All Activity

I have a Radxa Zero 3E and was facing the same issue. Since RK3566 is a stripped down version of RK3588. I figured the quirks on RK3588 might be able to apply to RK3566 as well. The issue seems to be fixed now after I implemented these quirks taken from rk3588-base.dtsi on top of rk356x-base.dtsi. tested on Linux 6.17.9 and 6.18.0-rc7(Armbian linux-edge-rockchip64) usb_host1_xhci: usb@fd000000 { compatible = "rockchip,rk3568-dwc3", "snps,dwc3"; reg = <0x0 0xfd000000 0x0 0x400000>; interrupts = <GIC_SPI 170 IRQ_TYPE_LEVEL_HIGH>; clocks = <&cru CLK_USB3OTG1_REF>, <&cru CLK_USB3OTG1_SUSPEND>, <&cru ACLK_USB3OTG1>; clock-names = "ref_clk", "suspend_clk", "bus_clk"; dr_mode = "host"; phys = <&usb2phy0_host>, <&combphy1 PHY_TYPE_USB3>; phy-names = "usb2-phy", "usb3-phy"; phy_type = "utmi_wide"; power-domains = <&power RK3568_PD_PIPE>; resets = <&cru SRST_USB3OTG1>; snps,dis_u2_susphy_quirk; + snps,dis_enblslpm_quirk; + snps,dis-u2-freeclk-exists-quirk; + snps,dis-del-phy-power-chg-quirk; + snps,dis-tx-ipgap-linecheck-quirk; status = "disabled"; }; If anyone else has other RK356x devices please try it out and I think it would be nice if someone could submit a patch to mainline kernel dts.

@Gavin Munday Did you compile a desktop or server image? Did you write the image onto a reliable sdcard? Remember there’s no framebuffer support. Server images need a usb to serial device attached to the UART. Having a usb serial device attached to the UART helps with these types of problems. https://docs.radxa.com/en/cubie/a7a/system-config/uart_debug Also, first boot might take awhile to boot. For desktop images wait at least 5 minutes for a login screen.

sadly my a7a doesnt doesnt boot with the image that outputs from the build

I don't understand then, what is the problem?

You don’t actually need a static route on PC1. Since PC1 already has a default gateway (192.168.100.254), anything outside 192.168.100.0/24 is automatically sent there. So PC1 isn’t the source of the routing problem. The real issue is on the 10.10.10.x side: Device1 has no gateway. Without a gateway, it can receive packets coming from the Linux box, but it has no way to send replies back to a different subnet. That’s why the communication fails. The fix is simply to give Device1 a valid gateway pointing to the Linux box. Device1: IP: 10.10.10.2 Mask: 255.255.255.0 Gateway: 10.10.10.1 On the Linux box: sudo sysctl -w net.ipv4.ip_forward=1 This is enough to route between NIC1 and NIC2. If Device1 cannot be configured manually, the Linux box can offer a small DHCP service on NIC2. A minimal dnsmasq example: interface=enp4s0 dhcp-range=10.10.10.50,10.10.10.150,255.255.255.0,12h dhcp-option=3,10.10.10.1 This simply ensures Device1 learns “send everything through 10.10.10.1”. Once Device1 has a gateway, the Linux box will forward traffic correctly and PC1 will reach it without needing any static route at all. Useful resources: https://wiki.archlinux.org/title/Router https://askubuntu.com/questions/205017/how-to-define-the-range-of-ips-using-dhcp https://pingmynetwork.com/network/ccna-200-301/dhcp-dynamic-host-configuration-protocol

I don't think so, then my Virtual Machine run would fail the same way as you describe, but it doesn't, it works fine. Of course it is the meson64 kernel running on top of KVM/virt machine, the RTC for example is emulated, while for your real N2+ is a specific Amlogic SoC kernel driver I assume, so there something could be wrong. Or what about clocking tree, there are many failures possible. I know for rockchip64 there were 2 RTC modules, at some point in time 1 was chosen over the other, so a change, could also introduce issues. You could browse kernel changes/patches And also what about an unknown bad/corrupted diskblock or so. I normally use Btrfs for rootfs, so easy to check for and detect corruption. So maybe re-image again. I am still not sure if I used the same image as you, but up to you to track that. You can do the same thing as I did to see if it is maybe a specific Amlogic driver issue; Run that image as VM on the same N2+, you need a network bridge with main eth port or else use NAT. Or you make it first efi bootable, then it can run easier in virt-manager GUI. N2+ is fast enough to run VMs, I even do that on ROCK3A (2GB RAM, 4x Cortex-A55). Or use standard Debian Trixie kernel on real N2+, not sure if that works good enough. Or remove/disable RTC kernel module. Or also remove fake-hwclock stuff if you haven't already done that. I did not do it for this VM run, but when real RTC+battery, I would remove or disable it.

@Gavin Mundayhttps://github.com/NickAlilovic/build/blob/Radxa-A7A/config/boards/radxa-cubie-a7z.csc

found it wasnt using correct tree..;p

i have, board not there unless it using some other name?

Try to select CSC/EOS/TVB or use EXPERT=yes

esalarm: personally, I tear out systemd-resolved and netplan, and just use /etc/resolv.conf. (and chattr +i) In my LAN the gateway is designated DNS resolver running Unbound, so my situation is more normal. But that's not the reason systemd-timesyncd and chrony are not working. Neither one can set the time properly, even as properly configured. Using tcpdump I see the 123/udp request going out to the proper machine, and the response coming back... but it has no effect on Armbian. No firewall blocks. Time stays on whatever I set it to, and counts from there, always a few minutes behind or ahead. There is something fundamentally wrong with current Armbian.

when i do ./compile.sh i cant see the board in the list???

Something wrong with xorg or x11 session on cinnamon desktop for my boxes, i can get into desktop when i choose experimental wayland session but that is not a good way to start. So, i switch to Mate and i like it

Cinnamon desktop environment is bugged for me, both my a95x r2 (amlogic s905w) and a95x f3 slim (amlogic s905x3) showing no display, blank screen like no input after loging in. My suggestion for desktop ubuntu based varian, better change it to Mate desktop. The armbian desktop cinnamon varian based on ubuntu is now work for me after switching to mate desktop

can this SoC boot from a USB drive? i've tried it multiple times and it didnt work.

moved to staging

Orange Pi 5 pro,whith Armbian_community_26.2.0-trunk.7_Orangepi5pro_noble_vendor_6.1.115_gnome_desktop 。 when I try to start Eclipse/idea/vscode or other JavaApp , java processes crash and gets core dump every time, tested on Jdk 17,21 25 ,both openjdk and oracleJDK。 It seems there is a compatibility issue between the kernel and the JDK. How can I fix it? Or I have to wait kenel update? some error like Current thread (0x0000ffffa802b9b0): JavaThread "main" [_thread_in_vm, id=6202, stack(0x0000ffffaeb62000,0x0000ffffaed60000) (2040K)] Stack: [0x0000ffffaeb62000,0x0000ffffaed60000], sp=0x0000ffffaed5ce40, free space=2027k Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code) V [libjvm.so+0xddc444] Symbol::as_C_string() const+0x14 V [libjvm.so+0xbef608] Method::print_external_name(outputStream*, Klass*, Symbol*, Symbol*)+0x40 V [libjvm.so+0xb07e7c] LinkResolver::resolve_method(LinkInfo const&, Bytecodes::Code, JavaThread*)+0x154 V [libjvm.so+0xb0a85c] LinkResolver::linktime_resolve_virtual_method(LinkInfo const&, JavaThread*)+0x3c V [libjvm.so+0xb0b204] LinkResolver::resolve_virtual_call(CallInfo&, Handle, Klass*, LinkInfo const&, bool, JavaThread*)+0x3c V [libjvm.so+0xb0b3a0] LinkResolver::resolve_invokevirtual(CallInfo&, Handle, constantPoolHandle const&, int, JavaThread*)+0xd4 V [libjvm.so+0x887cd4] InterpreterRuntime::resolve_invoke(JavaThread*, Bytecodes::Code)+0x250 V [libjvm.so+0x887f80] InterpreterRuntime::resolve_from_cache(JavaThread*, Bytecodes::Code)+0x3c j java.util.AbstractSet.removeAll(Ljava/util/Collection;)Z+8 java.base@25.0.1 j org.eclipse.osgi.container.ModuleResolver$ResolveProcess.resolveRevisions(Ljava/util/List;ZLorg/eclipse/osgi/container/ModuleResolver$ResolveProcess$ResolveLogger;Ljava/util/Map;)V+424 j org.eclipse.osgi.container.ModuleResolver$ResolveProcess.resolveRevisionsInBatch(Ljava/util/Collection;ZLorg/eclipse/osgi/container/ModuleResolver$ResolveProcess$ResolveLogger;Ljava/util/Map;)V+253 j org.eclipse.osgi.container.ModuleResolver$ResolveProcess.resolve()Lorg/eclipse/osgi/container/ModuleResolutionReport;+307 j org.eclipse.osgi.container.ModuleResolver.resolveDelta(Ljava/util/Collection;ZLjava/util/Collection;Ljava/util/Map;Lorg/eclipse/osgi/container/ModuleDatabase;)Lorg/eclipse/osgi/container/ModuleResolutionReport;+19 j org.eclipse.osgi.container.ModuleContainer.resolveAndApply(Ljava/util/Collection;ZZLorg/eclipse/osgi/container/ModuleContainer$ResolutionLock$Permits;)Lorg/eclipse/osgi/report/resolution/ResolutionReport;+256 j org.eclipse.osgi.container.ModuleContainer.resolve(Ljava/util/Collection;ZZ)Lorg/eclipse/osgi/report/resolution/ResolutionReport;+63 j org.eclipse.osgi.container.ModuleContainer.refresh(Ljava/util/Collection;)Lorg/eclipse/osgi/report/resolution/ResolutionReport;+34 j org.eclipse.osgi.storage.Storage.checkSystemBundle([Ljava/lang/String;)V+242 j org.eclipse.osgi.storage.Storage.createStorage(Lorg/eclipse/osgi/internal/framework/EquinoxContainer;)Lorg/eclipse/osgi/storage/Storage;+17 j org.eclipse.osgi.internal.framework.EquinoxContainer.<init>(Ljava/util/Map;Lorg/osgi/framework/connect/ModuleConnector;)V+146 j org.eclipse.osgi.launch.Equinox.<init>(Ljava/util/Map;Lorg/osgi/framework/connect/ModuleConnector;)V+10 j org.eclipse.osgi.launch.Equinox.<init>(Ljava/util/Map;)V+3

Good day, @jock the link of the Multitool is not valid anymore. Please help where I can download from? Thank you for your efford.

I also tried the latest image (https://github.com/armbian/community/releases/download/26.2.0-trunk.22/Armbian_community_26.2.0-trunk.22_Rock-3c_trixie_current_6.12.59_minimal.img.xz) and it works fine (with mainline u-boot). I think I may have experienced problems using radxa's software to erase the spi flash memory, but writing an image fine would always work. So you may be better off writing Radxa's u-boot image to the spi flash memory, as per their instructions (and to get rid of the Android image). You can subsequently manage the spi flash memory either directly from u-boot (by interrupting start-up) or from the linux terminal. See comments by @amazingfate in this post for instruction on how to write Armbian u-boot image to mtdblock0 from a linux terminal.

I went out and bought a waveshare clone 3.5 inch display, to try and confirm I haven't spent the last few weeks trying to do the impossible with broken GPIO pins or something - and got a partial success eventually, using fbtft, displaying basic console and plymouth boot logs. It has poor fps like you said - which I expected, its the reason I tried to use the tiny DRM driver from the beginning, but now I know the rockpi and DTS is functional. The same DTS also shows a broken, purple / blue output of bars on my red ili9488 display only if I adjust the bus width to 16 instead of 8 (0x10, instead of 0x08), otherwise it remains a blank white screen. A bus width of 18 and 24 also only gave me a white screen, so 16 might be correct. I've been unable to confirm touch works, since I only have a basic console and I can't figure out how to start LXDE on a framebuffer device, but I do see messages insinuating it initializes under dmesg | grep spi. I could probably just query the GPIO interrupt to find out if touch is working. Not that important right now. I did realize that the big "fixups" section at the bottom is clearly used to replace the empty 0xfffffff targets throughout. I'm unsure why the DTS was written this way, but it works so I am not going to touch it (changing the targets and switching from hex to decimal values broke the DTS for some reason.) Unfortunately, a simple drop in replacement of fb_ili9486 with panel_mipi_dbi did not make either display work, even with the added arguments you use in your latest DTS. Clearly I have much more work to do. BUT I've finally confirmed the DTS works, and all GPIO is claimed and SPI is initialized for both displays (even if the red LCD is showing gibberish), so now I need to nail down the subtle differences with the displays like the bus and reg width and find the right combo of spi clock speed, reg bus widths, and buffer length to get it all to work with fbtft (if only aliexpress sellers put up datasheets or example code!). Then I can follow the second half of your journey to hopefully translate a working fbtft setup to a panel_mipi_dbi setup. Maybe get backlight PWM working. All this still on edge kernel 6.18, on a rockpi 4b plus.

I have the NanoPi-R6C and that came out of the box with an OpenWRT variant on eMMC. Indeed many partitions, but I think I managed to get it running somehow with Debian/Armbian rootfs on some higher partition number, I don't remember exactly. As long it can load boot.scr. Indeed eMMC is number 2, I guess they avoid potential confusion that is sometimes there with SBCs and various OSses swapping numbers for SD-card and eMMC. I had that once, no data loss but costed a lot of time. It is a bit similar to that on 6-SATA PC's, Debian might name the OS dev (SSD) /dev/sdb and a large HDD /dev/sda while Opensuse names OS always /dev/sda. So what do we do with 2-MMC SBC's ? I see my HP x86-64 dualrole computer/tablet uses /dev/mmcblk1 (its eMMC) for OS (Linuxes at least, Windows10 I don't remenber). The SD-card slot I have normally empty but uses /dev/mmcblk0 if card is inserted. It has UEFI and SecureBoot and TPM, so no choice actually like there is with opensource U-Boot. For NanoPi-R6C I have written Tianocore EDK2-UEFI to eMMC and appears as /dev/mmcblk1. I do not use the rest of the 32GB storage, but the UEFI binary comes with 1 partition, so easy to use the rest of the 32GB, but currently I need the much larger and faster NVME. And for emergency, if I insert an SD-card with newly generated rolling/edge image or so, it boots from that and is then /dev/mmcblk0. So this is IMO the preferred numbering, but I already had created some shell script code that can identify the boot partition and the root partition, but it is not too generic, only works if certain mounts are there which is what I have as I focus on EFI bootable and Btrfs with subvolumes ('nested' and default). So I would at least use 0 for SD. 1 for eMMC is fine. For the LEDs, on my NanoPi-R6C, I actually don't know, I never look at them. I think it is not correct with my installation (in-place upgraded and also other Linuxes), but I guess correct with current downloadable images. At least they were correct with the FriendlyElec OpenWRT that was pre-installed. I still can't get maskROM mode to work, so I am a bit paranoid on messing too much with data in the Rockchip loader area. I am not sure what happens if I put rubbish data or broken U-Boot there, can I still start from SD-card is the question. I might need to apply a fixed voltage between 9-20V as powersupply, but need to look in PCB schematics again first. For the Zero2, I see removable eMMC, so no risk or fear for 'hard-bricking'. Although I would check if maskROM works. For network, I assume it is the on-SoC port, so fixed off-the-shelf, so end0 is fine. For optional WLAN, I am not sure, likely some PCIe/HWtree named string, I have seen multiple altnames for 'wlo1', at least 'wlp1s0' or 'wlx<MACaddress>'. systemd-networkd should be able to handle those, NM can do all anyway.

Hi everyone, how's it going? I'm new to the forum and I'm having trouble getting my Chinese T95 mini TV box to run Linux. I want to switch it to a basic server, but I'm having problems booting it from the SD card. I'm using the sunxi-fel mode method in Linux. I tested Armbian_23.8.1_Orangepipc_jammy_current_6.1.47_minimal.img and others like Armbian_23.8.1_Orangepipc_jammy_current_6.1.47.img. I used the .bin file generated on this forum. Boot it using the method `sudo sunxi-fel -v -p uboot u-boot-sunxi-with-spl.bin` in the terminal. TV BOX Model: T95mini | R69 - EMCP V2.0 - H3 Allwinner. The screenshots below show the screen freezing. I'm a total noob, guys, if you could please explain the procedures, indicate what needs to be done to make it work. Thanks in advance libretech_all_h3_cc_h3_mem_336 - u-boot-sunxi-with-spl.bin

Hi everyone, wanted to share what we found attempting Armbian support for Zero2. Zero2 utilises an Android-style partition layout from FriendlyElec, which I found to be fundamentally incompatible with Armbian's partition scheme. It seems to use an Android layout (boot at partition 4, rootfs at partition 8, etc.) vs Armbian's standard GPT layout. U-Boot is configured to look for partitions at specific Android-style locations Zero2 assigns SD=mmcblk0, eMMC=mmcblk2 (opposite of typical RK3528 boards) What I think is needed: Adapt hinlink_rk3528_defconfig for Android partition layout Modify partition table generation in the Armbian build system Update bootloader scripts to match partition locations Handle the inverted mmcblk device assignments LED mapping adjustments (Zero2 uses 2-LED system: sys_led, user_led vs typical 3-LED RK3528 boards) Network interface detection (single ethernet vs dual on R3S/hinlink) The device tree is already merged (rk3528-nanopi-zero2.dtb - thanks Jonas!), so that part is straightforward. It's specifically the bootloader/partition integration that's the blocker. Would love to hear thoughts. For now as a workaround, we've stayed with FriendlyElec's Ubuntu base to keep our project moving.

Glad I fell on this post Igor, Thank you for the share.. (download was indeed really fast!) I managed to install without any issues. (been using buster for way too long) Cheers, Again thanks in advance!

I already see what the issue is: I have configured things in such a way that strange/new machines get served differently then known/existing ones. In addition, the known/existing ones need an extra setting in case of systemd-resolved. This makes a quick test with some random VM like this appear doing things wrong, but is actually expected. If I purge --autoremove netplan.io and put a symlink ln -sf /usr/lib/systemd/network/89-ethernet.network.example /etc/systemd/network/89-ethernet.network, there is working network again after reboot or restart service.