汇编函数调用的参数传递

Posted on 2011-10-07 17:58 光铭阅读(4622) 评论(0) 编辑收藏举报

32bit 汇编函数调用参数传递是通过堆栈来实现的。

 1 .section .data
 2 output:
 3     .asciz "The processor Vendor ID is %s\n"
 4 .section .bss
 5     .lcomm buffer,12
 6 .section .text
 7 .globl main
 8 main:
 9     movl $0,%eax
10     cpuid
11     movl $buffer, %edi
12     movl %ebx, (%edi)
13     movl %edx, 4(%edi)
14     movl %ecx, 8(%edi)
15     pushl $buffer    #第二个参数先进栈，第一个后进栈，这样pop的时候第一个就会先出来,反正就是参数从右到左进栈
16     pushl $output
17     call printf
18     addl $8, %esp    #前面push了现在要堆栈平衡
19     pushl $0
20     call exit

64bit的参数传递会首先用到寄存器，寄存器不够，才用对堆栈

LINUX下面INTERGER参数（包括bool,char,short,int,long,long long)从左到右依次放到寄存器%rdi,%rsi,%rdx,%rcx,%r8,%r9，超过6个放堆栈里；SSE型的参数（float,double)，从左到右依次放到vector registers(%xmm0-%xmm7)

注意%rax的作用

1)temporary register;
2)with variable arguments passes information about the number of vector registers used;
3)1st return register
特别注意第二点
For calls that may call functions that use varargs or stdargs (prototype-lesscalls or calls to functions containing ellipsis (. . . ) in the declaration) %al is usedas hidden argument to specify the number of vector registers used.
对于变长参数的函数，要用%al指明用到的vector registers（存放float,double等类型）的个数，比如printf

 1 .section .data
 2 output:
 3     .asciz "The processor Vendor ID is %s\n"
 4 .section .bss
 5     .lcomm buffer, 12
 6 .section .text
 7 .globl main
 8 main:
 9     movl $0,%eax
10     cpuid
11     movl $buffer,%edi
12     movl %ebx,(%edi)
13     movl %edx,4(%edi)
14     movl %ecx,8(%edi)
15     movq $output,%rdi
16     movq $buffer,%rsi
17     movq $0,%rax    #没有用到vector registers，所以%al是0
18     call printf
19     mov $0,%ax   #这里的ax是返回值
20     call exit

Windows下面INTERGER参数从左到右依次放到rcx,rdx,r8,r9，超过4个放堆栈里，由于现在不搞windows，这个就不说了

刷新页面返回顶部

HelloWorld

公告

汇编函数调用的参数传递