2024-04-10 13:41:23 <iggy> it looks like the upcoming hifive system with the P550 should support hypervisor extensions
2024-04-16 04:06:01 <k02h3rparasites> https://appealforassistance.notion.site/Appeal-for-Assistance-545bafc95efa4ed6ba3c1738fbbddd0d
2024-04-18 00:33:58 <iggy> ncopa: I saw you mention something about the 3.20 riscv64 builders. Is there anything anybody in here can help with? (probably not, but couldn't hurt to ask)
2024-04-18 00:35:01 <celie> I think there was some discussion about what happened to the rv64 builders in #alpine-infra
2024-04-18 00:36:55 <celie> You can see the logs on irclogs.a.o, that's what i mean
2024-04-18 00:48:29 <iggy> what happened with the pioneer boxes?
2024-04-18 00:52:18 <celie> I think the #alpine-infra logs will explain it better than i can, also it's probably been summarized in a toot in https://fosstodon.org/@ncopa
2024-04-18 00:53:15 <celie> Looking at that, i think in summary, it's "network config messed up"
2024-04-18 00:53:22 <iggy> sorry, I'll scroll back more. I just read the tail end
2024-04-18 00:54:10 <celie> and from what i gather, no one has physical access to the pioneer boxes, at least not for now
2024-04-18 00:54:47 <iggy> wonder if some sort of ikvm connected to them would help (maybe with some sort of power switch control)
2024-04-18 04:45:44 <ikke> iggy: there is a remotely controllable power switch, but that does not help if the configuration is wrong 
2024-04-18 07:26:01 <ncopa> morning
2024-04-18 07:26:23 <ncopa> what happened was: i was about to set tup the build-3-20-riscv64 machine in lxc
2024-04-18 07:26:48 <ncopa> but the lxc container did not start, because it could not attach to the network bridge
2024-04-18 07:26:58 <ncopa> so I had a look at the network bridge
2024-04-18 07:27:28 <ncopa> and it was not existing. the machine was connected via eth0 and we needed a bridge over eth0
2024-04-18 07:28:13 <ncopa> so I uncommented the br0 config in /etc/network/interfaces, which would set the static ip address on br0, and let eth0 be a part of the br0
2024-04-18 07:28:25 <ncopa> easy peasy, done that 100 times before
2024-04-18 07:29:21 <ncopa> started: `sleep 2m && reboot -f` in case i'd mess up and lose connectivity
2024-04-18 07:29:42 <ncopa> and then from other window: `ifdown eth0 && ifup br0`
2024-04-18 07:30:26 <ncopa> directly after i pressed enter, clandmeter told me that he had kernel issues when he tried to configure a bridge.
2024-04-18 07:30:36 <ncopa> which explains why the br0 config was commented out
2024-04-18 07:31:25 <ncopa> so after 2 min, the machine did reboot (i assume), but it will configure the br0 on bootup.
2024-04-18 07:32:24 <ncopa> clandmeter will have to be physically at the machine and clean up my mess. meanwhile we only have the other pioneer machine
2024-04-18 07:32:43 <ncopa> but I also spun up two instances in scaleway that we cuse for CI now
2024-04-18 18:34:36 <iggy> ikke: an ikvm with that would I guess? I don't know the exact configuration, so just spitballing here
2024-04-18 18:35:07 <iggy> I've also done serial b2b between remote hosts before to help with certain situations
2024-04-18 18:35:08 <ikke> iggy: yeah, anything that would allow oob access, but that has not been setup yet
2024-04-18 18:35:44 <ikke> clandmeter said he'd perhaps has an opportunity on monday to fix things
2024-04-18 18:43:47 <iggy> wonder how hard it would be to do some serial between them... it looks like they don't have standard serial ports, but they have jtag and possibly some serial pins on the board... you'd have to fish some wires out the back of the machine and connect to a usb serial on the other box
2024-04-18 18:44:11 <iggy> janky, but maybe useful
2024-04-18 18:46:51 <mps> does riscv64 builder works? I think to upgrade linux-starfive to 6.8.7
2024-04-18 18:47:24 <ikke> mps: yes, I think it should work
2024-04-18 18:47:31 <ikke> except it has a big backlog
2024-04-18 18:47:41 <mps> ikke: ok, thanks
2024-04-18 18:53:20 <mps> pushed, lets see when builder will finish
2024-04-22 11:48:32 <clandmeter> ncopa: https://lore.kernel.org/linux-riscv/20240223-tidings-shabby-607f086cb4d7@spud/
2024-04-22 12:01:35 <ncopa> thank you!
2024-04-26 09:40:00 <ncopa> clandmeter: some update on sophgo kernel. I got response from upstream, and they said we could re-apply 3 commits that was reverted in their tree (we need to check if we enable HIGHMEM though).
2024-04-26 09:41:07 <ncopa> I  saw that they seem to have a patch for a kernel boot arg to disable the vecotr extensions for the 6.1 branch
2024-04-26 09:41:59 <ncopa> i noticed that the 6.6 branch has been fairly active recently. but it lack the commit for boot arg to disable v extensions, so maybe it lacks more things?
2024-04-26 09:42:55 <ncopa> i wonder if we should give the updated kernel another go? do we want continue with 6.1 branch or do we dare test out the 6.6?
2024-04-27 10:22:21 <clandmeter> ncopa: i remember 6.6 only runs on 1 core
2024-04-27 10:22:34 <clandmeter> due to some memory limitation or so
2024-04-27 10:22:58 <clandmeter> and numa is also disabled in 6.6 iirc
2024-04-29 10:59:02 <clandmeter> ncopa: https://github.com/sophgo/linux-riscv/commit/e2697baecc1569e2c25d679f15c52068b124ac3e
2024-04-29 11:10:07 <clandmeter> ncopa: there also seems some activity in https://github.com/milkv-community/linux