27 Jun 2023 |
@linus:schreibt.jetzt | but if it's a yubikey or something there's no kernel-side driver for it | 14:52:50 |
Fruity Passions | In reply to @linus:schreibt.jetzt what kind of USB device? Plain ol USB 2.0. It's model comes up as "FlashDisk" and it's very old (I think it's only a gig in size). | 15:26:03 |
Fruity Passions | In reply to @oddlama:matrix.org my guess is on usb_storage Already loaded by hardware-configuration but thanks for the guess :) | 15:27:10 |
Fruity Passions | In reply to @elvishjerricco:matrix.org If you want a brute force solution, you could probably just import the all-hardware.nix module out of nixpkgs I could try that? is it possible to, if it works, narrow it down afterwards? | 15:28:11 |
Fruity Passions | In reply to @linus:schreibt.jetzt what kind of USB device? * Plain ol USB 2.0. It's model comes up as "FlashDisk" and it's very old (I think it's only a gig in size). (note: i have other flash drives connected too and they aren't detected either (some significantly more modern)) | 15:29:50 |
ElvishJerricco | Yea I guess you could copy the code out of the module and then bisect it by commenting out one half at a time | 15:29:57 |
oddlama | In reply to @passion-fruit:matrix.org I could try that? is it possible to, if it works, narrow it down afterwards? Does the stick show up in stage2? If so, just check lsmod for relevant modules. | 15:37:52 |
oddlama | If the device doesn't show up at all, you might be missing xhci_hcd xhci_pci or something like that | 15:38:51 |
Fruity Passions | In reply to @oddlama:matrix.org If the device doesn't show up at all, you might be missing xhci_hcd xhci_pci or something like that hmm i'm missing xhci_hcd but not the pci variant. could that perhaps be it? | 15:49:18 |
oddlama | Depends on which controller is in use for the usb port | 15:52:33 |
oddlama | sometimes xhci_hcd is pulled in as a dependencies, so not sure if that one is the missing link | 15:53:26 |
oddlama | Try lsusb -t in stage 2 to see which driver is in use | 15:54:13 |
Fruity Passions | that did it! Thank you so much :) - i checked lsusb -t after booting with the live iso and the usb controllers were using hcd. added xhci_hcd and now it works without issue | 16:24:57 |
29 Jun 2023 |
@lily:lily.flowers | So why is boot.initrd.services.lvm.enable a hidden option if it is required to enable it for systemd-stage-1 whenever you have a LVM needed for root | 20:10:35 |
@lily:lily.flowers | (I, uh, spent longer than I wish I had the last few weeks banging my head against that) | 20:10:52 |
@lily:lily.flowers | (until I found that option) | 20:11:05 |
ElvishJerricco | I guess we forgot to unhide it in the unhide PR? | 20:11:25 |
@lily:lily.flowers | Maybe. I didn't think it was a sd-stage-1 only though, but LVM root in scripted stage 1 seems to work just fine (maybe it manually runs vgchange instead of relying on udev rules idk) | 20:12:16 |
@lily:lily.flowers | * Maybe. I didn't think it was a sd-stage-1 only option though, but LVM root in scripted stage 1 seems to work just fine (maybe it manually runs vgchange instead of relying on udev rules idk) | 20:12:22 |
ElvishJerricco | I thought it was only for sd-initrd | 20:12:45 |
@lily:lily.flowers | Yep scripted stage-1 absolutely does | 20:12:45 |
@lily:lily.flowers | Yeah I guess so | 20:12:52 |
@lily:lily.flowers | Yeah git blame says it was added in the initial sd-initrd commit | 20:13:30 |
@lily:lily.flowers | Gonna make a PR to unhide that then | 20:13:36 |
@lily:lily.flowers | I should probably backport that PR too then, right? | 20:17:14 |
ElvishJerricco | I guess? I tend not to think about stable when working on sd-initrd. Once we remove the "experimental" denotation in the docs, I'll have to change my ways though :P | 20:17:55 |
@lily:lily.flowers | I opened https://github.com/NixOS/nixpkgs/pull/240601, but I actually found a lot more options all over the place left hidden that I'm fixing now. So I'm gonna push those to that PR too and rename it in a minute | 20:25:00 |
ElvishJerricco | interesting... | 20:25:58 |
ElvishJerricco | I thought I had built the options JSON thingy with hidden options included and scanned over them all a while ago | 20:26:15 |
@lily:lily.flowers | Was boot.initrd.systemd.users and stuff specifically supposed to be left hidden ElvishJerricco? | 20:26:19 |