Types Of Addressing - Xerox 560 Reference Manual

Hide thumbs Also See for 560:
Table of Contents

Advertisement

reference address field of the instruction is used to
obtain a word, and the 17 or 20 low-order bits of the
word thus obtained effectively replace the initial ref-
erence address field; then indexing is carried out ac-
cording to the operation code of the instruction. See
Figures 7 and 9, later in this chapter.
6.
Index Reference Address.
If indexing is called for in
the instruction (a value other than zero in bits 12-14
of the instruction), the direct or indirect reference ad-
dress is modified by addition of the displacement value
in the general register (index) called for by the instruc-
tion (after scaling the displacement according to the
instruction type).
This final reference address value
(after indirect addressing, indexing, or both) is defined
as the effective virtual address of the instruction. In-
dexing after indirect addressing is ,ca lied postindexing.
See also Figures 7 and 9, later in this chapter.
7.
Displacements.
Displacements are the 16- to 22-bit
values used in index registers and by byte-string in-
structions to generate effective addresses of the appro-
priate size (byte, halfword, word, or doubleword).
8.
Register Address. If any instruction provides a virtual
address that is a memory reference (i .e., a direct,
indirect, or indexed reference address) in the range 0
through 15, the basic processor does not attempt to read
from or write into main memory locations 0 through 15.
Instead, the four low-order bits of the reference ad-
dress are used as a general register address and the gen-
era
I
reg ister correspond i ng to th is address is used as the
operand location or result destination. Thus, the in-
struction can use any of the first 16 registers in the cur-
renT regisTer biocK as Tne source or an operand, the
location of a direct address, or the destination of a re-
sult.
Such usage is called a "register-to-register"
operation.
9.
Actual Address. This is the address value actually used
by the basic processor to access main memory via the
memory address register (see Figure 5). If the effective
virtual address is in the range 0 through 15 (X10 through
X'F ' ), one of the first 16 general registers in the cur-
rent register block is being addressed. If the basic pro-
cessor is operating in the virtual addressing mode, all
addresses grea ter than 15 (X 1 F I) are transformed (usua IIy
into addresses in a different memory page) by the mem-
ory map into actual addresses. Contrarily, if the basic
processor is operating in either real or real extended
mode, no transformation via the memory map takes place.
10. Effective Address. The effective address is defined as
the final virtual address computed for an instruction.
Note, however, that some instructions do not use the
effective address as a location reference; instead, the
effective address is used to control the operation of
the instruction (as in a shift instruction), to designate
the address of an input/output device (as in an input/
output instruction), or to designate a specific element
of the system (as in a READ DIRECT or WRITE DIRECT
instruction) •
11. Effective Location. An effective location is defined
as the actual location (in main memory or in the current
register block) that is to receive the result of a memory-
referencing instruction, and is referenced by means of
an effective address.
Because an effective address
may be either an actual address or a virtual address,
when applicable, this definition of an effective loca-
tion assumes the transformation of a virtual address into
an actual address.
12. Effective Operand.
An effective operand is defined
as the contents of an actual location (in main memory
or in the current register block) that is to be used as
an operand by a memory-referencing instruction, and
is referred to by means of an effective address.
This
also presupposes the transformation of a virtual address
into an actual address.
TYPES OF ADDRESSING
Except for the special type of addressing performed by some
interrupt and trap instructi ons, all addressing within the
computer system is real, real extended, or virtual.
REAL ADDRESSING
In real addressing, a one-to-one relationship prevails be-
tween the effective virtual address of each instruction
and the actual address used to access main memory. Real
addressing has these characteristics:
1.
Each reference address is a 17-bit word address.
2.
The reference address may be direct or indirect, with
or without postindexing.
3.
Displacements associated with indexing are automati-
cally aligned, as required, using the full 32-bit contents
of the index register.
The final result is truncated to
the left of the high-order bit of the original 17-bit ref-
erence address, and the effective real address is a
16-bit doubleword address, 17-bit word address, 18-bit
halfword address, or a 19-bit byte address.
4.
If indirect addressing is invoked, the 17-bit reference
address in the instruction word is used to access the in-
direct address word in memory. The low-order 17 bits
of this word then replace the reference address of the
instruction word in the calculations described in (3),
above.
5.
Memory mapping and memory access protection are
never invoked.
6.
Memory write protection is automatically invoked.
7.
Leading zeros are automatically appended to the effec-
tive address to generate an actual word address as re-
quired by the main memory.
Main Memory
19

Advertisement

Table of Contents
loading

Table of Contents