1. 1 Threads, SMP, and Microkernels Chapter 4

2. 2 Process Resource ownership: process includes a virtual address space to hold the process image (fig 3.16) Scheduling/execution: follows an execution path that may be interleaved with other processes These two characteristics can be treated independently by the operating system

3. 3 Process Unit of dispatching is referred to as a thread or lightweight process Unit of resource ownership is referred to as a process or task

4. 4 Multithreading Operating system supports multiple threads of execution within a single process MS-DOS supports a single thread UNIX supports multiple user processes but only supports one thread per process Windows, Solaris, Linux, Mach, and OS/2 support multiple threads

5. 5

6. 6 Process Have a virtual address space which holds the process image Protected access to processors, other processes, files, and I/O resources

7. 7 Thread An execution state (running, ready, etc.) Saved thread context when not running Has an execution stack Some per-thread static storage for local variables Access to the memory and resources of its process  all threads of a process share this

8. 8

9. 9 Benefits of Threads Takes less time to create a new thread than a process Less time to terminate a thread than a process Less time to switch between two threads within the same process Since threads within the same process share memory and files, they can communicate with each other without invoking the kernel

10. 10 Uses of Threads in a Single- User Multiprocessing System Foreground to background work Asynchronous processing Speed of execution Modular program structure

11. 11 Threads Suspending a process involves suspending all threads of the process since all threads share the same address space Termination of a process, terminates all threads within the process

12. 12 Thread States States associated with a change in thread state  Spawn Spawn another thread  Block  Unblock  Finish Deallocate register context and stacks

13. 13 Remote Procedure Call Using Single Thread

14. 14 Remote Procedure Call Using Threads

15. 15 Multithreading

16. 16 Adobe PageMaker

17. 17 User-Level Threads All thread management is done by the application The kernel is not aware of the existence of threads

18. 18 User-Level Threads

19. 19

20. 20 Advantages of User-Level Threads Thread-switching does not require kernel mode privileges and switching Thread scheduling can be tailored to application ULTs can run on any OS. They can be implemented with a thread library that runs within the app.

21. 21 Disadvantages of ULTs If blocking system call (eg IO call), whole process blocked  can use ‘jacketing’ – call other process to check state of resource Cannot use multiprocessing since OS sees all threads as a single process.

22. 22 Kernel-Level Threads Windows is an example of this approach Kernel maintains context information for the process and the threads Scheduling is done on a thread basis

23. 23 Kernel-Level Threads

24. 24 VAX Running UNIX-Like Operating System

25. 25 Combined Approaches Example is Solaris Thread creation done in the user space Bulk of scheduling and synchronization of threads within application

26. 26 Combined Approaches

27. 27 Relationship Between Threads and Processes 1:M and M:N approaches are experimental.

28. 28 Symmetric Multiprocessing (SMP)

29. 29 Symmetric Multiprocessing Tradition: Single sequence of instructions  von Neumann machine concept Real machines do instruction pipelining, DMA and other parallel operations  Pipeline explanation and movieexplanation movie  Hyperthreading Hyperthreading With cheaper HW can extend this to multiple processors

30. 30 Categories of Computer Systems Single Instruction Single Data (SISD) stream  Single processor executes a single instruction stream to operate on data stored in a single memory.  ‘Normal’ computing Single Instruction Multiple Data (SIMD) stream  Each instruction is executed on a different set of data by the different processors.  Eg DSP or vector processing

31. 31 Categories of Computer Systems Multiple Instruction Single Data (MISD) stream  A sequence of data is transmitted to a set of processors, each of which executes a different instruction sequence.  Never implemented! Multiple Instruction Multiple Data (MIMD)  A set of processors simultaneously execute different instruction sequences on different data sets

32. 32

33. 33 Symmetric Multiprocessing Kernel can execute on any processor Typically each processor does self- scheduling from the pool of available process or threads

34. 34

35. 35 Multiprocessor Operating System Design Considerations Simultaneous concurrent processes or threads  Re-entrant kernel code, deadlock, etc. Scheduling (see ch 10 for more detail) Synchronization  Mutual exclusion, event ordering, locks etc (ch 5) Memory management  Coordination in paging systems when multiple processors accessing pages (ch 7) Reliability and fault tolerance  Graceful degradation on processor failure

36. 36 Distributed Processing Another alternative is distributed processing Each node has its own CPU, memory and communication channel  Slower communications, more independence EG. Cluster computing, Beowulf etc.  See Ch 14, 15

37. 37 Microkernels

38. 38 Microkernels Small operating system core (how small?) Concept: Contains only essential core operating systems functions Many services traditionally included in the operating system are now external subsystems  Device drivers  File systems  Virtual memory manager  Windowing system  Security services

39. 39

40. 40 Benefits of a Microkernel Organization Uniform interface on request made by a process  Don’t distinguish between kernel-level and user-level services  All services are provided by means of message passing Extensibility  Allows the addition of new services Flexibility  New features added  Existing features can be subtracted

41. 41 Benefits of a Microkernel Organization Portability  Changes needed to port the system to a new processor is changed in the microkernel - not in the other services Reliability  Modular design  Small microkernel can be rigorously tested Exhaustive testing vs. module testing

42. 42 Benefits of Microkernel Organization Distributed system support  Message are sent without knowing what the target machine is Object-oriented operating system  Components are objects with clearly defined interfaces that can be interconnected to form software

43. 43 Microkernel Design Low-level memory management  Mapping each virtual page to a physical page frame

44. 44 Microkernel Design Interprocess communication I/O and interrupt management

45. 45 MS-Windows Processes Implemented as objects An executable process may contain one or more threads Both processes and thread objects have built-in synchronization capabilities

46. 46

47. 47 Windows Process Object

48. 48 Windows Thread Object

49. 49 MS-Windows Thread States Ready Standby Running Waiting Transition Terminated

50. 50

51. 51 Solaris Process includes the user’s address space, stack, and process control block User-level threads Lightweight processes (LWP) Kernel threads

52. 52

53. 53

54. 54 Solaris Lightweight Data Structure Identifier Priority Signal mask Saved values of user-level registers Kernel stack Resource usage and profiling data Pointer to the corresponding kernel thread Pointer to the process structure

55. 55

56. 56 Linux Task Data Structure State Scheduling information Identifiers Interprocess communication Links Times and timers File system Address space Processor-specific context

57. 57 Linux States of a Process Running Interruptible Uninterruptible Stopped Zombie

58. 58

59. 59 Linux Threads Traditionally no threads  ULTs with pthread library Modern Linux supports KLT Process = thread (cloning)