t_dd_vbtmp.h

Mesa is an open-source implementation of the OpenGL specification - a system for rendering interacti

/*
 * Mesa 3-D graphics library
 * Version:  5.0.1
 *
 * Copyright (C) 1999-2003  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
 * BRIAN PAUL 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.
 *
 * Authors:
 *    Keith Whitwell <[e-mail]>
 */


/* Unlike the other templates here, this assumes quite a bit about the
 * underlying hardware.  Specifically it assumes a d3d-like vertex
 * format, with a layout more or less constrained to look like the
 * following:
 *
 * union {
 *    struct {
 *        float x, y, z, w;
 *        struct { char r, g, b, a; } color;
 *        struct { char r, g, b, fog; } spec;
 *        float u0, v0;
 *        float u1, v1;
 *        float u2, v2;
 *        float u3, v3;
 *    } v;
 *    struct {
 *        float x, y, z, w;
 *        struct { char r, g, b, a; } color;
 *        struct { char r, g, b, fog; } spec;
 *        float u0, v0, q0;
 *        float u1, v1, q1;
 *        float u2, v2, q2;
 *        float u3, v3, q3;
 *    } pv;
 *    struct {
 *        float x, y, z;
 *        struct { char r, g, b, a; } color;
 *    } tv;
 *    float f[16];
 *    unsigned int ui[16];
 *    unsigned char ub4[4][16];
 * }
 *
 
 * VERTEX:   hw vertex type as above
 * VERTEX_COLOR: hw color struct type in VERTEX
 *
 * DO_XYZW:  Emit xyz and maybe w coordinates.
 * DO_RGBA:  Emit color.
 * DO_SPEC:  Emit specular color.
 * DO_FOG:   Emit fog coordinate in specular alpha.
 * DO_TEX0:  Emit tex0 u,v coordinates.
 * DO_TEX1:  Emit tex1 u,v coordinates.
 * DO_TEX2:  Emit tex2 u,v coordinates.
 * DO_TEX3:  Emit tex3 u,v coordinates.
 * DO_PTEX:  Emit tex0,1,2,3 q coordinates where possible.
 *
 * HAVE_RGBA_COLOR: Hardware takes color in rgba order (else bgra).
 *
 * HAVE_HW_VIEWPORT:  Hardware performs viewport transform.
 * HAVE_HW_DIVIDE:  Hardware performs perspective divide.
 *
 * HAVE_TINY_VERTICES:  Hardware understands v.tv format.
 * HAVE_PTEX_VERTICES:  Hardware understands v.pv format.
 * HAVE_NOTEX_VERTICES:  Hardware understands v.v format with texcount 0.
 *
 * Additionally, this template assumes it is emitting *transformed*
 * vertices; the modifications to emit untransformed vertices (ie. to
 * t&l hardware) are probably too great to cooexist with the code
 * already in this file.
 *
 * NOTE: The PTEX vertex format always includes TEX0 and TEX1, even if
 * only TEX0 is enabled, in order to maintain a vertex size which is
 * an exact number of quadwords.
 */

#if (HAVE_HW_VIEWPORT)
#define VIEWPORT_X(dst,x) dst = x
#define VIEWPORT_Y(dst,y) dst = y
#define VIEWPORT_Z(dst,z) dst = z
#else
#define VIEWPORT_X(dst,x) dst = s[0]  * x + s[12]
#define VIEWPORT_Y(dst,y) dst = s[5]  * y + s[13]
#define VIEWPORT_Z(dst,z) dst = s[10] * z + s[14]
#endif

#if (HAVE_HW_DIVIDE && !HAVE_PTEX_VERTICES)
#error "can't cope with this combination" 
#endif 

#ifndef LOCALVARS
#define LOCALVARS
#endif

#ifndef CHECK_HW_DIVIDE
#define CHECK_HW_DIVIDE 1
#endif

#if (HAVE_HW_DIVIDE || DO_SPEC || DO_TEX0 || DO_FOG || !HAVE_TINY_VERTICES)

static void TAG(emit)( GLcontext *ctx,
		       GLuint start, GLuint end,
		       void *dest,
		       GLuint stride )
{
   LOCALVARS
      struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
   GLfloat (*tc0)[4], (*tc1)[4], (*fog)[4];
   GLfloat (*tc2)[4], (*tc3)[4];
   GLfloat (*col)[4], (*spec)[4];
   GLuint tc0_stride, tc1_stride, col_stride, spec_stride, fog_stride;
   GLuint tc2_stride, tc3_stride;
   GLuint tc0_size, tc1_size, col_size;
   GLuint tc2_size, tc3_size;
   GLfloat (*coord)[4];
   GLuint coord_stride;
   VERTEX *v = (VERTEX *)dest;
   const GLfloat *s = GET_VIEWPORT_MAT();
   const GLubyte *mask = VB->ClipMask;
   int i;

/*     fprintf(stderr, "%s(big) importable %d %d..%d\n",  */
/*  	   __FUNCTION__, VB->importable_data, start, end); */

   if (HAVE_HW_VIEWPORT && HAVE_HW_DIVIDE && CHECK_HW_DIVIDE) {
      (void) s;
      coord = VB->ClipPtr->data;
      coord_stride = VB->ClipPtr->stride;
   }
   else {
      coord = VB->NdcPtr->data;
      coord_stride = VB->NdcPtr->stride;
   }

   if (DO_TEX3) {
      const GLuint t3 = GET_TEXSOURCE(3);
      tc3 = VB->TexCoordPtr[t3]->data;
      tc3_stride = VB->TexCoordPtr[t3]->stride;
      if (DO_PTEX)
	 tc3_size = VB->TexCoordPtr[t3]->size;
   }

   if (DO_TEX2) {
      const GLuint t2 = GET_TEXSOURCE(2);
      tc2 = VB->TexCoordPtr[t2]->data;
      tc2_stride = VB->TexCoordPtr[t2]->stride;
      if (DO_PTEX)
	 tc2_size = VB->TexCoordPtr[t2]->size;
   }

   if (DO_TEX1) {
      const GLuint t1 = GET_TEXSOURCE(1);
      tc1 = VB->TexCoordPtr[t1]->data;
      tc1_stride = VB->TexCoordPtr[t1]->stride;
      if (DO_PTEX)
	 tc1_size = VB->TexCoordPtr[t1]->size;
   }

   if (DO_TEX0) {
      const GLuint t0 = GET_TEXSOURCE(0);
      tc0_stride = VB->TexCoordPtr[t0]->stride;
      tc0 = VB->TexCoordPtr[t0]->data;
      if (DO_PTEX) 
	 tc0_size = VB->TexCoordPtr[t0]->size;
   }

   if (DO_RGBA) {
      col_stride = VB->ColorPtr[0]->stride;
      col = VB->ColorPtr[0]->data;
      col_size = VB->ColorPtr[0]->size;
   }

   if (DO_SPEC) {
      if (VB->SecondaryColorPtr[0]) {
	 spec_stride = VB->SecondaryColorPtr[0]->stride;
	 spec = VB->SecondaryColorPtr[0]->data;
      } else {
	 spec = (GLfloat (*)[4])ctx->Current.Attrib[VERT_ATTRIB_COLOR1];
	 spec_stride = 0;
      }
   }

   if (DO_FOG) {
      if (VB->FogCoordPtr) {
	 fog = VB->FogCoordPtr->data;
	 fog_stride = VB->FogCoordPtr->stride;
      }
      else {
	 static GLfloat tmp[4] = {0, 0, 0, 0};
	 fog = &tmp;
	 fog_stride = 0;
      }
   }

   /* May have nonstandard strides:
    */
   if (start) {
      STRIDE_4F(coord, start * coord_stride);
      if (DO_TEX0)
	 STRIDE_4F(tc0, start * tc0_stride);
      if (DO_TEX1) 
	 STRIDE_4F(tc1, start * tc1_stride);
      if (DO_TEX2) 
	 STRIDE_4F(tc2, start * tc2_stride);
      if (DO_TEX3) 
	 STRIDE_4F(tc3, start * tc3_stride);
      if (DO_RGBA) 
	 STRIDE_4F(col, start * col_stride);
      if (DO_SPEC)
	 STRIDE_4F(spec, start * spec_stride);
      if (DO_FOG)
	 STRIDE_4F(fog, start * fog_stride);
   }

   for (i=start; i < end; i++, v = (VERTEX *)((GLubyte *)v + stride)) {
      if (DO_XYZW) {
	 if (HAVE_HW_VIEWPORT || mask[i] == 0) {
	    VIEWPORT_X(v->v.x, coord[0][0]);
	    VIEWPORT_Y(v->v.y, coord[0][1]);
	    VIEWPORT_Z(v->v.z, coord[0][2]);
	    v->v.w = coord[0][3];
	 }
	 STRIDE_4F(coord, coord_stride);
      }
      if (DO_RGBA) {
	 UNCLAMPED_FLOAT_TO_UBYTE(v->v.color.red, col[0][0]);
	 UNCLAMPED_FLOAT_TO_UBYTE(v->v.color.green, col[0][1]);
	 UNCLAMPED_FLOAT_TO_UBYTE(v->v.color.blue, col[0][2]);
	 if (col_size == 4) {
	    UNCLAMPED_FLOAT_TO_UBYTE(v->v.color.alpha, col[0][3]);
	 } else {
	    v->v.color.alpha = CHAN_MAX;
	 }
	 STRIDE_4F(col, col_stride);
      }
      if (DO_SPEC) {
	 UNCLAMPED_FLOAT_TO_UBYTE(v->v.specular.red, spec[0][0]);
	 UNCLAMPED_FLOAT_TO_UBYTE(v->v.specular.green, spec[0][1]);
	 UNCLAMPED_FLOAT_TO_UBYTE(v->v.specular.blue, spec[0][2]);
	 STRIDE_4F(spec, spec_stride);
      }
      if (DO_FOG) {
	 UNCLAMPED_FLOAT_TO_UBYTE(v->v.specular.alpha, fog[0][0]);
	 STRIDE_4F(fog, fog_stride);
      }
      if (DO_TEX0) {
	 v->v.u0 = tc0[0][0];
	 v->v.v0 = tc0[0][1];
	 if (DO_PTEX) {
	    if (HAVE_PTEX_VERTICES) {
	       if (tc0_size == 4) 
		  v->pv.q0 = tc0[0][3];
	       else
		  v->pv.q0 = 1.0;
	    } 
	    else if (tc0_size == 4) {
	       float rhw = 1.0 / tc0[0][3];
	       v->v.w *= tc0[0][3];
	       v->v.u0 *= rhw;
	       v->v.v0 *= rhw;
	    } 
	 } 
	 STRIDE_4F(tc0, tc0_stride);
      }
      if (DO_TEX1) {
	 if (DO_PTEX) {
	    v->pv.u1 = tc1[0][0];
	    v->pv.v1 = tc1[0][1];
	    if (tc1_size == 4) 
	       v->pv.q1 = tc1[0][3];
	    else
	       v->pv.q1 = 1.0;
	 } 
	 else {
	    v->v.u1 = tc1[0][0];
	    v->v.v1 = tc1[0][1];
	 }
	 STRIDE_4F(tc1, tc1_stride);
      } 
      else if (DO_PTEX) {
	 *(GLuint *)&v->pv.q1 = 0;	/* avoid culling on radeon */
      }
      if (DO_TEX2) {
	 if (DO_PTEX) {
	    v->pv.u2 = tc2[0][0];
	    v->pv.v2 = tc2[0][1];
	    if (tc2_size == 4) 
	       v->pv.q2 = tc2[0][3];
	    else
	       v->pv.q2 = 1.0;
	 } 
	 else {
	    v->v.u2 = tc2[0][0];
	    v->v.v2 = tc2[0][1];
	 }
	 STRIDE_4F(tc2, tc2_stride);
      } 
      if (DO_TEX3) {
	 if (DO_PTEX) {
	    v->pv.u3 = tc3[0][0];
	    v->pv.v3 = tc3[0][1];
	    if (tc3_size == 4) 
	       v->pv.q3 = tc3[0][3];
	    else
	       v->pv.q3 = 1.0;
	 } 
	 else {
	    v->v.u3 = tc3[0][0];
	    v->v.v3 = tc3[0][1];
	 }
	 STRIDE_4F(tc3, tc3_stride);
      } 
   }

}
#else

#if HAVE_HW_DIVIDE
#error "cannot use tiny vertices with hw perspective divide"