Re: Mach emulation

[Gernot Heiser]

On Wed, 15 Nov 2000 12:28:56 +0900, OKUJI Yoshinori <address@hidden> wrote,

okuji> > Would it be safe to assume that a multiprocessor ukernel multiplexes
okuji> > kernel-level threads to the available processors transparently?
okuji> 
okuji>   As far as I see L4/Alpha, that isn't.

You're right. The principle of separation of policy and mechanisms
requires that the kernel is free of policies (such as load
balancing).

It's also required for efficiency. Running a thread on the "wrong"
processor is costly (cache coherency) and migrating is expensive
(cold caches).

I've used that to start an L4 FAQ -
http://www.cse.unsw.edu.au/~disy/L4/faq.html

Gernot


Gernot Heiser                ,--_|\   School of Computer Sci. & Engin.
Phone:  +61 2 9385 5156     /      \  The University of NSW
Fax:    +61 2 9385 5533     \_,--._*  UNSW SYDNEY NSW 2052, Australia
E-mail: address@hidden        v   http://www.cse.unsw.edu.au/~gernot