How does the processor recognize the instruction length?

Question

assembler

Hello. Please help. I want to understand the essence of the processor. In Petzold's book, K.O.D. there is text:
In the computer from chapter 17, all commands (except Stop) occupied 3 bytes - the operation code and a two-byte address. In the 8080 processor, commands are 1, 2, and 3 bytes long. Some commands tell the processor to read a byte from a specific location memory, others - write a byte to a specific memory cell, others - perform some internal operations without accessing the RAM. Having completed the first command, the processor reads the next one from the memory, etc. Taken together , these commands are a computer program whose purpose is to get the processor to do something.
Questions:
1. As I understand it, the length of the command is different. Then how can the processor distinguish the instruction length (1, 2, and 3 bytes) to execute the next instruction?
2. Why, except for the Stop command, all commands take 3 bytes?
3. How to select a team? How to find the command fetch logic?
ps My Russian is bad. The text may contain spelling errors. I apologize in advance))

Answer 1

The easiest to understand:
for example, the two most significant bits of the command code determine its length
, respectively, the command
from 00 00000 to 00 111111 - will be single-byte
from 01 00000 to 01 111111 - will be double-byte
from 10 00000 to 10 111111 - will be three-byte
from 11 00000 to 11 111111 - will be four bytes
But this is not for 8080. Just to understand the general meaning.
And so - on the structure of the processor, the element "command decoder" apparently came across - in fact, it is his task to understand from the command code what the subsequent bytes mean.
2. The 8080 has more than one single-byte commands. Offhand: NOP, HALT, INT3. RET*, shifts, exchange between registers
3. if we talk about 8080, it’s more convenient and understandable not even logically, but in circuitry ... and so - one of the sources

Answer 2

The 8080 processor was 8-bit.
That is, it processed one command from one byte in one cycle, then it was already possible to understand how to process the next two bytes.
For such things there is an internal decoder.
Previously, there was just a special register (OPCODE), in which the first byte of the instruction was placed and the processor, decoding it, processed the following bytes according to this instruction

Answer 3

how can the processor distinguish the instruction length (1, 2, and 3 bytes) to execute the next instruction?

By opcode -- the part of the command that encodes the action of the command. The processor encoded the correspondence of the operation code and instruction length. The opcode may be contained entirely in the first byte of the instruction (as in the 8080), or it may be spread over multiple instruction bytes. In the latter case, the part of the opcode found in the first byte must be different from the opcode of any one-byte instruction in order to distinguish them.

Why, except for the Stop command, all commands take 3 bytes?

If we are talking about 8080, then this statement is not true. The 8080 has a few single byte commands (nop, bit shifts and rotations, push/pop, etc.), a few double byte commands (mvi and immediate value arithmetic -- adi, aci, etc.) and a few three byte commands ( directly addressed commands -- lda, sta, etc., jumps, subroutine calls, etc.). Because the values ​​that the instructions operate on are in the bytes following the first - in one if a 1-byte operand is needed and in two if a 2-byte operand or address is needed.

How to select a team? How to find the command fetch logic?

You can look at the source of the reversed 8080 .

Answer 4

The first byte indicates which command it is, or the type of command. Next comes the logic that either processes this particular one-byte command, or reads the next bytes to specify the command further. That. as soon as the processor (with the help of the command decoder) understands what the command is, it executes it - it gives current to the necessary elements in the right direction.

Answer 5

One of the options written above is to encode the length in separate bits. You can also use a prefix code