warning of pending commit attempt.

Julian Elischer julian at elischer.org
Thu Feb 28 18:37:22 UTC 2008

Marko Zec wrote:
> On Thursday 28 February 2008 16:27:41 Dag-Erling Smørgrav wrote:
>> Marko Zec <zec at icir.org> writes:
>>> Actually the patch provides certain level of support for
>>> virtualizing leaf sysctl nodes.  So far I have only introduced
>>> macros for methods / data types that I've found necessary to
>>> virtualize, such as SYSCTL_V_OID, SYSCTL_V_STRING, SYSCTL_V_INT,
>>> and SYSCTL_V_PROC.  [...]
>> Thanks, this is exactly what I was looking for.  Now all we need is a
>> way to start a vimage with hw.machine and hw.machine_arch set to a
>> vimage-specific value...
> So your question opens up a pandora's box...  Obviously it's trivial to 
> virtualize a sysctl, but I still don't have a clear idea on what would 
> be the most convenient way of specifying start-up constraints or 
> parameters when instatiating a new vimage.  At the moment each 
> virtualized variable is initialized to some system-wide compiled in 
> constant - we need to come up with a much more flexible / configurable 
> model...

Whooooa there!

Before we widen the scope of the vimage project to complete 
virtualisation of everything.
How about we get what we have now into the tree? :-)

BTW Marco, you might want to add some comments in vimage.h about how
you see the current framework growing to encompas such things as
dynamically assigned module numbers for kld modules and such so that
when we commit it, there is some sort of architetural guide for the
"thousands of people" who will want to improve it and extend it to
other views of virtualisation.

More information about the freebsd-current mailing list