Virtual Memory - CH 10

Chapter 10: Virtual Memory
Operating System Concepts – 10th Edition Silberschatz, Galvin and Gagne ©2018
Chapter 10: Virtual Memory
 Background
 Demand Paging
 Copy-on-Write
 Page Replacement
 Allocation of Frames
 Thrashing
 Memory-Mapped Files
 Allocating Kernel Memory
 Other Considerations
 Operating-System Examples
Operating System Concepts – 10th Edition 10.2 Silberschatz, Galvin and Gagne ©2018
Objectives
 Define virtual memory and describe its benefits.

 Illustrate how pages are loaded into memory using demand paging.
 Apply the FIFO, optimal, and LRU page-replacement algorithms.
 Describe the working set of a process, and explain how it is related to
program locality.
 Describe how Linux, Windows 10, and Solaris manage virtual memory.
 Design a virtual memory manager simulation in the C programming
language.
Background
 Code needs to be in memory to execute, but entire program rarely used
• Error code, unusual routines, large data structures
 Even in those cases where the entire program is needed, it may not all be
needed at the same time
 Consider ability to execute partially-loaded program
• Program no longer constrained by limits of physical memory
• Program and programs could be larger than physical memory
• Each program takes less memory while running -> more programs run
at the same time
 Increased CPU utilization and throughput with no increase in
response time or turnaround time
Virtual memory
 Virtual memory – separation of user logical memory (virtual address

space) from physical memory
• Only part of the program needs to be in memory for execution
• Logical address space can therefore be much larger than physical
address space
 Virtual address space – logical view of how process is stored in memory
• Usually start at address 0, contiguous addresses until end of space
• Meanwhile, physical memory organized in page frames
• MMU must map logical to physical
 Virtual memory can be implemented via:
• Demand paging
• Demand segmentation
Virtual Memory That is Larger Than Physical Memory
Demand Paging
 Pages are only loaded when they are demanded during
program execution
 Pages that are never accessed are thus never loaded into physical
memory
 A demand-paging system is similar to a paging system with
swapping
 Rather than swapping the entire process into memory, however, we
use a lazy swapper
 Lazy swapper: never swaps a page into memory unless that page
will be needed
 Swapper that deals with pages is a pager
Demand Paging
 Similar to paging system with

swapping (diagram on right)
Basic Concepts
 With swapping, pager guesses which pages will be used before swapping
out again
 Instead, pager brings in only those pages into memory
 How to determine that set of pages?
• Need new MMU functionality to implement demand paging
 If pages needed are already memory resident
• No difference from non demand-paging
 If page needed and not memory resident
• Need to detect and load the page into memory from storage
 Without changing program behavior
 Without programmer needing to change code
Valid-Invalid Bit
 With each page table entry a valid–invalid bit is associated
 v  in-memory – memory resident

 i  not-in-memory
 Initially valid–invalid bit is set to i on all entries
 During address translation, if valid–invalid bit in page table entry is i

 page fault trap
Valid-Invalid Bit
 Example of a page table snapshot:
 During MMU address translation, if valid–invalid bit in page table entry is i

 page fault
Page Table When Some Pages Are Not
in Main Memory
Steps in Handling a Page Fault (Cont.)
Steps in Handling Page Fault
1. If there is a reference to a page, first reference to that page will trap to

operating system
• Page fault
2. Operating system looks at another table to decide:
• Invalid reference  abort
• Just not in memory
3. Find free frame
4. Swap page into frame via scheduled disk operation
5. Reset tables to indicate page now in memory
Set validation bit = v
6. Restart the instruction that caused the page fault
Stages in Demand Paging – Worse Case
1. Trap to the operating system

2. Save the user registers and process state
3. Determine that the interrupt was a page fault
4. Check that the page reference was legal and determine the location of
the page on the disk
5. Issue a read from the disk to a free frame:
a) Wait in a queue for this device until the read request is serviced
b) Wait for the device seek and/or latency time
c) Begin the transfer of the page to a free frame
Stages in Demand Paging (Cont.)
6. While waiting, allocate the CPU to some other user
7. Receive an interrupt from the disk I/O subsystem (I/O completed)
8. Save the registers and process state for the other user
9. Determine that the interrupt was from the disk
10. Correct the page table and other tables to show page is now in memory
11. Wait for the CPU to be allocated to this process again
12. Restore the user registers, process state, and new page table, and then
resume the interrupted instruction
Pure Demand Paging
 Start process with no pages in memory
• OS sets instruction pointer to first instruction of process, non-memory-
resident  page fault occurs
• And for every other process pages, page fault occurs on first access
• Never bring a page into memory until it is required
 Theoretically, a given instruction could access multiple pages  multiple
page faults
• Consider fetch and decode of instruction which adds 2 numbers from
memory and stores result back to memory
• Pain decreased because of locality of reference
 Hardware support needed for demand paging
• Page table with valid / invalid bit
• Secondary memory (swap device with swap space)
• Instruction restart
Performance of Demand Paging
 Page Fault Rate 0  p  1
 Effective Access Time (EAT)
EAT = (1– p)  memory access + p  page fault overhead
 page-fault overhead
1. Service the page-fault interrupt
2. Read in the page (read the page from disk) – lots of time
3. Restart the process
 Example
 Memory access time = 200 nanoseconds
 Average page fault service time = 8 milliseconds
Demand Paging 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, then
P = 1/1000
EAT = 8.2 microseconds.
This is a slowdown by a factor of 40!! (the access time gets 1/40!!)
 If want performance degradation < 10 percent, then EAT = 200 + 20 220
• 220 > 200 + 7,999,800 x p
20 > 7,999,800 x p
• p < .0000025 => (25 page faults in any 10,000,000 memory accesses)
• < one page fault in every 400,000 memory accesses
Copy-on-Write
 Recall that the fork() system call creates a child process as a duplicate of
its parent
• It creates a copy of the parent‘s address space for the child
 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
 COW allows more efficient process creation as only modified pages are
copied
Page Replacement
 When virtual memory management over-allocates memory, it is possible

that all available memory is used by active processes
 In this situation, if a page fault occurs, there is no free frame to allocate it to

the requested page
 Solution
 Find some page in memory that is not currently being used and page it out
 We can free a frame by writing its contents to swap space and changing
the page table
 Same page may be brought into memory several times

Need For Page Replacement
Page Replacement
 Page replacement increases the effective access time
 Use modify (dirty) bit to reduce overhead of page transfers
 Each page has a modify bit associated with it
 The modify bit for a page is set by the hardware whenever any word
or byte in the page is written
 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
Demand Paging Algorithms
 Frame-allocation algorithm
 Determines how many frames to allocate to each process
 Page-replacement algorithm
 Selects the frames that are to be replaced
 Designing efficient algorithms is so important, because

disk I/O is so expensive
 In general, we want the algorithm with the lowest page-fault rate
Page Replacement Algorithms
 We evaluate algorithm by running it on a particular string of memory

references (reference string) and computing the number of page
faults on that string
 String is just page numbers, not full addresses
 Repeated access to the same page does not cause a page fault
 In all our examples, the reference string is

7, 0, 1, 2, 0, 3, 0, 4, 2, 3, 0, 3, 2, 1, 2, 0, 1, 7, 0, 1
and there are 3 frames

Page Replacement Algorithms
 Optimal
 FIFO
 LRU (Least Recently Used)

 LRU Approximation
 Additional-Reference-Bits Algorithm
 Second-Chance Algorithm
 Counting-Based Page Replacement

 LFU (Least Frequently Used)
 MFU (Most Frequently Used)
Optimal Page Replacement
 Replace page that will not be used for longest period of time
 Example: 9 page faults
 How do you know this?
 Can’t read the future
 Used for measuring how well your algorithm performs
7 0 1 2 0 3 0 4 2 3 0 3 2 1 2 0 1 7 0 1
7 7 7 2 2 2 2 2 7
0 0 0 0 4 0 0 0
1 1 3 3 3 1 1
First-In-First-Out (FIFO) Algorithm
 When a page must be replaced, the oldest page is chosen

 Can be implemented using a FIFO queue
 Example: 15 page faults
7 0 1 2 0 3 0 4 2 3 0 3 2 1 2 0 1 7 0 1
7 7 7 0 1 2 3 0 4 2 3 0 1 2 7
0 0 1 2 3 0 4 2 3 0 1 2 7 0
1 2 3 0 4 2 3 0 1 2 7 0 1
Graph of Page Faults Versus the Number of Frames
Belady’s Anomaly
 For some page-replacement algorithms, the page-fault rate may
increase as the number of allocated frames increases
 Example: reference string = 1, 2, 3, 4, 1, 2, 5, 1, 2, 3, 4, 5
Least Recently Used (LRU)
 LRU is an approximation of the optimal algorithm

 It uses the recent past as an approximation of the near future
 LRU replaces the page that has not been used for the
longest period of time
7 0 1 2 0 3 0 4 2 3 0 3 2 1 2 0 1 7 0 1
7 7 7 0 1 2 2 3 0 4 2 2 0 3 3 1 2 0 1 7
0 0 1 2 0 3 0 4 2 3 0 3 2 1 2 0 1 7 0
1 2 0 3 0 4 2 3 0 3 2 1 2 0 1 7 0 1
 Stack implementation
 Keep a stack of page numbers in a double link form
 Page referenced: move it to the top
 But each update more expensive
 No search for replacement
 Counter implementation
 Every page entry has a counter; every time page is referenced through this entry,
copy the clock into the counter
 When a page needs to be changed, look at the counters to find smallest value
 Search through table needed
 LRU and OPT are cases of stack algorithms that don’t

have Belady’s Anomaly
 An algorithm for which it can be shown that the set of pages in
memory for n frames is always a subset of the set of pages that
would be in memory with n + 1 frames
 Implementation of LRU needs hardware assistance

 The updating of the clock fields or stack must be done for every
memory reference
LRU Approximation Algorithms
 LRU needs special hardware and still slow
 Few computer systems provide sufficient hardware support for true LRU
 Many systems provide some help, in the form of a reference bit

 The reference bit for a page is set by the hardware whenever that page is
referenced
 Reference bits are associated with each entry in the page table
Additional-Reference-Bits Algorithm
 Keep an 8-bit byte for each page

 At regular intervals (say, 100 ms), a timer interrupt transfers control to
the operating system
 The operating system shifts the reference bit for each page into the
high-order bit of its 8-bit byte, shifting the other bits right by 1 bit
 These 8-bit shift registers contain the history of page use for the
last eight time periods
 Example: 00000000, 11111111, 11000100, 01110111
 The page with the lowest number is the LRU page, and it can be
replaced.
Second Chance Algorithm
 The basic algorithm is FIFO
 When a page has been selected, we inspect its reference bit

 If the value is 0, we proceed to replace this page
 If the value is 1, we give the page a second chance and move on to

select the next FIFO page, its reference bit is cleared, and its arrival
time is reset to the current
 Clock algorithm
 An implementation of the second-chance algorithm using a
circular queue
Enhanced Second-Chance Algorithm
 We can use reference bit and modify bit together
 Then we have four cases
 (0, 0) neither recently used nor modified - best page to replace
 (0, 1) not recently used but modified- not quite as good, need to be
written out
 (1, 0) recently used but clean - probably will be used again soon
 (1,1) recently used and modified - probably will be used again soon,
and need to be written out
 We replace the first page encountered in the lowest nonempty
class
Enhanced Second-Chance Algorithm
Combined Examples
Comparison
Counting Algorithms
 Keep a counter of the number of references that have been
made to each page
 Least Frequently Used (LFU)

 Requires that the page with the smallest count be replaced
 Problem: when a page is used heavily during the initial phase of a

process but then is never used again
 Solution: shift the counts right by 1 bit at regular intervals

Counting Algorithms
 Most Frequently Used (MFU)

 is based on the argument that the page with the smallest count was
probably just brought in and has yet to be used
 Neither MFU nor LFU replacement is common

 The implementation of these algorithms is expensive
 They do not approximate OPT replacement well

Page-Buffering Algorithms
 Keep a pool of free frames, always

 When a page fault occurs, a victim frame is chosen as before
 However, the desired page is read into a free frame from the pool
before the victim is written out
 When the victim is later written, its frame is added to the free-frame
pool
 Keep list of modified pages
 When backing store otherwise idle, write pages there and set to
non-dirty
Resident Set Management
 The OS must decide how many pages to bring into main

memory
 The smaller the amount of memory allocated to each process,
the more processes that can reside in memory.
 Small number of pages loaded increases page faults.
 Beyond a certain size, further allocations of pages will not affect

the page fault rate.
Resident Set Size
 Fixed-allocation
 Gives a process a fixed number of pages within which to execute
 When a page fault occurs, one of the pages of that process

must be replaced
 Variable-allocation
 Number of pages allocated to a process varies over the lifetime
of the process
Fixed Allocation
 Equal allocation
 If there are 100 frames (after allocating frames for the OS) and 5
processes, give each process 20 frames
 Proportional allocation
 Allocate according to the size of process
m  64
s i  size of process p i
s1  10
S   si s 2  127
m  total n u mb e r of frames a1 
10
 64  5
137
si
a i  allocationfor p i  m a2 
127
 64  59
S 137
Replacement Scope
 The scope of a replacement strategy can be categorized

as global or local.
 Both types are activated by a page fault when there are no free
page frames.
 A local replacement policy chooses only among the resident pages

of the process that generated the page fault
 A global replacement policy considers all unlocked pages

in main memory
Fixed Allocation, Local Scope
 Decide ahead of time the amount of allocation to give a

process
 If allocation is too small, there will be a high page fault rate
 If allocation is too large there will be too few programs in

main memory
 Increased processor idle time or
 Increased swapping
Variable Allocation, Global Scope
 Easiest to implement
 Adopted by many operating systems
 Operating system keeps list of free frames
 Free frame is added to resident set of process when a

page fault occurs
 If no free frame, replaces one from another process

 Therein lies the difficulty … which to replace.
Variable Allocation, Local Scope
 When new process added, allocate number of page

frames based on application type, program request, or
other criteria
 When page fault occurs, select page from among

the resident set of the process that suffers the fault
 Reevaluate allocation from time to time

Resident Set Management Summary
Thrashing
 Thrashing  a process is busy swapping pages in and out
 If a process does not have “enough” pages, the page-fault

rate is very high
 Page fault to get page
 Replace existing frame
 But quickly need replaced frame back

Thrashing
End of Chapter 10
Operating System Concepts – 10th Edition Silberschatz, Galvin and Gagne ©2018

Virtual Memory - CH 10

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Virtual Memory - CH 10

Uploaded by

Copyright:

Available Formats

Chapter 10: Virtual Memory

 Define virtual memory and describe its benefits.

 Virtual memory – separation of user logical memory (virtual address

 Similar to paging system with

 v  in-memory – memory resident

 During address translation, if valid–invalid bit in page table entry is i

 During MMU address translation, if valid–invalid bit in page table entry is i

1. If there is a reference to a page, first reference to that page will trap to

1. Trap to the operating system

• It creates a copy of the parent‘s address space for the child

 Copy-on-Write (COW) allows both parent and child processes to initially

• If either process modifies a shared page, only then is the page

 When virtual memory management over-allocates memory, it is possible

 In this situation, if a page fault occurs, there is no free frame to allocate it to

 Same page may be brought into memory several times

 Designing efficient algorithms is so important, because

 We evaluate algorithm by running it on a particular string of memory

 In all our examples, the reference string is

and there are 3 frames

 LRU (Least Recently Used)

 Counting-Based Page Replacement

 When a page must be replaced, the oldest page is chosen

 LRU is an approximation of the optimal algorithm

 LRU and OPT are cases of stack algorithms that don’t

 Implementation of LRU needs hardware assistance

 LRU needs special hardware and still slow

 Many systems provide some help, in the form of a reference bit

 Keep an 8-bit byte for each page

 The basic algorithm is FIFO

 When a page has been selected, we inspect its reference bit

 If the value is 1, we give the page a second chance and move on to

 Least Frequently Used (LFU)

 Problem: when a page is used heavily during the initial phase of a

 Solution: shift the counts right by 1 bit at regular intervals

 Most Frequently Used (MFU)

 Neither MFU nor LFU replacement is common

 They do not approximate OPT replacement well

 Keep a pool of free frames, always

 The OS must decide how many pages to bring into main

 Small number of pages loaded increases page faults.

 Beyond a certain size, further allocations of pages will not affect

 When a page fault occurs, one of the pages of that process

 The scope of a replacement strategy can be categorized

 A local replacement policy chooses only among the resident pages

 A global replacement policy considers all unlocked pages

 Decide ahead of time the amount of allocation to give a

 If allocation is too small, there will be a high page fault rate

 If allocation is too large there will be too few programs in

 Operating system keeps list of free frames

 Free frame is added to resident set of process when a

 If no free frame, replaces one from another process

 When new process added, allocate number of page

 When page fault occurs, select page from among

 Reevaluate allocation from time to time

 Thrashing  a process is busy swapping pages in and out

 If a process does not have “enough” pages, the page-fault

 Replace existing frame

 But quickly need replaced frame back

You might also like