Subject: Re: [vserver] official kernel inclusion
From: Herbert Poetzl <herbert@13thfloor.at>
Date: Thu, 10 Jan 2008 01:05:51 +0100

On Wed, Jan 09, 2008 at 12:20:42PM -0600, Adam Majer wrote:
> >>> What steps are required for it to be included?
> >> political hick-hack and a lot of convincing/knowing
> >> the core kernel developers ...
> > 
> > it might be worth the effort, though. OOT development is the main
> > obstacle for vserver acceptance here (compared to OpenVZ). 

OpenVZ is not OOT?

> > It would be a pity if vserver did go the reiserfs way.

> I think the correct way to proceed would be to isolate small changes
> that could be added to the kernel that would not be seen as too
> intrusive and definitely not in one go. The process could take months
> and months, but eventually the vserver patch would be reduced to,
> hopefully, zero.
> 
> My understanding is the changes vserver introduces are,
> 
>   * process isolation
>      + process space
>      + memory
>   * file system isolation
>   * privilege reduction in guest
>   * network isolation
> 
> I would propose that we try getting each one of these, in turn, 
> into the kernel. 

> The first would have to be the process space isolation.

mainline goes for the heavier process virtualization
instead of isolation ...

> But before even doing that, we need something better than one giant
> patch. 

http://vserver.13thfloor.at/Experimental/split-2.6.22.10-vs2.2.0.5.tar.bz2

> Is there a vserver repository where people can work and commit
> stuff? For example, arch/git/svn/hg/whatever ?

git is distributed and simple, create your own repository
and start working on it ...

best,
Herbert

> - Adam
> 
> BTW: I'm not a vserver developer or even a kernel developer, but I'll
> learn something in my hacking :)