MEMORY MANAGEMENT
Introduction
That part of the OS that manages memory is called the memory
manager. Its job is to keep track of which parts of memory are in
use, allocate memory to processes when they need it and
manage the swapping of processes between the main memory
and the hard disk when the memory is not big enough to hold all
the processes.
One of the earliest memory management policies was to use
virtual memory. The basic idea behind virtual memory is that
the combined size of the program, its data and the stack may
exceed the amount of physical memory available for it. What the
OS does is to keep those parts of the program currently in use in
memory and the rest on the disk. The two most common
technique of memory are paging and segmentation.
In early systems, another more straightforward method was to
use overlays. A large program would be split by the
programmers into pieces called overlays. The first overlay will
be loaded in memory, start running and when it was done it
would call another overlay in memory. Although the OS did the
actual work of swapping overlays in and out of memory, the
splitting of the program into small modules was the
programmer’s responsibility.
Memory Maps / Memory Allocation Methods
a) Fixed Partition Memory
Partition 4
Input Queue
Partition 3
(processes)
Partition 2
Partition 1
OS
Internal Memory
(RAM)
In a fixed memory partition system, the memory is divided into
unequal length partitions and, as each partition becomes free, the
job closest to the front of the queue that fits into the partition
will be loaded into it and executed.
Note that since a job will not exactly fit a partition size, there
will always be some wasted space referred to as internal
fragmentation. Another disadvantage is that a process may not
be able to run because it is larger than the largest available
partition.
b) Variable Partition Memory
In this scheme, the partitions are allowed to be variable in size at
load time. This is done in order to allocate the process the exact
amount of memory it requires. Processes are loaded into
consecutive areas until the memory is filled. When a process
terminates the space it occupied is freed and becomes available
for other processes.
E E E E
D
C C C C
B B Coalescing of
A A A “holes”
OS OS OS OS
When a process terminates the freed space is termed a
“hole” (Process D has terminated).
The distribution of the free memory space (holes) in the way
shown in the diagram is called external fragmentation.
It is possible to combine all the holes into one big chunk by
moving the processes downward as far as possible. This is
known as memory compaction and involves a substantial
amount of processing since all active process would have to be
suspended during the reshuffling.
Relocation and Protection
Multi programming introduces two problems that must be
solved, namely relocation and protection.
Relocation can be solved in two ways:
1) by a relocating loader, which modifies the actual jump
instructions so that the effect of the load address is taken
into account.
2) by the use of a base register (also called a relocation
register). The base register is a CPU register which is
loaded with the starting address in memory of where the
program is in memory. All “jumps” are then made relative
to this address, i.e. the effective address = value of base +
jump address.
Note: Some systems also have a limit register. This would be
loaded with the length of the partition (area in memory for
storing the process). Addresses are then checked against the
limit register to make sure they do not attempt to address
memory outside the current partition.
Note also that the base register technique has the advantage over
relocating loaders that a program can be moved in memory after
it has started execution. After it has been moved all that needs to
be done to make it ready to run is change the value of the base
register.
Paging
In a paged system each process is divided into a number of fixed
size chunks called pages. The memory space is viewed as a set
of page frames of the same size. The loading process now
involves transferring each process page to some memory page
frame.
A B C Internal Memory
Page 1 Page 1 Page 1 Page 1 of A
Aprocess
Page 2 Page 2 Page 2 Page 2 A
Page 3 Page 3 Page 3 Page 3
Page 4 Page 4 Page 4
Page 5 Page 5
Page 6 Page 6
Page 1 of B
Process
Page 2B
Page 3
Page 1 of C
CProcess
Page 2 C
Page 3
Page 4
The above diagram shows three processes of different sizes.
They are loaded in memory in consecutive locations.
If process B terminates and anther process (say process D)
consisting of 5 pages needs to be loaded into memory, then we
may end up with the situation shown below
Internal Memory
Page 1 of process A
Page 2
Page 3
Page 4
Page 5
Page 6
Page 1 of Process D
Page 2
Page 3
Page 1 of Process C
Page 2
Page 3
Page 4
Page 4 of process D
Page 5
With paging, an address is divided into 2 parts, a page number
and a displacement, or offset. This is called page addressing.
Memory Address
Page number Displacement
In the above example, the top 5 bits are used for the page
number and 11 bits are used for the displacement.
Hence with such a system, a maximum of 25 or 32 pages can be
supported. Each page can be up to 211 or 2048 memory
locations long.
Note that the relocation problem can be solved by simply
changing the page number when a page is loaded into a different
page frame.
