#include "vdp2.h"
#include "vdp1.h"

#include "vram.hpp"

void palette_data()
{
  vdp2.cram.u16[3] = rgb15( 0,  0,  0);
  vdp2.cram.u16[2] = rgb15(10, 10, 10);
  vdp2.cram.u16[1] = rgb15(21, 21, 21);
  vdp2.cram.u16[0] = rgb15(31, 31, 31);

  vdp2.cram.u16[16 + 3] = rgb15( 0,  0,  0);
  vdp2.cram.u16[16 + 2] = rgb15(21, 21, 21);
  vdp2.cram.u16[16 + 1] = rgb15(31, 31, 31);
  vdp2.cram.u16[16 + 0] = 0; // transparent
}

static void _2bpp_4bpp_vram_copy(uint32_t * vram, const start_size_t& buf)
{
  for (uint32_t ix = 0; ix < buf.size / 4; ix += 1) {
    const uint32_t pixels = reinterpret_cast<uint32_t const * const>(buf.start)[ix];
    const uint32_t px0 = pixels >> 16 & 0xffff;
    const uint32_t px1 = pixels >> 0  & 0xffff;

#define lshift(n) ((7 - n) * 2)
#define rshift(n) ((7 - n) * 4)
#define px(p, n) (((p >> lshift(n)) & 0b11) << rshift(n))
#define p0(n) (px(px0, n))
#define p1(n) (px(px1, n))
    vram[ix * 2 + 0] = p0(7) | p0(6) | p0(5) | p0(4) | p0(3) | p0(2) | p0(1) | p0(0);
    vram[ix * 2 + 1] = p1(7) | p1(6) | p1(5) | p1(4) | p1(3) | p1(2) | p1(1) | p1(0);
#undef p1
#undef p0
#undef px
#undef lshift
#undef rshift
  }
}

uint32_t character_pattern_table(const start_size_t& buf, const uint32_t top)
{
  // round to nearest multiple of 16
  const uint32_t table_size = ((buf.size * 2) + 0x10 - 1) & (-0x10);
  const uint32_t base_address = top - table_size;

  uint32_t * vram = &vdp1.vram.u32[(base_address / 4)];
  _2bpp_4bpp_vram_copy(vram, buf);

  return base_address;
}

uint32_t cell_data(const start_size_t& buf, const uint32_t top)
{
  // round to nearest multiple of 32
  const uint32_t table_size = ((buf.size * 2) + 0x20 - 1) & (-0x20);
  const uint32_t base_address = top - table_size; // in bytes

  uint32_t * vram = &vdp2.vram.u32[(base_address / 4)];
  _2bpp_4bpp_vram_copy(vram, buf);

  return base_address;
}