Process, Thread, Synchronization, Deadlock Review Note - Operating System

我的笔记地址：https://docs.google.com/document/d/19Lfkmvq_eU-KIHKhOrx4kYBOdL8sQpxWvtaod72Oaoc/edit?usp=sharing

 

Architectural support for Oses

Application: written by programmer, compiled by programmer, uses function calls

Libraries: written by elves, provided pre-compiled, defined in headers, input to linker, invoked like functions, maybe resolved when program is loaded

Portable OS layer: system calls (read, open, …), all high-level code

Machine-dependent layer: bootstrap, system initialization, interrupt and exception, I/O device driver, memory management, Kernel/user mode switching, processor management

 

1. Types of architecture support

1. Manipulating privileged machine state

2. Generating and handling events

3. Events: interrupts, exceptions, system calls, etc

2. Privileged instructions

1. What are privileged instructions?

A subset of instructions of every CPU is restricted to use only by the OS

1. Who gets to execute them? - OS

2. How does the CPU know whether they can be executed?

– When it runs the OS code. Kernel mode or user mode is indicated by a status bit in a protected control register. CPU checks mode bit when protected instruction executes. Attempts to execute in user mode are detected and prevented.

Only the operating system can

Directly access I/O devices (disks, printers, etc.)

Manipulate memory management state

Manipulate protected control registers

Halt instruction

3. Difference between user and kernal mode

1. User programs execute in user mode

2. OS executes in kernel mode

3. Why do they need to be privileged?

1. OS must be able to protect programs from each other

2. OS must protect itself from user programs

4. What do they manipulate?

1. Protected control registers

2. Memory management

1. OS must be able to protect programs from each other

2. OS must protect itself from user programs

3. Memory management hardware (MMU) provides memory protection mechanisms

4. Manipulating MMU uses protected (privileged) operations

3. I/O devices

3. Events

An event is an “unnatural” change in control flow.

Events immediately stop current execution.

Change mode, context (machine state), or both.

1. Events

1. Synchronous: fault (exceptions), system calls

2. Asynchronous: interrupts, software interrupt

2. What are faults, and how are they handled?

3. What are system calls, and how are they handled?

1. How do I/O devices use interrupts?

4. What is the difference between exceptions and interrupts?

1. Exceptions are caused by executing instructions

1. CPU requires software intervention to handle a fault or trap

2. Interrupts are caused by an external event: signal asynchronous events

1. Device finishes I/O, timer expires, etc.

 

OS modules, interfaces, and structures

Processes

1. Processes

1. What is a process?

The OS abstraction for execution.

2. What is the difference between a process and a program?

3. What is contained in a process?

2. Process data structures

1. Process Control Blocks (PCBs)

1. What information does it contain?

2. How is it used in a context switch?

2. State queues

1. What are process states?

2. What is the process state graph?

3. When does a process change state?

4. How does the OS use queues to keep track of processes?

3. Process manipulation

1. What does fork() on Unix do?

1. What does it mean for it to “return twice”?

2. What does exec() on Unix do?

1. How is it different from fork?

3. How are fork and exec used to implement shells?

 

Threads

1. Threads

1. What is a thread?

1. What is the difference between a thread and a process?

2. How are they related?

2. Why are threads useful?

3. What is the difference between user-level and kernel-level threads?

1. What are the advantages/disadvantages of one over another?

2. Thread implementation

1. How are threads managed by the run-time system?

1. Thread control blocks, thread queues

2. How is this different from process management?

2. What operations do threads support?

1. Fork, yield, sleep, etc.

2. What does thread yield do?

3. What is a context switch?

4. What is the difference between non-preemptive scheduling and preemptive thread scheduling?

1. Voluntary and involuntary context switches

 

Synchronization

1. Synchronization

1. Why do we need synchronization?

1. Coordinate access to shared data structures

2. Coordinate thread/process execution

2. What can happen to shared data structures if synchronization is not used?

1. Race condition

2. Corruption

3. Bank account example

3. When are resources shared?

1. Global variables, static objects

2. Heap objects

2. Mutual exclusion

1. What is mutual exclusion?

2. What is a critical section?

1. What guarantees do critical sections provide?

2. What are the requirements of critical sections?

1. Mutual exclusion (safety)

2. Progress (liveness)

3. Bounded waiting (no starvation: liveness)

4. Performance

3. How does mutual exclusion relate to critical sections?

4. What are the mechanisms for building critical sections?

1. Locks, semaphores, monitors, condition variables

3. Locks

1. What does acquire do?

2. What does release do?

3. What does it mean for acquire/release to be atomic?

4. How can locks be implemented?

1. Spinlocks

2. Disable/enable interrupts

3. Blocking (Nachos)

5. How does test-and-set/swap work?

1. What kind of lock does it implement?

6. What are the limitations of using spinlocks, interrupts?

1. Inefficient, interrupts turned off too long

4. Semaphores

1. What is a semaphore?

1. What does P/Decrement do?

2. What does V/Increment do?

3. How does a semaphore differ from a lock?

4. What is the difference between a binary semaphore and a counting semaphore?

2. When do threads block on semaphores?

3. When are they woken up again?

4. Using semaphores to solve synchronization problems

1. Readers/writers problem

2. Bounded buffers problem

5. Monitors

1. What is a monitor?

1. Shared data

2. Procedures

3. Synchronization

2. In what way does a monitor provide mutual exclusion?

1. To what extent is it provided?

3. How does a monitor differ from a semaphore?

4. How does a monitor differ from a lock?

5. What kind of support do monitors require?

1. Language, run-time support

6. Condition variables

1. What is a condition variable used for?

1. Coordinating the execution of threads

2. Not mutual exclusion

2. Operations

1. What are the semantics of Wait?

2. What are the semantics of Signal?

3. What are the semantics of Broadcast?

3. How are condition variables different from semaphores?

7. Implementing monitors

1. What does the implementation of a monitor look like?

1. Shared data

2. Procedures

3. A lock for mutual exclusion to procedures (w/ a queue)

4. Queues for the condition variables

2. What is the difference between Hoare and Mesa monitors?

1. Semantics of signal (whether the woken up waiter gets to run immediately or not)

2. What are their tradeoffs?

3. What does Java provide?

8. Locks and condition vars

1. In nachos, we don't have monitors

2. But we want to be able to use condition variables

3. So we isolate condition variables and make them independent (not associated with a monitor)

4. Instead, we have to associate them with a lock (mutex)

5. Now, to use a condition variable …

1. Threads must first acquire the lock (mutex)

2. CV::Wait releases the lock before blocking, acquires it after waking up

 

Scheduling

1. Scheduling

1. Components

1. Scheduler (dispatcher)

2. When does scheduling happen?

1. Job changes state (e.g., waiting to running)

2. Interrupt, exception

3. Job creation, termination

2. Scheduling goals

1. Goals

1. Maximaize CPU utilization

2. Maximize job throughput

3. Minimize ternaround time

4. Minimize waiting time

5. Minimize response time

2. What is the goal of a batch system?

3. What is the goal of an interactive system?

3. Starvation

1. Starvation

1. Indefinite denial of a resource (CPU, lock)

2. Causes

1. Side effect of scheduling

2. Side effect of synchronization

3. Operating systems try to prevent starvation

4. Scheduling algorithms

1. What are the properties advantages and disadvantages of the following scheduling algorithms?

1. First Come First Serve (FCFS)/First In First Out (FIFO)

2. Shortest Job first (SJF)/shortest remaining time first

3. Priorty

4. Round Robin

5. Multilevel feedback queues

2. What scheduling algorithm does Unix use? Why?

5. Some quick clarification

1. Round Robin schedule

1. If the time slice is 10ms, it means that a thread can at most use the CPU for 10 ms a time

2. If a thread gives up the CPU because it has to wait for I/O, condition variable, etc, the scheduler will switch to another thread in the ready queue

6. Deadlocks

1. Deadlock conditions

2. Deadlock detection

3. Deadlock prevention

4. How do today’s systems handle deadlocks? Why?