The sourcecode for rainbowcrack piqued my interest this weekend. I was looking through it and noticed the the whole thing was C++ with the exception of one part, see below


__asm__ __volatile__ ( "mov %2, %%eax;""xor %%edx, %%edx;"
"divl %3;"
"mov %%eax, %0;"
"mov %%edx, %1;"
: "=m"(nIndexOfX32), "=m"(nTemp)
: "m"(nIndexOfX32), "m"(nPlainCharsetLen)
: "%eax", "%edx"
);
</code></pre>
Well, I think I read through one or two pages of the Art of Assembly 3 or 4 years ago. Basically I know nothing about assembly.

I wanted to know what was so important about this code that it just had to be different from every other part of the app; so I started googling.

It was pretty easy to find out what the mov does. This website</a> was a god-send. They mentioned that a mov command in Intel ASM takes the value of the second argument and puts it in the first argument.

"Ordering of operands</strong> : Unlike to Intel convention (first operand is destination), the order of operands is source(s) first, and destination last. For example, Intel syntax "mov eax, edx</code>" will look like "mov %edx, %eax</code>" in AT&T assembly."

so


mov eax, nIndexOfX32 = Intel
mov %2, %%eax; = AT&T
</code></pre>
They both do the same thing, the arguments are just flipped.

Then I started wondering what the heck the lines starting with colons meant. Once again, that same article</a> had the answer. Apparently in extended assembly, you can specify the "input registers, output registers and a list of clobbered registers." Well, that's exactly what's going on.</span>

The first row is specifying the output data. The second row the input data, and the third row, a list of registers that are being "clobbered". I found this meant "the clobbered register after the third colon informs the GCC that the value of the register is to be modified inside "asm", so GCC won’t use this register to store any other value." Aha!</span>

Well, I was moving along, but was once again stuck. What the heck do the %0, %1, etc values mean. Well, apparently it's the order that the outputs and inputs are listed (starting from 0).

So in my example, the pointer to the memory location nIndexOfX32 is being put in the eax register (ok I'm not entirely sure this is what's going on, but from what I've read, that's what I've come up with.)

The next thing that confused me was that xor command. I know the theory behind an xor, but I couldn't figure out why they were xor'ing the same values. I also couldn't figure out the divl thing. I knew from reading</a> that the div had to do with division, and the 'l' part of that meant it was dividing a long. With regard to the div instruction, I couldn't figure out what was being divided by the %3 value. Ya know, cause normally in division you have a value1/value2. So where was the magical first value.

That's when I found this</a> page though, and the whole xor divl thing became instantly clear

"</a>The instructions for multiplication and division are mul and div. Both only operate on the accumulator register (eax) and use the data register (edx)</strong> as an overflow. The part of the registers affected are determined by the size of the operand. "

"</a>Therefore, to get expected results, it is recommended that you set edx to zero before calling mul or div</strong>."

Aha! Well, if you xor a value by its self, you get 0. So that xor command made the value in the edx register 0. Then, the divl command then divided the value in eax (remember that first mov command) by whatever is stored in %3 which, I'm guessing, is nPlainCharsetLen.

I think I have it all figured out then. The remaining mov's work under the same idea as the first one. The value of eax is the result of the division, and if it didn't divide cleanly, the overflow of the division is stored in edx.

For example, 12/7 is 1.71. The 1 would be stored in eax, and the 72 would be in edx.

So what is this block of assembler actually doing then? It looks like it's just dividing a number and sticking the quotient and remainder in nIndexOfX32 and nTemp respectively. They have a normal word for this, you may have heard of it before, it's called modulus division. Does this look familiar?

value = 12 % 7</pre>
Anyone who knows assembler care to help me out on this one? I hope I didn't totally misunderstand the code : )

Now that I really think about it though, that might understand why these comment lines were found above this assembly code


//m_Plain[i] = m_PlainCharset[nIndexOfX32 % m_nPlainCharsetLen];

//nIndexOfX32 /= m_nPlainCharsetLen;
</code></pre>
Modulus division... :-p