Thread: x86 Instruction SBB

Hello,
I'm trying to reverse some application from x86 asm to C but don't know how to translate this code:

mov ecx, the_loop_counter
mov esi, mem_Ptr
mov dl, byte_x
clc
@sbb_loop:

sbb [ecx+esi-1], dl

loop @sbb_loop
----------

to be specific I don't know how to check carry flag status, can anyone help me out please?
how to re-write this code into C?

big thanks in advance

sbb does the same as sub, but also subtracts 1 more if carry flag is set. clc clears the carry flag, but I think that somewhere within the loop it may be set in certain situations.

This is not enough information to actually tell what the loop does.

thanks, but I'm asking how to re-write this code into C, with sub instruction it would be easy something like:

Code:

for   (; the_loop_counter>0; the_loop_counter-- )  {
       
        *( BYTE *)((mem_Ptr + the_loop_counter) - 1) -= byte_x;

}

but I dont know how to write this "for" loop with sbb instruction, this code is part of a checksum algorithm.

To rewrite something to C you have to know the exact meaning of everything. There's no way to write the sbb instruction in C. If you know the exact meaning of the entire block of code, then there is probably an equivalent for C, for which this is an optimized compiled result. Not always though...

Wouldn't this work?

Code:

BYTE* b = ( BYTE *)((mem_Ptr + the_loop_counter) - 1);
*b -= byte_x - (*b < byte_x) ? 1 : 0;

If you are after converting a piece of optimized code, if that is the case, within the assembly code back to pre-compiled c code will takes lots of trial and error. For that will require much knowledge of how .asm code is created and what each instruction does. Also this would be very archetecturally dependant as well, but you have already eluded that piece.

Originally Posted by bithub

Wouldn't this work?

Code:

BYTE* b = ( BYTE *)((mem_Ptr + the_loop_counter) - 1);
*b -= byte_x - (*b < byte_x) ? 1 : 0;

thank you very much bithub seems this solve my problem, thanks & respect!