OS Fall’02 Memory Management Operating Systems Fall 2002.

OS Fall’02

Memory Management

Operating Systems Fall 2002

OS Fall’02

Background A compiler creates an executable code An executable code must be brought into

main memory to runOperating system maps the executable code onto the main memoryThe “mapping” depends on the hardware

Hardware accesses the memory locations in the course of the execution

OS Fall’02

Memory management Principal operation: Bringing

programs into the memory for execution by the processor

Requirements and issues:RelocationProtectionSharingLogical and physical organization

OS Fall’02

Relocation Compiler generated addresses are

relative A process can be loaded at different

memory location, swapped in/outActual physical mapping is not known at compile time

Simple if processes are allocated contiguous memory ranges

Complicated if paging is used

OS Fall’02

Protection Processes must be unable to

reference addresses of other processes or those of the operating system

Cannot be enforced at compile time (relocation)

Each memory access must be checked for validity

Enforced by hardware

OS Fall’02

Sharing Protection must be flexible enough

to allow controlled sharing of the same portion of the main memory

E.g., all running instances of a text editor can share the same code

OS Fall’02

Logical organization Main memory is organized as a linear

one-dimensional address space Programs are typically organized into

modulescan be written/compiled independently, have different degrees of protection, shared

A mechanism is desirable for supporting a logical structure of the user programs

OS Fall’02

Physical organization Two level hierarchical storage

Main memory: fast, expensive, scarce, volatileSecondary storage: slow, cheap, abundant, persistent

Memory management involves bi-directional information flow between the two

OS Fall’02

Memory management techniques

Partitioningfixed, dynamicnow obsoleteuseful to demonstrate basic principles

Paging and segmentation Paging and segmentation with

virtual memory

OS Fall’02

Fixed partitioning Main memory is divided into a

number of static partitionsall partitions of the same sizea collection of different sizes

A process can be loaded into a partition of equal or greater size

A process cannot be scattered among many partitions

OS Fall’02

Fixed partitioningOperating System

8 M

12 M

8 M

8 M

6 M

4 M

2 M

8 M

8 M

8 M

8 M

8 MOperating System

Internal fragmentation

wasted space is internal to an allocated region

Executing big programs requires explicit memory management by the programmer

overlays

OS Fall’02

Placement algorithms

NewProcesses

OperatingSystem

OperatingSystem

NewProcesses

OS Fall’02

Replacement algorithms If all partitions are occupied some

process should be swapped out Select a process that

occupies the smallest partitions that will hold the incoming processblocked process

OS Fall’02

Dynamic partitioning Partitions created dynamically Each process is loaded into a

partition of exactly the same size as that process

OS Fall’02

Dynamic partitioning example

Operating System

128 K

896 K

Operating System

Process 1 320 K

576 K

Operating System

Process 1 320 K

Process 2 224 K

352 K

OS Fall’02

Dynamic partitioning example

Operating System

Process 1 320 K

Process 2

Process 3

224 K

288 K

64 K

Operating System

Process 1 320 K

Process 3

224 K

288 K

64 K

Operating System

Process 1 320 K

Process 3 288 K

64 K

Process 4 128 K

96 K

OS Fall’02

Dynamic partitioning example

Operating System

320 K

Process 3 288 K

64 K

Process 4 128 K

96 K

Operating System

Process 3 288 K

64 K

Process 4 128 K

96 K

Process 2 224 k

96 K

OS Fall’02

External fragmentation Situation when the memory which is

external to the allocated partitions becomes fragmented

Can be reduced using compactionwastes the processor time

OS Fall’02

Placement algorithms Free regions are organized into a linked

list ordered by the memory addressessimplifies coalescing of free regionsincreases insertion time

First-fit: allocate the first region large enough to hold the process

optimization: use a roving pointer: next-fit

Best-fit: allocate the smallest region which is large enough to hold the process

OS Fall’02

An example

Lastallocatedblock (14K)

BeforeAfter

8K 8K

12K 12K

22K

18K

6K 6K

8K 8K

14K 14K

6K

2K

36K20K

Next Fit

Free block

Allocated block

Best Fit

First Fit

OS Fall’02

Placement algorithms discussion

First-fit:efficient and simpletends to create more fragmentation near the beginning of the list solved by next-fit

Best-fit:less efficient (might search the whole list)tends to create small unusable fragments

OS Fall’02

Buddy systems Approximates the best-fit principle Efficient search

first (best) fit has a linear complexity

Simplified coalescing Several variants exist

Binary buddy system allocates memory in blocks which are powers of 2

OS Fall’02

Binary buddy system Memory blocks are available in size

of 2K where L <= K <= U and where2L = smallest size of block allocated2U = largest size of block allocated(generally, the entire memory available)

Initially, all the available space is treated as a single block of size 2U

OS Fall’02

Creating buddies A request to allocate a block of size s:

If 2U-1 < s <= 2U, allocate the entire blockOtherwise, split the block into two equal buddies of size 2U-1

If 2U-2 < s <= 2U-1, then allocate the request to one of the buddies, otherwise split furtherProceed until the smallest block is allocated

OS Fall’02

Maintaining buddies Existing non-allocated buddies

(holes) are kept on several listsHoles of size 2i are kept on the ith listHole is removed from the (i+1)th list by splitting it into two bodies and putting them onto the ith listWhenever two adjacent buddies on the ith list are freed they are coalesced and moved to the (i+1)th list

OS Fall’02

Finding a free buddy

get_hole(i):if (i==U+1) then failure;if (list i is empty) {

get_hole(i+1);split hole into buddies;put buddies on list i;

}take first hole on list i;

OS Fall’02

An example

OS Fall’02

Remarks Buddy system is a reasonable

compromise between the fixed and dynamic partitioning

Internal fragmentation is a problemExpected case is about 28% which is high

BS is not used for memory management nowadays

Used for memory management by user level libraries (malloc)

OS Fall’02

Relocation with partitioning

Interrupt tooperating system

Process image inmain memory

Relative address

Absoluteaddress

Text

Data

Stack

Adder

Comparator

Base Register

Bounds Register

OS Fall’02

Paging The most efficient and flexible

method of memory allocation Process memory is divided into fixed

size chunks of the same size, called pages

Pages are mapped onto frames in the main memory

Internal fragmentation is possible with paging (negligible)

OS Fall’02

Paging support Process pages can be scattered all

over the main memory Process page table maintains

mapping of process pages onto frames

Relocation becomes complicatedHardware support is needed to support translation of relative addresses within a program into the memory addresses

OS Fall’02

Paging example01

2

3

4

5

6

7

8

9

10

11

12

13

14

A.0

A.1

A.2

A.3

C.0

C.1

C.2

C.3

D.0

D.1

D.2

D.3

D.4

0

1

2

3

Process A

0

1

2

3

0

1

2

Process B

---

---

---

0

1

2

3

4

0

1

2

3

Process C

7

8

9

10

4

5

6

11

12

Process D

Free Frame List

13

14

OS Fall’02

Address translation Page (frame) size is

a power of 2with page size = 2r, a logical address of l+r bits is interpreted as a tuple (l,r)l = page number, r = offset within the page

Page number is used as an index into the page table

OS Fall’02

Hardware support

Program Paging Main Memory

Logical address

Register

Page Table

PageFrame

Offset

P#

Frame #

Page Table Ptr

Page # Offset Frame # Offset

+

OS Fall’02

Segmentation A program can be divided into segments Segments are of variable size Segments reflect the logical (modular)

structure of a program Text segment, data segment, stack segment…

Similar to dynamic partitioning except segments of the same process can be scattered

subject to external fragmentation

OS Fall’02

Address translation

Maximum segment size is always a power of 2 Process’ segment table maps segment

numbers into their base addresses in the memory

With the maximum segment size of 2r, a logical address of l+r bits is interpreted as a pair (l,r):

l = segment number, r = offset within the segment

l is used as an index into the segment table

OS Fall’02

Hardware support

Base + d

Program Segmentation Main Memory

Virtual Address

Register

Segment Table

Seg

men

t

d

S#

Length Base

Seg Table Ptr

Seg # Offset = d

Segment Table

+

+

OS Fall’02

Remarks The price of paging/segmentation is a

sophisticated hardware But the advantages exceed by far

Paging decouples address translation and memory allocation

Not all the logical addresses are necessary mapped into physical memory at every given moment - virtual memory

Paging and segmentation are often combined to benefit from both worlds

OS Fall’02

Next: Virtual memory