/*
 * Mesa 3-D graphics library
 *
 * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included
 * in all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
 * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
 * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
 * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
 * OTHER DEALINGS IN THE SOFTWARE.
 */


/**
 * \file mipmap.c  mipmap generation and teximage resizing functions.
 */

#include "imports.h"
#include "formats.h"
#include "glformats.h"
#include "mipmap.h"
#include "mtypes.h"
#include "teximage.h"
#include "texobj.h"
#include "texstore.h"
#include "image.h"
#include "macros.h"
#include "util/half_float.h"
#include "util/format_rgb9e5.h"
#include "util/format_r11g11b10f.h"



static GLint
bytes_per_pixel(GLenum datatype, GLuint comps)
{
   GLint b;

   if (datatype == GL_UNSIGNED_INT_8_24_REV_MESA ||
       datatype == GL_UNSIGNED_INT_24_8_MESA)
      return 4;

   b = _mesa_sizeof_packed_type(datatype);
   assert(b >= 0);

   if (_mesa_type_is_packed(datatype))
      return b;
   else
      return b * comps;
}


/**
 * \name Support macros for do_row and do_row_3d
 *
 * The macro madness is here for two reasons.  First, it compacts the code
 * slightly.  Second, it makes it much easier to adjust the specifics of the
 * filter to tune the rounding characteristics.
 */
/*@{*/
#define DECLARE_ROW_POINTERS(t, e) \
      const t(*rowA)[e] = (const t(*)[e]) srcRowA; \
      const t(*rowB)[e] = (const t(*)[e]) srcRowB; \
      const t(*rowC)[e] = (const t(*)[e]) srcRowC; \
      const t(*rowD)[e] = (const t(*)[e]) srcRowD; \
      t(*dst)[e] = (t(*)[e]) dstRow

#define DECLARE_ROW_POINTERS0(t) \
      const t *rowA = (const t *) srcRowA; \
      const t *rowB = (const t *) srcRowB; \
      const t *rowC = (const t *) srcRowC; \
      const t *rowD = (const t *) srcRowD; \
      t *dst = (t *) dstRow

#define FILTER_SUM_3D(Aj, Ak, Bj, Bk, Cj, Ck, Dj, Dk) \
   ((unsigned) Aj + (unsigned) Ak \
    + (unsigned) Bj + (unsigned) Bk \
    + (unsigned) Cj + (unsigned) Ck \
    + (unsigned) Dj + (unsigned) Dk \
    + 4) >> 3

#define FILTER_3D(e) \
   do { \
      dst[i][e] = FILTER_SUM_3D(rowA[j][e], rowA[k][e], \
                                rowB[j][e], rowB[k][e], \
                                rowC[j][e], rowC[k][e], \
                                rowD[j][e], rowD[k][e]); \
   } while(0)

#define FILTER_SUM_3D_SIGNED(Aj, Ak, Bj, Bk, Cj, Ck, Dj, Dk) \
   (Aj + Ak \
    + Bj + Bk \
    + Cj + Ck \
    + Dj + Dk \
    + 4) / 8

#define FILTER_3D_SIGNED(e) \
   do { \
      dst[i][e] = FILTER_SUM_3D_SIGNED(rowA[j][e], rowA[k][e], \
                                       rowB[j][e], rowB[k][e], \
                                       rowC[j][e], rowC[k][e], \
                                       rowD[j][e], rowD[k][e]); \
   } while(0)

#define FILTER_F_3D(e) \
   do { \
      dst[i][e] = (rowA[j][e] + rowA[k][e] \
                   + rowB[j][e] + rowB[k][e] \
                   + rowC[j][e] + rowC[k][e] \
                   + rowD[j][e] + rowD[k][e]) * 0.125F; \
   } while(0)

#define FILTER_HF_3D(e) \
   do { \
      const GLfloat aj = _mesa_half_to_float(rowA[j][e]); \
      const GLfloat ak = _mesa_half_to_float(rowA[k][e]); \
      const GLfloat bj = _mesa_half_to_float(rowB[j][e]); \
      const GLfloat bk = _mesa_half_to_float(rowB[k][e]); \
      const GLfloat cj = _mesa_half_to_float(rowC[j][e]); \
      const GLfloat ck = _mesa_half_to_float(rowC[k][e]); \
      const GLfloat dj = _mesa_half_to_float(rowD[j][e]); \
      const GLfloat dk = _mesa_half_to_float(rowD[k][e]); \
      dst[i][e] = _mesa_float_to_half((aj + ak + bj + bk + cj + ck + dj + dk) \
                                      * 0.125F); \
   } while(0)
/*@}*/


/**
 * Average together two rows of a source image to produce a single new
 * row in the dest image.  It's legal for the two source rows to point
 * to the same data.  The source width must be equal to either the
 * dest width or two times the dest width.
 * \param datatype  GL_UNSIGNED_BYTE, GL_UNSIGNED_SHORT, GL_FLOAT, etc.
 * \param comps  number of components per pixel (1..4)
 */
static void
do_row(GLenum datatype, GLuint comps, GLint srcWidth,
       const GLvoid *srcRowA, const GLvoid *srcRowB,
       GLint dstWidth, GLvoid *dstRow)
{
   const GLuint k0 = (srcWidth == dstWidth) ? 0 : 1;
   const GLuint colStride = (srcWidth == dstWidth) ? 1 : 2;

   assert(comps >= 1);
   assert(comps <= 4);

   /* This assertion is no longer valid with non-power-of-2 textures
   assert(srcWidth == dstWidth || srcWidth == 2 * dstWidth);
   */

   if (datatype == GL_UNSIGNED_BYTE && comps == 4) {
      GLuint i, j, k;
      const GLubyte(*rowA)[4] = (const GLubyte(*)[4]) srcRowA;
      const GLubyte(*rowB)[4] = (const GLubyte(*)[4]) srcRowB;
      GLubyte(*dst)[4] = (GLubyte(*)[4]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i][0] = (rowA[j][0] + rowA[k][0] + rowB[j][0] + rowB[k][0]) / 4;
         dst[i][1] = (rowA[j][1] + rowA[k][1] + rowB[j][1] + rowB[k][1]) / 4;
         dst[i][2] = (rowA[j][2] + rowA[k][2] + rowB[j][2] + rowB[k][2]) / 4;
         dst[i][3] = (rowA[j][3] + rowA[k][3] + rowB[j][3] + rowB[k][3]) / 4;
      }
   }
   else if (datatype == GL_UNSIGNED_BYTE && comps == 3) {
      GLuint i, j, k;
      const GLubyte(*rowA)[3] = (const GLubyte(*)[3]) srcRowA;
      const GLubyte(*rowB)[3] = (const GLubyte(*)[3]) srcRowB;
      GLubyte(*dst)[3] = (GLubyte(*)[3]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i][0] = (rowA[j][0] + rowA[k][0] + rowB[j][0] + rowB[k][0]) / 4;
         dst[i][1] = (rowA[j][1] + rowA[k][1] + rowB[j][1] + rowB[k][1]) / 4;
         dst[i][2] = (rowA[j][2] + rowA[k][2] + rowB[j][2] + rowB[k][2]) / 4;
      }
   }
   else if (datatype == GL_UNSIGNED_BYTE && comps == 2) {
      GLuint i, j, k;
      const GLubyte(*rowA)[2] = (const GLubyte(*)[2]) srcRowA;
      const GLubyte(*rowB)[2] = (const GLubyte(*)[2]) srcRowB;
      GLubyte(*dst)[2] = (GLubyte(*)[2]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i][0] = (rowA[j][0] + rowA[k][0] + rowB[j][0] + rowB[k][0]) >> 2;
         dst[i][1] = (rowA[j][1] + rowA[k][1] + rowB[j][1] + rowB[k][1]) >> 2;
      }
   }
   else if (datatype == GL_UNSIGNED_BYTE && comps == 1) {
      GLuint i, j, k;
      const GLubyte *rowA = (const GLubyte *) srcRowA;
      const GLubyte *rowB = (const GLubyte *) srcRowB;
      GLubyte *dst = (GLubyte *) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i] = (rowA[j] + rowA[k] + rowB[j] + rowB[k]) >> 2;
      }
   }

   else if (datatype == GL_BYTE && comps == 4) {
      GLuint i, j, k;
      const GLbyte(*rowA)[4] = (const GLbyte(*)[4]) srcRowA;
      const GLbyte(*rowB)[4] = (const GLbyte(*)[4]) srcRowB;
      GLbyte(*dst)[4] = (GLbyte(*)[4]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i][0] = (rowA[j][0] + rowA[k][0] + rowB[j][0] + rowB[k][0]) / 4;
         dst[i][1] = (rowA[j][1] + rowA[k][1] + rowB[j][1] + rowB[k][1]) / 4;
         dst[i][2] = (rowA[j][2] + rowA[k][2] + rowB[j][2] + rowB[k][2]) / 4;
         dst[i][3] = (rowA[j][3] + rowA[k][3] + rowB[j][3] + rowB[k][3]) / 4;
      }
   }
   else if (datatype == GL_BYTE && comps == 3) {
      GLuint i, j, k;
      const GLbyte(*rowA)[3] = (const GLbyte(*)[3]) srcRowA;
      const GLbyte(*rowB)[3] = (const GLbyte(*)[3]) srcRowB;
      GLbyte(*dst)[3] = (GLbyte(*)[3]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i][0] = (rowA[j][0] + rowA[k][0] + rowB[j][0] + rowB[k][0]) / 4;
         dst[i][1] = (rowA[j][1] + rowA[k][1] + rowB[j][1] + rowB[k][1]) / 4;
         dst[i][2] = (rowA[j][2] + rowA[k][2] + rowB[j][2] + rowB[k][2]) / 4;
      }
   }
   else if (datatype == GL_BYTE && comps == 2) {
      GLuint i, j, k;
      const GLbyte(*rowA)[2] = (const GLbyte(*)[2]) srcRowA;
      const GLbyte(*rowB)[2] = (const GLbyte(*)[2]) srcRowB;
      GLbyte(*dst)[2] = (GLbyte(*)[2]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i][0] = (rowA[j][0] + rowA[k][0] + rowB[j][0] + rowB[k][0]) / 4;
         dst[i][1] = (rowA[j][1] + rowA[k][1] + rowB[j][1] + rowB[k][1]) / 4;
      }
   }
   else if (datatype == GL_BYTE && comps == 1) {
      GLuint i, j, k;
      const GLbyte *rowA = (const GLbyte *) srcRowA;
      const GLbyte *rowB = (const GLbyte *) srcRowB;
      GLbyte *dst = (GLbyte *) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i] = (rowA[j] + rowA[k] + rowB[j] + rowB[k]) / 4;
      }
   }

   else if (datatype == GL_UNSIGNED_SHORT && comps == 4) {
      GLuint i, j, k;
      const GLushort(*rowA)[4] = (const GLushort(*)[4]) srcRowA;
      const GLushort(*rowB)[4] = (const GLushort(*)[4]) srcRowB;
      GLushort(*dst)[4] = (GLushort(*)[4]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i][0] = (rowA[j][0] + rowA[k][0] + rowB[j][0] + rowB[k][0]) / 4;
         dst[i][1] = (rowA[j][1] + rowA[k][1] + rowB[j][1] + rowB[k][1]) / 4;
         dst[i][2] = (rowA[j][2] + rowA[k][2] + rowB[j][2] + rowB[k][2]) / 4;
         dst[i][3] = (rowA[j][3] + rowA[k][3] + rowB[j][3] + rowB[k][3]) / 4;
      }
   }
   else if (datatype == GL_UNSIGNED_SHORT && comps == 3) {
      GLuint i, j, k;
      const GLushort(*rowA)[3] = (const GLushort(*)[3]) srcRowA;
      const GLushort(*rowB)[3] = (const GLushort(*)[3]) srcRowB;
      GLushort(*dst)[3] = (GLushort(*)[3]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i][0] = (rowA[j][0] + rowA[k][0] + rowB[j][0] + rowB[k][0]) / 4;
         dst[i][1] = (rowA[j][1] + rowA[k][1] + rowB[j][1] + rowB[k][1]) / 4;
         dst[i][2] = (rowA[j][2] + rowA[k][2] + rowB[j][2] + rowB[k][2]) / 4;
      }
   }
   else if (datatype == GL_UNSIGNED_SHORT && comps == 2) {
      GLuint i, j, k;
      const GLushort(*rowA)[2] = (const GLushort(*)[2]) srcRowA;
      const GLushort(*rowB)[2] = (const GLushort(*)[2]) srcRowB;
      GLushort(*dst)[2] = (GLushort(*)[2]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i][0] = (rowA[j][0] + rowA[k][0] + rowB[j][0] + rowB[k][0]) / 4;
         dst[i][1] = (rowA[j][1] + rowA[k][1] + rowB[j][1] + rowB[k][1]) / 4;
      }
   }
   else if (datatype == GL_UNSIGNED_SHORT && comps == 1) {
      GLuint i, j, k;
      const GLushort *rowA = (const GLushort *) srcRowA;
      const GLushort *rowB = (const GLushort *) srcRowB;
      GLushort *dst = (GLushort *) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i] = (rowA[j] + rowA[k] + rowB[j] + rowB[k]) / 4;
      }
   }

   else if (datatype == GL_SHORT && comps == 4) {
      GLuint i, j, k;
      const GLshort(*rowA)[4] = (const GLshort(*)[4]) srcRowA;
      const GLshort(*rowB)[4] = (const GLshort(*)[4]) srcRowB;
      GLshort(*dst)[4] = (GLshort(*)[4]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i][0] = (rowA[j][0] + rowA[k][0] + rowB[j][0] + rowB[k][0]) / 4;
         dst[i][1] = (rowA[j][1] + rowA[k][1] + rowB[j][1] + rowB[k][1]) / 4;
         dst[i][2] = (rowA[j][2] + rowA[k][2] + rowB[j][2] + rowB[k][2]) / 4;
         dst[i][3] = (rowA[j][3] + rowA[k][3] + rowB[j][3] + rowB[k][3]) / 4;
      }
   }
   else if (datatype == GL_SHORT && comps == 3) {
      GLuint i, j, k;
      const GLshort(*rowA)[3] = (const GLshort(*)[3]) srcRowA;
      const GLshort(*rowB)[3] = (const GLshort(*)[3]) srcRowB;
      GLshort(*dst)[3] = (GLshort(*)[3]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i][0] = (rowA[j][0] + rowA[k][0] + rowB[j][0] + rowB[k][0]) / 4;
         dst[i][1] = (rowA[j][1] + rowA[k][1] + rowB[j][1] + rowB[k][1]) / 4;
         dst[i][2] = (rowA[j][2] + rowA[k][2] + rowB[j][2] + rowB[k][2]) / 4;
      }
   }
   else if (datatype == GL_SHORT && comps == 2) {
      GLuint i, j, k;
      const GLshort(*rowA)[2] = (const GLshort(*)[2]) srcRowA;
      const GLshort(*rowB)[2] = (const GLshort(*)[2]) srcRowB;
      GLshort(*dst)[2] = (GLshort(*)[2]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i][0] = (rowA[j][0] + rowA[k][0] + rowB[j][0] + rowB[k][0]) / 4;
         dst[i][1] = (rowA[j][1] + rowA[k][1] + rowB[j][1] + rowB[k][1]) / 4;
      }
   }
   else if (datatype == GL_SHORT && comps == 1) {
      GLuint i, j, k;
      const GLshort *rowA = (const GLshort *) srcRowA;
      const GLshort *rowB = (const GLshort *) srcRowB;
      GLshort *dst = (GLshort *) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i] = (rowA[j] + rowA[k] + rowB[j] + rowB[k]) / 4;
      }
   }

   else if (datatype == GL_FLOAT && comps == 4) {
      GLuint i, j, k;
      const GLfloat(*rowA)[4] = (const GLfloat(*)[4]) srcRowA;
      const GLfloat(*rowB)[4] = (const GLfloat(*)[4]) srcRowB;
      GLfloat(*dst)[4] = (GLfloat(*)[4]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i][0] = (rowA[j][0] + rowA[k][0] +
                      rowB[j][0] + rowB[k][0]) * 0.25F;
         dst[i][1] = (rowA[j][1] + rowA[k][1] +
                      rowB[j][1] + rowB[k][1]) * 0.25F;
         dst[i][2] = (rowA[j][2] + rowA[k][2] +
                      rowB[j][2] + rowB[k][2]) * 0.25F;
         dst[i][3] = (rowA[j][3] + rowA[k][3] +
                      rowB[j][3] + rowB[k][3]) * 0.25F;
      }
   }
   else if (datatype == GL_FLOAT && comps == 3) {
      GLuint i, j, k;
      const GLfloat(*rowA)[3] = (const GLfloat(*)[3]) srcRowA;
      const GLfloat(*rowB)[3] = (const GLfloat(*)[3]) srcRowB;
      GLfloat(*dst)[3] = (GLfloat(*)[3]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i][0] = (rowA[j][0] + rowA[k][0] +
                      rowB[j][0] + rowB[k][0]) * 0.25F;
         dst[i][1] = (rowA[j][1] + rowA[k][1] +
                      rowB[j][1] + rowB[k][1]) * 0.25F;
         dst[i][2] = (rowA[j][2] + rowA[k][2] +
                      rowB[j][2] + rowB[k][2]) * 0.25F;
      }
   }
   else if (datatype == GL_FLOAT && comps == 2) {
      GLuint i, j, k;
      const GLfloat(*rowA)[2] = (const GLfloat(*)[2]) srcRowA;
      const GLfloat(*rowB)[2] = (const GLfloat(*)[2]) srcRowB;
      GLfloat(*dst)[2] = (GLfloat(*)[2]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i][0] = (rowA[j][0] + rowA[k][0] +
                      rowB[j][0] + rowB[k][0]) * 0.25F;
         dst[i][1] = (rowA[j][1] + rowA[k][1] +
                      rowB[j][1] + rowB[k][1]) * 0.25F;
      }
   }
   else if (datatype == GL_FLOAT && comps == 1) {
      GLuint i, j, k;
      const GLfloat *rowA = (const GLfloat *) srcRowA;
      const GLfloat *rowB = (const GLfloat *) srcRowB;
      GLfloat *dst = (GLfloat *) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i] = (rowA[j] + rowA[k] + rowB[j] + rowB[k]) * 0.25F;
      }
   }

   else if (datatype == GL_HALF_FLOAT_ARB && comps == 4) {
      GLuint i, j, k, comp;
      const GLhalfARB(*rowA)[4] = (const GLhalfARB(*)[4]) srcRowA;
      const GLhalfARB(*rowB)[4] = (const GLhalfARB(*)[4]) srcRowB;
      GLhalfARB(*dst)[4] = (GLhalfARB(*)[4]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         for (comp = 0; comp < 4; comp++) {
            GLfloat aj, ak, bj, bk;
            aj = _mesa_half_to_float(rowA[j][comp]);
            ak = _mesa_half_to_float(rowA[k][comp]);
            bj = _mesa_half_to_float(rowB[j][comp]);
            bk = _mesa_half_to_float(rowB[k][comp]);
            dst[i][comp] = _mesa_float_to_half((aj + ak + bj + bk) * 0.25F);
         }
      }
   }
   else if (datatype == GL_HALF_FLOAT_ARB && comps == 3) {
      GLuint i, j, k, comp;
      const GLhalfARB(*rowA)[3] = (const GLhalfARB(*)[3]) srcRowA;
      const GLhalfARB(*rowB)[3] = (const GLhalfARB(*)[3]) srcRowB;
      GLhalfARB(*dst)[3] = (GLhalfARB(*)[3]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         for (comp = 0; comp < 3; comp++) {
            GLfloat aj, ak, bj, bk;
            aj = _mesa_half_to_float(rowA[j][comp]);
            ak = _mesa_half_to_float(rowA[k][comp]);
            bj = _mesa_half_to_float(rowB[j][comp]);
            bk = _mesa_half_to_float(rowB[k][comp]);
            dst[i][comp] = _mesa_float_to_half((aj + ak + bj + bk) * 0.25F);
         }
      }
   }
   else if (datatype == GL_HALF_FLOAT_ARB && comps == 2) {
      GLuint i, j, k, comp;
      const GLhalfARB(*rowA)[2] = (const GLhalfARB(*)[2]) srcRowA;
      const GLhalfARB(*rowB)[2] = (const GLhalfARB(*)[2]) srcRowB;
      GLhalfARB(*dst)[2] = (GLhalfARB(*)[2]) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         for (comp = 0; comp < 2; comp++) {
            GLfloat aj, ak, bj, bk;
            aj = _mesa_half_to_float(rowA[j][comp]);
            ak = _mesa_half_to_float(rowA[k][comp]);
            bj = _mesa_half_to_float(rowB[j][comp]);
            bk = _mesa_half_to_float(rowB[k][comp]);
            dst[i][comp] = _mesa_float_to_half((aj + ak + bj + bk) * 0.25F);
         }
      }
   }
   else if (datatype == GL_HALF_FLOAT_ARB && comps == 1) {
      GLuint i, j, k;
      const GLhalfARB *rowA = (const GLhalfARB *) srcRowA;
      const GLhalfARB *rowB = (const GLhalfARB *) srcRowB;
      GLhalfARB *dst = (GLhalfARB *) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         GLfloat aj, ak, bj, bk;
         aj = _mesa_half_to_float(rowA[j]);
         ak = _mesa_half_to_float(rowA[k]);
         bj = _mesa_half_to_float(rowB[j]);
         bk = _mesa_half_to_float(rowB[k]);
         dst[i] = _mesa_float_to_half((aj + ak + bj + bk) * 0.25F);
      }
   }

   else if (datatype == GL_UNSIGNED_INT && comps == 1) {
      GLuint i, j, k;
      const GLuint *rowA = (const GLuint *) srcRowA;
      const GLuint *rowB = (const GLuint *) srcRowB;
      GLuint *dst = (GLuint *) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i] = rowA[j] / 4 + rowA[k] / 4 + rowB[j] / 4 + rowB[k] / 4;
      }
   }

   else if (datatype == GL_UNSIGNED_SHORT_5_6_5 && comps == 3) {
      GLuint i, j, k;
      const GLushort *rowA = (const GLushort *) srcRowA;
      const GLushort *rowB = (const GLushort *) srcRowB;
      GLushort *dst = (GLushort *) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         const GLint rowAr0 = rowA[j] & 0x1f;
         const GLint rowAr1 = rowA[k] & 0x1f;
         const GLint rowBr0 = rowB[j] & 0x1f;
         const GLint rowBr1 = rowB[k] & 0x1f;
         const GLint rowAg0 = (rowA[j] >> 5) & 0x3f;
         const GLint rowAg1 = (rowA[k] >> 5) & 0x3f;
         const GLint rowBg0 = (rowB[j] >> 5) & 0x3f;
         const GLint rowBg1 = (rowB[k] >> 5) & 0x3f;
         const GLint rowAb0 = (rowA[j] >> 11) & 0x1f;
         const GLint rowAb1 = (rowA[k] >> 11) & 0x1f;
         const GLint rowBb0 = (rowB[j] >> 11) & 0x1f;
         const GLint rowBb1 = (rowB[k] >> 11) & 0x1f;
         const GLint red = (rowAr0 + rowAr1 + rowBr0 + rowBr1) >> 2;
         const GLint green = (rowAg0 + rowAg1 + rowBg0 + rowBg1) >> 2;
         const GLint blue = (rowAb0 + rowAb1 + rowBb0 + rowBb1) >> 2;
         dst[i] = (blue << 11) | (green << 5) | red;
      }
   }
   else if (datatype == GL_UNSIGNED_SHORT_4_4_4_4 && comps == 4) {
      GLuint i, j, k;
      const GLushort *rowA = (const GLushort *) srcRowA;
      const GLushort *rowB = (const GLushort *) srcRowB;
      GLushort *dst = (GLushort *) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         const GLint rowAr0 = rowA[j] & 0xf;
         const GLint rowAr1 = rowA[k] & 0xf;
         const GLint rowBr0 = rowB[j] & 0xf;
         const GLint rowBr1 = rowB[k] & 0xf;
         const GLint rowAg0 = (rowA[j] >> 4) & 0xf;
         const GLint rowAg1 = (rowA[k] >> 4) & 0xf;
         const GLint rowBg0 = (rowB[j] >> 4) & 0xf;
         const GLint rowBg1 = (rowB[k] >> 4) & 0xf;
         const GLint rowAb0 = (rowA[j] >> 8) & 0xf;
         const GLint rowAb1 = (rowA[k] >> 8) & 0xf;
         const GLint rowBb0 = (rowB[j] >> 8) & 0xf;
         const GLint rowBb1 = (rowB[k] >> 8) & 0xf;
         const GLint rowAa0 = (rowA[j] >> 12) & 0xf;
         const GLint rowAa1 = (rowA[k] >> 12) & 0xf;
         const GLint rowBa0 = (rowB[j] >> 12) & 0xf;
         const GLint rowBa1 = (rowB[k] >> 12) & 0xf;
         const GLint red = (rowAr0 + rowAr1 + rowBr0 + rowBr1) >> 2;
         const GLint green = (rowAg0 + rowAg1 + rowBg0 + rowBg1) >> 2;
         const GLint blue = (rowAb0 + rowAb1 + rowBb0 + rowBb1) >> 2;
         const GLint alpha = (rowAa0 + rowAa1 + rowBa0 + rowBa1) >> 2;
         dst[i] = (alpha << 12) | (blue << 8) | (green << 4) | red;
      }
   }
   else if (datatype == GL_UNSIGNED_SHORT_1_5_5_5_REV && comps == 4) {
      GLuint i, j, k;
      const GLushort *rowA = (const GLushort *) srcRowA;
      const GLushort *rowB = (const GLushort *) srcRowB;
      GLushort *dst = (GLushort *) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         const GLint rowAr0 = rowA[j] & 0x1f;
         const GLint rowAr1 = rowA[k] & 0x1f;
         const GLint rowBr0 = rowB[j] & 0x1f;
         const GLint rowBr1 = rowB[k] & 0x1f;
         const GLint rowAg0 = (rowA[j] >> 5) & 0x1f;
         const GLint rowAg1 = (rowA[k] >> 5) & 0x1f;
         const GLint rowBg0 = (rowB[j] >> 5) & 0x1f;
         const GLint rowBg1 = (rowB[k] >> 5) & 0x1f;
         const GLint rowAb0 = (rowA[j] >> 10) & 0x1f;
         const GLint rowAb1 = (rowA[k] >> 10) & 0x1f;
         const GLint rowBb0 = (rowB[j] >> 10) & 0x1f;
         const GLint rowBb1 = (rowB[k] >> 10) & 0x1f;
         const GLint rowAa0 = (rowA[j] >> 15) & 0x1;
         const GLint rowAa1 = (rowA[k] >> 15) & 0x1;
         const GLint rowBa0 = (rowB[j] >> 15) & 0x1;
         const GLint rowBa1 = (rowB[k] >> 15) & 0x1;
         const GLint red = (rowAr0 + rowAr1 + rowBr0 + rowBr1) >> 2;
         const GLint green = (rowAg0 + rowAg1 + rowBg0 + rowBg1) >> 2;
         const GLint blue = (rowAb0 + rowAb1 + rowBb0 + rowBb1) >> 2;
         const GLint alpha = (rowAa0 + rowAa1 + rowBa0 + rowBa1) >> 2;
         dst[i] = (alpha << 15) | (blue << 10) | (green << 5) | red;
      }
   }
   else if (datatype == GL_UNSIGNED_SHORT_5_5_5_1 && comps == 4) {
      GLuint i, j, k;
      const GLushort *rowA = (const GLushort *) srcRowA;
      const GLushort *rowB = (const GLushort *) srcRowB;
      GLushort *dst = (GLushort *) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         const GLint rowAr0 = (rowA[j] >> 11) & 0x1f;
         const GLint rowAr1 = (rowA[k] >> 11) & 0x1f;
         const GLint rowBr0 = (rowB[j] >> 11) & 0x1f;
         const GLint rowBr1 = (rowB[k] >> 11) & 0x1f;
         const GLint rowAg0 = (rowA[j] >> 6) & 0x1f;
         const GLint rowAg1 = (rowA[k] >> 6) & 0x1f;
         const GLint rowBg0 = (rowB[j] >> 6) & 0x1f;
         const GLint rowBg1 = (rowB[k] >> 6) & 0x1f;
         const GLint rowAb0 = (rowA[j] >> 1) & 0x1f;
         const GLint rowAb1 = (rowA[k] >> 1) & 0x1f;
         const GLint rowBb0 = (rowB[j] >> 1) & 0x1f;
         const GLint rowBb1 = (rowB[k] >> 1) & 0x1f;
         const GLint rowAa0 = (rowA[j] & 0x1);
         const GLint rowAa1 = (rowA[k] & 0x1);
         const GLint rowBa0 = (rowB[j] & 0x1);
         const GLint rowBa1 = (rowB[k] & 0x1);
         const GLint red = (rowAr0 + rowAr1 + rowBr0 + rowBr1) >> 2;
         const GLint green = (rowAg0 + rowAg1 + rowBg0 + rowBg1) >> 2;
         const GLint blue = (rowAb0 + rowAb1 + rowBb0 + rowBb1) >> 2;
         const GLint alpha = (rowAa0 + rowAa1 + rowBa0 + rowBa1) >> 2;
         dst[i] = (red << 11) | (green << 6) | (blue << 1) | alpha;
      }
   }

   else if (datatype == GL_UNSIGNED_BYTE_3_3_2 && comps == 3) {
      GLuint i, j, k;
      const GLubyte *rowA = (const GLubyte *) srcRowA;
      const GLubyte *rowB = (const GLubyte *) srcRowB;
      GLubyte *dst = (GLubyte *) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         const GLint rowAr0 = rowA[j] & 0x3;
         const GLint rowAr1 = rowA[k] & 0x3;
         const GLint rowBr0 = rowB[j] & 0x3;
         const GLint rowBr1 = rowB[k] & 0x3;
         const GLint rowAg0 = (rowA[j] >> 2) & 0x7;
         const GLint rowAg1 = (rowA[k] >> 2) & 0x7;
         const GLint rowBg0 = (rowB[j] >> 2) & 0x7;
         const GLint rowBg1 = (rowB[k] >> 2) & 0x7;
         const GLint rowAb0 = (rowA[j] >> 5) & 0x7;
         const GLint rowAb1 = (rowA[k] >> 5) & 0x7;
         const GLint rowBb0 = (rowB[j] >> 5) & 0x7;
         const GLint rowBb1 = (rowB[k] >> 5) & 0x7;
         const GLint red = (rowAr0 + rowAr1 + rowBr0 + rowBr1) >> 2;
         const GLint green = (rowAg0 + rowAg1 + rowBg0 + rowBg1) >> 2;
         const GLint blue = (rowAb0 + rowAb1 + rowBb0 + rowBb1) >> 2;
         dst[i] = (blue << 5) | (green << 2) | red;
      }
   }

   else if (datatype == MESA_UNSIGNED_BYTE_4_4 && comps == 2) {
      GLuint i, j, k;
      const GLubyte *rowA = (const GLubyte *) srcRowA;
      const GLubyte *rowB = (const GLubyte *) srcRowB;
      GLubyte *dst = (GLubyte *) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         const GLint rowAr0 = rowA[j] & 0xf;
         const GLint rowAr1 = rowA[k] & 0xf;
         const GLint rowBr0 = rowB[j] & 0xf;
         const GLint rowBr1 = rowB[k] & 0xf;
         const GLint rowAg0 = (rowA[j] >> 4) & 0xf;
         const GLint rowAg1 = (rowA[k] >> 4) & 0xf;
         const GLint rowBg0 = (rowB[j] >> 4) & 0xf;
         const GLint rowBg1 = (rowB[k] >> 4) & 0xf;
         const GLint r = (rowAr0 + rowAr1 + rowBr0 + rowBr1) >> 2;
         const GLint g = (rowAg0 + rowAg1 + rowBg0 + rowBg1) >> 2;
         dst[i] = (g << 4) | r;
      }
   }

   else if (datatype == GL_UNSIGNED_INT_2_10_10_10_REV && comps == 4) {
      GLuint i, j, k;
      const GLuint *rowA = (const GLuint *) srcRowA;
      const GLuint *rowB = (const GLuint *) srcRowB;
      GLuint *dst = (GLuint *) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         const GLint rowAr0 = rowA[j] & 0x3ff;
         const GLint rowAr1 = rowA[k] & 0x3ff;
         const GLint rowBr0 = rowB[j] & 0x3ff;
         const GLint rowBr1 = rowB[k] & 0x3ff;
         const GLint rowAg0 = (rowA[j] >> 10) & 0x3ff;
         const GLint rowAg1 = (rowA[k] >> 10) & 0x3ff;
         const GLint rowBg0 = (rowB[j] >> 10) & 0x3ff;
         const GLint rowBg1 = (rowB[k] >> 10) & 0x3ff;
         const GLint rowAb0 = (rowA[j] >> 20) & 0x3ff;
         const GLint rowAb1 = (rowA[k] >> 20) & 0x3ff;
         const GLint rowBb0 = (rowB[j] >> 20) & 0x3ff;
         const GLint rowBb1 = (rowB[k] >> 20) & 0x3ff;
         const GLint rowAa0 = (rowA[j] >> 30) & 0x3;
         const GLint rowAa1 = (rowA[k] >> 30) & 0x3;
         const GLint rowBa0 = (rowB[j] >> 30) & 0x3;
         const GLint rowBa1 = (rowB[k] >> 30) & 0x3;
         const GLint red = (rowAr0 + rowAr1 + rowBr0 + rowBr1) >> 2;
         const GLint green = (rowAg0 + rowAg1 + rowBg0 + rowBg1) >> 2;
         const GLint blue = (rowAb0 + rowAb1 + rowBb0 + rowBb1) >> 2;
         const GLint alpha = (rowAa0 + rowAa1 + rowBa0 + rowBa1) >> 2;
         dst[i] = (alpha << 30) | (blue << 20) | (green << 10) | red;
      }
   }

   else if (datatype == GL_UNSIGNED_INT_5_9_9_9_REV && comps == 3) {
      GLuint i, j, k;
      const GLuint *rowA = (const GLuint*) srcRowA;
      const GLuint *rowB = (const GLuint*) srcRowB;
      GLuint *dst = (GLuint*)dstRow;
      GLfloat res[3], rowAj[3], rowBj[3], rowAk[3], rowBk[3];
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         rgb9e5_to_float3(rowA[j], rowAj);
         rgb9e5_to_float3(rowB[j], rowBj);
         rgb9e5_to_float3(rowA[k], rowAk);
         rgb9e5_to_float3(rowB[k], rowBk);
         res[0] = (rowAj[0] + rowAk[0] + rowBj[0] + rowBk[0]) * 0.25F;
         res[1] = (rowAj[1] + rowAk[1] + rowBj[1] + rowBk[1]) * 0.25F;
         res[2] = (rowAj[2] + rowAk[2] + rowBj[2] + rowBk[2]) * 0.25F;
         dst[i] = float3_to_rgb9e5(res);
      }
   }

   else if (datatype == GL_UNSIGNED_INT_10F_11F_11F_REV && comps == 3) {
      GLuint i, j, k;
      const GLuint *rowA = (const GLuint*) srcRowA;
      const GLuint *rowB = (const GLuint*) srcRowB;
      GLuint *dst = (GLuint*)dstRow;
      GLfloat res[3], rowAj[3], rowBj[3], rowAk[3], rowBk[3];
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         r11g11b10f_to_float3(rowA[j], rowAj);
         r11g11b10f_to_float3(rowB[j], rowBj);
         r11g11b10f_to_float3(rowA[k], rowAk);
         r11g11b10f_to_float3(rowB[k], rowBk);
         res[0] = (rowAj[0] + rowAk[0] + rowBj[0] + rowBk[0]) * 0.25F;
         res[1] = (rowAj[1] + rowAk[1] + rowBj[1] + rowBk[1]) * 0.25F;
         res[2] = (rowAj[2] + rowAk[2] + rowBj[2] + rowBk[2]) * 0.25F;
         dst[i] = float3_to_r11g11b10f(res);
      }
   }

   else if (datatype == GL_FLOAT_32_UNSIGNED_INT_24_8_REV && comps == 1) {
      GLuint i, j, k;
      const GLfloat *rowA = (const GLfloat *) srcRowA;
      const GLfloat *rowB = (const GLfloat *) srcRowB;
      GLfloat *dst = (GLfloat *) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         dst[i*2] = (rowA[j*2] + rowA[k*2] + rowB[j*2] + rowB[k*2]) * 0.25F;
      }
   }

   else if (datatype == GL_UNSIGNED_INT_24_8_MESA && comps == 2) {
      GLuint i, j, k;
      const GLuint *rowA = (const GLuint *) srcRowA;
      const GLuint *rowB = (const GLuint *) srcRowB;
      GLuint *dst = (GLuint *) dstRow;
      /* note: averaging stencil values seems weird, but what else? */
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         GLuint z = (((rowA[j] >> 8) + (rowA[k] >> 8) +
                      (rowB[j] >> 8) + (rowB[k] >> 8)) / 4) << 8;
         GLuint s = ((rowA[j] & 0xff) + (rowA[k] & 0xff) +
                     (rowB[j] & 0xff) + (rowB[k] & 0xff)) / 4;
         dst[i] = z | s;
      }
   }
   else if (datatype == GL_UNSIGNED_INT_8_24_REV_MESA && comps == 2) {
      GLuint i, j, k;
      const GLuint *rowA = (const GLuint *) srcRowA;
      const GLuint *rowB = (const GLuint *) srcRowB;
      GLuint *dst = (GLuint *) dstRow;
      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         GLuint z = ((rowA[j] & 0xffffff) + (rowA[k] & 0xffffff) +
                     (rowB[j] & 0xffffff) + (rowB[k] & 0xffffff)) / 4;
         GLuint s = (((rowA[j] >> 24) + (rowA[k] >> 24) +
                      (rowB[j] >> 24) + (rowB[k] >> 24)) / 4) << 24;
         dst[i] = z | s;
      }
   }

   else {
      _mesa_problem(NULL, "bad format in do_row()");
   }
}


/**
 * Average together four rows of a source image to produce a single new
 * row in the dest image.  It's legal for the two source rows to point
 * to the same data.  The source width must be equal to either the
 * dest width or two times the dest width.
 *
 * \param datatype  GL pixel type \c GL_UNSIGNED_BYTE, \c GL_UNSIGNED_SHORT,
 *                  \c GL_FLOAT, etc.
 * \param comps     number of components per pixel (1..4)
 * \param srcWidth  Width of a row in the source data
 * \param srcRowA   Pointer to one of the rows of source data
 * \param srcRowB   Pointer to one of the rows of source data
 * \param srcRowC   Pointer to one of the rows of source data
 * \param srcRowD   Pointer to one of the rows of source data
 * \param dstWidth  Width of a row in the destination data
 * \param srcRowA   Pointer to the row of destination data
 */
static void
do_row_3D(GLenum datatype, GLuint comps, GLint srcWidth,
          const GLvoid *srcRowA, const GLvoid *srcRowB,
          const GLvoid *srcRowC, const GLvoid *srcRowD,
          GLint dstWidth, GLvoid *dstRow)
{
   const GLuint k0 = (srcWidth == dstWidth) ? 0 : 1;
   const GLuint colStride = (srcWidth == dstWidth) ? 1 : 2;
   GLuint i, j, k;

   assert(comps >= 1);
   assert(comps <= 4);

   if ((datatype == GL_UNSIGNED_BYTE) && (comps == 4)) {
      DECLARE_ROW_POINTERS(GLubyte, 4);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_3D(0);
         FILTER_3D(1);
         FILTER_3D(2);
         FILTER_3D(3);
      }
   }
   else if ((datatype == GL_UNSIGNED_BYTE) && (comps == 3)) {
      DECLARE_ROW_POINTERS(GLubyte, 3);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_3D(0);
         FILTER_3D(1);
         FILTER_3D(2);
      }
   }
   else if ((datatype == GL_UNSIGNED_BYTE) && (comps == 2)) {
      DECLARE_ROW_POINTERS(GLubyte, 2);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_3D(0);
         FILTER_3D(1);
      }
   }
   else if ((datatype == GL_UNSIGNED_BYTE) && (comps == 1)) {
      DECLARE_ROW_POINTERS(GLubyte, 1);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_3D(0);
      }
   }
   else if ((datatype == GL_BYTE) && (comps == 4)) {
      DECLARE_ROW_POINTERS(GLbyte, 4);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_3D_SIGNED(0);
         FILTER_3D_SIGNED(1);
         FILTER_3D_SIGNED(2);
         FILTER_3D_SIGNED(3);
      }
   }
   else if ((datatype == GL_BYTE) && (comps == 3)) {
      DECLARE_ROW_POINTERS(GLbyte, 3);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_3D_SIGNED(0);
         FILTER_3D_SIGNED(1);
         FILTER_3D_SIGNED(2);
      }
   }
   else if ((datatype == GL_BYTE) && (comps == 2)) {
      DECLARE_ROW_POINTERS(GLbyte, 2);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_3D_SIGNED(0);
         FILTER_3D_SIGNED(1);
       }
   }
   else if ((datatype == GL_BYTE) && (comps == 1)) {
      DECLARE_ROW_POINTERS(GLbyte, 1);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_3D_SIGNED(0);
      }
   }
   else if ((datatype == GL_UNSIGNED_SHORT) && (comps == 4)) {
      DECLARE_ROW_POINTERS(GLushort, 4);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_3D(0);
         FILTER_3D(1);
         FILTER_3D(2);
         FILTER_3D(3);
      }
   }
   else if ((datatype == GL_UNSIGNED_SHORT) && (comps == 3)) {
      DECLARE_ROW_POINTERS(GLushort, 3);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_3D(0);
         FILTER_3D(1);
         FILTER_3D(2);
      }
   }
   else if ((datatype == GL_UNSIGNED_SHORT) && (comps == 2)) {
      DECLARE_ROW_POINTERS(GLushort, 2);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_3D(0);
         FILTER_3D(1);
      }
   }
   else if ((datatype == GL_UNSIGNED_SHORT) && (comps == 1)) {
      DECLARE_ROW_POINTERS(GLushort, 1);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_3D(0);
      }
   }
   else if ((datatype == GL_SHORT) && (comps == 4)) {
      DECLARE_ROW_POINTERS(GLshort, 4);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_3D(0);
         FILTER_3D(1);
         FILTER_3D(2);
         FILTER_3D(3);
      }
   }
   else if ((datatype == GL_SHORT) && (comps == 3)) {
      DECLARE_ROW_POINTERS(GLshort, 3);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_3D(0);
         FILTER_3D(1);
         FILTER_3D(2);
      }
   }
   else if ((datatype == GL_SHORT) && (comps == 2)) {
      DECLARE_ROW_POINTERS(GLshort, 2);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_3D(0);
         FILTER_3D(1);
      }
   }
   else if ((datatype == GL_SHORT) && (comps == 1)) {
      DECLARE_ROW_POINTERS(GLshort, 1);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_3D(0);
      }
   }
   else if ((datatype == GL_FLOAT) && (comps == 4)) {
      DECLARE_ROW_POINTERS(GLfloat, 4);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_F_3D(0);
         FILTER_F_3D(1);
         FILTER_F_3D(2);
         FILTER_F_3D(3);
      }
   }
   else if ((datatype == GL_FLOAT) && (comps == 3)) {
      DECLARE_ROW_POINTERS(GLfloat, 3);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_F_3D(0);
         FILTER_F_3D(1);
         FILTER_F_3D(2);
      }
   }
   else if ((datatype == GL_FLOAT) && (comps == 2)) {
      DECLARE_ROW_POINTERS(GLfloat, 2);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_F_3D(0);
         FILTER_F_3D(1);
      }
   }
   else if ((datatype == GL_FLOAT) && (comps == 1)) {
      DECLARE_ROW_POINTERS(GLfloat, 1);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_F_3D(0);
      }
   }
   else if ((datatype == GL_HALF_FLOAT_ARB) && (comps == 4)) {
      DECLARE_ROW_POINTERS(GLhalfARB, 4);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_HF_3D(0);
         FILTER_HF_3D(1);
         FILTER_HF_3D(2);
         FILTER_HF_3D(3);
      }
   }
   else if ((datatype == GL_HALF_FLOAT_ARB) && (comps == 3)) {
      DECLARE_ROW_POINTERS(GLhalfARB, 3);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_HF_3D(0);
         FILTER_HF_3D(1);
         FILTER_HF_3D(2);
      }
   }
   else if ((datatype == GL_HALF_FLOAT_ARB) && (comps == 2)) {
      DECLARE_ROW_POINTERS(GLhalfARB, 2);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_HF_3D(0);
         FILTER_HF_3D(1);
      }
   }
   else if ((datatype == GL_HALF_FLOAT_ARB) && (comps == 1)) {
      DECLARE_ROW_POINTERS(GLhalfARB, 1);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_HF_3D(0);
      }
   }
   else if ((datatype == GL_UNSIGNED_INT) && (comps == 1)) {
      const GLuint *rowA = (const GLuint *) srcRowA;
      const GLuint *rowB = (const GLuint *) srcRowB;
      const GLuint *rowC = (const GLuint *) srcRowC;
      const GLuint *rowD = (const GLuint *) srcRowD;
      GLfloat *dst = (GLfloat *) dstRow;

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         const uint64_t tmp = (((uint64_t) rowA[j] + (uint64_t) rowA[k])
                               + ((uint64_t) rowB[j] + (uint64_t) rowB[k])
                               + ((uint64_t) rowC[j] + (uint64_t) rowC[k])
                               + ((uint64_t) rowD[j] + (uint64_t) rowD[k]));
         dst[i] = (GLfloat)((double) tmp * 0.125);
      }
   }
   else if ((datatype == GL_UNSIGNED_SHORT_5_6_5) && (comps == 3)) {
      DECLARE_ROW_POINTERS0(GLushort);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         const GLint rowAr0 = rowA[j] & 0x1f;
         const GLint rowAr1 = rowA[k] & 0x1f;
         const GLint rowBr0 = rowB[j] & 0x1f;
         const GLint rowBr1 = rowB[k] & 0x1f;
         const GLint rowCr0 = rowC[j] & 0x1f;
         const GLint rowCr1 = rowC[k] & 0x1f;
         const GLint rowDr0 = rowD[j] & 0x1f;
         const GLint rowDr1 = rowD[k] & 0x1f;
         const GLint rowAg0 = (rowA[j] >> 5) & 0x3f;
         const GLint rowAg1 = (rowA[k] >> 5) & 0x3f;
         const GLint rowBg0 = (rowB[j] >> 5) & 0x3f;
         const GLint rowBg1 = (rowB[k] >> 5) & 0x3f;
         const GLint rowCg0 = (rowC[j] >> 5) & 0x3f;
         const GLint rowCg1 = (rowC[k] >> 5) & 0x3f;
         const GLint rowDg0 = (rowD[j] >> 5) & 0x3f;
         const GLint rowDg1 = (rowD[k] >> 5) & 0x3f;
         const GLint rowAb0 = (rowA[j] >> 11) & 0x1f;
         const GLint rowAb1 = (rowA[k] >> 11) & 0x1f;
         const GLint rowBb0 = (rowB[j] >> 11) & 0x1f;
         const GLint rowBb1 = (rowB[k] >> 11) & 0x1f;
         const GLint rowCb0 = (rowC[j] >> 11) & 0x1f;
         const GLint rowCb1 = (rowC[k] >> 11) & 0x1f;
         const GLint rowDb0 = (rowD[j] >> 11) & 0x1f;
         const GLint rowDb1 = (rowD[k] >> 11) & 0x1f;
         const GLint r = FILTER_SUM_3D(rowAr0, rowAr1, rowBr0, rowBr1,
                                       rowCr0, rowCr1, rowDr0, rowDr1);
         const GLint g = FILTER_SUM_3D(rowAg0, rowAg1, rowBg0, rowBg1,
                                       rowCg0, rowCg1, rowDg0, rowDg1);
         const GLint b = FILTER_SUM_3D(rowAb0, rowAb1, rowBb0, rowBb1,
                                       rowCb0, rowCb1, rowDb0, rowDb1);
         dst[i] = (b << 11) | (g << 5) | r;
      }
   }
   else if ((datatype == GL_UNSIGNED_SHORT_4_4_4_4) && (comps == 4)) {
      DECLARE_ROW_POINTERS0(GLushort);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         const GLint rowAr0 = rowA[j] & 0xf;
         const GLint rowAr1 = rowA[k] & 0xf;
         const GLint rowBr0 = rowB[j] & 0xf;
         const GLint rowBr1 = rowB[k] & 0xf;
         const GLint rowCr0 = rowC[j] & 0xf;
         const GLint rowCr1 = rowC[k] & 0xf;
         const GLint rowDr0 = rowD[j] & 0xf;
         const GLint rowDr1 = rowD[k] & 0xf;
         const GLint rowAg0 = (rowA[j] >> 4) & 0xf;
         const GLint rowAg1 = (rowA[k] >> 4) & 0xf;
         const GLint rowBg0 = (rowB[j] >> 4) & 0xf;
         const GLint rowBg1 = (rowB[k] >> 4) & 0xf;
         const GLint rowCg0 = (rowC[j] >> 4) & 0xf;
         const GLint rowCg1 = (rowC[k] >> 4) & 0xf;
         const GLint rowDg0 = (rowD[j] >> 4) & 0xf;
         const GLint rowDg1 = (rowD[k] >> 4) & 0xf;
         const GLint rowAb0 = (rowA[j] >> 8) & 0xf;
         const GLint rowAb1 = (rowA[k] >> 8) & 0xf;
         const GLint rowBb0 = (rowB[j] >> 8) & 0xf;
         const GLint rowBb1 = (rowB[k] >> 8) & 0xf;
         const GLint rowCb0 = (rowC[j] >> 8) & 0xf;
         const GLint rowCb1 = (rowC[k] >> 8) & 0xf;
         const GLint rowDb0 = (rowD[j] >> 8) & 0xf;
         const GLint rowDb1 = (rowD[k] >> 8) & 0xf;
         const GLint rowAa0 = (rowA[j] >> 12) & 0xf;
         const GLint rowAa1 = (rowA[k] >> 12) & 0xf;
         const GLint rowBa0 = (rowB[j] >> 12) & 0xf;
         const GLint rowBa1 = (rowB[k] >> 12) & 0xf;
         const GLint rowCa0 = (rowC[j] >> 12) & 0xf;
         const GLint rowCa1 = (rowC[k] >> 12) & 0xf;
         const GLint rowDa0 = (rowD[j] >> 12) & 0xf;
         const GLint rowDa1 = (rowD[k] >> 12) & 0xf;
         const GLint r = FILTER_SUM_3D(rowAr0, rowAr1, rowBr0, rowBr1,
                                       rowCr0, rowCr1, rowDr0, rowDr1);
         const GLint g = FILTER_SUM_3D(rowAg0, rowAg1, rowBg0, rowBg1,
                                       rowCg0, rowCg1, rowDg0, rowDg1);
         const GLint b = FILTER_SUM_3D(rowAb0, rowAb1, rowBb0, rowBb1,
                                       rowCb0, rowCb1, rowDb0, rowDb1);
         const GLint a = FILTER_SUM_3D(rowAa0, rowAa1, rowBa0, rowBa1,
                                       rowCa0, rowCa1, rowDa0, rowDa1);

         dst[i] = (a << 12) | (b << 8) | (g << 4) | r;
      }
   }
   else if ((datatype == GL_UNSIGNED_SHORT_1_5_5_5_REV) && (comps == 4)) {
      DECLARE_ROW_POINTERS0(GLushort);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         const GLint rowAr0 = rowA[j] & 0x1f;
         const GLint rowAr1 = rowA[k] & 0x1f;
         const GLint rowBr0 = rowB[j] & 0x1f;
         const GLint rowBr1 = rowB[k] & 0x1f;
         const GLint rowCr0 = rowC[j] & 0x1f;
         const GLint rowCr1 = rowC[k] & 0x1f;
         const GLint rowDr0 = rowD[j] & 0x1f;
         const GLint rowDr1 = rowD[k] & 0x1f;
         const GLint rowAg0 = (rowA[j] >> 5) & 0x1f;
         const GLint rowAg1 = (rowA[k] >> 5) & 0x1f;
         const GLint rowBg0 = (rowB[j] >> 5) & 0x1f;
         const GLint rowBg1 = (rowB[k] >> 5) & 0x1f;
         const GLint rowCg0 = (rowC[j] >> 5) & 0x1f;
         const GLint rowCg1 = (rowC[k] >> 5) & 0x1f;
         const GLint rowDg0 = (rowD[j] >> 5) & 0x1f;
         const GLint rowDg1 = (rowD[k] >> 5) & 0x1f;
         const GLint rowAb0 = (rowA[j] >> 10) & 0x1f;
         const GLint rowAb1 = (rowA[k] >> 10) & 0x1f;
         const GLint rowBb0 = (rowB[j] >> 10) & 0x1f;
         const GLint rowBb1 = (rowB[k] >> 10) & 0x1f;
         const GLint rowCb0 = (rowC[j] >> 10) & 0x1f;
         const GLint rowCb1 = (rowC[k] >> 10) & 0x1f;
         const GLint rowDb0 = (rowD[j] >> 10) & 0x1f;
         const GLint rowDb1 = (rowD[k] >> 10) & 0x1f;
         const GLint rowAa0 = (rowA[j] >> 15) & 0x1;
         const GLint rowAa1 = (rowA[k] >> 15) & 0x1;
         const GLint rowBa0 = (rowB[j] >> 15) & 0x1;
         const GLint rowBa1 = (rowB[k] >> 15) & 0x1;
         const GLint rowCa0 = (rowC[j] >> 15) & 0x1;
         const GLint rowCa1 = (rowC[k] >> 15) & 0x1;
         const GLint rowDa0 = (rowD[j] >> 15) & 0x1;
         const GLint rowDa1 = (rowD[k] >> 15) & 0x1;
         const GLint r = FILTER_SUM_3D(rowAr0, rowAr1, rowBr0, rowBr1,
                                       rowCr0, rowCr1, rowDr0, rowDr1);
         const GLint g = FILTER_SUM_3D(rowAg0, rowAg1, rowBg0, rowBg1,
                                       rowCg0, rowCg1, rowDg0, rowDg1);
         const GLint b = FILTER_SUM_3D(rowAb0, rowAb1, rowBb0, rowBb1,
                                       rowCb0, rowCb1, rowDb0, rowDb1);
         const GLint a = FILTER_SUM_3D(rowAa0, rowAa1, rowBa0, rowBa1,
                                       rowCa0, rowCa1, rowDa0, rowDa1);

         dst[i] = (a << 15) | (b << 10) | (g << 5) | r;
      }
   }
   else if ((datatype == GL_UNSIGNED_SHORT_5_5_5_1) && (comps == 4)) {
      DECLARE_ROW_POINTERS0(GLushort);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         const GLint rowAr0 = (rowA[j] >> 11) & 0x1f;
         const GLint rowAr1 = (rowA[k] >> 11) & 0x1f;
         const GLint rowBr0 = (rowB[j] >> 11) & 0x1f;
         const GLint rowBr1 = (rowB[k] >> 11) & 0x1f;
         const GLint rowCr0 = (rowC[j] >> 11) & 0x1f;
         const GLint rowCr1 = (rowC[k] >> 11) & 0x1f;
         const GLint rowDr0 = (rowD[j] >> 11) & 0x1f;
         const GLint rowDr1 = (rowD[k] >> 11) & 0x1f;
         const GLint rowAg0 = (rowA[j] >> 6) & 0x1f;
         const GLint rowAg1 = (rowA[k] >> 6) & 0x1f;
         const GLint rowBg0 = (rowB[j] >> 6) & 0x1f;
         const GLint rowBg1 = (rowB[k] >> 6) & 0x1f;
         const GLint rowCg0 = (rowC[j] >> 6) & 0x1f;
         const GLint rowCg1 = (rowC[k] >> 6) & 0x1f;
         const GLint rowDg0 = (rowD[j] >> 6) & 0x1f;
         const GLint rowDg1 = (rowD[k] >> 6) & 0x1f;
         const GLint rowAb0 = (rowA[j] >> 1) & 0x1f;
         const GLint rowAb1 = (rowA[k] >> 1) & 0x1f;
         const GLint rowBb0 = (rowB[j] >> 1) & 0x1f;
         const GLint rowBb1 = (rowB[k] >> 1) & 0x1f;
         const GLint rowCb0 = (rowC[j] >> 1) & 0x1f;
         const GLint rowCb1 = (rowC[k] >> 1) & 0x1f;
         const GLint rowDb0 = (rowD[j] >> 1) & 0x1f;
         const GLint rowDb1 = (rowD[k] >> 1) & 0x1f;
         const GLint rowAa0 = (rowA[j] & 0x1);
         const GLint rowAa1 = (rowA[k] & 0x1);
         const GLint rowBa0 = (rowB[j] & 0x1);
         const GLint rowBa1 = (rowB[k] & 0x1);
         const GLint rowCa0 = (rowC[j] & 0x1);
         const GLint rowCa1 = (rowC[k] & 0x1);
         const GLint rowDa0 = (rowD[j] & 0x1);
         const GLint rowDa1 = (rowD[k] & 0x1);
         const GLint r = FILTER_SUM_3D(rowAr0, rowAr1, rowBr0, rowBr1,
                                       rowCr0, rowCr1, rowDr0, rowDr1);
         const GLint g = FILTER_SUM_3D(rowAg0, rowAg1, rowBg0, rowBg1,
                                       rowCg0, rowCg1, rowDg0, rowDg1);
         const GLint b = FILTER_SUM_3D(rowAb0, rowAb1, rowBb0, rowBb1,
                                       rowCb0, rowCb1, rowDb0, rowDb1);
         const GLint a = FILTER_SUM_3D(rowAa0, rowAa1, rowBa0, rowBa1,
                                       rowCa0, rowCa1, rowDa0, rowDa1);

         dst[i] = (r << 11) | (g << 6) | (b << 1) | a;
      }
   }
   else if ((datatype == GL_UNSIGNED_BYTE_3_3_2) && (comps == 3)) {
      DECLARE_ROW_POINTERS0(GLubyte);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         const GLint rowAr0 = rowA[j] & 0x3;
         const GLint rowAr1 = rowA[k] & 0x3;
         const GLint rowBr0 = rowB[j] & 0x3;
         const GLint rowBr1 = rowB[k] & 0x3;
         const GLint rowCr0 = rowC[j] & 0x3;
         const GLint rowCr1 = rowC[k] & 0x3;
         const GLint rowDr0 = rowD[j] & 0x3;
         const GLint rowDr1 = rowD[k] & 0x3;
         const GLint rowAg0 = (rowA[j] >> 2) & 0x7;
         const GLint rowAg1 = (rowA[k] >> 2) & 0x7;
         const GLint rowBg0 = (rowB[j] >> 2) & 0x7;
         const GLint rowBg1 = (rowB[k] >> 2) & 0x7;
         const GLint rowCg0 = (rowC[j] >> 2) & 0x7;
         const GLint rowCg1 = (rowC[k] >> 2) & 0x7;
         const GLint rowDg0 = (rowD[j] >> 2) & 0x7;
         const GLint rowDg1 = (rowD[k] >> 2) & 0x7;
         const GLint rowAb0 = (rowA[j] >> 5) & 0x7;
         const GLint rowAb1 = (rowA[k] >> 5) & 0x7;
         const GLint rowBb0 = (rowB[j] >> 5) & 0x7;
         const GLint rowBb1 = (rowB[k] >> 5) & 0x7;
         const GLint rowCb0 = (rowC[j] >> 5) & 0x7;
         const GLint rowCb1 = (rowC[k] >> 5) & 0x7;
         const GLint rowDb0 = (rowD[j] >> 5) & 0x7;
         const GLint rowDb1 = (rowD[k] >> 5) & 0x7;
         const GLint r = FILTER_SUM_3D(rowAr0, rowAr1, rowBr0, rowBr1,
                                       rowCr0, rowCr1, rowDr0, rowDr1);
         const GLint g = FILTER_SUM_3D(rowAg0, rowAg1, rowBg0, rowBg1,
                                       rowCg0, rowCg1, rowDg0, rowDg1);
         const GLint b = FILTER_SUM_3D(rowAb0, rowAb1, rowBb0, rowBb1,
                                       rowCb0, rowCb1, rowDb0, rowDb1);
         dst[i] = (b << 5) | (g << 2) | r;
      }
   }
   else if (datatype == MESA_UNSIGNED_BYTE_4_4 && comps == 2) {
      DECLARE_ROW_POINTERS0(GLubyte);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         const GLint rowAr0 = rowA[j] & 0xf;
         const GLint rowAr1 = rowA[k] & 0xf;
         const GLint rowBr0 = rowB[j] & 0xf;
         const GLint rowBr1 = rowB[k] & 0xf;
         const GLint rowCr0 = rowC[j] & 0xf;
         const GLint rowCr1 = rowC[k] & 0xf;
         const GLint rowDr0 = rowD[j] & 0xf;
         const GLint rowDr1 = rowD[k] & 0xf;
         const GLint rowAg0 = (rowA[j] >> 4) & 0xf;
         const GLint rowAg1 = (rowA[k] >> 4) & 0xf;
         const GLint rowBg0 = (rowB[j] >> 4) & 0xf;
         const GLint rowBg1 = (rowB[k] >> 4) & 0xf;
         const GLint rowCg0 = (rowC[j] >> 4) & 0xf;
         const GLint rowCg1 = (rowC[k] >> 4) & 0xf;
         const GLint rowDg0 = (rowD[j] >> 4) & 0xf;
         const GLint rowDg1 = (rowD[k] >> 4) & 0xf;
         const GLint r = FILTER_SUM_3D(rowAr0, rowAr1, rowBr0, rowBr1,
                                       rowCr0, rowCr1, rowDr0, rowDr1);
         const GLint g = FILTER_SUM_3D(rowAg0, rowAg1, rowBg0, rowBg1,
                                       rowCg0, rowCg1, rowDg0, rowDg1);
         dst[i] = (g << 4) | r;
      }
   }
   else if ((datatype == GL_UNSIGNED_INT_2_10_10_10_REV) && (comps == 4)) {
      DECLARE_ROW_POINTERS0(GLuint);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         const GLint rowAr0 = rowA[j] & 0x3ff;
         const GLint rowAr1 = rowA[k] & 0x3ff;
         const GLint rowBr0 = rowB[j] & 0x3ff;
         const GLint rowBr1 = rowB[k] & 0x3ff;
         const GLint rowCr0 = rowC[j] & 0x3ff;
         const GLint rowCr1 = rowC[k] & 0x3ff;
         const GLint rowDr0 = rowD[j] & 0x3ff;
         const GLint rowDr1 = rowD[k] & 0x3ff;
         const GLint rowAg0 = (rowA[j] >> 10) & 0x3ff;
         const GLint rowAg1 = (rowA[k] >> 10) & 0x3ff;
         const GLint rowBg0 = (rowB[j] >> 10) & 0x3ff;
         const GLint rowBg1 = (rowB[k] >> 10) & 0x3ff;
         const GLint rowCg0 = (rowC[j] >> 10) & 0x3ff;
         const GLint rowCg1 = (rowC[k] >> 10) & 0x3ff;
         const GLint rowDg0 = (rowD[j] >> 10) & 0x3ff;
         const GLint rowDg1 = (rowD[k] >> 10) & 0x3ff;
         const GLint rowAb0 = (rowA[j] >> 20) & 0x3ff;
         const GLint rowAb1 = (rowA[k] >> 20) & 0x3ff;
         const GLint rowBb0 = (rowB[j] >> 20) & 0x3ff;
         const GLint rowBb1 = (rowB[k] >> 20) & 0x3ff;
         const GLint rowCb0 = (rowC[j] >> 20) & 0x3ff;
         const GLint rowCb1 = (rowC[k] >> 20) & 0x3ff;
         const GLint rowDb0 = (rowD[j] >> 20) & 0x3ff;
         const GLint rowDb1 = (rowD[k] >> 20) & 0x3ff;
         const GLint rowAa0 = (rowA[j] >> 30) & 0x3;
         const GLint rowAa1 = (rowA[k] >> 30) & 0x3;
         const GLint rowBa0 = (rowB[j] >> 30) & 0x3;
         const GLint rowBa1 = (rowB[k] >> 30) & 0x3;
         const GLint rowCa0 = (rowC[j] >> 30) & 0x3;
         const GLint rowCa1 = (rowC[k] >> 30) & 0x3;
         const GLint rowDa0 = (rowD[j] >> 30) & 0x3;
         const GLint rowDa1 = (rowD[k] >> 30) & 0x3;
         const GLint r = FILTER_SUM_3D(rowAr0, rowAr1, rowBr0, rowBr1,
                                       rowCr0, rowCr1, rowDr0, rowDr1);
         const GLint g = FILTER_SUM_3D(rowAg0, rowAg1, rowBg0, rowBg1,
                                       rowCg0, rowCg1, rowDg0, rowDg1);
         const GLint b = FILTER_SUM_3D(rowAb0, rowAb1, rowBb0, rowBb1,
                                       rowCb0, rowCb1, rowDb0, rowDb1);
         const GLint a = FILTER_SUM_3D(rowAa0, rowAa1, rowBa0, rowBa1,
                                       rowCa0, rowCa1, rowDa0, rowDa1);

         dst[i] = (a << 30) | (b << 20) | (g << 10) | r;
      }
   }

   else if (datatype == GL_UNSIGNED_INT_5_9_9_9_REV && comps == 3) {
      DECLARE_ROW_POINTERS0(GLuint);

      GLfloat res[3];
      GLfloat rowAj[3], rowBj[3], rowCj[3], rowDj[3];
      GLfloat rowAk[3], rowBk[3], rowCk[3], rowDk[3];

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         rgb9e5_to_float3(rowA[j], rowAj);
         rgb9e5_to_float3(rowB[j], rowBj);
         rgb9e5_to_float3(rowC[j], rowCj);
         rgb9e5_to_float3(rowD[j], rowDj);
         rgb9e5_to_float3(rowA[k], rowAk);
         rgb9e5_to_float3(rowB[k], rowBk);
         rgb9e5_to_float3(rowC[k], rowCk);
         rgb9e5_to_float3(rowD[k], rowDk);
         res[0] = (rowAj[0] + rowAk[0] + rowBj[0] + rowBk[0] +
                   rowCj[0] + rowCk[0] + rowDj[0] + rowDk[0]) * 0.125F;
         res[1] = (rowAj[1] + rowAk[1] + rowBj[1] + rowBk[1] +
                   rowCj[1] + rowCk[1] + rowDj[1] + rowDk[1]) * 0.125F;
         res[2] = (rowAj[2] + rowAk[2] + rowBj[2] + rowBk[2] +
                   rowCj[2] + rowCk[2] + rowDj[2] + rowDk[2]) * 0.125F;
         dst[i] = float3_to_rgb9e5(res);
      }
   }

   else if (datatype == GL_UNSIGNED_INT_10F_11F_11F_REV && comps == 3) {
      DECLARE_ROW_POINTERS0(GLuint);

      GLfloat res[3];
      GLfloat rowAj[3], rowBj[3], rowCj[3], rowDj[3];
      GLfloat rowAk[3], rowBk[3], rowCk[3], rowDk[3];

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         r11g11b10f_to_float3(rowA[j], rowAj);
         r11g11b10f_to_float3(rowB[j], rowBj);
         r11g11b10f_to_float3(rowC[j], rowCj);
         r11g11b10f_to_float3(rowD[j], rowDj);
         r11g11b10f_to_float3(rowA[k], rowAk);
         r11g11b10f_to_float3(rowB[k], rowBk);
         r11g11b10f_to_float3(rowC[k], rowCk);
         r11g11b10f_to_float3(rowD[k], rowDk);
         res[0] = (rowAj[0] + rowAk[0] + rowBj[0] + rowBk[0] +
                   rowCj[0] + rowCk[0] + rowDj[0] + rowDk[0]) * 0.125F;
         res[1] = (rowAj[1] + rowAk[1] + rowBj[1] + rowBk[1] +
                   rowCj[1] + rowCk[1] + rowDj[1] + rowDk[1]) * 0.125F;
         res[2] = (rowAj[2] + rowAk[2] + rowBj[2] + rowBk[2] +
                   rowCj[2] + rowCk[2] + rowDj[2] + rowDk[2]) * 0.125F;
         dst[i] = float3_to_r11g11b10f(res);
      }
   }

   else if (datatype == GL_FLOAT_32_UNSIGNED_INT_24_8_REV && comps == 1) {
      DECLARE_ROW_POINTERS(GLfloat, 2);

      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
           i++, j += colStride, k += colStride) {
         FILTER_F_3D(0);
      }
   }

   else {
      _mesa_problem(NULL, "bad format in do_row()");
   }
}


/*
 * These functions generate a 1/2-size mipmap image from a source image.
 * Texture borders are handled by copying or averaging the source image's
 * border texels, depending on the scale-down factor.
 */

static void
make_1d_mipmap(GLenum datatype, GLuint comps, GLint border,
               GLint srcWidth, const GLubyte *srcPtr,
               GLint dstWidth, GLubyte *dstPtr)
{
   const GLint bpt = bytes_per_pixel(datatype, comps);
   const GLubyte *src;
   GLubyte *dst;

   /* skip the border pixel, if any */
   src = srcPtr + border * bpt;
   dst = dstPtr + border * bpt;

   /* we just duplicate the input row, kind of hack, saves code */
   do_row(datatype, comps, srcWidth - 2 * border, src, src,
          dstWidth - 2 * border, dst);

   if (border) {
      /* copy left-most pixel from source */
      assert(dstPtr);
      assert(srcPtr);
      memcpy(dstPtr, srcPtr, bpt);
      /* copy right-most pixel from source */
      memcpy(dstPtr + (dstWidth - 1) * bpt,
             srcPtr + (srcWidth - 1) * bpt,
             bpt);
   }
}


static void
make_2d_mipmap(GLenum datatype, GLuint comps, GLint border,
               GLint srcWidth, GLint srcHeight,
	       const GLubyte *srcPtr, GLint srcRowStride,
               GLint dstWidth, GLint dstHeight,
	       GLubyte *dstPtr, GLint dstRowStride)
{
   const GLint bpt = bytes_per_pixel(datatype, comps);
   const GLint srcWidthNB = srcWidth - 2 * border;  /* sizes w/out border */
   const GLint dstWidthNB = dstWidth - 2 * border;
   const GLint dstHeightNB = dstHeight - 2 * border;
   const GLubyte *srcA, *srcB;
   GLubyte *dst;
   GLint row, srcRowStep;

   /* Compute src and dst pointers, skipping any border */
   srcA = srcPtr + border * ((srcWidth + 1) * bpt);
   if (srcHeight > 1 && srcHeight > dstHeight) {
      /* sample from two source rows */
      srcB = srcA + srcRowStride;
      srcRowStep = 2;
   }
   else {
      /* sample from one source row */
      srcB = srcA;
      srcRowStep = 1;
   }

   dst = dstPtr + border * ((dstWidth + 1) * bpt);

   for (row = 0; row < dstHeightNB; row++) {
      do_row(datatype, comps, srcWidthNB, srcA, srcB,
             dstWidthNB, dst);
      srcA += srcRowStep * srcRowStride;
      srcB += srcRowStep * srcRowStride;
      dst += dstRowStride;
   }

   /* This is ugly but probably won't be used much */
   if (border > 0) {
      /* fill in dest border */
      /* lower-left border pixel */
      assert(dstPtr);
      assert(srcPtr);
      memcpy(dstPtr, srcPtr, bpt);
      /* lower-right border pixel */
      memcpy(dstPtr + (dstWidth - 1) * bpt,
             srcPtr + (srcWidth - 1) * bpt, bpt);
      /* upper-left border pixel */
      memcpy(dstPtr + dstWidth * (dstHeight - 1) * bpt,
             srcPtr + srcWidth * (srcHeight - 1) * bpt, bpt);
      /* upper-right border pixel */
      memcpy(dstPtr + (dstWidth * dstHeight - 1) * bpt,
             srcPtr + (srcWidth * srcHeight - 1) * bpt, bpt);
      /* lower border */
      do_row(datatype, comps, srcWidthNB,
             srcPtr + bpt,
             srcPtr + bpt,
             dstWidthNB, dstPtr + bpt);
      /* upper border */
      do_row(datatype, comps, srcWidthNB,
             srcPtr + (srcWidth * (srcHeight - 1) + 1) * bpt,
             srcPtr + (srcWidth * (srcHeight - 1) + 1) * bpt,
             dstWidthNB,
             dstPtr + (dstWidth * (dstHeight - 1) + 1) * bpt);
      /* left and right borders */
      if (srcHeight == dstHeight) {
         /* copy border pixel from src to dst */
         for (row = 1; row < srcHeight; row++) {
            memcpy(dstPtr + dstWidth * row * bpt,
                   srcPtr + srcWidth * row * bpt, bpt);
            memcpy(dstPtr + (dstWidth * row + dstWidth - 1) * bpt,
                   srcPtr + (srcWidth * row + srcWidth - 1) * bpt, bpt);
         }
      }
      else {
         /* average two src pixels each dest pixel */
         for (row = 0; row < dstHeightNB; row += 2) {
            do_row(datatype, comps, 1,
                   srcPtr + (srcWidth * (row * 2 + 1)) * bpt,
                   srcPtr + (srcWidth * (row * 2 + 2)) * bpt,
                   1, dstPtr + (dstWidth * row + 1) * bpt);
            do_row(datatype, comps, 1,
                   srcPtr + (srcWidth * (row * 2 + 1) + srcWidth - 1) * bpt,
                   srcPtr + (srcWidth * (row * 2 + 2) + srcWidth - 1) * bpt,
                   1, dstPtr + (dstWidth * row + 1 + dstWidth - 1) * bpt);
         }
      }
   }
}


static void
make_3d_mipmap(GLenum datatype, GLuint comps, GLint border,
               GLint srcWidth, GLint srcHeight, GLint srcDepth,
               const GLubyte **srcPtr, GLint srcRowStride,
               GLint dstWidth, GLint dstHeight, GLint dstDepth,
               GLubyte **dstPtr, GLint dstRowStride)
{
   const GLint bpt = bytes_per_pixel(datatype, comps);
   const GLint srcWidthNB = srcWidth - 2 * border;  /* sizes w/out border */
   const GLint srcDepthNB = srcDepth - 2 * border;
   const GLint dstWidthNB = dstWidth - 2 * border;
   const GLint dstHeightNB = dstHeight - 2 * border;
   const GLint dstDepthNB = dstDepth - 2 * border;
   GLint img, row;
   GLint bytesPerSrcImage, bytesPerDstImage;
   GLint srcImageOffset, srcRowOffset;

   (void) srcDepthNB; /* silence warnings */

   bytesPerSrcImage = srcRowStride * srcHeight * bpt;
   bytesPerDstImage = dstRowStride * dstHeight * bpt;

   /* Offset between adjacent src images to be averaged together */
   srcImageOffset = (srcDepth == dstDepth) ? 0 : 1;

   /* Offset between adjacent src rows to be averaged together */
   srcRowOffset = (srcHeight == dstHeight) ? 0 : srcRowStride;

   /*
    * Need to average together up to 8 src pixels for each dest pixel.
    * Break that down into 3 operations:
    *   1. take two rows from source image and average them together.
    *   2. take two rows from next source image and average them together.
    *   3. take the two averaged rows and average them for the final dst row.
    */

   /*
   printf("mip3d %d x %d x %d  ->  %d x %d x %d\n",
          srcWidth, srcHeight, srcDepth, dstWidth, dstHeight, dstDepth);
   */

   for (img = 0; img < dstDepthNB; img++) {
      /* first source image pointer, skipping border */
      const GLubyte *imgSrcA = srcPtr[img * 2 + border]
         + srcRowStride * border + bpt * border;
      /* second source image pointer, skipping border */
      const GLubyte *imgSrcB = srcPtr[img * 2 + srcImageOffset + border]
         + srcRowStride * border + bpt * border;

      /* address of the dest image, skipping border */
      GLubyte *imgDst = dstPtr[img + border]
         + dstRowStride * border + bpt * border;

      /* setup the four source row pointers and the dest row pointer */
      const GLubyte *srcImgARowA = imgSrcA;
      const GLubyte *srcImgARowB = imgSrcA + srcRowOffset;
      const GLubyte *srcImgBRowA = imgSrcB;
      const GLubyte *srcImgBRowB = imgSrcB + srcRowOffset;
      GLubyte *dstImgRow = imgDst;

      for (row = 0; row < dstHeightNB; row++) {
         do_row_3D(datatype, comps, srcWidthNB, 
                   srcImgARowA, srcImgARowB,
                   srcImgBRowA, srcImgBRowB,
                   dstWidthNB, dstImgRow);

         /* advance to next rows */
         srcImgARowA += srcRowStride + srcRowOffset;
         srcImgARowB += srcRowStride + srcRowOffset;
         srcImgBRowA += srcRowStride + srcRowOffset;
         srcImgBRowB += srcRowStride + srcRowOffset;
         dstImgRow += dstRowStride;
      }
   }


   /* Luckily we can leverage the make_2d_mipmap() function here! */
   if (border > 0) {
      /* do front border image */
      make_2d_mipmap(datatype, comps, 1,
                     srcWidth, srcHeight, srcPtr[0], srcRowStride,
                     dstWidth, dstHeight, dstPtr[0], dstRowStride);
      /* do back border image */
      make_2d_mipmap(datatype, comps, 1,
                     srcWidth, srcHeight, srcPtr[srcDepth - 1], srcRowStride,
                     dstWidth, dstHeight, dstPtr[dstDepth - 1], dstRowStride);

      /* do four remaining border edges that span the image slices */
      if (srcDepth == dstDepth) {
         /* just copy border pixels from src to dst */
         for (img = 0; img < dstDepthNB; img++) {
            const GLubyte *src;
            GLubyte *dst;

            /* do border along [img][row=0][col=0] */
            src = srcPtr[img * 2];
            dst = dstPtr[img];
            memcpy(dst, src, bpt);

            /* do border along [img][row=dstHeight-1][col=0] */
            src = srcPtr[img * 2] + (srcHeight - 1) * srcRowStride;
            dst = dstPtr[img] + (dstHeight - 1) * dstRowStride;
            memcpy(dst, src, bpt);

            /* do border along [img][row=0][col=dstWidth-1] */
            src = srcPtr[img * 2] + (srcWidth - 1) * bpt;
            dst = dstPtr[img] + (dstWidth - 1) * bpt;
            memcpy(dst, src, bpt);

            /* do border along [img][row=dstHeight-1][col=dstWidth-1] */
            src = srcPtr[img * 2] + (bytesPerSrcImage - bpt);
            dst = dstPtr[img] + (bytesPerDstImage - bpt);
            memcpy(dst, src, bpt);
         }
      }
      else {
         /* average border pixels from adjacent src image pairs */
         assert(srcDepthNB == 2 * dstDepthNB);
         for (img = 0; img < dstDepthNB; img++) {
            const GLubyte *srcA, *srcB;
            GLubyte *dst;

            /* do border along [img][row=0][col=0] */
            srcA = srcPtr[img * 2 + 0];
            srcB = srcPtr[img * 2 + srcImageOffset];
            dst = dstPtr[img];
            do_row(datatype, comps, 1, srcA, srcB, 1, dst);

            /* do border along [img][row=dstHeight-1][col=0] */
            srcA = srcPtr[img * 2 + 0]
               + (srcHeight - 1) * srcRowStride;
            srcB = srcPtr[img * 2 + srcImageOffset]
               + (srcHeight - 1) * srcRowStride;
            dst = dstPtr[img] + (dstHeight - 1) * dstRowStride;
            do_row(datatype, comps, 1, srcA, srcB, 1, dst);

            /* do border along [img][row=0][col=dstWidth-1] */
            srcA = srcPtr[img * 2 + 0] + (srcWidth - 1) * bpt;
            srcB = srcPtr[img * 2 + srcImageOffset] + (srcWidth - 1) * bpt;
            dst = dstPtr[img] + (dstWidth - 1) * bpt;
            do_row(datatype, comps, 1, srcA, srcB, 1, dst);

            /* do border along [img][row=dstHeight-1][col=dstWidth-1] */
            srcA = srcPtr[img * 2 + 0] + (bytesPerSrcImage - bpt);
            srcB = srcPtr[img * 2 + srcImageOffset] + (bytesPerSrcImage - bpt);
            dst = dstPtr[img] + (bytesPerDstImage - bpt);
            do_row(datatype, comps, 1, srcA, srcB, 1, dst);
         }
      }
   }
}


/**
 * Down-sample a texture image to produce the next lower mipmap level.
 * \param comps  components per texel (1, 2, 3 or 4)
 * \param srcData  array[slice] of pointers to source image slices
 * \param dstData  array[slice] of pointers to dest image slices
 * \param srcRowStride  stride between source rows, in bytes
 * \param dstRowStride  stride between destination rows, in bytes
 */
void
_mesa_generate_mipmap_level(GLenum target,
                            GLenum datatype, GLuint comps,
                            GLint border,
                            GLint srcWidth, GLint srcHeight, GLint srcDepth,
                            const GLubyte **srcData,
                            GLint srcRowStride,
                            GLint dstWidth, GLint dstHeight, GLint dstDepth,
                            GLubyte **dstData,
                            GLint dstRowStride)
{
   int i;

   switch (target) {
   case GL_TEXTURE_1D:
      make_1d_mipmap(datatype, comps, border,
                     srcWidth, srcData[0],
                     dstWidth, dstData[0]);
      break;
   case GL_TEXTURE_2D:
   case GL_TEXTURE_CUBE_MAP_POSITIVE_X:
   case GL_TEXTURE_CUBE_MAP_NEGATIVE_X:
   case GL_TEXTURE_CUBE_MAP_POSITIVE_Y:
   case GL_TEXTURE_CUBE_MAP_NEGATIVE_Y:
   case GL_TEXTURE_CUBE_MAP_POSITIVE_Z:
   case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z:
      make_2d_mipmap(datatype, comps, border,
                     srcWidth, srcHeight, srcData[0], srcRowStride,
                     dstWidth, dstHeight, dstData[0], dstRowStride);
      break;
   case GL_TEXTURE_3D:
      make_3d_mipmap(datatype, comps, border,
                     srcWidth, srcHeight, srcDepth,
                     srcData, srcRowStride,
                     dstWidth, dstHeight, dstDepth,
                     dstData, dstRowStride);
      break;
   case GL_TEXTURE_1D_ARRAY_EXT:
      assert(srcHeight == 1);
      assert(dstHeight == 1);
      for (i = 0; i < dstDepth; i++) {
	 make_1d_mipmap(datatype, comps, border,
			srcWidth, srcData[i],
			dstWidth, dstData[i]);
      }
      break;
   case GL_TEXTURE_2D_ARRAY_EXT:
   case GL_TEXTURE_CUBE_MAP_ARRAY:
      for (i = 0; i < dstDepth; i++) {
	 make_2d_mipmap(datatype, comps, border,
			srcWidth, srcHeight, srcData[i], srcRowStride,
			dstWidth, dstHeight, dstData[i], dstRowStride);
      }
      break;
   case GL_TEXTURE_RECTANGLE_NV:
   case GL_TEXTURE_EXTERNAL_OES:
      /* no mipmaps, do nothing */
      break;
   default:
      _mesa_problem(NULL, "bad tex target in _mesa_generate_mipmaps");
      return;
   }
}


/**
 * compute next (level+1) image size
 * \return GL_FALSE if no smaller size can be generated (eg. src is 1x1x1 size)
 */
GLboolean
_mesa_next_mipmap_level_size(GLenum target, GLint border,
                       GLint srcWidth, GLint srcHeight, GLint srcDepth,
                       GLint *dstWidth, GLint *dstHeight, GLint *dstDepth)
{
   if (srcWidth - 2 * border > 1) {
      *dstWidth = (srcWidth - 2 * border) / 2 + 2 * border;
   }
   else {
      *dstWidth = srcWidth; /* can't go smaller */
   }

   if ((srcHeight - 2 * border > 1) && 
       target != GL_TEXTURE_1D_ARRAY_EXT &&
       target != GL_PROXY_TEXTURE_1D_ARRAY_EXT) {
      *dstHeight = (srcHeight - 2 * border) / 2 + 2 * border;
   }
   else {
      *dstHeight = srcHeight; /* can't go smaller */
   }

   if ((srcDepth - 2 * border > 1) &&
       target != GL_TEXTURE_2D_ARRAY_EXT &&
       target != GL_PROXY_TEXTURE_2D_ARRAY_EXT &&
       target != GL_TEXTURE_CUBE_MAP_ARRAY &&
       target != GL_PROXY_TEXTURE_CUBE_MAP_ARRAY) {
      *dstDepth = (srcDepth - 2 * border) / 2 + 2 * border;
   }
   else {
      *dstDepth = srcDepth; /* can't go smaller */
   }

   if (*dstWidth == srcWidth &&
       *dstHeight == srcHeight &&
       *dstDepth == srcDepth) {
      return GL_FALSE;
   }
   else {
      return GL_TRUE;
   }
}


/**
 * Helper function for mipmap generation.
 * Make sure the specified destination mipmap level is the right size/format
 * for mipmap generation.  If not, (re) allocate it.
 * \return GL_TRUE if successful, GL_FALSE if mipmap generation should stop
 */
static GLboolean
prepare_mipmap_level(struct gl_context *ctx,
                     struct gl_texture_object *texObj, GLuint level,
                     GLsizei width, GLsizei height, GLsizei depth,
                     GLsizei border, GLenum intFormat, mesa_format format)
{
   const GLuint numFaces = _mesa_num_tex_faces(texObj->Target);
   GLuint face;

   if (texObj->Immutable) {
      /* The texture was created with glTexStorage() so the number/size of
       * mipmap levels is fixed and the storage for all images is already
       * allocated.
       */
      if (!texObj->Image[0][level]) {
         /* No more levels to create - we're done */
         return GL_FALSE;
      }
      else {
         /* Nothing to do - the texture memory must have already been
          * allocated to the right size so we're all set.
          */
         return GL_TRUE;
      }
   }

   for (face = 0; face < numFaces; face++) {
      struct gl_texture_image *dstImage;
      const GLenum target = _mesa_cube_face_target(texObj->Target, face);

      dstImage = _mesa_get_tex_image(ctx, texObj, target, level);
      if (!dstImage) {
         /* out of memory */
         return GL_FALSE;
      }

      if (dstImage->Width != width ||
          dstImage->Height != height ||
          dstImage->Depth != depth ||
          dstImage->Border != border ||
          dstImage->InternalFormat != intFormat ||
          dstImage->TexFormat != format) {
         /* need to (re)allocate image */
         ctx->Driver.FreeTextureImageBuffer(ctx, dstImage);

         _mesa_init_teximage_fields(ctx, dstImage,
                                    width, height, depth,
                                    border, intFormat, format);

         ctx->Driver.AllocTextureImageBuffer(ctx, dstImage);

         /* in case the mipmap level is part of an FBO: */
         _mesa_update_fbo_texture(ctx, texObj, face, level);

         ctx->NewState |= _NEW_TEXTURE;
      }
   }

   return GL_TRUE;
}


/**
 * Prepare all mipmap levels beyond 'baseLevel' for mipmap generation.
 * When finished, all the gl_texture_image structures for the smaller
 * mipmap levels will be consistent with the base level (in terms of
 * dimensions, format, etc).
 */
void
_mesa_prepare_mipmap_levels(struct gl_context *ctx,
                            struct gl_texture_object *texObj,
                            unsigned baseLevel, unsigned maxLevel)
{
   const struct gl_texture_image *baseImage =
      _mesa_select_tex_image(texObj, texObj->Target, baseLevel);
   const GLint border = 0;
   GLint width = baseImage->Width;
   GLint height = baseImage->Height;
   GLint depth = baseImage->Depth;
   const GLenum intFormat = baseImage->InternalFormat;
   const mesa_format texFormat = baseImage->TexFormat;
   GLint newWidth, newHeight, newDepth;

   /* Prepare baseLevel + 1, baseLevel + 2, ... */
   for (unsigned level = baseLevel + 1; level <= maxLevel; level++) {
      if (!_mesa_next_mipmap_level_size(texObj->Target, border,
                                        width, height, depth,
                                        &newWidth, &newHeight, &newDepth)) {
         /* all done */
         break;
      }

      if (!prepare_mipmap_level(ctx, texObj, level,
                                newWidth, newHeight, newDepth,
                                border, intFormat, texFormat)) {
         break;
      }

      width = newWidth;
      height = newHeight;
      depth = newDepth;
   }
}


static void
generate_mipmap_uncompressed(struct gl_context *ctx, GLenum target,
			     struct gl_texture_object *texObj,
			     const struct gl_texture_image *srcImage,
			     GLuint maxLevel)
{
   GLuint level;
   GLenum datatype;
   GLuint comps;

   _mesa_uncompressed_format_to_type_and_comps(srcImage->TexFormat, &datatype, &comps);

   for (level = texObj->BaseLevel; level < maxLevel; level++) {
      /* generate image[level+1] from image[level] */
      struct gl_texture_image *srcImage, *dstImage;
      GLint srcRowStride, dstRowStride;
      GLint srcWidth, srcHeight, srcDepth;
      GLint dstWidth, dstHeight, dstDepth;
      GLint border;
      GLint slice;
      GLubyte **srcMaps, **dstMaps;
      GLboolean success = GL_TRUE;

      /* get src image parameters */
      srcImage = _mesa_select_tex_image(texObj, target, level);
      assert(srcImage);
      srcWidth = srcImage->Width;
      srcHeight = srcImage->Height;
      srcDepth = srcImage->Depth;
      border = srcImage->Border;

      /* get dest gl_texture_image */
      dstImage = _mesa_select_tex_image(texObj, target, level + 1);
      if (!dstImage) {
         break;
      }
      dstWidth = dstImage->Width;
      dstHeight = dstImage->Height;
      dstDepth = dstImage->Depth;

      if (target == GL_TEXTURE_1D_ARRAY) {
	 srcDepth = srcHeight;
	 dstDepth = dstHeight;
	 srcHeight = 1;
	 dstHeight = 1;
      }

      /* Map src texture image slices */
      srcMaps = calloc(srcDepth, sizeof(GLubyte *));
      if (srcMaps) {
         for (slice = 0; slice < srcDepth; slice++) {
            ctx->Driver.MapTextureImage(ctx, srcImage, slice,
                                        0, 0, srcWidth, srcHeight,
                                        GL_MAP_READ_BIT,
                                        &srcMaps[slice], &srcRowStride);
            if (!srcMaps[slice]) {
               success = GL_FALSE;
               break;
            }
         }
      }
      else {
         success = GL_FALSE;
      }

      /* Map dst texture image slices */
      dstMaps = calloc(dstDepth, sizeof(GLubyte *));
      if (dstMaps) {
         for (slice = 0; slice < dstDepth; slice++) {
            ctx->Driver.MapTextureImage(ctx, dstImage, slice,
                                        0, 0, dstWidth, dstHeight,
                                        GL_MAP_WRITE_BIT,
                                        &dstMaps[slice], &dstRowStride);
            if (!dstMaps[slice]) {
               success = GL_FALSE;
               break;
            }
         }
      }
      else {
         success = GL_FALSE;
      }

      if (success) {
         /* generate one mipmap level (for 1D/2D/3D/array/etc texture) */
         _mesa_generate_mipmap_level(target, datatype, comps, border,
                                     srcWidth, srcHeight, srcDepth,
                                     (const GLubyte **) srcMaps, srcRowStride,
                                     dstWidth, dstHeight, dstDepth,
                                     dstMaps, dstRowStride);
      }

      /* Unmap src image slices */
      if (srcMaps) {
         for (slice = 0; slice < srcDepth; slice++) {
            if (srcMaps[slice]) {
               ctx->Driver.UnmapTextureImage(ctx, srcImage, slice);
            }
         }
         free(srcMaps);
      }

      /* Unmap dst image slices */
      if (dstMaps) {
         for (slice = 0; slice < dstDepth; slice++) {
            if (dstMaps[slice]) {
               ctx->Driver.UnmapTextureImage(ctx, dstImage, slice);
            }
         }
         free(dstMaps);
      }

      if (!success) {
         _mesa_error(ctx, GL_OUT_OF_MEMORY, "mipmap generation");
         break;
      }
   } /* loop over mipmap levels */
}


static void
generate_mipmap_compressed(struct gl_context *ctx, GLenum target,
			   struct gl_texture_object *texObj,
			   struct gl_texture_image *srcImage,
			   GLuint maxLevel)
{
   GLuint level;
   mesa_format temp_format;
   GLint components;
   GLuint temp_src_row_stride, temp_src_img_stride; /* in bytes */
   GLubyte *temp_src = NULL, *temp_dst = NULL;
   GLenum temp_datatype;
   GLenum temp_base_format;
   GLubyte **temp_src_slices = NULL, **temp_dst_slices = NULL;

   /* only two types of compressed textures at this time */
   assert(texObj->Target == GL_TEXTURE_2D ||
	  texObj->Target == GL_TEXTURE_2D_ARRAY ||
	  texObj->Target == GL_TEXTURE_CUBE_MAP ||
          texObj->Target == GL_TEXTURE_CUBE_MAP_ARRAY);

   /*
    * Choose a format for the temporary, uncompressed base image.
    * Then, get number of components, choose temporary image datatype,
    * and get base format.
    */
   temp_format = _mesa_get_uncompressed_format(srcImage->TexFormat);

   components = _mesa_format_num_components(temp_format);

   switch (_mesa_get_format_datatype(srcImage->TexFormat)) {
   case GL_FLOAT:
      temp_datatype = GL_FLOAT;
      break;
   case GL_SIGNED_NORMALIZED:
      /* Revisit this if we get compressed formats with >8 bits per component */
      temp_datatype = GL_BYTE;
      break;
   default:
      temp_datatype = GL_UNSIGNED_BYTE;
   }

   temp_base_format = _mesa_get_format_base_format(temp_format);


   /* allocate storage for the temporary, uncompressed image */
   temp_src_row_stride = _mesa_format_row_stride(temp_format, srcImage->Width);
   temp_src_img_stride = _mesa_format_image_size(temp_format, srcImage->Width,
                                                 srcImage->Height, 1);
   temp_src = malloc(temp_src_img_stride * srcImage->Depth);

   /* Allocate storage for arrays of slice pointers */
   temp_src_slices = malloc(srcImage->Depth * sizeof(GLubyte *));
   temp_dst_slices = malloc(srcImage->Depth * sizeof(GLubyte *));

   if (!temp_src || !temp_src_slices || !temp_dst_slices) {
      _mesa_error(ctx, GL_OUT_OF_MEMORY, "generate mipmaps");
      goto end;
   }

   /* decompress base image to the temporary src buffer */
   {
      /* save pixel packing mode */
      struct gl_pixelstore_attrib save = ctx->Pack;
      /* use default/tight packing parameters */
      ctx->Pack = ctx->DefaultPacking;

      /* Get the uncompressed image */
      assert(srcImage->Level == texObj->BaseLevel);
      ctx->Driver.GetTexSubImage(ctx,
                                 0, 0, 0,
                                 srcImage->Width, srcImage->Height,
                                 srcImage->Depth,
                                 temp_base_format, temp_datatype,
                                 temp_src, srcImage);
      /* restore packing mode */
      ctx->Pack = save;
   }

   for (level = texObj->BaseLevel; level < maxLevel; level++) {
      /* generate image[level+1] from image[level] */
      const struct gl_texture_image *srcImage;
      struct gl_texture_image *dstImage;
      GLint srcWidth, srcHeight, srcDepth;
      GLint dstWidth, dstHeight, dstDepth;
      GLint border;
      GLuint temp_dst_row_stride, temp_dst_img_stride; /* in bytes */
      GLint i;

      /* get src image parameters */
      srcImage = _mesa_select_tex_image(texObj, target, level);
      assert(srcImage);
      srcWidth = srcImage->Width;
      srcHeight = srcImage->Height;
      srcDepth = srcImage->Depth;
      border = srcImage->Border;

      /* get dest gl_texture_image */
      dstImage = _mesa_select_tex_image(texObj, target, level + 1);
      if (!dstImage) {
         break;
      }
      dstWidth = dstImage->Width;
      dstHeight = dstImage->Height;
      dstDepth = dstImage->Depth;

      /* Compute dst image strides and alloc memory on first iteration */
      temp_dst_row_stride = _mesa_format_row_stride(temp_format, dstWidth);
      temp_dst_img_stride = _mesa_format_image_size(temp_format, dstWidth,
                                                    dstHeight, 1);
      if (!temp_dst) {
	 temp_dst = malloc(temp_dst_img_stride * dstDepth);
	 if (!temp_dst) {
	    _mesa_error(ctx, GL_OUT_OF_MEMORY, "generate mipmaps");
            goto end;
	 }
      }

      /* for 2D arrays, setup array[depth] of slice pointers */
      for (i = 0; i < srcDepth; i++) {
         temp_src_slices[i] = temp_src + temp_src_img_stride * i;
      }
      for (i = 0; i < dstDepth; i++) {
         temp_dst_slices[i] = temp_dst + temp_dst_img_stride * i;
      }

      /* Rescale src image to dest image.
       * This will loop over the slices of a 2D array.
       */
      _mesa_generate_mipmap_level(target, temp_datatype, components, border,
                                  srcWidth, srcHeight, srcDepth,
                                  (const GLubyte **) temp_src_slices,
                                  temp_src_row_stride,
                                  dstWidth, dstHeight, dstDepth,
                                  temp_dst_slices, temp_dst_row_stride);

      /* The image space was allocated above so use glTexSubImage now */
      ctx->Driver.TexSubImage(ctx, 2, dstImage,
                              0, 0, 0, dstWidth, dstHeight, dstDepth,
                              temp_base_format, temp_datatype,
                              temp_dst, &ctx->DefaultPacking);

      /* swap src and dest pointers */
      {
	 GLubyte *temp = temp_src;
	 temp_src = temp_dst;
	 temp_dst = temp;
         temp_src_row_stride = temp_dst_row_stride;
         temp_src_img_stride = temp_dst_img_stride;
      }
   } /* loop over mipmap levels */

end:
   free(temp_src);
   free(temp_dst);
   free(temp_src_slices);
   free(temp_dst_slices);
}

/**
 * Automatic mipmap generation.
 * This is the fallback/default function for ctx->Driver.GenerateMipmap().
 * Generate a complete set of mipmaps from texObj's BaseLevel image.
 * Stop at texObj's MaxLevel or when we get to the 1x1 texture.
 * For cube maps, target will be one of
 * GL_TEXTURE_CUBE_MAP_POSITIVE/NEGATIVE_X/Y/Z; never GL_TEXTURE_CUBE_MAP.
 */
void
_mesa_generate_mipmap(struct gl_context *ctx, GLenum target,
                      struct gl_texture_object *texObj)
{
   struct gl_texture_image *srcImage;
   GLint maxLevel;

   assert(texObj);
   srcImage = _mesa_select_tex_image(texObj, target, texObj->BaseLevel);
   assert(srcImage);

   maxLevel = _mesa_max_texture_levels(ctx, texObj->Target) - 1;
   assert(maxLevel >= 0);  /* bad target */

   maxLevel = MIN2(maxLevel, texObj->MaxLevel);

   _mesa_prepare_mipmap_levels(ctx, texObj, texObj->BaseLevel, maxLevel);

   if (_mesa_is_format_compressed(srcImage->TexFormat)) {
      generate_mipmap_compressed(ctx, target, texObj, srcImage, maxLevel);
   } else {
      generate_mipmap_uncompressed(ctx, target, texObj, srcImage, maxLevel);
   }
}