If you set `vt=none` in the configuration file things should work with 
`CONFIG_VT=n`.

It just so happened to have been discussed on IRC a moment ago.

Pushed to master, so it'll be in the next release when I get around to it.

Your configuration is not using agreety, it's using greetd and tuigreet. 
Neither of these know about displa ydevices. However, you can use a udev 
rule and a service dependency to make the greetd service itself wait for 
a particular device to start.

The udev rule would look something like this:

         ACTION=="add", SUBSYSTEM=="drm", KERNEL=="card2", TAG+="systemd"

The greetd service dependency looks like this:

         After=systemd-user-sessions.service plymouth-quit-wait.service 
dev-dri-card2.device

The issue is that pam_systemd_loadkey.so, like pam_gdm.so is that they 
try to auto-fill the password when authentication is attempted. We are 
not trying authentication, because PAM could ask any number of questions 
that we don't want during auto-login.

Supporting these modules would require greetd to do a dummy 
authentication attempt with all null answers, bailing auto-login if any 
question got asked. Not impossible, but not something we currently do.

I happen to have thrown together 
https://git.sr.ht/~kennylevinsen/pam_fde_boot_pw which aims to do the 
job without relying on dummy authentication. Completely untested, all I 
know is that it compiles.

Announcing the release of greetd 0.10.3.

https://git.sr.ht/~kennylevinsen/greetd/refs/0.10.3

---

greetd 0.10.3

This bugfix release fixes an incorrect cast in the PAM ffi code that
could lead to periodic failures.

Kenny Levinsen (2):
       greetd: Use pointer::is_null instead of cast
       Bump version to 0.10.3

The issue has been identified - a stupid bug in the PAM FFI code leading 
to periodic login failures.

I will be rolling out a 0.10.3 shortly once the fix has been validated.

I am currently looking into an issue reported in the IRC channel just 
after I pushed 0.10.2.

If the issue is with the 0.10.x, we might end up rolling a 0.10.3 with a 
bugfix, so maybe hold on updating for a moment.

Announcing the release of greetd 0.10.2.

https://git.sr.ht/~kennylevinsen/greetd/refs/0.10.2

---

greetd 0.10.2

This bugfix release fixes builds on stable rust toolchains.

Kenny Levinsen (2):
       greetd_ipc: Only set feature flag for docsrs
       Bump version to 0.10.2

> Updating to 0.10.1 now throws error[E0554] related to the doc_cfg line
> in greetd_ipc/src/lib.rs:34:1. Removal of the doc_cfg line results in a
> successful build. Tested on openSUSE and Arch, rustc 1.78.0.


Thanks for highlighting this - I might not have noticed otherwise, as I 
have had to run nightly due to cargo libgit issues. I'll roll out a 
0.5.2 with a fix for stable compatibility in a bit.

I'm not particularly amused by docs.rs requiring nightly features to 
build such basic documentation, especially since the exact same code 
built just fine when 0.9.0 was released... Oh well.

Announcing the release of greetd 0.10.1.

https://git.sr.ht/~kennylevinsen/greetd/refs/0.10.1

---

greetd 0.10.1

This bugfix release reintroduces the XDG_VTNR environment variable which
some X11 sessions were relying on. It also changes the error behavior
when a greeter incorrectly sends a user response when there is no
question to answer.

Kenny Levinsen (6):