BUGS DEVELOPERS

The information below has not been updated since the end of 2000 and may be out of date. However, if you wanted to use the library in your application it may still be helpful, hence the reason for keeping that information available.

19/11/00
London, England.

Sylvain Martinez

BUGS V4.0.0

Here is an explanation of the libcrypt function you can use :

In all these functions, we return 1 if the action succeeded and 0 if it failed. (INTEGER)

I FIRST STEPS

1) Include the following headers which can be found in bugs-4.0.0/include/
#include “bstandard.h”
#include “extra.h”
#include “wrapper.h”

If you are using other functions than the one defines in wrapper.h then you will need to include other headers such as main.h, utils.h, misc.h, etc

2) You need the following GLOBAL variable:
globalvar *varinit;
Which you usually declare just after the #include and just before the main()

3) you MUST CALL binit() in your application before you do ANYTHING !
This will initialise important variables needed by the library.
You can also assign your own value to these variables but it should not be necessary.
To do that, you will have to overwrite the default value assign to these variables by binit() just after you called this function if:
a) You want to use another value for KEY_BUFFER than 16 (default)
b) Overwrite the value of MISC if you want to deactivate some of the dynamic options, for example, If you only want the ROUND and SHUFFLE to be dynamic:
varinit->MISC = 0;
varinit->MISC ^= BMASK_ROUND;
varinit->MISC ^= BMASK_SHUFFLE;

Below is a small example:

/* ———– START ———– */

#include
#include “bstandard.h”
#include “extra.h”
#include “wrapper.h”

globalvar *varinit;

main()
{
binit(128,1,”test.log”,0,varinit);
printf(“\n Hello World. This is just an initialisation.\n”);
}

/* ———– END ———&##8211; */

varinit is just a structure:

/*
* NB_BYTE : Lenght in bytes of the int we are using
* NB_BITS : Number of bits of the int we are using
* NB_SHIFT : shift to use for the division in bcrypt_swap()
* KEYLENGTH : Length of the key used to crypt
* NB_INDEX : Number of index of the aray that will contain the cipher text
* NB_CHAR : Number of characters of the cipher text and of the clear text
* USER_LENGTH : Length of the username you can use
* RANDON : Random Number Generator algorithm to use
* SEED : Initial Seed to use with the RNG
* MISC : if set to 1 then it will exit most of the crypt functions
* (Useful if you are creating a multitask application)
* PROGRESS : This should show an estimation of the encryption
* /decryption progress
* PLEASE NOTE IT COULD BE >100
* KEY_BUFFER : Nb of Key to generate in order to create the buffer used in
* seed(), this creates to key dependancy.
* BCRYPT_ENDIAN : If set to 1 then your computer is using
* big endian to store data.
* VERSION : Libcrypt version number (as a string)
*/

typedef struct
{
int NB_BYTE;
int NB_BITS;
int NB_SHIFT;
int KEYLENGTH;
int NB_INDEX;
int NB_CHAR;
int USER_LENGTH;
int RANDOM;
int SEED;
int MISC;
int PROGRESS;
int KEY_BUFFER;
int BCRYPT_ENDIAN;
char LIB_VERSION[10];
}

You should not need to use these variables but this enable the different cryptography functions to have the same set-up.

II FUNCTIONS

—–> binit(int length, int random, char *file_name,int mode, globalvar *varinit)

In this function we initialise the variable.

Parameters:
1) Length of the key you are going to use
2) Random algorithm to use:
random = 0, you want to use the standard C random function
random = 1, you want to use the ISAAC random function which is supposed to be better !
3) You can specify the filename where the error output will be written if you choose mode = 1 for the next parameter
4) mode = 0, error are written in the standard error output
mode = 1, error are written in the file specified above or in ‘bugslib.log’ if the parameter 3) is an empty string.
mode = 2, no error are written
these mode are useful if you want to generate a log

5) this structure contains all the global variable used by the library.
You have to declare a variable with this type at the begining of your program.

This function MUST be called before any other function !
Only one call is required, but you can call this function several times if you want to change the key while you are crypting.
(That may crash the application if you do that …)

Rem: See the example “bcrypt.c” or anyother applications in this package

—–> RETURN_TYPE bssl(int level, int *round, int *block_crypt, int *block_shuffle, globalvar *varinit, int mode)

This function automatically set some parameters use by the cryptography library and RETURN THE POWER LEVEL. (on the 19/11/00 you have 5 power levels: 0 to 4)

You should call this function after binit()

It will set the variable sent as a parameter and some global variables as follow:

levels currently available:
D_ = Dynamic
BSSL Level Keylength Key Buffer D_Buff D_Round D_Swap D_Shuff Power Round Block_Crypt Block_Shuffle
BSSL_VLOW 128 8 no no no no 2 2 0 4
BSSL_LOW 128 8 no no yes yes 3 2 0 4
BSSL_MEDIUM 128 16 yes yes yes yes 4 2 0 4
BSSL_HIGH 256 16 yes yes yes yes 4 2 0 4
BSSL_VHIGH 512 32 yes yes yes yes 4 4 0 4

The BSSL level are defined in bstandard.h, and start from 0(NO BSSL), 1(VLOW) to 5(VHIGH)

If you specify the BSSL level to 0, then this function won’t change any parameters and send back a negative value for the POWER level.

Parameters:
1) BSSL level you want to use
2) this will set the nb of round
3) this will set the block crypt
4) this will set the block shuffle
5) Your global variable
6) mode = 0, quiet (no warnings, no information)
mode = 1, verbose (few warnings and information)
mode = 2, debug (all warnings and information, can generated HUGE ouput files !)
7) Your global variable.

—–> bkey_generator (unsigned char *pass_param, int length, int complexity, char *file_path, int power, int random, int mode, globalvar *varinit)

This function is useful to generate key.

Parameters:
1) If you want to initialise the key generation with a password, you have to send it in this parameter.
2) You have to send the length of the string send in the first parameter if you do not want to use a password as the initialisation, just sent the 0 value in this parameter.
3) The complexity that is going to be used during the swapping process should be 2. Higher value may increase security but there is no statistical proof of this and it would slow down slightly the generation. The choice is yours.
4) location of the file that will contain the key : key file
5) Power you want to use to generate the key. (POWER 0,1,2,3 or 4)
6) random = 0, you use a password for the initialisation. So you have to be sure that the first parameter contain some characters.
random = 1, you want to use the random initialisation. We use the standard C random function.
7) mode = 0, quiet (no warnings, no information)
mode = 1, verbose (few warnings and information)
mode = 2, debug (all warnings and information, can generated HUGE ouput files !)
8) Your global variable.

Rem: See the example “bkey.c”

——> blogin (TYPE_INT *code_file, unsigned char *pass_clear, int length, int power, int complexity, int mode, globalvar *varinit)

This function is used if you want to manage several user, in fact to replace or to add a new login step. this function check if the clear password send in parameter will have the same cipher form than the cipher passwd sent in parameter.

So you have to use the function read_passwd before to get the cipher password that matche a user.

Parameters:
1) you have to send the password of the user you get with passwd_funtion
2) you send the clear password that you want to check if he will generate the same cipher password sent in the first parameter
3) length of the clear password sent in the second parameter.
4) Power you want to use to crypt. (POWER 0,1,2,3 or 4)
5) The complexity that is going to be used during the swapping process should be 2. Higher value may increase security but there is no statistical proof of this and it would slow down slightly the generation. The choice is yours.
6) mode = 0, quiet (no warnings, no information)
mode = 1, verbose (few warnings and information)
mode = 2, debug (all warnings and information, can generated HUGE ouput files !)
7) global variable

Once the clear password has been crypted, we compare with the cipher password sent in parameter. And we return 1 if it is the same cipher form.

Rem: See the example “blogin.c”

—–>bpass (TYPE_INT *pass_code, unsigned char *pass_clear, int length, int power, int complexity, int mode, globalvar *varinit)

This function generate a cipher password from a clear string.

Parameters:
1) This is were the cipher password will be stored
2) it is your clear string you want to crypt
3) it is the length of the string sent in the third parameter
4) Power you want to use to crypt. (POWER 0,1,2,3 or 4)
5) The complexity that is going to be used during the swapping process should be 2. Higher value may increase security but ther is no statistical proof of this and it would slow down slightly the generation. The choice is yours.
6) mode = 0, quiet (no warnings, no information)
mode = 1, verbose (few warnings and information)
mode = 2, debug (all warnings and information, can generated HUGE ouput files !)
7) global variable

Rem: See the example “bpass.c”

—–> bfile (int choice, char *name_clear, char *name_code, char *name_key unsigned char *pass_param, int length_pass, int power, int complexity, int block_crypt, int block_shuffle, int memory, int mode, globalvar *varinit)

This function will crypt a file
Please note that if you choose the ASCII mode, the crypted data will be placed between the 2 following KEYWORDS:
[BUGS_ASCII_MODE_v02_START]
[BUGS_ASCII_MODE_v02_END]

Where v02 is the ASCII algorithm version.

Parameters:
1) choice = 0, you want to crypt
choice = 1, you want to uncrypt
choice = 2, you want to crypt and generate an ASCII crypted file
choice = 3, you want to decrypt a file which is an ASCII crypted file.
2) name of the file you want to crypt
3) name of the file crypted
4) name of the file that contain the key, if you want to use a key file instead of a password. You can also use any kind of file to crypt! if you do not want to use a key file, just send “”
5) Password you want to use to crypt the file. if you use a key file, just send “”
6) length of the string sent in the previous parameter
7) Power you want to use to crypt. (POWER 0,1,2,3 or 4)
8) The complexity that is going to be used during the swapping process should be 2. Higher value may increase security but ther is no statistical proof of this and it would slow down slightly the generation. The choice is yours.
9) The crypt’s block is the size of your “work space” which mean the algorithm will crypt a file working within block having the size you specified in this parameter. The block crypt’s has to be at least as big as the size of the key used. (if you are using a 128 bits key, you need this parameter to be >= 16).
If this parameter is set to 0 it means the algorithm will consider the file as only ONE big block (default value).
10) The shuffle’s block is the size of the block used during the shuffle process. It has to be at least has big as the size of the type of integer you are using. By default, the TYPE_INT is set to “long” in libcrypt.h. On Linux the long type is 4 BYTES, in this case this parameter has to be >=4.
For security and efficiency reason this parameter has to be at least 6 times smaller than the block’s crypt value. (previous parameter).
When you use a 128 bits keylength and the probability seed (power 1 or 4) then the actual key length is not 16 but 32 as with the probability seed the length of the string is increase by the size of the key. In other words with the power 1 or 4 you can have a shuffle’s block of 4 and a crypt’s block of 16.
11) memory = 0, The cryptography algo will use less memory and will do more disk access (slow)
memory = 1, The cryptography algo will load all the blocks into memory and will do less disk access. This method is the fastest but could generate an error if you try to crypt a very very large file… This should anyway be the DEFAULT MODE !
12) mode = 0, quiet (no warnings, no information)
mode = 1, verbose (few warnings and information)
mode = 2, debug (all warnings and information, can generated HUGE ouput files !)
13) global variable

Rem: See the example “bcrypt.c”

—–> bstream (int choice, unsigned char *stringtocrypt, int length_string, char *name_key, unsigned char *pass_param, int length_pass, int power, int complexity, int block_shuffle, int mode, globalvar *varinit)

This function is similar to the bfile() one but here you are
crypting a string.

Parameters:
1) choice = 0, you want to crypt
choice = 1, you want to uncrypt
2) String you want to crypt
3) Length of the string to crypt (previous parameter)
4) Name of the Key file, if you are using a password instead just send an empty string as a parmaeter “”
5) Password to use to crypt the string
6) length of the string sent in the previous parameter
7) Power you want to use to crypt. (POWER 0,1,2,3 or 4)
8) The complexity that is going to be used during the swapping process should be 2. Higher value may increase security but ther is no statistical proof of this and it would slow down slightly the generation.
The choice is yours.
9) The shuffle’s block is the size of the block used during the shufflevprocess. It has to be at least has big as the size of the type of integer you are using. By default, the TYPE_INT is set to “long”
in libcrypt.h. On Linux the long type is 4 BYTES, in this case this parameter has to be >=4.
For security and efficiency reason this parameter has to be at least 6 times smaller than the block’s crypt value. (previous parameter).
When you use a 128 bits keylength and the probability seed (power 1 or 4) then the actual key length is not 16 but 32 as with the probability seed the length of the string is increase by the size of the key. In other words with the power 1 or 4 you can have a shuffle’s block of 4 and a crypt’s block of 16.
6) mode = 0, quiet (no warnings, no information)
mode = 1, verbose (few warnings and information)
mode = 2, debug (all warnings and information, can generated HUGE ouput files !)
11) global variable

Rem: See the example “bchat.c”

—– >bcrypt_read_passwd (char *user, char *file_path, TYPE_INT *code_file, int mode, globalvar *varinit)

This function read a password of a user in a password file. You generate the password file with the write_password function.

Parameters:
1) name of the user that you want to get the cipher password in the password file
2) location of the password file
3) it is where the cipher password will be stored
4) mode = 0, quiet (no warnings, no information)
mode = 1, verbose (few warnings and information)
mode = 2, debug (all warnings and information, can generated HUGE ouput files !)
5) global variable

Rem: See the example “bpass.c” or “blogin.c”

—–> bcrypt_write_passwd (char *user, TYPE_INT *pass_code, char *file_path, int mode, globalvar *varinit)

This function writes a password in a password file, similar /etc/password.

Parameters:
1) name of the user
2) that should contain the cipher password of the user
3) location of the password file
4) mode = 0, quiet (no warnings, no information)
mode = 1, verbose (few warnings and information)
mode = 2, debug (all warnings and information, can generated HUGE ouput files !)
5) global variable

Rem: See the example “bpass.c”

—–> bcrypt_delete_passwd (char *pass_file, char *user, int keylength, int mode, globalvar *varinit)

This function delete a user from the password file. The old password file is saved in the file pass.old

Parameters:
1) location of the password file
2) name of the user you want to delete
3) Length of the key used to generate the cipher password of the user.
Be very careful with this parameter, indeed, if you specify a length superior of the key’s length used to generate the cipher password, you may destroy other user’s data contained in the password file.
It is why you should not use differents key’s length in your password file. You can, but you have to manage it carefully.
4) mode = 0, quiet (no warnings, no information)
mode = 1, verbose (few warnings and information)
mode = 2, debug (all warnings and information, can generated HUGE ouput files !)
5) global variable.

Rem: See the example “bpassdel.c”

—–> bcrypt_write_hide (int choice, char *source_file, char *dest_file, int mode)

This function hide some data in a file. You can use this function to hide your cipher text.

Parameters:
1) choice = 0, hide data at the begining of a file
choice = 1, hide data at the end of a file
2) location of the file you want to hide
3) location of the file that recevie the data to hide.
4) mode = 0, quiet (no warnings, no information)
mode = 1, verbose (few warnings and information)
mode = 2, debug (all warnings and information, can generated HUGE ouput files !)

This function works quite simply, but is enough powerful to hide some data in picture, sound, zip, exe, etc … To make that invisible choose to hide at the end of a file.

We may work later on a *real* steganograpyh algorithm, but as usual time is the problem.

Rem: See the example “bhide.c”

—–> bcrypt_read_hide (int choice, char *source_file, char *dest_file, int mode)

This function extract data from a file. Use it if you want to recover some hidden data.

Parameters:
1) choice = 0, extract data from the begening of a file
choice = 1, extract data from the end of a file
2) Location of the file that contain the hidden data
3) Location of the file that will receive the hidden data that has been extracted
4) mode = 0, quiet (no warnings, no information)
mode = 1, verbose (few warnings and information)
mode = 2, debug (all warnings and information, can generated HUGE ouput files !)

Rem: See the example “bhide.c”

——> bpow(int base, int n)

This function calculate a power : base^n
The standard function on Linux seems to have some problems to work properly (back in 2000), so we decided to write our own.

Parameters:
1) The number you want to calculate the power
2) The value of the power