bitswap.h

Go to the documentation of this file.

#pragma once
 
#ifndef __INC_BITSWAP_H
#define __INC_BITSWAP_H
 
#include "FastLED.h"
#include "fl/force_inline.h"
#include "fl/int.h"

 
FASTLED_NAMESPACE_BEGIN
 
 
#if defined(FASTLED_ARM) || defined(FASTLED_ESP8266) || defined(FASTLED_DOXYGEN)
typedef union {
  fl::u8 raw;   
  struct {
  fl::u32 a0:1;  
  fl::u32 a1:1;  
  fl::u32 a2:1;  
  fl::u32 a3:1;  
  fl::u32 a4:1;  
  fl::u32 a5:1;  
  fl::u32 a6:1;  
  fl::u32 a7:1;  
  };
} just8bits;

typedef struct {
  fl::u32 a0:1;  
  fl::u32 a1:1;  
  fl::u32 a2:1;  
  fl::u32 a3:1;  
  fl::u32 a4:1;  
  fl::u32 a5:1;  
  fl::u32 a6:1;  
  fl::u32 a7:1;  
  fl::u32 b0:1;  
  fl::u32 b1:1;  
  fl::u32 b2:1;  
  fl::u32 b3:1;  
  fl::u32 b4:1;  
  fl::u32 b5:1;  
  fl::u32 b6:1;  
  fl::u32 b7:1;  
  fl::u32 c0:1;  
  fl::u32 c1:1;  
  fl::u32 c2:1;  
  fl::u32 c3:1;  
  fl::u32 c4:1;  
  fl::u32 c5:1;  
  fl::u32 c6:1;  
  fl::u32 c7:1;  
  fl::u32 d0:1;  
  fl::u32 d1:1;  
  fl::u32 d2:1;  
  fl::u32 d3:1;  
  fl::u32 d4:1;  
  fl::u32 d5:1;  
  fl::u32 d6:1;  
  fl::u32 d7:1;  
} sub4;

typedef union {
  fl::u32 word[2];  
  fl::u8 bytes[8];  
  struct {
    sub4 a;  
    sub4 b;  
  };
} bitswap_type;
 

#define SWAPSA(X,N) out.  X ## 0 = in.a.a ## N; \
  out.  X ## 1 = in.a.b ## N; \
  out.  X ## 2 = in.a.c ## N; \
  out.  X ## 3 = in.a.d ## N;

#define SWAPSB(X,N) out.  X ## 0 = in.b.a ## N; \
  out.  X ## 1 = in.b.b ## N; \
  out.  X ## 2 = in.b.c ## N; \
  out.  X ## 3 = in.b.d ## N;

#define SWAPS(X,N) out.  X ## 0 = in.a.a ## N; \
  out.  X ## 1 = in.a.b ## N; \
  out.  X ## 2 = in.a.c ## N; \
  out.  X ## 3 = in.a.d ## N; \
  out.  X ## 4 = in.b.a ## N; \
  out.  X ## 5 = in.b.b ## N; \
  out.  X ## 6 = in.b.c ## N; \
  out.  X ## 7 = in.b.d ## N;
 

FASTLED_FORCE_INLINE void swapbits8(bitswap_type in, bitswap_type & out) {
 
  // SWAPS(a.a,7);
  // SWAPS(a.b,6);
  // SWAPS(a.c,5);
  // SWAPS(a.d,4);
  // SWAPS(b.a,3);
  // SWAPS(b.b,2);
  // SWAPS(b.c,1);
  // SWAPS(b.d,0);
 
  // SWAPSA(a.a,7);
  // SWAPSA(a.b,6);
  // SWAPSA(a.c,5);
  // SWAPSA(a.d,4);
  //
  // SWAPSB(a.a,7);
  // SWAPSB(a.b,6);
  // SWAPSB(a.c,5);
  // SWAPSB(a.d,4);
  //
  // SWAPSA(b.a,3);
  // SWAPSA(b.b,2);
  // SWAPSA(b.c,1);
  // SWAPSA(b.d,0);
  // //
  // SWAPSB(b.a,3);
  // SWAPSB(b.b,2);
  // SWAPSB(b.c,1);
  // SWAPSB(b.d,0);
 
  for(int i = 0; i < 8; ++i) {
    just8bits work;
    work.a3 = in.word[0] >> 31;
    work.a2 = in.word[0] >> 23;
    work.a1 = in.word[0] >> 15;
    work.a0 = in.word[0] >> 7;
    in.word[0] <<= 1;
    work.a7 = in.word[1] >> 31;
    work.a6 = in.word[1] >> 23;
    work.a5 = in.word[1] >> 15;
    work.a4 = in.word[1] >> 7;
    in.word[1] <<= 1;
    out.bytes[i] = work.raw;
  }
}

FASTLED_FORCE_INLINE void slowswap(unsigned char *A, unsigned char *B) {
 
  for(int row = 0; row < 7; ++row) {
    fl::u8 x = A[row];
 
    fl::u8 bit = (1<<row);
    unsigned char *p = B;
    for(fl::u32 mask = 1<<7 ; mask ; mask >>= 1) {
      if(x & mask) {
        *p++ |= bit;
      } else {
        *p++ &= ~bit;
      }
    }
    // B[7] |= (x & 0x01) << row; x >>= 1;
    // B[6] |= (x & 0x01) << row; x >>= 1;
    // B[5] |= (x & 0x01) << row; x >>= 1;
    // B[4] |= (x & 0x01) << row; x >>= 1;
    // B[3] |= (x & 0x01) << row; x >>= 1;
    // B[2] |= (x & 0x01) << row; x >>= 1;
    // B[1] |= (x & 0x01) << row; x >>= 1;
    // B[0] |= (x & 0x01) << row; x >>= 1;
  }
}

void transpose8x1_noinline(unsigned char *A, unsigned char *B);

FASTLED_FORCE_INLINE void transpose8x1(unsigned char *A, unsigned char *B) {
  fl::u32 x, y, t;
 
  // Load the array and pack it into x and y.
  y = *(unsigned int*)(A);
  x = *(unsigned int*)(A+4);
 
  // pre-transform x
  t = (x ^ (x >> 7)) & 0x00AA00AA;  x = x ^ t ^ (t << 7);
  t = (x ^ (x >>14)) & 0x0000CCCC;  x = x ^ t ^ (t <<14);
 
  // pre-transform y
  t = (y ^ (y >> 7)) & 0x00AA00AA;  y = y ^ t ^ (t << 7);
  t = (y ^ (y >>14)) & 0x0000CCCC;  y = y ^ t ^ (t <<14);
 
  // final transform
  t = (x & 0xF0F0F0F0) | ((y >> 4) & 0x0F0F0F0F);
  y = ((x << 4) & 0xF0F0F0F0) | (y & 0x0F0F0F0F);
  x = t;
 
  *((uint32_t*)B) = y;
  *((uint32_t*)(B+4)) = x;
}

FASTLED_FORCE_INLINE void transpose8x1_MSB(unsigned char *A, unsigned char *B) {
  fl::u32 x, y, t;
 
  // Load the array and pack it into x and y.
  y = *(unsigned int*)(A);
  x = *(unsigned int*)(A+4);
 
  // pre-transform x
  t = (x ^ (x >> 7)) & 0x00AA00AA;  x = x ^ t ^ (t << 7);
  t = (x ^ (x >>14)) & 0x0000CCCC;  x = x ^ t ^ (t <<14);
 
  // pre-transform y
  t = (y ^ (y >> 7)) & 0x00AA00AA;  y = y ^ t ^ (t << 7);
  t = (y ^ (y >>14)) & 0x0000CCCC;  y = y ^ t ^ (t <<14);
 
  // final transform
  t = (x & 0xF0F0F0F0) | ((y >> 4) & 0x0F0F0F0F);
  y = ((x << 4) & 0xF0F0F0F0) | (y & 0x0F0F0F0F);
  x = t;
 
  B[7] = y; y >>= 8;
  B[6] = y; y >>= 8;
  B[5] = y; y >>= 8;
  B[4] = y;
 
  B[3] = x; x >>= 8;
  B[2] = x; x >>= 8;
  B[1] = x; x >>= 8;
  B[0] = x; /* */
}

template<int m, int n>
FASTLED_FORCE_INLINE void transpose8(unsigned char *A, unsigned char *B) {
  fl::u32 x, y, t;
 
  // Load the array and pack it into x and y.
  if(m == 1) {
    y = *(unsigned int*)(A);
    x = *(unsigned int*)(A+4);
  } else {
    x = (A[0]<<24)   | (A[m]<<16)   | (A[2*m]<<8) | A[3*m];
    y = (A[4*m]<<24) | (A[5*m]<<16) | (A[6*m]<<8) | A[7*m];
  }
 
  // pre-transform x
  t = (x ^ (x >> 7)) & 0x00AA00AA;  x = x ^ t ^ (t << 7);
  t = (x ^ (x >>14)) & 0x0000CCCC;  x = x ^ t ^ (t <<14);
 
  // pre-transform y
  t = (y ^ (y >> 7)) & 0x00AA00AA;  y = y ^ t ^ (t << 7);
  t = (y ^ (y >>14)) & 0x0000CCCC;  y = y ^ t ^ (t <<14);
 
  // final transform
  t = (x & 0xF0F0F0F0) | ((y >> 4) & 0x0F0F0F0F);
  y = ((x << 4) & 0xF0F0F0F0) | (y & 0x0F0F0F0F);
  x = t;
 
  B[7*n] = y; y >>= 8;
  B[6*n] = y; y >>= 8;
  B[5*n] = y; y >>= 8;
  B[4*n] = y;
 
  B[3*n] = x; x >>= 8;
  B[2*n] = x; x >>= 8;
  B[n] = x; x >>= 8;
  B[0] = x;
  // B[0]=x>>24;    B[n]=x>>16;    B[2*n]=x>>8;  B[3*n]=x>>0;
  // B[4*n]=y>>24;  B[5*n]=y>>16;  B[6*n]=y>>8;  B[7*n]=y>>0;
}
 
#endif
 
FASTLED_NAMESPACE_END
 
#endif