maxsub

You mean the M.2 PCIe slot? My RV2 board has no oculink slot, only 2 PCIe slots, one for NVMe and one for SSD.

So this is built fully from source by patching the xunlong source onto the edge code base? Armbian-unofficial_26.02.0-trunk_Orangepirv2_trixie_edge_6.18.8_minimal+Orangepirv2_1.0.0_ubuntu_noble_desktop_gnome_linux6.6.63.img.xz

In your downloads site, does this one have the GPU hack? Armbian-unofficial_26.02.0-trunk_Orangepirv2_trixie_current_6.6.99_minimal.img. I can do some testing. Thank you.

I started a build of mesa-spacemit-kt from source on the RV2 and ran into only a couple of issues: ../src/amd/llvm/ac_llvm_helper.cpp:102:10: error: no matching function for call to ‘unwrap(LLVMOpaqueModule*&)’ 102 | unwrap(module)->setTargetTriple(TM->getTargetTriple().getTriple()); | ~~~~~~^~~~~~~~ In file included from /usr/include/llvm-19/llvm/IR/Type.h:18, from /usr/include/llvm-19/llvm/IR/DerivedTypes.h:23, from /usr/include/llvm-19/llvm/IR/InstrTypes.h:26, from /usr/include/llvm-19/llvm/Analysis/TargetLibraryInfo.h:14, from ../src/amd/llvm/ac_llvm_helper.cpp:8:

The second time I built it, the image deb worked and the bcmdhd dkms loaded. > dkms status bcmdhd-spacemit-sdio/101.10.591.84.37-4, 6.18.9-edge-spacemit, riscv64: installed

The bcmdhd dkms is getting built and the debs are being updated if I use SKIP_ARMBIAN_REPO=yes. [🐳|🔨] Building for 6.18.9-edge-spacemit [🐳|🔨] [🐳|🔨] Building initial module bcmdhd-spacemit-sdio/101.10.591.84.37-4 for 6.18.9-edge-spacemit [🐳|🔨] The kernel is built without module signing facility, modules won't be signed [🐳|🔨] [🐳|🔨] Building module(s)............................... done. [🐳|🔨] Installing /lib/modules/6.18.9-edge-spacemit/updates/dkms/bcmdhd.ko.xz [🐳|🔨] Installing /lib/modules/6.18.9-edge-spacemit/updates/dkms/dhd_static_buf.ko.xz [🐳|🔨] Running depmod.... done. [🐳|🔨] Processing triggers for man-db (2.13.1-1) ... When installing the image deb the previous dkms gets deleted. Preparing to unpack linux-image-edge-spacemit_26.02.0-trunk_riscv64__6.18.9-S3aec-D0000-P87b2-C50afHfe35-HK01ba-Vc222-B052a-R448a.deb ... Armbian 'linux-image-edge-spacemit' for '6.18.9-edge-spacemit': 'prerm' starting. dkms: removing module bcmdhd-spacemit-sdio/101.10.591.84.37-4 for kernel 6.18.9-edge-spacemit (riscv64) Module bcmdhd-spacemit-sdio/101.10.591.84.37-4 for kernel 6.18.9-edge-spacemit (riscv64): Before uninstall, this module version was ACTIVE on this kernel. Deleting /lib/modules/6.18.9-edge-spacemit/updates/dkms/bcmdhd.ko.xz Deleting /lib/modules/6.18.9-edge-spacemit/updates/dkms/dhd_static_buf.ko.xz ......... It gets built during install of the image deb: Autoinstall of module bcmdhd-spacemit-sdio/101.10.591.84.37-4 for kernel 6.18.9-edge-spacemit (riscv64) Building module(s).............................................. Installing /lib/modules/6.18.9-edge-spacemit/updates/dkms/bcmdhd.ko.xz Installing /lib/modules/6.18.9-edge-spacemit/updates/dkms/dhd_static_buf.ko.xz Running depmod.....

@sven-ola The dkms did not get installed by the build. I manually installed it and the wifi works. I am trying the build with SKIP_ARMBIAN_REPO now as I did not have that before.

Got it! But this DKMS should be loaded by the build: function post_family_tweaks_bsp__orangepirv2_wifi() { display_alert "$BOARD" "Force load bmcdhd wireless" "info" run_host_command_logged mkdir -pv "${destination}"/etc/modules-load.d run_host_command_logged echo "bcmdhd" > "${destination}"/etc/modules-load.d/${BOARD}.conf }

On RV2 with the edge branch WiFi does not work.

I have been testing the R2S extensively. The 1GbE is very stable, but the 2.5GbE drops frequently. I do not have debug logs yet but it is something I need to analyze. I also have a wireless AP running on the R2S. Performance is poor: 1GbE WAN connection, but I am getting only 20Mbps throughput but it is stable.

Thank you. Looking forward to all these contributions hitting the main branch so I can depend on the standard edge and current builds!

The patch brought the idle load average drastically down on the R2S from 2.0 to 0.18. However the temp remains at 57-58 deg C. orangepi@orangepir2s:~$ cat /sys/class/thermal/thermal_zone1/temp 56000 orangepi@orangepir2s:~$ htop orangepi@orangepir2s:~$ cat /sys/class/thermal/thermal_zone1/temp 57000

I patched that, am rebuilding and will post the R2S temps after that patch.

Is this specific patch that reverts the busy looping thread? https://github.com/armbian/build/commit/99ad08839eed42d647e716a3ed090ea318d01d2e

I compiled two different versions on the 8GB RV2: edge and current, and targeted two boards: R2S and RV2. Both compiles worked. Now I am testing the R2S as an access point with a USB wifi. The wifi adapter is working fine. The RV2 reports a stable temperature of 35C peaking at 47C when building. The R2S runs hot: 58C with a heatsink when doing nothing, peaking at 70C when doing .deb updates.

I have managed to compile this on two different platforms (Ubuntu and MacOS) for both boards: RV2 and R2S. My RV2 has 8GB and has a 64GB eMMC. I am hoping to be able to build it on the R2S natively. Will let you know how that goes. Looking forward to seeing this PR merged into armbian:main. I think the GPU will remain orphaned. OrangePi decided to use a unsupported GPU and it seems mostly a check box for them v/s them truly intending to support it.

I am so thrilled to be able to get this working. Donation to Armbian on the way.

I found the problem: [ 2299.766819] Out of memory: Killed process 43216 (pahole) total-vm:8235296kB, anon-rss:4667816kB, file-rss:0kB, shmem-rss:0kB, UID:0 pgtables:12912kB oom_score_adj:0

pahole is current. The build fails inside the docker-shell when manually invoked as well. Looks like the BTF invocation is getting killed (maybe needs more RAM)? 🐳|🔨] BTF .tmp_vmlinux1.btf.o [🐳|🔨] Killed root@c408bbb4ccde:/armbian# apt show pahole Package: pahole Version: 1.25-0ubuntu3 Priority: optional Section: universe/utils Source: dwarves Origin: Ubuntu Maintainer: Ubuntu Developers <ubuntu-devel-discuss@lists.ubuntu.com> Original-Maintainer: Thomas Girard <thomas.g.girard@free.fr> Bugs: https://bugs.launchpad.net/ubuntu/+filebug Installed-Size: 2012 kB Depends: libbpf1 (>= 1:1.0.0), libc6 (>= 2.38), libdw1t64 (>= 0.157), libelf1t64 (>= 0.142), zlib1g (>= 1:1.1.4) Breaks: dwarves (<< 1.22) Replaces: dwarves (<< 1.22) Homepage: https://git.kernel.org/pub/scm/devel/pahole/pahole.git Download-Size: 275 kB APT-Manual-Installed: no APT-Sources: http://ports.ubuntu.com/ubuntu-ports noble/universe arm64 Packages Description: set of advanced DWARF utilities

@sven-ola I have docker. The build BRANCH=current works but BRANCH=edge errors in the BTF section.

@sven-ola I did try the BRANCH=edge build and got an ELF error: 🐳|🔨] libbpf: failed to find '.BTF' ELF section in vmlinux.unstripped [🐳|🔨] FAILED: load BTF from vmlinux.unstripped: No data available [🐳|🔨] make[2]: *** [scripts/Makefile.vmlinux:72: vmlinux.unstripped] Error 255 [🐳|🔨] make[2]: *** Deleting file 'vmlinux.unstripped' [🐳|🔨] make[1]: *** [/armbian/cache/sources/linux-kernel-worktree/6.18__spacemit__riscv64/Makefile:1242: vmlinux] Error 2 [🐳|🔨] make: *** [Makefile:248: __sub-make] Error 2

This is fantastic work! Thank you so much. I was able to build from source. Tested it on RV2 and R2S and both work great. Your changes are patched on to the 6.6.99 kernel as far as I can tell. Are these going to be merged upstream and make their way into 6.17? Thanks again for your amazing work.