[trustees] Kernels newer than 2.6.18

[Andrew R. <an...@ae...>]

Ok guys,

I'm still tracking down what is making the hash rebuild go crazy, but at
the very least, I can at least verify that the problem exists (this is
good, so it should be fixed in the near future).

Currently, I am continuing to base my changes against 2.6.18 kernel, but
I took at look at what was coming up with the 2.6.19 (and already the
2.6.20 kernel is out), and the future is not good.  Basically, with all
the interest in virtualization they are changing around the
namespace/vfs stuff quite a bit (and of course, never in a backwards
compatible way, such is life).  

So I've decided that in the future, I will *probably* be making all
releases as patches to the kernels rather than an external project
(we're practically there already as newer kernels require a patch
anyway).  I guess I'm just asking what the preferred policy on which
kernel will be patched against should be.  Anyone have any suggestions?
I tend to believe in just keeping it up to date with whatever is newest
stable, but I'd be open to other views.  (although I'm staying against
2.6.18 for now, I need to refigure out what the heck they all changed in
2.6.19 and newer).

At some point the subversion repository will probably become obsolete as
I'm going to have to use git to track the kernel development more
closely, but I'll let everyone know when that happens.  (git has quite a
learning curve it seems and i've barely started up that curve :)

- Andy

-- 
Andrew Ruder <an...@ae...>
http://www.aeruder.net