Calling printf from COFF

Index > High Level Languages > Calling printf from COFF

Author

Thread

Constantine

Joined: 29 Oct 2016
Posts: 7

Constantine

I know how to call printf from usual PE executable. But in current project the entrypoint is in C++ file, it calls function from COFF written in FASM. And I need to call printf from FASM code. I tried following:

Code:

format MS COFF

include 'win32axp.inc'

extrn '__imp__printf' as printf:dword

public cutOutTags as '_cutOutTags'

; void cutOutTags(const char * html, char ** tags, int tagsCount, char * res);
        
proc cutOutTags c uses ECX EDX ESI EDI EBP, html:DWORD, tagsToDelete:DWORD, tagsCount:DWORD, res:DWORD
        
        cinvoke printf, "test"
        ret
endp

Compilation command line is

Code:

fasm.exe cutTags.asm
g++ -std=c++11 -static -Wall -o sem.exe  sem.cpp cutTags.obj

Compilation & linking is OK, but app crashes. What am I doing wrong?

25 Dec 2016, 06:29

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 17849
Location: In your JS exploiting you and your system

revolution

You probably need ccall instead of cinvoke. Invoke is for the Windows API, and ccall is for the C library functions.

25 Dec 2016, 06:50

Constantine

Joined: 29 Oct 2016
Posts: 7

Constantine

Thanks, but it is still the same. Also documentation says that
cinvoke printf, ...
and
ccall [printf], ...
are equivalents.

25 Dec 2016, 07:25

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 17849
Location: In your JS exploiting you and your system

revolution

Drop the brackets.

ccall printf,...

25 Dec 2016, 07:35

Constantine

Joined: 29 Oct 2016
Posts: 7

Constantine

I tried. It still crashes.

25 Dec 2016, 07:39

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 17849
Location: In your JS exploiting you and your system

revolution

What does the disassembler/debugger show?

Are you sure it crashes on the code you show above?

25 Dec 2016, 07:42

Constantine

Joined: 29 Oct 2016
Posts: 7

Constantine

I do not use disassembler/debugger.

Yes, if I comment the printf line program runs correctly.

25 Dec 2016, 11:07

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 17849
Location: In your JS exploiting you and your system

revolution

Constantine wrote:

I do not use disassembler/debugger.

It could help you a lot. Especially in cases like this where you don't know what is happening.

Ollydbg

25 Dec 2016, 11:14

Constantine

Joined: 29 Oct 2016
Posts: 7

Constantine

And how do I find my asm code in the debug?

25 Dec 2016, 15:51

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 17849
Location: In your JS exploiting you and your system

revolution

When it crashes you will see where it stops.

Or, you can insert int3 just before the call and the debugger will stop there for you to single step onwards.

25 Dec 2016, 15:54

system error

Joined: 01 Sep 2013
Posts: 671

system error

I don't know about other codes you're linking to but from the source above;

1. Your test string "test" is not null-terminated.
2. Try to get rid of decorated names first. Just use extrn printf, public cutOutTags.

25 Dec 2016, 17:18

system error

Joined: 01 Sep 2013
Posts: 671

system error

If you are to test on cutOutTags alone as a single source, try something like this below and see what more clues you can get from it. I use GoLink to produce the .exe.

Code:

format MS COFF
include 'win32axp.inc'

extrn printf
extrn exit

public cutOutTags

; void cutOutTags(const char * html, char ** tags, int tagsCount, char * res); 
         
proc cutOutTags c uses ECX EDX ESI EDI EBP, html:DWORD, tagsToDelete:DWORD, tagsCount:DWORD, res:DWORD 
         
        push 0          ;C-string null
        push "test"
        push esp
        call printf
        add esp,12
        push 0
        call exit
        ;ret
endp

25 Dec 2016, 17:24

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 17849
Location: In your JS exploiting you and your system

revolution

system error wrote:

1. Your test string "test" is not null-terminated.

It is properly terminated by both the cinvoke and the ccall macros.

A debugger/disassembler will show you such things. No need to assume here. Wink

25 Dec 2016, 17:37

system error

Joined: 01 Sep 2013
Posts: 671

system error

revo. I don't know much about macros but no, I don't think so or at least doesn't appear that way to me.

25 Dec 2016, 17:47

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 17849
Location: In your JS exploiting you and your system

revolution

system error wrote:

... no, I don't think so or at least doesn't appear that way to me.

Ya, it is. No need to manually add a null if you use the macros as intended.

25 Dec 2016, 17:51

system error

Joined: 01 Sep 2013
Posts: 671

system error

you're right. The string works fine without explicit terminator.

25 Dec 2016, 17:53

Constantine

Joined: 29 Oct 2016
Posts: 7

Constantine

The problem was in outer code. cutOutTags() function was called from C++ code as

std::cout << cutOutTags(input.c_str(), argv + 1, argc - 1, out);

Simple call (without std::cout) solves the problem. Thank you guys for helping!

P. S. Still I could not get any debugger to work. I tried Ollydbg & x32dbg. Either of them just stops in some obscure place not giving any crash message or even any reason why it have stopped and didn't want to continue tracing. And also I couldn't find any of my fasm instructions in exe file. Is there some good manual on debugging?

25 Dec 2016, 21:41

< Last Thread | Next Thread >

Forum Rules:

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum