Armbian 26.08 Release

[igorpecovnik]

Armbian 26.08 – Release Month: Start Releasing

We are now in the Armbian 26.08 release month. It is time to stop adding features, stabilize the code, build images and start testing.

Release focus

• U-Boot 2026.07
• Linux 6.18 LTS remains our primary kernel
• Focus on stabilization and fixing regressions
• Incomplete or unfinished features should be moved to 26.11
• Test released images

Start releasing

Please start generating release images today under version 26.8.1.

What is missing?

Check summary of those jobs:
https://github.com/armbian/armbian.github.io/actions/workflows/reporting-download-images.yml

Follow the release image preparation procedure.

Important

If you don't have permission to run those release workflows, ping @igorpecovnik and I'll grant it.

Generated images will appear in incoming under your GitHub username. Please test the boards you maintain and report/fix regressions.

From this point, only critical fixes and release-blocking changes should go into 26.8.1. Everything else should move to the 26.11 development cycle.

Target: Armbian 26.8.1

Promote your well-working CSC boards to Supported

If you maintain a CSC (community-supported) board and you're confident it works well on 26.08 — it boots and the hardware/features you care about are solid — you're welcome to promote it to supported:

• rename config/boards/<board>.csc → config/boards/<board>.conf, and
• set BOARD_MAINTAINER to your GitHub handle.

Release month is the right time to graduate boards that have proven themselves — but only promote what you actually stand behind and will keep an eye on.

---

Board maintainers — please test the boards you maintain and report/fix regressions for 26.8.1:

@adeepn @alnahian2011 @amazingfate @belegdol @chenxmqqqq @cjkoenig @code002-2 @crackerjacques @CT1IQI @d93921012 @deece @defcom5-rockchip @diegodsgarcia @dlitz @dust-7 @efectn @eldandaman @enromytase @EvilOlaf @farwayer @fkpwolf @frank-f @ginkage @govindsi @gpaesano @Grippy98 @HeyMeco @hqnicolas @iav @ifroncy01 @jjg @JohnTheCoolingFan @jomadeto @juliovendramini @k-popov @kasimling @lch08 @lexfrei @lukaszsobala @mingzhangqun @Mkiring @neheb @nyanmisaka @ojafuenf @okrc @Ovaday @paolosabatino @pdapandapda @pierg75

Summary by CodeRabbit

• Chores
  • Updated the application version to 26.11.0-trunk.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: 9d114e21-4c71-4188-8581-542d2603e240

📥 Commits

Reviewing files that changed from the base of the PR and between e6b5653 and 855a1e3.

📒 Files selected for processing (1)

• VERSION

---

📝 Walkthrough

Walkthrough

The project version changes from 26.08.0-trunk to 26.11.0-trunk in the VERSION file.

Changes

Version update

Layer / File(s)	Summary
Version declaration VERSION	The project version changes from 26.08.0-trunk to 26.11.0-trunk.

Estimated code review effort: 1 (Trivial) | ~2 minutes

Merge Risk: ⚪ Minimal · up to 855a1

This PR only advances the project version to the 26.11 development cycle; no actionable merge-blocking risk remains beyond normal checks and review.

Suggested reviewers: evilolaf

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly identifies the Armbian 26.08 release effort described in the objectives, although the version bump itself advances to 26.11.0-trunk.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch bump/version-26.11

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[igorpecovnik]

@pumuckly @pyavitz @rbqvq @redrathnure @retro98boy @rorystandley @rpardini @ruabmbua @saschabuehrle @sdeleeuw @Shadowrom2020 @shkolnik @ssilnicki-dev @SuperKali @TallGuy74 @tenox7 @ValdikSS @vidplace7 @yisding @zevlaoxie2010

[ValdikSS]

Just confirmed that MGLRU on sunxi is now compiled and enabled by default. Thanks!
Tested on Orange Pi Zero3.
#10133

[JohnTheCoolingFan]

1. Bigtreetech CB1 - tested trixie minimal. Boots, but no ethernet (interface shows up and a local ipv6 is assigned, but dhcp does not work). Didn't test wifi because can't install iwd via apt ethernet works, issue was a faulty cable.
2. Bigtreetech CB2 - tested resolute minimal. Boots, ethernet and wifi work.

The cb1 ethernet issues should've been resolved by #10155 ? There may have been changes since then, will need to re-test.

[deece]

1. Bigtreetech CB1 - tested trixie minimal. Boots, but no ethernet (interface shows up and a local ipv6 is assigned, but dhcp does not work). Didn't test wifi because can't install iwd via apt
2. Bigtreetech CB2 - tested resolute minimal. Boots, ethernet and wifi work.

The cb1 ethernet issues should've been resolved by #10155 ? There may have been changes since then, will need to re-test.

This sounds like the 25MHz clock is no longer getting to the PHY

[EvilOlaf]

Bigtreetech CB1 - tested trixie minimal. Boots, but no ethernet (interface shows up and a local ipv6 is assigned, but dhcp does not work). Didn't test wifi because can't install iwd via apt

Is this the case for both 6.18 and 7.1?

[lexfrei]

Tested Sovol Zero (a CB1 clone, same sun50i-h616-bigtreetech-cb1.dtsi) on 6.18.44 / 26.11.0-trunk.7. Ethernet works: 100Mb/s full duplex, autoneg on, RX alive (1497 packets, no errors), and the PHY binds as Allwinner AC300 EPHY rather than generic C22.

@JohnTheCoolingFan worth checking zcat /proc/config.gz | grep PWM_SUNXI_ENHANCE on the image you tested. The symptom you describe is what we chased in #10242: ac300 is built in, but its probe needs the pwm clock, and while the PWM driver was a module the probe hit -EPROBE_DEFER. If networkd touched the interface in that window, phy core force-bound generic C22 to an uninitialised EPHY, giving 10/Half parallel-detect and RX=0. It was a race, so a board could win or lose it depending on boot timing, which is why it looked board-specific. Fixed by building the pwm driver in, merged 21 July.

One more trap when retesting: the mainline driver derives a different MAC than the BSP one, so DHCP hands out a new address and the old IP goes quiet. That looks like a dead link if you only ping the previous address.

[JohnTheCoolingFan]

zcat /proc/config.gz | grep PWM_SUNXI: CONFIG_PWM_SUNXI_ENHANCE=y @lexfrei
also, ethtool says Link detected: yes and shows 100BaseT/Full as the highest mode. Ethernet PHY binds as Allwinner AC300 EPHY. So now the suspect is MAC address? In ip addr it shows up the same mac as in my router's dhcp list, but even after a reboot the interface only has a local ipv6 address. When I first tested I don't remember seeing the bigtreetech-cb1 device in the router dhcp list at all though.

[igorpecovnik]

Board works https://docs.armbian.com/Contribute/Datacenter/#boards but I notice that it dies after a while - led stop blinking. Sadly don't have any console attached to ATM. Will try to catch what is going on.

[JohnTheCoolingFan]

how recent is the build on that board?

[lexfrei]

@JohnTheCoolingFan ran the same kernel on a genuine CB1 (Manta M4P bench, 26.11.0-trunk.7 / 6.18.44): DHCP works, networkctl shows routable (configured), lease from the router. So the kernel and the ac300 driver look fine, and my earlier guess about the probe race doesn't apply to your case either, since your PHY already binds correctly.

That leaves the image's network config. On my bench the DHCP setup comes from /etc/netplan/10-dhcp-all-interfaces.yaml, and that file is not shipped by any package (dpkg -S finds no owner), it is written when the image is built. So it can differ between image builds while the BSP package stays identical.

Worth checking on yours: systemctl is-active NetworkManager systemd-networkd, ls /etc/netplan/, and networkctl status end0. If neither manager owns end0, the link comes up and nothing ever sends a DISCOVER, which matches what you see.

[paolosabatino]

Tested RockPi-E (rk3328) - Trixie minimal

• Basics: boot, reboot and shutdown all work fine
• eMMC: access is ok, 120MByte/s read speed, boot ok
• sdcard: access is ok, boot ok
• Gigabit Ethernet: ok, throughput above 900MBit/s
• Fast Ethernet: ok, throughput above 90MBit/s
• Wifi: rtl8822cu works fine out of the box, TX throughput 24MBit/s, RX throughput 5MBit/s (router is crappy and far)
• USB: USB3 port works fine in USB3 and USB2 mode
• Analog audio codec: works fine

Summary: ready for production ✅

[JohnTheCoolingFan]

@lexfrei tested images with both NETWORKING_STACK=systemd-networkd and NETWORKING_STACK=network-manager and both behave the same. networkctl reports the interface as "degraded (configuring)" and nmcli reports connecting (getting IP configuration). I have not tested on 7.1 yet, building and flashing images takes a while.

Edit: 7.1 is doing the same

[paolosabatino]

Tested Asus Tinkerboard-S (rk3288) - Tested Ubuntu Resolute XFCE:

• Basics: boot, reboot and shutdown all work fine
• eMMC: access is ok, ~ 140MByte/s read speed, boot ok
• sdcard: access is ok, boot ok
• Gigabit Ethernet: ok, throughput above 900MBit/s
• HDMI: works fine
• HDMI Audio: works fine
• Wifi: rtl8723bs, able to connect to wifi, but zero throughput; probably router too far away - update: move the board near the router, performance is very good: TX: ~75MBit/s, RX: ~90MBit/s
• Bluetooth: works fine, able to stream audio from paired smartphone
• USB2: all ports work fine
• Analog audio codec: works fine

Summary: ready for production ✅

edit: updated Wifi throughput results

[lexfrei]

@JohnTheCoolingFan built a clean trixie minimal from current master (26.08.0-trunk, 6.18.44) and flashed it to a genuine CB1. DHCP works: lease from the router, networkctl says routable (configured), RX counters moving. So a fresh image off master doesn't reproduce it here.

Two things that would narrow it down: ip -s link show end0 while it sits in configuring (if TX grows and RX stays 0, frames leave and nothing comes back), and the exact compile.sh arguments you used for the image.

Also worth ruling out the router side, since you mentioned the board never appeared in the DHCP list on the first test. A lease is two-way, and configuring is what you get when DISCOVER goes out unanswered.

[lexfrei]

Sovol Zero (supported, I maintain it), tested on 6.18.44:

• Basics: boot and reboot fine
• eMMC: ok, runs at 40MHz via the board overlay, boots from it
• Ethernet: ok, 100Mb/s full duplex, DHCP lease, PHY binds as Allwinner AC300 EPHY
• WiFi: onboard 8189fs is dead on this particular unit (hardware), USB MT7601U dongle works
• CAN: ok, 1Mbit, ERROR-ACTIVE
• USB: ok, all devices enumerate
• Thermals: 38-39C idle

On this board that was a kernel upgrade over an existing rootfs. A clean trixie minimal built from current master (26.08.0-trunk, 6.18.44) I flashed to a genuine CB1 instead, and it comes up fine there: ethernet, DHCP, both wifi adapters. The printer the Sovol lives in is disassembled, so the print stack is untested.

sovol-sv08 and sovol-sv08-max: no hardware here, can't test. They stay csc.

[JohnTheCoolingFan]

BTT CB1 ethernet problem was caused by a faulty cable, the image doesn't actually have issues.

[Shadowrom2020]

Tested Qidi X-6:

• Performance test passed
• Ethernet and Wifi Test passed
• eMMC works as intended
• USB works for 3D printing MCU
• Serial connections work for 3D printing MCU
• Stability test passed
  Ready for production ✔️

[paolosabatino]

Radxa RockPi-S (rk3308) - Trixie Minimal (I'm not the maintainer though, the RC image is still in the workflow, I tested using a freshly build local image from up to date master branch):

• Basics: boot, reboot and shutdown all work fine
• eMMC: access is ok, ~ 20MByte/s read speed, boot ok
• sdcard: access is ok, boot ok
• Fast Ethernet: ok, throughput above 90MBit/s
• Wifi: rtl8723ds, works. performance is TX: ~16MBit/s, RX: ~6MBit/s (router is crappy and far away)
• Bluetooth: did not test
• USB2: the port works fine
• Analog audio codec: works fine

Summary: ready for production ✅

[redrathnure]

mkspi (MKS PI and SKIPR boards)

Summary: ready for production ✅

MKS PI (Debian Trixie, current 6.18 and edge 7.1 images):

• ✅ boot from microSD/eMMC/USB
• ✅ TS35 and HDMI video
• ✅ LAN and WIFI dongle
• ✅ SPI, UART and debug ports
• ✅ USB 2 and USB 3
• ❌ console=both does not work. A both option prints boot logs only to the connected screen, debug port output is empty after U-Boot messages

SKIPR (Debian Trixie, edge 7.1 image):

• ✅ boot from eMMC
• ✅ TS35 video
• ✅ WIFI dongle
• ✅ SPI, UART and debug ports
• ✅ CAN bus network.
• ✅ USB 2 and USB 3, including 5Gb/s with external USB storage. Issues with UAS, need to disable via usbstoragequirks. Most likely it is NOT Armbian related issue.
• ❌ console=both does not work. A both option prints boot logs only to the connected screen, debug port output is empty after U-Boot messages

Side notes:

• U-boot 2026.x gives less errors and warnings
• Trixie with kernel 7.1 subjectively faster:
  1. lower CPU temp
  2. Better performance in heavy CPU scenarios (e.g. video encoding)
  3. better performance in heavy IO (USB 3 + network) scenarios

[defcom5-rockchip]

If you want,
I could post an OrangePi-5B report in paolosabatino's format (boot/eMMC/SD/ethernet/wifi/BT/USB/audio ✅) to help the 26.08 release — edge (mainline ~7.x) image and a vendor (Rockchip BSP 6.1) image.?. I have a 2.4ghz wifi for testing against bluetooth audio through the WIIM. Not an absolute bluetooth sensor test but is close. I could get it over the weekend.

[paolosabatino]

Tested OrangePi 4 LTS (rk3399) - Tested Ubuntu Resolute XFCE:

• Basics: boot, reboot and shutdown all work fine
• eMMC: access is ok, ~330MByte/s read speed, boot ok
• sdcard: access is ok, boot ok
• Gigabit Ethernet: ok, throughput > 900MBit/s
• HDMI: works fine
• HDMI Audio: works fine
• Wifi: uwe5622 performance is very good: TX: ~80MBit/s, RX: ~30MBit/s (same distant crappy router of the other tests of mine)
• Bluetooth: works fine, able to stream audio from paired smartphone
• USB2: all ports work fine
• USB3: works ok
• Analog audio codec: works fine

Summary: ready for production ✅

[Ovaday]

X88 PRO TV box based on RK3566 - non-standard board, new DTS added.

As no standard image will be created, I built it locally.

The board lacks modern U-Boot, so the image required disabling Armbian boot scripts:

sudo mv /mnt/armbian/boot/boot.scr /mnt/armbian/boot/boot.scr.disabled 2>/dev/null || true
sudo mv /mnt/armbian/boot/boot.scr.uimg /mnt/armbian/boot/boot.scr.uimg.disabled 2>/dev/null || true

This is expected for this board, potentially in the future I will build some patch specifically for this board. Or maybe manufacturer will release a bord with an updated U-Boot. At least we have a compatable DTS/DTB.

Summary: ready for production ✅
Board continues to be with limited support, without standard image.

[deece]

Mellow Fly C5 tested with both current and edge, boots to a prompt and ethernet works

Summary: ready for production ✅

[alnahian2011]

I tested TX6S-axp313 with the lastest armbian I builded in my local machine.
Now,
What's Working:

• IR
• WiFi
• HDMI
• USB
• Ethernet (Fixed it in Fix ethernet on tx6s-axp-313 #10443 )
• Tried Emmc It works.
• Audio and GUI works ( tested via xfce de and played video from chrome ; just need to go pulseaudio and select 2nd stereo audio)

[retro98boy]

Since the CSC boards cannot be built using GitHub Actions, I built and tested them on my own machine.

Build steps

git checkout -b 26.08-test
git fetch --prune upstream

# Automatic board configs status synchronise
# cc32cd6c6dfdf36bfe89bfe2f3e881f592122fc8
git reset --hard upstream/main

./compile.sh build BOARD=bestv-r3300-l BRANCH=current BUILD_DESKTOP=no BUILD_MINIMAL=yes DOCKER_ARMBIAN_BASE_IMAGE=debian:trixie KERNEL_CONFIGURE=no RELEASE=resolute

./compile.sh build BOARD=cainiao-cniot-core BRANCH=current BUILD_DESKTOP=no BUILD_MINIMAL=yes DOCKER_ARMBIAN_BASE_IMAGE=debian:trixie KERNEL_CONFIGURE=no RELEASE=resolute

./compile.sh build BOARD=smart-am40 BRANCH=current BUILD_DESKTOP=no BUILD_MINIMAL=yes DOCKER_ARMBIAN_BASE_IMAGE=debian:trixie KERNEL_CONFIGURE=no RELEASE=resolute

./compile.sh build BOARD=norco-emb-3531 BRANCH=current BUILD_DESKTOP=no BUILD_MINIMAL=yes DOCKER_ARMBIAN_BASE_IMAGE=debian:trixie KERNEL_CONFIGURE=no RELEASE=resolute

Test results

bestv-r3300-l:

• USB Type-A ok
• FE ok
• RTL8189FTV WiFi ok
• eMMC ok
• microSD ok
• HDMI ok
• audio in 3.5mm ok
• CVBS in 3.5mm not test

cainiao-cniot-core:

I no longer have this device, so I only confirmed that the build completes successfully.

smart-am40:

• USB Type-A ok
• GBE ok
• M.2 WiFi/BT ok
• eMMC ok
• microSD ok
• front HDMI (cdn_dp) ok

norco-emb-3531:

• USB Type-A ok
• GBE ok
• on board USB WiFi ok
• eMMC ok
• microSD ok
• HDMI ok
• DP (cdn_dp) ok
• PCIe x4 slot
• ES8316 sound ok

[jomadeto]

Magicsee C400 Plus not boot, working on it

[lukaszsobala]

I've added Milk-V Duo S very recently, so pretty much everything mentioned in the PR #10326 still works. Not sure if I should do anything more before promoting it. But I think it's too fresh, nobody else seems to have tested it yet. I stand by the board, but I've made a lot of fixes to it which should be upstreamed.

[lch08]

Tested LubanCat-5 IO (RK3588) - Ubuntu Resolute KDE Plasma (Armbian 26.8.1, kernel 6.18.44-current-rockchip64, 2026-08-16):

• ✅ Basics: boot, reboot and shutdown
• ✅ eMMC: ~315MByte/s read, ~249MByte/s write
• ✅ sdcard: ~91MByte/s read, ~42MByte/s write
• ✅ 2x 1000M Ethernet: throughput > 940Mbit/s in both directions
• ✅ HDMI TX0 and TX1
• ✅ HDMI RX
• ✅ USB(4xType-A, 1xType-C)
• ✅ PCIe WiFi
• ✅ PCIe NVMe: ok, ~1.2-1.3GByte/s read, ~719MByte/s write
• ✅ UART4 and UART7
• ✅ CAN0 and CAN1: CAN and CAN-FD

Summary: ready for production ✅

[SuperKali]

Tested on Armbian 26.8.1, Debian Trixie/resolute minimal @igorpecovnik

Board	Kernel	Works	Issues	Untested	Log
arduino-uno-q	7.1.8-edge-qrb2210	eMMC, WiFi, BT, audio, GPU, RTC	cpuidle PSCI fail no KVM	USB, display, video codec	log
nanopct6-lts	6.1.115-vendor-rk35xx	2x 2.5GbE (RTL8125B via PCIe), NVMe, eMMC, SD, SPI NOR, WiFi + BT, USB 2.0/3.0, HDMI x2 + DP + HDMI-in, audio (rt5616/HDMI/DP), GPU (mali), NPU, RGA/VPU, RTC, KVM	IRQ 28 storm (rk_pwm_irq disabled) outdated ATF	MIPI-CSI	log
nanopct6-lts	6.18.44-current-rockchip64	2x 2.5GbE (RTL8125B via PCIe), NVMe, eMMC, SD, SPI NOR, WiFi + BT, USB 2.0/3.0, HDMI x2 + HDMI-in, audio (rt5616/HDMI), GPU (panthor), NPU (rocket), RGA/VPU (V4L2), crypto, IR, RTC, KVM	DP support incomplete in mainline random MACs	MIPI-CSI	log
nanopi-m5	6.1.115-vendor-rk35xx	2x GbE (RTL8211F), HDMI + audio, analog audio (rt5616), UFS, SD, SPI NOR, M.2 NVMe, WiFi, BT, GPU (panfrost), NPU, RGA/VPU, RTC, KVM	—	USB, SATA	log
nanopi-m5	6.18.44-current-rockchip64	2x GbE (RTL8211F), HDMI + audio, analog audio (rt5616), UFS, SD, SPI NOR, M.2 NVMe, WiFi, BT, GPU (panfrost), RTC, KVM	no NPU no encoder/RGA (decode only) no SATA no MIPI-CSI	USB	log
nanopi-r76s	6.1.115-vendor-rk35xx	2x 2.5GbE (RTL8125B via PCIe), HDMI + audio, eMMC, SD, WiFi, BT, GPU (panfrost), NPU, RTC, KVM	—	USB	log
nanopi-r76s	7.1.8-edge-rockchip64	2x 2.5GbE (RTL8125B via PCIe), HDMI + audio, eMMC, SD, WiFi, BT, GPU (panfrost), RTC, KVM	no NPU no encoder/RGA (decode only)	USB	log
youyeetoo-r1-v3	6.1.115-vendor-rk35xx	GbE (RGMII, RTL8211F), eMMC 64GB, SD, NVMe, USB 2.0/3.0, HDMI, MIPI-DSI0/DSI1 (7" display, via overlay), audio (es8323 + HDMI), WiFi + BT (rtl8852be)*, GPU (mali G610), NPU (rknpu, 3 cores), RGA/VPU/IEP, RTC, KVM	CAN broken (gpio0-28 conflict with pwm) random MAC	MIPI-CSI	log
youyeetoo-r1-v3	7.1.8-edge-rockchip64	GbE (RGMII, RTL8211F), eMMC 64GB, SD, NVMe, USB 2.0/3.0, HDMI, audio (es8323), WiFi + BT (rtl8852be)*, GPU (panthor), NPU (rocket, 3 cores), RGA, VPU (hantro dec + vepu121 enc + AV1 dec), crypto (rk2), RTC, KVM	random MAC no IEP (no mainline driver) no MIPI-DSI (no mainline overlay, missing panel driver for the Youyeetoo 7")	CAN (enableable via overlay), MIPI-CSI	log
youyeetoo-yy3588	6.1.115-vendor-rk35xx	2.5GbE (RTL8125B via PCIe), GbE (RGMII, RTL8211F), eMMC 256GB, SD, WiFi (rtw89_8852be via PCIe) + BT (RTL8852 USB), USB 2.0/3.0, HDMI + DP, MIPI-DSI (7" display, via overlay), HDMI-in, audio (es8388 + HDMI + DP), GPU (mali G610), NPU (rknpu, 3 cores), RGA (2x rga3 + rga2), full VPU/mpp stack (vdpu, vepu, 4x jpege, jpegd, 2x rkvdec2, 2x rkvenc2, AV1), IEP, Type-C PD (fusb302), SARADC + adc-keys, pwm-fan, RTC, KVM	random MAC on both NICs no devfreq for VOP/VENC/VDEC (regulators missing in DT)	SATA, MIPI-CSI (camera overlays missing), M.2 NVMe (PCIe 3.0 x4 slot empty)	log
youyeetoo-yy3588	6.18.44-current-rockchip64	2.5GbE (RTL8125B via PCIe), GbE (RGMII, RTL8211F), eMMC 256GB, SD, WiFi (rtw89_8852be via PCIe) + BT (RTL8852 USB), USB 2.0/3.0, HDMI + audio, analog audio (es8388), GPU (panthor), NPU (rocket, 3 cores), RGA, VPU (hantro dec + vepu121 enc + AV1 dec), crypto (rk2), adc-keys, pwm-fan, RTC, KVM	random MAC on both NICs DP support incomplete in mainline HDMI-in not working (snps_hdmirx probes, PHY init times out) no IEP (no mainline driver) no MIPI-DSI (no mainline overlay)	SATA, MIPI-CSI, M.2 NVMe (PCIe 3.0 x4 slot empty)	log
radxa-dragon-q6a	6.18.2-current-qcs6490	SD, GbE (RTL8168 via PCIe), USB 2.0/3.0, HDMI, BT, WiFi (AIC8800D80, vendor driver via dkms), GPU (Adreno A660 + GMU fw), video codec (venus enc/dec), audio (wcd938x, headset + DP jack), aDSP + cDSP + fastrpc, crypto (qcrypto, qcom_rng), RTC (ds1307, battery-backed), UEFI boot	no KVM (HYP mode not available)	DP over USB-C alt mode (Type-C features disabled in firmware), eMMC/UFS module slot (removable, no module fitted in this log), M.2 (second PCIe controller, slot empty)	log
radxa-dragon-q8b	7.0.11-vendor-sc8280xp	SD, NVMe (PCIe Gen3 x2), 2x 2.5GbE (TC956x PCIe switch + QCA8081 PHY), USB 2.0 (hub, HID, mass storage), HDMI (via DP controller), GPU (Adreno A660 + GMU fw), video codec (qcom_iris), audio (wcd938x, headset + 3x DP jack), aDSP + cDSP + fastrpc, EEPROM (at24), qcomtee, RTC (ds1307, battery-backed), UEFI boot	no KVM (HYP mode not available)	DP out, Type-C (nothing connected; pmic-glink fails to link both connectors), USB 3.x (SuperSpeed root hubs up, no device), WiFi/BT (M.2 slot empty), UFS module (not fitted), second Ethernet port (no link)	log
radxa-dragon-q8b	7.1.8-edge-sc8280xp	SD, NVMe (PCIe Gen3 x2), 2x 2.5GbE (TC956x PCIe switch + QCA8081 PHY), USB 2.0 (hub, HID, mass storage), HDMI (via DP controller), GPU (Adreno A660 + GMU fw), video codec (qcom_iris), audio (wcd938x, headset + 3x DP jack), aDSP + cDSP + fastrpc, EEPROM (at24), qcomtee, RTC (ds1307, battery-backed), UEFI boot	no KVM (HYP mode not available) qseecom skipped as untested machine: no efivars, no TPM client no radxa_svc_glink (platform_profile)	DP out, Type-C (nothing connected; pmic-glink fails to link both connectors), USB 3.x (SuperSpeed root hubs up, no device), WiFi/BT (M.2 slot empty), UFS module (not fitted), second Ethernet port (no link)	log

* WiFi/BT tested on a similar board with the M.2 module fitted; the module was not present during this log.

Untested means the driver is there but nothing was connected during boot, not a defect. Empty slots, disconnected monitors and missing RTC batteries are not counted as issues.

[efectn]

Here are the results of the tests i carried out with the boards i have with me currently:

Orange Pi 5:

Vendor Kernel:

• Ethernet works.
• USB ports work as expected.
• Hardware decoding on Chrome doesn't work OOTB (perhaps related to https://github.com/armbian/configng/pull/887/changes )

Current Kernel:

• Ethernet works.
• USB ports except USB-C work as expected.
• VP8, H264, HEVC, AV1 decoders show up in Chromium and AV1 decoding works well under 4k@24 fps.

Khadas Edge 2:

Vendor Kernel:

• WiFi works.
• Bluetooth works.
• Khadas MCU fan control works.
• USB ports work as expected.
• MMC boot works.
• Have some dw_dp "AUX timeout" floods in dmesg output, but not able to test DP and DP audio as i don't have DP-capable type-c cable right now. And it was broken according to some reports in forum, will check it out when i have one.

Current Kernel:

• WiFi works.
• Bluetooth works.
• Khadas MCU fan control works.
• USB ports except USB-C work as expected.
• HW decoding seems present at Chromium.

Banana Pi M5 Pro:
Vendor Kernel:

• USB ports work.
• Wifi and BT work.
• Both ETH ports work.

Edge Kernel:

• All USB ports work, including USB-C.
• Wifi works, BT doesn't. (board: add BT support to bananpim5pro edge kernel #10445 fixes BT stuff)
• Both eth ports work.
• HW decoding seems present at Chromium (HEVC and H264)

Orange Pi 5 Plus:
Current Kernel:

• All USB ports work.
• Both eth ports work.
• Both HDMI ports work, HDMIRX is untested.
• PCIE2 works, did not test PCIe3.
• VP8, H264, HEVC, AV1 decoders show up in Chromium.

Vendor Kernel:

• All USB ports work.
• Both eth ports work.
• Both HDMI ports work, HDMIRX is untested.
• PCIE2 works, did not test PCIe3.

NanoPi R6S/M6:
#10442

General Notes:

• armbian-install is broken for the all images i've tested. Since i use Turkish language, i have to use LC_ALL=C to make it work due to some regex issues.
• Chromium doesn't have any HW decoding on vendor images, same as OPi 5's problem. It seems that, the PPA does not support Resolute.
• USB-C is not stable at the boards running mainline kernel (host mode), except OPi 5 Plus and Bananapi M5 Pro. Will check it out what's wrong with USB-C under mainline later.
• All tests have been conducted with Ubuntu Resolute.

[code002-2]

Testing report for Xiaomi Pad 6S Pro (sheng) — 26.8.1

Tested on the current image build. The following works:

Boot & storage: boots from UFS, stable
CPU: all cores, stable under load
Display: 3048×2032 panel, desktop rendering OK
GPU: Adreno 740 working
WiFi: works
Bluetooth: works
USB-C: host mode + DP (DisplayPort) output detected
Audio: speaker output works
Battery: charging/power status OK
Touchscreen: works (desktop session)
Power/volume keys: work
No release-blocking regressions found from my side on 26.8.1. Thanks!