Blowfish help

[Deepaak]

below is the code which is producing error, but i am unable to understand that actually what is the problem, plz someone help.

Code: [Select]

Type
  PBLOWFISH_CTX = ^BLOWFISH_CTX;
  BLOWFISH_CTX = record
    P:array [0..18] of LongInt;
    S:array [0..4,0..256] of LongInt;
  End;

Procedure Blowfish_Init(ctx : PBLOWFISH_CTX; Key : PByteArray; KeyLen : Integer); cdecl; external;

.....

Var
  ctx : PBLOWFISH_CTX;
  S : String;
  Key : TByteArray;

Begin
    S := 'TESTKEY'; 
    FillChar(Key,SizeOf(Key),0);
    Move(S[1],Key,Min(SizeOf(Key),Length(Key)));

=> Error : Blowfish_Init(ctx,@Key,Length(S)); 


Assembler Code

Code: [Select]

Blowfish_Init
00423220 55                       push   %ebp
00423221 57                       push   %edi
00423222 56                       push   %esi
00423223 53                       push   %ebx
00423224 83ec24                   sub    $0x24,%esp
00423227 8b742438                 mov    0x38(%esp),%esi
0042322B 8b6c243c                 mov    0x3c(%esp),%ebp
0042322F 31c9                     xor    %ecx,%ecx
00423231 31d2                     xor    %edx,%edx
00423233 8d3c31                   lea    (%ecx,%esi,1),%edi
00423236 6690                     xchg   %ax,%ax
00423238 8b9c9100b04200           mov    0x42b000(%ecx,%edx,4),%ebx
=>0042323F 895c9748                 mov    %ebx,0x48(%edi,%edx,4)
00423243 42                       inc    %edx
00423244 81fa00010000             cmp    $0x100,%edx
0042324A 75ec                     jne    0x423238 <Blowfish_Init+24>



[howardpc]

Why not use the native FCL Blowfish implementation (fpc/packages/fcl-base/src/blowfish.pp) rather than an external routine?
Then it would be easier to debug, and this community would be more familiar with it.

[Leledumbo]

Procedure Blowfish_Init(ctx : PBLOWFISH_CTX; Key : PByteArray; KeyLen : Integer); cdecl; external;

.....

Var
  ctx : PBLOWFISH_CTX;
  S : String;
  Key : TByteArray;

Are you sure this code even compile?

[Deepaak]

Why not use the native FCL Blowfish implementation (fpc/packages/fcl-base/src/blowfish.pp) rather than an external routine?
Then it would be easier to debug, and this community would be more familiar with it.

I would love to use it, but i cannot, i have to use this external library only. If i had to do in 'c', then it is simple, but pascal is new to me. Till the date when i had started lazarus, i had faced small challenges, which i had successfully completed with the help of this forum. This is only time i had got stuck,

[Deepaak]

Procedure Blowfish_Init(ctx : PBLOWFISH_CTX; Key : PByteArray; KeyLen : Integer); cdecl; external;

.....

Var
  ctx : PBLOWFISH_CTX;
  S : String;
  Key : TByteArray;

Are you sure this code even compile?

sir this is the part of code. Program successfully compile  and run till Blowfish_Init encounters..

[Leledumbo]

sir this is the part of code. Program successfully compile  and run till Blowfish_Init encounters..

Oops, sorry. I didn't see that the Blowfish_Init is an external declaration.

What is TByteArray? If it's dynamic array then it might be the problem. Another possibility is the calling convention of Blowfish_Init. Ensure it's really cdecl and not stdcall (or even something else though it's very unlikely).

[Deepaak]

sir this is the part of code. Program successfully compile  and run till Blowfish_Init encounters..

Oops, sorry. I didn't see that the Blowfish_Init is an external declaration.

What is TByteArray? If it's dynamic array then it might be the problem. Another possibility is the calling convention of Blowfish_Init. Ensure it's really cdecl and not stdcall (or even something else though it's very unlikely).

This is C Header File Declaration

Code: [Select]


typedef struct {
  unsigned long P[16 + 2];
  unsigned long S[4][256];
} BLOWFISH_CTX;

void Blowfish_Init(BLOWFISH_CTX *ctx, unsigned char *key, int keyLen);
void Blowfish_Encrypt(BLOWFISH_CTX *ctx, unsigned long *xl, unsigned long *xr);
void Blowfish_Decrypt(BLOWFISH_CTX *ctx, unsigned long *xl, unsigned long *xr);


and i had converted it as follows

Code: [Select]

Unit unit1;

{$mode objfpc}
{$LinkLib 'ex_bf_0001.a'} //blowfish lib

Interface

Uses
  Classes, Sysutils;

Type
 // PBLOWFISH_CTX = ^BLOWFISH_CTX;
  BLOWFISH_CTX = record
    P:array [0..18] of LongInt;
    S:array [0..4, 0..256] of LongInt;
  End;


Type TByteArr = Array [0..32] of Byte;
Type PByteArr = ^TByteArr;

Procedure Blowfish_Init(ctx : BLOWFISH_CTX; Key : TByteArr; KeyLen : Integer); cdecl; external;
Procedure Blowfish_Encrypt(ctx: BLOWFISH_CTX; xl: LongInt; xr: LongInt);cdecl;external;
Procedure Blowfish_Decrypt(ctx: BLOWFISH_CTX; xl: LongInt; xr: LongInt);cdecl;external;


I know, i am messing some important thing, but i am unable to figure it out.

[Leledumbo]

This

Code: [Select]

BLOWFISH_CTX = record
    P:array [0..18] of LongInt;
    S:array [0..4, 0..256] of LongInt;
  End;
is not a correct translation of this

Code: [Select]

typedef struct {
  unsigned long P[16 + 2];
  unsigned long S[4][256];
} BLOWFISH_CTX;
C arrays are 0 based, and the one in the square bracket is the LENGTH, not UPPER BOUND. Therefore [16 + 2] translates to [0 .. 17], [4][256] translates to [0 .. 3][0 .. 255].

unsigned long converts to LongWord. It's better to use ctype unit in order to have type compatibility with C. You can use culong for unsigned long, cint for int, etc.

[Deepaak]

This

Code: [Select]

BLOWFISH_CTX = record
    P:array [0..18] of LongInt;
    S:array [0..4, 0..256] of LongInt;
  End;
is not a correct translation of this

Code: [Select]

typedef struct {
  unsigned long P[16 + 2];
  unsigned long S[4][256];
} BLOWFISH_CTX;
C arrays are 0 based, and the one in the square bracket is the LENGTH, not UPPER BOUND. Therefore [16 + 2] translates to [0 .. 17], [4][256] translates to [0 .. 3][0 .. 255].

unsigned long converts to LongWord. It's better to use ctype unit in order to have type compatibility with C. You can use culong for unsigned long, cint for int, etc.

I had done as you said, but the error continues to be the same...

[Leledumbo]

I can't help any further without having the dll... well, assuming your translation is correct and you do everything as expected in the library documentation.

[User137]

Type TByteArr = Array [0..32] of Byte;
Type PByteArr = ^TByteArr;

Use PByte type with Lazarus. It works the way you assume PByteArr would. In Delphi, the same used to be PByteArray.

Code: [Select]

Blowfish_Init(ctx,@Key,Length(S));Pointer to pointer on Key, and ctx is uninitialized. I think it should be

Code: [Select]

Var
  ctx : BLOWFISH_CTX; // without P*
...
Blowfish_Init(@ctx, Key, Length(S));Also, don't you need to put something in ctx record first? (I don't know)

[Deepaak]

I can't help any further without having the dll... well, assuming your translation is correct and you do everything as expected in the library documentation.

 

Code: [Select]

var
 Key : AnsiString;
begin
  Key := 'TESTKEY';

  uBlowFish.Blowfish_Init(@ctx,@Key[1],Length(Key));



hi, i had done this according to your valuable guidance, but i got struct is something very new to me. that is

To encrypt a 64-bit block, call Blowfish_Encrypt with a pointer to
       BLOWFISH_CTX, a pointer to the 32-bit left half of the plaintext
      and a pointer to the 32-bit right half.  The plaintext will be
      overwritten with the ciphertext.

Procedure Blowfish_Encrypt(ctx: PBLOWFISH_CTX; xl: PLongWord; xr: PLongWord);cdecl;external;

Here x1 : 32 bit and Xr : 32 bit.

I am consfused how to convert a value into two 32 bit value (xl.xr) and vice versa for this operation.


I had found the same thing in fcl-blowfish implemention.

Code: [Select]

TBFBlock     = array[0..1] of LongInt;     { BlowFish }
But how to convert a TStream to TBFBlock, i got nothing in return. please help..

[Leledumbo]

I am consfused how to convert a value into two 32 bit value (xl.xr) and vice versa for this operation.

I'm quite confused as well since xl and xr is a pointer to 32-bit value (instead of PChar), in what encoding should the plaintext be? Nevertheless, assuming the plaintext is really stored in a 32-bit array, you can split to xl and xr with (say the whole plaintext is kept in variable pt):

Code: [Select]

xl := @pt[0];
xr := @pt[Length[pt] div 2];


But how to convert a TStream to TBFBlock

You can't, both are different thing. Unless you specify what "convert" will do to the TStream such that it can produce a TBFBlock.

[Deepaak]

I am consfused how to convert a value into two 32 bit value (xl.xr) and vice versa for this operation.

I'm quite confused as well since xl and xr is a pointer to 32-bit value (instead of PChar), in what encoding should the plaintext be? Nevertheless, assuming the plaintext is really stored in a 32-bit array, you can split to xl and xr with (say the whole plaintext is kept in variable pt):

Code: [Select]

xl := @pt[0];
xr := @pt[Length[pt] div 2];


But how to convert a TStream to TBFBlock

You can't, both are different thing. Unless you specify what "convert" will do to the TStream such that it can produce a TBFBlock.

here  64bit(qword) is to splitted into two 32bit(LongWord).

I found many example but cannot understand upto mark, or maybe i am not satisfied whether it is the correct way or not.
does fpc/lazarus has some inbuilt function to split 64bit value into two 32 bit (longword) values. If yes, then it will be good as gold. 

[User137]

There is propably some shl/shr bitshift way, but you can also use record.

Code: [Select]

TInt64Parts = record
  a, b: dword;
end;
...
var i64: int64; lowint: dword;
begin
  lowint:=TInt64Parts(i64).a;Hope it works.