[Solved] FAILED TO OPEN /dev/mcblk1p2 LUKS

[kucklehead]

I have gentoo amd64 installed on a sd card. I followed these guides here:
https://wiki.gentoo.org/wiki/Dm-crypt#Kernel_Configuration
https://wiki.gentoo.org/wiki/Systemd
https://forums.gentoo.org/viewtopic-t-1136343-postdays-0-postorder-asc-start-50.html
https://amedeos.github.io/gentoo/2019/01/14/install-gentoo-with-luks-lvm-and-systemd.html
https://wiki.gentoo.org/wiki/Genkernel

NOTE: It is installed on a sd card. I don't know if that's an issue

Here is my config:
http://dpaste.com/ATENS5FAS

Here is my grub.conf:
http://dpaste.com/BC2L2FFXL

I am using genkernel which I have it configured to literally do everything for me:
http://dpaste.com/H8MLJD24B

Here is my /etc/crypttab:
http://dpaste.com/3Z8S9R2RL

Here is my etc/fstab:
http://dpaste.com/BQPTA8KCT

The issue: The issue I have is after typing in my passphrase. It will pop up "Killed" and followed up by FAILED TO OPEN /dev/mmcblk1p2 LUKS
Can't find root

Edit:
there are two partitions. One is called cryptlvm which is available once I unlocked it. That needs to be mounted first for everything else to work, I am assuming

[alamahant]

It might be the case that
/dev/mmcblk
changes to /dev/sdx
Plz use PARTUUID or UUID
_________________

[Hu]

Killed is not normal here. My suspicion is that your cryptsetup dies badly, before it can finish its work. If you drop to a rescue shell and do the initramfs steps by hand, do you also get Killed? What is the full command that is killed to produce that output?

[kucklehead]

When I do drop into shell and do the initramfs by hand, I still get the same results.
Example: If I type in cryptsetup open /dev/mmcblk1p2 cryptlvm
It'll ask for my password, and the same error is "reproduced"

If I go into shell and poke around such as:

[kucklehead]

I was wrong about the command. How do I debug it and find out why it is outputting "Killed" whenever I try to manually open it or open it on boot. This is for a Chromebook, and I don't know if that affects anything exactly?

[alamahant]

If you have logical volumes you do NOT need UUID
Maybe only for the cryptdevice.
_________________

[Hu]

Normally, Killed would be printed by your shell to advise you that the child process was terminated by a SIGKILL. It's rather unusual for that to happen in interactive contexts. The only case I can think of would be that your cryptsetup was built with buffer overrun detection, and a buffer overrun occurred, which triggered the handler to SIGKILL the process to contain further damage. If so, an strace should confirm that it is sending itself a SIGKILL. Normally, I would expect this sequence of events to also print a fairly obvious diagnostic about "stack smashing detected" just before the SIGKILL is sent.

[kucklehead]

Unfortunately strace isn't built with my busybox shell. Been following this guide here: https://linux.die.net/man/1/strace
Is there any debugging tools that come with it or do I have to include strace into it?

I tried installing it with dracut and it doesn't load past loading initramfs <gentoo>

These are the commands I used to build the init with dracut:

[kucklehead]

How do I trace it? I was googling it and found pgrep? Is there an article I can read. How do I do this exactly?

[Hu]

strace -o /tmp/crypt.strace -f -tt crypsetup luksOpen /dev/mmcblk1p2 cryptlvm. Before posting, edit your password out of the strace output. If cryptsetup crashes even on a wrong password, that would be ideal, as you could just enter a dummy password that it does not matter if we know.

[kucklehead]

I found out why it was not working. It is because it is on a chromebook which has low memory on it. Here is my proof and how I figured it out: https://sleeplessbeastie.eu/2021/07/05/how-to-open-luks2-encrypted-volume-on-a-low-memory-device/
Now I have another question, can I remotely update and upgrade gentoo using a more powerful machine? I was thinking SSH?