OS-Chap-9 Virtual-Memory

Chapter 9: Virtual Memory
Operating System Concepts – 9th Edition Silberschatz, Galvin and Gagne ©2013
Chapter 9: 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 – 9th Edition 9.2 Silberschatz, Galvin and Gagne ©2013
Objectives
To describe the benefits of a virtual memory system.

To explain the concepts of demand paging, page-replacement
algorithms, and allocation of page frames.
To discuss the principle of the working-set model.
To examine the relationship between shared memory and
memory-mapped files.
To explore how kernel memory is managed.
Background
Code needs to be in memory to execute, but
Code for rarely happening errors.
entire program rarely used.
Error code, unusual routines, large data Routines that are rarely used.
structures.
Large data structures such as a
Entire program code not needed at same time. large array. However, it is
Consider ability to execute partially-loaded partially used.
program.
Program no longer constrained by limits of The process needs partial
physical memory. loading before starting to run
Each program takes less memory while which is faster than waiting
running -> more programs run at the same for fully loading the process.
time.
Less remaining part of the
 Increased CPU utilization and throughput process to load (Less I/O).
with no increase in response time or
turnaround time. Less loaded parts to be paged
back to disk (Less I/O).
Less I/O needed to load or swap programs
into memory -> each user program runs faster.
Background (Cont.)
Virtual memory – Separation of user logical memory 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.
Allows address spaces to be shared by several processes.
Allows for more efficient process creation.
More programs running concurrently.
Less I/O needed to load or swap processes.
The concept of “Swapping” is usually used when we swap an entire

process from RAM to Disk and bring in another entire process.
The concept of “Paging” is usually used when we swap a page from

RAM to Disk and bring in another page.
Background (Cont.)
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
Part of the space on disk

is used as a “Swap space”.
It is faster than other
parts of the disk in which
we store ordinary files and
data. This swap space is
used for page movement
between disk and physical
memory (RAM).
Such as Disk
Page
Table
What is the difference between logical and virtual memory?
They are the same except that when we say logical memory, then all process pages
are loaded into physical memory (none demand-paging). However, in virtual
memory we use the concept of swapping pages out of memory to disk and swapping
other pages in from disk to memory (partially loaded process )(on demand-paging).
Virtual-Address Space
Usually design logical address space for Remember, one way
stack to start at Max logical address and for processes to
communicate is via
grow “down” while heap grows “up”.
shared memory. So, a
Maximizes address space use. given part of unused
hole can be used as
Unused address space between
shared memory
the two is hole. between processes.
 No physical memory needed
until heap or stack grows to a Also, part of the
given new page. unused hole can be
used for a library to
Enables sparse address spaces with be shared between
holes left for growth, dynamically linked several processes.
libraries, etc.
Sparse address
System libraries shared via mapping into space: Address space
virtual address space. containing scattered
Shared memory by mapping pages read- holes.
write into virtual address space.
Pages can be shared during fork(),
speeding process creation.
We will see this point later
Shared Library Using Virtual Memory
Demand Paging
Let us say that a process occupies a few
Could bring entire process into memory at
pages, then the first few instructions to start
load time.
the process reside in the first page. So,
Or bring a page into memory only when it paging in the first page and starting process
is needed. execution is faster than loading the whole
Less I/O needed, no unnecessary I/O. process into memory (faster response). Also,
this way we can bring several partially loaded
Less memory needed. processes instead of bringing few fully loaded
Faster response. processes (more users). Clearly, this requires
less memory and less I/O.
More users.
Similar to paging system with swapping
(diagram on right).
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 will be needed.
Swapper that deals with pages is a
pager.
Basic Concepts
With swapping, pager guesses which pages will be Pager brings in the pages that
will be used.
used before swapping out again.
Instead, pager brings in only those pages into “valid-invalid” bit is added to page
table entries. Its content is “i”
memory. (invalid) when:
How to determine that set of pages?
1. Page is not loaded in a memory
Need new MMU functionality to implement frame.
demand paging. 2. Page is outside process valid
logical space.
If pages needed are already memory resident.
Otherwise, its content is “v” (valid).
No difference from none demand-paging.
If all pages are loaded in
If page needed and not memory resident. memory frames, then there
Need to detect and load the page into memory is no difference between on
demand paging and none
from storage. demand paging. This is
 Without changing program behavior. because when a page is
needed it will be there
 Without programmer needing to change code. anyway in physical memory.
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.
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
Suppose a process needs ‘D’. It

is in page 3. By referring to page
table we see it is not mapped to
a given frame yet. This why the
‘valid-invalid bit’ next to it is i
(invalid). So, the page is
demanded in (paged in)
(swapped in) from disk to
physical memory and the page
table entry is updated.
Another reason for valid-invalid

bit to be “i” is when the process
tries to access a page that is not
in its allowed address space.
Here, access is denied.
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. Find free frame.
3. Swap page into frame via scheduled disk operation.
4. Reset tables to indicate page now in memory.
Set validation bit = v.
5. Restart the instruction that caused the page fault.
Steps in Handling a Page Fault
Aspects of Demand Paging
Extreme case – start process with no pages in memory: In pure demand paging, don’t
page in a page unless when it is
OS sets instruction pointer to first instruction of first access (needed). So, every
access to a new page results in
process, non-memory-resident -> page fault. page fault before paging in that
page.
And for every other process pages on first access.
Pure demand paging. Because of the principle of
locality of reference, the data
Actually, a given instruction could access multiple pages needed for a given instruction
will most likely be in the same
-> multiple page faults. page or at least minimum
number of pages.
Consider fetch and decode of instruction which adds
2 numbers from memory and stores result back to C=A+B
memory. - Fetch and decode instruction
- Fetch A
Pain decreased because of locality of reference. - Fetch B
- Perform addition
Hardware support needed for demand paging: - Store result in C
Page table with valid / invalid bit. What if C is not in memory

(Page Fault), then page will be
Secondary memory (swap device with swap space). brought to memory and
instruction is restarted.
Instruction restart.
Instruction Restart
For example, assume one instruction is able to
Consider an instruction move a big block from current location to another
that could access several location (source block to destination block).
different locations:
Assume for example, the source block is all fit in one page except
block move the very last part which had to be stored in a second page. Suppose
first page is in memory and second one is not (page fault). So, data
in the first page are moved but the operation is not complete
because when we reach the second page, it will result in page fault
and the instruction is restarted.
What if the source block and the second block

overlaps? In this case, when the move is performed,
auto part of the source block data will be overridden, so, if
increment/decrement we face a page fault before completing the move
location. operation, we cannot simply restart the instruction.
Restart the whole
operation? Solution 1: Test the pages at the boundaries of each
block before start execution, if a page fault occurs,
 What if source and then bring it to memory then start instruction.
destination
overlap? Solution 2: Old data of the source block is temporarily
stored in registers, so, if we face the situation
explained above, we use the temporary data to bring
back the source block to its original values.
Performance of Demand Paging
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:
1. Wait in a queue for this device until the read request is serviced
2. Wait for the device seek and/or latency time
3. Begin the transfer of the page to a free frame
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
Performance of Demand Paging (Cont.)
Three major activities
Service the interrupt – careful coding means just several hundred
instructions needed
Read the page – lots of time
Restart the process – again just a small amount of time
Page Fault Rate 0  p  1
if p = 0 no page faults
if p = 1, every reference is a fault
Effective Access Time (EAT)
EAT = (1 – p) x memory access
+ p (page fault overhead
+ swap page out
+ swap page in )
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
EAT = 8.2 microseconds.
This is a slowdown by a factor of 40!!
If want performance degradation < 10 percent
220 > 200 + 7,999,800 x p
20 > 7,999,800 x p
p < .0000025
< one page fault in every 400,000 memory accesses
Demand Paging Optimizations
Swap space is usually on the hard disk,
same as file system. However, dealing
Swap space I/O faster than file system I/O even if on the
with swap space is faster.
same device, why?
Because swap allocated in larger chunks, less The advantage of this is that entire
process is moved at one time to the swap
management needed than file system.
space, then paging needs to interact with
Copy entire process image to swap space at process load the fast swap space without needing to
time. go to the file system. However, the
downside here is that not needed pages
Then page in and out of swap space. are loaded into swap space.
Used in older BSD Unix.
When a page is needed, it is demanded
Demand page in from program binary on disk but discard directly from the file system on the hard
rather than paging out when freeing frame. disk, then moved to physical memory.
When a page needs to be replaced, then:
Used in Solaris and current BSD. 1- If the page is not modified, then no
Still need to write to swap space. need to move it to swap space (Just
discard it).
 Pages not associated with a file (like stack and 2- If the page is modified, then it is
heap) – anonymous memory. written to the swap space in case we
need to page it in again.
 Pages modified in memory but not yet written back 3- If the page is related to anonymous
to the file system. data (such as data in stack), then it is
also moved to swap space because there
Mobile systems.
is not a dedicated space for that in the
Typically, don’t support swapping. file system.
Instead, demand page from file system and reclaim Here, only the needed pages are
read-only pages (such as code). demanded. However, the downside is that
interacting with slower file system is
needed. Silberschatz, Galvin and Gagne ©2013
Operating System Concepts – 9th Edition 9.21
Copy-on-Write
Copy-on-Write (COW) allows both parent and
Remember that fork() system call initially creates a child
child processes to initially share the same with an address space identical to its parent. However,
pages in memory. many processes call exec() immediately after fork() which
If either process modifies a shared page, means that their address space will be replaced with a
new one. So, the previous duplicate of parent address
only then is the page copied.
space goes for nothing (unnecessary). To alleviate this,
COW allows more efficient process creation as we use fork() with copy-on-write.
only modified pages are copied.
In general, free pages are allocated from a fork() with copy-on-write: Instead of duplicating
address space of parent, both parent and child share the
pool of zero-fill-on-demand pages.
address space, then, if a page is modified by parent or
Pool should always have free frames for child, then this page is duplicated into the address space
fast demand page execution. of the process (parent or child) that modified it. This way,
we save memory. The duplicate page is stored in a free
 Don’t want to have to free a frame as frame. A pool of free frames contains zeroed-out frames
well as other processing on page (Frames with erased (blank) content) and they can be
fault. used when performing copy-on-write.
Why zero-out a page before allocating it?

vfork() is a version of fork() that works differently, and it
vfork() variation on fork()system call has
does not use copy-on-write. Here, parent and child share
parent suspend and child using copy-on-write the address space. If for example, child wants to modify
address space of parent a page, then parent is suspended until modify operation
takes place, then parent is resumed again. This way no
Designed to have child call exec().
pages are duplicated (Very efficient memory allocation).
Very efficient.
Before Process 1 Modifies Page C
After Process 1 Modifies Page C
What happens if there is no free frame?
Used up by process pages.
Also, in demand from the kernel, I/O buffers, etc.
How much to allocate to each?
Page replacement – find some page in memory, but not really in use,
page it out.
Algorithm – terminate? swap out? replace the page?
Performance – want an algorithm which will result in minimum
number of page faults.
Same page may be brought into memory several times.
There might be no free frames because frames can be used by

processes, and it can be used by I/O buffers, … etc. So, we may run
out of free frames.
What to do in this case?

1. Terminate a process to free all its occupied frames or
2. Page replacement, swap some currently unused page out and
swap in the page that caused page fault.
Page Replacement
Prevent over-allocation of memory by modifying page-fault service

routine to include page replacement.
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/logical memory can be provided on a
smaller physical memory.
If a dirty bit is 1, this means the page is updated, therefore, if

swapped out, then it has to go to disk. However, if dirty bit is 0,
then page is not modified, then we can simply discard it because a
copy of it is already there on disk.
Need For Page Replacement
Reserved for parts
of operating system
Process 1(User 1) needs to load M. This leads

to a page fault because a reference to M does
not exist in the page table. So, M is brought
from disk and loaded into the last free frame in
memory. Now Process 2(User 2) wants to load
B which also leads to a page fault. However,
this time, no free frames exist in physical
memory, so, we need page replacement.
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.
▪ If there is no free frame, use a page replacement algorithm to select a
victim frame.
o Write victim frame to disk if dirty.
3. Bring the desired page into the (newly) free frame; update the page and
frame tables.
4. Continue the process by restarting the instruction that caused the trap.
Note: Now potentially two-page transfers for page fault – increasing EAT.
EAT (Effective Access Time): The average time needed to access a
page in memory.
Page Replacement
1. Select a victim page and swap it out.

2. Change “valid-invalid bit” for this
page to invalid because it is no longer
available in memory.
3. Swap in the desired page.
4. Update page table to reflect the new
page mapping.
Page and Frame Replacement Algorithms
Frame-allocation algorithm determines:

How many frames to give each process.
Which frames to replace.
Page-replacement algorithm:
Want lowest page-fault rate on both first access and re-access.
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.
Results depend on number of frames available.
In all our examples, the reference string of referenced page numbers is:
7, 0, 1, 2, 0, 3, 0, 4, 2, 3, 0, 3, 0, 3, 2, 1, 2, 0, 1, 7, 0, 1
Graph of Page Faults Versus The Number of Frames
In general, more frames means less page faults.
However, there is Belady’s Anomaly which we will explain later.
First-In-First-Out (FIFO) Algorithm
Reference string: 7, 0, 1, 2, 0, 3, 0, 4, 2, 3, 0, 3, 0, 3, 2, 1, 2, 0, 1, 7, 0, 1.
3 frames (3 pages can be in memory at a time per process):
Look next slide
15-page faults
Can vary by reference string: consider 1, 2, 3, 4, 1, 2, 5, 1, 2, 3, 4, 5:
Adding more frames can cause more page faults!
 Belady’s Anomaly
How to track ages of pages?
Just use a FIFO queue.
First-In-First-Out (FIFO) Algorithm
Reference string: 7, 0, 1, 2, 0, 3, 0, 4, 2, 3, 0, 3, 0, 3, 2, 1, 2, 0, 1, 7, 0, 1
Reference 7 0 1 2 0 3 0 4 2 3 0 3 0 3 2 1 2 0 1 7 0 1
Fault x x x x x x x x x x x x x x x
7 7 7 0 1 2 3 0 4 2 3 0 1 2 7
Queue 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
First in First out (FIFO) queue. Enter from the bottom of the queue and
leaves from the top of the queue.
Number of page faults = 15
FIFO Illustrating Belady’s Anomaly
In this example, number of page faults when total number of frames is 4

is higher than number of page faults when total number of frames is 3.
Optimal Algorithm
Replace page that will not be used for longest period of time.
9 is optimal for the example.
How do you know this?
Can’t read the future.
Used for measuring how well your algorithm performs.
Look next slide
Optimal Algorithm (OPT) cannot suffer from Belady’s Anomaly.
Optimal Algorithm
Reference 7 0 1 2 0 3 0 4 2 3 0 3 0 3 2 1 2 0 1 7 0 1
Fault x x x x x x x x x
7 7 7 2 2 2 2 2 7
Memory 0 0 0 0 4 0 0 0
1 1 3 3 3 1 1
The page to be replaced is the one that will not be used for longer time.
For example, here (in red font), 2 results in page fault. Which page to replace (7 or 0 or 1)?
7: start at the 2, 7 will be used after 16 references.
0: start at the 2, 0 will be used after 1 reference.
1: start at the 2, 1 will be used after 12 references.
So, page 7 is the one that won’t be used for longer time, so it should be replaced.
Least Recently Used (LRU) Algorithm
Use past knowledge rather than future.

Replace page that has not been used in the most amount of time.
Associate time of last use with each page:
12 faults – better than FIFO but worse than OPT. Look next slide
Generally good algorithm and frequently used.

But how to implement?
Least Recently Used (LRU) Algorithm
Reference 7 0 1 2 0 3 0 4 2 3 0 3 0 3 2 1 2 0 1 7 0 1
Fault x x x x x x x x x x x x
7 7 7 2 2 4 4 4 0 1 1 1
0 0 0 0 0 0 3 3 3 0 0
1 1 3 3 2 2 2 2 2 7
Replace page that has not been used in the most amount of time.
For example, here (in red font), 1 results in page fault. Which page to replace (0 or 3 or 2)?
0: start at the 1, 0 has not been used since 3 references back.
3: start at the 1, 3 has not been used since 2 references back.
2: start at the 1, 2 has not been used since 1 reference back.
So, page 0 is the one that has no been used for longer time, so it should be replaced.
LRU Algorithm (Cont.)
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
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
But each update more expensive
No search for replacement
LRU and OPT are cases of stack algorithms that don’t have
Belady’s Anomaly
Use Of A Stack to Record Most Recent Page References
LRU Approximation Algorithms
LRU needs special hardware and still slow
Reference bit
With each page associate a bit, initially = 0
When page is referenced bit set to 1
Replace any with reference bit = 0 (if one exists)
 We do not know the order, however
Second-chance algorithm
Generally FIFO, plus hardware-provided reference bit
Clock replacement
If page to be replaced has
 Reference bit = 0 -> replace it
 reference bit = 1 then:
– set reference bit 0, leave page in memory
– replace next page, subject to same rules
Second-Chance (clock) Page-Replacement Algorithm
Enhanced Second-Chance Algorithm
Improve algorithm by using reference bit and modify bit (if

available) in concert
Take ordered pair (reference, modify)
1. (0, 0) neither recently used not modified – best page to replace
2. (0, 1) not recently used but modified – not quite as good, must
write out before replacement
3. (1, 0) recently used but clean – probably will be used again soon
4. (1, 1) recently used and modified – probably will be used again
soon and need to write out before replacement
When page replacement called for, use the clock scheme but
use the four classes replace page in lowest non-empty class
Might need to search circular queue several times
Counting Algorithms
Keep a counter of the number of references that have been made

to each page
Not common
Lease 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
Page-Buffering Algorithms
Keep a pool of free frames, always
Then frame available when needed, not found at fault time
Read page into free frame and select victim to evict and add
to free pool
When convenient, evict victim
Possibly, keep list of modified pages
When backing store otherwise idle, write pages there and set
to non-dirty
Possibly, keep free frame contents intact and note what is in them
If referenced again before reused, no need to load contents
again from disk
Generally useful to reduce penalty if wrong victim frame
selected
Applications and Page Replacement
All of these algorithms have OS guessing about future page

access
Some applications have better knowledge – i.e. databases
Memory intensive applications can cause double buffering
OS keeps copy of page in memory as I/O buffer
Application keeps page in memory for its own work
Operating system can given direct access to the disk, getting out
of the way of the applications
Raw disk mode
Bypasses buffering, locking, etc
Allocation of Frames
Each process needs minimum number of frames
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
Maximum of course is total frames in the system
Two major allocation schemes
fixed allocation
priority allocation
Many variations
Fixed Allocation
Equal allocation – For example, if there are 100 frames (after
allocating frames for the OS) and 5 processes, give each process
20 frames
Keep some as free frame buffer pool
Proportional allocation – Allocate according to the size of process

Dynamic as degree of multiprogramming, process sizes
change
m = 64
si = size of process pi s1 = 10
S =  si s2 = 127
m = total number of frames a1 =
10
´ 62 » 4
137
s
ai = allocation for pi = i  m 127
S a2 = ´ 62 » 57
137
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
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
But then process execution time can vary greatly
But greater throughput so more common
Local replacement – each process selects from only its own

set of allocated frames
More consistent per-process performance
But possibly underutilized memory
Non-Uniform Memory Access
So far all memory accessed equally
Many systems are NUMA – speed of access to memory varies
Consider system boards containing CPUs and memory,
interconnected over a system bus
Optimal performance comes from allocating memory “close to”
the CPU on which the thread is scheduled
And modifying the scheduler to schedule the thread on the
same system board when possible
Solved by Solaris by creating lgroups
 Structure to track CPU / Memory low latency groups
 Used my schedule and pager
 When possible schedule all threads of a process and
allocate all memory for that process within the lgroup
Thrashing
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
This leads to:
 Low CPU utilization
 Operating system thinking that it needs to increase the
degree of multiprogramming
 Another process added to the system
Thrashing  a process is busy swapping pages in and out
Thrashing (Cont.)
Demand Paging and Thrashing
Why does demand paging work?
Locality model
Process migrates from one locality to another
Localities may overlap
Why does thrashing occur?

 size of locality > total memory size
Limit effects by using local or priority page replacement
Locality In A Memory-Reference Pattern
Working-Set Model
  working-set window  a fixed number of page references
Example: 10,000 instructions
WSSi (working set of Process Pi ) =
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
Approximation of locality
if D > m  Thrashing
Policy if D > m, then suspend or swap out one of the processes
Keeping Track of the Working Set
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
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
Page-Fault Frequency
More direct approach than WSS
Establish “acceptable” page-fault frequency (PFF) rate
and use local replacement policy
If actual rate too low, process loses frame
If actual rate too high, process gains frame
Working Sets and Page Fault Rates
n Direct relationship between working set of a process and its
page-fault rate
n Working set changes over time
n Peaks and valleys over time
Memory-Mapped Files
n Memory-mapped file I/O allows file I/O to be treated as routine
memory access by mapping a disk block to a page in memory
n A file is initially read using demand paging
l A page-sized portion of the file is read from the file system into
a physical page
l Subsequent reads/writes to/from the file are treated as
ordinary memory accesses
n Simplifies and speeds file access by driving file I/O through
memory rather than read() and write() system calls
n Also allows several processes to map the same file allowing the
pages in memory to be shared
n But when does written data make it to disk?
l Periodically and / or at file close() time
l For example, when the pager scans for dirty pages
Memory-Mapped File Technique for all I/O
n Some OSes uses memory mapped files for standard I/O

n Process can explicitly request memory mapping a file via mmap()
system call
l Now file mapped into process address space
n For standard I/O (open(), read(), write(), close()), mmap
anyway
l But map file into kernel address space
l Process still does read() and write()
 Copies data to and from kernel space and user space
l Uses efficient memory management subsystem
 Avoids needing separate subsystem
n COW can be used for read/write non-shared pages
n Memory mapped files can be used for shared memory (although
again via separate system calls)
Memory Mapped Files
Shared Memory via Memory-Mapped I/O
Shared Memory in Windows API
n First create a file mapping for file to be mapped
l Then establish a view of the mapped file in process’s virtual
address space
n Consider producer / consumer
l Producer create shared-memory object using memory mapping
features
l Open file via CreateFile(), returning a HANDLE
l Create mapping via CreateFileMapping() creating a
named shared-memory object
l Create view via MapViewOfFile()
n Sample code in Textbook
Allocating Kernel Memory
n Treated differently from user memory
n Often allocated from a free-memory pool
l Kernel requests memory for structures of varying sizes
l Some kernel memory needs to be contiguous
 I.e. for device I/O
Buddy System
n Allocates memory from fixed-size segment consisting of physically-
contiguous pages
n Memory allocated using power-of-2 allocator
l Satisfies requests in units sized as power of 2
l Request rounded up to next highest power of 2
l When smaller allocation needed than is available, current chunk
split into two buddies of next-lower power of 2
 Continue until appropriate sized chunk available
n For example, assume 256KB chunk available, kernel requests 21KB
l Split into AL and AR of 128KB each
 One further divided into BL and BR of 64KB
– One further into CL and CR of 32KB each – one used to
satisfy request
n Advantage – quickly coalesce unused chunks into larger chunk
n Disadvantage - fragmentation
Buddy System Allocator
Slab Allocator
n Alternate strategy
n Slab is one or more physically contiguous pages
n Cache consists of one or more slabs
n Single cache for each unique kernel data structure
l Each cache filled with objects – instantiations of the data
structure
n When cache created, filled with objects marked as free
n When structures stored, objects marked as used
n If slab is full of used objects, next object allocated from empty
slab
l If no empty slabs, new slab allocated
n Benefits include no fragmentation, fast memory request
satisfaction
Slab Allocation
Slab Allocator in Linux
n For example process descriptor is of type struct task_struct
n Approx 1.7KB of memory
n New task -> allocate new struct from cache
l Will use existing free struct task_struct
n Slab can be in three possible states
1. Full – all used
2. Empty – all free
3. Partial – mix of free and used
n Upon request, slab allocator
1. Uses free struct in partial slab
2. If none, takes one from empty slab
3. If no empty slab, create new empty
Slab Allocator in Linux (Cont.)
n Slab started in Solaris, now wide-spread for both kernel mode and
user memory in various OSes
n Linux 2.2 had SLAB, now has both SLOB and SLUB allocators
l SLOB for systems with limited memory
 Simple List of Blocks – maintains 3 list objects for small,
medium, large objects
l SLUB is performance-optimized SLAB removes per-CPU
queues, metadata stored in page structure
Other Considerations -- Prepaging
n Prepaging
l To reduce the large number of page faults that occurs at
process startup
l Prepage all or some of the pages a process will need, before
they are referenced
l But if prepaged pages are unused, I/O and memory was wasted
l Assume s pages are prepaged and α of the pages is used
 Is cost of s * α save pages faults > or < than the cost of
prepaging
s * (1- α) unnecessary pages?
 α near zero  prepaging loses
Other Issues – Page Size
n Sometimes OS designers have a choice
l Especially if running on custom-built CPU
n Page size selection must take into consideration:
l Fragmentation
l Page table size
l Resolution
l I/O overhead
l Number of page faults
l Locality
l TLB size and effectiveness
n Always power of 2, usually in the range 2 12 (4,096 bytes) to 222
(4,194,304 bytes)
n On average, growing over time
Other Issues – TLB Reach
n TLB Reach - The amount of memory accessible from the TLB
n TLB Reach = (TLB Size) X (Page Size)
n Ideally, the working set of each process is stored in the TLB

l Otherwise there is a high degree of page faults
n Increase the Page Size

l This may lead to an increase in fragmentation as not all
applications require a large page size
n Provide Multiple Page Sizes

l This allows applications that require larger page sizes the
opportunity to use them without an increase in fragmentation
Other Issues – Program Structure
n Program structure
l int[128,128] data;
l Each row is stored in one page
l Program 1
for (j = 0; j <128; j++)
for (i = 0; i < 128; i++)
data[i,j] = 0;
128 x 128 = 16,384 page faults
l Program 2
for (i = 0; i < 128; i++)
for (j = 0; j < 128; j++)
data[i,j] = 0;
128 page faults
Other Issues – I/O interlock
n I/O Interlock – Pages must

sometimes be locked into memory
n Consider I/O - Pages that are used
for copying a file from a device
must be locked from being selected
for eviction by a page replacement
algorithm
n Pinning of pages to lock into
memory
Operating System Examples
n Windows
n Solaris
Windows
n Uses demand paging with clustering. Clustering brings in pages
surrounding the faulting page
n Processes are assigned working set minimum and working set
maximum
n Working set minimum is the minimum number of pages the
process is guaranteed to have in memory
n A process may be assigned as many pages up to its working set
maximum
n When the amount of free memory in the system falls below a
threshold, automatic working set trimming is performed to
restore the amount of free memory
n Working set trimming removes pages from processes that have
pages in excess of their working set minimum
Solaris
n Maintains a list of free pages to assign faulting processes
n Lotsfree – threshold parameter (amount of free memory) to
begin paging
n Desfree – threshold parameter to increasing paging
n Minfree – threshold parameter to being swapping
n Paging is performed by pageout process
n Pageout scans pages using modified clock algorithm
n Scanrate is the rate at which pages are scanned. This ranges
from slowscan to fastscan
n Pageout is called more frequently depending upon the amount of
free memory available
n Priority paging gives priority to process code pages
Solaris 2 Page Scanner
End of Chapter 9
Operating System Concepts – 9th Edition Silberschatz, Galvin and Gagne ©2013

OS-Chap-9 Virtual-Memory

Uploaded by

Copyright:

Available Formats

OS-Chap-9 Virtual-Memory

Uploaded by

Document Information

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

OS-Chap-9 Virtual-Memory

Uploaded by

Copyright:

Available Formats

Chapter 9: Virtual Memory

To describe the benefits of a virtual memory system.

The concept of “Swapping” is usually used when we swap an entire

The concept of “Paging” is usually used when we swap a page from

Virtual address space – Logical view of how process is stored in

Part of the space on disk

What is the difference between logical and virtual memory?

Suppose a process needs ‘D’. It

Another reason for valid-invalid

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

Page table with valid / invalid bit. What if C is not in memory

What if the source block and the second block

Why zero-out a page before allocating it?

There might be no free frames because frames can be used by

What to do in this case?

Prevent over-allocation of memory by modifying page-fault service

If a dirty bit is 1, this means the page is updated, therefore, if

Process 1(User 1) needs to load M. This leads

1. Select a victim page and swap it out.

Frame-allocation algorithm determines:

In general, more frames means less page faults.

However, there is Belady’s Anomaly which we will explain later.

Look next slide

Number of page faults = 15

In this example, number of page faults when total number of frames is 4

Optimal Algorithm (OPT) cannot suffer from Belady’s Anomaly.

Number of page faults = 9

Use past knowledge rather than future.

Generally good algorithm and frequently used.

Number of page faults = 12

Improve algorithm by using reference bit and modify bit (if

Keep a counter of the number of references that have been made

Lease Frequently Used (LFU) Algorithm: replaces page with

Most Frequently Used (MFU) Algorithm: based on the argument

All of these algorithms have OS guessing about future page

Proportional allocation – Allocate according to the size of process

Use a proportional allocation scheme using priorities rather

If process Pi generates a page fault,

Local replacement – each process selects from only its own

Thrashing  a process is busy swapping pages in and out

Why does thrashing occur?

n Some OSes uses memory mapped files for standard I/O

n TLB Reach - The amount of memory accessible from the TLB

n TLB Reach = (TLB Size) X (Page Size)

n Ideally, the working set of each process is stored in the TLB

n Increase the Page Size

n Provide Multiple Page Sizes

128 x 128 = 16,384 page faults

128 page faults

n I/O Interlock – Pages must

You might also like