* per call, ie, v_samp_factor block rows for each component in the image.
 * This amount of data is read from the source buffer, DCT'd and quantized,
 * and saved into the virtual arrays.  We also generate suitable dummy blocks
 * as needed at the right and lower edges.  (The dummy blocks are constructed
 * in the virtual arrays, which have been padded appropriately.)  This makes
 * it possible for subsequent passes not to worry about real vs. dummy blocks.
 *
 * We must also emit the data to the entropy encoder.  This is conveniently
 * done by calling compress_output() after we've loaded the current strip
 * of the virtual arrays.
 *
 * NB: input_buf contains a plane for each component in image.  All
 * components are DCT'd and loaded into the virtual arrays in this pass.
 * However, it may be that only a subset of the components are emitted to
 * the entropy encoder during this first pass; be careful about looking
 * at the scan-dependent variables (MCU dimensions, etc).
 */

METHODDEF(boolean)
compress_first_pass (j_compress_ptr cinfo, JSAMPIMAGE input_buf)
{
  my_coef_ptr coef = (my_coef_ptr) cinfo->coef;
  JDIMENSION last_iMCU_row = cinfo->total_iMCU_rows - 1;
  JDIMENSION blocks_across, MCUs_across, MCUindex;
  int bi, ci, h_samp_factor, block_row, block_rows, ndummy;
  JCOEF lastDC;
  jpeg_component_info *compptr;
  JBLOCKARRAY buffer;
  JBLOCKROW thisblockrow, lastblockrow;

  for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components;
       ci++, compptr++) {
    /* Align the virtual buffer for this component. */
    buffer = (*cinfo->mem->access_virt_barray)
      ((j_common_ptr) cinfo, coef->whole_image[ci],
       coef->iMCU_row_num * compptr->v_samp_factor,
       (JDIMENSION) compptr->v_samp_factor, TRUE);
    /* Count non-dummy DCT block rows in this iMCU row. */
    if (coef->iMCU_row_num < last_iMCU_row)
      block_rows = compptr->v_samp_factor;
    else {
      /* NB: can't use last_row_height here, since may not be set! */
      block_rows = (int) (compptr->height_in_blocks % compptr->v_samp_factor);
      if (block_rows == 0) block_rows = compptr->v_samp_factor;
    }
    blocks_across = compptr->width_in_blocks;
    h_samp_factor = compptr->h_samp_factor;
    /* Count number of dummy blocks to be added at the right margin. */
    ndummy = (int) (blocks_across % h_samp_factor);
    if (ndummy > 0)
      ndummy = h_samp_factor - ndummy;
    /* Perform DCT for all non-dummy blocks in this iMCU row.  Each call
     * on forward_DCT processes a complete horizontal row of DCT blocks.
     */
    for (block_row = 0; block_row < block_rows; block_row++) {
      thisblockrow = buffer[block_row];
      (*cinfo->fdct->forward_DCT) (cinfo, compptr,
				   input_buf[ci], thisblockrow,
				   (JDIMENSION) (block_row * DCTSIZE),
				   (JDIMENSION) 0, blocks_across);
      if (ndummy > 0) {
	/* Create dummy blocks at the right edge of the image. */
	thisblockrow += blocks_across; /* => first dummy block */
	jzero_far((void FAR *) thisblockrow, ndummy * SIZEOF(JBLOCK));
	lastDC = thisblockrow[-1][0];
	for (bi = 0; bi < ndummy; bi++) {
	  thisblockrow[bi][0] = lastDC;
	}
      }
    }
    /* If at end of image, create dummy block rows as needed.
     * The tricky part here is that within each MCU, we want the DC values
     * of the dummy blocks to match the last real block's DC value.
     * This squeezes a few more bytes out of the resulting file...
     */
    if (coef->iMCU_row_num == last_iMCU_row) {
      blocks_across += ndummy;	/* include lower right corner */
      MCUs_across = blocks_across / h_samp_factor;
      for (block_row = block_rows; block_row < compptr->v_samp_factor;
	   block_row++) {
	thisblockrow = buffer[block_row];
	lastblockrow = buffer[block_row-1];
	jzero_far((void FAR *) thisblockrow,
		  (size_t) (blocks_across * SIZEOF(JBLOCK)));
	for (MCUindex = 0; MCUindex < MCUs_across; MCUindex++) {
	  lastDC = lastblockrow[h_samp_factor-1][0];
	  for (bi = 0; bi < h_samp_factor; bi++) {
	    thisblockrow[bi][0] = lastDC;
	  }
	  thisblockrow += h_samp_factor; /* advance to next MCU in row */
	  lastblockrow += h_samp_factor;
	}
      }
    }
  }
  /* NB: compress_output will increment iMCU_row_num if successful.
   * A suspension return will result in redoing all the work above next time.
   */

  /* Emit data to the entropy encoder, sharing code with subsequent passes */
  return compress_output(cinfo, input_buf);
}


/*
 * Process some data in subsequent passes of a multi-pass case.
 * We process the equivalent of one fully interleaved MCU row ("iMCU" row)
 * per call, ie, v_samp_factor block rows for each component in the scan.
 * The data is obtained from the virtual arrays and fed to the entropy coder.
 * Returns TRUE if the iMCU row is completed, FALSE if suspended.
 *
 * NB: input_buf is ignored; it is likely to be a NULL pointer.
 */

METHODDEF(boolean)
compress_output (j_compress_ptr cinfo, JSAMPIMAGE input_buf)
{
  my_coef_ptr coef = (my_coef_ptr) cinfo->coef;
  JDIMENSION MCU_col_num;	/* index of current MCU within row */
  int blkn, ci, xindex, yindex, yoffset;
  JDIMENSION start_col;
  JBLOCKARRAY buffer[MAX_COMPS_IN_SCAN];
  JBLOCKROW buffer_ptr;
  jpeg_component_info *compptr;

  /* Align the virtual buffers for the components used in this scan.
   * NB: during first pass, this is safe only because the buffers will
   * already be aligned properly, so jmemmgr.c won't need to do any I/O.
   */
  for (ci = 0; ci < cinfo->comps_in_scan; ci++) {
    compptr = cinfo->cur_comp_info[ci];
    buffer[ci] = (*cinfo->mem->access_virt_barray)
      ((j_common_ptr) cinfo, coef->whole_image[compptr->component_index],
       coef->iMCU_row_num * compptr->v_samp_factor,
       (JDIMENSION) compptr->v_samp_factor, FALSE);
  }

  /* Loop to process one whole iMCU row */
  for (yoffset = coef->MCU_vert_offset; yoffset < coef->MCU_rows_per_iMCU_row;
       yoffset++) {
    for (MCU_col_num = coef->mcu_ctr; MCU_col_num < cinfo->MCUs_per_row;
	 MCU_col_num++) {
      /* Construct list of pointers to DCT blocks belonging to this MCU */
      blkn = 0;			/* index of current DCT block within MCU */
      for (ci = 0; ci < cinfo->comps_in_scan; ci++) {
	compptr = cinfo->cur_comp_info[ci];
	start_col = MCU_col_num * compptr->MCU_width;
	for (yindex = 0; yindex < compptr->MCU_height; yindex++) {
	  buffer_ptr = buffer[ci][yindex+yoffset] + start_col;
	  for (xindex = 0; xindex < compptr->MCU_width; xindex++) {
	    coef->MCU_buffer[blkn++] = buffer_ptr++;
	  }
	}
      }
      /* Try to write the MCU. */
      if (! (*cinfo->entropy->encode_mcu) (cinfo, coef->MCU_buffer)) {
	/* Suspension forced; update state counters and exit */
	coef->MCU_vert_offset = yoffset;
	coef->mcu_ctr = MCU_col_num;
	return FALSE;
      }
    }
    /* Completed an MCU row, but perhaps not an iMCU row */
    coef->mcu_ctr = 0;
  }
  /* Completed the iMCU row, advance counters for next one */
  coef->iMCU_row_num++;
  start_iMCU_row(cinfo);
  return TRUE;
}

#endif /* FULL_COEF_BUFFER_SUPPORTED */


/*
 * Initialize coefficient buffer controller.
 */

GLOBAL(void)
jinit_c_coef_controller (j_compress_ptr cinfo, boolean need_full_buffer)
{
  my_coef_ptr coef;

  coef = (my_coef_ptr)
    (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_IMAGE,
				SIZEOF(my_coef_controller));
  cinfo->coef = (struct jpeg_c_coef_controller *) coef;
  coef->pub.start_pass = start_pass_coef;

  /* Create the coefficient buffer. */
  if (need_full_buffer) {
#ifdef FULL_COEF_BUFFER_SUPPORTED
    /* Allocate a full-image virtual array for each component, */
    /* padded to a multiple of samp_factor DCT blocks in each direction. */
    int ci;
    jpeg_component_info *compptr;

    for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components;
	 ci++, compptr++) {
      coef->whole_image[ci] = (*cinfo->mem->request_virt_barray)
	((j_common_ptr) cinfo, JPOOL_IMAGE, FALSE,
	 (JDIMENSION) jround_up((long) compptr->width_in_blocks,
				(long) compptr->h_samp_factor),
	 (JDIMENSION) jround_up((long) compptr->height_in_blocks,
				(long) compptr->v_samp_factor),
	 (JDIMENSION) compptr->v_samp_factor);
    }
#else
    ERREXIT(cinfo, JERR_BAD_BUFFER_MODE);
#endif
  } else {
    /* We only need a single-MCU buffer. */
    JBLOCKROW buffer;
    int i;

    buffer = (JBLOCKROW)
      (*cinfo->mem->alloc_large) ((j_common_ptr) cinfo, JPOOL_IMAGE,
				  C_MAX_BLOCKS_IN_MCU * SIZEOF(JBLOCK));
    for (i = 0; i < C_MAX_BLOCKS_IN_MCU; i++) {
      coef->MCU_buffer[i] = buffer + i;
    }
    coef->whole_image[0] = NULL; /* flag for no virtual arrays */
  }
}