Can I try to change the Linux kernel?

Question

So, I want to try a new concept, which includes deep modifications to the Linux kernel, and I wanted to know whether it is worth trying or too hard?

Answer 1

Go to LKLM and check the archives if anyone else is working on something similar. If not, let us know what you plan to do and ask about the easiest way to do it (kernel module, etc.). Create your own git repository to make changes and allow other users interested to exit your repo. If your work is successful, ask the maintainer to pull your changes and add them to the trunk.

This, of course, is not wise to do, but only you will find out if you should work on these profound changes.

Answer 2

In the absence of additional information, it sounds as if you are suggesting a better description of the kernel modification, rather than forking.

What aspect of your ambition requires forking? Viking is a thing whose essence is mainly determined by political terms, that is, renaming a project, setting up another strategic direction, etc.

Answer 3

I want to stick to a question that is related to difficulties. If you just want to add a system call to provide limited additional features, such as configuration file descriptors, then this is what any trained programmer can do. If you want to play in real time, performance, process processing, beware that it can be very difficult due to how deep you enter the core.

For many courses, course students need to add a system call or add some trivial fields / parameters to the design. About 20 hours for something trivial, 60 for something easy, but useful.

Answer 4

Like any piece of truly open source software, forcing the kernel is as simple as forking from the version control repository and starting work on your own version. Just like any piece of truly open source software, if you intend to branch it, it is best to have a solid reason for branching in place. And if you ultimately plan to unite, then you really are not branching, but instead you are branching, and then ask that it be united. Just remember to regularly replace this instance.

For the core, however, it will be difficult ... the Linux kernel is a very large and complex software. You will be stuck with GPLv2 unless you delete all the GPLv2 code that cannot get permission to upgrade to GPLv3, so licensing should stay the same. But if you want to make significant changes to the Linux kernel because you think you can do it better, absolutely no one is stopping you. Add new drivers, change the way the charts work, remove the old APIs, do whatever you want. As soon as you prove something, people will begin to pay attention, especially if you do it better than at present. And you can achieve the final good result: your work will ultimately be merged back into the highway.

Answer 5

Hiding open source software should be seen as a last resort. First you should try to make a kernel branch and send it for merging. It will make life a lot easier for everyone.

Answer 6

Deep modifications will present you with a dilemma: do you track new Linux kernels and merge your changes, or do you use the kernel from which you originally aimed and fix critical fixes?

If your changes cannot be included in the main core, you will come across the above. Can you support these efforts, and is it worth it?

Answer 7

Yes, it is difficult, but if you have a new concept, and you think that building it on top of the kernel is a good idea, then it may be worth the time.

If you are still disconnecting from the Linux kernel, there is always a Minix kernel ...

Answer 8

IIRC properly support the modified Linux kernel - a set of patches as in Andrew Morton mm .

Answer 9

most of what you can do should be done using the kernel module. If you really have something big, then you can try and develop it. Hey, open source, no one will tell you what not :) Just make sure you support often.

Answer 10

You should consider whether your idea really needs to be executed in kernel space or whether it can be efficiently executed in user space. That would be a reasonable question to ask yourself (with details of the idea, of course).

I often saw people posting kernel-related ideas with LKML that really can be done entirely in user space, sometimes more easily and efficiently.