Question

I have 2 assembly methods:

1. DWToHex PROTO:DWORD modifies binary integer number of type DWORD passed as in input parameter
   inside EAX into a hexadecimal representation, pointer to which is passed inside the single parameter. It is function used for debug purpose and it works fine.
2. ProcA PROTO:DWORD – does nothing except for corrupting the content of EAX below is the code:
   (sorry for the formatting -c seems like the regular code formatting did not work here)

; #########################################################################

`.386                       ; set processor type`
`.model flat, stdcall       ; memory model & calling convention`
`option casemap :none       ; case sensitive`

`.code`

; #########################################################################

procA public addrSTD:DWORD

ret

ProcA   endp

; #########################################################################

end

Here is the debug trace:

.data

MsgBoxCaption4 db  "before calling", 0
MsgBoxCaption5 db  "after   calling", 0

.data?

N       DWORD   ?
NTXT        db 16 dup(?)

.code

mov eax, N
invoke  cx_DWToHex, addr NTXT
invoke  MessageBox, NULL, addr NTXT, addr MsgBoxCaption4, MB_OK ; show eax as hexadecimal

mov eax, N                      ;   ***** point A ****
invoke  ProcA, addr sdt         ;   ***** point B ****
;   mov eax, N
invoke  cx_DWToHex, addr NTXT
invoke  MessageBox, NULL, addr NTXT, addr MsgBoxCaption5, MB_OK ; show eax as hexadecimal

……………………………………………………………………………….

Question:
Why the content of EAX in point A in point B is not the same?

Answer 1

The invoke directive you use for calling ProcA is supposed to get the address of ‘sdt’, and push it on the stack. More than likely, it generates code similar to

lea eax, sdt
push eax
call ProcA

which destroys the contents of eax. You can verify this by looking at the disassembly of the generated code.