Daily Ubuntu Kernel Builds - Now With 100% More APT Repo!

I’m on the latest release fw from Starfive (Release VisionFive2 Software v2.11.5 · starfive-tech/VisionFive2 · GitHub)

1 Like

Please ensure your sdcard is unmounted before doing any of the wiping/imaging to the sdcard.

umount /run/media/davidm/rootpart
umount /run/media/davidm/root
umount /run/media/davidm/305E-3140

Wipe your sdcard, then image it with ubuntu image.

wipefs --all /dev/sdb
xz -dk ubuntu-23.04-preinstalled-server-riscv64+visionfive2.img.xz
dd if=./ubuntu-23.04-preinstalled-server-riscv64+visionfive2.img of=/dev/sdb bs=4M status=progress conv=fdatasync

Unless you are giving the SDcard away and worried about leaking data, the wipe command only adds wear to your SDcard and is entirely unnecessary. Also, you can just decompress directly to the SDcard. This is Linux, not SysV: xzcat ubuntu-23.04-preinstalled-server-riscv64+visionfive2.img.xz > /dev/sdb


Reboot is inconsistent for me. Sometimes it hangs after a few lines of output, other times it’s fine. Not sure if I’m on the latest firmware but can check tomorrow.

:frowning: Indeed, even with the latest firmware my board just failed to reboot. It hung after printing the below on the serial console:

[ESC[0;32m  OK  ESC[0m] Stopped ESC[0;1;39mmotd-news.timerESC[0m - Message of the Day.
[ESC[0;32m  OK  ESC[0m] Stopped ESC[0;1;39msystemd-tmpfiles-c_ Cleanup of Temporary Directories.
[ESC[0;32m  OK  ESC[0m] Stopped ESC[0;1;39mua-timer.timerESC[0m_e Timer for running re

On latest firmware v2.11.5 u-boot and payload, booting Ubuntu 23.04 with Kernel #72.

a. does anyone else get anything from the video out over hdmi?
b. anyone had it booting with an nvme connected?
c. does the second led go green for anyone else (ubuntu 6.2 kernel and 6.4.0 #72)
d. is iperf3 performance much lower (with more collisions) on ubuntu vs debian 5.15?

Sorry still new, so feeling like it could all be me causing the issues still.

Having some trouble installing newer kernels. I’m downloading from the release section manually and then running a dpkg -i *.deb. after rebooting the board won’t come back up. After following the guide posted here, https://agreenbhm.github.io/linux/repo/riscv64 12 / results in a repo that cannot be found. Any help would be greatly appreciated.

Not sure where the “12” before the “/” in the repo URL is coming from, but remove it. It should read:

https://agreenbhm.github.io/linux/repo/riscv64 /

Edit, I see. Stupid formatting from this forum software is adding that “12” . When you copy it it gets copied too. Remove that and try running apt update.

1 Like

Thank you very much! Yeah I figured it was a formatting error as well, and it isnt present in my file. I’ll give it another go tonight and report back asap. Thank you for the help! Greatly appreciate the work you have been doing.

Please report back the exact error you get from apt if it still gives you issues when you test again. Hopefully it’ll work but if not it shouldn’t be too difficult to diagnose.


Sorry for the delayed response. I removed my /custom_kernel.list file, and then recreated it to make sure there wasn’t any issues. After recreating the file, and running linux-image-rolling linux-headers-rolling it worked as intended. Kernel was updated successfully. Thank you again @agreenbhm! Greatly appreciate the help!

Excellent, glad it’s working now!


Thank you again for your help!


Your Linux kernel “6.4.0-rc3+ #90” is behaving correctly for me. Thank you. I got wired ethernet and have been successfully building some existing projects with rust like salvo web server.
Linux vf2-UbuSrv-YOW 6.4.0-rc3+ #90 SMP Thu May 25 00:40:09 UTC 2023 riscv64 riscv64 riscv64 GNU/Linux

I would like to start playing with the gpu driver, what is the best way to consider tweaking this to simply run it in ubuntu server?

Has anybody attempted this? I wish there was more visibility as to the progress of this work and whether they wanted individuals to run and report their tests.

1 Like