Orange Pi One+: only CPU0 comes online with current bootloader stack

[AbdulKus]

What happened?

Orange Pi One+ boots current Armbian userspace/kernel with only CPU0 online.

The kernel detects four CPUs, but secondary CPU bring-up fails through PSCI during early boot.

Before the fix the board showed:

smp: Bringing up secondary CPUs ...
CPU1: failed to come online
CPU2: failed to come online
CPU3: failed to come online
SMP: Total of 1 processors activated.

After manually rebuilding and flashing a newer bootloader stack on real hardware:

• U-Boot: v2026.01
• TF-A / BL31: lts-v2.12.9

all four CPUs come online correctly.

After the fix the board shows:

psci: PSCIv1.1 detected in firmware.
smp: Bringing up secondary CPUs ...
SMP: Total of 4 processors activated.
CPU: All CPU(s) started at EL2

How to reproduce?

1. Flash a current Armbian community (26.2.0 Trixie 6.12.74) image for orangepioneplus
2. Boot Orange Pi One+
3. Check lscpu or journalctl -k -b

Branch

main / current

On which host OS are you observing this problem?

Orange Pi One+ target board, Armbian current runtime

Relevant log output

Before:

[    0.006142] smp: Bringing up secondary CPUs ...
[    5.092199] CPU1: failed to come online
[    5.092212] CPU1: failed in unknown state : 0x0
[   10.216455] CPU2: failed to come online
[   10.216467] CPU2: failed in unknown state : 0x0
[   15.340723] CPU3: failed to come online
[   15.340735] CPU3: failed in unknown state : 0x0
[   15.340810] SMP: Total of 1 processors activated.
[ 1733.461108] psci: failed to boot CPU1 (-22)

After pinning orangepioneplus to U-Boot v2026.01 and TF-A lts-v2.12.9:

Mar 13 18:36:12 orangepioneplus kernel: psci: PSCIv1.1 detected in firmware.
Mar 13 18:36:12 orangepioneplus kernel: smp: Bringing up secondary CPUs ...
Mar 13 18:36:12 orangepioneplus kernel: SMP: Total of 4 processors activated.
Mar 13 18:36:12 orangepioneplus kernel: CPU: All CPU(s) started at EL2

I am opening a PR with the minimal board config change that reproduces this working bootloader combination for orangepioneplus only.

[github-actions]

Jira ticket: AR-2837

[AbdulKus]

PR opened: #9522

[EvilOlaf]

Okay but why does mine work just fine without this adjustment? https://paste.armbian.com/olayawekun

[    0.005778] smp: Bringing up secondary CPUs ...
[    0.007708] Detected VIPT I-cache on CPU1
[    0.007902] CPU1: Booted secondary processor 0x0000000001 [0x410fd034]
[    0.009983] Detected VIPT I-cache on CPU2
[    0.010151] CPU2: Booted secondary processor 0x0000000002 [0x410fd034]
[    0.011737] Detected VIPT I-cache on CPU3
[    0.011883] CPU3: Booted secondary processor 0x0000000003 [0x410fd034]
[    0.011999] smp: Brought up 1 node, 4 CPUs
[    0.012090] SMP: Total of 4 processors activated.
[    0.012103] CPU: All CPU(s) started at EL2

[AbdulKus]

Okay but why does mine work just fine without this adjustment? https://paste.armbian.com/olayawekun

[    0.005778] smp: Bringing up secondary CPUs ...
[    0.007708] Detected VIPT I-cache on CPU1
[    0.007902] CPU1: Booted secondary processor 0x0000000001 [0x410fd034]
[    0.009983] Detected VIPT I-cache on CPU2
[    0.010151] CPU2: Booted secondary processor 0x0000000002 [0x410fd034]
[    0.011737] Detected VIPT I-cache on CPU3
[    0.011883] CPU3: Booted secondary processor 0x0000000003 [0x410fd034]
[    0.011999] smp: Brought up 1 node, 4 CPUs
[    0.012090] SMP: Total of 4 processors activated.
[    0.012103] CPU: All CPU(s) started at EL2

I checked it over a month ago and just now on the latest release. Every time, one core is running and the other three are offline, even visible in htop. Maybe different board/processor revisions?

You're apparently using Armbian noble. I had issue on Armbian Trixie cli.

[EvilOlaf]

It is very unlikely that the userspace matters because uboot and kernel are always the very same, independent from userspace.
I may have found something else but need to dig deeper. May take a few days.

[AbdulKus]

It is very unlikely that the userspace matters because uboot and kernel are always the very same, independent from userspace. I may have found something else but need to dig deeper. May take a few days.

If necessary, I can give SSH access to the board with a freshly installed image.