How do I develop a kernel? I want to get rid of the bloated mess made by Linus and his sjw team

did it fail? if not, then it should probably be ok

keep one kernel you know its correct at least

Read LInux From Scratch, load only the stuff you want.

You can also look into MINIX

No it seems to be working so far.

Just use MINIX.

Just needs USB drivers

>Jow Forums nazi can't program
how surprising

drivers will be problem in the end, not new kernel.

i have haswell g3258 and an msi h81m-p33 and i can't get anything hurd based to boot. h81 and haswell are 6 years old at this point.

hurd feels like it's one of those "works on bare metal only if you have an early 00's thinkpad" stuff.

i'm open to listening other user's experiences as i'm curious.

Attached: stallman2.jpg (1536x1536, 1.03M)

Get seL4

This is ryzen is the better option, AMD sells cheap old CPUs.

Importing drivers is quite easy we might just need a better kernel actually
Something like the bsd kernel

Everything sponsored by Stallmann and his foundation is garbage including hurd. He waste all donation money in political agitation

Try porting Sel4 to the GNU userland.

Try linux 5.1.15. If that doesn't work, drop back to the longterm version 4.19.56+.
Also, install the appropriate firmware package on your distro.
You might want to consider using more of a beginner distro. None of these errors or warnings should be happening if you don't understand them well enough to deal with them yourself.

If you read the output it clearly shows an error updating a virtualbox module, and a warning about possibly missing firmware. The system will of course work, but if you need virtualbox you'll probably eventually have problems. The i915 warning was only a warning about possibly missing firmware, if you are using intel video check if performance has degraded, if you have a video card then it can be totally ignored.
I'd recommend upgrading or downgrading from that particular kernel version. 5.1.15 is the currently supported stable kernel, and there are several longterm support kernels from the 4.x series, if you want to have a bleeding edge kernel (which I recommend against since you don't seem very technically inclined) you'd want to try 5.2+ but if you don't understand DKMS and kernel firmware blobs, then you are gonna have a bad time. What distro are you on?

If you create a kernel with a compatible ABI & API, you'll end up with the same sort of problems and less manpower to fix them. FreeBSD kind of skirts the issue with the linux compatibility layer, but the truth is it's not an ideal solution and basically anyone that needs the linux software (&drivers) is pretty much accepting that they lose the big benefits of using the freebsd kernel. DESU an approach like openBSD of saying "we don't support X or Y and have no intention of doing so, and will not support any attempt to do so" is better because it means they can focus on making what they do support work well

>we might just need a better kernel actually
Who is this we?
retarded Jow Forumsfags, I swear, take a look at your device tree and tell me how many components do you see, now think of porting each one of them and that's for your system alone.

Or you could just use something that has a lot of people working on it like Linux and not resort to mental gymnastics trying to defend your hipster OS that nobody uses

>4.19 long term
>must update twice a week
JUST

Serenity OS is a from-scratch implemention of an x86 Unix kernel plus GUI, implemented in C++, with YouTube screencasts of many features being implemented. Should at least wet your beak!

github.com/SerenityOS/serenity
youtube.com/channel/UC3ts8coMP645hZw9JSD3pqQ

>How do I develop a kernel?
arjunsreedharan.org/post/82710718100/kernels-101-lets-write-a-kernel
Literally 1st result on google.
The pure state of linux fags on Jow Forums.

Attached: Screenshot_20190630-052434_Via.png (720x1280, 194K)

>C++

Attached: 1555634180928.jpg (640x480, 138K)

> C++
Condemned

Where's that screenshot from? I want to read the rest.

Ha! 2 seconds ahead bro!

Attached: 10-10.gif (500x288, 528K)

Commentaries section on the fucking link hypersummerfagget

projects to read:
>pre-1.0 linux releases
>9front.org/
>github.com/HelenOS/helenos
>github.com/seL4/seL4
>github.com/mit-pdos/xv6-public
>zircon from fuschia
>github.com/SerenityOS/serenity/
>github.com/dbader/potatoes
>gitlab.com/sortix/sortix
>github.com/littleosbook/aenix
>github.com/rhtyd/tantra
>templeos
>github.com/froggey/Mezzano
>github.com/dwelch67/raspberrypi
>github.com/s-matyukevich/raspberry-pi-os
>github.com/bztsrc/raspi3-tutorial/

references:
>osdev wiki
>intel's manuals, arm's manuals
>0xax.gitbooks.io/linux-insides/content/
>littleosbook.github.io/
>jamesmolloy.co.uk/tutorial_html/
>web.archive.org/web/20190227160201/http://www.osdever.net/bkerndev/Docs/intro.htm
>os161.eecs.harvard.edu/

>intel
>arm
riscv or bust

I don't care for imaginary architectures

it's hardly imaginary when it already has a working toolchain

yeah only existing hardware is still missing

I'm interested in doing something like this.

How good will kernel written in JS be?

I've got almost an year experience studying programming so I think with a little effort and luck I could pull it out.

It will be amazing I'm sure.

You don't think they're going to make an a sequel?

>Try porting Sel4 to the GNU userland.

See genode.org

make a js kernel for node-os.com/ an undoubtedly a new era of computing will be ushered in. Javascript is basically LISP except it actually gets used. Imagine how far ahead we'd be if instead of C and Objective C and C++ everyone just used lisp? We'd already have autoerotic sexbots. But just because history made mistakes doesn't mean we must continue down the stray path. We can redefine the OS as something hypertrophic and beautiful.

Attached: based_os.com.png (1884x1560, 129K)

>Javascript is basically LISP

Attached: TheHomer.png (315x183, 63K)

There is no comment section, unless it requires javascript. You wouldn't happen to be browsing the internet with js enabled, would you?

you can just modify it you dont need to develop

I am interested in programming a (hobby/toy) OS using existing bootloader and kernel. If I want to use the linux kernel is LFS a good route to understand where to start ? Only goal at the moment is to produce a hello world OS that uses the linux kernel.

Serenity is a treasure.

Attached: A5A576D6-365A-4B70-A2BD-B3DC59551AF3.jpg (624x893, 95K)

One person trying to cover that much shit is probably unrealistic, unless he has a lot of free time and resources at his disposal. Retard.

Writing a kernel isn't the same work as writing a userspace application where you can do your own implementation in the same time it takes to understand an existing project. You probably want to exhaust your code reuse from an existing kernel before embarking on a brand new project.

You can't. You're just a sour consumer. Even if you were a genius, you''d run into the same problems as Linus&friends and make worse, single-minded decisions.

Writing your own kernel is unrealistic.

you can write your own kernel, its not as difficult as you would think. creating something with feature parity to present windows, mac, or linux on the other hand is years of full time work at best case.

There are people on Jow Forums writing their own kernels right now, it is not unrealistic.

Microsoft just released the source code for their WSL 2 kernel.
github.com/microsoft/WSL2-Linux-Kernel
Are you excited too?

it's interesting because on WSL 2 supporting machine its actually running a hypervisor on bare metal and windows itself along with your linux instance are running as lighweight VM's. So its not a simple case of MS writing some userspace app that emulates linux behavior.

made by /our swede/

the problem with drivers is domain knowledge, API is not a problem as FreeBSD shows.

The kernel isn't so bad because the x86 platform is documented. But imagine having to write and maintaining drivers for 50 billion pieces of glitchy, undocumented, non-spec-conforming hardware. What a nightmare.

Yeah some trash that will never work. Waste of effort.

Just join a sane project, defined as sjw free.
Openbsd, Dragonfly, Genode, Haiku... you do have choices.

This, unironically. Hell you could just use the templeos kernel and build your owm graphics system on top.

If you're asking for this you're obviously not ready for doing it.

>sour grapes

Let's see yoru kernel.

let’s see yours