Intel 80287, 80286 manual Memory Segmentation and Segment Registers

80286 BASE ARCHITECTURE

stack frame. The use of these general-purpose registers for operand addressing is discussed in section 2.3.3, "Index, Pointer, and Base Registers." Register usage for individual instructions is discussed in chapters 3 and 4.

As shown in figure 2-4, eight byte registers overlap four of the 16-bit general registers. These registers are named AH, BH, CH, and DH (high bytes); and AL, BL, CL, and DL (low bytes); they overlap AX, BX, CX, and DX. These registers can be used either in their entirety or as individual 8-bit regis- ters. This dual interpretation simplifies the handling of both 8- and 16-bit data elements.

2.3.2 Memory Segmentation and Segment Registers

Complete programs generally consist of many different code modules (or segments), and different types of data segments. However, at any given time during program execution, only a small subset of a program's segments are actually in use. Generally, this subset will include code, data, and possibly a stack. The 80286 architecture takes advantage of this by providing mechanisms to support direct access to the working set of a program's execution environment and access to additional segments on demand.

At any given instant, four segments of memory are immediately accessible to an executing 80286 program. The segment registers DS, ES, SS, and CS are used to identify these four current segments. Each of these registers specifies a particular kind of segment, as characterized by the associated mnemonics ("code," "stack," "data," or "extra") shown in figure 2-4.

An executing program is provided with concurrent access to the four individual segments of memory- a code segment, a stack segment, and two data segments-by means of the four segment registers. Each may be said to select a segment, since it uniquely determines the one particular segment from among the numerous segments in memory, which is to be immediately accessible at highest speed. Thus, the 16-bit contents of a segment register is called a segment selector.

Once a segment is selected, a base address is associated with it. To address an element within a segment, a 16-bit offset from the segment's base address must be supplied. The 16-bit segment selector and the 16-bit offset taken together form the high and low order halves, respectively, of a 32-bit virtual address pointer. Once a segment is selected, only the lower 16-bits of the pointer, called the offset, generally need to be specified by an instruction. Simple rules define which segment register is used to form an address when only a 16-bit offset is specified.

An executing program requires, first of all, that its instructions reside somewhere in memory. The segment of memory containing the currently executing sequence of instructions is known as the current code segment; it is specified by means of the CS register. All instructions are fetched from this code segment, using as an offset the contents of the instruction pointer (IP). The CS:IP register combination therefore forms the full 32-bit pointer for the next sequential program instruction. The CS register is manipulated indirectly. Transitions from one code segment to another (e.g., a procedure call) are effected implicitly as the result of control-transfer instructions, interrupts, and trap operations.

Stacks playa fundamental role in the 80286 architecture; subroutine calls, for example, involve a number of implicit stack operations. Thus, an executing program will generally require a region of memory for its stack. The segment containing this region is known as the current stack segment, and it is specified by means of the SS register. All stack operations are performed within this segment, usually in terms of address offsets contained in the stack pointer (SP) and stack frame base (BP) registers. Unlike CS, the SS register can be loaded explicitly for dynamic stack definition.

2-8