I'm very amazed by this. How can this even be possible? Is there a standard formula that breaks every copy protection out there?
I'm very amazed by this. How can this even be possible? Is there a standard formula that breaks every copy protection out there?
A most basic protection, for example, might make the user enter a serial number to register a piece of software. If you enter the correct number then all is OK, if you enter the wrong number a message-box might appear saying "Sorry registration number is invalid".
Displaying that message-box almost immediately makes the cracking job 80% simpler. Disassemble the program, look for the reference to the text-string. Then look for a "compare + jump" that goes to that area of the code. 99% of the time the comparison will be "Is the serial number OK?" and if you remove the jump, or invert it ("jmpz -> jumpnz"), you're done.
If you want to read more you can find mirrors of +fravias documentation and demonstration site online which were live back in the day and introduce this stuff, and lots more:
The countdown was displayed in red digits in the corner of the screen. I used SoftICE to set a memory write breakpoint on the video framebuffer at the location where the timer was displayed. That allowed me to locate the code printing out the timer. That in turn gave me the address of the timer variable, against which I set another memory breakpoint, which brought me to the routine that was counting down the timer. A 1 byte patch effectively disabled the timer and allowed me to play for hours.
Most of Fravias work was easily understandable, and pretty obvious, from that background. But he really was an interesting character and I was genuinely saddened to learn of his death.
I wish I could have met him.
Or better, I got rid of the annoying popup.
Consumer drives could not make a bad sector, only firmware modified drives could, and I didn't have one at the time.
Based on some successful cracks I did on the Apple ][, I started working through the code, only to find out they didn't check for a specific error, just an error.
Upon seeing that, I restarted my Atari with the copied disk in the drive. Ataris made a beep for each sector loaded, which was very handy. About 19, or maybe it was 23 beeps later, I opened the drive door waited for the error, closed it, and the game worked just fine.
No need for a HEX patch.
I also cracked Ultima 6. That took me 3 days of intense work. The protection was hidden in the dialog engine. So you were not seeing the actual code, but the interpreter of dialogs actually interpreting the protection in the dialog engine (that was a super simple interpreter, more like a state machine). In the end, it was just a matter of understanding the dialog data and extracting the answer to the protection question ("on page 34, what is the third word ?").
I also cracked some football game by writing a program that would intercept disk access and the replay the disk-errors to cheat it. An emulation layer if you like.
The games that I've never cracked were Elite 2 frontier (the damn thing had checks all over the place and that was too hard); "le manoir de mortevielle" had super nasty decryption stuff. IIRC they were running decrypting code inside the interrupt table (so if you had to use the keyboard or timer to trace the code, then woosh :-)). I also went mad on Leisure Suit Larry were the protection was written in an interpreted language. At that time I was 16 or 17 and had no idea of what an interpreter was = >I've never understood the scheme :-(
You had also games with special messages in it such as "we worked very hard on this so don't crack us).
But the best thing I've seen is when I've tried to look into a already cracked game .the previous hacker Mr. Detergent I think, send a message along the line of "you're weaker than me". I Was sooooo upset :-)
I guess I cracked 20 or 30 games on PC. Most of them were very simple. But in the end things were getting harder. I had to use the Soft-ICE debugger (massive thing :-)) and then, well some had these hardware debugger and came the protected mode and, well, I got lost and started working on more creative stuff (Imphobia if one remembers :-))
stF
In the Z80 processor there is a magic register, the R-register, which has a value which increments magically after each instruction has executed. It is possible to calculate how much it goes up by, for example "XOR A" increments R by one, but "LD IX,23000" increments R by two.
If R is given a predetermined value at the start of the protection system, its value is known after every instruction until the loading system starts.
This was used in many of the protection routines on Spectrums, usually as part of code that would decrypt routines, and it was a pain for the same reason that the interrupt would have been - if you're in a debugger you mess up the global state of that register.
stF
Interesting thread - well, for former BBC Micro programmers perhaps! - on the topic: http://stardot.org.uk/forums/viewtopic.php?f=1&t=696
/* complicated auth check */
...
/* more complicated auth check */
...
/* etc */
if(!isValid) {
/* do nasty things */
...
}
You just need to find isValid and flip its value. Easy cheese in a good debugger.I used to keep Soft-ICE installed on all my boxes. There were a few copy protected apps I didn't bother writing out the cracked file since I used them infrequently. Instead I'd set a breakpoint right before the e.g. isValid flag. Whenever I'd run those apps, Soft-ICE would pop up and I'd know I need to flip the bit for the full version.
For the most basic protection, find out where the copy protection check takes place, and patch the executable to skip over that check.
Obviously this would only work for the most trivial of copy protection. The basic principle remains the same, though.