diesel, electric , horse power ..
if you don't give more infos, actually our crystal ball is broken. Which wifi chip if ? ssv6051, spe... 8023cs ??
And last thing, if it is working under armbian-config and not in user-space programs ( docker) try to ask in docker forums ;-)

@tommy
there is a lot of interest in retro gaming and graphics use of those box but if you have read the whole 3ad (  and I'm not sure you did.. )  you should also have read that gpu hardware acceleration is in some way conflicting with x11 server and this is the reason why it runs smoothly on android or libreelec that were compiled with good performance on graphics but not so well on linux desktop. You can try to mixt up things among armbian and libreelec that uses wayland instead of xorg, but this is on yourself
And NO at the actual state of progress there is few possibilities that the things will go better on 322x

@dotbg
 

Quote

Is there any limits to the maximum usable size of RAM?

pardon I didn't understand ! What you mean ? That your box is claiming have mega mega mega tons or gigatons of ram and you see only one or two ?

 

[dotbg]

Nope! There are some kind of boxes claiming 8GB of RAM for example! 

 

[RaptorSDS]

vor 4 Stunden schrieb dotbg:

Nope! There are some kind of boxes claiming 8GB of RAM for example! 

 

more like 8 Gbit or ram =1Gbyte ram ; or they have ram and rom swap or like some box its only  fake advertise

 

please show the link to example

 

[jock]

@dotbg Some boxes with rk3566/rk3568 are sold with 8GB (gigabytes) of RAM. That chip is capable to address such amount of ram, yet you have to trust the chinese manufacturers and vendors that they really put 8Gb of RAM in the box

 

rk322x instead is not able to go above 2GB (gigabytes). Everything beyond on a rk322x is surely fake.

[fabiobassa]

Mostly just FAKE specs

[jock]

@tommy I Didn't ever try retroarch. I took a look on snes emulation on libreelec and surprisingly it was working quite well on rk322x. I tried once to compile snes9x for rk322x and also was working fairly well in X.org, but performance dropped when I tried any scaling 2X filter.

 

That said, I would suggest to you, as @fabiobassa said, to try and see if you can avoid X.org, maybe trying to compile with GBM/DRM support so you get fullscreen without passing through the X server.

Also you should tell which kernel are you using, it may be important.

[tommy]

Thanks fabiobassa and jock for your help.

 

 

 

 

11 hours ago, jock said:

@tommy I Didn't ever try retroarch. I took a look on snes emulation on libreelec and surprisingly it was working quite well on rk322x. I tried once to compile snes9x for rk322x and also was working fairly well in X.org, but performance dropped when I tried any scaling 2X filter.

 

That said, I would suggest to you, as @fabiobassa said, to try and see if you can avoid X.org, maybe trying to compile with GBM/DRM support so you get fullscreen without passing through the X server.

Also you should tell which kernel are you using, it may be important.

 

I use legacy kernel 4.4.194

~$ cat /proc/version
Linux version 4.4.194-rk322x (root@nott-xubuntu) (gcc version 8.3.0 (GNU Toolchain for the A-profile Architecture 8.3-2019.03 (arm-rel-8.36)) ) #2 SMP Sat Oct 2 12:44:11 +07 2021

 

 

I try a dirty hack by change soft link of /usr/lib/arm-linux-gnueabihf/libMali.so from X11 version "libmali-utgard-400-r7p0-r0p0-x11-gbm.so" to "libmali-utgard-400-r7p0-r0p0-gbm.so" for retroarch call library and set default login terminal by "$ sudo systemctl set-default multi-user". I set display resolution to 720x480 for reduce processing load.  The result is acceptable to snes (use snes9x2010) level.

 

I am an amature in Linux build system and C programing. I will try to learn to compile armbian avoiding X11 version but I not sure my knowledge and skill is enough. I have seen someone edit EmuElec(LibreElect with Retroarch interface) to could be run rk322x. I try to edit Emuelect source that almost done but dependencies are very fraustrated. Full Linux system like this armbian project is the best.

[fabiobassa]

@tommy
Even if the main purpose of this 3ad is " general" speaking of the process to bring pure linux on 322x boxes, without going deep in user space apps, people as you, with programming skills, are always welcome to discuss and apport ideas to the project.

I am more the " hardware man", @jock is the " code man" , but what is now clear, even to me, is that doesn't matter the mali version, the bottle neck is the necessity to copy the result of gpu elaboration into x11 windows server. This is where the whole process slows up.

This is true for videos and jock ( but other people too ) is actively searching a solution , a work around but now mostly efforts are on mainline kernel

Example given : if you do init3, unload all mali libraries, insmod all gbm libraries you will have FULL gpu power and you can run even lot of frames per seconds, but the you loose the x11 features. If you control games by game pad or keyboard this could be not such a big problem , once you stop the game you unload gbm go back on mali and startx. Android uses the gbm way so good hardware acceleration and poor desktop experience. On armbian we experiment the opposite .
But this switch  isn't convenient for normal use.
@jock could explain this 1.000.000 times better than  I did, but the core of problem is exactly what i wrote

 

[fabiobassa]

@tommy

post scriptum.. I usually dont like or play games but i have had 29 giga of arcade games  that run on amlogin 905 . it is aarch64 and our little box is armhf 32 bit but i guess games are no-arch so if you go further on the process of emulation please share your experience

 

[jock]

@tommy

I would say you to give a chance to the mainline kernel too. Lima driver has decent performance and you don't have to lose time around proprietary Mali OpenGLES-only drivers swapping links under the hood. (BTW for legacy kernel there is the media script link on first page that should install the libraries in the right places).

 

For snes9x you should not need particular load on the GPU since it has just to render the framebuffer. Lima performances should suffice.