Re: [PATCH 1/4] Virtualization/containers: introduction

From: Hubertus Franke
Date: Thu Feb 09 2006 - 11:36:44 EST


Jeff Dike wrote:
On Wed, Feb 08, 2006 at 05:24:16PM -0700, Eric W. Biederman wrote:

To deal with networking there are currently a significant number of
variables with static storage duration. Making those variables global
and placing them in structures is neither as efficient as it could be
nor is it as maintainable as it should be. Other subsystems have
similar problems.


BTW, there is another solution, which you may or may not consider to
be clean.

That is to load a separate copy of the subsystem (code and data) as a
module when you want a new instance of it. The code doesn't change,
but you probably have to move it around some and provide some sort of
interface to it.

I did this to the scheduler last year - see
http://marc.theaimsgroup.com/?l=linux-kernel&m=111404726721747&w=2

Jeff



Jeff, interesting, but won't that post some serious scalability issue?
Imaging 100s of container/namespace ?

The namespace is mainly there to identify which data needs to be private
when multiple instances of a subsystem are considered and
encapsulate that data in an object/datastructure !

-- Hubertus

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/