如何在树莓派上使用 ARM 程序集查找字符串的长度

Question

我目前正在使用具有ARM架构的RPI学习汇编。我已经创建了下面的代码。但是，当我输入字符串（例如"hi"）时，会出现一些错误。我没有收到正确的长度。示例："Hi"->长度为 8（显然不正确）。当输入大于 4 个字符时，我也会出现段错误。示例："Hello" ->长度为 8，然后是段错误。

我希望我做错了什么！ :)提前感谢您的帮助。

.data
/*Message1 */
.balign 4
message1: .asciz "Enter a string: "
/* Message 2 */
.balign 4
message2: .asciz "Your string has a length of %dn"
/* Message 3 */
.balign 4
message3: .asciz "Your string is '%s'.n"
/* format for scanf */
.balign 4
pattern: .asciz "%s"
/*Where we will store the length of the string */
.balign 4
strLen: .word 0
/*where scanf will store the string */
.balign 4
string: .asciz ""
/*Return value */
.balign 4
return: .word 0
/* Text Section */
.text
end:
        ldr lr, add_of_return /*load in the return address to lr */
        ldr lr, [lr] /* load the value at the address lr into lr */
        bx lr  /* branch exchange (basically end the program) */
add_of_return: .word return
calcStrLen:
        /*load in the first character to reg 2
          compare it to 0 and branch to label
          'done' if the two are the same */
        ldr r2, [r1]
        beq done
        /* increment the counter (r0) and the
           string pointer (r1) */
        add r0, r0, #1
        add r1, r1, #1
        /* store counter value into strLen */
           str r0, [r3]
        /* branch back to top of 'calcStrLen' */
           b calcStrLen
done:
        /* load the address of strLen to reg 1
           and then get the value at reg 1 and
           store it into reg 1 */
        ldr r1, add_of_strLen
        ldr r1, [r1]
        /* load in the address of message2 into
           reg 0 */
        ldr r0, add_of_message2
        /* print */
        bl printf
        /* branch to label 'end' */
        b end
.global main
main:
        /* load return address into reg 1
           and then get the value at the
           address and store it into reg 1 */
        ldr r1, add_of_return
        str lr, [r1]
        /* load address of message1 into reg 0
           and then print message1 */
        ldr r0, add_of_message1
        bl printf
        /* load in the necessary elements to
           scan in a string from the user */
        ldr r0, add_of_pattern
        ldr r1, add_of_string
        bl scanf
        /* prints the string given by the user */
        ldr r0, add_of_message3
        ldr r1, add_of_string
        bl printf
        /* load in the address of the user's string */
        ldr r1, add_of_string
        /*load in the address of strLen, then
          go to the address and store the value
          into r0 */
        ldr r0, add_of_strLen
        ldr r0, [r0]
        /* store strLen's address in reg 3 */
        ldr r3, add_of_strLen
        /* branch to label 'calcStrLen' */
        bl calcStrLen
/*Definitions of address values used */
add_of_message1: .word message1
add_of_message2: .word message2
add_of_message3: .word message3
add_of_pattern: .word pattern
add_of_string: .word string
add_of_strLen: .word strLen
/* external */
.global printf
.global scanf            

Answer 1

string: .asciz ""

我很确定只是为字符串""分配足够的内存（即一个字节）并填充它，这意味着您将获得内存：

string: .byte 0
next:   ; whatever comes next

如果你然后去那个位置存储两个或更多的字节（例如通过scanf()），你将处于一个痛苦的世界，因为会有某些描述的损坏（在next）。

它损坏的内容取决于您的部分在内存中的布局方式。

现在，至于如何修复它，这取决于您的汇编程序。大多数都有.defs类型的伪操作码来定义任意大小的空间，因此您可以执行以下操作：

string: .defs 128

取决于您的汇编程序支持的内容。