Assembly

Assembly Language

Down to the very core of computers, every signal sent to any part of the machine is a series of electrical pulses. If the electrical pulse is high, for example 5 volts, that means a binary 1 digit is sent. If the electrical pulse is low, for example 0 volts, that means a binary 0 digit is sent. Put 4 binary digits together and you get a nibble. Put 8 together, you get a byte. Put 16 together, you get a word. Computers understand commands in such segments, but humans would rather not. So, they developed a better way of viewing such information and that was with hexadecimal. First of all, it didn't take 16 characters to show one command and in general, it became more convient to remember commands like:

90 - No Operation
CC - Break
E9 - 16 bit jump
CD - Interrupt
etc.

But soon, it was understood that such codes needed to be user friendly, so they invented assembly code. With assembly code, you can do any operation with special codes that made sense. For example, JMP jumped, INT interrupted, NOP did no operation, etc.

An assembler is the program that reads an assembly file (usually .asm) and converts it to a binary executable. This is very much akin to a compiler generating a executable program. However, assemblers are considered separately from compilers because compilers hide the low-level details of your system, while assemblers expose you to these details.

Assembly Syntax

In assembly, every instruction (or operation/operator) is followed by a comma-separated list of parameters (or operands). These operands can be either immediate values (i.e. numbers such as 1, 2, 0x16, 0101b), or registers, or a memory location. For example,

mov eax, 123

The instruction is mov, the operands are eax and 123. Here eax is a register and 123 is an immediate value.

There are two main different syntaxes for assembly: the Intel syntax, and the AT&T syntax.

Intel Syntax

Generally speaking, the first operand of an operation is the destination operand, and the other operand (or operands) is the source operand. For example:

mov eax, 123

The mov instruction takes a value from the source operand, and places it in the destination operand, so the value 123 would be placed into the register eax. The Intel syntax is used by Intel in all documentation, descriptions and examples.

Could probably use more information. -- FMota

AT&T Syntax

AT&T syntax is the reverse of Intel syntax. The data operated on moves from left to right. Here is the same statement as above in AT&T syntax.

mov $0x123, %eax

There are also some other differences with operands:

• Literal values such as 0x123 are prefixed with '$' (see example above).
• Memory locations have no prefix: mov 0x123, %eax. Moves the value stored at 0x123 to eax.
• Registers are prefixed with '%'
• When using a register as a pointer it is put in parenthesis: mov (%eax), %ebx. Moves the value stored at the pointer stored in eax to ebx.
• When using a statement with memory locations (with no register present), a suffix is needed after the instruction. For example: movl $0x123, 0x123. The 'l' (stands for long) is needed to tell the assembler that the values are 32 bits.

Assemblers

The two most widely used assemblers nowadays are GAS (AT&T or Intel syntax) and NASM (Intel syntax). There is also fasm (Intel syntax). See Tool Comparison for comparisons between common OSDev tools such as these.

See Also

Articles

• Inline Assembly

External Links

• GNU Binutils - An open source collection of software including the GNU assembler (gas)
• gas manual - Official manual for the GNU assembler (gas)
• NASM - sourceforge.net project site for NASM
• x86_Assembly - A wikibook on writing assembly for x86 based PCs