assembly - MASM32 x86 窗口程序集 GetCommandLineToArgvW

Question

请问如何在x86汇编中使用CommandLineToArgvW函数。我遇到了麻烦。截至今天，我所能打印的只是参数的数量以及 cmd 中的程序执行。我想将参数保存在不同的变量中。我该怎么做？

我的代码是这样的:

include \masm32\include\masm32rt.inc

.data
  Format db "%d", 10, 0

.data?
  Arguments db 100 dup(?)

.code

start:

  mov esi, offset Arguments

  push ebp
  mov ebp, esp
  sub esp, 4

  call GetCommandLineW

  lea ecx, dword ptr[ebp - 4]
  push ecx
  push eax
  call CommandLineToArgvW

  mov esi, eax
  push offset Arguments
  call StdOut

  push dword ptr [ebp - 4]
  push offset Format
  call crt_printf
  add esp, 8


  push 0
  call ExitProcess

end start

我现在的输出是参数的数量。例如:

• D:\masm32>Hello.exe 我是你好
• 4
• D:\masm32>

Answer 1

CommandLineToArgvW至少要注意三个怪癖:

1. 结果是指向宽字符串的指针数组。

   MASM32 函数 crt_printf 使用 Microsoft VC 运行时库 (msvcrt.dll) 中的函数 printf。因此，您可以使用大写的“S”作为类型字段字符。看看printf Type Field Characters on MSDN .

2. 结果是指向字符串的指针数组的第一个元素的地址。

   大多数打印函数需要一个指向字符串的指针，而不是指向字符串指针的指针。您必须取消引用该地址才能获得指向该字符串的指针。命令行“Hello.exe I am Hello”将被拆分为四个字符串:“Hello.exe”、“I”、“am”、“Hello”。指向这些字符串的指针是在一个有4个指针的数组中找到的:[指向“Hello.exe”的指针]、[指向“I”的指针]，依此类推。假设函数 CommandLineToArgvW 有一个返回值 EAX=0x001445A8。 Hexdump 看起来像

   Address   Hex dump                                         ASCII
   001445A8  B8 45 14 00|CC 45 14 00|D0 45 14 00|D6 45 14 00| ¸E.ÌE.ÐE.ÖE.
   001445B8  48 00 65 00|6C 00 6C 00|6F 00 2E 00|65 00 78 00| H.e.l.l.o...e.x.
   001445C8  65 00 00 00|49 00 00 00|61 00 6D 00|00 00 48 00| e...I...a.m...H.
   001445D8  65 00 6C 00|6C 00 6F 00|00 00 00 00|00 00 00 00| e.l.l.o.........
   

   地址 0x001445A8 是指向 0x001445B8 的指针(以小端格式显示在转储中)，这是宽字符格式的“Hello.exe”的开头。下一个指针在 0x001445A8 后面 4 个字节:0x001445CC - 指向“I”。下一个指针距离 4 个字节，依此类推。您只需添加 4 即可快速遍历该数组。您可以通过将索引乘以 4 轻松获得列表中间字符串的地址 - 指向第三个字符串的指针(“am”，索引:2)位于 0x001445A8 + 2 * 4 = 0x001445B0 => 0x001445D0 => “上午”。

3. 该函数分配内存，必须使用 LocalFree 手动释放。

我尽可能少地改动了你的程序:

include \masm32\include\masm32rt.inc

.data
    Format db "argc: %d", 10, 0
    fmt db "%S",10,0                ; %S: printf wide-character string / wprintf single-character string
    szArglist dd ?

.code

start:

    push ebp
    mov ebp, esp
    sub esp, 4

    ; https://msdn.microsoft.com/library/windows/desktop/ms683156.aspx
    call GetCommandLineW        ; EAX = pointer to the command line

    ; https://msdn.microsoft.com/library/windows/desktop/bb776391.aspx
    lea ecx, dword ptr[ebp - 4] ; Get the current address of [ebp-4]
    push ecx                    ; int *pNumArgs (Pointer to a SDWORD, here at ebp-4)
    push eax                    ; LPCWSTR lpCmdLine (from GetCommandLineW)
    call CommandLineToArgvW

    mov [szArglist], eax        ; Store the result of CommandLineToArgvW (at least for LocalFree)

    mov esi, eax                ; ESI = address of a pointer (the first element in szArglist)
    mov ebx, [ebp-4]            ; Countdown the number of arguments

    @@:                         ; Loop
    push dword ptr [esi]        ; Pointer to a string (dereferenced esi)
    push OFFSET fmt             ; Format string
    call crt_printf             ; printf (""%S\n", esi)
    add esp, 8                  ; Clear the stack after printf
    add esi, 4                  ; Next address of a pointer (next element of szArglist)
    dec ebx                     ; Countdown the number of arguments
    jne @B                      ; Loop to the last @@

    push dword ptr [szArglist]
    call LocalFree              ; Free the memory occupied by CommandLineToArgvW

    push dword ptr [ebp - 4]    ; Value that is stored in [ebp-4]
    push offset Format          ; Pointer to format string
    call crt_printf             ; printf ("argc: %d\n", [ebp-4])
    add esp, 8                  ; Clear the stack after printf

    push 0
    call ExitProcess

end start

MASM32 函数 StdOut 无法处理宽字符字符串。您必须先将它们转换为 ANSI 字符串。用于此目的的 Windows 函数是 WideCharToMultiByte :

include \masm32\include\masm32rt.inc

.data
    szArglist dd ?
    buf db 1024 DUP (?)
    crlf db 13, 10, 0           ; New line

.code

start:

    push ebp
    mov ebp, esp
    sub esp, 4

    ; https://msdn.microsoft.com/library/windows/desktop/ms683156.aspx
    call GetCommandLineW        ; EAX = pointer to the command line

    ; https://msdn.microsoft.com/library/windows/desktop/bb776391.aspx
    lea ecx, dword ptr[ebp - 4] ; Get the current address of [ebp-4]
    push ecx                    ; int *pNumArgs (Pointer to a SDWORD, here at ebp-4)
    push eax                    ; LPCWSTR lpCmdLine (from GetCommandLineW)
    call CommandLineToArgvW

    mov [szArglist], eax        ; Store the result of CommandLineToArgvW (at least for LocalFree)

    mov esi, eax                ; ESI = address of a pointer (the first element in szArglist)
    mov ebx, [ebp-4]            ; Countdown the number of arguments

    @@:                         ; Loop

    ; https://msdn.microsoft.com/library/windows/desktop/dd374130.aspx
    push NULL                   ; LPBOOL  lpUsedDefaultChar
    push NULL                   ; LPCSTR  lpDefaultChar
    push SIZEOF buf             ; int     cbMultiByte
    push OFFSET buf             ; LPSTR   lpMultiByteStr
    push -1                     ; int     cchWideChar
    push [esi]                  ; LPCWSTR lpWideCharStr (dereferenced esi)
    push 0                      ; DWORD   dwFlags
    push 0                      ; UINT    CodePage
    call WideCharToMultiByte

    push OFFSET buf             ; Pointer to an ANSI string
    call StdOut
    push OFFSET crlf            ; New line
    call StdOut

    add esi, 4                  ; Next address of a pointer (next element of szArglist)
    dec ebx                     ; Countdown the number of arguments
    jne @B                      ; Loop to the last @@

    push dword ptr [szArglist]
    call LocalFree              ; Free the memory occupied by CommandLineToArgvW

    push OFFSET buf
    push dword ptr [ebp - 4]
    call dwtoa
    push OFFSET buf             ; Pointer to a string
    call StdOut                 ; printf (""%S\n", esi)
    push OFFSET crlf
    call StdOut

    push 0
    call ExitProcess

end start