Arjan van de Ven wrote:

quoted

Since this code is so rather, um, custom, I was going to reimplement 
stop_machine in the module.
    

that sounds like a big mistake. I assume you want your VMI module to be
part of mainline for one.

And this is the sort of thing that if we want to support it, we better
support it inside the main kernel, eg provide an api to modules to use
it, rather than having each module hack their own....
  

Yes, after discussion with Rusty, it appears that beefing up 
stop_machine_run is the right way to go.  And it has benefits for 
non-paravirt code as well, such as allowing plug-in kprobes or oprofile 
extension modules to be loaded without having to deal with a debug 
exception or NMI during module load/unload.

Thanks,

Zach