Miscellaneous testing

Hello, I’ve just gotten my Radxa Orion O6 in the mail a few days ago and have just finished setting it up.
There are a few things i tested and ran into issues with so i would like to ask:

  1. Is it normal for the new CIX open source kernel to perform about 2x worse than the one from the prebuilt one in debian image?
    CPU scaling seems to be completely messed up, In the debian image the CPU clocks up higher when there’s more load (As one would expect) but if you build the kernel from the CIX sources the clocks go down the more load is put on?

  2. If the board is powered via an ATX PSU the power supply does not seem to shut off when the board is off, Could this be fixed via a future UEFI update?

  3. The board refuses to reboot occasionally if you try from the OS, Perhaps this is an issue on my install?

  4. My USB devices seem to randomly act up (Disconnects at random or refuses to work until i unplug and replug) Additionally my 60% keyboard’s function buttons act as media and brightness buttons for some reason?

  5. EFI does not seem to properly work yet, At least in the linux environment side (Lack of EFI FB and EFI vars)

Update, I’ve just tried pure mainline linux (linux-next) and a few things were resolved,
Keyboard now works as intended and so does CPU scaling.

Another update:
After trying to compile EDK2 (And failing) i’ve had MarioB compile the latest beta for me (Many thanks), Once flashed and booted i noticed that EFI started working in linux, Including the RTC and other fun stuff.
Sadly i can’t create new boot entries from efibootmgr but at least i can use it.

The fan no longer spins at max speed and i’m pretty sure efifb works now? Can’t tell since my install boots too quickly.

I’ve noticed that the power button shows up now in dmesg
Pressing the power button does nothing though?
Would be nice if the fan was also exposed to linux, Perhaps someday.

Finally i would like to see some sort of progress page (Perhaps i missed it) listing what works/doesn’t in mainline, Something akin to the RK3588 mainline page

Thank you for reading :smile:

7 Likes

I would also find a progress page highly useful!
It would probably also help attract new customers.

4 Likes

Note, the fan is exposed to linux when working with the BSP kernel, I’m silencing it like this:

echo 17 > /sys/bus/platform/devices/4070000.i2c:ec@76:pwm-fan/hwmon/hwmon1/pwm1

It’s just that mainline doesn’t seem to have the CIX hwmon driver yet.

Correct, I’ve since switched to mainline and updated to the 9.0.0 firmware.
Fan is automatic now so there’s no need to manually set it’s speed.

@Mario, for anyone also trying to do this, was anything special/surprising needed to get the released components (kernel, EDK2, etc.) to build?

:100: on the need for a Wiki page listing what works (in which versions), what doesn’t, and what’s being worked on!

Since nobody seems to be interested in creating a list of what works and what doesn’t i decided to take the liberty of doing that myself.
List can be found here.

Do note that i only care about mainline.

If anyone wants to expand the list please feel free to do so, Hope that helps!

9 Likes

How would like contributions to the list? Pull requests?

The front USB connections work on mainline, but only with firmware 9.0.0. With firmware 0.2.2 the ACPI nodes for the relevant USB ports show up as “status 0” and are not probed by the kernel.

Yeah pull requests are welcome, My case has USB 3.0 ports so i can’t really test that, Would have assumed it worked but rathered not put “works” unless i actually tested it.

@System64, that is a really nice table, thanks! I have one small suggestion - you should add entries for the video encoder and decoder as well. Another idea is to track support for the second USB-C port as video output (to compliment the DP, eDP, and HDMI entries).

1 Like

@volyrique Done.
If you have any other suggestions feel free to open a merge request.

2 Likes

After using my orion for a bit i feel like i should report some of the things i’ve noticed,

  1. Hard powering off the machine (Holding the power button) does not fully bring the board down, Some parts remain active such as the GPU, CPU fan or sometimes USB.
  2. Ethernet seems to occasionally crash with the following:
[  184.737236] r8169 0000:31:00.0 eth0: NETDEV WATCHDOG: CPU: 1: transmit queue 0 timed out 5173 ms
[  184.737304] Internal error: synchronous external abort: 0000000096000010 [#1]  SMP
[  184.737322] Modules linked in: cdns3_pci_wrap amdgpu i2c_algo_bit drm_ttm_helper ttm drm_panel_backlight_quirks drm_exec drm_suballoc_helper cec amdxcp drm_buddy gpu_sched drm_display_helper drm_client_lib drm_kms_helper
[  184.737423] CPU: 1 UID: 0 PID: 1559 Comm: kworker/1:2 Tainted: G S                  6.15.0-rc3-baka+ #14 PREEMPT 
[  184.737436] Tainted: [S]=CPU_OUT_OF_SPEC
[  184.737442] Hardware name: Radxa Computer (Shenzhen) Co., Ltd. Radxa Orion O6/Radxa Orion O6, BIOS 0.3.0-1 2025-04-28T03:35:34+00:00
[  184.737453] Workqueue: events rtl_task
[  184.737477] pstate: 23400009 (nzCv daif +PAN -UAO +TCO +DIT -SSBS BTYPE=--)
[  184.737488] pc : readl+0x0/0x18
[  184.737501] lr : rtl_task+0x40/0x138
[  184.737510] sp : ffff8000889d3d60
[  184.737516] x29: ffff8000889d3d60 x28: ffff000080010205 x27: 0000000000000000
[  184.737535] x26: ffff000125bb2ac0 x25: 61c8864680b583eb x24: ffff800081167000
[  184.737554] x23: ffff000120ee1dc0 x22: ffff000081332468 x21: ffff0087b7f40c40
[  184.737573] x20: ffff000081330868 x19: ffff000081332468 x18: 0000000000000000
[  184.737591] x17: 0000000000000000 x16: 0000000000000000 x15: 0000000000000000
[  184.737609] x14: 0000000000000000 x13: 0000000000000000 x12: 0000000000000000
[  184.737628] x11: 0000000000000000 x10: 6326bb1d31e0fc6d x9 : 0000000000000000
[  184.737647] x8 : 0000000000000000 x7 : ffff000120ee1e00 x6 : fefefefefefefeff
[  184.737665] x5 : 000073746e657665 x4 : ffff000080fb4780 x3 : ffff0087b7f40c40
[  184.737683] x2 : ffffffffffffe400 x1 : 0000000000000002 x0 : ffff8000829c0040
[  184.737704] Call trace:
[  184.737712]  readl+0x0/0x18 (P)
[  184.737726]  process_scheduled_works+0x18c/0x238
[  184.737742]  worker_thread+0x140/0x1c0
[  184.737752]  kthread+0x12c/0x19c
[  184.737764]  ret_from_fork+0x10/0x20
[  184.737783] Code: d50332bf 5280e001 72a80001 17fffff5 (b9400000) 
[  184.737795] ---[ end trace 0000000000000000 ]---
[  189.921231] r8169 0000:31:00.0 eth0: NETDEV WATCHDOG: CPU: 1: transmit queue 0 timed out 10357 ms
[  195.041216] r8169 0000:31:00.0 eth0: NETDEV WATCHDOG: CPU: 1: transmit queue 0 timed out 15477 ms
  1. Not really sure if this is an issue on my board? Or an issue with my USB Hub but USB occasionally starts to glitch out, Either disconnecting my keyboard, Mouse, Or other things, I did not experience this with any other machine using my USB Hub. (It’s a powered USB hub)
  2. UEFI boot time is very very long, It takes quite a bit for UEFI to start loading i’m hoping something can be done about that. (Not talking about UEFI scanning for stuff, You can disable scanning for network devices from settings, I’m referring to the time it takes from powering on the board until something shows on screen)
  3. The CPU sometimes seems to crash, Likely due to not enough voltage for my board? (Hence why the stock frequency is 2.5 with an overclock to 2.6) though this is very very rare but does happen sometimes.
  4. The power supply remains active despite the machine being off, I’m really hoping this can be fixed with a software update as i would rather not hear my fans (even though quiet) run all the time, I would also prefer to not wear out the bearings for no reason.

I’ve personally switched to the orion a month ago, It will be my daily driver for at least 3-5 years,
I hope to provide extended use feedback here and actually see some of the bugs get fixed since this is a really great board.

1 Like

Have you by any chance tested the latest 0.3.0 firmware? I am curios to see if the psu is still active.

Hello,
I appreciate the testing because I also plan to switch either to the Orion O6 or to a Ampere system.
When the schematics were first released I read them and found out that the PSU will never turn off.
Because the ATX header just shorts PS_ON with GND to always supply 12V.
So it isn’t a real ATX power input just a 12V input with a different plug that happens to short out two pins.

1 Like

Yup i am on the 0.3.0 firmware
PSU remains active even post shutdown, PCIE devices seem to shut off as far as i’m aware though?
I think, Fans on my rx580 stop spinning whenever i shut my board off though i manually power off the PSU just in case.

That is a huge shame… I should consider getting a pico PSU or a silent one instead then…

You don’t even need a pico PSU because you can feed it 12V over the ATX header without any smarts or the lower voltage rails

1 Like

Humm, Got any suggestions for a PSU then?
Ideally with something that can power a D-GPU

You need a high powered 12V PSU (a GPU also only takes 12V) and the cheapest kind of 12V PSU I can think of is a ATX PSU.
There are some PSUs out there that turn off their fan if it isn’t needed so you could use on of those.
Or alternatively there are also 12V PSUs on amazon that output a lot of power without a fan, the only downside is that you have to wire them up yourself

1 Like