* try to setup advance gl context attribute. If driver doesn't support it fallback to the default.

zzogl:
* use glsl2h to generate an header shader as GSdx. Much easier to install
* Get GSUniformBufferOGL & GSVertexArrayOGL speed improvement from GSdx

cmake:
* detect current gcc version. Yield  a warning if < 4.7 or an error if < 4.5

glsl2h:
* support zzogl too
* compute md5sum to avoid useless relink


git-svn-id: http://pcsx2.googlecode.com/svn/trunk@5696 96395faa-99c1-11dd-bbfe-3dabce05a288
This commit is contained in:
gregory.hainaut 2013-07-06 09:42:46 +00:00
parent ba2ec6fe59
commit 544c84a344
18 changed files with 1081 additions and 129 deletions

View File

@ -90,6 +90,8 @@ include(Pcsx2Utils)
# Detect current OS
detectOperatingSystem()
check_compiler_version("4.7" "4.5")
#-------------------------------------------------------------------------------
# Include specific module
# BuildParameters Must be done before SearchForStuff
@ -109,6 +111,9 @@ include_directories(${PROJECT_SOURCE_DIR}/common/include
${CMAKE_BINARY_DIR}/common/include
# WORKAROUND Some issue with multiarch on Debian/Ubuntu
/usr/include/i386-linux-gnu
# WORKAROUND Clang integration issue with multiarch on Debian
#/usr/include/i386-linux-gnu/c++
#/usr/include/i386-linux-gnu/c++/4.8
)
# make the translation

View File

@ -50,3 +50,19 @@ function(write_svnrev_h)
file(WRITE ${CMAKE_BINARY_DIR}/common/include/svnrev.h "#define SVN_REV 0 \n#define SVN_MODS 0")
endif()
endfunction()
function(check_compiler_version version_warn version_err)
if("${CMAKE_CXX_COMPILER_ID}" MATCHES "GNU")
execute_process(COMMAND ${CMAKE_C_COMPILER} -dumpversion OUTPUT_VARIABLE GCC_VERSION)
string(STRIP "${GCC_VERSION}" GCC_VERSION)
if(GCC_VERSION VERSION_LESS ${version_err})
message(FATAL_ERROR "PCSX2 doesn't support your old GCC ${GCC_VERSION}! Please upgrade it !
The minimum version is ${version_err} but ${version_warn} is warmly recommended")
else()
if(GCC_VERSION VERSION_LESS ${version_warn})
message(WARNING "PCSX2 will stop to support GCC ${GCC_VERSION} in a near future. Please upgrade it to GCC ${version_warn}.")
endif()
endif()
endif()
endfunction()

View File

@ -4,8 +4,12 @@ use strict;
use warnings;
use File::Spec;
use File::Basename;
use File::Copy;
use Cwd 'abs_path';
use Digest::file qw/digest_file_hex/;
use Digest::MD5 qw(md5_hex);
my @res = qw/convert interlace merge shadeboost tfx/;
my $path = File::Spec->catdir(dirname(abs_path($0)), "..", "plugins", "GSdx", "res");
@ -14,6 +18,9 @@ foreach my $r (@res) {
}
glsl2h($path, "fxaa", "fx");
my $zz_path = File::Spec->catdir(dirname(abs_path($0)), "..", "plugins", "zzogl-pg", "opengl");
glsl2h($zz_path, "ps2hw_gl4", "glsl");
sub glsl2h {
my $path = shift;
my $glsl = shift;
@ -22,13 +29,16 @@ sub glsl2h {
my $in = File::Spec->catfile($path, "${glsl}.$ext");
my $out = File::Spec->catfile($path, "${glsl}.h");
open(my $GLSL, "<$in") or die;
open(my $H, ">$out") or die;
my $data = "";
my $include = "";
if ($in =~ /GSdx/) {
$include = "#include \"stdafx.h\""
}
my $header = <<EOS;
/*
* Copyright (C) 2011-2013 Gregory hainaut
* Copyright (C) 2007-2009 Gabest
*
* This file was generated by glsl2h.pl script
*
* This Program is free software; you can redistribute it and/or modify
@ -50,18 +60,29 @@ sub glsl2h {
#pragma once
#include "stdafx.h"
$include
static const char* ${glsl}_${ext} =
EOS
print $H $header;
$data = $header;
my $line;
while(defined($line = <$GLSL>)) {
chomp $line;
$line =~ s/\\/\\\\/g;
$line =~ s/"/\\"/g;
print $H "\t\"$line\\n\"\n";
$data .= "\t\"$line\\n\"\n";
}
$data .= "\t;\n";
# Rewriting the file will trigger a relink (even if the content is the
# same). So we check first the content with md5 digest
my $old_md5 = digest_file_hex($out, "MD5");
my $new_md5 = md5_hex($data);
if ($old_md5 ne $new_md5) {
open(my $H, ">$out") or die;
print $H $data;
}
print $H "\t;\n";
}

View File

@ -1272,12 +1272,8 @@ void GSDeviceOGL::CompileShaderFromSource(const std::string& glsl_file, const st
}
if (GLLoader::found_GL_ARB_separate_shader_objects) {
version += "#extension GL_ARB_separate_shader_objects : require\n";
// REMOVE ME: Emulate open source driver
//if (!GLLoader::found_GL_ARB_shading_language_420pack) {
// version += "#define NO_STRUCT 1\n";
//}
} else {
if (GLLoader::found_only_gl30)
if (!GLLoader::fglrx_buggy_driver)
version += "#define DISABLE_SSO\n";
}
if (GLLoader::found_only_gl30) {

View File

@ -35,17 +35,15 @@ void GSWndEGL::CreateContext(int major, int minor)
EGLint numConfigs;
EGLint contextAttribs[] =
{
// Not yet supported by Radeon/Gallium
#if 0
EGL_CONTEXT_MAJOR_VERSION_KHR, major,
EGL_CONTEXT_MINOR_VERSION_KHR, minor,
// Keep compatibility for old cruft
//EGL_CONTEXT_OPENGL_PROFILE_MASK_KHR, EGL_CONTEXT_OPENGL_COMPATIBILITY_PROFILE_BIT_KHR,
// FIXME : Request a debug context to ease opengl development
EGL_CONTEXT_FLAGS_KHR, EGL_CONTEXT_OPENGL_DEBUG_BIT_KHR | EGL_CONTEXT_OPENGL_FORWARD_COMPATIBLE_BIT_KHR,
#endif
EGL_NONE
};
EGLint NullContextAttribs[] = { EGL_NONE };
EGLint attrList[] = {
EGL_RED_SIZE, 8,
EGL_GREEN_SIZE, 8,
@ -70,11 +68,18 @@ void GSWndEGL::CreateContext(int major, int minor)
throw GSDXRecoverableError();
}
m_eglContext = eglCreateContext(m_eglDisplay, eglConfig, EGL_NO_CONTEXT, contextAttribs );
m_eglContext = eglCreateContext(m_eglDisplay, eglConfig, EGL_NO_CONTEXT, contextAttribs);
EGLint status = eglGetError();
if (status == EGL_BAD_ATTRIBUTE) {
// Radeon/Gallium don't support advance attribute. Fallback to random value
fprintf(stderr, "EGL: warning your driver doesn't suport advance openGL context attributes\n");
m_eglContext = eglCreateContext(m_eglDisplay, eglConfig, EGL_NO_CONTEXT, NullContextAttribs);
status = eglGetError();
}
if ( m_eglContext == EGL_NO_CONTEXT )
{
fprintf(stderr,"EGL: Failed to create the context\n");
fprintf(stderr,"EGL STATUS: %x\n", eglGetError());
fprintf(stderr,"EGL STATUS: %x\n", status);
throw GSDXRecoverableError();
}

View File

@ -1,7 +1,4 @@
/*
* Copyright (C) 2011-2013 Gregory hainaut
* Copyright (C) 2007-2009 Gabest
*
* This file was generated by glsl2h.pl script
*
* This Program is free software; you can redistribute it and/or modify

View File

@ -1,7 +1,4 @@
/*
* Copyright (C) 2011-2013 Gregory hainaut
* Copyright (C) 2007-2009 Gabest
*
* This file was generated by glsl2h.pl script
*
* This Program is free software; you can redistribute it and/or modify

View File

@ -1,7 +1,4 @@
/*
* Copyright (C) 2011-2013 Gregory hainaut
* Copyright (C) 2007-2009 Gabest
*
* This file was generated by glsl2h.pl script
*
* This Program is free software; you can redistribute it and/or modify

View File

@ -1,7 +1,4 @@
/*
* Copyright (C) 2011-2013 Gregory hainaut
* Copyright (C) 2007-2009 Gabest
*
* This file was generated by glsl2h.pl script
*
* This Program is free software; you can redistribute it and/or modify

View File

@ -1,7 +1,4 @@
/*
* Copyright (C) 2011-2013 Gregory hainaut
* Copyright (C) 2007-2009 Gabest
*
* This file was generated by glsl2h.pl script
*
* This Program is free software; you can redistribute it and/or modify

View File

@ -1,7 +1,4 @@
/*
* Copyright (C) 2011-2013 Gregory hainaut
* Copyright (C) 2007-2009 Gabest
*
* This file was generated by glsl2h.pl script
*
* This Program is free software; you can redistribute it and/or modify

View File

@ -110,6 +110,7 @@ set(zzoglHeaders
GS.h
GSDump.h
HostMemory.h
ps2hw_gl4.h
Mem.h
Mem_Swizzle.h
Mem_Transmit.h
@ -153,6 +154,9 @@ set(zzoglLinuxHeaders
include_directories(.
Linux)
# Generate Glsl header file
add_custom_command(OUTPUT ps2hw_gl4.h COMMAND perl ${PROJECT_SOURCE_DIR}/linux_various/glsl2h.pl)
# add library
add_library(${Output} SHARED
${zzoglSources}
@ -167,32 +171,24 @@ set_target_properties(${Output} PROPERTIES COMPILE_DEFINITIONS USE_GSOPEN2)
# link target with project internal libraries
target_link_libraries(${Output} Utilities)
# link target with Cg
if(NOT GLSL_API)
target_link_libraries(${Output} ${CG_LIBRARIES})
endif(NOT GLSL_API)
# link target with glew
target_link_libraries(${Output} ${GLEW_LIBRARY})
# link target with opengl
# link target with the various opengl flavor
target_link_libraries(${Output} ${OPENGL_LIBRARIES})
if(EGL_API AND EGL_FOUND AND EGL_GL_CONTEXT_SUPPORT)
target_link_libraries(${Output} ${EGL_LIBRARIES})
endif()
if(NOT GLSL_API)
target_link_libraries(${Output} ${CG_LIBRARIES})
endif(NOT GLSL_API)
target_link_libraries(${Output} ${GLEW_LIBRARY})
# link target with X11
target_link_libraries(${Output} ${X11_LIBRARIES})
# link target with jpeg
target_link_libraries(${Output} ${JPEG_LIBRARIES})
if(Linux)
# link target with gtk2
target_link_libraries(${Output} ${GTK2_LIBRARIES})
endif(Linux)
# link target with zlib
target_link_libraries(${Output} ${ZLIB_LIBRARIES})
# User flags options
@ -202,25 +198,17 @@ endif(NOT USER_CMAKE_LD_FLAGS STREQUAL "")
if(PACKAGE_MODE)
install(TARGETS ${Output} DESTINATION ${PLUGIN_DIR})
if(GLSL_API)
install(FILES ${PROJECT_SOURCE_DIR}/plugins/zzogl-pg/opengl/ps2hw.glsl DESTINATION ${GLSL_SHADER_DIR})
install(FILES ${PROJECT_SOURCE_DIR}/plugins/zzogl-pg/opengl/ps2hw_gl4.glsl DESTINATION ${GLSL_SHADER_DIR})
else(GLSL_API)
if(NOT REBUILD_SHADER)
install(FILES ${PROJECT_SOURCE_DIR}/plugins/zzogl-pg/opengl/ps2hw.dat DESTINATION ${PLUGIN_DIR})
endif(NOT REBUILD_SHADER)
endif(GLSL_API)
else(PACKAGE_MODE)
else()
install(TARGETS ${Output} DESTINATION ${CMAKE_SOURCE_DIR}/bin/plugins)
if(GLSL_API)
install(FILES ${PROJECT_SOURCE_DIR}/plugins/zzogl-pg/opengl/ps2hw.glsl DESTINATION ${CMAKE_SOURCE_DIR}/bin/plugins)
install(FILES ${PROJECT_SOURCE_DIR}/plugins/zzogl-pg/opengl/ps2hw_gl4.glsl DESTINATION ${CMAKE_SOURCE_DIR}/bin/plugins)
else(GLSL_API)
if(NOT REBUILD_SHADER)
install(FILES ${PROJECT_SOURCE_DIR}/plugins/zzogl-pg/opengl/ps2hw.dat DESTINATION ${CMAKE_SOURCE_DIR}/bin/plugins)
endif(NOT REBUILD_SHADER)
endif(GLSL_API)
endif(PACKAGE_MODE)
endif()
if(NOT GLSL_API AND NOT REBUILD_SHADER)
if(PACKAGE_MODE)
install(FILES ${PROJECT_SOURCE_DIR}/plugins/zzogl-pg/opengl/ps2hw.dat DESTINATION ${PLUGIN_DIR})
else()
install(FILES ${PROJECT_SOURCE_DIR}/plugins/zzogl-pg/opengl/ps2hw.dat DESTINATION ${CMAKE_SOURCE_DIR}/bin/plugins)
endif()
endif()
################################### Replay Loader
if(BUILD_REPLAY_LOADERS)

View File

@ -58,10 +58,14 @@ public:
void upload(const void* src)
{
uint32 flags = GL_MAP_WRITE_BIT | GL_MAP_INVALIDATE_BUFFER_BIT;
uint8* dst = (uint8*) glMapBufferRange(target, 0, size, flags);
memcpy(dst, src, size);
glUnmapBuffer(target);
// uint32 flags = GL_MAP_WRITE_BIT | GL_MAP_INVALIDATE_BUFFER_BIT;
// uint8* dst = (uint8*) glMapBufferRange(target, 0, size, flags);
// memcpy(dst, src, size);
// glUnmapBuffer(target);
// glMapBufferRange allow to set various parameter but the call is
// synchronous whereas glBufferSubData could be asynchronous.
// TODO: investigate the extension ARB_invalidate_subdata
glBufferSubData(target, 0, size, src);
}
~GSUniformBufferOGL() {

View File

@ -34,13 +34,12 @@ struct GSInputLayoutOGL {
};
class GSBufferOGL {
size_t m_stride;
const size_t m_stride;
size_t m_start;
size_t m_count;
size_t m_limit;
GLenum m_target;
const GLenum m_target;
GLuint m_buffer;
size_t m_default_size;
public:
GSBufferOGL(GLenum target, size_t stride) :
@ -52,12 +51,12 @@ class GSBufferOGL {
{
glGenBuffers(1, &m_buffer);
// Opengl works best with 1-4MB buffer.
m_default_size = 2 * 1024 * 1024 / m_stride;
m_limit = 2 * 1024 * 1024 / m_stride;
}
~GSBufferOGL() { glDeleteBuffers(1, &m_buffer); }
void allocate() { allocate(m_default_size); }
void allocate() { allocate(m_limit); }
void allocate(size_t new_limit)
{
@ -73,6 +72,7 @@ class GSBufferOGL {
void upload(const void* src, uint32 count)
{
#if 0
// Upload the data to the buffer
void* dst;
if (Map(&dst, count)) {
@ -81,15 +81,25 @@ class GSBufferOGL {
memcpy(dst, src, m_stride*m_count);
Unmap();
}
#else
m_count = count;
// Current GPU buffer is really too small need to allocate a new one
if (m_count > m_limit) {
allocate(std::max<int>(m_count * 3 / 2, m_limit));
} else if (m_count > (m_limit - m_start) ) {
// Not enough left free room. Just go back at the beginning
m_start = 0;
// Orphan the buffer to avoid synchronization
allocate(m_limit);
}
glBufferSubData(m_target, m_stride * m_start, m_stride * m_count, src);
#endif
}
bool Map(void** pointer, uint32 count ) {
#ifdef OGL_DEBUG
GLint b_size = -1;
glGetBufferParameteriv(m_target, GL_BUFFER_SIZE, &b_size);
if (b_size <= 0) return false;
#endif
m_count = count;
@ -99,7 +109,7 @@ class GSBufferOGL {
// Current GPU buffer is really too small need to allocate a new one
if (m_count > m_limit) {
allocate(std::max<int>(m_count * 3 / 2, m_default_size));
allocate(std::max<int>(m_count * 3 / 2, m_limit));
} else if (m_count > (m_limit - m_start) ) {
// Not enough left free room. Just go back at the beginning
@ -116,13 +126,7 @@ class GSBufferOGL {
// Upload the data to the buffer
*pointer = (uint8*) glMapBufferRange(m_target, m_stride*m_start, m_stride*m_count, map_flags);
//fprintf(stderr, "Map %x from %d to %d\n", *pointer, m_start, m_start+m_count);
#ifdef OGL_DEBUG
if (*pointer == NULL) {
fprintf(stderr, "CRITICAL ERROR map failed for vb!!!\n");
return false;
}
#endif
return true;
}

View File

@ -457,10 +457,7 @@ bool ZZCreate(int _width, int _height)
GPU_TEXWIDTH = min (g_MaxTexWidth/8, 1024);
g_fiGPU_TEXWIDTH = 1.0f / GPU_TEXWIDTH;
// FIXME: not clean maybe re integrate the function in shader files --greg
#if defined(GLSL_API) || defined(GLSL4_API)
if (!ZZshCreateOpenShadersFile()) return false;
#else
#if !(defined(GLSL_API) || defined(GLSL4_API))
if (!CreateOpenShadersFile()) return false;
#endif

View File

@ -50,7 +50,7 @@
#include "zpipe.h"
#include <map>
#include <fcntl.h> // this for open(). Maybe linux-specific
#include <sys/mman.h> // and this for mmap
#include "ps2hw_gl4.h"
// ----------------- Defines
@ -88,9 +88,6 @@ u8* s_lpShaderResources = NULL;
ZZshShaderLink pvs[16] = {sZero}, g_vsprog = sZero, g_psprog = sZero; // 2 -- ZZ
ZZshParameter g_vparamPosXY[2] = {pZero}, g_fparamFogColor = pZero;
char* ZZshSource; // Shader's source data.
off_t ZZshSourceSize;
bool g_bCRTCBilinear = true;
float4 g_vdepth, vlogz;
@ -175,34 +172,10 @@ bool ZZshStartUsingShaders() {
// open shader file according to build target
bool ZZshCreateOpenShadersFile() {
std::string ShaderFileName("plugins/ps2hw_gl4.glsl");
int ShaderFD = open(ShaderFileName.c_str(), O_RDONLY);
struct stat sb;
if ((ShaderFD == -1) || (fstat(ShaderFD, &sb) == -1)) {
// Each linux distributions have his rules for path so we give them the possibility to
// change it with compilation flags. -- Gregory
#ifdef GLSL_SHADER_DIR_COMPILATION
#define xGLSL_SHADER_DIR_str(s) GLSL_SHADER_DIR_str(s)
#define GLSL_SHADER_DIR_str(s) #s
ShaderFileName = string(xGLSL_SHADER_DIR_str(GLSL_SHADER_DIR_COMPILATION)) + "/ps2hw_gl4.glsl";
ShaderFD = open(ShaderFileName.c_str(), O_RDONLY);
#endif
if ((ShaderFD == -1) || (fstat(ShaderFD, &sb) == -1)) {
ZZLog::Error_Log("No source for %s: \n", ShaderFileName.c_str());
return false;
}
}
ZZshSourceSize = sb.st_size;
ZZshSource = (char*)mmap(NULL, sb.st_size, PROT_READ | PROT_WRITE, MAP_PRIVATE, ShaderFD, 0); // This function directly maped file into memory.
ZZshSource[ ZZshSourceSize - 1] = 0; // Made source null-terminated.
close(ShaderFD);
return true;
}
void ZZshExitCleaning() {
munmap(ZZshSource, ZZshSourceSize);
delete constant_buffer;
delete common_buffer;
delete vertex_buffer;
@ -299,10 +272,25 @@ static bool ValidateProgram(ZZshProgram Prog) {
char* InfoLog = new char[infologlength];
glGetProgramInfoLog(Prog, infologlength, &lenght, InfoLog);
ZZLog::Error_Log("Validation %d... %d:\t %s", Prog, infologlength, InfoLog);
delete[] InfoLog;
}
return (isValid != 0);
}
static void ValidatePipeline(GLuint pipeline) {
glValidateProgramPipeline(pipeline);
GLint isValid;
glGetProgramPipelineiv(pipeline, GL_VALIDATE_STATUS, &isValid);
if (!isValid) {
int lenght, infologlength;
glGetProgramPipelineiv(pipeline, GL_INFO_LOG_LENGTH, &infologlength);
char* InfoLog = new char[infologlength];
glGetProgramPipelineInfoLog(pipeline, infologlength, &lenght, InfoLog);
ZZLog::Error_Log("Validation %d... %d:\t %s", pipeline, infologlength, InfoLog);
delete[] InfoLog;
}
}
inline bool CompileShaderFromFile(ZZshProgram& program, const std::string& DefineString, std::string main_entry, GLenum ShaderType)
{
std::string header("");
@ -315,10 +303,18 @@ inline bool CompileShaderFromFile(ZZshProgram& program, const std::string& Defin
const GLchar* ShaderSource[2];
#if 0
// It sucks because it doesn't report the good line for error/warnings!
// But at least this stupid AMD drivers doesn't crash...
ShaderSource[0] = header.append(ps2hw_gl4_glsl).c_str();
program = glCreateShaderProgramv(ShaderType, 1, &ShaderSource[0]);
#else
ShaderSource[0] = header.c_str();
ShaderSource[1] = (const GLchar*)ZZshSource;
ShaderSource[1] = ps2hw_gl4_glsl;
program = glCreateShaderProgramv(ShaderType, 2, &ShaderSource[0]);
#endif
ZZLog::Debug_Log("Creating program %d for %s", program, main_entry.c_str());
@ -343,10 +339,7 @@ void ZZshSetupShader() {
// shaders given the current GL state"
// It might be a good idea to validate the pipeline also in release mode???
#if defined(DEVBUILD) || defined(_DEBUG)
glValidateProgramPipeline(s_pipeline);
GLint isValid;
glGetProgramPipelineiv(s_pipeline, GL_VALIDATE_STATUS, &isValid);
if (!isValid) ZZLog::Error_Log("Something weird happened on pipeline validation.");
ValidatePipeline(s_pipeline);
#endif
PutParametersInProgram(vs, ps);

View File

@ -58,8 +58,8 @@
struct vertex
{
vec4 color;
TEX_DECL tex;
vec4 Z;
TEX_DECL tex;
float fog;
};

View File

@ -0,0 +1,941 @@
/*
* This file was generated by glsl2h.pl script
*
* 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, 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 GNU Make; see the file COPYING. If not, write to
* the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA USA.
* http://www.gnu.org/copyleft/gpl.html
*
*/
#pragma once
static const char* ps2hw_gl4_glsl =
"//#version 420 Keep it for text editor detection\n"
"\n"
"// ZZ Open GL graphics plugin\n"
"// Copyright (c)2009-2010 zeydlitz@gmail.com, arcum42@gmail.com, gregory.hainaut@gmail.com\n"
"// Based on Zerofrog's ZeroGS KOSMOS (c)2005-2008\n"
"//\n"
"// This program is free software; you can redistribute it and/or modify\n"
"// it under the terms of the GNU General Public License as published by\n"
"// the Free Software Foundation; either version 2 of the License, or\n"
"// (at your option) any later version.\n"
"//\n"
"// This program is distributed in the hope that it will be useful,\n"
"// but WITHOUT ANY WARRANTY; without even the implied warranty of\n"
"// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the\n"
"// GNU General Public License for more details.\n"
"//\n"
"// You should have received a copy of the GNU General Public License\n"
"// along with this program; if not, write to the Free Software\n"
"// Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA\n"
"\n"
"// divides by z for every pixel, instead of in vertex shader\n"
"// fixes kh textures\n"
"\n"
"#extension GL_ARB_texture_rectangle: require\n"
"#extension GL_ARB_shading_language_420pack: require\n"
"#extension GL_ARB_separate_shader_objects : require\n"
"\n"
"\n"
"//#define TEST_AEM // tests AEM for black pixels\n"
"//#define REGION_REPEAT // set if texture wrapping mode is region repeat\n"
"//#define WRITE_DEPTH // set if depth is also written in a MRT\n"
"//#define ACCURATE_DECOMPRESSION // set for less capable hardware ATI Radeon 9000 series\n"
"//#define EXACT_COLOR // make sure the output color is clamped to 1/255 boundaries (for alpha testing)\n"
"//#define NO_LOGZ // disable logz\n"
"\n"
"#define PERSPECTIVE_CORRECT_TEX\n"
"\n"
"#ifdef PERSPECTIVE_CORRECT_TEX\n"
"#define TEX_XY tex.xy/tex.z\n"
"#define TEX_DECL vec3\n"
"#else\n"
"#define TEX_XY tex.xy\n"
"#define TEX_DECL vec2\n"
"#endif\n"
"\n"
"// NVidia CG-data types\n"
"#define half2 vec2\n"
"#define half3 vec3\n"
"#define half4 vec4\n"
"#define float2 vec2\n"
"#define float3 vec3\n"
"#define float4 vec4\n"
"\n"
"////////////////////////////////////////////////////////////////////\n"
"// INPUT/OUTPUT\n"
"////////////////////////////////////////////////////////////////////\n"
"// NOTE: Future optimization tex.w is normally useless (in cg it is a float3) so it can contains the fog value\n"
"struct vertex\n"
"{\n"
" vec4 color;\n"
" vec4 Z;\n"
" TEX_DECL tex;\n"
" float fog;\n"
"};\n"
"\n"
"#ifdef VERTEX_SHADER\n"
"out gl_PerVertex {\n"
" invariant vec4 gl_Position;\n"
" float gl_PointSize;\n"
" float gl_ClipDistance[];\n"
"};\n"
"\n"
"layout(location = 0) in ivec4 Vert;\n"
"layout(location = 1) in vec4 Color;\n"
"layout(location = 2) in vec4 SecondaryColor;\n"
"layout(location = 3) in vec3 TexCoord;\n"
"\n"
"layout(location = 0) out vertex VSout;\n"
"\n"
"#endif\n"
"\n"
"#ifdef FRAGMENT_SHADER\n"
"\n"
"layout(location = 0) in vertex PSin;\n"
"\n"
"// FIXME: host only do glDrawBuffers of 1 buffers not 2. I think this is a major bug\n"
"layout(location = 0) out vec4 FragData0;\n"
"layout(location = 1) out vec4 FragData1;\n"
"\n"
"#endif\n"
"\n"
"////////////////////////////////////////////////////////////////////\n"
"// Texture SAMPLER\n"
"////////////////////////////////////////////////////////////////////\n"
"// // main ps2 memory, each pixel is stored in 32bit color\n"
"// uniform sampler2DRect g_sMemory[2];\n"
"//\n"
"// // used to get the tiled offset into a page given the linear offset\n"
"// uniform sampler2DRect g_sSrcFinal;\n"
"// uniform sampler2D g_sBlocks;\n"
"// uniform sampler2D g_sBilinearBlocks;\n"
"// uniform sampler2D g_sConv16to32;\n"
"// uniform sampler3D g_sConv32to16;\n"
"// uniform sampler2DRect g_sBitwiseANDX;\n"
"// uniform sampler2DRect g_sBitwiseANDY;\n"
"// uniform sampler2DRect g_sInterlace;\n"
"//\n"
"// // used only on rare cases where the render target is PSMT8H\n"
"// uniform sampler2D g_sCLUT;\n"
"// main ps2 memory, each pixel is stored in 32bit color\n"
"layout(binding = 10) uniform sampler2DRect g_sMemory; // dual context\n"
"\n"
"// used to get the tiled offset into a page given the linear offset\n"
"layout(binding = 1) uniform sampler2DRect g_sSrcFinal;\n"
"layout(binding = 2) uniform sampler2D g_sBlocks;\n"
"layout(binding = 3) uniform sampler2D g_sBilinearBlocks;\n"
"layout(binding = 4) uniform sampler2D g_sConv16to32;\n"
"layout(binding = 5) uniform sampler3D g_sConv32to16;\n"
"layout(binding = 6) uniform sampler2DRect g_sBitwiseANDX;\n"
"layout(binding = 7) uniform sampler2DRect g_sBitwiseANDY;\n"
"layout(binding = 8) uniform sampler2DRect g_sInterlace;\n"
"\n"
"// used only on rare cases where the render target is PSMT8H\n"
"layout(binding = 9) uniform sampler2D g_sCLUT;\n"
"\n"
"////////////////////////////////////////////////////////////////////\n"
"// UNIFORM BUFFER\n"
"////////////////////////////////////////////////////////////////////\n"
"layout(std140, binding = 0) uniform constant_buffer\n"
"{\n"
" // Both shader\n"
" // .z is used for the addressing fn\n"
" // FIXME: not same value between c and shader...\n"
" // float4 g_fBilinear = float4(-0.7f, -0.65f, 0.9,1/32767.0f);\n"
" float4 g_fBilinear;\n"
" float4 g_fZBias;\n"
" float4 g_fc0;\n"
" float4 g_fMult;\n"
" // Vertex\n"
" float4 g_fZ; // transforms d3dcolor z into float z\n"
" // Pixel\n"
" half4 g_fExactColor;\n"
"};\n"
"layout(std140, binding = 1) uniform common_buffer\n"
"{\n"
" float4 g_fPosXY;\n"
" float4 g_fFogColor;\n"
"};\n"
"layout(std140, binding = 2) uniform vertex_buffer\n"
"{\n"
" float4 g_fBitBltPos;\n"
" float4 g_fBitBltTex;\n"
" float4 g_fBitBltTrans;\n"
"};\n"
"layout(std140, binding = 3) uniform fragment_buffer\n"
"{\n"
" half4 fTexAlpha2;\n"
"\n"
" float4 g_fTexOffset; // converts the page and block offsets into the mem addr/1024\n"
" float4 g_fTexDims; // mult by tex dims when accessing the block texture\n"
" float4 g_fTexBlock;\n"
"\n"
" float4 g_fClampExts; // if clamping the texture, use (minu, minv, maxu, maxv)\n"
" float4 TexWrapMode; // 0 - repeat/clamp, 1 - region rep (use fRegRepMask)\n"
"\n"
" float4 g_fRealTexDims; // tex dims used for linear filtering (w,h,1/w,1/h)\n"
"\n"
" // (alpha0, alpha1, 1 if highlight2 and tcc is rgba, 1-y)\n"
" half4 g_fTestBlack; // used for aem bit\n"
"\n"
" float4 g_fPageOffset;\n"
"\n"
" half4 fTexAlpha;\n"
"\n"
" float4 g_fInvTexDims; // similar to g_fClutOff\n"
"\n"
" // used for rectblitting\n"
" float4 g_fBitBltZ;\n"
"\n"
" half4 g_fOneColor; // col*.xxxy+.zzzw\n"
"};\n"
"\n"
"\n"
"// given a local tex coord, returns the coord in the memory\n"
"float2 ps2memcoord(float2 realtex)\n"
"{\n"
" float4 off;\n"
"\n"
" // block off\n"
" realtex.xy = realtex.xy * g_fTexDims.xy + g_fTexDims.zw;\n"
" realtex.xy = (realtex.xy - fract(realtex.xy)) * g_fMult.zw;\n"
" float2 fblock = fract(realtex.xy);\n"
" off.xy = realtex.xy-fblock.xy;\n"
"\n"
"#ifdef ACCURATE_DECOMPRESSION\n"
" off.z = texture(g_sBlocks, g_fTexBlock.xy*fblock + g_fTexBlock.zw).r;\n"
" off.x = dot(off.xy, g_fTexOffset.xy);\n"
" float r = g_fTexOffset.w;\n"
" float f = fract(off.x);\n"
" float fadd = g_fTexOffset.z * off.z;\n"
" off.w = off.x + fadd + r;\n"
" off.x = fract(f + fadd + r);\n"
" off.w -= off.x ;\n"
"#else\n"
" off.z = texture(g_sBlocks, g_fTexBlock.xy*fblock + g_fTexBlock.zw).r;\n"
"\n"
" // combine the two\n"
" off.x = dot(off.xyz, g_fTexOffset.xyz)+g_fTexOffset.w;\n"
" off.x = modf(off.x, off.w);\n"
"#endif\n"
"\n"
" off.xy = off.xw * g_fPageOffset.zy + g_fPageOffset.wx;\n"
" //off.y = off.w * g_fPageOffset.y + g_fPageOffset.x;\n"
" return off.xy;\n"
"}\n"
"\n"
"// find all texcoords for bilinear filtering\n"
"// assume that orgtex are already on boundaries\n"
"void ps2memcoord4(float4 orgtex, out float4 off0, out float4 off1)\n"
"{\n"
" //float4 off0, off1, off2, off3;\n"
" float4 realtex;\n"
"\n"
" // block off\n"
" realtex = (orgtex * g_fTexDims.xyxy + g_fTexDims.zwzw);// * g_fMult.zwzw;\n"
" float4 fblock = fract(realtex.xyzw);\n"
" float4 ftransblock = g_fTexBlock.xyxy*fblock + g_fTexBlock.zwzw;\n"
" realtex -= fblock;\n"
"\n"
" float4 transvals = g_fTexOffset.x * realtex.xzxz + g_fTexOffset.y * realtex.yyww + g_fTexOffset.w;\n"
"\n"
" float4 colors;// = texture(g_sBilinearBlocks, ftransblock.xy);\n"
"\n"
" // this is faster on ffx ingame\n"
" colors.x = texture(g_sBlocks, ftransblock.xy).r;\n"
" colors.y = texture(g_sBlocks, ftransblock.zy).r;\n"
" colors.z = texture(g_sBlocks, ftransblock.xw).r;\n"
" colors.w = texture(g_sBlocks, ftransblock.zw).r;\n"
"\n"
" float4 fr, rem;\n"
"\n"
"#ifdef ACCURATE_DECOMPRESSION\n"
" fr = fract(transvals);\n"
" float4 fadd = colors * g_fTexOffset.z;\n"
" rem = transvals + fadd;\n"
" fr = fract(fr + fadd);\n"
" rem -= fr;\n"
"#else\n"
" transvals += colors * g_fTexOffset.z;\n"
"\n"
" fr = modf(transvals, rem);\n"
"#endif\n"
"\n"
" rem = rem * g_fPageOffset.y + g_fPageOffset.x;\n"
" fr = fr * g_fPageOffset.z + g_fPageOffset.w;\n"
"\n"
" // combine\n"
" off0 = g_fc0.yxyx * fr.xxyy + g_fc0.xyxy * rem.xxyy;\n"
" off1 = g_fc0.yxyx * fr.zzww + g_fc0.xyxy * rem.zzww;\n"
"}\n"
"\n"
"void ps2memcoord4_fast(float4 orgtex, out float4 off0, out float4 off1)\n"
"{\n"
" float4 realtex;\n"
"\n"
" realtex = (orgtex * g_fTexDims.xyxy + g_fTexDims.zwzw);// * g_fMult.zwzw;\n"
" float4 fblock = fract(realtex.xyzw);\n"
" float2 ftransblock = g_fTexBlock.xy*fblock.xy + g_fTexBlock.zw;\n"
" realtex -= fblock;\n"
"\n"
" float4 transvals = g_fTexOffset.x * realtex.xzxz + g_fTexOffset.y * realtex.yyww + g_fTexOffset.w;\n"
"\n"
" float4 colors = texture(g_sBilinearBlocks, ftransblock.xy);\n"
" float4 fr, rem;\n"
"\n"
"#ifdef ACCURATE_DECOMPRESSION\n"
" fr = fract(transvals);\n"
" float4 fadd = colors * g_fTexOffset.z;\n"
" rem = transvals + fadd;\n"
" fr = fract(fr + fadd);\n"
" rem -= fr;\n"
"#else\n"
" transvals += colors * g_fTexOffset.z;\n"
"\n"
" fr = modf(transvals, rem);\n"
"#endif\n"
"\n"
" rem = rem * g_fPageOffset.y + g_fPageOffset.x;\n"
" fr = fr * g_fPageOffset.z;\n"
"\n"
" off0 = g_fc0.yxyx * fr.xxyy + g_fc0.xyxy * rem.xxyy;\n"
" off1 = g_fc0.yxyx * fr.zzww + g_fc0.xyxy * rem.zzww;\n"
"}\n"
"\n"
"// Wrapping modes\n"
"#if defined(REPEAT)\n"
"\n"
"float2 ps2addr(float2 coord)\n"
"{\n"
" return fract(coord.xy);\n"
"}\n"
"\n"
"#elif defined(CLAMP)\n"
"\n"
"float2 ps2addr(float2 coord)\n"
"{\n"
" return clamp(coord.xy, g_fClampExts.xy, g_fClampExts.zw);\n"
"}\n"
"\n"
"#elif defined(REGION_REPEAT)\n"
"\n"
"// computes the local tex coord along with addressing modes\n"
"float2 ps2addr(float2 coord)\n"
"{\n"
" float2 final = fract(clamp(coord.xy, g_fClampExts.xy, g_fClampExts.zw));\n"
"\n"
" if( TexWrapMode.x > g_fBilinear.z ) // region repeat mode for x (umsk&x)|ufix\n"
" final.x = texture(g_sBitwiseANDX, abs(coord.x)*TexWrapMode.zx).x * g_fClampExts.x + g_fClampExts.z;\n"
" if( TexWrapMode.y > g_fBilinear.z ) // region repeat mode for x (vmsk&x)|vfix\n"
" final.y = texture(g_sBitwiseANDY, abs(coord.y)*TexWrapMode.wy).x * g_fClampExts.y + g_fClampExts.w;\n"
"\n"
" return final;\n"
"}\n"
"\n"
"#else\n"
"\n"
"float2 ps2addr(float2 coord)\n"
"{\n"
" return fract(clamp(coord.xy, g_fClampExts.xy, g_fClampExts.zw));\n"
"}\n"
"\n"
"#endif\n"
"\n"
"half4 tex2DPS_32(float2 tex0)\n"
"{\n"
" return texture(g_sMemory, ps2memcoord(tex0));\n"
"}\n"
"\n"
"// use when texture is not tiled -- shader 1\n"
"half4 tex2DPS_tex32(float2 tex0)\n"
"{\n"
" return texture(g_sMemory, g_fTexDims.xy*tex0+g_fTexDims.zw)*g_fZBias.zzzw+g_fPageOffset.w;\n"
"}\n"
"\n"
"// use when texture is not tiled -- shader 2\n"
"half4 tex2DPS_clut32(float2 tex0)\n"
"{\n"
" float index = texture(g_sMemory, g_fTexDims.xy*tex0+g_fTexDims.zw).a+g_fPageOffset.w;\n"
" return texture(g_sCLUT, index*g_fExactColor.xz+g_fExactColor.yz);\n"
"}\n"
"\n"
"// Shader 3\n"
"// use when texture is not tiled and converting from 32bit to 16bit\n"
"// don't convert on the block level, only on the column level\n"
"// so every other 8 pixels, use the upper bits instead of lower\n"
"half4 tex2DPS_tex32to16(float2 tex0)\n"
"{\n"
" bool upper = false;\n"
" tex0.y += g_fPageOffset.z;\n"
" float2 ffrac = mod(tex0, g_fTexOffset.xy);\n"
" tex0.xy = g_fc0.ww * (tex0.xy + ffrac);\n"
" if( ffrac.x > g_fTexOffset.z ) {\n"
" tex0.x -= g_fTexOffset.z;\n"
" upper = true;\n"
" }\n"
" if( ffrac.y >= g_fTexOffset.w ) {\n"
" tex0.y -= g_fTexOffset.w;\n"
" tex0.x += g_fc0.w;\n"
" }\n"
"\n"
" half4 color = texture(g_sMemory, g_fTexDims.xy*tex0+g_fTexDims.zw)*g_fZBias.zzzw+g_fPageOffset.w;\n"
" float2 uv = upper ? color.xw : color.zy;\n"
" return texture(g_sConv16to32, uv+g_fPageOffset.xy);\n"
"}\n"
"\n"
"// Shader 4\n"
"// used when a 16 bit texture is used an 8h\n"
"half4 tex2DPS_tex16to8h(float2 tex0)\n"
"{\n"
" float4 final;\n"
" float2 ffrac = mod(tex0+g_fPageOffset.zw, g_fTexOffset.xy);\n"
" tex0.xy = g_fPageOffset.xy * tex0.xy - ffrac * g_fc0.yw;\n"
"\n"
" if( ffrac.x > g_fTexOffset.x*g_fc0.w )\n"
" tex0.x += g_fTexOffset.x*g_fc0.w;\n"
" if( tex0.x >= g_fc0.y ) tex0 += g_fTexOffset.zw;\n"
"\n"
" float4 upper = texture(g_sMemory, g_fTexDims.xy*tex0+g_fTexDims.zw);\n"
"\n"
" // only need alpha\n"
" float index = texture(g_sConv32to16, upper.zyx-g_fc0.z).y + upper.w*g_fc0.w*g_fc0.w;\n"
" return texture(g_sCLUT, index+g_fExactColor.yz);\n"
"}\n"
"\n"
"// Shader 5\n"
"// used when a 16 bit texture is used a 32bit one\n"
"half4 tex2DPS_tex16to32(float2 tex0)\n"
"{\n"
" float4 final;\n"
" float2 ffrac = mod(tex0+g_fPageOffset.zw, g_fTexOffset.xy);\n"
" //tex0.xy = g_fPageOffset.xy * tex0.xy - ffrac * g_fc0.yw;\n"
" tex0.y += g_fPageOffset.y * ffrac.y;\n"
"\n"
" if( ffrac.x > g_fTexOffset.z ) {\n"
" tex0.x -= g_fTexOffset.z;\n"
" tex0.y += g_fTexOffset.w;\n"
" }\n"
"\n"
" float fconst = g_fc0.w*g_fc0.w;\n"
" float4 lower = texture(g_sSrcFinal, g_fTexDims.xy*tex0);\n"
" float4 upper = texture(g_sMemory, g_fTexDims.xy*tex0+g_fTexDims.zw);\n"
"\n"
" final.zy = texture(g_sConv32to16, lower.zyx).xy + lower.ww*fconst;\n"
" final.xw = texture(g_sConv32to16, upper.zyx).xy + upper.ww*fconst;\n"
" return final;\n"
"}\n"
"\n"
"half4 tex2DPS_tex16to32h(float2 tex0)\n"
"{\n"
" float4 final = vec4(0.0, 0.0, 0.0, 0.0);\n"
" return final;\n"
"}\n"
"\n"
"//half4 f;\n"
"//f.w = old.y > (127.2f/255.0f) ? 1 : 0;\n"
"//old.y -= 0.5f * f.w;\n"
"//f.xyz = fract(old.yyx*half3(2.002*255.0f/256.0f, 64.025f*255.0f/256.0f, 8.002*255.0f/256.0f));\n"
"//f.y += old.x * (0.25f*255.0f/256.0f);\n"
"\n"
"////////////////////////////////\n"
"// calculates the texture color\n"
"////////////////////////////////\n"
"\n"
"#define decl_ps2shade(num) \\\n"
"decl_ps2shade_##num(_32) \\\n"
"decl_ps2shade_##num(_tex32) \\\n"
"decl_ps2shade_##num(_clut32) \\\n"
"decl_ps2shade_##num(_tex32to16) \\\n"
"decl_ps2shade_##num(_tex16to8h) \\\n"
"decl_ps2shade_##num(_tex16to32h)\n"
"\n"
"// nearest\n"
"#define decl_ps2shade_0(bit) \\\n"
"float4 ps2shade0##bit( TEX_DECL tex) \\\n"
"{ \\\n"
" return tex2DPS##bit( ps2addr(TEX_XY)); \\\n"
"}\n"
"\n"
"// do fast memcoord4 calcs when textures behave well\n"
"#ifdef REPEAT\n"
"#define PS2MEMCOORD4 ps2memcoord4\n"
"#else\n"
"#define PS2MEMCOORD4 ps2memcoord4\n"
"#endif\n"
"\n"
"\n"
"#define decl_BilinearFilter(bit, addrfn) \\\n"
"half4 BilinearFilter##bit(float2 tex0) \\\n"
"{ \\\n"
" float4 off0, off1; \\\n"
" float4 ftex; \\\n"
" float2 ffrac; \\\n"
" ftex.xy = tex0 + g_fBilinear.xy * g_fRealTexDims.zw; \\\n"
" ffrac = fract(ftex.xy*g_fRealTexDims.xy); \\\n"
" ftex.xy -= ffrac.xy * g_fRealTexDims.zw; \\\n"
" \\\n"
" ftex.zw = ps2addr(ftex.xy + g_fRealTexDims.zw); \\\n"
" ftex.xy = ps2addr(ftex.xy); \\\n"
" \\\n"
" PS2MEMCOORD4(ftex, off0, off1); \\\n"
" half4 c0 = texture(g_sMemory, off0.xy); \\\n"
" half4 c1 = texture(g_sMemory, off0.zw); \\\n"
" half4 c2 = texture(g_sMemory, off1.xy); \\\n"
" half4 c3 = texture(g_sMemory, off1.zw); \\\n"
" return mix( mix(c0, c1, vec4(ffrac.x)), mix(c2, c3, ffrac.x), vec4(ffrac.y) ); \\\n"
"}\n"
"\n"
"decl_BilinearFilter(_32, ps2addr)\n"
"decl_BilinearFilter(_tex32, ps2addr)\n"
"decl_BilinearFilter(_clut32, ps2addr)\n"
"decl_BilinearFilter(_tex32to16, ps2addr)\n"
"decl_BilinearFilter(_tex16to8h, ps2addr)\n"
"decl_BilinearFilter(_tex16to32h, ps2addr)\n"
"\n"
"//TODO! For mip maps, only apply when LOD >= 0\n"
"// lcm == 0, LOD = log(1/Q)*L + K, lcm == 1, LOD = K\n"
"\n"
"// bilinear\n"
"#define decl_ps2shade_1(bit) \\\n"
"half4 ps2shade1##bit(TEX_DECL tex) \\\n"
"{ \\\n"
" return BilinearFilter##bit(TEX_XY); \\\n"
"}\n"
"\n"
"// nearest, mip nearest\n"
"#define decl_ps2shade_2(bit) \\\n"
"half4 ps2shade2##bit(TEX_DECL tex) \\\n"
"{ \\\n"
" return tex2DPS##bit( ps2addr(TEX_XY)); \\\n"
"}\n"
"\n"
"// nearest, mip linear\n"
"#define decl_ps2shade_3(bit) \\\n"
"half4 ps2shade3##bit(TEX_DECL tex) \\\n"
"{ \\\n"
" return tex2DPS##bit(ps2addr(TEX_XY)); \\\n"
"}\n"
"\n"
"// linear, mip nearest\n"
"#define decl_ps2shade_4(bit) \\\n"
"half4 ps2shade4##bit(TEX_DECL tex) \\\n"
"{ \\\n"
" return BilinearFilter##bit(TEX_XY); \\\n"
"}\n"
"\n"
"// linear, mip linear\n"
"#define decl_ps2shade_5(bit) \\\n"
"half4 ps2shade5##bit(TEX_DECL tex) \\\n"
"{ \\\n"
" return BilinearFilter##bit(TEX_XY); \\\n"
"}\n"
"\n"
"decl_ps2shade(0)\n"
"decl_ps2shade(1)\n"
"decl_ps2shade(2)\n"
"decl_ps2shade(3)\n"
"decl_ps2shade(4)\n"
"decl_ps2shade(5)\n"
"\n"
"\n"
"half4 ps2CalcShade(half4 texcol, half4 color)\n"
"{\n"
"#ifdef TEST_AEM\n"
" if( dot(texcol.xyzw, g_fTestBlack.xyzw) <= g_fc0.z )\n"
" texcol.w = g_fc0.x;\n"
" else\n"
"#endif\n"
" texcol.w = texcol.w * fTexAlpha.y + fTexAlpha.x;\n"
"\n"
" texcol = texcol * (fTexAlpha2.zzzw * color + fTexAlpha2.xxxy) + fTexAlpha.zzzw * color.wwww;\n"
"\n"
" return texcol;\n"
"}\n"
"\n"
"// final ops on the color\n"
"#ifdef EXACT_COLOR\n"
"\n"
"half4 ps2FinalColor(half4 col)\n"
"{\n"
" // g_fOneColor has to scale by 255\n"
" half4 temp = col * g_fOneColor.xxxy + g_fOneColor.zzzw;\n"
" temp.w = floor(temp.w)*g_fExactColor.w;\n"
" return temp;\n"
"}\n"
"\n"
"#else\n"
"half4 ps2FinalColor(half4 col)\n"
"{\n"
" return col * g_fOneColor.xxxy + g_fOneColor.zzzw;\n"
"}\n"
"#endif\n"
"\n"
"\n"
"#ifdef FRAGMENT_SHADER // This is code only for FRAGMENTS (pixel shader)\n"
"\n"
"#ifdef WRITE_DEPTH\n"
"void write_depth_target()\n"
"{\n"
" FragData1 = PSin.Z; \n"
"}\n"
"#else\n"
"void write_depth_target() { }\n"
"#endif\n"
"\n"
"void RegularPS() {\n"
" // whenever outputting depth, make sure to mult by 255/256 and 1\n"
" FragData0 = ps2FinalColor(PSin.color);\n"
" write_depth_target();\n"
"}\n"
"\n"
"#define DECL_TEXPS(num, bit) \\\n"
"void Texture##num##bit##PS() \\\n"
"{ \\\n"
" FragData0 = ps2FinalColor(ps2CalcShade(ps2shade##num##bit(PSin.tex), PSin.color)); \\\n"
" write_depth_target(); \\\n"
"}\n"
"\n"
"#define DECL_TEXPS_(num) \\\n"
"DECL_TEXPS(num, _32) \\\n"
"DECL_TEXPS(num, _tex32) \\\n"
"DECL_TEXPS(num, _clut32) \\\n"
"DECL_TEXPS(num, _tex32to16) \\\n"
"DECL_TEXPS(num, _tex16to8h)\n"
"\n"
"DECL_TEXPS_(0)\n"
"DECL_TEXPS_(1)\n"
"DECL_TEXPS_(2)\n"
"DECL_TEXPS_(3)\n"
"DECL_TEXPS_(4)\n"
"DECL_TEXPS_(5)\n"
"\n"
"void RegularFogPS() {\n"
" half4 c;\n"
" c.xyz = mix(g_fFogColor.xyz, PSin.color.xyz, vec3(PSin.fog));\n"
" c.w = PSin.color.w;\n"
" FragData0 = ps2FinalColor(c);\n"
" write_depth_target();\n"
"}\n"
"\n"
"#define DECL_TEXFOGPS(num, bit) \\\n"
"void TextureFog##num##bit##PS() \\\n"
"{ \\\n"
" half4 c = ps2CalcShade(ps2shade##num##bit(PSin.tex), PSin.color); \\\n"
" c.xyz = mix(g_fFogColor.xyz, c.xyz, vec3(PSin.fog)); \\\n"
" FragData0 = ps2FinalColor(c); \\\n"
" write_depth_target(); \\\n"
"}\n"
"\n"
"#define DECL_TEXFOGPS_(num) \\\n"
"DECL_TEXFOGPS(num, _32) \\\n"
"DECL_TEXFOGPS(num, _tex32) \\\n"
"DECL_TEXFOGPS(num, _clut32) \\\n"
"DECL_TEXFOGPS(num, _tex32to16) \\\n"
"DECL_TEXFOGPS(num, _tex16to8h)\n"
"\n"
"DECL_TEXFOGPS_(0)\n"
"DECL_TEXFOGPS_(1)\n"
"DECL_TEXFOGPS_(2)\n"
"DECL_TEXFOGPS_(3)\n"
"DECL_TEXFOGPS_(4)\n"
"DECL_TEXFOGPS_(5)\n"
"\n"
"//-------------------------------------------------------\n"
"// Techniques not related to the main primitive commands\n"
"half4 BilinearBitBlt(float2 tex0)\n"
"{\n"
" float4 ftex;\n"
" float2 ffrac;\n"
"\n"
" ffrac.xy = fract(tex0*g_fRealTexDims.xy);\n"
" ftex.xy = tex0 - ffrac.xy * g_fRealTexDims.zw;\n"
" ftex.zw = ftex.xy + g_fRealTexDims.zw;\n"
"\n"
" float4 off0, off1;\n"
" ps2memcoord4_fast(ftex, off0, off1);\n"
" half4 c0 = texture(g_sMemory, off0.xy);\n"
" half4 c1 = texture(g_sMemory, off0.zw);\n"
" half4 c2 = texture(g_sMemory, off1.xy);\n"
" half4 c3 = texture(g_sMemory, off1.zw);\n"
"\n"
" return mix( mix(c0, c1, vec4(ffrac.x)), mix(c2, c3, vec4(ffrac.x)), vec4(ffrac.y) );\n"
"}\n"
"\n"
"void BitBltPS() {\n"
" FragData0 = texture(g_sMemory, ps2memcoord(PSin.tex.xy))*g_fOneColor.xxxy;\n"
"}\n"
"\n"
"// used when AA\n"
"void BitBltAAPS() {\n"
" FragData0 = BilinearBitBlt(PSin.tex.xy) * g_fOneColor.xxxy;\n"
"}\n"
"\n"
"void BitBltDepthPS() {\n"
" vec4 data;\n"
" data = texture(g_sMemory, ps2memcoord(PSin.tex.xy));\n"
" FragData0 = data + g_fZBias.y;\n"
"#ifdef NO_LOGZ\n"
" gl_FragDepth = dot(data, g_fBitBltZ);\n"
"#else\n"
" gl_FragDepth = log(g_fc0.y + dot(data, g_fBitBltZ)) * g_fOneColor.w;\n"
"#endif\n"
"}\n"
"\n"
"void BitBltDepthMRTPS() {\n"
" vec4 data;\n"
" data = texture(g_sMemory, ps2memcoord(PSin.tex.xy));\n"
" FragData0 = data + g_fZBias.y;\n"
" FragData1.x = g_fc0.x;\n"
"#ifdef NO_LOGZ\n"
" gl_FragDepth = dot(data, g_fBitBltZ);\n"
"#else\n"
" gl_FragDepth = log(g_fc0.y + dot(data, g_fBitBltZ)) * g_fOneColor.w;\n"
"#endif\n"
"}\n"
"\n"
"// static const float BlurKernel[9] = {\n"
"// 0.027601,\n"
"// 0.066213,\n"
"// 0.123701,\n"
"// 0.179952,\n"
"// 0.205065,\n"
"// 0.179952,\n"
"// 0.123701,\n"
"// 0.066213,\n"
"// 0.027601\n"
"// };\n"
"\n"
"half4 BilinearFloat16(float2 tex0)\n"
"{\n"
" return texture(g_sSrcFinal, tex0.xy);\n"
"}\n"
"\n"
"void CRTCTargInterPS() {\n"
" float finter = texture(g_sInterlace, PSin.Z.yy).x * g_fOneColor.z + g_fOneColor.w + g_fc0.w;\n"
" float4 c = BilinearFloat16(PSin.tex.xy);\n"
" c.w = ( g_fc0.w*c.w * g_fOneColor.x + g_fOneColor.y ) * finter;\n"
" FragData0 = c;\n"
"}\n"
"\n"
"void CRTCTargPS() {\n"
" float4 c = BilinearFloat16(PSin.tex.xy);\n"
" // FIXME DEBUG: to validate tex coord on blit\n"
" //c = vec4(PSin.tex.x/512.0f, PSin.tex.y/512.0f, 0.0, 1.0);\n"
" c.w = g_fc0.w * c.w * g_fOneColor.x + g_fOneColor.y;\n"
" FragData0 = c;\n"
"}\n"
"\n"
"void CRTCInterPS() {\n"
" float finter = texture(g_sInterlace, PSin.Z.yy).x * g_fOneColor.z + g_fOneColor.w + g_fc0.w;\n"
" float2 filtcoord = trunc(PSin.tex.xy) * g_fInvTexDims.xy + g_fInvTexDims.zw;\n"
" half4 c = BilinearBitBlt(filtcoord);\n"
" c.w = (c.w * g_fOneColor.x + g_fOneColor.y)*finter;\n"
" FragData0 = c;\n"
"}\n"
"\n"
"// simpler\n"
"void CRTCInterPS_Nearest() {\n"
" float finter = texture(g_sInterlace, PSin.Z.yy).x * g_fOneColor.z + g_fOneColor.w + g_fc0.w;\n"
" half4 c = texture(g_sMemory, ps2memcoord(PSin.tex.xy));\n"
" c.w = (c.w * g_fOneColor.x + g_fOneColor.y)*finter;\n"
" FragData0 = c;\n"
"}\n"
"\n"
"void CRTCPS() {\n"
" float2 filtcoord = PSin.tex.xy * g_fInvTexDims.xy+g_fInvTexDims.zw;\n"
" half4 c = BilinearBitBlt(filtcoord);\n"
" c.w = c.w * g_fOneColor.x + g_fOneColor.y;\n"
" FragData0 = c;\n"
"}\n"
"\n"
"// simpler\n"
"void CRTCPS_Nearest() {\n"
" half4 c = texture(g_sMemory, ps2memcoord(PSin.tex.xy));\n"
" c.w = c.w * g_fOneColor.x + g_fOneColor.y;\n"
" FragData0 = c;\n"
"}\n"
"\n"
"void CRTC24InterPS() {\n"
" float finter = texture(g_sInterlace, PSin.Z.yy).x * g_fOneColor.z + g_fOneColor.w + g_fc0.w;\n"
" float2 filtcoord = trunc(PSin.tex.xy) * g_fInvTexDims.xy + g_fInvTexDims.zw;\n"
"\n"
" half4 c = texture(g_sMemory, ps2memcoord(filtcoord));\n"
" c.w = (c.w * g_fOneColor.x + g_fOneColor.y)*finter;\n"
" FragData0 = c;\n"
"}\n"
"\n"
"void CRTC24PS() {\n"
" float2 filtcoord = trunc(PSin.tex.xy) * g_fInvTexDims.xy + g_fInvTexDims.zw;\n"
" half4 c = texture(g_sMemory, ps2memcoord(filtcoord));\n"
" c.w = c.w * g_fOneColor.x + g_fOneColor.y;\n"
" FragData0 = c;\n"
"}\n"
"\n"
"void ZeroPS() {\n"
" FragData0 = g_fOneColor;\n"
"}\n"
"\n"
"void ZeroDebugPS() {\n"
" FragData0 = vec4(PSin.tex.x, PSin.tex.y, PSin.tex.z, 0.5);\n"
"}\n"
"\n"
"void ZeroDebug2PS() {\n"
" vec2 xy = ps2memcoord(fract(PSin.tex.xy/PSin.tex.z)) * vec2(1/4096.0f, 1/48.0f);\n"
" FragData0 = vec4(xy.x, xy.y, 0.0, 0.5);\n"
"}\n"
"\n"
"void ZeroDebug3PS() {\n"
" //FragData0 = vec4(PSin.position.x/2.0f + 0.5f, PSin.position.y/2.0f + 0.5f, 1.0, 0.5);\n"
"}\n"
"\n"
"void BaseTexturePS() {\n"
" FragData0 = texture(g_sSrcFinal, PSin.tex.xy) * g_fOneColor;\n"
"}\n"
"\n"
"void Convert16to32PS() {\n"
" float4 final;\n"
" float2 ffrac = mod ( PSin.tex.xy + g_fTexDims.zw, g_fTexOffset.xy);\n"
" float2 tex0 = g_fTexDims.xy * PSin.tex.xy - ffrac * g_fc0.yw;\n"
"\n"
" if (ffrac.x > g_fTexOffset.x*g_fc0.w)\n"
" tex0.x += g_fTexOffset.x*g_fc0.w;\n"
" if (tex0.x >= g_fc0.y)\n"
" tex0 += g_fTexOffset.zw;\n"
"\n"
" float4 lower = texture(g_sSrcFinal, tex0);\n"
" float4 upper = texture(g_sSrcFinal, tex0 + g_fPageOffset.xy);\n"
"\n"
" final.zy = texture(g_sConv32to16, lower.zyx).xy + lower.ww*g_fPageOffset.zw;\n"
" final.xw = texture(g_sConv32to16, upper.zyx).xy + upper.ww*g_fPageOffset.zw;\n"
"\n"
" FragData0= final;\n"
"}\n"
"\n"
"// use when texture is not tiled and converting from 32bit to 16bit\n"
"// don't convert on the block level, only on the column level\n"
"// so every other 8 pixels, use the upper bits instead of lower\n"
"void Convert32to16PS() {\n"
" bool upper = false;\n"
" float2 ffrac = mod(PSin.tex.xy + g_fTexDims.zw, g_fTexOffset.xy);\n"
" float2 tex0 = g_fc0.ww * (PSin.tex.xy + ffrac);\n"
" if( ffrac.x > g_fTexOffset.z ) {\n"
" tex0.x -= g_fTexOffset.z;\n"
" upper = true;\n"
" }\n"
" if( ffrac.y >= g_fTexOffset.w ) {\n"
" tex0.y -= g_fTexOffset.w;\n"
" tex0.x += g_fc0.w;\n"
" }\n"
"\n"
" half4 color = texture(g_sSrcFinal, tex0*g_fTexDims.xy)*g_fc0.yyyw;\n"
" float2 uv = upper ? color.xw : color.zy;\n"
" FragData0 = texture(g_sConv16to32, uv*g_fPageOffset.xy+g_fPageOffset.zw)*g_fTexDims.xxxy;\n"
"}\n"
"#endif //FRAGMENT_SHADER\n"
"\n"
"#ifdef VERTEX_SHADER\n"
"\n"
"void SetColor() {\n"
" VSout.color = Color;\n"
"}\n"
"\n"
"void SetTex() {\n"
"#ifdef PERSPECTIVE_CORRECT_TEX\n"
" VSout.tex = TexCoord;\n"
"#else\n"
" VSout.tex.xy = TexCoord.xy/TexCoord.z;\n"
"#endif\n"
"}\n"
"\n"
"void SetZ() {\n"
"#ifdef WRITE_DEPTH\n"
" VSout.Z = SecondaryColor * g_fZBias.x + g_fZBias.y;\n"
" VSout.Z.w = 1.0f;\n"
"#endif\n"
"}\n"
"\n"
"void SetPosition() {\n"
" float4 position;\n"
" position.xy = vec2(Vert.xy) * g_fPosXY.xy + g_fPosXY.zw;\n"
" // FIXME: the factor in normal mode seem bogus. They don't have same order than in log mode. Or I failed to understand the logic\n"
" //// normal mode.\n"
" // -> dot(g_fZ, SecondaryColor.zyxw)) => reconstruct a float from normalized char. The float range from 0 to 1\n"
" // position.z = dot(g_fZ, SecondaryColor.zyxw);\n"
" //// logz mode\n"
" // -> dot(g_fZ, SecondaryColor.zyxw)) => reconstruct a float from normalized char. The float range from 0 to 2**32\n"
" // position.z = log(g_fc0.y + dot(g_fZ, SecondaryColor.zyxw)) * g_fZNorm.x\n"
" // position.z = log(1 + Z_INT) * 0.999f / (32 * log(2.0)) = log2(1 + Z_INT) * 0.999f / 32\n"
" // log2(...) will range from 0 to 32\n"
"\n"
" // position.z = (log(g_fc0.y + dot(g_fZ, SecondaryColor.zyxw)) * g_fZNorm.x + g_fZNorm.y) * g_fZMin.y + dot(g_fZ, SecondaryColor.zyxw) * g_fZMin.x ;\n"
"#ifdef NO_LOGZ\n"
" position.z = dot(g_fZ, SecondaryColor.zyxw);\n"
"#else\n"
" position.z = log2(1.0f + dot(g_fZ, SecondaryColor.zyxw)) * 0.999f/32.0f;\n"
"#endif\n"
" position.w = 1.0f;\n"
"\n"
" gl_Position = position;\n"
"}\n"
"\n"
"void SetFog() {\n"
" VSout.fog = float(Vert.z) * g_fBilinear.w;\n"
"}\n"
"\n"
"// just smooth shadering\n"
"void RegularVS() {\n"
" SetPosition();\n"
" SetColor();\n"
" SetZ();\n"
"}\n"
"\n"
"// diffuse texture mapping\n"
"void TextureVS() {\n"
" SetPosition();\n"
" SetColor();\n"
" SetTex();\n"
" SetZ();\n"
"}\n"
"\n"
"void RegularFogVS() {\n"
" SetPosition();\n"
" SetColor();\n"
" SetZ();\n"
" SetFog();\n"
"}\n"
"\n"
"void TextureFogVS() {\n"
" SetPosition();\n"
" SetColor();\n"
" SetTex();\n"
" SetZ();\n"
" SetFog();\n"
"}\n"
"\n"
"void BitBltVS() {\n"
" vec4 position;\n"
" position.xy = vec2(Vert.xy) * g_fBitBltPos.xy + g_fBitBltPos.zw;\n"
" position.zw = g_fc0.xy;\n"
" gl_Position = position;\n"
"\n"
" VSout.tex.xy = TexCoord.xy * g_fBitBltTex.xy + g_fBitBltTex.zw;\n"
" VSout.Z.xy = position.xy * g_fBitBltTrans.xy + g_fBitBltTrans.zw;\n"
"}\n"
"\n"
"#endif\n"
;