AT&T mnemonic conventions

(1) AT&T mnemonics and register names are invariably(总是，一定地，不变地) lowercase, Intel(nasm) suggest uppercase but will accept lowercase, AT&T requires lowercase;

AT&T助记符和寄存器名字一定是小写字母，而Intel（nasm）建议是大写字母，但是也接受小写字母，但是AT&T要求必须是小写字母；

(2) Register names are always preceded by the percent symbol %;

Intel: AX or EBX

AT&T: %ax or %ebx

寄存器名字之前总是有百分号%；

(3) Every AT&T machine instruction mnemonic that has operands has a single-character suffix(后缀) indicating how large its operands are;

b ---> byte, 8bits

w--->word, 16bits

l--->long, 32bits

每一个AT&T机器指令助记符，如果有操作数的话，都有一个单字母的后缀，表明它操作数的位数；

Intel: MOV BX, AX

AT&T: movw %ax, %bx

(4) In the AT&T syntax, source and destination operands are placed in the opposite order from Intel syntax:

Intel: MOV BX, AX

AT&T: movw %ax, %bx

在AT&T语法中，源操作数和目的操作数的顺序与Intel语法中是刚好相反的；

(5) In the AT&T syntax, immediate operands are always preceded by the dollar sign($):

Intel: PUSH 32

AT&T: pushl $32

在AT&T语法中，立即数的前面总是有美元符号$；

(6) Not all AT&T instruction mnemonics are generated by gcc:

Equivalents to Intel's JCXZ, JECXZ, LOOP, LOOPZ, LOOPE, LOOPNZ, and LOOPNE have only recently been added to the AT&T mnemonic set, and gcc never generates code that uses them

不是所有的AT&T指令助记符都由gcc产生，有些指令gcc不产生；

(7) In the AT&T syntax, displacements in memory reference are signed quantities placed outstide parentheses contaning the base, index, and scale values;

相关日志：AT&T Memory Reference Syntax