Memory Management

Memory Management Chapter 7 CSNB224 – AAG 2007

Memory Management • In uniprogramming system, memory is divided into two parts, one part for the OS and one part for the program that currently being executed (user part). • In multiprogramming system, the user part of the memory must be further subdivided to accommodate multiple processes. • The task of subdivision is carried out dynamically by the OS memory management. CSNB224 – AAG 2007

Memory Management • Memory needs to be allocated efficiently to pack as many processes into memory as possible. • This is because effective memory management is important in multiprogramming system. CSNB224 – AAG 2007

Memory Management Requirements • Five requirements of memory management: • Relocation • Protection • Sharing • Logical Organization • Physical Organization CSNB224 – AAG 2007

Memory Management Requirements • Relocation • Programmer does not know where the program will be placed in memory when it is executed • While the program is executing, it may be swapped to disk and returned to memory at a different location (relocate process at different area) • The location for the process is unknown and the process must be allowed to moved in and out due to swapping. CSNB224 – AAG 2007

Memory Management Requirements • These facts raise some technical concerns related to addressing, as illustrated in Figure 7.1. • Figure 7.1 shows a process image. • The OS will need to know the location of: • process control information • the execution stack • the entry point to begin the execution of the program for the process. • Because OS is managing memory and is responsible for bringing this process into memory, these addresses are easy to recognize. CSNB224 – AAG 2007

Memory Management Requirements • However the processor need to deal with memory references within the program. • Branch inst. contain an address to reference the instruction to be executed next. • Data reference inst. contain the address of the byte or word of data referenced. • So the processor and OS must be able to translate the memory references found in the code of the program into actual physical memory address • reflecting the current location of the program in memory. CSNB224 – AAG 2007

Memory Management Requirements • Protection • Processes should not be able to reference memory locations in another process without permission • Impossible to check absolute addresses in programs since the program could be relocated. • All memory references must be checked at run time to ensure that it refer to the only memory space allocated to the process CSNB224 – AAG 2007

Memory Management Requirements • Memory protection requirement must be satisfied by the processor rather than OS. • OS cannot anticipate all of the memory references a program will make • Time consuming to screen each program in advance for possible memory referenced violation. • It is only possible to assess the permissibility of a memory reference at the time of execution of the instruction making reference. CSNB224 – AAG 2007

Memory Management Requirements • Sharing • Allow several processes to access the same portion of memory. • So, any protection mechanism must have the flexibility, example: • If a number of processes are executing the same program, it is advantageous to allow each process access to the same copy of the program rather than have their own separate copy  e.g. ECHO procedure CSNB224 – AAG 2007

Memory Management Requirements • Memory management system must therefore allow controlled access to shared areas of memory without compromising essential protection. CSNB224 – AAG 2007

Memory Management Requirements • Logical Organization • Programs are written in modules • Modules can be written and compiled independently • Different degrees of protection given to modules (read-only, execute-only) • Share modules ##Read on your own CSNB224 – AAG 2007

Memory Management Requirements • Physical Organization • Memory available for a program plus its data may be insufficient • Overlaying allows various modules to be assigned the same region of memory • Programmer does not know how much space will be available ##Read on your own CSNB224 – AAG 2007

Memory Partitioning • Principal operation of memory management is to bring programs into memory for execution by the processor. • In most modern multiprogramming system it invokes virtual memory (VM) that use both segmentation and paging techniques. • First look at simpler technique that do not use VM  Partitioning • Two types of partitioning: • Fixed Partitioning • Dynamic Partitioning CSNB224 – AAG 2007

Fixed Partitioning • Most schemes of memory management, assume that the OS occupies some fixed partition of memory and the rest of memory is available for use by multiple processes. • The simplest scheme for managing this available memory is to partition it into regions with fixed boundaries. • Figure 7.2 shows example of two alternatives for fixed partitioning: • Equal-size partition • Unequal-size partition CSNB224 – AAG 2007

Fixed Partitioning • Equal-size partition (Figure 7.2a) • any process whose size is less than or equal to the partition size can be loaded into an available partition • if all partitions are full, the operating system can swap a process out of a partition CSNB224 – AAG 2007

Fixed Partitioning • Two difficulties: • A program may not fit in a partition. The programmer must design the program with “overlays” • Program and data are organized in such away that various modules can be assigned the same region of memory with a main program responsible for switching in and out as needed • Memory use is inefficient. Any program, no matter how small, occupies an entire partition. This is called internal fragmentation. • Eg: Program length 2MB, it occupies an 8MB partition. So 6MB is internal fragmentation. CSNB224 – AAG 2007

Fixed Partitioning • Unequal-size partition (Figure 7.2b) • A program as large as 16MB can be accommodate without need to overlays it. • Partition smaller than 8MBallow smaller program to be accommodate with less internal fragmentation. CSNB224 – AAG 2007

Fixed Partitioning - Placement Algorithm  Where to put / load process in memory. • Equal-size partitions • If there is an available partition, a process can be loaded into that partition • because all partitions are of equal size, it does not matter which partition is used • If all partitions are occupied by blocked processes, choose one process to swap out to make room for the new process CSNB224 – AAG 2007

Unequal-size partitions: Two ways: Use of multiple queues Assign each process to the smallest partition within which it will fit A queue for each partition size Tries to minimize internal fragmentation Problem: some queues will be empty if no processes within a size range is present Fixed Partitioning - Placement Algorithm CSNB224 – AAG 2007

Fixed Partitioning - Placement Algorithm CSNB224 – AAG 2007

Use of a single queue When its time to load a process into memory the smallest available partition that will hold the process is selected If all partitions is occupied: Preference to swapping out the smallest partition that will hold the incoming process. Also have to consider other factors such as priority and blocked vs. ready process. Increases the level of multiprogramming at the expense of internal fragmentation Fixed Partitioning - Placement Algorithm CSNB224 – AAG 2007

Fixed Partitioning - Placement Algorithm CSNB224 – AAG 2007

Fixed Partitioning • The used of unequal-size partition provides a degree of flexibility to fixed partitioning. • Fixed partitioning schemes are relatively simple and require minimal OS software and processing overhead. • But it still have disadvantages: • The number of partitions specified at system generation time limits the number of active (not suspended) processes in the system. • Because partition sizes are preset at system generation time, small jobs will not utilize partition space efficiently. CSNB224 – AAG 2007

Fixed Partitioning • The used of fixed partitioning is almost unknown today. • Example of OS that use this technique was an early IBM mainframe OS, OS/MFT (Multiprogramming with a Fixed Number of Tasks. CSNB224 – AAG 2007

Dynamic Partitioning • Partitions are of variable length and number • Each process is allocated exactly as much memory as it requires. • Used in IBM’s OS/MVT (Multiprogramming with a Variable number of Tasks) • An example using 1 MB (1024KB) of memory is shown in Figure 7.4. CSNB224 – AAG 2007

Dynamic Partitioning CSNB224 – AAG 2007

Dynamic Partitioning • Initially memory is empty, except for the OS(refer figure a) • The first three process are loaded in, starting where the OS ends and occupying just enough space for each process (refer figure b,c,d). • This leaves a “hole” at the end of memory that is too small for a fourth process. CSNB224 – AAG 2007

Dynamic Partitioning CSNB224 – AAG 2007

Dynamic Partitioning • At some point none of the processes in memory is ready. OS swaps process 2 (figure e), which leaves sufficient room to load a new process, process 4 (figure f). Because process 4 is smaller than process 2, another small hole is created. • Later, a point is reached at which none of the processes in memory is ready, but process 2, in the Ready-Suspend state is available. • So OS swaps process 1 out (figure g) and swaps process 2 back in (figure h). CSNB224 – AAG 2007

Dynamic Partitioning • Eventually holes are formed in the memory. • Memory will become more and more fragmented, and memory utilization declines. • This hole is called external fragmentation. • Referring to fact that the memory is external to all partitions becomes increasingly fragmented. • Technique to overcome external fragmentation is called compaction. • Time to time OS shifts the processes so that they are contiguous and so that all of free memory is together in one block. CSNB224 – AAG 2007

Dynamic Partitioning • Example in figure 7.4h, compaction will result in a block of free memory of length 256K, and this maybe sufficient to load an additional process. • Difficulties of compaction is time consuming procedure and wasteful of processor time. CSNB224 – AAG 2007

Dynamic Partitioning - Placement Algorithm  Where to put / load process in memory. • OS must decide which free block to allocate to a process. • 3 placement algorithms that might be considered: • Best Fit • First Fit • Next Fit CSNB224 – AAG 2007

Dynamic Partitioning - Placement Algorithm • Best Fit • Chooses the block that is closest in size to the request. • Chooses smallest hole • First-fit: • Begins to scan memory from the beginning and chooses the first available block that is large enough. • Chooses first hole from beginning CSNB224 – AAG 2007

Dynamic Partitioning - Placement Algorithm • Next-fit: • Begins to scan memory from the last placement and chooses the next available block that is large enough. • Chooses first hole from last placement CSNB224 – AAG 2007

Dynamic Partitioning- Placement Algorithm • Which of these algorithms is best will depend on the exact sequence of process swapping that occurs and the size of those processes. • Remarks for each of the algorithms: • First Fit • The simplest and usually the best and fastest. • Next-Fit • Tends to produce slightly worse result than first fit. • More frequently lead to an allocation from a free block at the end of memory • Compaction maybe required more frequently with next fit CSNB224 – AAG 2007

Dynamic Partitioning- Placement Algorithm • Best Fit • Worse performer, because this algorithm will looks for the smallest block that will satisfy the requirement, it guarantees that the fragment left behind is small as possible. • The memory is quickly littered by blocks too small to satisfy memory allocation requests. • Compaction must be done more frequently than the other two algorithms. CSNB224 – AAG 2007

Replacement Algorithm • When all processes in memory are blocked, the OS must choose which process to replace • A process must be swapped out (to a Blocked-Suspend state) and be replaced by a new process or a process from the Ready-Suspend queue • We will discuss later such algorithms for memory management schemes using virtual memory CSNB224 – AAG 2007

Buddy System • Both fixed and dynamic partitioning schemes have drawbacks. • A fixed partitioning scheme limits the number of active processes and may used space inefficiently if there is a poor match between available partition sizes and process sizes. • A dynamic partitioning scheme is more complex to maintain and includes the overhead of compaction. • An interesting compromise is the buddy system. CSNB224 – AAG 2007

Buddy System • Memory blocks are available of size 2K , L<=K<=U, where: 2L= smallest size block that is allocated 2U= largest size block that is allocated, generally 2U is the size of the entire memory available for allocation. • Begin with entire space available is treated as a single block of 2U • If a request of size s such that 2U-1< s <= 2U, entire block is allocated • Otherwise block is split into two equal buddies • Process continues until smallest block greater than or equal to s is generated CSNB224 – AAG 2007

Buddy System • At any time, the buddy system maintains a list of holes (unallocated blocks) of each size 2i . • A hole may be removed from the (i+1) list by splitting it in half to create two buddies of size 2i in the i list. • Whenever a pair of buddies on the i list both become unallocated, they are removed from that list and combine back into a single block on the (i+1) list. CSNB224 – AAG 2007

Buddy System • Example shows in Figure 7.6 using 1MB initial block. • The first request, A, is for 100kb is needed. • The initial block is divided into two 512k buddies. • The first of these is divided into two 256K buddies, and the first these is divided into two 128K buddies, one of which is allocated to A. • The next request,B, requires 256K block. Such block is already available and is allocated. • The process continues with splitting and combine back into 256K, which immediately combine back with its buddy. CSNB224 – AAG 2007

Buddy System • Figure 7.7 shows a binary tree representation of the buddy allocation immediately after the release B request. • The leaf nodes represent the current partitioning of the memory. • If two buddies are leaf nodes, then at least one must be allocated, otherwise they would be combine into a larger block. CSNB224 – AAG 2007

Buddy System • The buddy system is a reasonable compromise to overcome the disadvantages of both partitioning schemes. • In contemporary OS, VM based on paging and segmentation is superior. • This buddy system has found application in parallel systems as an efficient means of allocation and release for parallel programs. • A modified form of buddy system is used for UNIX kernel memory allocation. CSNB224 – AAG 2007

Relocation • When program loaded into memory the actual (absolute) memory locations are determined • A process may occupy different partitions which means different absolute memory locations during execution (from swapping) • Compaction will also cause a program to occupy a different partition which means different absolute memory locations • Thus the locations (instruction and data) referenced by a process are not fixed. CSNB224 – AAG 2007

Memory Management