Segmented Memory Management

Overview of Segmented Memory Management

In segmented memory management, a program is composed of several segments. Each segment has an independent and complete logical meaning. For example, the main program, subroutine, symbol table, stack, data, etc. can be divided into different segments.
The length of each segment can be different, and each segment can be written, loaded and executed independently. Segmented memory management is divided and managed based on the logical structure of the program, which can better meet the needs of the program.
Segmented memory management supports segmentation viewpoint, making the memory structure of the program clearer and easier to manage, with strong logical integrity.

Segment division: The address space of the program is divided into multiple segments according to logical relationships, and each segment usually contains a functional module (such as a function). The distribution of each segment in memory corresponds to the logical structure of the program.
Segment independence: Each segment has an independent storage space in the memory, and the storage addresses between segments are usually discontinuous, while the addresses within each segment are continuous. The address space of each segment starts addressing from "0".
Its logical address consists of two parts: the segment number and the address within the segment:

When a segment cannot find a large enough free area, the system can merge scattered free areas through mobile technology to facilitate loading larger jobs.
When the job is completed, the main memory space it occupies will be recycled and registered in the free partition table. When recycling, it will also check whether there are adjacent free partitions and merge them into a new free area for management.
The segment table is stored in the main memory. When accessing data or instructions, it may be necessary to access the main memory at least twice. In order to increase the speed, high-speed cache registers are often used to store commonly used segment table entries.

Segment storage management uses dynamic relocation to load jobs. When the job is executed, the conversion from logical address to physical address is realized through the hardware address conversion mechanism.
The table entries in the segment table act as base address registers and limit length registers, responsible for hardware address conversion.

Sharing of segments is a prominent advantage of segment management, which allows multiple processes to share the same segment of code or data. The processes set the same segment name in the segment table and specify appropriate read and write permissions.
For example, in a multi-user system, multiple users can share the same program segment (such as the code of a text editor). If the segment is reentrant, only one copy of the code needs to be kept in main memory.
When sharing, you need to pay attention to the modification and replacement of shared segments to ensure that the shared segments in use will not be replaced out of main memory. The sharing bit can help determine whether the segment is in use.