[SOLVED](kind of) [linux-raspberrypi] RAM corruption

[nullptr]

I'd like help with narrowing down a RAM corruption on my Raspberry Pi (1) Model B rev 2.

This ram corruption leads irregularly to segfaulting or double free-ing programs, corrupted files and unresponsive init daemon so i had to use the --force or --force --force option to reboot.

It is not the power supply nor the sd card because I can easily reproduce this error with $this->bbcode_second_pass_code('', 'sudo memtester 240M 1') 240M being just a few MB less than what I have available after the services for which i use the Pi.

To further test this I build a 4.19.81 kernel (following to raspberrypi.org cross-compile guide [as the pi wasn't able to build a kernel even via distcc]) with CONFIG_MEMTEST=y and added memtest=17 to /boot/cmdline.txt and to my surprise memtest didn't report any bad memory regions and on this kernel "sudo memtester 240M 10" only failed once.
I therefore think that something in the (too long to include here) diff between the upstream config (with "make bcmrpi_defconfig") and the archlinuxarm config could be what leads to this.
So what config options can i fiddle with too further close in on this potential bug?

tl;dr:
100% repro ram corruption, but only ~10% (currently running a longer test) with kernel with upstream default config
What config options should I test to narrow this down?

[nullptr]

I'm considering this solved, as further testing showed that this is not an issue with the kernel package.
I am currently using "disable_pvt=1" in /boot/config.txt and reinstalled a fresh sd card image to make sure I don't have any corrupted files, hoping this will help.

Using memtester and the code from https://fivelinesofcode.blogspot.com/20 ... sical.html it looks like the error is/was not bound to a fixed physical address so I am not any closer to finding the cause of that error