Embedded Operating Systems - II

Transcription

1 Embedded Operating Systems - II 馮立琪 lcfeng@mail.cgu.edu.tw 長庚大學資訊工程系計算機系統實驗室

2 Agenda A Brief History of Operating Systems Defining an RTOS Defining a Task Task States and Scheduling Typical Task Operations The Scheduler Objects Services Key Characteristics of an RTOS 計算機系統實驗室 2

3 A Brief History of Operating Systems In the early days, developers created software applications that included lowlevel machine code To initialize and interact with the hardware Result in no-portable, bug-prone applications Operating systems thus provide the basic software foundation Facilitate the abstraction of the underlying hardware from the application code 計算機系統實驗室 3

4 A Brief History of Operating Systems (Cont.) The evolution of operating systems helped shift the design of software applications from large, monolithic applications to more modular, interconnected applications that could run on top of the OS environment. 計算機系統實驗室 4

5 A Brief History of Operating Systems (Cont.) General-purpose operating systems (GPOS) UNIX In the 60s and 70s, multi-user access to expensive mid-sized and mainframe computing system Eventually, ported to all types of machines Microsoft Windows Drive the personal-computing era Real-time operating systems For post-pc, embedded-computing era VxWork, uclinux 計算機系統實驗室 5

6 Functional Similarities Between a RTOS and GPOS Abstracting the hardware from the software application Software and hardware resource management Provision of underlying OS services to applications Some level of multitasking 計算機系統實驗室 6

7 Functional Differences Between a RTOS and GPOS Better reliability in embedded application contexts The ability to scale up and down to meeting application needs Faster performance Reduced memory requirements Scheduling policies tailored for real-time embedded systems Support for diskless embedded systems by allowing executables to boot and run from ROM or RAM Better portability to different hardware platforms 計算機系統實驗室 7

8 GPOSes vs. RTOSes GPOSes typically require a lot more memory are not well suited to real-time embedded devices with limited memory and high performance requirements RTOSes Reliable, compact, scalable, and perform well in real-time embedded systems Can be easily tailored to use only those components required for a particular applications 計算機系統實驗室 8

9 Defining an RTOS RTOS is a program that Schedules execution in a timely manner Manage system resources Provide a consistent foundation for developing application code A good RTOS should be scalable To meeting different requirements for different applications 計算機系統實驗室 9

10 Defining an RTOS (Cont.) In some applications, an RTOS comprises only a kernel The core supervisory software that provides minimal logic, scheduling, and resource-management algorithms. An RTOS can be a combination of various modules Kernel, File system, Networking protocol stacks, Other components required for a particular application 計算機系統實驗室 10

11 High-Level View of an RTOS 計算機系統實驗室 11

12 RTOS Kernel Most RTOS kernels contain the following components Scheduler Algorithms that determine which task executes Objects Round-robin, preemptive scheduling Special kernel construct that helps developer create applications Tasks, semaphores, and message queues Services Operations that the kernel performs on an object Generally operations such as timing, interrupt handling, and resource management 計算機系統實驗室 12

13 Common Components in an RTOS kernel 計算機系統實驗室 13

14 The Scheduler At the heart of every kernel Provides the algorithms to determine which task executes when Related topics Schedulable entities Multitasking Context switching Dispatcher Scheduling algorithms 計算機系統實驗室 14

15 Schedulable Entities A kernel object that can compete for execution time on a system Task An independent thread of execution that contains a sequence of independently schedulable instructions Processes differ from tasks in that They provide better memory protection features, at the expense of performance and memory overhead. 計算機系統實驗室 15

16 Multitasking The ability of the OS to handle multiple activities within set deadlines Many threads of execution appear to be running concurrently However, they are actually interleaved executions sequentially 計算機系統實驗室 16

17 Multitasking Using a Context Switch 計算機系統實驗室 17

18 The Context Switch Each task has its own context the state of the CPU registers required for tasks running When a task running, its context is highly dynamic Context switch Occurred when a scheduler switches from one task to another TCB: Task Control Block A data structure used by the kernel to maintain task-specific information The dynamic context of a task is stored in its TCB 計算機系統實驗室 18

19 The Context Switch (Cont.) Steps in context switches The kernel saves task 1 s context information in its TCB It loads task 2 s context information from its TCB, which becomes the current thread of execution The context of task 1 is frozen while task 2 executes If the scheduler needs to run task 1 again, task 1 continues from where it left off just before context switch 計算機系統實驗室 19

20 The Dispatcher The part of scheduler that performs context switching and changes the flow of execution Three areas that is passed through by the flow of execution (or flow of control) Through an application task Through an ISR Through the kernel 計算機系統實驗室 20

21 The Dispatcher (Cont.) A task or ISR makes a system call Flow of control passes to the kernel to execute one of system routines provided by the kernel When leaving the kernel Dispatcher passes control to one of the tasks in user s application Notably, not necessary the original task that made the system call Depend on the scheduling algorithms 計算機系統實驗室 21

22 Defining a Task Task An independent thread of execution Compete with other concurrent tasks for processor execution time An application can be decomposed into multiple concurrent tasks 計算機系統實驗室 22

23 Task Defined by its set of parameters and supporting data structures Associated name Unique ID Priority Task control block Stack Task routine 計算機系統實驗室 23

24 A Task, its Associated Parameters, and Supporting Data Structures 計算機系統實驗室 24

25 System Tasks Created when the kernel first starts The parameters are assigned from a set of reserved priority levels Used internally by the RTOS User applications should avoid using these priority levels 計算機系統實驗室 25

26 System Tasks (Cont.) Examples Initialization or startup task Initialize the system and create and start system tasks Idle task Used up processor idle cycles when no other activity is present Logging task Logs system messages Exception-handling task Handle exceptions Debug agent task Allow debug with a host debugger 計算機系統實驗室 26

27 Idle Task Set to the lowest priority and execute in an endless loop Run when either no other task can run or when no other tasks exist CPU needs to be continued running unless it can be suspended Idle task may be replaced by a userconfigured routine to implement special function Power conservation code that suspends system after a period of idle time 計算機系統實驗室 27

28 Create a Task Developer must assign a new task via kernel APIs Task name Priority Stack size Task routine Then, the kernel Assign each task a unique ID Create an associated TCB and stack space in memory 計算機系統實驗室 28

29 Task States and Scheduling Ready state The task is ready to run but cannot because a higher priority task is executing Blocked state Request a resource but is not yet available Request to wait until some event occurs Delay itself for some duration Running state The task is currently the highest priority task and is running 計算機系統實驗室 29

30 A Typical Finite State Machine for Task Execution States 計算機系統實驗室 30

31 Ready State When a task is first created Kernel put it into the ready state Cannot move to the blocking state A task first needs to run so it can make a blocking call Can only move to the running state 計算機系統實驗室 31

32 Ready State (Cont.) Two implementations Only one task per priority level The highest priority task that is ready runs More than one tasks per priority levles Need a task-ready list 計算機系統實驗室 32

33 Running State A running task can move to the blocked states in Making a call that requests an unavailable resource Making a call that requests to wait for an event to occur Making a call to delay the task for some duration 計算機系統實驗室 33

34 Blocked State The possibility of blocked states is important in a system Lower priority tasks have chances to run Examples of unblocking conditions A semaphore token for which a task is waiting is released A message, on which the task is waiting, arrives in a message queue A time delay imposed on the task expires 計算機系統實驗室 34

35 Typical Task Operations Kernel must provide task-management services The actions that a kernel performs to support task Create and maintain the TCB and task stacks The API that a kernel provides for developer to manipulate tasks Create and delete tasks Control task scheduling Obtain task information 計算機系統實驗室 35

36 Scheduling Algorithms Also called scheduling policy Two common scheduling algorithms Preemptive priority-based scheduling Round-robin scheduling Developers can create and define their own scheduling algorithms 計算機系統實驗室 36

37 Preemptive Priority-Based Scheduling Most real-time kernels use preemptive priority-based scheduling by default Real-time kernel generally support 256 priority levels Priority assignment Static: assign priority to tasks when created Dynamic: priority can be changed dynamically 計算機系統實驗室 37

38 Preemptive Priority-Based Scheduling 計算機系統實驗室 38

39 Preemptive Priority-Based Scheduling (Cont.) The ability to change task priorities dynamically allows an embedded application the flexibility to adjust to external events as they occur creating a true real-time, responsive system However, the misuse of this capability can lead to priority inversion, deadlock, and eventual system failure. 計算機系統實驗室 39

40 Priority Inversion Priority inversion is a situation in which a low-priority task executes while a higher priority task waits on it due to resource contentions 計算機系統實驗室 40

41 Resource Access Control Protocol A resource access control protocol is a set of rules that defines the conditions under which a resource can be granted to a requesting task and governs the execution scheduling property of the task holding the resource 計算機系統實驗室 41

42 Priority Inheritance Protocol Priority Inheritance Protocol rules. when a task T requests a resource R Rule # Description 1 If R is in use, T is blocked. 2 If R is free, R is allocated to T. 3 When a task of a higher priority requests the same resource, T's execution priority is raised to the requesting task's priority level. 4 The task returns to its previous priority when it releases R. 計算機系統實驗室 42

43 Priority Inheritance Protocol 計算機系統實驗室 43

44 Transitive priority promotion example 計算機系統實驗室 44

45 Round-Robin Scheduling Provide each task an equal share of the CPU execution time Pure RR (Round-Robin) scheduling cannot satisfy real-time system requirements Preemptive, priority-based scheduling can be augmented with round-robin scheduling Equal allocation of CPU time for tasks of the same priority 計算機系統實驗室 45

46 Round-Robin Scheduling (Cont.) If a task in a round-robin cycle is preempted by a higher-priority task its time quantum left is saved and then restored when the interrupted task is again eligible for execution. 計算機系統實驗室 46

47 Round-Robin and Preemptive Scheduling 計算機系統實驗室 47

48 Objects Kernel objects are special constructs that are the building blocks for application development Tasks concurrent and independent threads of execution that can compete for CPU execution time Semaphores token-like objects that can be incremented or decremented by tasks for synchronization or mutual exclusion Message Queues buffer-like data structures that can be used for synchronization, mutual exclusion, and data exchange by passing messages between tasks 計算機系統實驗室 48

49 Services Along with objects, most kernels provide services that help developers create applications for real-time embedded systems. Sets of API calls that can be used to Perform operations on kernel objects Facilitate Timer management Interrupt handling Device I/O Memory management 計算機系統實驗室 49

50 Key Characteristics of an RTOS An application's requirements define the requirements of its underlying RTOS. Some of the more common attributes are Reliability Predictability Performance Compactness Scalability 計算機系統實驗室 50

51 Reliability A reliable system Be available (continue to provide service) and does not fail Different degrees of reliability might be acceptable Quantified by the downtime per year Number of 9s: the percent of the total time that a system must be available To determine system reliability The combination of all system elements determines the reliability of a system Hardware, BSP, RTOS, and application 計算機系統實驗室 51

52 Categorizing Highly Available Systems by Allowable Downtime 計算機系統實驗室 52

53 Predictability Many embedded systems are also real-time systems meeting time requirements is key to ensuring proper operation For real-time systems, the RTOS needs to be predictable The completion of operating system calls occurs within known timeframes The variance of the response times for each type of system call should be small 計算機系統實驗室 53

54 Performance An embedded system must perform fast enough to fulfill its timing requirements Processor s performance MIPS: million instructions per second Throughput may be used to measure the overall performance of a system the rate at which a system can generate output based on the inputs coming in Call-by-call method may be used to measure RTOS performance Produce timestamps when a system call starts and when it completes 計算機系統實驗室 54

55 Compactness To determine how compact an embedded system be Application design constraints, Cost constraints For example, a cell phone must be small, portable and low cost Limit system memory Limit the size of application and operating system To meet total system requirements, designers must understand both the static and dynamic memory consumption of the RTOS and the application that will run on it. 計算機系統實驗室 55

56 Scalability RTOSes can be used in a wide variety of embedded systems Be able to scale up or down to meet applicationspecific requirements An RTOS should be capable of adding or deleting modular components, including file systems and protocol stacks. Depending on how much functionality is required For example, an RTOS may be used in both a cellular phone project and a base station project To save time and money 計算機系統實驗室 56