Scheduler Activations

Parallelism

PL uses both user and kernel threads

Let's say we run the code on a 4-CPU machine, the system will create 4 kernel threads first. Then all the use level threads will be put in a work queue, and run on the 4 kernel threads. It is multiplexed.

Problem

When a user-level thread makes a blocking I/O request or takes a page fault, the kernel thread serving as its virtual processor also blocks. As a result, the processor is idle.

Plausible Solution

Create more kernel threads than number of physical processor; When one kernel thread blocks because its user-level thread blocks in the kernel, another kernel thread is available to run user-level threads on that processor. But when the blocked user-level thread returns, there will be more runnable kernel threads than processors.

Scheduler Activations

"Fancy" kernel level threads. When blocked, notify, and run another use-level thread.

It notifies the user-level thread system of a kernel event, such as the user-level thread blocks or wakes up. It also provides space in the kernel for saving the processor context of the activation's current user-level thread.

When a user-level thread is blocked, create a new scheduler activation, upcall into the user-level scheduler. The user-level scheduler decides what is the next to run.