Operating System Principles

1. Operating System Principles Ku-Yaw Chang canseco@mail.dyu.edu.tw Assistant Professor, Department of Computer Science and Information Engineering Da-Yeh University

2. Chapter 8 Memory-Management Strategies • Keep several processes in memory to increase CPU utilization • Memory management algorithms • Require hardware support • Common strategies • Paging • Segmentation Chapter 8 Memory-Management Strategies

3. Background Swapping Contiguous Memory Allocation Paging Structure of the Page Table Segmentation Example: The Intel Pentium Summary Exercises Chapter 8 Memory-Management Strategies Chapter 8 Memory-Management Strategies

4. 8.1 Background • Program must be brought (loaded) into memory and placed within a process for it to be run. • Address binding • A mapping from one address space to another • A typical instruction-execution cycle • Fetch an instruction from memory • Decode the instruction • May cause operands to be fetched from memory • Execute the instruction • Store results back into memory Chapter 8 Memory-Management Strategies

5. 8.1.1 Basic Hardware • CPU can directly access • Registers built into the processor • Generally within one cycle of the CPU clock • Main memory • May take many cycles of the CPU clock to complete • The processor normally needs to stall • Cache • A memory buffer used to accommodate a speed differential Chapter 8 Memory-Management Strategies

6. 8.1.1 Basic Hardware • Protection must be provided by the hardware • Protect the OS from access by user processes • Protect user processes fromone another • One possible implementation • Base register • Hold the smallest legal physical memory address • Limit register • Specify the size of the range Chapter 8 Memory-Management Strategies

7. 8.1.1 Basic Hardware • The OS is given unrestricted access to both operating system and users’ memory Chapter 8 Memory-Management Strategies

8. 8.1.2 Address Binding • Input queue • A collection of processes on the disk that are waiting to be brought into memory to run the program. • A user program will go through several steps before being executed • Addresses in source program are symbolic • A compiler binds these symbolic addresses to relocatable addresses • A loader binds these relocatable addresses to absolute addresses Chapter 8 Memory-Management Strategies

9. 8.1.2 Address Binding • Compile time • Absolute code can be generated • Know at compile time where the process will reside in memory • MS-DOS .COM-format programs are absolute code Chapter 8 Memory-Management Strategies

10. 8.1.2 Address Binding • Load time • Relocatable code can be generated • Not known at compile time where the process will reside in memory • Final binding is delayed until load time Chapter 8 Memory-Management Strategies

11. 8.1.2 Address Binding • Execution time • The process can be moved from one memory segment to another • Binding must be delayed until run time • Special hardware must be available Chapter 8 Memory-Management Strategies

12. 8.1.3 Logical- Versus Physical-Address Space • Logical address • An address generated by the CPU • Compile-time and load-time • Also called virtual address • Logical-address space • The set of all logical addresses • Physical address • An address seen by the memory unit • The one loaded into the memory-address unit • Execution time • Logical and physical address spaces differ • Physical-address space Chapter 8 Memory-Management Strategies

13. 8.1.3 Logical- Versus Physical-Address Space • Memory-Management Unit (MMU) • A hardware device • Run-time mapping from virtual to physical addresses • Different methods to accomplish such a mapping • Logical addresses • 0 to max • Physical addresses • R + 0 to R + max • The user program • Deal with logical addresses • Never see the real physical addresses Chapter 8 Memory-Management Strategies

14. Dynamic RelocationUsing a Relocation Register Chapter 8 Memory-Management Strategies

15. 8.1.4 Dynamic Loading • The entire program and data must be in memory for its execution • The size of a process is limited to the size of physical memory. • Dynamic Loading • All routines are kept on disk in a relocatable load format • The main program is loaded into memory and is executed • A routine is not loaded until it is called • Advantage • An unused routine is never loaded Chapter 8 Memory-Management Strategies

16. 8.1.5 Dynamic Linking andShared Libraries • Dynamic Linking • Linking is postponed until execution time • Small piece of code, called stub, used to locate the appropriate memory-resident library routine • OS checks if routine is in processes’ memory address • If yes, load the routine into memory • Stub replaces itself with the address of the routine, and executes the routine. • Dynamic linking is particularly useful for libraries • Library updates Chapter 8 Memory-Management Strategies

17. Background Swapping Contiguous Memory Allocation Paging Structure of the Page Table Segmentation Example: The Intel Pentium Summary Exercises Chapter 8 Memory-Management Strategies Chapter 8 Memory-Management Strategies

18. Swapping • A process can be • Swapped temporarily out of memory to a backing store • Commonly a fast disk • Brought back into memory for continued execution • A process swapped back into • The same memory space • Binding is done at assembly or load time • A different memory space • Execution-time binding Chapter 8 Memory-Management Strategies

19. Swapping of two processes Chapter 8 Memory-Management Strategies

20. Swapping • Context-switch time is fairly high • User process size: 10MB • Transfer rate: 40MB per second • Actual transfer: • 10000KB / 40000 KB per second = 250 milliseconds • An average latency: 8 ms • Total swap time • 258 + 258 = 516 ms • Time quantum should be substantially larger than 0.516 seconds. Chapter 8 Memory-Management Strategies

21. Swapping • Major part of the swap time is transfer time • Directly proportional to the amount of memory swapped • Factors • How much memory is actually used • To reduce swap time • Be sure the process is completely idle • Pending I/O • Never swap a process with pending I/O • Execute I/O operations only into operating-system buffers Chapter 8 Memory-Management Strategies

22. Background Swapping Contiguous Memory Allocation Paging Structure of the Page Table Segmentation Example: The Intel Pentium Summary Exercises Chapter 8 Memory-Management Strategies Chapter 8 Memory-Management Strategies

23. 8.3 Contiguous Memory Allocation • Memory • One for the resident operating system • In either low or high memory • Location of the interrupt vector • One for the user processes • Contiguous memory allocation • Each process is contained in a single contiguous section of memory • For efficiency purpose Chapter 8 Memory-Management Strategies

24. 8.3.1 Memory Mapping and Protection • Hardware support • A relocation register • The value of the smallest physical address • A limit register • The range of logical addresses Chapter 8 Memory-Management Strategies

25. 8.3.2 Memory Allocation • Fixed-sized partitions • Simplest • Each partition contain exactly one process • Degree of multiprogramming is bounded by the number of partitions • Strategies • First fit (faster) • Best fit (better storage utilization) • Worst fit Chapter 8 Memory-Management Strategies

26. 8.3.2 Memory Allocation • Multiple-partition allocation • Hole– block of available memory; holes of various size are scattered throughout memory • When a process arrives, it is allocated memory from a hole large enough to accommodate it • Operating system maintains information about:a) allocated partitions b) free partitions (hole) OS OS OS OS process 5 process 5 process 5 process 5 process 9 process 9 process 8 process 10 process 2 process 2 process 2 process 2 Chapter 8 Memory-Management Strategies

27. 8.3.2 Memory Allocation • First fit • Allocate the first hole that is big enough • Best fit • Allocate the smallest hole that is big enough • Must search entire list, unless ordered by size • Produce the smallest leftover hole • Worst fit • Allocate the largest hole • Must search entire list • Produce the largest leftover hole Chapter 8 Memory-Management Strategies

28. 8.3.3 Fragmentation • External Fragmentation • Enough total memory space to satisfy a request, but not contiguous • Reduced by compaction • Shuffle memory contents to place all free memory together in one large block • Compaction is possible only if relocation is dynamic, and is done at execution time • Internal Fragmentation • Allocated memory may be slightly larger than requested memory • This size difference is memory internal to a partition, but not being used Chapter 8 Memory-Management Strategies

29. Background Swapping Contiguous Memory Allocation Paging Structure of the Page Table Segmentation Example: The Intel Pentium Summary Exercises Chapter 8 Memory-Management Strategies Chapter 8 Memory-Management Strategies

30. 8.4 Paging • A memory-management scheme that permits the physical-address space of a process to be noncontiguous • Avoid the problem of fitting memory chunks of varying sizes onto the backing store Chapter 8 Memory-Management Strategies

31. 8.4.1 Basic Method • Frames (physical memory) • Fixed-sized blocks • Pages (logical memory) • Blocks of the same size • Every address generated by the CPU is divided into • A page number (p) • An index to a page table • contain the base address of each page in the physical memory • A page offset (d) • The displacement with the page Chapter 8 Memory-Management Strategies

32. 8.4.1 Basic Method • Page size • Defined by the hardware • Typically a power of 2 • Varying between 512 bytes to 16MB per page • Logical address space is 2m • A page size is 2n addressing units • High-order m-n bits designate the page number • n low-order bits designates the page offset page number page offset p d m-n n Chapter 8 Memory-Management Strategies

33. Paging hardware Chapter 8 Memory-Management Strategies

34. Paging Model Chapter 8 Memory-Management Strategies

35. Paging Example • Page size of 4 bytes • Physical memory of 32 bytes • 8 pages • Paging itself is a form of dynamic relocation • No external fragmentation • Have internal fragmentation Chapter 8 Memory-Management Strategies

36. 8.4.1 Basic Method • Internal fragmentation • Average one-half page per process • Suggest that small page sizes are desirable • Overhead is involved in each page-table entry • Disk I/O is more efficient when the data being transferred is larger • Generally, page sizes have grown over time • Processes, data sets and main memory have become larger • Between 4KB and 8KB in size • Some CPUs and kernels support multiple page sizes Chapter 8 Memory-Management Strategies

37. Free Frames After Allocation Before Allocation Chapter 8 Memory-Management Strategies

38. 8.4.2 Hardware Support • OS • Allocate a page table for each process • A pointer to the page table is stored in the PCB • Hardware implementation • The page table is implemented as a set of dedicated registers • Efficiency is a major consideration • Satisfactory if the page table is reasonably small • Page-table base register (PTBR) • Point to the page table • The page table is kept in main memory • Problem: two memory access are needed to access a byte • One for the page table entry • One for the byte Chapter 8 Memory-Management Strategies

39. 8.4.2 Hardware Support • Translation look-aside buffer (TLB) • A special, small, fast-lookup hardware cache • The search is fast • Be used with page tales • Each entry consists of two parts: a key and a value • Number of entries: between 64 and 1,024 • A TLB miss • If the page number is not in the TLB • Entries can be wired down • Cannot be removed from the TLB Chapter 8 Memory-Management Strategies

40. Paging Hardware with TLB Chapter 8 Memory-Management Strategies

41. 8.4.2 Hardware Support • Hit ratio • The percentage of times that a particular page number is found in the TLB • Example: • Hit ratio: 80% • 20 nanoseconds to search the TLB • 100 nanoseconds to access memory • If the page is in the TLB: 120 nanoseconds • If not, 220 nanoseconds • Effective access time = 0.8 * 120 + 0.2 * 220 = 140 nanoseconds Chapter 8 Memory-Management Strategies

42. 8.4.3 Protection • Memory protection • Accomplished by protection bits associated with each frame • One bit to define a page to be read-write or read-only • An attempt to write to a read-only page causes a hardware trap to the OS • One bit – valid-invalid bit • OS sets this bit to allow or disallow access to the page Chapter 8 Memory-Management Strategies

43. 8.4.3 Protection Chapter 8 Memory-Management Strategies

44. 8.4.4 Shared Pages • An advantage of paging • Sharing common code • Only one copy of the code need be kept in physical memory • Example • Compilers, window systems, run-time libraries, database systems, and so on. • Reentrant code (or pure code) • Non-self-modifying code • Never change during execution Chapter 8 Memory-Management Strategies

45. Sharing of codein a paging environment Chapter 8 Memory-Management Strategies

46. Background Swapping Contiguous Memory Allocation Paging Structure of the Page Table Segmentation Example: The Intel Pentium Summary Exercises Chapter 8 Memory-Management Strategies Chapter 8 Memory-Management Strategies

47. Be skipped Chapter 8 Memory-Management Strategies

48. Background Swapping Contiguous Memory Allocation Paging Structure of the Page Table Segmentation Example: The Intel Pentium Summary Exercises Chapter 8 Memory-Management Strategies Chapter 8 Memory-Management Strategies

49. 8.6 Segmentation • Paging • The separation of the user’s view of memory and the actual physical memory Chapter 8 Memory-Management Strategies

50. 8.6.1 Basic Method • User’s view of memory • A collection of variable-sized segments, with no necessary ordering among segments Chapter 8 Memory-Management Strategies