quakeforge/libs/video/renderer/gl/gl_lightmap.c
Bill Currie 1e65ec22f9 [gl] Use a scrap for lightmaps
This gives a rather significant speed boost to timedemo demo1: from
about 2300-2360fps up to 2520-2600fps, at least when using
multi-texture.

Since it was necessary for testing the scrap, gl got the ability to set
the console background texture, too.
2022-05-11 00:33:47 +09:00

659 lines
16 KiB
C

/*
gl_lightmap.c
surface-related refresh code
Copyright (C) 1996-1997 Id Software, Inc.
Copyright (C) 2000 Joseph Carter <knghtbrd@debian.org>
This program 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 2
of the License, or (at your option) any later version.
This program 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 this program; if not, write to:
Free Software Foundation, Inc.
59 Temple Place - Suite 330
Boston, MA 02111-1307, USA
*/
#ifdef HAVE_CONFIG_H
# include "config.h"
#endif
#ifdef HAVE_STRING_H
# include <string.h>
#endif
#ifdef HAVE_STRINGS_H
# include <strings.h>
#endif
#include <math.h>
#include <stdio.h>
#include "QF/cvar.h"
#include "QF/render.h"
#include "QF/sys.h"
#include "QF/scene/entity.h"
#include "QF/GL/defines.h"
#include "QF/GL/funcs.h"
#include "QF/GL/qf_lightmap.h"
#include "QF/GL/qf_rmain.h"
#include "QF/GL/qf_rsurf.h"
#include "QF/GL/qf_sky.h"
#include "QF/GL/qf_textures.h"
#include "QF/GL/qf_vid.h"
#include "compat.h"
#include "r_internal.h"
static scrap_t *light_scrap;
static int dlightdivtable[8192];
static int gl_internalformat; // 1 or 3
static int lightmap_bytes; // 1, 3, or 4
GLuint gl_lightmap_textures[MAX_LIGHTMAPS];
// keep lightmap texture data in main memory so texsubimage can update properly
// LordHavoc: changed to be allocated at runtime (typically lower memory usage)
static byte *lightmaps[MAX_LIGHTMAPS];
static unsigned *blocklights;
static int bl_extents[2];
static int allocated[MAX_LIGHTMAPS][BLOCK_WIDTH];
qboolean gl_lightmap_modified[MAX_LIGHTMAPS];
instsurf_t *gl_lightmap_polys;
glRect_t gl_lightmap_rectchange[MAX_LIGHTMAPS];
static int lmshift = 7;
void (*gl_R_BuildLightMap) (const transform_t *transform, mod_brush_t *brush,
msurface_t *surf);
void
gl_lightmap_init (void)
{
int s;
memset (&lightmaps, 0, sizeof (lightmaps));
dlightdivtable[0] = 1048576 >> 7;
for (s = 1; s < 8192; s++)
dlightdivtable[s] = 1048576 / (s << 7);
}
static inline void
R_AddDynamicLights_1 (const transform_t *transform, msurface_t *surf)
{
float dist;
unsigned int maxdist, maxdist2, maxdist3;
int smax, smax_bytes, tmax,
grey, s, t;
unsigned int lnum, td, i, j;
unsigned int sdtable[18];
unsigned int *bl;
vec3_t impact, local;
vec4f_t entorigin = { 0, 0, 0, 1 };
smax = (surf->extents[0] >> 4) + 1;
smax_bytes = smax * gl_internalformat;
tmax = (surf->extents[1] >> 4) + 1;
if (transform) {
//FIXME give world entity a transform
entorigin = Transform_GetWorldPosition (transform);
}
for (lnum = 0; lnum < r_maxdlights; lnum++) {
if (!(surf->dlightbits[lnum / 32] & (1 << (lnum % 32))))
continue; // not lit by this light
VectorSubtract (r_dlights[lnum].origin, entorigin, local);
dist = DotProduct (local, surf->plane->normal) - surf->plane->dist;
VectorMultSub (r_dlights[lnum].origin, dist, surf->plane->normal,
impact);
i = DotProduct (impact, surf->texinfo->vecs[0]) +
surf->texinfo->vecs[0][3] - surf->texturemins[0];
// reduce calculations
t = dist * dist;
for (s = 0; s < smax; s++, i -= 16)
sdtable[s] = i * i + t;
i = DotProduct (impact, surf->texinfo->vecs[1]) +
surf->texinfo->vecs[1][3] - surf->texturemins[1];
// for comparisons to minimum acceptable light
maxdist = (int) (r_dlights[lnum].radius * r_dlights[lnum].radius);
// clamp radius to avoid exceeding 8192 entry division table
if (maxdist > 1048576)
maxdist = 1048576;
maxdist3 = maxdist - t;
// convert to 8.8 blocklights format
grey = (r_dlights[lnum].color[0] + r_dlights[lnum].color[1] +
r_dlights[lnum].color[2]) * maxdist / 3.0;
bl = blocklights;
for (t = 0; t < tmax; t++, i -= 16) {
td = i * i;
if (td < maxdist3) { // ensure part is visible on this line
maxdist2 = maxdist - td;
for (s = 0; s < smax; s++) {
if (sdtable[s] < maxdist2) {
j = dlightdivtable[(sdtable[s] + td) >> 7];
*bl++ += (grey * j) >> 7;
} else
bl++;
}
} else
bl += smax_bytes; // skip line
}
}
}
static inline void
R_AddDynamicLights_3 (const transform_t *transform, msurface_t *surf)
{
float dist;
unsigned int maxdist, maxdist2, maxdist3;
int smax, smax_bytes, tmax,
red, green, blue, s, t;
unsigned int lnum, td, i, j;
unsigned int sdtable[18];
unsigned int *bl;
vec3_t impact, local;
vec4f_t entorigin = { 0, 0, 0, 1 };
smax = (surf->extents[0] >> 4) + 1;
smax_bytes = smax * gl_internalformat;
tmax = (surf->extents[1] >> 4) + 1;
if (transform) {
entorigin = Transform_GetWorldPosition (transform);
}
for (lnum = 0; lnum < r_maxdlights; lnum++) {
if (!(surf->dlightbits[lnum / 32] & (1 << (lnum % 32))))
continue; // not lit by this light
VectorSubtract (r_dlights[lnum].origin, entorigin, local);
dist = DotProduct (local, surf->plane->normal) - surf->plane->dist;
VectorMultSub (r_dlights[lnum].origin, dist, surf->plane->normal,
impact);
i = DotProduct (impact, surf->texinfo->vecs[0]) +
surf->texinfo->vecs[0][3] - surf->texturemins[0];
// reduce calculations
t = dist * dist;
for (s = 0; s < smax; s++, i -= 16)
sdtable[s] = i * i + t;
i = DotProduct (impact, surf->texinfo->vecs[1]) +
surf->texinfo->vecs[1][3] - surf->texturemins[1];
// for comparisons to minimum acceptable light
maxdist = (int) (r_dlights[lnum].radius * r_dlights[lnum].radius);
// clamp radius to avoid exceeding 8192 entry division table
if (maxdist > 1048576)
maxdist = 1048576;
maxdist3 = maxdist - t;
// convert to 8.8 blocklights format
red = r_dlights[lnum].color[0] * maxdist;
green = r_dlights[lnum].color[1] * maxdist;
blue = r_dlights[lnum].color[2] * maxdist;
bl = blocklights;
for (t = 0; t < tmax; t++, i -= 16) {
td = i * i;
if (td < maxdist3) { // ensure part is visible on this line
maxdist2 = maxdist - td;
for (s = 0; s < smax; s++) {
if (sdtable[s] < maxdist2) {
j = dlightdivtable[(sdtable[s] + td) >> 7];
*bl++ += (red * j) >> 7;
*bl++ += (green * j) >> 7;
*bl++ += (blue * j) >> 7;
} else
bl += 3;
}
} else
bl += smax_bytes; // skip line
}
}
}
static void
R_BuildLightMap_1 (const transform_t *transform, mod_brush_t *brush,
msurface_t *surf)
{
byte *dest;
int maps, size, smax, tmax, i;
unsigned int scale;
unsigned int *bl;
surf->cached_dlight = (surf->dlightframe == r_framecount);
smax = (surf->extents[0] >> 4) + 1;
tmax = (surf->extents[1] >> 4) + 1;
size = smax * tmax * gl_internalformat;
// set to full bright if no light data
if (!brush->lightdata) {
memset (&blocklights[0], 0xff, size * sizeof(int));
goto store;
}
// clear to no light
memset (&blocklights[0], 0, size * sizeof(int));
// add all the lightmaps
if (surf->samples) {
byte *lightmap;
lightmap = surf->samples;
for (maps = 0; maps < MAXLIGHTMAPS && surf->styles[maps] != 255;
maps++) {
scale = d_lightstylevalue[surf->styles[maps]];
surf->cached_light[maps] = scale; // 8.8 fraction
bl = blocklights;
for (i = 0; i < size; i++) {
*bl++ += *lightmap++ * scale;
}
}
}
// add all the dynamic lights
if (surf->dlightframe == r_framecount)
R_AddDynamicLights_1 (transform, surf);
store:
// bound and shift
// Also, invert because we're using a diff blendfunc now
bl = blocklights;
dest = (byte *) blocklights;
for (i = 0; i < tmax * smax; i++) {
*dest++ = min (*bl >> lmshift, 255);
bl++;
}
GL_SubpicUpdate (surf->lightpic, (byte *) blocklights, 1);
}
static void
R_BuildLightMap_3 (const transform_t *transform, mod_brush_t *brush,
msurface_t *surf)
{
byte *dest;
int maps, size, smax, tmax, i;
unsigned int scale;
unsigned int *bl;
surf->cached_dlight = (surf->dlightframe == r_framecount);
smax = (surf->extents[0] >> 4) + 1;
tmax = (surf->extents[1] >> 4) + 1;
size = smax * tmax * gl_internalformat;
// set to full bright if no light data
if (!brush->lightdata) {
memset (&blocklights[0], 0xff, size * sizeof(int));
goto store;
}
// clear to no light
memset (&blocklights[0], 0, size * sizeof(int));
// add all the lightmaps
if (surf->samples) {
byte *lightmap;
lightmap = surf->samples;
for (maps = 0; maps < MAXLIGHTMAPS && surf->styles[maps] != 255;
maps++) {
scale = d_lightstylevalue[surf->styles[maps]];
surf->cached_light[maps] = scale; // 8.8 fraction
bl = blocklights;
for (i = 0; i < smax * tmax; i++) {
*bl++ += *lightmap++ * scale;
*bl++ += *lightmap++ * scale;
*bl++ += *lightmap++ * scale;
}
}
}
// add all the dynamic lights
if (surf->dlightframe == r_framecount)
R_AddDynamicLights_3 (transform, surf);
store:
// bound and shift
// and invert too
bl = blocklights;
dest = (byte *) blocklights;
for (i = 0; i < tmax * smax; i++) {
*dest++ = min (*bl >> lmshift, 255);
bl++;
*dest++ = min (*bl >> lmshift, 255);
bl++;
*dest++ = min (*bl >> lmshift, 255);
bl++;
}
GL_SubpicUpdate (surf->lightpic, (byte *) blocklights, 1);
}
static void
R_BuildLightMap_4 (const transform_t *transform, mod_brush_t *brush,
msurface_t *surf)
{
byte *dest;
int maps, size, smax, tmax, i;
unsigned int scale;
unsigned int *bl;
surf->cached_dlight = (surf->dlightframe == r_framecount);
smax = (surf->extents[0] >> 4) + 1;
tmax = (surf->extents[1] >> 4) + 1;
size = smax * tmax * gl_internalformat;
// set to full bright if no light data
if (!brush->lightdata) {
memset (&blocklights[0], 0xff, size * sizeof(int));
goto store;
}
// clear to no light
memset (&blocklights[0], 0, size * sizeof(int));
// add all the lightmaps
if (surf->samples) {
byte *lightmap;
lightmap = surf->samples;
for (maps = 0; maps < MAXLIGHTMAPS && surf->styles[maps] != 255;
maps++) {
scale = d_lightstylevalue[surf->styles[maps]];
surf->cached_light[maps] = scale; // 8.8 fraction
bl = blocklights;
for (i = 0; i < smax * tmax; i++) {
*bl++ += *lightmap++ * scale;
*bl++ += *lightmap++ * scale;
*bl++ += *lightmap++ * scale;
}
}
}
// add all the dynamic lights
if (surf->dlightframe == r_framecount)
R_AddDynamicLights_3 (transform, surf);
store:
// bound and shift
// and invert too
bl = blocklights;
dest = (byte *) blocklights;
for (i = 0; i < tmax * smax; i++) {
*dest++ = min (*bl >> lmshift, 255);
bl++;
*dest++ = min (*bl >> lmshift, 255);
bl++;
*dest++ = min (*bl >> lmshift, 255);
bl++;
*dest++ = 255;
}
GL_SubpicUpdate (surf->lightpic, (byte *) blocklights, 1);
}
// BRUSH MODELS ===============================================================
static inline void
do_subimage_2 (int i)
{
byte *block, *lm, *b;
int stride, width;
glRect_t *rect = &gl_lightmap_rectchange[i];
width = rect->w * lightmap_bytes;
stride = BLOCK_WIDTH * lightmap_bytes;
b = block = Hunk_TempAlloc (0, rect->h * width);
lm = lightmaps[i] + (rect->t * BLOCK_WIDTH + rect->l) * lightmap_bytes;
for (i = rect->h; i > 0; i--) {
memcpy (b, lm, width);
b += width;
lm += stride;
}
qfglTexSubImage2D (GL_TEXTURE_2D, 0, rect->l, rect->t, rect->w, rect->h,
gl_lightmap_format, GL_UNSIGNED_BYTE, block);
}
void
gl_R_BlendLightmaps (void)
{
float *v;
instsurf_t *sc;
glpoly_t *p;
qfglDepthMask (GL_FALSE); // don't bother writing Z
qfglBlendFunc (lm_src_blend, lm_dest_blend);
qfglBindTexture (GL_TEXTURE_2D, gl_R_LightmapTexture ());
for (sc = gl_lightmap_polys; sc; sc = sc->lm_chain) {
if (sc->transform) {
qfglPushMatrix ();
qfglLoadMatrixf (sc->transform);
}
for (p = sc->surface->polys; p; p = p->next) {
qfglBegin (GL_POLYGON);
v = p->verts[0];
for (int j = 0; j < p->numverts; j++, v += VERTEXSIZE) {
qfglTexCoord2fv (&v[5]);
qfglVertex3fv (v);
}
qfglEnd ();
}
if (sc->transform)
qfglPopMatrix ();
}
// Return to normal blending --KB
qfglBlendFunc (GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA);
qfglDepthMask (GL_TRUE); // back to normal Z buffering
}
void
gl_overbright_f (void *data, const cvar_t *cvar)
{
mod_brush_t *brush;
if (!cvar)
return;
if (gl_overbright) {
if (!gl_combine_capable && gl_mtex_capable) {
Sys_Printf ("Warning: gl_overbright has no effect with "
"gl_multitexture enabled if you don't have "
"GL_COMBINE support in your driver.\n");
lm_src_blend = GL_ZERO;
lm_dest_blend = GL_SRC_COLOR;
lmshift = 7;
gl_rgb_scale = 1.0;
} else {
lm_src_blend = GL_DST_COLOR;
lm_dest_blend = GL_SRC_COLOR;
switch (gl_overbright) {
case 2:
lmshift = 9;
gl_rgb_scale = 4.0;
break;
case 1:
lmshift = 8;
gl_rgb_scale = 2.0;
break;
default:
lmshift = 7;
gl_rgb_scale = 1.0;
break;
}
}
} else {
lm_src_blend = GL_ZERO;
lm_dest_blend = GL_SRC_COLOR;
lmshift = 7;
gl_rgb_scale = 1.0;
}
if (!gl_R_BuildLightMap)
return;
brush = &r_refdef.worldmodel->brush;
for (unsigned i = 0; i < brush->numsurfaces; i++) {
msurface_t *surf = brush->surfaces + i;
if (surf->flags & (SURF_DRAWTURB | SURF_DRAWSKY))
continue;
gl_R_BuildLightMap (0, brush, surf);
}
}
// LIGHTMAP ALLOCATION ========================================================
static void
GL_CreateSurfaceLightmap (mod_brush_t *brush, msurface_t *surf)
{
int smax, tmax;
if (surf->flags & (SURF_DRAWSKY | SURF_DRAWTURB))
return;
smax = (surf->extents[0] >> 4) + 1;
tmax = (surf->extents[1] >> 4) + 1;
surf->lightpic = GL_ScrapSubpic (light_scrap, smax, tmax);
if (!surf->lightpic) {
Sys_Error ("FIXME taniwha is being lazy");
}
if (smax > bl_extents[0]) {
bl_extents[0] = smax;
}
if (tmax > bl_extents[1]) {
bl_extents[1] = tmax;
}
}
/*
GL_BuildLightmaps
Builds the lightmap texture with all the surfaces from all brush models
*/
void
GL_BuildLightmaps (model_t **models, int num_models)
{
model_t *m;
mod_brush_t *brush;
memset (allocated, 0, sizeof (allocated));
r_framecount = 1; // no dlightcache
switch (r_lightmap_components) {
case 1:
gl_internalformat = 1;
gl_lightmap_format = GL_LUMINANCE;
lightmap_bytes = 1;
gl_R_BuildLightMap = R_BuildLightMap_1;
break;
case 3:
gl_internalformat = 3;
if (gl_use_bgra)
gl_lightmap_format = GL_BGR;
else
gl_lightmap_format = GL_RGB;
lightmap_bytes = 3;
gl_R_BuildLightMap = R_BuildLightMap_3;
break;
case 4:
default:
gl_internalformat = 3;
if (gl_use_bgra)
gl_lightmap_format = GL_BGRA;
else
gl_lightmap_format = GL_RGBA;
lightmap_bytes = 4;
gl_R_BuildLightMap = R_BuildLightMap_4;
break;
}
if (!light_scrap) {
light_scrap = GL_CreateScrap (2048, gl_lightmap_format, 1);
} else {
GL_ScrapClear (light_scrap);
}
bl_extents[1] = bl_extents[0] = 0;
for (int j = 1; j < num_models; j++) {
m = models[j];
if (!m)
break;
if (m->path[0] == '*' || m->type != mod_brush) {
// sub model surfaces are processed as part of the main model
continue;
}
brush = &m->brush;
gl_currentmodel = m;
// non-bsp models don't have surfaces.
for (unsigned i = 0; i < brush->numsurfaces; i++) {
if (brush->surfaces[i].flags & (SURF_DRAWTURB | SURF_DRAWSKY))
continue;
GL_CreateSurfaceLightmap (brush, brush->surfaces + i);
GL_BuildSurfaceDisplayList (brush, brush->surfaces + i);
}
}
int size = bl_extents[0] * bl_extents[1] * 3; // * 3 for rgb support
blocklights = realloc (blocklights, size * sizeof (blocklights[0]));
// upload all lightmaps that were filled
for (int j = 1; j < num_models; j++) {
m = models[j];
if (!m)
break;
if (m->path[0] == '*' || m->type != mod_brush) {
// sub model surfaces are processed as part of the main model
continue;
}
brush = &m->brush;
// non-bsp models don't have surfaces.
for (unsigned i = 0; i < brush->numsurfaces; i++) {
msurface_t *surf = brush->surfaces + i;
if (surf->lightpic) {
gl_R_BuildLightMap (0, brush, surf);
}
}
}
}
int
gl_R_LightmapTexture (void)
{
return GL_ScrapTexture (light_scrap);
}
void
gl_R_FlushLightmaps (void)
{
GL_ScrapFlush (light_scrap);
}