| Programs Under the Hood...Our Friend the CPU Posted by: dargueta in Untagged on Jun 25, 2008 |
(Part 3)
Hi, everyone, it's me again with the latest installment of Programs Under the Hood. Today we're going to get to know the inner workings of the processor, and start learning a little assembly language.
INTRODUCING THE GENERIC INTEL CPU
AL 
Starting with the 80386, each of the 16-bit registers was extended to 32 bits. Now you have EAX, EBX, ECX, and EDX. The low word of EAX is…EAL? No! It’s still AX. The low byte of EAX is the same as the low byte of AX, which is AL. (If you’re confused, look at the diagram.) What about the high word? Sadly, you can’t access it directly. You either need to write to the entire register, or use a mask and shift bits around. As a side note, the general registers are the only registers that can be subdivided into smaller registers.
*There is also a hidden register that points to the current instruction in the current code segment called IP, which stands for instruction pointer. You can’t access it directly. Together, they make a full pointer, CS:IP. Read on in the next issue to see what I mean by “pointer”.
I know what you’re thinking…in a COM program, these registers all have the same value, because the program takes up one segment. In an EXE program, they usually point to different segments.
TESTING THE WATERS
mov ah,09
What do you think that does? It means move the byte 09h into register AH. MOV is the mnemonic, or name of the instruction, AH is the destination operand, and the number 09h is the source operand. MOV actually copies data from one place to another because it leaves the source untouched. Different instructions have different numbers of operands; they can have anywhere from none to two. Very few have three, and they’re all floating-point instructions that you probably won’t use for a great while. With that knowledge in mind, try and guess what this does:
add ax,bx
If you guessed that it adds BX to AX, you’re partially right. Where does it put the result? In the destination operand, in this case AX. In C-style pseudocode, this is equivalent to AX += BX. What if I wanted to add the contents of, say, ECX and the 32-bit integer at memory location 48C3h? Easy. You can use either of the following, depending on where you want to store the result:
add ecx,[48c3h] ;store result of addition in ecx
Try figuring out this one:
mov ax,920fh
those registers, but only two at once.
This does nothing useful, really, but you kind of get the idea of how this is going to work. Some more instructions we’re going to use are SUB, AND, OR, XOR. These all take two arguments, just like MOV and ADD.
Well, that’s it for today. I didn’t get to do everything that I wanted to do, but I’ll keep plugging along. Next time we’ll get to know debug.exe a little better, and I promise you then we’ll write our first somewhat useful program.
Hi, everyone, it's me again with the latest installment of Programs Under the Hood. Today we're going to get to know the inner workings of the processor, and start learning a little assembly language.
INTRODUCING THE GENERIC INTEL CPU
Starting with the 80386, each of the 16-bit registers was extended to 32 bits. Now you have EAX, EBX, ECX, and EDX. The low word of EAX is…EAL? No! It’s still AX. The low byte of EAX is the same as the low byte of AX, which is AL. (If you’re confused, look at the diagram.) What about the high word? Sadly, you can’t access it directly. You either need to write to the entire register, or use a mask and shift bits around. As a side note, the general registers are the only registers that can be subdivided into smaller registers.
*There is also a hidden register that points to the current instruction in the current code segment called IP, which stands for instruction pointer. You can’t access it directly. Together, they make a full pointer, CS:IP. Read on in the next issue to see what I mean by “pointer”.
I know what you’re thinking…in a COM program, these registers all have the same value, because the program takes up one segment. In an EXE program, they usually point to different segments.
Wait a minute…doesn’t SS point to the stack? What do we need SP or BP for? Aha…now we’re getting into segmentation, or the division of memory into segments. That'll be left for the next blog.
TESTING THE WATERS
mov ah,09
What do you think that does? It means move the byte 09h into register AH. MOV is the mnemonic, or name of the instruction, AH is the destination operand, and the number 09h is the source operand. MOV actually copies data from one place to another because it leaves the source untouched. Different instructions have different numbers of operands; they can have anywhere from none to two. Very few have three, and they’re all floating-point instructions that you probably won’t use for a great while. With that knowledge in mind, try and guess what this does:
If you guessed that it adds BX to AX, you’re partially right. Where does it put the result? In the destination operand, in this case AX. In C-style pseudocode, this is equivalent to AX += BX. What if I wanted to add the contents of, say, ECX and the 32-bit integer at memory location 48C3h? Easy. You can use either of the following, depending on where you want to store the result:
add ecx,[48c3h] ;store result of addition in ecx
Try figuring out this one:
mov ax,920fh
those registers, but only two at once.
This does nothing useful, really, but you kind of get the idea of how this is going to work. Some more instructions we’re going to use are SUB, AND, OR, XOR. These all take two arguments, just like MOV and ADD.
Well, that’s it for today. I didn’t get to do everything that I wanted to do, but I’ll keep plugging along. Next time we’ll get to know debug.exe a little better, and I promise you then we’ll write our first somewhat useful program.
Set as favorite
Bookmark
Email This
Hits: 257
Trackback(0)
TrackBack URI for this entry
3 Comments 
How far are you going to take us with these blogs?Write comment
