PLEX86  x86- Virtual Machine (VM) Program
 Plex86  |  CVS  |  Mailing List  |  Download  |  Linux  |  Newsgroups

ancient kernels 2575


Your Ad Here

Your Ad Here

ancient kernels 2576
Thanks to Michael Black for his many comments and suggestions. If I choose to disagree with some (but not all) of his remarks...

Keep in mind that Linux was in effect well developed, given that it was no period of defining the kernel (unlike Windows where it took Microsoft various commercial releases to get it right), since the target goal was already well developed.

But, it was bootstrapping. Linus let the kernel loose pretty early, because when he'd done that then others could work on it. And of course, it often makes sense to make something minimal, for which later work can be built on. I seem to recall reading that there was a period when there was a new kernel release every day for a period of time.

So those early releases may be stunted versions, with only incremental changes between versions. There will be a time when it slowed down, and that's likely where you want to be, because only then do you have a kernel that is close to the "finished" kernel.

But really, you're all over the map. First you want to study the kernel. But it shouldn't matter what how big the kernel has grown, because the basics have always been there. You don't look at all the code at one time, you look at the functions one by one. You can simply ignore the code added later that simply adds more hardware support, because once you understand the basic of one driver for a floppy drive (for example) you don't need to understand all the drivers that do the same function. Indeed, you may just need to grasp how one driver, for anything, works (what it does, how it interfaces with the rest of the kernel, how it is interfaced with the actual hardware, etc) and that will give you enough insight.

Likewise, unless you can have a clear goal for why you want to study the kernel, you may just be thrashing away. Most people will never have to "understand the kernel", with a secondary number wanting to understand some specific part (because they need to fix something make a new driver for some bit of hardware they have), and a relatively tiny number have a need to fully understand the kernel. Yes "need" is a tricky term, but learning often is propelled by the concrete world. If you have no reason to know how the kernel works other than to know, the process may not go as well as when you have something you need to fix, or whatever.

And it's still not clear to me why you need to understand the kernel in order to make your own distribution. Looking at the code isn't going to help you there, or at least there should be documentation about how to plug the kernel into other things, because you need that step in order to understand the kernel.

And if you don't need to study the source code, you don't need to worry about getting an early version of the kernel, which will carry a penalty in itself.

Look for documentation from the early days, because the only way back then was to "do it yourself". The distributions just took the process and mbutt produced it. The Linux From Scratch HOWTO is likely what you really want, and there are likely some other HOWTOs that will be helpful, including one about going from boot to BASH prompt.

Note also that some recent "live" distributions have documentation about customizing them, and that may serve your purpose.

Michael



Your Ad Here

List | Previous | Next

ancient kernels 2576

Linux groups from Newsgroups

The #1 Usenet Provider on the Internet

ancient kernels 2574