SMS Text messages, character count and cellphones

I've been playing around with SMS PDUs encodings and recently observed some curious things regarding the way that cellphones treat the text messages when they include special characters.First look at these two tables of the GSM alphabet:GSM7 Alphabet

GSM7 Alphabet (special)
The tables above show the GSM Alphabet using a 7-bit encoding which implies a maximum number of 160 characters per text message. However, sometimes this is a little bit more tricky and the count is not so straightforward:
If, for instance, you type a character from the second table, it needs to be escaped with the escape character 1) 0xB1, and it will take two characters instead of one. What happens if your text message contains 160 characters and one of them is a ']'?
Simple: You will get charged for two text messages because you're exceeding the maximum length of a simple PDU.
Usually your cellphone won't warn you about this and you will send it anyways without knowing the fact that this text message will cost twice than you think. The same happens with the '€' symbol and some other not showing up in the table above.If you have a closer look at the tables, you might realize that the 'é' symbol appears but where are 'á','í','ó' and 'ú'? The GSM alphabet was originally designed by France and they only use 'é' so if you want to use accents, there's no way using this encoding.
I have tested some cellphones and they behave in two different ways:

  1. Removing those characters with accents (except 'é') and substituting them by the same character without accent.
  2. Using a 16-bit UNICODE encoding to allow sending every character (in this case, the maximum length of the textmessage is 70 characters).

In case 1, the only 'side-effect' is that the recipient of the text message won't get your accents and you can send up to 160 characters.
In case 2, your cellphone won't warn you and you might send up to 160 characters thinking that it will take just one text message.However, again, you will be charged for up to 3 text messages without knowing it! The recipient will get a multi-part message showing all the characters you sent with no modification.
Here you can see the decoding of a PDU (using PDUSpy) of a text message sent with accents and encoded using UCS2:

  • PROTOCOL USED : Implicit / SC-specific
  • ALPHABET USED : 16bit UCS2

If the accents are removed from the original text message, the cellphone will automatically use the GSM7 alphabet and you will be allowed to send up to 160 characters in just one PDU (you will get charged once).All in all, be careful and if possible make some research to figure out what your cellphone does and check it against your bill because you will probably save some (or a lot of) money.

Setting up iPhone SDK on a Virtual Machine

Disclaimer: All the information posted is intended for illustrative and educational purposes only. I just want to show you that it's possible to set up the iPhone SDK on a Virtual Machine. Please, BUY an Apple Mac OS X License if you are going to use this and BUY a Mac computer (Apple's EULA agreement states that you cannot run Mac OS X under non Apple hardware).

In this post I will try to explain how to set up the SDK for iPhone OS 3.1 on a PC running Windows (Vista 64 in my case). From the readme file of the SDK you can read:

Xcode 3.1.4, when used for Mac-only development, is compatible with Intel and PowerPC Macs running Mac OS X Leopard 10.5 and later. Use of the iPhone SDK requires an Intel-based Mac running Mac OS X Leopard version 10.5.7 or later.

So we need a Leopard 10.5. By googling a little bit you will realize that there are some modified ready-to-use Leopard images out there that you can download and run out of the box on VMWare.

I'm running a 10.5.2 version (which takes about 5 mins to boot on my quad core). Once you get a Mac OS X running on your PC, you can download the free iPhone SDK from and install it.

Mac OS X Leopard on VmWare under Vista

The iPhone SDK for OS 3.1 won't install under a version prior to 10.5.7 so I had to 'trick' the installer rather to update the Mac OS X which seems to be a painful process. To do this, you have to modify the /System/Library/Core Services/SystemVersion.plist and change both the ProductUserVisibleVersion and ProductVersion keys to 10.5.7.

With this done, I selected just the SDK for 3.1 (to save space in my hard disk) and the installation process begins.

Custom Install  And 3.5 hours later....  Install Succeeded

Once you have the iPhone SDK installed, you can run Xcode (from SpotLight) and launch a new project using a template just to try it out on the iPhone Simulator:

Xcode and simulator   Xcode and simulator (2)

At this point you can develop your own iPhone applications and test them on the simulator. Also, if you joined the Apple developer program (the standard one is $99) you can test them in your iPhone as well.

If you're planning to play around with the SDK I strongly recommend you to sign up on the iPhone Dev Center because there are lots of resources available: Getting started documents, videos,  sample code, etc.

More to come,


Windows Mobile - Routing audio through the Earpiece (PPC / SmartPhone)


I've developed a little application that switches the audio output from the rear speaker to the front one and viceversa. This is useful for VoIP applications which are quite unusable without headphones since the audio comes from the back speaker. It just runs for 10 minutes and it's supposed to work at least with the latest HTC models.

Download PocketPC Trial version:

Download SmartPhone Trial version:

Don't hesitate to contact me for any further information.


LaTeX in WordPress

When I had to write some formulas in the blog I used to render them from LaTeX to images using some websites such as Online LaTeX Equation Editor. It has the disadvantage that you've got to write the equation, render it, save it to your harddisk and then upload to the WordPress file manager. Now, thanks to an interesting plugin written by Steve Mayer you can embed its usage into WordPress just by typing the LaTeX commands between special tags.


It uses MimeTeX  which is an standalone program that directly renders LaTeX expressions into images without using the entire TeX package or its fonts. Thus, it's a simple, lightweight and elegant solution ready to be used in your websites or blogs.


Line Following Robot - DIEZ

DIEZ_fDIEZ is a line following robot which has been designed for several national contests being held in Spain throughout the year.

The goal for this kind of robots is to follow a black line over a white floor as quick as possible. The track has also branches which should be taken properly by the robot. It should be able to read some little strips located at both sides of the track which shows the robot the right direction to take in the next branch. In almost all the competitions, taking the wrong side of the branch makes the robot to travel a longer distance and, sometimes involves a 10-15 seconds penalty over the total time achieved.

The tracks usually have many turns and 90 degrees corners so DIEZ has been designed to be lightweight and small in order to achieve these movements as fast as possible with the less moment of inertia.

Prior to the final assembly of the robot, a CAD design was performed when all the decisions about the design were made. Here you can see a GIF showing the first steps from CAD to assembly phases:


Sensors, Control Board and Algorithm:

The main board is an LPC2138-01 based on a 32-bit LPC2138 microcontroller at 60 MIPS.

The sensor board's got 10 tiny SMD reflective sensors which are read by the microcontroller with one of its A/D converters. Due to the different light conditions in the contests and sometimes, the camera flashes and the reflection ratios of the track surfaces, DIEZ performs an ambient light cancelation by reading the sensors with the IR led switched off. Besides, the IR led light is modulated to achieve an even more reliable reading from the sensors.

The control algorithm used by DIEZ is a PD controller. The input of the controller is the position of the center of the line, which is estimated using an interpolation function of the values read from the sensor board. A C# application's Control Centerbeen developed to test and debug the algorithm before coding it into the microcontroller. Here you can see an screenshot of this program which allowed us to monitor in Real Time what the robot saw in every moment. Also, thanks to the SD socket of the LPC2138-01 board, the robot was recording all the data and parameters on it for a later study with this program. This was specially useful to detect some of the problems of the algorithm because we had all the telemetry. This recording was quite feasible using the Embedded File System Library since the robot just had to create a new telemetry file on a FAT32 filesystem which was perefctly readable from any computer.

Weight and Dimensions

The total weight of DIEZ is less than 400 gr (about 14 oz) with two Li-Po batteries (1450 mAh).

Its size is 15 x 15 x 7 cm (5.9 x 5.9 x 2.7 inches).


Here you can see a video with DIEZ in action. Those 'ugly' pink thingies on its wheels are stripes of a water balloon which we had handy to increase the adherence and, thus, the maximum speed to reach a corner. Due to its light weight, these rubbers were necessary to avoid the skids.

That's all 🙂

Again, hope you liked the article.


ARM Development Board - LPC2138-01


I have designed a new development board which is supposed to be used in many of the projects I am/will be involved in. It's been built around an NXP LPC2138 microcontroller which has the following main features:

  • 16/32 bit ARM7TDMI-S microcontroller (LQFP64 package)
  • 32 KB of internal SRAM and 512 KB of on-chip flash program memory. 128-bit wide interface/accelerator enables high-speed 60 Mhz operation.
  • In-System Programming/In-Application Programming (ISP/IAP) via on-chip bootloader software. Single flash sector or full chip erase in 400 ms and programming of 256 B in 1 ms.
  • Two 8-channel 10-bit ADCs with conversion times as low as 2.44 us per channel.
  • Two 32-bit timers/external event counters (with four capture and four compare channels each), PWM unit (six outputs) and watchdog.
  • Multiple serial interfaces including two UARTs (16C550), two Fast I2C-bus (400 kbit/s), SPI and SSP with buffering and variable data length capabilities.
  • Vectored interrupt controller with configurable priorities and vector addresses.
  • Up to forty-seven 5 V tolerant general purpose I/O pins in tiny LQFP64 package.
  • Up to nine edge or level sensitive external interrupt pins available.
  • 60 MHz maximum CPU clock available from programmable on-chip PLL with settling time of 100 us.
  • On-chip integrated oscillator operates with external crystal in range of 1 MHz to 30 MHz and with external oscillator up to 50 MHz.
  • Power saving modes include Idle and Power-down.
  • Individual enable/disable of peripheral functions as well as peripheral clock scaling down for additional power optimization.
  • Processor wake-up from Power-down mode via external interrupt or BOD.
  • Single power supply chip with POR and BOD circuits:
    • CPU operating voltage range of 3.0 V to 3.6 V (3.3 V +- 10 pct) with 5 V tolerant I/O pads.

This microcontroller is suitable for most of my projects due to its high processing power (up to 60 MIPS) and its large amount of memory. Also, the board itself has some kind of features that I couldn't find in any other development board:

  • Audio playing support with earphones connector. Connected to the DAC output of the microcontroller through an audio amplifier circuit. It is possible to use the DAC output for another purposes.
  • Audio recording support with microphone connector. Connected to one of the ADCs channels of the microcontroller through an audio amplifier and signal conditioning circuit. It is possible to use the ADC input for another purposes.
  • Selectable power supply input: directly from USB or from external battery for standalone applications (Vin ranging from 3.3V to 20V).
  • RS232 interface through DB9 connector.
  • USB communication through USB Type-B connector (on-board USB-Serial converter). RX and TX leds.
  • 8 General Purpose LED's connected to the microcontroller through an output buffer which can be disabled through a jumper in case that these GPIO pins are needed.
  • Reset Button
  • General purpose button connected to one of the external interrupts of the microcontroller.
  • In Circuit Programming interface up to 115200 bps.
  • 12 MHz crystal for operation up to 60 MHz.
  • MMC/SD memory card interface connected to the SPI bus of the microcontroller.
  • Integrated Li-Ion / Li-Po Battery charger. It can charge batteries from USB or external power supply with LED charge indicators.
  • All the GPIO pins are available in the board. Including JTAG interface for programming / debugging in circuit.
  • Battery level monitoring. The input voltage is connected to one of the ADC inputs of the microcontroller through a voltage divisor so that it can be measured by software.


As the ARM architecture is being more and more popular there are lots of development tools available for all the platforms.

On Windows machines you can use WinARM which is a collection of tools to develop software for the ARM-family of controllers/processors including GNU C/C++ compiler (gcc / g++), GNU Binutils, GDB server, ...

Most of these tools are part of GNU arm-elf toolchain distributions so you can easily develop /program the LPC2138 microcontroller under GNU/Linux.

Also, you can find some good commercial software for ARM microcontrollers/microprocessors like the ARM RealView Development Suite.

In some applications an RTOS (Real Time Operating System) is needed and there are some of them available for this microcontroller. I would stand out the following ones: FreeRTOS (free for non-commercial purposes) and RealView RL-ARM.



Crackme 1

In this tutorial I will explain how to solve a crackme rated level 4 at from DrSpliff which I found very interesting.

You can download the original author's file from here.
You can download both the original file and the patched one created by me from here.

The readme file says:

This crackme uses the basic principals and building blocks of software protection on Unix systems with a little bit of crypto thrown in to piss you off.

Difficulty: 4 - Needs special knowledge
Platform: Unix/linux etc.
Language: C/C++

Actually this is not a Linux dependent crackme since it's got no specific Unix/Linux protections but it was compiled for it though.

First thing we do is executing the binary (although I trust I use to have a quick look at the binaries to avoid malicious code being executed on my machine).

daniel@gargamel:~/crackme$ ./drscm1
Dr.Spliff's Crackme - v0.1 (Alpha) (Linux/ia32)
Name: daniel
Serial: 123
Sorry daniel, your license is invalid

Having a quick look at the binary and loading it into gdb it seems that the author's not used any antidebugging techniques nor ELF header corruption so we can freely analyze / disassemble it.

ida-graphviewFirst I opened it with IDA (my favourite disassembler. It's for Windows but you can run it under wine) to analyze the program flow with its nice Graph View. Quickly I can see where in the code it reads our name / serial (see image).

After this, some memory is allocated via malloc and a buffer is copied into it. This buffer is 56 bytes long and it's located at address 0x8049a20 (this may change in your computer). After the copy, this buffer is modified somehow by this piece of code:

At the beggining of the loop, edi is pointing to our buffer and it's being processed in blocks of 8 bytes (64 bits).

.text:08048742                 push    edi
.text:08048743                 add     edi, 8
.text:08048746                 push    offset 08049A5C
.text:0804874B                 call    sub_080485AF
.text:08048750                 add     [ebp+var_410], 8
.text:08048757                 pop     eax
.text:08048758                 pop     edx
.text:08048759                 movsx   eax, byte_8049A58
.text:08048760                 cmp     [ebp+var_410], eax
.text:08048766                 jl      short loc_8048742

If we look further into sub_080485AF, we can identify its main loop with a 'suspicious' structure:

.text:080485DB                 mov     edx, ecx
.text:080485DD                 lea     eax, [esi+ecx]
.text:080485E0                 shl     edx, 4
.text:080485E3                 add     edx, [ebp+var_14]
.text:080485E6                 xor     edx, eax
.text:080485E8                 mov     eax, ecx
.text:080485EA                 shr     eax, 5
.text:080485ED                 add     eax, [ebp+var_18]
.text:080485F0                 xor     edx, eax
.text:080485F2                 sub     ebx, edx
.text:080485F4                 mov     edx, ebx
.text:080485F6                 shl     edx, 4
.text:080485F9                 lea     eax, [esi+ebx]
.text:080485FC                 add     edx, [edi]
.text:080485FE                 add     esi, 61C88647h
.text:08048604                 xor     edx, eax
.text:08048606                 mov     eax, ebx
.text:08048608                 shr     eax, 5
.text:0804860B                 add     eax, [edi+4]
.text:0804860E                 xor     edx, eax
.text:08048610                 sub     ecx, edx
.text:08048612                 dec     [ebp+var_10]
.text:08048615                 jns     short loc_80485DB

Notice the 4-bit shifts to the left and the 5-bit shifts to the right. It looks familiar... hmm, maybe TEA? That's it! Having a look at a public implementation of TEA algorithm we can say for sure it is and this is actually the decoding function. If this is the decoding function, the key used shouldn't be too far away and the stack is a good place to look at first. Indeed, it's there and it looks like it's been hardcoded by the author (at least there are no x-refs to it).

Okay, let's go along the code and see what happens after all this crypto stuff. We find this piece of code:

.text:08048768                 push    0               ; prot
.text:0804876A                 movsx   eax, byte_8049A58
.text:08048771                 push    eax             ; len
.text:08048772                 push    ds:addr         ; addr
.text:08048778                 call    _mprotect
.text:0804877D                 push    offset sub_804855B
.text:08048782                 push    offset sub_80484FC
.text:08048787                 push    ebx
.text:08048788                 call    ds:addr

Interesting huh? The address at 'ds:addr' is actually our decrypted buffer. The mprotect function call is used to set protection of a memory mapping and as we can see in the code above the protection argument is zero or (PROT_NONE) to allow execution of the memory which is about to be executed in the 'call ds:addr' instruction.

The arguments of the function which is supposed to be in 'ds:addr' are two procedures: sub_804855B and sub_80484FC. Clearly these procedures are the 'bad and good' boys. I could try to change the code so that the good boy address is pushed twice onto the stack but the function at 'ds:addr' could check that the two parameters are distinct and this wouldn't be a real crack 🙂 I'd like to dig more into the executable to find out what the author really pretended.

So, the way to go is to load the executable with GDB and break just after the data is already decrypted so that we can analyze what it really holds:

 (gdb) x/56i *0x8049b8c
0x804a028:      push   %ebp
0x804a029:      mov    %esp,%ebp
0x804a02b:      push   %esi
0x804a02c:      sub    $0x4,%esp
0x804a02f:      mov    0x8(%ebp),%esi      ; take a ptr to a ptr to the name
0x804a032:      mov    $0x0,%ecx           ; ecx = 0
0x804a037:      mov    (%esi),%edx         ; edx = ptr to the name
0x804a039:      cmpb   $0x0,(%edx)         ; compare ''
0x804a03c:      je     0x804a049           ; if string's finished jump to 0x804a049
0x804a03e:      movsbl (%edx),%eax         ; eax = name[i]
0x804a041:      add    %eax,%ecx           ; ecx += name[i]
0x804a043:      inc    %edx                ; ptr++
0x804a044:      cmpb   $0x0,(%edx)         ; name[i] == '' ?
0x804a047:      jne    0x804a03e           ; no -> jump to loop
0x804a049:      cmp    0x4(%esi),%ecx      ; compare with the serial when string's fully proccessed
0x804a04c:      jne    0x804a057           ; if we nop this jump, then we'll always have the goodboy
0x804a04e:      sub    $0xc,%esp
0x804a051:      push   %esi
0x804a052:      call   *0xc(%ebp)          ; call good boy
0x804a055:      nop
0x804a056:      nop
0x804a057:      sub    $0xc,%esp
0x804a05a:      push   %esi
0x804a05b:      call   *0x10(%ebp)         ; call bad boy
0x804a05e:      nop
0x804a05f:      nop

The good boy is at ebp+0x0C and the bad boy at ebp+0x10C. The routine performs a simple algorithm and then compares the result to the serial we entered. If they match, then we'll have the good boy. We can see that writing a keygen is pretty easy now and that's why I won't do it.

So, by simply nopping the 'jne 0x804a057' instruction located at 0x804a04c we'll get the application to show the good boy no matter whether the serial is right or not. To do this we have to dump these 56 bytes, patch and re-cypher them using TEA with the same key.

Key used by the executable for decryption:


db 75h, 8, 0D3h, 16h, 0D1h, 3, 0Eh, 0 ; 0
db 0E5h, 2Bh, 96h, 0, 8Fh, 45h, 1Eh, 0 ; 8

TEA cyphered data:

.data:08049A20 db 0Ch, 26h, 0F5h, 42h, 0A3h, 0ACh, 0D3h, 3Ch; 0
.data:08049A20 db 0DCh, 0B3h, 86h, 0B2h, 4Bh, 0C5h, 3Ch, 29h; 8
.data:08049A20 db 88h, 66h, 49h, 73h, 0A4h, 0CFh, 7Bh, 0B1h; 16
.data:08049A20 db 3Ah, 5Bh, 0CBh, 37h, 0BDh, 2Ah, 35h, 0FCh; 24
.data:08049A20 db 95h, 0E7h, 0E8h, 0BAh, 2Bh, 66h, 67h, 0BCh; 32
.data:08049A20 db 2Dh, 0AFh, 34h, 32h, 2Eh, 5Bh, 0C1h, 0Eh; 40
.data:08049A20 db 1Eh, 0FAh, 8Eh, 63h, 6Dh, 0C1h, 51h, 78h; 48

Putting this all together we need a TEA implementation. I will use the following source code:

void code(long *v, long *k)

{   unsigned long y = v[0],
                  z = v[1],
                  sum = 0,		/* set up */
                  delta = 0x9e3779b9,
                  n = 32,
                  a = k[0],
                  b = k[1],
                  c = k[2],
                  d = k[3];
while (n-- > 0)			/* basic cycle start */
    {   sum += delta;
        y   += (z <> 5) + b;
        z   += (y <> 5) + d;		/* end cycle */
v[0] = y;
    v[1] = z;
void decode(long *v, long *k)
{   unsigned long n = 32,
                  y = v[0],
                  z = v[1],
                  delta = 0x9e3779b9,
                  a = k[0],
                  b = k[1],
                  c = k[2],
                  d = k[3];
sum = delta < 0)
    {   z   -= (y <> 5) + d;
        y   -= (z <> 5) + b;
        sum -= delta;
 					/* end cycle */
    v[0] = y;
    v[1] = z;
int main(int argc, char **argv)
long cyphered[] = {0x42F5260C, 0x3CD3ACA3, 0x0B286B3DC, 0x293CC54B, 0x73496688, 0x0B17BCFA4, 0x37CB5B3A, 0x0FC352ABD, 0x0BAE8E795, 0x0BC67662B, 0x3234AF2D, 0x0EC15B2E, 0x638EFA1E, 0x7851C16D };
long key[] = {  0x16D30875, 0x0E03D1, 0x962BE5, 0x1E458F };
long *ptrc;
int i;
 	decode((unsigned long *)ptrc, (unsigned long *)key);
 return 1;

Executing this code we will get our buffer decrypted (we could have got this data by debugging the executable itself but this way we are also checking that everything is going all right). Having a look at the 'cyphered' buffer with gdb on the above program:

(gdb) p/x cyphered
{0x56e58955, 0x8b04ec83, 0xb90875, 0x8b000000,
0x3a8016, 0xbe0f0b74, 0x42c10102, 0x75003a80,
0x44e3bf5, 0xec830975, 0x55ff560c, 0x8390900c,
0xff560cec, 0x90901055}

Okay here is the decyphered buffer with actually holds the code executed by our binary. It contains the jnz to be nopped by us so we can clearly identify the JNZ opcodes (0x0975) which should be changed by two NOP's (0x9090) and the buffer would look like this:

{0x56e58955, 0x8b04ec83, 0xb90875, 0x8b000000,
0x3a8016, 0xbe0f0b74, 0x42c10102, 0x75003a80,
0x44e3bf5, 0xec839090, 0x55ff560c, 0x8390900c,
0xff560cec, 0x90901055}

So now it's time to cypher this data using the same key so that when our program decrypts it, we always get the good boy.

int main(int argc, char **argv)
{long decyphered[] = {	0x56e58955, 0x8b04ec83, 0xb90875, 0x8b000000,0x3a8016, 0xbe0f0b74, 0x42c10102, 0x75003a80,0x44e3bf5, 0xec839090, 0x55ff560c, 0x8390900c,0xff560cec, 0x90901055 };
long key[] = {  0x16D30875, 0x0E03D1, 0x962BE5, 0x1E458F };
long *ptrd;
char *ptr;
int i;
 	code((unsigned long *)ptrd, (unsigned long *)key);
return 1;

The output of this program:

0x0C, 0x26, 0xF5, 0x42, 0xA3, 0xAC, 0xD3, 0x3C,
0xDC, 0xB3, 0x86, 0xB2, 0x4B, 0xC5, 0x3C, 0x29,
0x88, 0x66, 0x49, 0x73, 0xA4, 0xCF, 0x7B, 0xB1,
0x3A, 0x5B, 0xCB, 0x37, 0xBD, 0x2A, 0x35, 0xFC,
0x05, 0x51, 0x3F, 0xE3, 0xCF, 0x7F, 0x2D, 0x78,
0x2D, 0xAF, 0x34, 0x32, 0x2E, 0x5B, 0xC1, 0x0E,
0x1E, 0xFA, 0x8E, 0x63, 0x6D, 0xC1, 0x51, 0x78,

Actually the only block (8 bytes long) which has changed's been obviously the one containing the NOP'ed JNZ so its bytes are to be substituted in the original binary file. We open it with our favourite hex-editor and patch the file.

Once we've patched the file it's time to try and see what's going on after all this job:

daniel@gargamel:~/crackme$ ./drscm1-patched
Dr.Spliff's Crackme - v0.1 (Alpha) (Linux/ia32)
Name: daniel
Serial: 123
Congrats daniel, your license is valid
Now, go post your solution on so others can learn

daniel@gargamel:~/crackme$ ./drscm1-patched
Dr.Spliff's Crackme - v0.1 (Alpha) (Linux/ia32)
Name: dani
Serial: 4535
Congrats dani, your license is valid
Now, go post your solution on so others can learn

daniel@gargamel:~/crackme$ ./drscm1-patched
Dr.Spliff's Crackme - v0.1 (Alpha) (Linux/ia32)
Name: aa
Serial: 194
Congrats aa, your license is valid

Great ! It worked! Now let's have a look at the modified code to see how it looks like after the patch:

(gdb) x/24i 0x804a028
0x804a028:      push   %ebp
0x804a029:      mov    %esp,%ebp
0x804a02b:      push   %esi
0x804a02c:      sub    $0x4,%esp
0x804a02f:      mov    0x8(%ebp),%esi
0x804a032:      mov    $0x0,%ecx
0x804a037:      mov    (%esi),%edx
0x804a039:      cmpb   $0x0,(%edx)
0x804a03c:      je     0x804a049
0x804a03e:      movsbl (%edx),%eax
0x804a041:      add    %eax,%ecx
0x804a043:      inc    %edx
0x804a044:      cmpb   $0x0,(%edx)
0x804a047:      jne    0x804a03e
0x804a049:      cmp    0x4(%esi),%ecx
0x804a04c:      nop    			<---- Our NOP-ed JNZ
0x804a04d:      nop    			<---- Our NOP-ed JNZ
0x804a04e:      sub    $0xc,%esp
0x804a051:      push   %esi
0x804a052:      call   *0xc(%ebp)	<---- Always call the good boy
0x804a055:      nop

This is the hardest part of the crackme. However we assumed for simplicity that the serial entered is a number. However the program checks the return of an atoi() call to the serial. If atoi fails then the 'Invalid serial' message appears. To avoid this, just patch the jne after the atoi call and replace it by an unconditional jump:

80486d8:     jne    0x80486f8

Open the file with an hexeditor and change the 0x75 (JNE opcode) byte at 0x6d8 by a 0xEB (JMP opcode).

Also, the program performs a check to the length of the input name:

or      edx, 0FFFFFFFFh
xor     eax, eax
mov     ecx, edx
repne scasb
not     ecx
dec     ecx
cmp     ecx, 1
ja      short loc_804869A

This is a typical 'strlen' and the jump to 0x804869A is taken just if the length of the name is > 0.
So if we want to skip this restriction, we could patch the JA opcode the same way: Replacing the 0x77 (JA) byte at 0x68C offset of the file by a 0xEB (JMP).

The End.

Hope you enjoyed this article and found it useful.