Re: Mach's in-kernel scheduling

[Thomas Bushnell, BSG]

Igor Khavkine <i_khavki@alcor.concordia.ca> writes:

> It is my understanding that Mach supports "in-kernel" threads,
> i.e. threads that run concurrently in the kernel, not being visible
> in userspace.
> 
> The first kernel thread, after it does all the initialization and starts
> userspace initialization, runs vm_pageout() to become the garbage
> collection daemon.

Please don't call this garbage collection, it isn't.  

> Turns out that zones are susceptible to garbage collection. However,
> I'm really clueless about Mach's way of scheduling these "in-kernel"
> threads. When does this GC thread actually get to run, can it run
> concurrently while a system call is executed, or does it need
> a yield statement somewhere?

It's a thread just like any other.  It's in the same state as if a
system call started doing things but never returned.

Threads in kernel mode like this (either this one, or threads during
syscall execution) are not pre-emptively scheduled.  Instead, they
must explicitly block to relinquish the CPU.  

The basic code for this thread is vm_pageout.c: vm_pageout_continue.
If you look at that function, you can see where it blocks.