4.1 Memory Management

Memory management is the part of the operating system that controls the computer's main memory (RAM): it keeps track of which parts are in use and by which process, gives memory to programs when they start or ask for more, and takes it back when they finish. Because RAM is limited and shared by many running programs, good memory management decides how fast and how reliably the whole system runs. Example: when a records clerk at a land registry office in Machakos opens a large scanned title deed while the payroll system is running, the operating system finds free memory for the new document without disturbing the payroll program.

4.1.1 Definition

Memory management is the function of an operating system that handles and coordinates primary memory, moving processes and data between main memory and secondary storage during execution. It decides which process gets memory, when and how much, tracks every memory location as allocated or free, and frees memory when a process ends.

Key terms

• Main memory (RAM): Fast, volatile storage that holds the programs and data the CPU is using now.
• Logical (virtual) address: The address a program uses, generated by the CPU.
• Physical address: The actual location in RAM where the data is stored.
• Address space: The set of addresses a process may use.
• Memory allocation and deallocation: Giving memory to a process and taking it back when the process no longer needs it.
• Swapping: Moving a whole process, or parts of it, between RAM and disk to free memory for others.

4.1.2 Objectives of Memory management

The memory manager has several objectives, all aimed at letting many programs share limited RAM safely and efficiently.

• Relocation: Programs can be loaded anywhere in memory, and moved during execution, because logical addresses are translated to physical addresses at run time.
• Protection: A process must not read or change memory that belongs to another process or to the operating system; the hardware checks every access.
• Sharing: Where useful, several processes may share the same memory, such as one copy of a library or a shared data area, under control.
• Logical organisation: Programs are written as modules (code, data, stack); memory management supports this view, for example through segmentation.
• Physical organisation: Data is moved between main memory and secondary storage so that programs larger than RAM, or more programs than fit at once, can run.
• Efficient utilisation: Memory is kept as fully and usefully occupied as possible, reducing waste through fragmentation.
• Support for multiprogramming: Several processes are kept in memory at once so that the CPU always has work.

4.1.3 Components of the Memory Management unit

The memory management unit (MMU) is the hardware, usually built into the CPU, that translates the logical addresses used by programs into physical addresses in RAM and enforces memory protection. The operating system sets up the tables the MMU uses, and the MMU applies them on every memory access.

• Relocation (base) register: Holds the starting physical address of the process; it is added to every logical address to give the physical address.
• Limit register: Holds the size of the process's address space; any address beyond it causes a trap to the operating system, which protects other processes.
• Page table and page table base register: The page table maps each page of a process to a frame in RAM; the base register points to the current process's page table.
• Translation lookaside buffer (TLB): A small, fast cache of recent page-to-frame translations, so most addresses are translated without reading the page table in memory.
• Segment table: In segmented systems, holds the base and limit of each segment.
• Protection and status bits: Bits in each table entry mark a page as read-only, read-write or executable, present or not present in RAM, and modified (dirty) or not.
• Page fault handling: When a program uses a page not in RAM, the MMU raises a page fault and the operating system loads the page from disk.