Index
> Windows > "call" doing wrongGoto page Previous 1, 2 |
| Author |
|
|
revolution 04 Aug 2020, 13:53
Overclick wrote: about call: |
|||
04 Aug 2020, 13:53 |
|
|
Overclick 04 Aug 2020, 14:03
Just to test it
|
|||
|
04 Aug 2020, 14:03 |
|
|
Overclick 04 Aug 2020, 15:18
revolution, why JMP works fine in same condition? What have I done to corrupt call addressing for "call" only?
|
|||
|
04 Aug 2020, 15:18 |
|
|
revolution 04 Aug 2020, 15:56
It isn't call that is corrupted, it is your stack that is corrupted, because ret is doing more than you expect.
Try bitRAKEs suggestion to use retn for any internal function within your proc. Or try my suggestion to move the function out of the proc. |
|||
|
04 Aug 2020, 15:56 |
|
|
DimonSoft 04 Aug 2020, 16:02
bitRAKE wrote: but when assemblers starting making decisions what a return instruction is based on the context They don’t. It’s a macro thing, the source is available, the behaviour is documented and the feature is done in a way that makes it feel quite natural. One should always know what is going on under the hood of any library or syntax sugar feature to avoid cargo cults. Overclick wrote: Why it is not documented anyhow? FASM.pdf, end of the section 3.1.3. Computers never do what you want, they always do what you asked for. |
|||
|
04 Aug 2020, 16:02 |
|
|
Overclick 04 Aug 2020, 16:22
Once again
 :
I don't have any issues with ret or stack itself. Error happens before ret. For example: Code: must_be_executed: <any visible operations> invoke ExitProcess,rax ; or MessageBox or anything BEFORE ret and stack request. ret ..... DlgProc ... call must_be_executed ; jumps to hell jmp must_be_executed ; works just fine ... ret endp |
|||
|
04 Aug 2020, 16:22 |
|
|
revolution 04 Aug 2020, 16:37
Show us a minimal example that exhibits your problem, something complete we can compile without having to guess any other parts.
|
|||
|
04 Aug 2020, 16:37 |
|
|
Overclick 04 Aug 2020, 20:47
Here we go:
Code: format PE64 GUI 5.0 entry start include 'win64a.inc' section '.data' data readable writeable message db 'WHERE IS MY MESSAGE?',0 section '.text' code readable executable start: invoke GetModuleHandle,0 invoke DialogBoxParam,rax,37,HWND_DESKTOP,DialogProc,0 invoke ExitProcess,0 LOOK_AT_ME: invoke MessageBox,HWND_DESKTOP,message,0,MB_OK ret proc DialogProc uses rbx rsi rdi,hWnd,wMsg,wParam,lParam mov [hWnd],rcx mov [wMsg],rdx mov [wParam],r8 mov [lParam],r9 cmp [wMsg],WM_COMMAND je .wmcommand cmp [wMsg],WM_CLOSE je .wmclose xor rax,rax jmp .finish .wmcommand: cmp [wParam],BN_CLICKED shl 16 + IDOK jne .processed call LOOK_AT_ME ret .wmclose: invoke EndDialog,[hWnd],0 .processed: mov rax,1 .finish: ret endp section '.idata' import data readable writeable library kernel,'KERNEL32.DLL',\ user,'USER32.DLL' import kernel,\ GetModuleHandle,'GetModuleHandleA',\ ExitProcess,'ExitProcess' import user,\ DialogBoxParam,'DialogBoxParamA',\ CheckRadioButton,'CheckRadioButton',\ GetDlgItemText,'GetDlgItemTextA',\ IsDlgButtonChecked,'IsDlgButtonChecked',\ MessageBox,'MessageBoxA',\ EndDialog,'EndDialog' section '.rsrc' resource data readable directory RT_DIALOG,dialogs resource dialogs,\ 37,LANG_ENGLISH+SUBLANG_DEFAULT,demonstration dialog demonstration,'Create message box',70,70,190,175,WS_CAPTION+WS_POPUP+WS_SYSMENU+DS_MODALFRAME dialogitem 'BUTTON','OK',IDOK,85,150,45,15,WS_VISIBLE+WS_TABSTOP+BS_DEFPUSHBUTTON enddialog It was ok at 32bit but 64 just crashing |
|||
|
04 Aug 2020, 20:47 |
|
|
revolution 04 Aug 2020, 23:14
This is a case of unaligned stack. Try this
Code: ;... start: push rbp ;<-- add me invoke GetModuleHandle,0 invoke DialogBoxParam,rax,37,HWND_DESKTOP,DialogProc,0 invoke ExitProcess,0 LOOK_AT_ME: push rbp ;<-- add me invoke MessageBox,HWND_DESKTOP,message,0,MB_OK pop rbp ;<-- add me ret ;... |
|||
|
04 Aug 2020, 23:14 |
|
| Goto page Previous 1, 2 < Last Thread | Next Thread > |
Forum Rules:
|