| 1 Jun 2023 |
 ElvishJerricco | is that normal for live cd images? | 02:36:37 |
| ElvishJerricco | because I hate it | 02:36:40 |
 @janne.hess:helsinki-systems.de | Since it's likely that somebody has already tried: Does anyone know if there's a systemd-cryptenroll way to use fido2+password? In a way where I need both to unlock? | 09:56:49 |
| @janne.hess:helsinki-systems.de | ah looks like it's not possible: https://github.com/systemd/systemd/issues/21088 | 10:29:02 |
 @lily:lily.flowers | In reply to @elvishjerricco:matrix.org
is that normal for live cd images? That specifically I don't think is normal. We are a tad special there. But doing special things for ISO images is (unfortunately) normal | 10:35:53 |
 Arian | we missed announcing systemd-initrd in the NixOS 23.05 release notes! | 11:55:00 |
| Arian | do we still want to add it? | 11:55:04 |
 @hexa:lossy.network | wow, yeah! | 11:56:08 |
| ElvishJerricco | Arian: my intention is that systemd initrd is still considered experimental in 23.05, and that it will (hopefully) reach stability for 23.11, with a slight possibility of becoming default in 24.05 | 11:57:37 |
| Arian | We could announce it as experimental? :D | 11:57:56 |
| ElvishJerricco | If that warrants an announcement in 23.05, then cool :P | 11:57:57 |
| Arian | I think so | 11:58:02 |
| ElvishJerricco | In reply to @janne.hess:helsinki-systems.de
Since it's likely that somebody has already tried: Does anyone know if there's a systemd-cryptenroll way to use fido2+password? In a way where I need both to unlock? So the only reason it supports a passphrase for the tpm is because it can literally pass the passphrase to the tpm and let it control the security | 12:07:52 |
|  @raphi:tapesoftware.net changed their display name from raphi to raphi (element unread channel fix when). | 13:03:16 |
 lukegb (he/him) | Can you set a PIN on a FIDO2 key you use for systemd-cryptenroll or does it not support that | 14:27:54 |
| lukegb (he/him) | Admittedly that's *for the entire authenticator* and not just that credential but still | 14:29:16 |
| ElvishJerricco | lukegb (he/him): According to man systemd-cryptenroll, there's --fido2-with-client-pin=BOOL | 14:29:23 |
| 2 Jun 2023 |
 Charles ⚡️ | In reply to @charles:computer.surgery
it just kinda started happening well, it just kinda stopped happening | 20:05:57 |
| Charles ⚡️ | so that's cool i guess | 20:06:00 |
|  @ckie:ckie.dev changed their display name from ckie (they/them; limited keyboard usage, voice preferred) to ckie (they/them). | 22:15:27 |
| 4 Jun 2023 |
|  @mberndt:matrix.org left the room. | 11:56:53 |
| 5 Jun 2023 |
 @nikstur:matrix.org | I just came across something curious: I'm currently building a custom kernel for an Nvidia BlueField2 and when I try to boot it with the systemd initrd, it fails to load the overlay module, giving me this error: overlay: module has no symbols (stripped?). When I use the legacy initrd, it is able to load the module. Could this happen because of the way makeInitrdNG works? | 12:25:12 |
| ElvishJerricco |
Could this happen because of the way makeInitrdNG works?
That would be pretty surprising, but I have no idea
| 12:26:14 |
| ElvishJerricco | we copy the whole modulesClosure derivation in | 12:26:23 |
| @janne.hess:helsinki-systems.de | In reply to @nikstur:matrix.org
I just came across something curious: I'm currently building a custom kernel for an Nvidia BlueField2 and when I try to boot it with the systemd initrd, it fails to load the overlay module, giving me this error: overlay: module has no symbols (stripped?). When I use the legacy initrd, it is able to load the module. Could this happen because of the way makeInitrdNG works? we have the same issue on rpi4 | 12:26:30 |
| @janne.hess:helsinki-systems.de | solution is to disable stripping | 12:26:35 |
| @nikstur:matrix.org | Is there an issue for that already? | 12:26:45 |
| @nikstur:matrix.org | I guess its an AARCH64 thing then | 12:26:53 |
| @janne.hess:helsinki-systems.de | boot.initrd.systemd.strip | 12:26:55 |
| @janne.hess:helsinki-systems.de | I don't think so, we thought it was a random oddity with our kernel | 12:27:05 |
