All Activity

Last photo just to the right of the rk3528 at 3 o'clock are the most likely (the two round gold pins). With power off check with a multi meter if one of them is connected to ground. Then with power applied use an oscilloscope to check if the other one has a clock signal on it. If it has these are probably the maskrom pins. Credit to @Hqnicolas

Many thanks for your investigations! SD card issues are solved, I obviously stopped the system during a write operation - after a shutdown (and card check with an PCs adapter) the errors are gone. But this had no impact. What I did for comparison: I applied the same semi-manual netplan / networkd / hostapd setup sequence to an OrangePiZero, also trixie ( v26.8 rolling for Orange Pi Zero running Armbian Linux 6.18.37-current-sunxi). And guess what: this works without a hiccup! So I looked into the syslog output when connecting. This is the output for the OPi (success!): 2026-07-07T20:47:35.333876+02:00 orangepizero hostapd: wlan0: STA ee:79:bc:4f:11:45 IEEE 802.11: authenticated 2026-07-07T20:47:35.336530+02:00 orangepizero hostapd: wlan0: STA ee:79:bc:4f:11:45 IEEE 802.11: associated (aid 1) 2026-07-07T20:47:35.406476+02:00 orangepizero hostapd: wlan0: STA ee:79:bc:4f:11:45 RADIUS: starting accounting session 01344FB459DCA3B1 2026-07-07T20:47:35.407484+02:00 orangepizero hostapd: wlan0: STA ee:79:bc:4f:11:45 WPA: pairwise key handshake completed (RSN) 2026-07-07T20:47:35.409156+02:00 orangepizero kernel: ieee80211 phy0: CCMP_PAIRWISE keylen=16! And this is the output for the failing BPi: 2026-07-06T17:54:46.957014+02:00 bananapipro hostapd: wlan0: STA 22:50:7b:ac:e3:2e IEEE 802.11: associated 2026-07-06T17:54:46.983182+02:00 bananapipro hostapd: wlan0: STA 22:50:7b:ac:e3:2e RADIUS: starting accounting session 9B014BB4D748D33B 2026-07-06T17:54:46.984065+02:00 bananapipro hostapd: wlan0: STA 22:50:7b:ac:e3:2e WPA: pairwise key handshake completed (RSN) 2026-07-06T17:55:28.848201+02:00 bananapipro hostapd: wlan0: STA 22:50:7b:ac:e3:2e IEEE 802.11: disassociated or: (worst case) 2026-07-07T16:52:15.236724+02:00 bananapipro hostapd: wlan0: STA 22:50:7b:ac:e3:2e IEEE 802.11: associated 2026-07-07T16:52:15.270595+02:00 bananapipro hostapd: wlan0: STA 22:50:7b:ac:e3:2e RADIUS: starting accounting session E501E4B1A6220EF1 2026-07-07T16:52:15.271648+02:00 bananapipro hostapd: wlan0: STA 22:50:7b:ac:e3:2e WPA: pairwise key handshake completed (RSN) 2026-07-07T16:52:30.104160+02:00 bananapipro kernel: ieee80211 phy0: brcmf_psm_watchdog_notify: PSM's watchdog has fired! 2026-07-07T16:55:07.185663+02:00 bananapipro kernel: brcmfmac: brcmf_sdio_bus_rxctl: resumed on timeout 2026-07-07T16:55:07.185794+02:00 bananapipro kernel: ieee80211 phy0: brcmf_cfg80211_get_channel: chanspec failed (-110) 2026-07-07T16:55:09.745805+02:00 bananapipro kernel: brcmfmac: brcmf_sdio_bus_rxctl: resumed on timeout 2026-07-07T16:55:09.749180+02:00 bananapipro kernel: ieee80211 phy0: brcmf_cfg80211_get_tx_power: error (-110) 2026-07-07T16:57:18.001649+02:00 bananapipro kernel: brcmfmac: brcmf_sdio_bus_rxctl: resumed on timeout 2026-07-07T16:57:20.561670+02:00 bananapipro kernel: ieee80211 phy0: brcmf_proto_bcdc_query_dcmd: brcmf_proto_bcdc_msg failed w/status -110 2026-07-07T16:57:20.562221+02:00 bananapipro kernel: ieee80211 phy0: brcmf_cfg80211_get_station: GET STA INFO failed, -110 2026-07-07T16:58:23.281661+02:00 bananapipro kernel: ieee80211 phy0: brcmf_proto_bcdc_query_dcmd: brcmf_proto_bcdc_msg failed w/status -110 2026-07-07T16:58:23.282246+02:00 bananapipro kernel: ieee80211 phy0: brcmf_cfg80211_get_channel: chanspec failed (-110) .... (more errors) The ip link/ ip route is identical, for the OPi also the networkctl output is clean: root@orangepizero:/home/thomas/work# networkctl IDX LINK TYPE OPERATIONAL SETUP 1 lo loopback carrier unmanaged 2 end0 ether enslaved configured 3 wlan0 wlan enslaved configured 5 br0 bridge routable configured 4 links listed. I also started to compare the log outputs for the start of hostapd, many things look equal / "similar" (maybe due to different underlaying hardware). I could provide the 2 (streamlined) logs, here only the most obvious deltas: nl80211: flush -> DEL_STATION wlan0 (all) --> nl80211: Station flush failed: ret=-14 (Bad address) (BPi) <none> (OPi) <TX queue cmds OPi> nl80211: TX queue param set: queue=0 aifs=1 cw_min=3 cw_max=7 burst_time=15 --> res=0 nl80211: TX queue param set: queue=1 aifs=1 cw_min=7 cw_max=15 burst_time=30 --> res=0 nl80211: TX queue param set: queue=2 aifs=3 cw_min=15 cw_max=63 burst_time=0 --> res=0 nl80211: TX queue param set: queue=3 aifs=7 cw_min=15 cw_max=1023 burst_time=0 --> res=0 <TX queue cmds BPi> nl80211: TX queue param set: queue=0 aifs=1 cw_min=3 cw_max=7 burst_time=15 --> res=-95 Failed to set TX queue parameters for queue 0. nl80211: TX queue param set: queue=1 aifs=1 cw_min=7 cw_max=15 burst_time=30 --> res=-95 Failed to set TX queue parameters for queue 1. nl80211: TX queue param set: queue=2 aifs=3 cw_min=15 cw_max=63 burst_time=0 --> res=-95 Failed to set TX queue parameters for queue 2. nl80211: TX queue param set: queue=3 aifs=7 cw_min=15 cw_max=1023 burst_time=0 --> res=-95 Failed to set TX queue parameters for queue 3. Question is, if I have to give up on the hardware or try a SW workaround (as my old solution running on BPiPro Bookworm / NM: wifi set to umanaged and using a hand-made hostapd / dnsmasq / iptables solution - will this work for Trixie? Do not know.)

You disable the zram swap, I forgot how, I think look in /etc/default/ambian* files. Else disable or mask the specific .service file.

This is no success, it id wrong implementation, it is AI, from concept/principel it seems OK, but implementation is NOK So I think I stop experimenting and will use NM if bridged WiPi AP is needed, I have several optional working in Debian Trixie and other distro. For wired lines, systemd-networkd with brideges and VLANs work OK. As said already, Debian has netplan optional, so in Armbain Trixie I simply do apt purge --autoremove netplan.io (and install NM via apt install network-manager). Then via nmtui it is easy to setup a bridged WiFi AP, no hostapd needed. By the way, you have several errors w.r.t. your SD-card, might be a sunxi kernel issue, but also you brad SD-card or so.

Problem seems to be that the wlan0 somehow must have a carrier, else it cannot be added to the bridge. A quick hack in a running system, done via serial terminal is: root@raspi7:/etc/systemd/network# systemctl stop systemd-networkd ; systemctl restart hostapd ; systemctl start systemd-networkd Then: root@raspi7:/etc/systemd/network# networkctl IDX LINK TYPE OPERATIONAL SETUP 1 lo loopback carrier unmanaged 2 eth0 ether enslaved configured 3 br0 bridge routable configured 4 wlan0 wlan enslaved configured I did some more testing and also with wlan0 stated as fail, I could browse the internet on my smarthone via the AP. Note that in the meantime, I also simplified the 10-* files a bit, added a MACAddress same as eth0 for br0, so my routers issues same IP address when DHCP, but that is not fundamental to this issue I think, i more how I do it also via NM and how it was automatically in old Linuxes. Google/Gemini suggests to add Before=systemd-networkd.service to hostapd.service, maybe I try, but I find that a dirty hack, it should be already in the OS, although Trixie is not the latest Linux.

I made a comparable setup on a raspberrypi3b+ and I see in the journal: Jul 07 17:37:01 raspi7 systemd-networkd[292]: wlan0: Configuring with /etc/systemd/network/10-wlan0.network. Jul 07 17:37:01 raspi7 systemd-networkd[292]: wlan0: Failed to set master interface: Device does not allow enslaving to a bridge. Operation not supported Jul 07 17:37:01 raspi7 systemd-networkd[292]: wlan0: Failed It worked many years ago when it was buster or bullseye and using ifupdown interfaces network setup. Also ported that to NetworkManager before Bookworm and that works, also in Trixie. This is my first manual setup, as it is Debian Trixie based, I can avoid the use of netplan.io. I remember I have seen this issue somewhere, It might be a newer release systemd issue. But have not searched internet now, will see later why this is.

Well, netplan is the advertised / provided solution for Armbian. So when you look into /etc/systemd/network you find a readme from armbian pointing you to /etc/netplan. I think what netplan does is generating the following files for systemd-networkd under /run/systemd/network: 10-netplan-end0.network: [Match] Name=end0 [Link] MACAddress=02:cd:08:83:3c:f3 [Network] LinkLocalAddressing=no Bridge=br0 10-netplan-wlan0.network [Match] Name=wlan0 [Network] LinkLocalAddressing=no Bridge=br0 10-netplan-br0.network [Match] Name=br0 [Network] LinkLocalAddressing=ipv6 Address=192.168.178.29/24 DNS=9.9.9.9 DNS=1.1.1.1 ConfigureWithoutCarrier=yes [Route] Destination=0.0.0.0/0 Gateway=192.168.178.1 Metric=200 10-netplan-br0.netdev [NetDev] Name=br0 Kind=bridge And this is the ip output: ip route: default via 192.168.178.1 dev br0 proto static metric 200 192.168.178.0/24 dev br0 proto kernel scope link src 192.168.178.29 ip link: 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 2: end0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq master br0 state UP mode DEFAULT group default qlen 1000 link/ether 02:cd:08:83:3c:f3 brd ff:ff:ff:ff:ff:ff altname enx02cd08833cf3 3: br0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP mode DEFAULT group default qlen 1000 link/ether a2:67:ed:53:cb:f1 brd ff:ff:ff:ff:ff:ff 4: wlan0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast master br0 state UP mode DEFAULT group default qlen 1000 link/ether e0:76:d0:03:53:bc brd ff:ff:ff:ff:ff:ff altname wlxe076d00353bc

Thanks for the reply @eselarm. I was wondering if the install script was doing more than just flashing the SPI and cloning the partition, but if that's all it does I'm good. Regarding Zram and ramlog, I'm not concerned about how much the Nvme SSD gets written, I mainly use the Rock 5B as a low power NAS for temporary files, it downloads a ton of stuff that get shared on the other devices on my LAN, and after a while the files get deleted, so the work load is already more intense than what a bit of swapping and logging is going to do. But I sometimes use it to host simple game servers for my friends, so from time to time I need all the possible RAM, would disabling ZRAM and ramlog be beneficial, considering the SSD is fast enough to act as an efficient swap?

AI development is shifting from implementing models from scratch to composing powerful capabilities via APIs, enabling developers to integrate speech recognition, language models, and tool execution into useful applications with far less effort. View the full article

I had a quick look at the script, but it is too complex for me to see what could be wrong. Also there is netplan layer in between that I am not familiar with anymore. Maybe you can look at what is generated in terms of systemd-networkd files (e.g. in /etc/systemd/network/ ). And also use commands like ip route etc to see the actual state.

As far as my experience is, there is no difference in final result if you do manual 'install' like you did. It runs the usual (user) setup scripts. It is even the same when you start an image as container, done that many times for various platforms. In fact writing SPI-flash is non-automatic, is manual as can be quite invasive. With modern ARM64 SBC like ROCK5B also people might threat it like a PC, with difference that PC comes with a motherboard that has already firmware/bootloader/UEFI/BIOS written by the manufacturer. Although some motherbaords also can use open source, like coreboot I think, but guessing. What to do as further tuning and customization depends on what you want to do with the ROCK5B. There is 1.5 decades between the single core ARMv6 based RaspberryPi1 and the current high-end SBC's that are 100x faster and can be treated like an Intel/AMD PC. I actually have no NVME for x86-64 (is still SATA SSD although also M.2 slot as well), but use my 500GB Samsung 970 EVO+ in the ROCK5B. I use Btrfs and compress-force=zstd for almost all Linux rootfs, so always create an extra swap partition, although a swapfile also works nowadays with special create options. I rather not use the (currently very expensive) RAM for swap when 4-lane PCIE v3 NVME is there. But that is also because my ROCK5B acts as a server, containers and KVM etc and also I use the RAM and NVME for caching a large HDD. So in terms of wear of flash storage, that caching is just way more demanding than what happens on a decade old SD-card only SBC. But NVME/SATA SSD's have perfectly readable stats and many good brands can write more TeraBytes than the specs mention from manufacturer. SD-cards can be very poorly design internally, and have no stats. So you simply don't know when they will break. If the SBC only does some simple things like camera streaming or some sensor's on GPIO pins, it makes sense to keep as much as possible in RAM. So then just keep it like it is. Or even disable journal to non-volatile storage if never any issues and only enable if troubles. One problem I had with zram swap is that on a 256MiB NanoPi-NEO, it failed when media streaming under memory pressure. mp3 or opus or h264 etc don't compress and can still buffer a lot, so the OOM killer started acting. I was not able to change the whole media stuff so stopped zram swap and put swap on normal extra swap partition.

Hmm... that's strange. Does no one really want to participate in this discussion ? Has any of you really never used the NanoPi M5 before ? Don't you want us to join forces and work together to make this SBC run great ? It is a great piece of hardware who deserves a software to match. Let's show the world that there is a community for other SBCs besides Raspberry Pi ! So far we know for sure that the NanoPi M5 hardware is capable of full 4K. It is already working in Android14 OS. We just need to move the video drivers and software library froma Android to Armbian.

Greetings! My first foray into Armbian. Apparently the Link Star H68K is hardware equivalent to the Hinlink H68K. I'm trying to install one of the community supported Armbian images (Debian Trixie, Ubuntu 26.04) for the Hinlink onto a Link Star H68K, using the RKDevTool from the Seeed Studio wiki. And failing. The only image that loads into RKDevTool is the Lubuntu 20.04 from the Seeed Studio wiki. It's just a tad out of date. Any pointers on getting a more up-to-date OS installed would be most welcome. AFAIK my device is OG, not V2. I've also tried booting from TF card, but the device refuses to boot at all with a TF card inserted. Some Googling suggests that a current TF card is too fast for the device and causes boot failure. Card spec is Class 10 / UHS-1. I've sourced some Sandisk Class 4 cards and currently awaiting delivery, to see if that solves the boot issue and gives me a path to transfer the OS from TF to EMMC. Any thoughts/advice most welcome.

This week's cycle emphasizes broad U-Boot modernization, new board and SoC enablement, and kernel and wireless driver consolidation. A large-scale U-Boot bump moves sunxi 32-bit and 64-bit targets from v2024.01 to v2026.07-rc4, with follow-on updates for self-pinned H616/H618 boards (Zero2W, Zero3, Longan Pi 3H), Mixtile Edge2, NanoPi R5S (now patch-less), and the Youyeetoo YY3588 switching to mainline v2026.04. The imx6 line (UDOO, Cubox-i) was modernized to U-Boot v2026.07 with legacy 6.12, current 6.18, and edge 7.1 kernels. Related toolchain work fixes ODROID-C1, ODROID-XU4, Recore, and X96Q builds under Trixie's GCC 14, and resolves errexit failures on Rockchip SPI boards. Platform expansion introduces community support for the Allwinner A733-based Radxa Cubie A7Z and Orange Pi Zero 3W, Rockchip Graperain G3568 v2, and Anbernic RG Vita Pro and Lubancat-5IO image entries. BeagleY-AI gained USB, PCIe, ISP + IMX219, and VPAC patches on the vendor kernel, alongside GPU acceleration fixes for TI K3 targets and TI Wave5 VPU firmware. Rockchip RV1106 support was split into distinct RV1103G and RV1103B families, and new SPI/NVMe boot and Maskrom recovery paths were added. On the kernel and driver side, sunxi received an H3/H5 DVFS RCU-stall fix, MMC/I2C PM deadlock resolution, MGLRU enablement, and LTE modem USB serial support. Meson64 gained a GPIO pinctrl cansleep series and v7.2-rc1 via bleedingedge, while SpacemiT K1 was updated to linux-7.2.y. The RTL8189ES, RTL8189FS, and RTL8192EU wireless drivers were migrated to dedicated forks with 7.2 compatibility and patch cleanup, and an RTW88 SDIO interrupt storm was addressed. User-visible improvements include swapfile creation fixes, useradd-based first-login provisioning, and video-group access to Rockchip MPP codec devices. #Armbian #EmbeddedLinux #UBoot #Rockchip #Allwinner Changesboard: visionfive2: optimize config for modern U-Boot and single-partition booting. by @zevlaoxie2010 in armbian/build#10097Add Anbernic RG Vita Pro board image. by @crackerjacques in armbian/armbian.github.io#351Add Lubancat-5IO board image and vendor logo. by @lch08 in armbian/armbian.github.io#348Add TI Wave5 VPU firmware. by @Grippy98 in armbian/firmware#132beaglebadge: drop vendor-edge u-boot target (not in ti-u-boot-2026.01). by @igorpecovnik in armbian/build#10126BeagleY-AI Add Patches for ISP + IMX219 & VPAC. by @Grippy98 in armbian/build#10129board: anbernic-rg-ds: make OTG-C port dual-role so USB hubs work. by @crackerjacques in armbian/build#10124Build-Switches: document KERNEL_DO_STUBBLE. by @iav in armbian/documentation#928Bump/rk uboot swig fleet mkspi 2. by @redrathnure in armbian/build#10122chore(deps): patch vulnerable transitive deps. by @SuperKali in armbian/website#33ci(docs-preview): skip publish/comment steps on fork PRs. by @iav in armbian/documentation#934ci: auto-regenerate build lists when image-info.json changes. by @igorpecovnik in armbian/armbian.github.io#349ci: bump image build cap 30m → 45m. by @igorpecovnik in armbian/os-ci-test#24ci: widen auto-retry budget and the systemic-failure threshold. by @igorpecovnik in armbian/os-ci-test#23clk: sunxi-ng: ccu_common: Replace ktime-dependent lock wait loop with udelay. by @deece in armbian/build#10119config: rockpi-e: change rkbin repository and use spl-blobs. by @chenxmqqqq in armbian/build#10027cubie-a7z: Add community support for Radxa Cubie A7Z (Allwinner A733). by @shkolnik in armbian/build#10036docs: add Datacenter access section (Netbird VPN + board access). by @igorpecovnik in armbian/documentation#936docs: document PREFER_NATIVE_ARMHF and enrich arm64-compat-vdso entry. by @iav in armbian/documentation#921docs: exclude docs/README.md from mkdocs build. by @iav in armbian/documentation#933docs: use unpinned python3-dev in setup instructions. by @igorpecovnik in armbian/documentation#935Drop testing-wireless-performance-test.yml. by @igorpecovnik in armbian/armbian.github.io#350Enable USB and PCIe on BeagleY-AI vendor kernel. by @Grippy98 in armbian/build#10115fake-ubuntu-advantage-tools: provide ubuntu-pro-client. by @Grippy98 in armbian/build#10116family-rv1106: Split RV1103G and RV1103B, cleanup overrides. by @vidplace7 in armbian/build#10011feat(api): serve QDL board metadata and proxy-cache firehose assets. by @SuperKali in armbian/website#31feat(qdl): serve the board registry from the API. by @SuperKali in armbian/imager#163firstlogin: use useradd/groupadd instead of the adduser suite. by @igorpecovnik in armbian/build#10127Fix swapfile creation. by @pierg75 in armbian/build#10065Fix TI K3 Suite Matching for Ubuntu Hosts and Enable GPU Acceleration for BeagleBoards. by @Grippy98 in armbian/build#10112fix: stop setting LC_ALL in armbian-lang.sh. by @saschabuehrle in armbian/build#10132Guard config-disabled static helpers to silence -Wunused-function. by @iav in armbian/wifi-rtl8852bs#9Import long-carried armbian/build regd-init fix, with authorship. by @iav in armbian/wifi-rtl8852bs#8imx6 modernization: udoo + cubox-i u-boot v2026.07 + kernel (legacy 6.12 / current 6.18 / edge 7.1). by @igorpecovnik in armbian/build#10098kernel-rust: ship prebuilt Rust crate artifacts in linux-headers. by @iav in armbian/build#10117Makena: add SC8280XP firehose and Radxa Dragon Q8B UFS provisioning. by @SuperKali in armbian/qcombin#3meson64: add GPIO shared-proxy/pinctrl cansleep series (upstream v3). by @adeepn in armbian/build#10090meson64: enable v7.2-rc1 via bleedingedge. by @EvilOlaf in armbian/build#10079mixtile-edge2: u-boot: main: bump (generic) u-boot to v2026.07-rc5. by @rpardini in armbian/build#10071nanopi-r5s: u-boot: main: bump u-boot to v2026.07-rc5 + make it patch-less. by @rpardini in armbian/build#10072odroidc1: fix u-boot build on trixie (GCC 14) via config.mk -fpermissive. by @igorpecovnik in armbian/build#10111odroidxu4: fix u-boot build on trixie (GCC 14 implicit-decl errors). by @igorpecovnik in armbian/build#10110orangepizero3w: Add community support for Orange Pi Zero 3W (Allwinner A733). by @shkolnik in armbian/build#10103recore: fix ATF v2.8.0 build - drop -Wl, prefix from LD flag. by @igorpecovnik in armbian/build#10100refactor(api): drop min_imager_version from QDL enrichment. by @SuperKali in armbian/website#32rockchip64: fix errexit build failure on SPI boards (undefined hook). by @igorpecovnik in armbian/build#10101rockchip64_common: add board_uboot postprocess hooks. by @mingzhangqun in armbian/build#10031rockchip: add support for Graperain G3568 v2. by @rbqvq in armbian/build#10056rockchip: udev: grant video group access to MPP codec devices and DMA-heaps. by @yisding in armbian/build#10085rtl8189es: switch fork, clean up patches, bump to 7.2. by @EvilOlaf in armbian/build#10086rtl8189fs: switch to own fork, add 7.2 compatibility, drop patches from misc. by @EvilOlaf in armbian/build#10092rtl8192eu: switch to own fork, add 7.2 compat, remove integrated patches from misc. by @EvilOlaf in armbian/build#10095shellfmt: run lib/tools/shellfmt.sh, no changes. by @rpardini in armbian/build#10087SpacemiT K1: Update EDGE to linux-7.2.y. by @pyavitz in armbian/build#10123sunxi (32+64-bit): bump u-boot v2024.01 -> v2026.07-rc4. by @igorpecovnik in armbian/build#10089sunxi, sunxi64, sun60iw2: enable MGLRU in kernel config. by @ValdikSS in armbian/build#10133sunxi-6.18: fix H3/H5 RCU stall / CPU hang during DVFS (current kernel). by @igorpecovnik in armbian/build#10083sunxi64: bump self-pinned H616/H618 boards to v2026.07 (zero2w, zero3, longanpi-3h). by @igorpecovnik in armbian/build#10102sunxi64: enable USB_SERIAL_OPTION (module) for GSM/UMTS/LTE modem support. by @pumuckly in armbian/build#10050sunxi: pin mmc and i2c controllers to active to break circular PM deadlock. by @deece in armbian/build#10120uboot: add SPI/NVMe boot support and Maskrom SD-card/flash fixes. by @mingzhangqun in armbian/build#10030wifi: rtw88: sdio: Fix unhandled RX request interrupt storm. by @deece in armbian/build#10121WifiPerformance: update infrastructure section to autotests framework. by @igorpecovnik in armbian/documentation#931x96q: fix u-boot build hang — add missing CONFIG_DRAM_SUNXI_TPR6. by @igorpecovnik in armbian/build#10114youyeetoo-yy3588: switch to mainline U-Boot v2026.04. by @SuperKali in armbian/build#10125youyeetoo: fix 7" touch with TPC2059 config, add YY3588 display overlays. by @SuperKali in armbian/linux-rockchip#507yy3588: fix eDP video port and timing, document panel requirement. by @SuperKali in armbian/linux-rockchip#508View the full article

This week's cycle emphasizes broad U-Boot modernization, new board and SoC enablement, and kernel and wireless driver consolidation. A large-scale U-Boot bump moves sunxi 32-bit and 64-bit targets from v2024.01 to v2026.07-rc4, with follow-on updates for self-pinned H616/H618 boards (Zero2W, Zero3, Longan Pi 3H), Mixtile Edge2, NanoPi R5S (now patch-less), and the Youyeetoo YY3588 switching to mainline v2026.04. The imx6 line (UDOO, Cubox-i) was modernized to U-Boot v2026.07 with legacy 6.12, current 6.18, and edge 7.1 kernels. Related toolchain work fixes ODROID-C1, ODROID-XU4, Recore, and X96Q builds under Trixie's GCC 14, and resolves errexit failures on Rockchip SPI boards. Platform expansion introduces community support for the Allwinner A733-based Radxa Cubie A7Z and Orange Pi Zero 3W, Rockchip Graperain G3568 v2, and Anbernic RG Vita Pro and Lubancat-5IO image entries. BeagleY-AI gained USB, PCIe, ISP + IMX219, and VPAC patches on the vendor kernel, alongside GPU acceleration fixes for TI K3 targets and TI Wave5 VPU firmware. Rockchip RV1106 support was split into distinct RV1103G and RV1103B families, and new SPI/NVMe boot and Maskrom recovery paths were added. On the kernel and driver side, sunxi received an H3/H5 DVFS RCU-stall fix, MMC/I2C PM deadlock resolution, MGLRU enablement, and LTE modem USB serial support. Meson64 gained a GPIO pinctrl cansleep series and v7.2-rc1 via bleedingedge, while SpacemiT K1 was updated to linux-7.2.y. The RTL8189ES, RTL8189FS, and RTL8192EU wireless drivers were migrated to dedicated forks with 7.2 compatibility and patch cleanup, and an RTW88 SDIO interrupt storm was addressed. User-visible improvements include swapfile creation fixes, useradd-based first-login provisioning, and video-group access to Rockchip MPP codec devices. #Armbian #EmbeddedLinux #UBoot #Rockchip #Allwinner Changesboard: visionfive2: optimize config for modern U-Boot and single-partition booting. by @zevlaoxie2010 in armbian/build#10097Add Anbernic RG Vita Pro board image. by @crackerjacques in armbian/armbian.github.io#351Add Lubancat-5IO board image and vendor logo. by @lch08 in armbian/armbian.github.io#348Add TI Wave5 VPU firmware. by @Grippy98 in armbian/firmware#132beaglebadge: drop vendor-edge u-boot target (not in ti-u-boot-2026.01). by @igorpecovnik in armbian/build#10126BeagleY-AI Add Patches for ISP + IMX219 & VPAC. by @Grippy98 in armbian/build#10129board: anbernic-rg-ds: make OTG-C port dual-role so USB hubs work. by @crackerjacques in armbian/build#10124Build-Switches: document KERNEL_DO_STUBBLE. by @iav in armbian/documentation#928Bump/rk uboot swig fleet mkspi 2. by @redrathnure in armbian/build#10122chore(deps): patch vulnerable transitive deps. by @SuperKali in armbian/website#33ci(docs-preview): skip publish/comment steps on fork PRs. by @iav in armbian/documentation#934ci: auto-regenerate build lists when image-info.json changes. by @igorpecovnik in armbian/armbian.github.io#349ci: bump image build cap 30m → 45m. by @igorpecovnik in armbian/os-ci-test#24ci: widen auto-retry budget and the systemic-failure threshold. by @igorpecovnik in armbian/os-ci-test#23clk: sunxi-ng: ccu_common: Replace ktime-dependent lock wait loop with udelay. by @deece in armbian/build#10119config: rockpi-e: change rkbin repository and use spl-blobs. by @chenxmqqqq in armbian/build#10027cubie-a7z: Add community support for Radxa Cubie A7Z (Allwinner A733). by @shkolnik in armbian/build#10036docs: add Datacenter access section (Netbird VPN + board access). by @igorpecovnik in armbian/documentation#936docs: document PREFER_NATIVE_ARMHF and enrich arm64-compat-vdso entry. by @iav in armbian/documentation#921docs: exclude docs/README.md from mkdocs build. by @iav in armbian/documentation#933docs: use unpinned python3-dev in setup instructions. by @igorpecovnik in armbian/documentation#935Drop testing-wireless-performance-test.yml. by @igorpecovnik in armbian/armbian.github.io#350Enable USB and PCIe on BeagleY-AI vendor kernel. by @Grippy98 in armbian/build#10115fake-ubuntu-advantage-tools: provide ubuntu-pro-client. by @Grippy98 in armbian/build#10116family-rv1106: Split RV1103G and RV1103B, cleanup overrides. by @vidplace7 in armbian/build#10011feat(api): serve QDL board metadata and proxy-cache firehose assets. by @SuperKali in armbian/website#31feat(qdl): serve the board registry from the API. by @SuperKali in armbian/imager#163firstlogin: use useradd/groupadd instead of the adduser suite. by @igorpecovnik in armbian/build#10127Fix swapfile creation. by @pierg75 in armbian/build#10065Fix TI K3 Suite Matching for Ubuntu Hosts and Enable GPU Acceleration for BeagleBoards. by @Grippy98 in armbian/build#10112fix: stop setting LC_ALL in armbian-lang.sh. by @saschabuehrle in armbian/build#10132Guard config-disabled static helpers to silence -Wunused-function. by @iav in armbian/wifi-rtl8852bs#9Import long-carried armbian/build regd-init fix, with authorship. by @iav in armbian/wifi-rtl8852bs#8imx6 modernization: udoo + cubox-i u-boot v2026.07 + kernel (legacy 6.12 / current 6.18 / edge 7.1). by @igorpecovnik in armbian/build#10098kernel-rust: ship prebuilt Rust crate artifacts in linux-headers. by @iav in armbian/build#10117Makena: add SC8280XP firehose and Radxa Dragon Q8B UFS provisioning. by @SuperKali in armbian/qcombin#3meson64: add GPIO shared-proxy/pinctrl cansleep series (upstream v3). by @adeepn in armbian/build#10090meson64: enable v7.2-rc1 via bleedingedge. by @EvilOlaf in armbian/build#10079mixtile-edge2: u-boot: main: bump (generic) u-boot to v2026.07-rc5. by @rpardini in armbian/build#10071nanopi-r5s: u-boot: main: bump u-boot to v2026.07-rc5 + make it patch-less. by @rpardini in armbian/build#10072odroidc1: fix u-boot build on trixie (GCC 14) via config.mk -fpermissive. by @igorpecovnik in armbian/build#10111odroidxu4: fix u-boot build on trixie (GCC 14 implicit-decl errors). by @igorpecovnik in armbian/build#10110orangepizero3w: Add community support for Orange Pi Zero 3W (Allwinner A733). by @shkolnik in armbian/build#10103recore: fix ATF v2.8.0 build - drop -Wl, prefix from LD flag. by @igorpecovnik in armbian/build#10100refactor(api): drop min_imager_version from QDL enrichment. by @SuperKali in armbian/website#32rockchip64: fix errexit build failure on SPI boards (undefined hook). by @igorpecovnik in armbian/build#10101rockchip64_common: add board_uboot postprocess hooks. by @mingzhangqun in armbian/build#10031rockchip: add support for Graperain G3568 v2. by @rbqvq in armbian/build#10056rockchip: udev: grant video group access to MPP codec devices and DMA-heaps. by @yisding in armbian/build#10085rtl8189es: switch fork, clean up patches, bump to 7.2. by @EvilOlaf in armbian/build#10086rtl8189fs: switch to own fork, add 7.2 compatibility, drop patches from misc. by @EvilOlaf in armbian/build#10092rtl8192eu: switch to own fork, add 7.2 compat, remove integrated patches from misc. by @EvilOlaf in armbian/build#10095shellfmt: run lib/tools/shellfmt.sh, no changes. by @rpardini in armbian/build#10087SpacemiT K1: Update EDGE to linux-7.2.y. by @pyavitz in armbian/build#10123sunxi (32+64-bit): bump u-boot v2024.01 -> v2026.07-rc4. by @igorpecovnik in armbian/build#10089sunxi, sunxi64, sun60iw2: enable MGLRU in kernel config. by @ValdikSS in armbian/build#10133sunxi-6.18: fix H3/H5 RCU stall / CPU hang during DVFS (current kernel). by @igorpecovnik in armbian/build#10083sunxi64: bump self-pinned H616/H618 boards to v2026.07 (zero2w, zero3, longanpi-3h). by @igorpecovnik in armbian/build#10102sunxi64: enable USB_SERIAL_OPTION (module) for GSM/UMTS/LTE modem support. by @pumuckly in armbian/build#10050sunxi: pin mmc and i2c controllers to active to break circular PM deadlock. by @deece in armbian/build#10120uboot: add SPI/NVMe boot support and Maskrom SD-card/flash fixes. by @mingzhangqun in armbian/build#10030wifi: rtw88: sdio: Fix unhandled RX request interrupt storm. by @deece in armbian/build#10121WifiPerformance: update infrastructure section to autotests framework. by @igorpecovnik in armbian/documentation#931x96q: fix u-boot build hang — add missing CONFIG_DRAM_SUNXI_TPR6. by @igorpecovnik in armbian/build#10114youyeetoo-yy3588: switch to mainline U-Boot v2026.04. by @SuperKali in armbian/build#10125youyeetoo: fix 7" touch with TPC2059 config, add YY3588 display overlays. by @SuperKali in armbian/linux-rockchip#507yy3588: fix eDP video port and timing, document panel requirement. by @SuperKali in armbian/linux-rockchip#508View the full article

I tried to build an bridged AP so that I can reach the Internet from a WiFi net spawned from the BananaPiPro. By this I followed the procedure from armbian-config (config.network.sh). After install the AP is shown in the login screen: WiFi AP: SSID: (BPi), channel 7 (2442 MHz), width: 20 MHz, center1: 2442 MHz When I try to connect from my mobile phone it either says "connecting without Internet". The output in syslog is then: 026-07-06T12:35:02.645970+02:00 bananapipro hostapd: wlan0: STA 22:50:7b:ac:e3:2e IEEE 802.11: associated 2026-07-06T12:35:02.690671+02:00 bananapipro hostapd: wlan0: STA 22:50:7b:ac:e3:2e RADIUS: starting accounting session 500A81EDD1715454 2026-07-06T12:35:02.691497+02:00 bananapipro hostapd: wlan0: STA 22:50:7b:ac:e3:2e WPA: pairwise key handshake completed (RSN) Ort otherwise the phone just says "cannot connect". My netplan file is as follows: network: version: 2 renderer: networkd ethernets: end0: dhcp4: false dhcp6: false macaddress: "02:cd:08:83:3c:f3" wlan0: dhcp4: false dhcp6: false bridges: br0: addresses: - "192.168.178.29/24" nameservers: addresses: - 9.9.9.9 - 1.1.1.1 interfaces: - wlan0 - end0 routes: - metric: 200 to: default via: "192.168.178.1" The hostapd.conf is: interface=wlan0 driver=nl80211 ssid=BPi hw_mode=g channel=7 wmm_enabled=0 macaddr_acl=0 auth_algs=1 ignore_broadcast_ssid=0 wpa=2 wpa_passphrase=secret wpa_key_mgmt=WPA-PSK wpa_pairwise=TKIP rsn_pairwise=CCMP ctrl_interface=/var/run/hostapd ctrl_interface_group=0 ieee80211n=1 ht_capab=[DSSS_CK-40][HT20+] preamble=1 bridge=br0 The output of network list is: root@bananapipro:/etc/hostapd# networkctl list IDX LINK TYPE OPERATIONAL SETUP 1 lo loopback carrier unmanaged 2 end0 ether enslaved configured 3 br0 bridge routable configured 4 wlan0 wlan enslaved failed Does anybody have an idea where to look further for a solution? As I have an AP running on my other BananaPiPro (Bookworm CLI / Networkmanager, manual config, not bridged) it cannot be a complete HW limitation, I think.

Hi, I just switched from an ancient ubuntu-rockchip install to armbian ubuntu 26.04 with the mainline kernel. I did a bit of an unorthodox installation, I wanted to install on an nvme drive, and instead of using a microSD and then copying the system with armbian-install, since I didn't have a working microsd available I did a "manual" installation, I flashed rkspi_loader.img into the SPI taken from the armbian image and then flashed the armbian img directly to the nvme drive with an usb adapter from my PC. All is good and working fine, but I'm wondering if I'm missing any settings that gets applied during the armbian-install process (are there any?) that are specific for nvme. And regarding Zram, ram2log and other optimizations that armbian uses for SDs, how should I configure them? My setup is a Radxa Rock 5B with 8GB of RAM with a Samsung 970 evo plus, should I disable zram and create a swapfile? Any other setting to look out for? I think my hdparm -t speed is a bit low, around 1300 MB/s. Thanks in advance

I was wrong, the dafault baudrate is 100 kHz - the sensor is not responding at first try, when it is in sleep mode.

Hi Dodeval, so I am having trouble accessing Gemini 3.1 pro chat about this but this sounds vaguely familiar. Things that could be the cause are the "mkfs.ext4 -b 4096" option and "debootstrap --arch=arm64". Try some of these top tier models https://artificialanalysis.ai/leaderboards/models they should be able to crack the code even though you might have to wrestle a bit to get them to produce the right solution. Most things seem to work like magic I guess it is about the device tree, battery management seems fine etc.. touch needs drivers, no audio from internal speakers which is no big deal anyway as bluetooth audio works everything there works as expected. I did get this aliexpress.com/item/1005009649297050.html at a discount but I not completely get the touchpad to work comfortably with a patched input driver, a lot of miss-clicks goosh nice keyboard but still need bluetooth mouse in reality. Rii i4 Mini Bluetooth would be fine but then again it is not quite the desktop experience I was looking for.

Hi, I need help identifying the Maskrom/eMMC short points for an RK3528 TV box. Device: - TV box sold as H20 Pro - SoC confirmed as Rockchip RK3528 - Board marking: M16-R6B-L3-V1.0-260121 - RAM detected in Linux: 963 MiB - eMMC detected in Linux: 7.83 GB / 7.30 GiB - eMMC name from Linux: P1J95K - eMMC CID: fe014e50314a39354b12739109744200 What happened: - The box originally booted Android from eMMC. - Armbian booted successfully from SD using a customized joilg/x88pro-based image. - Later I wrote U-Boot/idbloader to eMMC: - idbloader.img at sector 64 - u-boot.itb at sector 16384 - After reboot, the board no longer boots from SD or eMMC. - LED stays red. - No HDMI output. - No Ethernet link/activity. - RKDevTool on Windows shows "No Devices Found". Recovery attempts: - Tried the reset/recovery button behind the AV/P2 jack while connecting power and USB. - Tried both USB ports. - Tried RKDevTool v2.86 with Rockchip drivers installed. - No Loader or Maskrom device appears. - Tried UART pads marked GND/TX/RX using ESP32-C3 as bridge, but output is not readable. Question: Can anyone identify the eMMC Maskrom short point for this board? I need to force RK3528 into Maskrom mode so I can restore the first bootloader area of the eMMC. Attached photos: - Front of board - Back of board - Macro photo of the eMMC/storage chip area - Macro photo of test pads near the storage chip

I tried to downgrade to 6.18, unfortunately it froze during the process, which left the kernel and boot loader in an incomplete state. I have decided to re-image the whole thing with 6.18.x and for now it appears to be stable again.

Your previous post reminded me that the board supplier changed the BOM of the ODROID-N2+ by using a different SPI flash vendor. He hacked the support for it into his legacy firmware build without making any further note about it. It took some effort of reverse engineering to figure that out. Mainline hasn't picked up this additional driver activation to this day, but I still keep it in my builds anyway. In my latest build for the ODROID-HC4, I also included this driver to see if they might have gone about it in the same way. But your confirmation shows that this is probably not the case. Since I don't have an ODROId-HC4 with the behavior you described on hand, I can't analyze any further what the cause of it is. These days, I also mostly avoid devices with Amlogic SoCs because of their strict closed-source policy and lack of mainline support. And the board manufacturer isn't much more helpful on this point either. Devices powered by Rockchip are much more appealing objects. So you have to help yourself if you want to find a solution.

These most recent edge updates for K1 based boards are really great! With all this mainline additions of kernel 7.2 it's finally becoming a trustworthy clean alternative to outdated vendor images. At least for headless micro server usage everything is now already provided by mainline code (e.g. temperature and CPU frequency control are indispensable for serious work on this board!) Only WiFi and BT still don't work out of the box. I personally would prefer utilizing mainline brcmfmac drivers instead of fighting with all those unmaintainable bmcdhd patches. That's working for other boards with AP6256 modules as well.

OK I see: # dmesg | grep watchdog [ 0.462093] sunxi-wdt 1c20c90.watchdog: Watchdog enabled (timeout=16 sec, nowayout=0) That would mean study 'bark' or so, forgot most of it. I guess the board should restart if IPv6 LL connection is not there for more than 1 minute or so, but in relation to NAS makes then only sense if I move from NBD to iSCSI or so.

The forum returned 'nginx bad gateway' but I see draft of may message is still there. In the meantime, I did a simple 'apt update && apt full-upgrade && reboot' and now I have: U-Boot SPL 2024.01-armbian-2024.01-S866c-P7738-Hb9d3-Vf23c-Bb703-R448a (Jun 21 2025 - 02:53:13 +0000) Kernel: Linux 6.18.35-current-sunxi I might look into HW watchdog things, no clue if it is there and able to work, but used it successfully for Raspberrypi4 to workaround USB3 lockups/crashes.