Linux Kernel Development Hits An All-Time High

[phoronix]

Phoronix: Linux Kernel Development Hits An All-Time High

The advancement of the Linux kernel in 2014 was nothing short of fantastic! The kernel added so many new features and is now more than 19.1 million lines of code...

http://www.phoronix.com/vr.php?view=MTg3OTE

[duby229]

The bigger the better is -not- true in most cases, and this is one of them. There is a hell of a lot of junk in the kernel that -needs- to be branched out.

[_deepfire]

The bigger the better is -not- true in most cases, and this is one of them. There is a hell of a lot of junk in the kernel that -needs- to be branched out.

Yes, "lines of code spent".

[opensource]

It moves with time, once hardware gets too outdated, code gets removed. But too much code does indeed slows everything down, and Linus has already been asked/knows about this ultimate problem. Speed of hardware increases too, so at the moment.. IDK.

[Ferdinand]

The bigger the better is -not- true in most cases, and this is one of them. There is a hell of a lot of junk in the kernel that -needs- to be branched out.

I thought you could you choose what you compile in the kernel?

[mazumoto]

Most of it is drivers, I think. And that part will always tend to grow, as there is a continuous stream of new hardware being released. Additionally you do want to support old hardware, that's a strong point of Linux. So there will be not so much opportunity for cleanup and removal. As it's all relatively modular, I don't think it's a big problem. Similar with file systems.

Of course there is also the tendency to pull more and more into the kernel, it seems (drm, kdbus, etc.). But Linux is for performance. If you want a really lean kernel, go Hurd or Minix.

[not.sure]

The bigger the better is -not- true in most cases, and this is one of them. There is a hell of a lot of junk in the kernel that -needs- to be branched out.

Like what?
The larger additions are typically file sytems and drivers, both of which are pretty self-contained, and add real features.

[Ferdinand]

Most of it is drivers, I think. And that part will always tend to grow, as there is a continuous stream of new hardware being released. Additionally you do want to support old hardware, that's a strong point of Linux. So there will be not so much opportunity for cleanup and removal. As it's all relatively modular, I don't think it's a big problem. Similar with file systems.

Of course there is also the tendency to pull more and more into the kernel, it seems (drm, kdbus, etc.). But Linux is for performance. If you want a really lean kernel, go Hurd or Minix.

But if you wanted to use the kernel for a simple smartwatch you could only choose the 'features' you wanted right? So you would get a lean kernel.

[Licaon]

http://www.phoronix.net/image.php?id=0x2015&image=linux_kernel2014_4

What is the purpose of this seemingly futuristic graph?

[duby229]

Oh, yeah please don't misunderstand me. Even though the kernel is monolithic, it's awesome that the modular architecture works so well. It really is one of the best things about linux from user perspective. It's great! I love it.

From the perspective of developer though, It's huge. It's not just my opinion that there is a lot of code in the kernel that doesn't belong there. Plus there is the issue of old code too.