mirror of
https://github.com/dhewm/dhewm3.git
synced 2024-11-23 12:53:09 +00:00
736ec20d4d
Don't include the lazy precompiled.h everywhere, only what's required for the compilation unit. platform.h needs to be included instead to provide all essential defines and types. All includes use the relative path to the neo or the game specific root. Move all idlib related includes from idlib/Lib.h to precompiled.h. precompiled.h still exists for the MFC stuff in tools/. Add some missing header guards.
354 lines
11 KiB
C++
354 lines
11 KiB
C++
/*
|
|
===========================================================================
|
|
|
|
Doom 3 GPL Source Code
|
|
Copyright (C) 1999-2011 id Software LLC, a ZeniMax Media company.
|
|
|
|
This file is part of the Doom 3 GPL Source Code ("Doom 3 Source Code").
|
|
|
|
Doom 3 Source Code is free software: you can redistribute it and/or modify
|
|
it under the terms of the GNU General Public License as published by
|
|
the Free Software Foundation, either version 3 of the License, or
|
|
(at your option) any later version.
|
|
|
|
Doom 3 Source Code is distributed in the hope that it will be useful,
|
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
GNU General Public License for more details.
|
|
|
|
You should have received a copy of the GNU General Public License
|
|
along with Doom 3 Source Code. If not, see <http://www.gnu.org/licenses/>.
|
|
|
|
In addition, the Doom 3 Source Code is also subject to certain additional terms. You should have received a copy of these additional terms immediately following the terms and conditions of the GNU General Public License which accompanied the Doom 3 Source Code. If not, please request a copy in writing from id Software at the address below.
|
|
|
|
If you have questions concerning this license or the applicable additional terms, you may contact in writing id Software LLC, c/o ZeniMax Media Inc., Suite 120, Rockville, Maryland 20850 USA.
|
|
|
|
===========================================================================
|
|
*/
|
|
|
|
#include "sys/platform.h"
|
|
|
|
#include "renderer/tr_local.h"
|
|
|
|
int c_turboUsedVerts;
|
|
int c_turboUnusedVerts;
|
|
|
|
/*
|
|
=====================
|
|
R_CreateVertexProgramTurboShadowVolume
|
|
|
|
are dangling edges that are outside the light frustum still making planes?
|
|
=====================
|
|
*/
|
|
srfTriangles_t *R_CreateVertexProgramTurboShadowVolume( const idRenderEntityLocal *ent,
|
|
const srfTriangles_t *tri, const idRenderLightLocal *light,
|
|
srfCullInfo_t &cullInfo ) {
|
|
int i, j;
|
|
srfTriangles_t *newTri;
|
|
silEdge_t *sil;
|
|
const glIndex_t *indexes;
|
|
const byte *facing;
|
|
|
|
R_CalcInteractionFacing( ent, tri, light, cullInfo );
|
|
if ( r_useShadowProjectedCull.GetBool() ) {
|
|
R_CalcInteractionCullBits( ent, tri, light, cullInfo );
|
|
}
|
|
|
|
int numFaces = tri->numIndexes / 3;
|
|
int numShadowingFaces = 0;
|
|
facing = cullInfo.facing;
|
|
|
|
// if all the triangles are inside the light frustum
|
|
if ( cullInfo.cullBits == LIGHT_CULL_ALL_FRONT || !r_useShadowProjectedCull.GetBool() ) {
|
|
|
|
// count the number of shadowing faces
|
|
for ( i = 0; i < numFaces; i++ ) {
|
|
numShadowingFaces += facing[i];
|
|
}
|
|
numShadowingFaces = numFaces - numShadowingFaces;
|
|
|
|
} else {
|
|
|
|
// make all triangles that are outside the light frustum "facing", so they won't cast shadows
|
|
indexes = tri->indexes;
|
|
byte *modifyFacing = cullInfo.facing;
|
|
const byte *cullBits = cullInfo.cullBits;
|
|
for ( j = i = 0; i < tri->numIndexes; i += 3, j++ ) {
|
|
if ( !modifyFacing[j] ) {
|
|
int i1 = indexes[i+0];
|
|
int i2 = indexes[i+1];
|
|
int i3 = indexes[i+2];
|
|
if ( cullBits[i1] & cullBits[i2] & cullBits[i3] ) {
|
|
modifyFacing[j] = 1;
|
|
} else {
|
|
numShadowingFaces++;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
if ( !numShadowingFaces ) {
|
|
// no faces are inside the light frustum and still facing the right way
|
|
return NULL;
|
|
}
|
|
|
|
// shadowVerts will be NULL on these surfaces, so the shadowVerts will be taken from the ambient surface
|
|
newTri = R_AllocStaticTriSurf();
|
|
|
|
newTri->numVerts = tri->numVerts * 2;
|
|
|
|
// alloc the max possible size
|
|
#ifdef USE_TRI_DATA_ALLOCATOR
|
|
R_AllocStaticTriSurfIndexes( newTri, ( numShadowingFaces + tri->numSilEdges ) * 6 );
|
|
glIndex_t *tempIndexes = newTri->indexes;
|
|
glIndex_t *shadowIndexes = newTri->indexes;
|
|
#else
|
|
glIndex_t *tempIndexes = (glIndex_t *)_alloca16( tri->numSilEdges * 6 * sizeof( tempIndexes[0] ) );
|
|
glIndex_t *shadowIndexes = tempIndexes;
|
|
#endif
|
|
|
|
// create new triangles along sil planes
|
|
for ( sil = tri->silEdges, i = tri->numSilEdges; i > 0; i--, sil++ ) {
|
|
|
|
int f1 = facing[sil->p1];
|
|
int f2 = facing[sil->p2];
|
|
|
|
if ( !( f1 ^ f2 ) ) {
|
|
continue;
|
|
}
|
|
|
|
int v1 = sil->v1 << 1;
|
|
int v2 = sil->v2 << 1;
|
|
|
|
// set the two triangle winding orders based on facing
|
|
// without using a poorly-predictable branch
|
|
|
|
shadowIndexes[0] = v1;
|
|
shadowIndexes[1] = v2 ^ f1;
|
|
shadowIndexes[2] = v2 ^ f2;
|
|
shadowIndexes[3] = v1 ^ f2;
|
|
shadowIndexes[4] = v1 ^ f1;
|
|
shadowIndexes[5] = v2 ^ 1;
|
|
|
|
shadowIndexes += 6;
|
|
}
|
|
|
|
int numShadowIndexes = shadowIndexes - tempIndexes;
|
|
|
|
// we aren't bothering to separate front and back caps on these
|
|
newTri->numIndexes = newTri->numShadowIndexesNoFrontCaps = numShadowIndexes + numShadowingFaces * 6;
|
|
newTri->numShadowIndexesNoCaps = numShadowIndexes;
|
|
newTri->shadowCapPlaneBits = SHADOW_CAP_INFINITE;
|
|
|
|
#ifdef USE_TRI_DATA_ALLOCATOR
|
|
// decrease the size of the memory block to only store the used indexes
|
|
R_ResizeStaticTriSurfIndexes( newTri, newTri->numIndexes );
|
|
#else
|
|
// allocate memory for the indexes
|
|
R_AllocStaticTriSurfIndexes( newTri, newTri->numIndexes );
|
|
// copy the indexes we created for the sil planes
|
|
SIMDProcessor->Memcpy( newTri->indexes, tempIndexes, numShadowIndexes * sizeof( tempIndexes[0] ) );
|
|
#endif
|
|
|
|
// these have no effect, because they extend to infinity
|
|
newTri->bounds.Clear();
|
|
|
|
// put some faces on the model and some on the distant projection
|
|
indexes = tri->indexes;
|
|
shadowIndexes = newTri->indexes + numShadowIndexes;
|
|
for ( i = 0, j = 0; i < tri->numIndexes; i += 3, j++ ) {
|
|
if ( facing[j] ) {
|
|
continue;
|
|
}
|
|
|
|
int i0 = indexes[i+0] << 1;
|
|
shadowIndexes[2] = i0;
|
|
shadowIndexes[3] = i0 ^ 1;
|
|
int i1 = indexes[i+1] << 1;
|
|
shadowIndexes[1] = i1;
|
|
shadowIndexes[4] = i1 ^ 1;
|
|
int i2 = indexes[i+2] << 1;
|
|
shadowIndexes[0] = i2;
|
|
shadowIndexes[5] = i2 ^ 1;
|
|
|
|
shadowIndexes += 6;
|
|
}
|
|
|
|
return newTri;
|
|
}
|
|
|
|
/*
|
|
=====================
|
|
R_CreateTurboShadowVolume
|
|
=====================
|
|
*/
|
|
srfTriangles_t *R_CreateTurboShadowVolume( const idRenderEntityLocal *ent,
|
|
const srfTriangles_t *tri, const idRenderLightLocal *light,
|
|
srfCullInfo_t &cullInfo ) {
|
|
int i, j;
|
|
idVec3 localLightOrigin;
|
|
srfTriangles_t *newTri;
|
|
silEdge_t *sil;
|
|
const glIndex_t *indexes;
|
|
const byte *facing;
|
|
|
|
R_CalcInteractionFacing( ent, tri, light, cullInfo );
|
|
if ( r_useShadowProjectedCull.GetBool() ) {
|
|
R_CalcInteractionCullBits( ent, tri, light, cullInfo );
|
|
}
|
|
|
|
int numFaces = tri->numIndexes / 3;
|
|
int numShadowingFaces = 0;
|
|
facing = cullInfo.facing;
|
|
|
|
// if all the triangles are inside the light frustum
|
|
if ( cullInfo.cullBits == LIGHT_CULL_ALL_FRONT || !r_useShadowProjectedCull.GetBool() ) {
|
|
|
|
// count the number of shadowing faces
|
|
for ( i = 0; i < numFaces; i++ ) {
|
|
numShadowingFaces += facing[i];
|
|
}
|
|
numShadowingFaces = numFaces - numShadowingFaces;
|
|
|
|
} else {
|
|
|
|
// make all triangles that are outside the light frustum "facing", so they won't cast shadows
|
|
indexes = tri->indexes;
|
|
byte *modifyFacing = cullInfo.facing;
|
|
const byte *cullBits = cullInfo.cullBits;
|
|
for ( j = i = 0; i < tri->numIndexes; i += 3, j++ ) {
|
|
if ( !modifyFacing[j] ) {
|
|
int i1 = indexes[i+0];
|
|
int i2 = indexes[i+1];
|
|
int i3 = indexes[i+2];
|
|
if ( cullBits[i1] & cullBits[i2] & cullBits[i3] ) {
|
|
modifyFacing[j] = 1;
|
|
} else {
|
|
numShadowingFaces++;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
if ( !numShadowingFaces ) {
|
|
// no faces are inside the light frustum and still facing the right way
|
|
return NULL;
|
|
}
|
|
|
|
newTri = R_AllocStaticTriSurf();
|
|
|
|
#ifdef USE_TRI_DATA_ALLOCATOR
|
|
R_AllocStaticTriSurfShadowVerts( newTri, tri->numVerts * 2 );
|
|
shadowCache_t *shadowVerts = newTri->shadowVertexes;
|
|
#else
|
|
shadowCache_t *shadowVerts = (shadowCache_t *)_alloca16( tri->numVerts * 2 * sizeof( shadowVerts[0] ) );
|
|
#endif
|
|
|
|
R_GlobalPointToLocal( ent->modelMatrix, light->globalLightOrigin, localLightOrigin );
|
|
|
|
int *vertRemap = (int *)_alloca16( tri->numVerts * sizeof( vertRemap[0] ) );
|
|
|
|
SIMDProcessor->Memset( vertRemap, -1, tri->numVerts * sizeof( vertRemap[0] ) );
|
|
|
|
for ( i = 0, j = 0; i < tri->numIndexes; i += 3, j++ ) {
|
|
if ( facing[j] ) {
|
|
continue;
|
|
}
|
|
// this may pull in some vertexes that are outside
|
|
// the frustum, because they connect to vertexes inside
|
|
vertRemap[tri->silIndexes[i+0]] = 0;
|
|
vertRemap[tri->silIndexes[i+1]] = 0;
|
|
vertRemap[tri->silIndexes[i+2]] = 0;
|
|
}
|
|
|
|
newTri->numVerts = SIMDProcessor->CreateShadowCache( &shadowVerts->xyz, vertRemap, localLightOrigin, tri->verts, tri->numVerts );
|
|
|
|
c_turboUsedVerts += newTri->numVerts;
|
|
c_turboUnusedVerts += tri->numVerts * 2 - newTri->numVerts;
|
|
|
|
#ifdef USE_TRI_DATA_ALLOCATOR
|
|
R_ResizeStaticTriSurfShadowVerts( newTri, newTri->numVerts );
|
|
#else
|
|
R_AllocStaticTriSurfShadowVerts( newTri, newTri->numVerts );
|
|
SIMDProcessor->Memcpy( newTri->shadowVertexes, shadowVerts, newTri->numVerts * sizeof( shadowVerts[0] ) );
|
|
#endif
|
|
|
|
// alloc the max possible size
|
|
#ifdef USE_TRI_DATA_ALLOCATOR
|
|
R_AllocStaticTriSurfIndexes( newTri, ( numShadowingFaces + tri->numSilEdges ) * 6 );
|
|
glIndex_t *tempIndexes = newTri->indexes;
|
|
glIndex_t *shadowIndexes = newTri->indexes;
|
|
#else
|
|
glIndex_t *tempIndexes = (glIndex_t *)_alloca16( tri->numSilEdges * 6 * sizeof( tempIndexes[0] ) );
|
|
glIndex_t *shadowIndexes = tempIndexes;
|
|
#endif
|
|
|
|
// create new triangles along sil planes
|
|
for ( sil = tri->silEdges, i = tri->numSilEdges; i > 0; i--, sil++ ) {
|
|
|
|
int f1 = facing[sil->p1];
|
|
int f2 = facing[sil->p2];
|
|
|
|
if ( !( f1 ^ f2 ) ) {
|
|
continue;
|
|
}
|
|
|
|
int v1 = vertRemap[sil->v1];
|
|
int v2 = vertRemap[sil->v2];
|
|
|
|
// set the two triangle winding orders based on facing
|
|
// without using a poorly-predictable branch
|
|
|
|
shadowIndexes[0] = v1;
|
|
shadowIndexes[1] = v2 ^ f1;
|
|
shadowIndexes[2] = v2 ^ f2;
|
|
shadowIndexes[3] = v1 ^ f2;
|
|
shadowIndexes[4] = v1 ^ f1;
|
|
shadowIndexes[5] = v2 ^ 1;
|
|
|
|
shadowIndexes += 6;
|
|
}
|
|
|
|
int numShadowIndexes = shadowIndexes - tempIndexes;
|
|
|
|
// we aren't bothering to separate front and back caps on these
|
|
newTri->numIndexes = newTri->numShadowIndexesNoFrontCaps = numShadowIndexes + numShadowingFaces * 6;
|
|
newTri->numShadowIndexesNoCaps = numShadowIndexes;
|
|
newTri->shadowCapPlaneBits = SHADOW_CAP_INFINITE;
|
|
|
|
#ifdef USE_TRI_DATA_ALLOCATOR
|
|
// decrease the size of the memory block to only store the used indexes
|
|
R_ResizeStaticTriSurfIndexes( newTri, newTri->numIndexes );
|
|
#else
|
|
// allocate memory for the indexes
|
|
R_AllocStaticTriSurfIndexes( newTri, newTri->numIndexes );
|
|
// copy the indexes we created for the sil planes
|
|
SIMDProcessor->Memcpy( newTri->indexes, tempIndexes, numShadowIndexes * sizeof( tempIndexes[0] ) );
|
|
#endif
|
|
|
|
// these have no effect, because they extend to infinity
|
|
newTri->bounds.Clear();
|
|
|
|
// put some faces on the model and some on the distant projection
|
|
indexes = tri->silIndexes;
|
|
shadowIndexes = newTri->indexes + numShadowIndexes;
|
|
for ( i = 0, j = 0; i < tri->numIndexes; i += 3, j++ ) {
|
|
if ( facing[j] ) {
|
|
continue;
|
|
}
|
|
|
|
int i0 = vertRemap[indexes[i+0]];
|
|
shadowIndexes[2] = i0;
|
|
shadowIndexes[3] = i0 ^ 1;
|
|
int i1 = vertRemap[indexes[i+1]];
|
|
shadowIndexes[1] = i1;
|
|
shadowIndexes[4] = i1 ^ 1;
|
|
int i2 = vertRemap[indexes[i+2]];
|
|
shadowIndexes[0] = i2;
|
|
shadowIndexes[5] = i2 ^ 1;
|
|
|
|
shadowIndexes += 6;
|
|
}
|
|
|
|
return newTri;
|
|
}
|