Questions tagged [intel-syntax]

Intel Syntax is used in Intel's (and AMD's) manuals, and by many assemblers. See the x86 tag wiki for links to those and to assembler manuals.

The other major syntax for x86 assembly is AT&T (att). Other syntaxes include HLA, and the Go assembler's x86 syntax (which looks like 16-bit, using AX to actually mean AL/AX/EAX/RAX depending on the operand-size).

All flavours of Intel Syntax share these characteristics:

• Parameter order: destination <- source. pinsrd xmm0, eax, 2
• Square brackets indicate a memory operand: add eax, [esi] (but beware the differences between NASM and MASM for symbols, offset vs. just omitting the []).
• Operand size: implied by a register name, or specified explicitly with dword (NASM) or dword ptr (MASM) in instructions with no register operand.
  add qword [fs:rdi + r10], 123. (Put the size operator on the memory operand.)
• Immediate values and other numeric constants: No $ prefix. For hex, use a trailing h, and make sure the number starts with a 0. e.g. 0deadbeefH. Some Intel-syntax assemblers (e.g. NASM) also support C-style 0xdeadbeef constants, but all support trailing-h. Binary constants can use a b suffix.

The two major flavours of Intel syntax are NASM-style and MASM/TASM style.
How to know if an assembly code has particular syntax (emu8086, NASM, TASM, ...)?
and x86, difference between BYTE and BYTE PTR

The GNU assembler, as/gas (and compatible ones like clang's built-in assembler) supports a .intel_syntax noprefix directive to switch to a mode with MASM-like syntax. There isn't official documentation for GAS's intel-syntax. See also Distinguishing memory from constant in GNU as .intel_syntax. If you're not sure, encode the machine-code you want somehow (e.g. using another assembler) and disassemble in Intel syntax with objdump -drwC -Mintel.

• MASM: when [] is ignored, and how to get addresses as immediate constants. Confusing brackets in MASM32
• Assembly difference between [var], and var for NASM vs. MASM.