01:11 <geist> as a side note does anyone have luck booting the current ubuntu riscv server images according to their wiki directions?

01:12 <geist> i always end up with either an EFI failure (as if uboot image didn't fully support something) or an issue loading the device tree

01:12 <geist> i'm thinking it's one of those versioning things where it's only officially tested againt some very specific version of qemu and/or opensbi and/or uboot

09:42 <xypron> HdkR: see https://lore.kernel.org/linux-riscv/20220420112408.155561-1-apatel@ventanamicro.com/ PATCH v3] RISC-V: Increase range and default value of NR_CPUS

09:45 <xypron> geist: Are you refering to the presinstalled image or the the server install image? The preinstalled images use u-boot-menu and create extlinux/extlinux.conf. Ther server install image uses EFI booting via GRUB. U-Boot does not support setting EFI variables from Linux. You will have to use the efidebug command to set the boot option to point to GRUB. Please, open an issue on launchpad.net if you problems persist.

09:46 <geist> aaaah that sounds like maybe what is missing

09:46 <geist> will give that a try tomorrow, thanks

09:56 <mps> yesterday I've got alpine fully working on visionfive, except sound

09:57 <mps> only couldn't get grub to load DTB

09:58 <mps> anyone know why stock grub (not from big distros) can't work with devicetree directive

12:30 <conchuod> mmind00: does the ethernet on a d1 work in uboot?

14:06 <mmind00> conchuod: yes

14:06 <mmind00> conchuod: give me a sec, I'll check if I added something

14:07 <HdkR> xypron: Nice. Very fresh patches. Nice to see that is the only change required

14:09 <mmind00> conchuod: nope, the only thing I added was support for pxe and that it loads my kernel-fit-image my boardfarm provides ... ethernet itself ran out of the box

14:20 <conchuod> :thinking: board farm? how many ya got?

14:20 <conchuod> and thanks

14:25 <mmind00> conchuod: of the d1 - one ... but the thing is also full of Rockchip boards ;-) ... and I don't like to have boards sitting on my table and be netbootable via tftp

14:32 <mmind00> conchuod: and if you're curious ... https://chemnitzer.linux-tage.de/2017/media/programm/papers/208_Linux-Kernel_nderungen_Konzepte_um_Testbarrieren_abzubauen_folien_kernel-testing.pdf ... pages 37 (hardware) + 43 (software) for impressions ;-)

14:33 <conchuod> ye i meant of "random" boards not just the d1

14:36 <conchuod> mmind00: is that bank of laptops yours?

14:37 <mmind00> conchuod: nope :-D ... that was a photo of the chrome-os testlab at Google, that I found

14:38 <conchuod> Is it just the bits after "my boardfarm" that are yours then?

14:38 <mmind00> i.e. the Ikea Ivar thingy is mine, the others were examples for other boardfarms ;-)

14:40 <conchuod> i need to get an ethernet controlled pdu

15:33 <geertu> mmind00: The one on p31 looks quite different nowadays

15:35 <mmind00> geertu: hehe ... though I found the lego-usage really cool back in the day, and do still remember your talk about herding boards ;-)

15:37 <geertu> mmind00: next iteration at https://drive.google.com/file/d/1MVuOf07tBOYKacT3alByRSVYcCgjCX9B/view

15:38 <geertu> That's from 4.5 years ago, so in the mean time more boards were added

15:41 <geertu> and the LEGO power distribution box has been replaced by a DIN rail with Phoenix Contact terminal blocks

15:42 <geertu> I still use LEGO to hold PCbs, though. Haven't found time and a better solution to improve.

15:43 <mmind00> :-) ... it's pretty cool to see all the different ways people organize

15:43 <geertu> Indeed. NIH rules ;-)

16:02 <conchuod> does the pc powersupply power everything there?

16:27 <geertu> conchuod: It does for me. Board farm control parts (2x boneblack, acme, 2 x teensy3-bcu2, ...) are powered from 5V standby-power.

16:38 <conchuod> teensys :)

16:40 <conchuod> I really like Paul and his work.

17:02 <geertu> conchuod: BTW, are there plans to upstream the pac1394 driver?

17:02 <geertu> pac1934

17:08 <conchuod> thats a complicated question geertu

17:09 <conchuod> The answer is probably yes but not right now

17:12 <conchuod> (trying to figure out how I can expand on that)

17:13 <conchuod> The team that wrote it hasn't upstreamed a driver before, we've kicked off some review internally with the goal of getting it cleaned up & upstreamed