| 7 Apr 2022 |
 @janne.hess:helsinki-systems.de | go hard or go home | 19:22:15 |
| @janne.hess:helsinki-systems.de | (which I can‘t because the flights are expensive) | 19:22:24 |
 @elvishjerricco:matrix.org | I'm curious how hibernation was so easy | 19:23:49 |
| @janne.hess:helsinki-systems.de | just add a resume= parameter | 19:26:19 |
| @elvishjerricco:matrix.org | oh | 19:26:31 |
| @elvishjerricco:matrix.org | huh | 19:26:33 |
| @elvishjerricco:matrix.org | I thought that'd be a whole thing :P | 19:26:47 |
| @janne.hess:helsinki-systems.de | nope, the generator just handles that properly | 19:27:20 |
| @elvishjerricco:matrix.org | Well awesome. One more chunk of logic we get from upstream for free | 19:27:38 |
| @elvishjerricco:matrix.org | does it work correctly with LUKS root too? | 19:28:11 |
| @janne.hess:helsinki-systems.de | ah if you do the TimeoutStartSec=infinity thing for the cryptsetup unit, it now timeouts because the device that it has to decrypt times out :D | 19:28:26 |
| @elvishjerricco:matrix.org | d'oh | 19:28:42 |
| @janne.hess:helsinki-systems.de | In reply to @elvishjerricco:matrix.org
does it work correctly with LUKS root too? yes. I have luks -> lvm -> thin lvm -> xfs and my swap next to the thin volumes | 19:28:59 |
| @elvishjerricco:matrix.org | that's amazing | 19:29:14 |
| @janne.hess:helsinki-systems.de | I‘m surprised as well, yeah+ | 19:29:21 |
| @elvishjerricco:matrix.org | Janne Heß: Is the stuff you did today for udev/lvm/luks/handover today pushed anywhere? I'd like to play with it :) | 23:03:19 |
| @janne.hess:helsinki-systems.de | I can push it tomorrow | 23:04:46 |
| @janne.hess:helsinki-systems.de | I just met some spanish people I'm currently getting drunk with 👀 | 23:05:07 |
| @elvishjerricco:matrix.org | lol fair enough | 23:05:22 |
| @janne.hess:helsinki-systems.de | 🇪🇸 | 23:05:42 |
| 8 Apr 2022 |
| @janne.hess:helsinki-systems.de | In reply to @janne.hess:helsinki-systems.de
I can push it tomorrow I forgot to push it. Damn it | 18:18:02 |
| @janne.hess:helsinki-systems.de | After discussion with Arian, we will need a small systemd patch now | 18:18:17 |
| @elvishjerricco:matrix.org | Janne Heß: What for? | 18:28:51 |
| @janne.hess:helsinki-systems.de | So I don't have to write to /run/current-system below the mount | 18:31:06 |
| @elvishjerricco:matrix.org | what's the reason for just not unmounting /sysroot/run after nixos-activation? | 19:20:29 |
| @elvishjerricco:matrix.org | * what's the reason for not just not unmounting /sysroot/run after nixos-activation? | 19:20:42 |
| @janne.hess:helsinki-systems.de | We do that but we need the binary in /run/current-system during the switch | 19:26:26 |
| @elvishjerricco:matrix.org | I'm confused. Shouldn't that be there if you don't unmount /sysroot/run? | 19:27:27 |
 gkleen | I probably don't have the required context but isn't it kinda fragile to address anything below /run/current-system as oppose to under /nix/store? | 20:10:53 |
| @elvishjerricco:matrix.org | gkleen: These are special circumstances. In stage 1, we choose a system configuration to boot by looking at the init= kernel param. We chroot into the real root. We run the NixOS activation script, which places that configuration at /run/current-system. That's how we know for sure it has what we need. And for the specific feature we're after, we need the systemd binary for PID 1 to be built with knowledge of the absolute path it'll be used from, so we have to use something generic like /run/currentys-sytem/... | 20:13:37 |
