I just tested the installation media using another board (KCMA-D8) over serial as well and it hung at the same point. Perhaps there's some issue with serial during init?

thanks for doing these experiments - some computers have problems with coreboot and libreboot, and some do not - it is still not obvious which is which though - according to your remark on IRC, you seem to have noticed that yourself

we need people who find those problems with coreboot and libreboot, to do more experiments - they have been difficult to support because people who report problems, most often do not give enough information, or follow-through with more experiments - what is apparent is that there is some important difference with EFI vs BIOS mode, and between the openrc and systemd ISOS

have you tried any of the other ISOs? - the openrc/LXDE ISO, any of the systenmd ISOs?
have you tried booting with either EFI and BIOS mode?

someone noted on ticket #2628 that coreboot/libreboot can boot the openrc ISOs; but only in EFI mode - ticket #2108 has some example grub entries, which people have used to coax libreboot into booting the ISO - comment 23 of #2108 is a list of more experiments that could be done to narrow in on the symptoms

Tested the other ISO images and the results are below:

• OpenRC/LXDE: Hangs at the same point during bootup
• SystemD/LXDE (non-talking parabola): Works!
• SystemD/CLI (talking parabola): Works!

I only tried booting with BIOS, not EFI.

so the APU2 with coreboot can boot the systemd/LXDE ISO, but not any of the openrc ISOs, or the talking ISO ?

Both the APU2 and KCMA-D8 can boot the SystemD ISOs (both Talking and non-Talking) but not the OpenRC ones (CLI and LXDE).

Seems to point to OpenRC as the issue.

thanks, that is the most interesting result in a while - the talking and non-talking SystemD ISOs were made from a different code branch - the talking ISO was made from the same code branch as the openrc ISOs - that is what i have been assuming was the most important difference; but this suggests that the problem is specific to the openrc boot process, and not the EFI

note to self: this probably means that the unified branch can be merged with the master branch; then finally rebased onto the arch32 upstream

If no one is able to find the bug, I'm willing to purchase an old monitor to hook up to my KCMA-D8 and see if there's a difference having a display instead of serial-only output w/ the OpenRC ISOs. This could be why it's possible to boot the OpenRC ISOs on a Librebooted x200 and Librebooted iMac5,2 (tested both myself) with no issue. But we'll cross that bridge when we get to it, I suppose.

tl;dr everything, but I suspect that you have to wait a bit more for OpenRC to boot everything. You seem to be stucked when starting the 'local' service, but I'm sure it's just that it's just taking longer, nothing else. There's a 'local' script which starts the pacman keyring, and sometimes it takes its time. If you have a kind of slow PC, you'll need some patience.

https://git.parabola.nu/parabolaiso.git/tree/configs/releng-openrc/airootfs/etc/local.d/pacman-init.start

tl;dr everything, but I suspect that you have to wait a bit more for OpenRC to boot everything.

I've let it sit for hours to no avail so, no, this is not the case. It indeed hangs at this point. I'll leave it running today while I'm at work and report back on the results after I'm home ~6 hours later but I can guarantee it won't make a difference (I've left it running while I slept before and there was no change).

Megver83 wrote:

You seem to be stucked when starting the 'local' service, but I'm sure it's just that it's just taking longer, nothing else.

No, it's not stuck starting the local service. If you read the log I attached in the OP you'll see that it completes. Yes, it takes a few minutes (at most) but it does finish initializing the pacman keyring. What is happening is that it hangs right after.

I let it run while at work and 8+ hours later it's still hung at the same point. This is not a matter of it just "taking longer".

UPDATE: I tested the OpenRC ISOs (CLI and LXDE) using a physical monitor connected to my KCMA-D8's VGA output and was able to boot successfully. The KCMA-D8 was running the latest release of Libreboot configured for textmode and the SeaBIOS payload.

Thus, the issue is wholly unrelated to coreboot/Libreboot and should be unassociated from issue #2108. The issue lies with serial during the OpenRC boot process, causing a hang after ostensibly initializing the pacman keyring. In my opinion until this is fixed a note should be added somewhere that warns users of broken serial communication when booting from the OpenRC ISOs since this affects headless setups.