How can I pad my md5 message with c/c++

Question

I'm working on a program in c++ to do md5 checksums. I'm doing this mainly because I think I'll learn a lot of different things about c++, checksums, OOP, and whatever else I run into.

I'm having trouble the check sums and I think the problem is in the function padbuff which does the message padding.

#include "HashMD5.h"

int leftrotate(int x, int y);
void padbuff(uchar * buffer);

//HashMD5 constructor
HashMD5::HashMD5()
{
  Type = "md5";
  Hash = "";
}

HashMD5::HashMD5(const char * hashfile)
{
  Type = "md5";
  std::ifstream filestr;
  filestr.open(hashfile, std::fstream::in | std::fstream::binary);
  if(filestr.fail())
  {
    std::cerr << "File " << hashfile << " was not opened.\n";
    std::cerr << "Open failed with error ";
  }
}

std::string HashMD5::GetType()
{
  return this->Type;
}

std::string HashMD5::GetHash()
{
  return this->Hash;
}

bool HashMD5::is_open()
{
  return !((this->filestr).fail());
}

void HashMD5::CalcHash(unsigned int * hash)
{
  unsigned int *r, *k;
  int r2[4] = {0, 4, 9, 15};
  int r3[4] = {0, 7, 12, 19};
  int r4[4] = {0, 4, 9, 15};
  uchar * buffer;
  int bufLength = (2<<20)*8;
  int f,g,a,b,c,d, temp;
  int *head;
  uint32_t maxint = 1<<31;

  //Initialized states
  unsigned int h[4]{ 0x67452301, 0xefcdab89, 0x98badcfe, 0x10325476};

  r = new unsigned int[64];
  k = new unsigned int[64];
  buffer = new uchar[bufLength];
  if(r==NULL || k==NULL || buffer==NULL)
  {
    std::cerr << "One of the dyn alloc failed\n";
  }

  // r specifies the per-round shift amounts
  for(int i = 0; i<16; i++)
    r[i] = 7 + (5 * ((i)%4) );

  for(int i = 16; i < 32; i++)
    r[i] = 5 + r2[i%4];

  for(int i = 32; i< 48; i++)
    r[i] = 4 + r3[i%4];

  for(int i = 48; i < 63; i++)
    r[i] = 6 + r4[i%4];

  for(int i = 0; i < 63; i++)
  {
    k[i] = floor( fabs( sin(i + 1)) * maxint);
  }

  while(!(this->filestr).eof())
  {
    //Read in 512 bits
    (this->filestr).read((char *)buffer, bufLength-512);

    padbuff(buffer);

    //The 512 bits are now 16 32-bit ints
    head = (int *)buffer;

    for(int i = 0; i < 64; i++)
    {
      if(i >=0 && i <=15)
      {
        f = (b & c) | (~b & d);
        g = i;
      }
      else if(i >= 16 && i <=31)
      {
        f = (d & b) | (~d & b);
        g = (5*i +1) % 16;
      }
      else if(i >=32 && i<=47)
      {
        f = b ^ c ^ d;
        g = (3*i + 5 ) % 16;
      }
      else
      {
        f = c ^ (b | ~d);
        g = (7*i) % 16;
      }

      temp = d;
      d = c;
      c = b;
      b = b + leftrotate((a + f + k[i] + head[g]), r[i]);
      a = temp;

    }

    h[0] +=a;
    h[1] +=b;
    h[2] +=c;
    h[3] +=d;         
  }


  delete[] r;
  delete[] k;

  hash = h;
}

int leftrotate(int x, int y)
{
   return(x<<y) | (x >> (32 -y));
}

void padbuff(uchar* buffer)
{
   int lack;
   int length = strlen((char *)buffer);
   uint64_t mes_size = length % UINT64_MAX;
   if((lack = (112 - (length % 128) ))>0)
   {
     *(buffer + length) = ('\0'+1 ) << 3;
     memset((buffer + length + 1),0x0,lack);  
     memcpy((void*)(buffer+112),(void *)&mes_size, 64);
   }

}

In my test program I run this on the an empty message. Thus length in padbuff is 0. Then when I do *(buffer + length) = ('\0'+1 ) << 3;, I'm trying to pad the message with a 1. In the Netbeans debugger I cast buffer as a uint64_t and it says buffer=8. I was trying to put a 1 bit in the most significant spot of buffer so my cast should have been UINT64_MAX. Its not, so I'm confused about how my padding code works. Can someone tell me what I'm doing and what I'm supposed to do in padbuff? Thanks, and I apologize for the long freaking question.

Just to be clear about what the padding is supposed to be doing, here is the padding excerpt from Wikipedia: The message is padded so that its length is divisible by 512. The padding works as follows: first a single bit, 1, is appended to the end of the message. This is followed by as many zeros as are required to bring the length of the message up to 64 bits fewer than a multiple of 512. The remaining bits are filled up with 64 bits representing the length of the original message, modulo 264.

I'm mainly looking for help for padbuff, but since I'm trying to learn all comments are appreciated.

Answer 1

The first question is what you did:

1. length % UINT64_MAX doesn't make sense at all because length is in bytes and MAX is the value you can store in UINT64.
2. You thought that putting 1 bit in the most significant bit would give the maximum value. In fact, you need to put 1 in all bits to get it.
3. You shift 1 by 3. It's only half the length of the byte.
4. The byte pointed by buffer is the least significant in little endian. (I assume you have little endian since the debugger showed 8).

The second question how it should work. I don't know what exactly padbuff should do but if you want to pad and get UINT64_MAX, you need something like this:

int length = strlen((char *)buffer);
int len_of_padding = sizeof(uint64_t) - length % sizeof(uint64_t);
if(len_of_padding > 0)
{
  memset((void*)(buffer + length), 0xFF, len_of_padding);
}

You worked with the length of two uint64 values. May be you wanted to zero the next one:

uint64_t *after = (uint64_t*)(buffer + length + len_of_padding);
*after = 0;