Operating Systems

Nguyen Tri Thanh

[email protected]

1
 Question
What is incorrect about overlays?
A. overlays allows a large program to run in a
smaller MEM
B. Overlays only loads codes on demand (when
they are used)
C. Programmers need to split the program into
modules
D. Overlays is supported in all high level
programming languages

2
 Question
What is incorrect about swapping?
A. swapping is the same as overlays
B. swapping uses hard disk as the backing store
C. swapping allows many processes whose size
is even larger than MEM to run
D. a lower priority process is rolled out for a
higher priority one to run (when needed)

3
 Review
Which is incorrect about non-contiguous
MEM allocation?
A. split logical memory into parts
B. utilize MEM more effectively in comparison
with contiguous allocation method
C. need a Memory Management Unit
D. only suitable for some types of processes

4
 Review
Which is correct about MMU of paging and
segmentation allocation methods?
A. they are the same
B. MMU of paging needs more information than
that of segmentation
C. they use different resolution methods
D. MMU of segmentation is faster than that of
paging

5
 Question
Suppose a process in contiguous allocation:
 the base address is 10400
 the limit register is 1200
 the reference is 246;
Which of the following is the correct physical
address of the reference ?
A. 10154
B. 10646
C. 1446
D. 954
6
 Question
A system uses paging
 the frame size of 2KB;
 the address register is 32 bits
Which of the following is correct about register
segmentation?
A. (page:offset) = (19:13)
B. (page:offset) = (21:11)
C. (page:offset) = (22:10)
D. (page:offset) = (20:12)

7
 Question
Frame
A system uses paging 56
 the frame size of 4KB; 120

 the address register is 32 bits 3

 Which of the following is the correct physical

address of the reference (2,1296)?
A. 560*4096+1296
B. 120*4096+1296
C. 3*4096+1296
D. 120*1024+1296

8
Virtual Memory
Paging on demand
Page replacement
Frame allocation
Thrashing

9
Objectives
 Introduce paging method
 Introduce segmentation method

10
Reference
 Chapter 9 of Operating System Concepts

11
Virtual memory

12
 Virtual memory
 Separation of user logical memory from
physical memory.
 Only a part of the program needs to be in memory
for execution
 Logical address space can therefore be much
larger than physical address space
 Allows address spaces to be shared by several
processes
 Allows for more efficient process creation
 Virtual memory can be implemented via
 Paging on demand
13
 Segmentation on demand
Virtual memory
 Linux is, of course, a virtual memory system, meaning that
the addresses seen by user programs do not directly
correspond to the physical addresses used by the
hardware. Virtual memory introduces a layer of indirection that
allows a number of nice things. With virtual memory,
programs running on the system can allocate far more
memory than is physically available; indeed, even a single
process can have a virtual address space larger than the
system's physical memory. Virtual memory also allows the
program to play a number of tricks with the
process's address space, including mapping the program's
memory to device memory.
 http://www.makelinux.net/ldd3/chp-15-sect-1
14
Virtual-address Space

15
Virtual-address Space

16
Virtual Memory That is Larger Than
Physical Memory

17
Shared Library Using Virtual
Memory

18
 Process Creation
 Virtual memory allows other benefits
 Copy-on-Write during process creation
 Memory-Mapped Files

19
 Copy-on-Write
 Copy-on-Write (COW) allows both parent and
child processes to initially share the same
pages in memory
 If either process modifies a shared page, only
then is the page copied
 refer to Sect. 2, Chapter 3 of “Lập trình C/C++ …”
 COW allows more efficient process creation
as only modified pages are copied

20
 Memory Mapped Files
 A file is considered as a memory segment
 Read/write operations are performed via memory
 not read/write file system calls
 Allow multiple processes shared a file
 refer to Sect. 5, Chapter 4 of “Lập trình /C++ trên
Linux”

21
Memory Mapped Files

22
 Question
 What is the correct advantage of memory
mapped file?
A. reduces the task of the system’s OS
B. treats as the buffer for manipulating the file
C. allows programmers to organize the file
D. uses as shared resource among processes

23
 Question
 Which of the following is incorrect about virtual
memory?
A. it is separated from physical memory
B. it is mapped into physical memory during process
execution
C. it gives additional benefits, e.g., COW, file mapping
D. an address in virtual memory is preserved when
mapped into physical memory

24
 Dynamic loading

 Routine is not loaded until it is called

 Better memory-space utilization
 unused routine is never loaded
 Useful when
 large amounts of code are needed to handle
infrequently occurring cases
 No special support from the operating system
 refer to Section 3.4, Chapter 7 of “Lập trình C/C++ …”

25
Dynamic linking and shared library

 Linking postponed until execution time

 Small piece of code, stub, used to locate the
appropriate memory-resident library routine
 Stub replaces itself with the address of the routine, and
executes the routine
 Operating system needed to check if routine is in
processes’ memory address
 Dynamic linking is particularly useful for libraries
 Also known as shared libraries in Linux
 refer to Sect. 3, Chapter 7 of “Lập trình C/C++ trên Linux”
26
Paging on demand

27
Food order

28
Food order

29
Food order

30
Food order

Each person prefers different food 31

 Food order

One more steak

Each person prefers different food 32

 Food order
One more ice cream

One more steak

Each person prefers different food 33

 Food order
One more ice cream

One more steak One more beer

Each person prefers different food 34

 Food order
One more ice cream

One more steak One more beer

Each person prefers different food 35

 Food order
One more ice cream

One more steak One more beer

In a minute

Each person prefers different food 36

 Paging on demand
 Bring a page into memory only when needed
 Less I/O needed
 Less memory needed
 Faster response
 More users
 Page is needed  reference to it
 invalid reference  abort
 not-in-memory  bring to memory
 Lazy swapper – never swaps a page into
memory unless page is needed
 Swapper that deals with pages is a pager 37
 Question
 Why it is possible when only a part of a
process is loaded into MEM?
A. Because instructions of a process are
independent
B. Because we can indicate which instructions to
run
C. Because only one instruction is executed at a
time
D. Because related instructions are always in the
same group
38
Transfer of a Paged Memory to
Contiguous Disk Space

39
 Valid-Invalid Bit
 With each page table entry a valid–invalid bit is
associated
 (v  in-memory, i  not-in-memory)
 Initially valid–invalid bit is set to i on all entries
 Example of a page table snapshot:
Frame # valid-invalid bit
v
 During address translation, if v
v
valid–invalid bit in page table
v
entry is I  page fault i
….

i
i
page table 40
Page Table When Some Pages
Are Not in Main Memory

41
 Page Fault
 If there is a reference to a page,
 first reference to that page will trap to operating
system: page fault
1. Operating system looks at another table to decide:
 Invalid reference  abort
 Just not in memory
2. Get empty frame
3. Swap page into frame
4. Update the page table
5. Set validation bit = v
6. Restart the instruction that caused the page fault

42
 Page Fault (Cont.)
 Restart instruction MEM

 block move
NEW
NEW
OLD OLD

 auto increment/decrement location

43
Steps in Handling a Page Fault

44
 Process
information

45
Page fault in Windows

46
 Question
 Which of the following is incorrect about a
page fault?
A. it happens in paging on demand
B. it happens when a reference to a page that is not
in MEM
C. when a page fault occurs the corresponding
process will be terminated
D. a page fault handler is called whenever it occurs

47
 Question
 Which of the following is incorrect order of
steps in handling a page fault?
A. check if the valid bit is invalid  raise a page
fault
B. a page fault is raised  find the page in
backing store
C. a page fault is raised  find a free frame
D. load the page into memory  restart the
instruction

48
 If no free frame available
 Call page replacement procedure
 swap out an unused page from MEM
 Algorithms
 Optimal, FIFO, LRU, LRU-approximation, counting
 Performance of the algorithm
 page-fault rate
 which algorithm is better?

49
 Performance of paging on
demand
 Page Fault Rate 0  p  1.0
 if p = 0 no page faults
 if p = 1, every reference is a page fault
 Effective Access Time (EAT)
EAT = (1 – p) * memory access
+ p (page fault overhead
+ swap page out
+ swap page in
+ restart overhead)
50
Paging on Demand Example
 Memory access time = 200 nanoseconds
 Average page-fault service time = 8
milliseconds
 EAT = (1 – p) x 200 + p (8 milliseconds)
= (1 – p) x 200 + p x 8,000,000
= 200 + p x 7,999,800
 If one access out of 1,000 causes a page
fault
 EAT = 8.2 microseconds.
 slowdown by a factor of 41!!
51
 Page Replacement
 Prevent over-allocation of memory
 include page replacement in page-fault service
routine
 Use modify (dirty) bit to reduce overhead of
page transfers
 only modified pages are written to disk
 Page replacement completes separation
between logical memory and physical memory
 large virtual memory can be provided on a smaller
physical memory
52
Page Replacement

53
Need For Page Replacement

54
 Basic Page Replacement
1. Find the location of the desired page on disk
2. Find a free frame
 If there is a free frame, use it
 Else use a page replacement algorithm to select a
victim frame (and swap it out)
3. Bring the desired page into the (newly) free
frame; update the page table
4. Resume the process

55
Page Replacement

56
 Question
Which of the following is incorrect about page
replacement?
A. a victim frame is selected to be swapped out
B. the page table which is the victim will be updated
C. the victim frame is always written into the backing
store
D. the victim frame is only written into the backing
store if it is dirty

57
Page Replacement Algorithms
 Want lowest page-fault rate
 Evaluate algorithm
 run it on a particular string of memory references
(reference string)
 compute the number of page faults on that string
 In all our examples, the reference string is
1, 2, 3, 4, 1, 2, 5, 1, 2, 3, 4, 5

58
Graph of Page Faults Versus The
Number of Frames

59
 Optimal Algorithm
 Replace page that will not be used for longest
period of time
 4 frames example
1, 2, 3, 4, 1, 2, 5, 1, 2, 3, 4, 5

1
How do you know this? 4
2 6 page faults
 Used for measuring
3
how well the algorithm
4 5
performs
60
 Question
 A reference string 7 0 1 2 0 3 0 4 2 3 0 3 2 1 2 0
1 7 0 1, Optimal algorithm is used with 3 frames.
Which of the following is the correct order of
swapped out pages?
A. 710342
B. 701432
C. 710432
D. 714132

61
 Question
 A reference string 7 0 1 2 0 3 0 4 2 3 0 3 2 1 2 0
1 7 0 1, Optimal algorithm is used with 3 frames.
Which of the following is the correct number of
page faults?
A. 8
B. 9
C. 10
D. 11

62
 First-In-First-Out (FIFO)
Algorithm
 Reference string: 1, 2, 3, 4, 1, 2, 5, 1, 2,
3, 4, 5
1 1 4 5
2 2 1 3 9 page faults
 3 frames
3 3 2 4

1 1 5 4
2 2 1 5 10 page faults

 4 frames 3 3 2

4 4 3  Belady’s anomaly
Belady’s Anomaly: more frames  more page faults 63
FIFO Illustrating Belady’s
Anomaly

64
FIFO Page Replacement

65
 Question
 A reference string 7 0 1 2 0 3 0 4 2 3 0 3 2 1 2 0
1 7 0 1, FIFO is used with 3 frames. Which of
the following is the correct order of swapped out
pages?
A. 701230423012
B. 701230432012
C. 701320423012
D. 701230423102

66
 Question
 A reference string 7 0 1 2 0 3 0 4 2 3 0 3 2 1 2 0
1 7 0 1, FIFO is used with 3 frames. Which of
the following is the correct number of page
faults?
A. 13
B. 14
C. 15
D. 16

67
 Least Recently Used (LRU)
Algorithm
 Least recently used page is swapped out first
 Reference string: 1, 2, 3, 4, 1, 2, 5, 1, 2, 3, 4, 5
 4 frames

1 1 1 1 5
2 2 2 2 2
3 5 5 4 4
4 4 3 3 3

68
 Question
 A reference string 7 0 1 2 0 3 0 4 2 3 0 3 2 1 2 0
1 7 0 1, LRU is used with 3 frames.
 Which of the following is the correct order of
swapped out pages?
A. 712314132
B. 721304232
C. 712304123
D. 712304032

69
 Question
 A reference string 7 0 1 2 0 3 0 4 2 3 0 3 2 1 2 0
1 7 0 1, LRU is used with 3 frames.
 Which of the following is the correct number of
page faults?
A. 13
B. 12
C. 11
D. 10

70
Least Recently Used Algorithm
 Counter implementation
 Every page entry has a counter;
 every time page is referenced, copy the clock
into the counter
 When a page needs to be swapped
 look at the counters to determine

71
 LRU Algorithm (Cont.)
 Stack implementation
 keep a stack of page numbers in a double link
form
 Page referenced:
 move it to the top
 requires 6 pointers to be changed
 No search for replacement

72
Use Of A Stack to Record The Most
Recent Page References

73
LRU Approximation Algorithms
 Reference bit
 With each page associate a bit, initially = 0
 When page is referenced bit set to 1
 Replace the one which is 0 (if one exists)
 We do not know the order

 Second chance (follow clock order)

 Need reference bit
 Clock replacement
 If page at the pointer has reference bit = 1 then:
 set reference bit 0
 leave page in memory
 Move the pointer to next page, subject to same rules
74
 Else select the page as the victim
Second-Chance (clock) Page-
Replacement Algorithm

75
 Question
 Suppose the second chance is used;
 the reference bits of frames are: 1 1 0 1 1 0
 the head is at second frame
 Which of the following are the reference bits
after a page replacement is done
A. 000110
B. 101110
C. 100110
D. 101010

76
 Question
 Suppose the LRU counter implementation
without additional support, which of the
following is incorrect?
A. Timestamp is used to mark the referred time
B. The smallest timestamp is selected for
replacement
C. Searching is need to find the smallest timestamp
D. No need to search to find the smallest timestamp

77
 Question
 Suppose the LRU counter implementation
without additional support, which of the
following is incorrect?
A. Timestamp is used to mark the referred time
B. The smallest timestamp is selected for
replacement
C. Searching is need to find the smallest timestamp
D. No need to search to find the smallest timestamp

78
 Counting Algorithms
 Keep a counter of the number of references
that have been made to each page
 Least Frequently Used (LFU) Algorithm
 replaces page with smallest count
 Most Frequently Used (MFU) Algorithm
 based on the argument that the page with the
smallest count was probably just brought in and
has yet to be used

79
 Allocation of Frames
 Each process needs minimum number of
pages
 Example: IBM 370 – 6 pages to handle SS
MOVE instruction:
 instruction is 6 bytes, might span 2 pages
 2 pages to handle from
 2 pages to handle to
 Two major allocation schemes
 fixed allocation
 priority allocation
80
 Fixed Allocation
 Equal allocation
 For example, if there are 100 frames and 5
processes, give each process 20 frames.
 Proportional allocation
 Allocate according to the size of process
 si  size of process pi m  64
 S   si s1  10
s2  127
 m  total number of frames
10
s a1   64  5
 ai  allocation for ai  i  m 137
S 127
a2   64  59
137 81
 Question
 A system uses proportional allocation and
has
 90 frames x 2KB
 3 processes with size of (138KB, 96KB, 164KB)
 Which of the following is the correct number
of allocated frames of (P1, P2, P3)
A. 32, 21, 37
B. 31, 22, 37
C. 30, 22, 38
D. 33, 22, 35
82
 Priority Allocation
 Use a proportional allocation scheme using
priorities rather than size

 If process Pi generates a page fault,

 select for replacement one of its frames
 select for replacement a frame from a process
with lower priority number

83
 Global vs. Local Allocation
 Global replacement
 process selects a replacement frame from the set
of all frames;
 one process can take a frame from another
 Local replacement
 each process selects from only its own set of
allocated frames

84
 Thrashing
 If a process does not have “enough” frames,
the page-fault rate is very high.
 Thrashing
 a process is busy swapping pages in and out
 This caused by:
 low CPU utilization
 operating system thinks that it needs to increase
the degree of multiprogramming
 another process added to the system

85
Thrashing (Cont.)

86
 Solutions to Thrashing
 Use local allocation
 Use priority allocation
 not good solution
 Working set model
 A suitable solution

87
 Question
 Which of the following is incorrect about
priority allocation?
A. higher priority process is allocated first
B. it prevents thrashing from happening
C. frames are allocated globally
D. it does not prevent thrashing from
happening

88
 Working-Set Model
   working-set window
 a number of page references, e.g. 10,000
 Working set of Process Pi
 WSSi =total number of pages referenced in the most
recent  (varies in time)
 if  too small will not encompass entire locality
 if  too large will encompass several localities
 if  =   will encompass entire program
 D =  WSSi  total demand frames
 if D > m (total of frames) Thrashing
 Policy if D > m, then suspend one process
89
Working-set model

90
 Keep Track of the WSet
 Approximate with interval timer + a reference bit
 Example:  = 10,000
 Timer interrupts after every 5000 time units
 Keep in memory 2 bits for each page in last 10000
and 15000
 15000--..bit1..-- 10000 ---..bit2..--- 5000 ---..ref.. --- now
 Whenever a timer interrupts copy and sets the values
of all reference bits to 0
 If one of the bits in memory = 1  page in working set
 Why is this not completely accurate?
 Improvement = 10 bits and interrupt every 1000
time units
91
 Question
 Suppose a delta =10; reference string
 2615777751623444344413234443
444…
 Which of the following is the correct WSS at
20th reference?
A. {2 3 4 6}
B. {2 3 4 5 6}
C. {1 2 3 4 6}
D. {7 1 2 3 4 6}

92
Question?

93