GIOS P2L4 - Part 2

What is the purpose of the set_concurrency system call?

• To allocate kernel level threads to a process

• To request more user level threads

• To signal the kernel before blocking

• To interact and coordinate between the kernel and the user level library

Why is it helpful for the kernel to signal to the user level library before blocking?

• To allocate another thread to the process temporarily

• To deallocate a thread when it becomes idle

• To allow the user level library to request more kernel level threads

• To coordinate and make progress even when some threads are blocking

What does the user level library see?

• Kernel level threads

• CPUs

• Kernel level scheduler

• User level threads

What can the user level library do to ensure a user level thread always executes on a specific kernel level thread?

• Request binding of the user level thread to a specific kernel level thread

• Pin the kernel level thread to a particular CPU

• Specify the mapping between ULT and KLT in the many to many case

• Use mutexes and wait queues for synchronization

When does the process jump to the user level library scheduler?

• When ULTs explicitly yield

• When a timer set by the UL library expires

• When ULTs call library functions like lock/unlock

• When blocked threads become runnable

What is the issue when a user level thread acquires a lock and the kernel level thread gets preempted?

• The user level library scheduler cycles through remaining user level threads

• The user level threads waste time until the thread holding the lock is scheduled again

• The ULT/KLT mapping in the many to many case changes

• The kernel is unaware of the data structures used by the user level

In a multi CPU system, how can a user level library impact what is running on another CPU?

• By directly modifying the registers of the other CPU

• By sending a signal from one CPU to the other

• By requesting the kernel to preempt a thread on the other CPU

• By blocking the thread on one CPU and scheduling another thread on the other CPU

What are adaptive mutexes used for?

• To block threads when the critical section is short

• To spin and try to acquire the mutex in a loop

• To determine if the owner of the mutex is currently running on a CPU

• To store information about the length of the critical section

What happens to the memory associated with a thread when it exits?

• It is immediately freed

• It is marked as being on death row

• It is reused for new thread creation

• It is garbage collected periodically by a reaper thread

What are interrupts?

• Events generated externally by components other than the CPU

• Notifications that some external event has occurred

• Signals triggered by the CPU and the software running on it

• Events that occur in response to specific actions on the CPU