Devterm a06 and kernel 6.x no video

[Greg_E]

I have a Dev term a06 and have been having problems with the newer releases based on kernel 6.x. After booting one of these, the display does not have backlight or any text/image displayed.  On the last 5.x it boots and works fine. I have not yet tried the HDMI output on the 6.x versions, need to get an adapter, but I'm guessing that the built in display is not discovered. The Dev term does not normally have a serial port, I'd need to solder up an adapter, or I would have looked at it to see if anything was happening.

 

Any suggestions for a very basic Linux user to try? Especially something I can maybe edit after flashing (before first boot). I should be able to get a Linux image up on WSL if needed to read/write the card. I had Kali on WSL, but it broke during an update, but shouldn't be too hard to stand up a new VM.

Edited July 12, 2024 by Greg_E
autocorrect is stupid

[Greg_E]

Just realized that autocorrect got me in the title, grrrr!

 

Devterm 

[Igor]

19 hours ago, Greg_E said:

Any suggestions for a very basic Linux user to try?

If non basic Linux user which you are addressing here would know what needs to be fixed, fix is usually sent to the build system and we would all have an image that boots fine. Until then, you can only join in exploring or use some older image that works.

 

AFAIK currently nobody knows what is the problem or problem is being discussed, solution being tested. If you have a custom device and keep on with recent kernel, this is most common thing that happens. Things starts to fall apart and until time, will and interest comes together, its usually fixed. And sometimes, if there are not enough people, which eventually happens, support stops / functioning on upgrades becomes random.

[Greg_E]

Well, I would suggest no longer building any OS for this device, there doesn't seem to be enough interest to run this OS, so might as well not spend the time on building updates. I think I'm just going to ditch mine since the hardware designers have very little desire to see it stay operational.

 

I tried with an hdmi, and still nothing. The system also doesn't try to expand the filesystem to fill the free space. And since the hardware designers didn't bother to include a serial port, can't see what is happening. I'm guessing something with GRUB, but that's a guess at this point. It is definitely something different between kernel v5 and v6, every one of the v5 flavors boot right up as expected.

[Igor]

3 hours ago, Greg_E said:

Well, I would suggest no longer building any OS for this device, there doesn't seem to be enough interest to run this OS, so might as well not spend the time on building updates.

Generating images, once device is in the system, does not burn our expensive time. It costs a lot less then answering this support question. Updates are generated for all devices (of one family) at once. Adding and removing is expensive and perhaps someone fixed this problem? Then we need to put it back. Those are already maintaining activities, which we don't have resources for ... keeping devices in auto-build costs us close to nothing (until compilation succeeded) and this is the same way all others distros do, while they are marking those random auto-build as "supported by Linux X" ... We at least tell you "we don't know if it works", as checking is expensive. Or even impossible as nobody from the team has this device. It was added by someone like you, user, that wanted to keep this device in the Armbian system, which system provides a lot of common fixes and generally helps maintaining those devices. 

 

Our interest is that Armbian runs well on as many devices as possible, but the costs of that is super extreme - as you know, we can't sell you our work, you don't want to pay our bills .... Once we bring support on some device, we have copycats that does absolutely nothing but also "supports" this device. It is really difficult as costs dealing with those devices is close to impossible to cover. If you want that device function perfectly and when there is close to nobody helping, one needs to spent serious time / money to support HW dealers business, copycats and you. Its not sustainable. The same applies to other devices that were thrown to the market ... We are maintaining this system, some devices, while the rest are on you - community.

Providing images, keeping this device in the system that is getting common updates all the time, is already great added value. Bugs are shared among similar devices, which means when this bug will be fixed for Pinebook PRO, this device will also have working images ...

 

3 hours ago, Greg_E said:

hardware designers have very little desire

 

Its pure economy. They can't afford that. Device is here - and if you want to keep it operational, its your problem. Well, our common problem. We add our share and we can't add more. We don't have this device, we don't have anyone maintaining it. Not anymore. Now, a lot of those devices will still work even nobody maintain them.

 

3 hours ago, Greg_E said:

The system also doesn't try to expand the filesystem to fill the free space.

 

This indicates it stops before loading user space, probably kernel is crashing.

 

3 hours ago, Greg_E said:

And since the hardware designers didn't bother to include a serial port

Sometimes you need to solder those pins. HW designers aim is to make a device that they can sell well ... Does this device matches that? Yes, its a nice toy.

 

3 hours ago, Greg_E said:

I'm guessing something with GRUB

There is no GRUB on those devices, but yes, problems with boot loader are possible.

 

3 hours ago, Greg_E said:

It is definitely something different between kernel v5 and v6

Every major kernel bump renders some devices into not usable state. That is why we have so much work that nobody notice and very little help. If we could allocate needed resources, all of those "Community (not) supported" will be working. This is certainly not in HW designers interests (as they want to sell you new device), while end users can't understand that buying hardware is just a fists step and means nothing. You need to add a lot more to have this device functional. This is not RPi, who has masses of people that will maintain it for free.

[seanieb]

Hello, I have the same issue. UART output: 

Trying a 2023 build to see if it starts, but looking online it looks like this is caused by a kernel build being incomplete and when it's wrong the device specific kernel files are smaller than usual. I noticed the 2023 image is about double the size. Same error on that, though

 

channel 0
CS = 0
MR0=0x98
MR4=0x3
MR5=0xFF
MR8=0x8
MR12=0x72
MR14=0x72
MR18=0x0
MR19=0xS = 1
MR0=0x18
MR4=0x3
MR5=0xFF
MR8=0x8
MR12=0x72
MR14=0x72
MR18=0x0
MR19=0x0
R0=0x988
MR4=0x1
MR5=0xFF
MR8=nel 1 training pass!
change freng done
channel 0, cs 1, advanced training done
ng done
ng done
change freq to 856MHz 1ng done
ng done
ng done
ng done
ddr_set_rate to 416MHZ,Boot1 Release Time: May 29 2020 17:36:36, version: 1.26
CPUId =mmc: ERROR: SDHCI ERR:cmd:0x102,stat:0x18000
mmc: ERROR: Card dOR: SDHCI ERR:cmd:0x102,stat:0x1mmc: ERROR: SDHCI ERR:cmd:0x102,stat:0x18000
mmc: ERROR: Card did not respond to voltage select!
SdmmcInit=2 1
mmc0:cmd5,20
SdmmcInit=0 0
BootCapSize=0
Us2000 , 0
StorageInit ok = 58036adLba = 2000
8fa0
NOTICE:  BL31: v1.3(release):845ee93
NOTICE:  BL31: Built : 15:INFO:    plat_rockchip_pmu_init(1196): pd status 3e
INFO:    BLded by BL2 boot loader, Booting device without OPTEE initializatError initializing runtime service opteed_fast
INFO:    BL31: Pit to normal world
INFO:    Ent

U-Boot 2022.04-armbian-2022.04-Se4b6-Pbcc6-Hc48d-V50f4-Bb703Model: Clockworkpi A06
DRAM:  3.9 GiB
PMIC:  RK808  
Core:  215 devices, 18 uclasses, devicetree: separate
MMC:   mmc@fe320000: 1, mmc@fe330000: 0
Loading Environment from MMC..USB EHCI 1.00
Bus usb@fe3c0000: USB EHCI 1.00
scanning bus usb@3 USB Device(s) found
scanning bus usb@fe3c0000 for devices... 1 USB Device(s) found
      scce
Scanning mmc 1:1...s #0, OK
Found U-Boot script /boot/boot.scr
3906 bytes read in 5 ms (762.7 K500000
Boot script loaded from 144 bytes read in 4 ms (35.2 KiB/s)
16738941 bytes read in 711 ms (22.5 MiB/s)
38343168 bytes read in 1625 ms (22.5 MiB/s)
ip/rk3399clockw_qkpi-a06.dtb'
2825 bytes read in 12 ms (229.5 ## Loading init Ramdisk from Legacy Image at 06000000 ...
  ImOK
ERROR: Did not find a cmdline Flattened Device Tree

 

Edited 53 minutes ago by seanieb