On Thu, Feb 22, 2007 at 10:42:51AM -0800, Bucky Katz wrote:
> Joerg Sonnenberger <joerg@britannica.bec.de> writes:
> 
> > That's the point where I disagree. I do not think it is necessary to do
> > that or even helpful. Let blocking events block the lwp.  When you took
> > the time for a trap, the context switching overhead is reasonably small.
> 
> In the uniprocessor case M:N degenerates to N:1, in practice. If I let
> the 1 lwp block, how do I get any of the other threads scheduled?

I'm talking about two different things. The number of LWPs can be
higher, the kernel just keeps only one on the active list of each CPU at
most. Optimising the number of kernel contexts is IMO mostly irrelevant,
since the kernel context is quite light compared to other resources
bound to thread.

Joerg