Is a upstream Slackware (Porteus) kernel possible?

Post#1 by **fu11m00n** » 21 Apr 2026, 08:39

I'm in the process of collating the Aarch64 (Stuart Winter) and x86_64 (upstream) repos into my project... for my Q4 release (June).
The packages come directly from Slackware upstream. It's a walk in the park.

However... everything grinds to a halt when I get to the kernel.
The upstream Slackware kernel cannot be shoehorned into Porteus (without a kludge).

So what is the kludge that lets the "real" Slackware kernel replace the Porteus one?

Unfortunately... my project requires a pure upstream kernel.
To transition Porteus to Aarch64 a normal Slackware kernel is essential.

Thanks in advance... for input that resolves this issue.

Post#2 by **ncmprhnsbl** » 22 Apr 2026, 02:36

porteus at present requires kernels built with an AUFS patch to function. (the patch is applied to the kernel source before compilation)
neko has managed to create versions of porteus (basicly using a modified initrd and some other tweaks) that make use of OverlayFS, which presumably, a standard slackware kernel would have enabled.

Post#3 by **fu11m00n** » 22 Apr 2026, 05:33

@ ncmprhnsbl... thanks.

Hmmm... sounds very interesting. So a "stock" Linux kernel (source code) gets a patch before compile time.
Neko's approach also sounds golden. Is it his tool in the "Community" section that does all the heavy lifting?

Or is it a different approach found in the "Tutorials" section?

I'm still wading thru Stuart Winter's (Aarch64) creation and it's going great.
I made my 1st post of the Aarch64 (and x86_64)... "Linux From Scratch" modules on my Sourceforge project page.

Painstakingly by hand I have reorganized (using depfinder and slakfinder) the .txz packages (for Aarch64 and x86_64) into package "bundle" repos.
My curating has been slowed down this past week because COTUS (Bill HR 8250) has forced me to offshore my project to Switzerland and Taiwan.

One of the main goals of my project (which has drawn the BDFL's attention) is to make Slackware completely (pre)resolve.
AlienBob's project seems to have resolved Slackware but in a way that is a blackbox (I can't wrap my head around it).

My modular approach is... Kernel --> Directory --> Library --> Linux From Scratch --> X11 --> "/a" --> Api --> Cli --> Gui.
I should be done collating by week's end. I think the missing link ("the last mile") is... a real Slackware (OverlayFs?) kernel (and some initrd kludges).

Resurrecting Ahau's ancient ARM project is a lot harder that I thought it would be.