[klibc] klibc-1.2.1: kinit works...

[H. Peter Anvin]

Aaron Griffin wrote:
> 
> Just curious, is there any plan to add some sort of system to load
> modules before running kinit.  I've been doing some work
> here-and-there with a similar initramfs setup, where all the pre-root
> modules are added to the initramfs rather than built-in.
> 
> What are your opinions of loading modules, creating device nodes, then
> passing things off to kinit.  Would you recommend something like this?
>  Or would it make more sense to ignore kinit in a case like this?
> 

Well, kinit is primarily intended as a replacement for the in-kernel 
root device handling, all of it, with all the ugly warts that are there. 
  It's also intended to operate totally standalone, which leads to a 
whopping 20K compressed initramfs.

That being said, there is no reason that you can't run code before it; 
to make that simpler, in fact, I'm trying to set it up so that the 
default initramfs looks like:

/dev/console
/dev/zero
/dev/null
/kinit
/init -> /kinit

... so that you can override /init, put your own code there, and then 
run kinit.

I would be happy to discuss conventions for initramfs, but if we're 
going to do so, we should keep in mind that they need to be generic 
enough that they can be included in the stock kernel.  At this stage I 
really rather want to get the current compatibility stuff done and 
released, rather than try to add things to it.

	-hpa