You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

1410 lines
49 KiB

/*-------------------------------------------------------------------------
* drawElements Quality Program OpenGL (ES) Module
* -----------------------------------------------
*
* Copyright 2014 The Android Open Source Project
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*
*//*!
* \file
* \brief Compiler test case.
*//*--------------------------------------------------------------------*/
#include "glsShaderLibraryCase.hpp"
#include "tcuTestLog.hpp"
#include "tcuRenderTarget.hpp"
#include "tcuTextureUtil.hpp"
#include "tcuSurface.hpp"
#include "tcuStringTemplate.hpp"
#include "gluShaderProgram.hpp"
#include "gluPixelTransfer.hpp"
#include "gluDrawUtil.hpp"
#include "gluContextInfo.hpp"
#include "gluStrUtil.hpp"
#include "glwFunctions.hpp"
#include "glwEnums.hpp"
#include "deRandom.hpp"
#include "deInt32.h"
#include "deMath.h"
#include "deString.h"
#include "deStringUtil.hpp"
#include "deSharedPtr.hpp"
#include <map>
#include <vector>
#include <string>
#include <sstream>
namespace deqp
{
namespace gls
{
using namespace tcu;
using namespace glu;
using namespace glu::sl;
using std::vector;
using std::string;
using std::ostringstream;
using std::map;
using std::pair;
using de::SharedPtr;
// OpenGL-specific specialization utils
static vector<RequiredExtension> checkAndSpecializeExtensions (const vector<RequiredExtension>& src,
const ContextInfo& ctxInfo)
{
vector<RequiredExtension> specialized;
for (size_t extNdx = 0; extNdx < src.size(); ++extNdx)
{
const RequiredExtension& extension = src[extNdx];
int supportedAltNdx = -1;
for (size_t alternativeNdx = 0; alternativeNdx < extension.alternatives.size(); ++alternativeNdx)
{
if (ctxInfo.isExtensionSupported(extension.alternatives[alternativeNdx].c_str()))
{
supportedAltNdx = (int)alternativeNdx;
break;
}
}
if (supportedAltNdx >= 0)
{
specialized.push_back(RequiredExtension(extension.alternatives[supportedAltNdx], extension.effectiveStages));
}
else
{
// no extension(s). Make a nice output
std::ostringstream extensionList;
for (size_t ndx = 0; ndx < extension.alternatives.size(); ++ndx)
{
if (!extensionList.str().empty())
extensionList << ", ";
extensionList << extension.alternatives[ndx];
}
if (extension.alternatives.size() == 1)
throw tcu::NotSupportedError("Test requires extension " + extensionList.str());
else
throw tcu::NotSupportedError("Test requires any extension of " + extensionList.str());
}
}
return specialized;
}
static void checkImplementationLimits (const vector<RequiredCapability>& requiredCaps,
const ContextInfo& ctxInfo)
{
for (size_t capNdx = 0; capNdx < requiredCaps.size(); ++capNdx)
{
const RequiredCapability& capability = requiredCaps[capNdx];
if (capability.type != CAPABILITY_LIMIT)
continue;
const deUint32 pname = capability.enumName;
const int requiredValue = capability.referenceValue;
const int supportedValue = ctxInfo.getInt((int)pname);
if (supportedValue <= requiredValue)
throw tcu::NotSupportedError("Test requires " + de::toString(glu::getGettableStateStr(pname)) + " (" + de::toString(supportedValue) + ") >= " + de::toString(requiredValue));
}
}
// Shader source specialization
// This functions builds a matching vertex shader for a 'both' case, when
// the fragment shader is being tested.
// We need to build attributes and varyings for each 'input'.
static string genVertexShader (const ShaderCaseSpecification& spec)
{
ostringstream res;
const bool usesInout = glslVersionUsesInOutQualifiers(spec.targetVersion);
const char* const vtxIn = usesInout ? "in" : "attribute";
const char* const vtxOut = usesInout ? "out" : "varying";
res << glu::getGLSLVersionDeclaration(spec.targetVersion) << "\n";
// Declarations (position + attribute/varying for each input).
res << "precision highp float;\n";
res << "precision highp int;\n";
res << "\n";
res << vtxIn << " highp vec4 dEQP_Position;\n";
for (size_t ndx = 0; ndx < spec.values.inputs.size(); ndx++)
{
const Value& val = spec.values.inputs[ndx];
const DataType basicType = val.type.getBasicType();
const DataType floatType = getDataTypeFloatScalars(basicType);
const char* const typeStr = getDataTypeName(floatType);
res << vtxIn << " " << typeStr << " a_" << val.name << ";\n";
if (getDataTypeScalarType(basicType) == TYPE_FLOAT)
res << vtxOut << " " << typeStr << " " << val.name << ";\n";
else
res << vtxOut << " " << typeStr << " v_" << val.name << ";\n";
}
res << "\n";
// Main function.
// - gl_Position = dEQP_Position;
// - for each input: write attribute directly to varying
res << "void main()\n";
res << "{\n";
res << " gl_Position = dEQP_Position;\n";
for (size_t ndx = 0; ndx < spec.values.inputs.size(); ndx++)
{
const Value& val = spec.values.inputs[ndx];
const string& name = val.name;
if (getDataTypeScalarType(val.type.getBasicType()) == TYPE_FLOAT)
res << " " << name << " = a_" << name << ";\n";
else
res << " v_" << name << " = a_" << name << ";\n";
}
res << "}\n";
return res.str();
}
static void genCompareOp (ostringstream& output, const char* dstVec4Var, const ValueBlock& valueBlock, const char* nonFloatNamePrefix, const char* checkVarName)
{
bool isFirstOutput = true;
for (size_t ndx = 0; ndx < valueBlock.outputs.size(); ndx++)
{
const Value& val = valueBlock.outputs[ndx];
// Check if we're only interested in one variable (then skip if not the right one).
if (checkVarName && val.name != checkVarName)
continue;
// Prefix.
if (isFirstOutput)
{
output << "bool RES = ";
isFirstOutput = false;
}
else
output << "RES = RES && ";
// Generate actual comparison.
if (getDataTypeScalarType(val.type.getBasicType()) == TYPE_FLOAT)
output << "isOk(" << val.name << ", ref_" << val.name << ", 0.05);\n";
else
output << "isOk(" << nonFloatNamePrefix << val.name << ", ref_" << val.name << ");\n";
}
if (isFirstOutput)
output << dstVec4Var << " = vec4(1.0);\n"; // \todo [petri] Should we give warning if not expect-failure case?
else
output << dstVec4Var << " = vec4(RES, RES, RES, 1.0);\n";
}
static inline bool supportsFragmentHighp (glu::GLSLVersion version)
{
return version != glu::GLSL_VERSION_100_ES;
}
static string genFragmentShader (const ShaderCaseSpecification& spec)
{
ostringstream shader;
const bool usesInout = glslVersionUsesInOutQualifiers(spec.targetVersion);
const bool customColorOut = usesInout;
const char* const fragIn = usesInout ? "in" : "varying";
const char* const prec = supportsFragmentHighp(spec.targetVersion) ? "highp" : "mediump";
shader << glu::getGLSLVersionDeclaration(spec.targetVersion) << "\n";
shader << "precision " << prec << " float;\n";
shader << "precision " << prec << " int;\n";
shader << "\n";
if (customColorOut)
{
shader << "layout(location = 0) out mediump vec4 dEQP_FragColor;\n";
shader << "\n";
}
genCompareFunctions(shader, spec.values, true);
shader << "\n";
// Declarations (varying, reference for each output).
for (size_t ndx = 0; ndx < spec.values.outputs.size(); ndx++)
{
const Value& val = spec.values.outputs[ndx];
const DataType basicType = val.type.getBasicType();
const DataType floatType = getDataTypeFloatScalars(basicType);
const char* const floatTypeStr = getDataTypeName(floatType);
const char* const refTypeStr = getDataTypeName(basicType);
if (getDataTypeScalarType(basicType) == TYPE_FLOAT)
shader << fragIn << " " << floatTypeStr << " " << val.name << ";\n";
else
shader << fragIn << " " << floatTypeStr << " v_" << val.name << ";\n";
shader << "uniform " << refTypeStr << " ref_" << val.name << ";\n";
}
shader << "\n";
shader << "void main()\n";
shader << "{\n";
shader << " ";
genCompareOp(shader, customColorOut ? "dEQP_FragColor" : "gl_FragColor", spec.values, "v_", DE_NULL);
shader << "}\n";
return shader.str();
}
// Specialize a shader for the vertex shader test case.
static string specializeVertexShader (const ShaderCaseSpecification& spec, const std::string& src, const vector<RequiredExtension>& extensions)
{
ostringstream decl;
ostringstream setup;
ostringstream output;
const bool usesInout = glslVersionUsesInOutQualifiers(spec.targetVersion);
const char* const vtxIn = usesInout ? "in" : "attribute";
const char* const vtxOut = usesInout ? "out" : "varying";
// generated from "both" case
DE_ASSERT(spec.caseType == CASETYPE_VERTEX_ONLY);
// Output (write out position).
output << "gl_Position = dEQP_Position;\n";
// Declarations (position + attribute for each input, varying for each output).
decl << vtxIn << " highp vec4 dEQP_Position;\n";
for (size_t ndx = 0; ndx < spec.values.inputs.size(); ndx++)
{
const Value& val = spec.values.inputs[ndx];
const DataType basicType = val.type.getBasicType();
const DataType floatType = getDataTypeFloatScalars(basicType);
const char* const floatTypeStr = getDataTypeName(floatType);
const char* const refTypeStr = getDataTypeName(basicType);
if (getDataTypeScalarType(basicType) == TYPE_FLOAT)
{
decl << vtxIn << " " << floatTypeStr << " " << val.name << ";\n";
}
else
{
decl << vtxIn << " " << floatTypeStr << " a_" << val.name << ";\n";
setup << refTypeStr << " " << val.name << " = " << refTypeStr << "(a_" << val.name << ");\n";
}
}
// \todo [2015-07-24 pyry] Why are uniforms missing?
for (size_t ndx = 0; ndx < spec.values.outputs.size(); ndx++)
{
const Value& val = spec.values.outputs[ndx];
const DataType basicType = val.type.getBasicType();
const DataType floatType = getDataTypeFloatScalars(basicType);
const char* const floatTypeStr = getDataTypeName(floatType);
const char* const refTypeStr = getDataTypeName(basicType);
if (getDataTypeScalarType(basicType) == TYPE_FLOAT)
decl << vtxOut << " " << floatTypeStr << " " << val.name << ";\n";
else
{
decl << vtxOut << " " << floatTypeStr << " v_" << val.name << ";\n";
decl << refTypeStr << " " << val.name << ";\n";
output << "v_" << val.name << " = " << floatTypeStr << "(" << val.name << ");\n";
}
}
// Shader specialization.
map<string, string> params;
params.insert(pair<string, string>("DECLARATIONS", decl.str()));
params.insert(pair<string, string>("SETUP", setup.str()));
params.insert(pair<string, string>("OUTPUT", output.str()));
params.insert(pair<string, string>("POSITION_FRAG_COLOR", "gl_Position"));
StringTemplate tmpl (src);
const string baseSrc = tmpl.specialize(params);
const string withExt = injectExtensionRequirements(baseSrc, extensions, SHADERTYPE_VERTEX);
return withExt;
}
// Specialize a shader for the fragment shader test case.
static string specializeFragmentShader (const ShaderCaseSpecification& spec, const std::string& src, const vector<RequiredExtension>& extensions)
{
ostringstream decl;
ostringstream setup;
ostringstream output;
const bool usesInout = glslVersionUsesInOutQualifiers(spec.targetVersion);
const bool customColorOut = usesInout;
const char* const fragIn = usesInout ? "in" : "varying";
const char* const fragColor = customColorOut ? "dEQP_FragColor" : "gl_FragColor";
// generated from "both" case
DE_ASSERT(spec.caseType == CASETYPE_FRAGMENT_ONLY);
genCompareFunctions(decl, spec.values, false);
genCompareOp(output, fragColor, spec.values, "", DE_NULL);
if (customColorOut)
decl << "layout(location = 0) out mediump vec4 dEQP_FragColor;\n";
for (size_t ndx = 0; ndx < spec.values.inputs.size(); ndx++)
{
const Value& val = spec.values.inputs[ndx];
const DataType basicType = val.type.getBasicType();
const DataType floatType = getDataTypeFloatScalars(basicType);
const char* const floatTypeStr = getDataTypeName(floatType);
const char* const refTypeStr = getDataTypeName(basicType);
if (getDataTypeScalarType(basicType) == TYPE_FLOAT)
decl << fragIn << " " << floatTypeStr << " " << val.name << ";\n";
else
{
decl << fragIn << " " << floatTypeStr << " v_" << val.name << ";\n";
std::string offset = isDataTypeIntOrIVec(basicType) ? " * 1.0025" : ""; // \todo [petri] bit of a hack to avoid errors in chop() due to varying interpolation
setup << refTypeStr << " " << val.name << " = " << refTypeStr << "(v_" << val.name << offset << ");\n";
}
}
// \todo [2015-07-24 pyry] Why are uniforms missing?
for (size_t ndx = 0; ndx < spec.values.outputs.size(); ndx++)
{
const Value& val = spec.values.outputs[ndx];
const DataType basicType = val.type.getBasicType();
const char* const refTypeStr = getDataTypeName(basicType);
decl << "uniform " << refTypeStr << " ref_" << val.name << ";\n";
decl << refTypeStr << " " << val.name << ";\n";
}
/* \todo [2010-04-01 petri] Check all outputs. */
// Shader specialization.
map<string, string> params;
params.insert(pair<string, string>("DECLARATIONS", decl.str()));
params.insert(pair<string, string>("SETUP", setup.str()));
params.insert(pair<string, string>("OUTPUT", output.str()));
params.insert(pair<string, string>("POSITION_FRAG_COLOR", fragColor));
StringTemplate tmpl (src);
const string baseSrc = tmpl.specialize(params);
const string withExt = injectExtensionRequirements(baseSrc, extensions, SHADERTYPE_FRAGMENT);
return withExt;
}
static void generateUniformDeclarations (std::ostream& dst, const ValueBlock& valueBlock)
{
for (size_t ndx = 0; ndx < valueBlock.uniforms.size(); ndx++)
{
const Value& val = valueBlock.uniforms[ndx];
const char* const typeStr = getDataTypeName(val.type.getBasicType());
if (val.name.find('.') == string::npos)
dst << "uniform " << typeStr << " " << val.name << ";\n";
}
}
static map<string, string> generateVertexSpecialization (const ProgramSpecializationParams& specParams)
{
const bool usesInout = glslVersionUsesInOutQualifiers(specParams.caseSpec.targetVersion);
const char* vtxIn = usesInout ? "in" : "attribute";
ostringstream decl;
ostringstream setup;
map<string, string> params;
decl << vtxIn << " highp vec4 dEQP_Position;\n";
for (size_t ndx = 0; ndx < specParams.caseSpec.values.inputs.size(); ndx++)
{
const Value& val = specParams.caseSpec.values.inputs[ndx];
const DataType basicType = val.type.getBasicType();
const char* const typeStr = getDataTypeName(val.type.getBasicType());
if (getDataTypeScalarType(basicType) == TYPE_FLOAT)
{
decl << vtxIn << " " << typeStr << " " << val.name << ";\n";
}
else
{
const DataType floatType = getDataTypeFloatScalars(basicType);
const char* const floatTypeStr = getDataTypeName(floatType);
decl << vtxIn << " " << floatTypeStr << " a_" << val.name << ";\n";
setup << typeStr << " " << val.name << " = " << typeStr << "(a_" << val.name << ");\n";
}
}
generateUniformDeclarations(decl, specParams.caseSpec.values);
params.insert(pair<string, string>("VERTEX_DECLARATIONS", decl.str()));
params.insert(pair<string, string>("VERTEX_SETUP", setup.str()));
params.insert(pair<string, string>("VERTEX_OUTPUT", string("gl_Position = dEQP_Position;\n")));
return params;
}
static map<string, string> generateFragmentSpecialization (const ProgramSpecializationParams& specParams)
{
const bool usesInout = glslVersionUsesInOutQualifiers(specParams.caseSpec.targetVersion);
const bool customColorOut = usesInout;
const char* const fragColor = customColorOut ? "dEQP_FragColor" : "gl_FragColor";
ostringstream decl;
ostringstream output;
map<string, string> params;
genCompareFunctions(decl, specParams.caseSpec.values, false);
genCompareOp(output, fragColor, specParams.caseSpec.values, "", DE_NULL);
if (customColorOut)
decl << "layout(location = 0) out mediump vec4 dEQP_FragColor;\n";
for (size_t ndx = 0; ndx < specParams.caseSpec.values.outputs.size(); ndx++)
{
const Value& val = specParams.caseSpec.values.outputs[ndx];
const char* const refTypeStr = getDataTypeName(val.type.getBasicType());
decl << "uniform " << refTypeStr << " ref_" << val.name << ";\n";
decl << refTypeStr << " " << val.name << ";\n";
}
generateUniformDeclarations(decl, specParams.caseSpec.values);
params.insert(pair<string, string>("FRAGMENT_DECLARATIONS", decl.str()));
params.insert(pair<string, string>("FRAGMENT_OUTPUT", output.str()));
params.insert(pair<string, string>("FRAG_COLOR", fragColor));
return params;
}
static map<string, string> generateGeometrySpecialization (const ProgramSpecializationParams& specParams)
{
ostringstream decl;
map<string, string> params;
decl << "layout (triangles) in;\n";
decl << "layout (triangle_strip, max_vertices=3) out;\n";
decl << "\n";
generateUniformDeclarations(decl, specParams.caseSpec.values);
params.insert(pair<string, string>("GEOMETRY_DECLARATIONS", decl.str()));
return params;
}
static map<string, string> generateTessControlSpecialization (const ProgramSpecializationParams& specParams)
{
ostringstream decl;
ostringstream output;
map<string, string> params;
decl << "layout (vertices=3) out;\n";
decl << "\n";
generateUniformDeclarations(decl, specParams.caseSpec.values);
output << "gl_out[gl_InvocationID].gl_Position = gl_in[gl_InvocationID].gl_Position;\n"
"gl_TessLevelInner[0] = 2.0;\n"
"gl_TessLevelInner[1] = 2.0;\n"
"gl_TessLevelOuter[0] = 2.0;\n"
"gl_TessLevelOuter[1] = 2.0;\n"
"gl_TessLevelOuter[2] = 2.0;\n"
"gl_TessLevelOuter[3] = 2.0;";
params.insert(pair<string, string>("TESSELLATION_CONTROL_DECLARATIONS", decl.str()));
params.insert(pair<string, string>("TESSELLATION_CONTROL_OUTPUT", output.str()));
params.insert(pair<string, string>("GL_MAX_PATCH_VERTICES", de::toString(specParams.maxPatchVertices)));
return params;
}
static map<string, string> generateTessEvalSpecialization (const ProgramSpecializationParams& specParams)
{
ostringstream decl;
ostringstream output;
map<string, string> params;
decl << "layout (triangles) in;\n";
decl << "\n";
generateUniformDeclarations(decl, specParams.caseSpec.values);
output << "gl_Position = gl_TessCoord[0] * gl_in[0].gl_Position + gl_TessCoord[1] * gl_in[1].gl_Position + gl_TessCoord[2] * gl_in[2].gl_Position;\n";
params.insert(pair<string, string>("TESSELLATION_EVALUATION_DECLARATIONS", decl.str()));
params.insert(pair<string, string>("TESSELLATION_EVALUATION_OUTPUT", output.str()));
params.insert(pair<string, string>("GL_MAX_PATCH_VERTICES", de::toString(specParams.maxPatchVertices)));
return params;
}
static void specializeShaderSources (ProgramSources& dst,
const ProgramSources& src,
const ProgramSpecializationParams& specParams,
glu::ShaderType shaderType,
map<string, string> (*specializationGenerator) (const ProgramSpecializationParams& specParams))
{
if (!src.sources[shaderType].empty())
{
const map<string, string> tmplParams = specializationGenerator(specParams);
for (size_t ndx = 0; ndx < src.sources[shaderType].size(); ++ndx)
{
const StringTemplate tmpl (src.sources[shaderType][ndx]);
const std::string baseGLSLCode = tmpl.specialize(tmplParams);
const std::string sourceWithExts = injectExtensionRequirements(baseGLSLCode, specParams.requiredExtensions, shaderType);
dst << glu::ShaderSource(shaderType, sourceWithExts);
}
}
}
static void specializeProgramSources (glu::ProgramSources& dst,
const glu::ProgramSources& src,
const ProgramSpecializationParams& specParams)
{
specializeShaderSources(dst, src, specParams, SHADERTYPE_VERTEX, generateVertexSpecialization);
specializeShaderSources(dst, src, specParams, SHADERTYPE_FRAGMENT, generateFragmentSpecialization);
specializeShaderSources(dst, src, specParams, SHADERTYPE_GEOMETRY, generateGeometrySpecialization);
specializeShaderSources(dst, src, specParams, SHADERTYPE_TESSELLATION_CONTROL, generateTessControlSpecialization);
specializeShaderSources(dst, src, specParams, SHADERTYPE_TESSELLATION_EVALUATION, generateTessEvalSpecialization);
dst << ProgramSeparable(src.separable);
}
enum
{
VIEWPORT_WIDTH = 128,
VIEWPORT_HEIGHT = 128
};
class BeforeDrawValidator : public glu::DrawUtilCallback
{
public:
enum TargetType
{
TARGETTYPE_PROGRAM = 0,
TARGETTYPE_PIPELINE,
TARGETTYPE_LAST
};
BeforeDrawValidator (const glw::Functions& gl, glw::GLuint target, TargetType targetType);
void beforeDrawCall (void);
const std::string& getInfoLog (void) const;
glw::GLint getValidateStatus (void) const;
private:
const glw::Functions& m_gl;
const glw::GLuint m_target;
const TargetType m_targetType;
glw::GLint m_validateStatus;
std::string m_logMessage;
};
BeforeDrawValidator::BeforeDrawValidator (const glw::Functions& gl, glw::GLuint target, TargetType targetType)
: m_gl (gl)
, m_target (target)
, m_targetType (targetType)
, m_validateStatus (-1)
{
DE_ASSERT(targetType < TARGETTYPE_LAST);
}
void BeforeDrawValidator::beforeDrawCall (void)
{
glw::GLint bytesWritten = 0;
glw::GLint infoLogLength;
std::vector<glw::GLchar> logBuffer;
int stringLength;
// validate
if (m_targetType == TARGETTYPE_PROGRAM)
m_gl.validateProgram(m_target);
else if (m_targetType == TARGETTYPE_PIPELINE)
m_gl.validateProgramPipeline(m_target);
else
DE_ASSERT(false);
GLU_EXPECT_NO_ERROR(m_gl.getError(), "validate");
// check status
m_validateStatus = -1;
if (m_targetType == TARGETTYPE_PROGRAM)
m_gl.getProgramiv(m_target, GL_VALIDATE_STATUS, &m_validateStatus);
else if (m_targetType == TARGETTYPE_PIPELINE)
m_gl.getProgramPipelineiv(m_target, GL_VALIDATE_STATUS, &m_validateStatus);
else
DE_ASSERT(false);
GLU_EXPECT_NO_ERROR(m_gl.getError(), "get validate status");
TCU_CHECK(m_validateStatus == GL_TRUE || m_validateStatus == GL_FALSE);
// read log
infoLogLength = 0;
if (m_targetType == TARGETTYPE_PROGRAM)
m_gl.getProgramiv(m_target, GL_INFO_LOG_LENGTH, &infoLogLength);
else if (m_targetType == TARGETTYPE_PIPELINE)
m_gl.getProgramPipelineiv(m_target, GL_INFO_LOG_LENGTH, &infoLogLength);
else
DE_ASSERT(false);
GLU_EXPECT_NO_ERROR(m_gl.getError(), "get info log length");
if (infoLogLength <= 0)
{
m_logMessage.clear();
return;
}
logBuffer.resize(infoLogLength + 2, '0'); // +1 for zero terminator (infoLogLength should include it, but better play it safe), +1 to make sure buffer is always larger
if (m_targetType == TARGETTYPE_PROGRAM)
m_gl.getProgramInfoLog(m_target, infoLogLength + 1, &bytesWritten, &logBuffer[0]);
else if (m_targetType == TARGETTYPE_PIPELINE)
m_gl.getProgramPipelineInfoLog(m_target, infoLogLength + 1, &bytesWritten, &logBuffer[0]);
else
DE_ASSERT(false);
// just ignore bytesWritten to be safe, find the null terminator
stringLength = (int)(std::find(logBuffer.begin(), logBuffer.end(), '0') - logBuffer.begin());
m_logMessage.assign(&logBuffer[0], stringLength);
}
const std::string& BeforeDrawValidator::getInfoLog (void) const
{
return m_logMessage;
}
glw::GLint BeforeDrawValidator::getValidateStatus (void) const
{
return m_validateStatus;
}
// ShaderCase.
ShaderLibraryCase::ShaderLibraryCase (tcu::TestContext& testCtx, RenderContext& renderCtx, const glu::ContextInfo& contextInfo, const char* name, const char* description, const ShaderCaseSpecification& specification)
: tcu::TestCase (testCtx, name, description)
, m_renderCtx (renderCtx)
, m_contextInfo (contextInfo)
, m_spec (specification)
{
}
ShaderLibraryCase::~ShaderLibraryCase (void)
{
}
void ShaderLibraryCase::init (void)
{
DE_ASSERT(isValid(m_spec));
if (!isGLSLVersionSupported(m_renderCtx.getType(), m_spec.targetVersion))
TCU_THROW(NotSupportedError, (string(getGLSLVersionName(m_spec.targetVersion)) + " is not supported").c_str());
checkImplementationLimits(m_spec.requiredCaps, m_contextInfo);
// log the expected result
switch (m_spec.expectResult)
{
case EXPECT_PASS:
// Don't write anything
break;
case EXPECT_COMPILE_FAIL:
m_testCtx.getLog() << tcu::TestLog::Message << "Expecting shader compilation to fail." << tcu::TestLog::EndMessage;
break;
case EXPECT_LINK_FAIL:
m_testCtx.getLog() << tcu::TestLog::Message << "Expecting program linking to fail." << tcu::TestLog::EndMessage;
break;
case EXPECT_COMPILE_LINK_FAIL:
m_testCtx.getLog() << tcu::TestLog::Message << "Expecting either shader compilation or program linking to fail." << tcu::TestLog::EndMessage;
break;
case EXPECT_VALIDATION_FAIL:
m_testCtx.getLog() << tcu::TestLog::Message << "Expecting program validation to fail." << tcu::TestLog::EndMessage;
break;
case EXPECT_BUILD_SUCCESSFUL:
m_testCtx.getLog() << tcu::TestLog::Message << "Expecting shader compilation and program linking to succeed. Resulting program will not be executed." << tcu::TestLog::EndMessage;
break;
default:
DE_ASSERT(false);
break;
}
}
static void setUniformValue (const glw::Functions& gl, const std::vector<deUint32>& pipelinePrograms, const std::string& name, const Value& val, int arrayNdx, tcu::TestLog& log)
{
bool foundAnyMatch = false;
for (int programNdx = 0; programNdx < (int)pipelinePrograms.size(); ++programNdx)
{
const DataType dataType = val.type.getBasicType();
const int scalarSize = getDataTypeScalarSize(dataType);
const int loc = gl.getUniformLocation(pipelinePrograms[programNdx], name.c_str());
const int elemNdx = arrayNdx * scalarSize;
DE_ASSERT(elemNdx+scalarSize <= (int)val.elements.size());
if (loc == -1)
continue;
foundAnyMatch = true;
DE_STATIC_ASSERT(sizeof(Value::Element) == sizeof(glw::GLfloat));
DE_STATIC_ASSERT(sizeof(Value::Element) == sizeof(glw::GLint));
gl.useProgram(pipelinePrograms[programNdx]);
switch (dataType)
{
case TYPE_FLOAT: gl.uniform1fv(loc, 1, &val.elements[elemNdx].float32); break;
case TYPE_FLOAT_VEC2: gl.uniform2fv(loc, 1, &val.elements[elemNdx].float32); break;
case TYPE_FLOAT_VEC3: gl.uniform3fv(loc, 1, &val.elements[elemNdx].float32); break;
case TYPE_FLOAT_VEC4: gl.uniform4fv(loc, 1, &val.elements[elemNdx].float32); break;
case TYPE_FLOAT_MAT2: gl.uniformMatrix2fv(loc, 1, GL_FALSE, &val.elements[elemNdx].float32); break;
case TYPE_FLOAT_MAT3: gl.uniformMatrix3fv(loc, 1, GL_FALSE, &val.elements[elemNdx].float32); break;
case TYPE_FLOAT_MAT4: gl.uniformMatrix4fv(loc, 1, GL_FALSE, &val.elements[elemNdx].float32); break;
case TYPE_INT: gl.uniform1iv(loc, 1, &val.elements[elemNdx].int32); break;
case TYPE_INT_VEC2: gl.uniform2iv(loc, 1, &val.elements[elemNdx].int32); break;
case TYPE_INT_VEC3: gl.uniform3iv(loc, 1, &val.elements[elemNdx].int32); break;
case TYPE_INT_VEC4: gl.uniform4iv(loc, 1, &val.elements[elemNdx].int32); break;
case TYPE_BOOL: gl.uniform1iv(loc, 1, &val.elements[elemNdx].int32); break;
case TYPE_BOOL_VEC2: gl.uniform2iv(loc, 1, &val.elements[elemNdx].int32); break;
case TYPE_BOOL_VEC3: gl.uniform3iv(loc, 1, &val.elements[elemNdx].int32); break;
case TYPE_BOOL_VEC4: gl.uniform4iv(loc, 1, &val.elements[elemNdx].int32); break;
case TYPE_UINT: gl.uniform1uiv(loc, 1, (const deUint32*)&val.elements[elemNdx].int32); break;
case TYPE_UINT_VEC2: gl.uniform2uiv(loc, 1, (const deUint32*)&val.elements[elemNdx].int32); break;
case TYPE_UINT_VEC3: gl.uniform3uiv(loc, 1, (const deUint32*)&val.elements[elemNdx].int32); break;
case TYPE_UINT_VEC4: gl.uniform4uiv(loc, 1, (const deUint32*)&val.elements[elemNdx].int32); break;
case TYPE_FLOAT_MAT2X3: gl.uniformMatrix2x3fv(loc, 1, GL_FALSE, &val.elements[elemNdx].float32); break;
case TYPE_FLOAT_MAT2X4: gl.uniformMatrix2x4fv(loc, 1, GL_FALSE, &val.elements[elemNdx].float32); break;
case TYPE_FLOAT_MAT3X2: gl.uniformMatrix3x2fv(loc, 1, GL_FALSE, &val.elements[elemNdx].float32); break;
case TYPE_FLOAT_MAT3X4: gl.uniformMatrix3x4fv(loc, 1, GL_FALSE, &val.elements[elemNdx].float32); break;
case TYPE_FLOAT_MAT4X2: gl.uniformMatrix4x2fv(loc, 1, GL_FALSE, &val.elements[elemNdx].float32); break;
case TYPE_FLOAT_MAT4X3: gl.uniformMatrix4x3fv(loc, 1, GL_FALSE, &val.elements[elemNdx].float32); break;
case TYPE_SAMPLER_2D:
case TYPE_SAMPLER_CUBE:
DE_FATAL("implement!");
break;
default:
DE_ASSERT(false);
}
}
if (!foundAnyMatch)
log << tcu::TestLog::Message << "WARNING // Uniform \"" << name << "\" location is not valid, location = -1. Cannot set value to the uniform." << tcu::TestLog::EndMessage;
}
static bool isTessellationPresent (const ShaderCaseSpecification& spec)
{
if (spec.programs[0].sources.separable)
{
const deUint32 tessellationBits = (1 << glu::SHADERTYPE_TESSELLATION_CONTROL) |
(1 << glu::SHADERTYPE_TESSELLATION_EVALUATION);
for (int programNdx = 0; programNdx < (int)spec.programs.size(); ++programNdx)
if (spec.programs[programNdx].activeStages & tessellationBits)
return true;
return false;
}
else
return !spec.programs[0].sources.sources[glu::SHADERTYPE_TESSELLATION_CONTROL].empty() ||
!spec.programs[0].sources.sources[glu::SHADERTYPE_TESSELLATION_EVALUATION].empty();
}
static bool isTessellationSupported (const glu::RenderContext& renderCtx, const glu::ContextInfo& ctxInfo)
{
if (renderCtx.getType().getProfile() == PROFILE_ES)
{
const int majorVer = renderCtx.getType().getMajorVersion();
const int minorVer = renderCtx.getType().getMinorVersion();
return (majorVer > 3) || (majorVer == 3 && minorVer >= 2) ||
ctxInfo.isExtensionSupported("GL_EXT_tessellation_shader");
}
else
return false;
}
static bool checkPixels (tcu::TestLog& log, const tcu::ConstPixelBufferAccess& surface)
{
bool allWhite = true;
bool allBlack = true;
bool anyUnexpected = false;
for (int y = 0; y < surface.getHeight(); y++)
{
for (int x = 0; x < surface.getWidth(); x++)
{
const tcu::IVec4 pixel = surface.getPixelInt(x, y);
// Note: we really do not want to involve alpha in the check comparison
// \todo [2010-09-22 kalle] Do we know that alpha would be one? If yes, could use color constants white and black.
const bool isWhite = (pixel[0] == 255) && (pixel[1] == 255) && (pixel[2] == 255);
const bool isBlack = (pixel[0] == 0) && (pixel[1] == 0) && (pixel[2] == 0);
allWhite = allWhite && isWhite;
allBlack = allBlack && isBlack;
anyUnexpected = anyUnexpected || (!isWhite && !isBlack);
}
}
if (!allWhite)
{
if (anyUnexpected)
log << TestLog::Message << "WARNING: expecting all rendered pixels to be white or black, but got other colors as well!" << TestLog::EndMessage;
else if (!allBlack)
log << TestLog::Message << "WARNING: got inconsistent results over the image, when all pixels should be the same color!" << TestLog::EndMessage;
return false;
}
return true;
}
bool ShaderLibraryCase::execute (void)
{
const float quadSize = 1.0f;
static const float s_positions[4*4] =
{
-quadSize, -quadSize, 0.0f, 1.0f,
-quadSize, +quadSize, 0.0f, 1.0f,
+quadSize, -quadSize, 0.0f, 1.0f,
+quadSize, +quadSize, 0.0f, 1.0f
};
static const deUint16 s_indices[2*3] =
{
0, 1, 2,
1, 3, 2
};
TestLog& log = m_testCtx.getLog();
const glw::Functions& gl = m_renderCtx.getFunctions();
// Compute viewport.
const tcu::RenderTarget& renderTarget = m_renderCtx.getRenderTarget();
de::Random rnd (deStringHash(getName()));
const int width = deMin32(renderTarget.getWidth(), VIEWPORT_WIDTH);
const int height = deMin32(renderTarget.getHeight(), VIEWPORT_HEIGHT);
const int viewportX = rnd.getInt(0, renderTarget.getWidth() - width);
const int viewportY = rnd.getInt(0, renderTarget.getHeight() - height);
const int numVerticesPerDraw = 4;
const bool tessellationPresent = isTessellationPresent(m_spec);
const bool separablePrograms = m_spec.programs[0].sources.separable;
bool allCompilesOk = true;
bool allLinksOk = true;
const char* failReason = DE_NULL;
vector<ProgramSources> specializedSources (m_spec.programs.size());
deUint32 vertexProgramID = -1;
vector<deUint32> pipelineProgramIDs;
vector<SharedPtr<ShaderProgram> > programs;
SharedPtr<ProgramPipeline> programPipeline;
GLU_EXPECT_NO_ERROR(gl.getError(), "ShaderCase::execute(): start");
if(isCapabilityRequired(CAPABILITY_ONLY_GLSL_ES_100_SUPPORT, m_spec) && glu::IsES3Compatible(gl))
return true;
if(isCapabilityRequired(CAPABILITY_EXACTLY_ONE_DRAW_BUFFER, m_spec))
{
// on unextended ES2 there is only one draw buffer
// and there is no GL_MAX_DRAW_BUFFERS query
glw::GLint maxDrawBuffers = 0;
gl.getIntegerv(GL_MAX_DRAW_BUFFERS, &maxDrawBuffers);
if ((gl.getError() == GL_NO_ERROR) && (maxDrawBuffers > 1))
throw tcu::NotSupportedError("Test requires exactly one draw buffer");
}
// Specialize shaders
if (m_spec.caseType == CASETYPE_VERTEX_ONLY)
{
const vector<RequiredExtension> reqExt = checkAndSpecializeExtensions(m_spec.programs[0].requiredExtensions, m_contextInfo);
DE_ASSERT(m_spec.programs.size() == 1 && m_spec.programs[0].sources.sources[SHADERTYPE_VERTEX].size() == 1);
specializedSources[0] << glu::VertexSource(specializeVertexShader(m_spec, m_spec.programs[0].sources.sources[SHADERTYPE_VERTEX][0], reqExt))
<< glu::FragmentSource(genFragmentShader(m_spec));
}
else if (m_spec.caseType == CASETYPE_FRAGMENT_ONLY)
{
const vector<RequiredExtension> reqExt = checkAndSpecializeExtensions(m_spec.programs[0].requiredExtensions, m_contextInfo);
DE_ASSERT(m_spec.programs.size() == 1 && m_spec.programs[0].sources.sources[SHADERTYPE_FRAGMENT].size() == 1);
specializedSources[0] << glu::VertexSource(genVertexShader(m_spec))
<< glu::FragmentSource(specializeFragmentShader(m_spec, m_spec.programs[0].sources.sources[SHADERTYPE_FRAGMENT][0], reqExt));
}
else
{
DE_ASSERT(m_spec.caseType == CASETYPE_COMPLETE);
const int maxPatchVertices = isTessellationPresent(m_spec) && isTessellationSupported(m_renderCtx, m_contextInfo)
? m_contextInfo.getInt(GL_MAX_PATCH_VERTICES) : 0;
for (size_t progNdx = 0; progNdx < m_spec.programs.size(); progNdx++)
{
const ProgramSpecializationParams progSpecParams (m_spec, checkAndSpecializeExtensions(m_spec.programs[progNdx].requiredExtensions, m_contextInfo), maxPatchVertices);
specializeProgramSources(specializedSources[progNdx], m_spec.programs[progNdx].sources, progSpecParams);
}
}
if (!separablePrograms)
{
de::SharedPtr<glu::ShaderProgram> program (new glu::ShaderProgram(m_renderCtx, specializedSources[0]));
vertexProgramID = program->getProgram();
pipelineProgramIDs.push_back(program->getProgram());
programs.push_back(program);
// Check that compile/link results are what we expect.
DE_STATIC_ASSERT(glu::SHADERTYPE_VERTEX == 0);
for (int stage = glu::SHADERTYPE_VERTEX; stage < glu::SHADERTYPE_LAST; ++stage)
if (program->hasShader((glu::ShaderType)stage) && !program->getShaderInfo((glu::ShaderType)stage).compileOk)
allCompilesOk = false;
if (!program->getProgramInfo().linkOk)
allLinksOk = false;
log << *program;
}
else
{
// Separate programs
for (size_t programNdx = 0; programNdx < m_spec.programs.size(); ++programNdx)
{
de::SharedPtr<glu::ShaderProgram> program(new glu::ShaderProgram(m_renderCtx, specializedSources[programNdx]));
if (m_spec.programs[programNdx].activeStages & (1u << glu::SHADERTYPE_VERTEX))
vertexProgramID = program->getProgram();
pipelineProgramIDs.push_back(program->getProgram());
programs.push_back(program);
// Check that compile/link results are what we expect.
DE_STATIC_ASSERT(glu::SHADERTYPE_VERTEX == 0);
for (int stage = glu::SHADERTYPE_VERTEX; stage < glu::SHADERTYPE_LAST; ++stage)
if (program->hasShader((glu::ShaderType)stage) && !program->getShaderInfo((glu::ShaderType)stage).compileOk)
allCompilesOk = false;
if (!program->getProgramInfo().linkOk)
allLinksOk = false;
// Log program and active stages
{
const tcu::ScopedLogSection section (log, "Program", "Program " + de::toString(programNdx+1));
tcu::MessageBuilder builder (&log);
bool firstStage = true;
builder << "Pipeline uses stages: ";
for (int stage = glu::SHADERTYPE_VERTEX; stage < glu::SHADERTYPE_LAST; ++stage)
{
if (m_spec.programs[programNdx].activeStages & (1u << stage))
{
if (!firstStage)
builder << ", ";
builder << glu::getShaderTypeName((glu::ShaderType)stage);
firstStage = true;
}
}
builder << tcu::TestLog::EndMessage;
log << *program;
}
}
}
switch (m_spec.expectResult)
{
case EXPECT_PASS:
case EXPECT_VALIDATION_FAIL:
case EXPECT_BUILD_SUCCESSFUL:
if (!allCompilesOk)
failReason = "expected shaders to compile and link properly, but failed to compile.";
else if (!allLinksOk)
failReason = "expected shaders to compile and link properly, but failed to link.";
break;
case EXPECT_COMPILE_FAIL:
if (allCompilesOk && !allLinksOk)
failReason = "expected compilation to fail, but shaders compiled and link failed.";
else if (allCompilesOk)
failReason = "expected compilation to fail, but shaders compiled correctly.";
break;
case EXPECT_LINK_FAIL:
if (!allCompilesOk)
failReason = "expected linking to fail, but unable to compile.";
else if (allLinksOk)
failReason = "expected linking to fail, but passed.";
break;
case EXPECT_COMPILE_LINK_FAIL:
if (allCompilesOk && allLinksOk)
failReason = "expected compile or link to fail, but passed.";
break;
default:
DE_ASSERT(false);
return false;
}
if (failReason != DE_NULL)
{
// \todo [2010-06-07 petri] These should be handled in the test case?
log << TestLog::Message << "ERROR: " << failReason << TestLog::EndMessage;
if (isCapabilityRequired(CAPABILITY_FULL_GLSL_ES_100_SUPPORT, m_spec))
{
log << TestLog::Message
<< "Assuming build failure is caused by implementation not supporting full GLSL ES 100 specification, which is not required."
<< TestLog::EndMessage;
if (allCompilesOk && !allLinksOk)
{
// Used features are detectable at compile time. If implementation parses shader
// at link time, report it as quality warning.
m_testCtx.setTestResult(QP_TEST_RESULT_QUALITY_WARNING, failReason);
}
else
m_testCtx.setTestResult(QP_TEST_RESULT_NOT_SUPPORTED, "Full GLSL ES 100 is not supported");
}
else if (m_spec.expectResult == EXPECT_COMPILE_FAIL && allCompilesOk && !allLinksOk)
{
// If implementation parses shader at link time, report it as quality warning.
m_testCtx.setTestResult(QP_TEST_RESULT_QUALITY_WARNING, failReason);
}
else
m_testCtx.setTestResult(QP_TEST_RESULT_FAIL, failReason);
return false;
}
// Return if shader is not intended to be run
if (m_spec.expectResult == EXPECT_COMPILE_FAIL ||
m_spec.expectResult == EXPECT_COMPILE_LINK_FAIL ||
m_spec.expectResult == EXPECT_LINK_FAIL ||
m_spec.expectResult == EXPECT_BUILD_SUCCESSFUL)
return true;
// Setup viewport.
gl.viewport(viewportX, viewportY, width, height);
if (separablePrograms)
{
programPipeline = de::SharedPtr<glu::ProgramPipeline>(new glu::ProgramPipeline(m_renderCtx));
// Setup pipeline
gl.bindProgramPipeline(programPipeline->getPipeline());
for (int programNdx = 0; programNdx < (int)m_spec.programs.size(); ++programNdx)
{
deUint32 shaderFlags = 0;
for (int stage = glu::SHADERTYPE_VERTEX; stage < glu::SHADERTYPE_LAST; ++stage)
if (m_spec.programs[programNdx].activeStages & (1u << stage))
shaderFlags |= glu::getGLShaderTypeBit((glu::ShaderType)stage);
programPipeline->useProgramStages(shaderFlags, pipelineProgramIDs[programNdx]);
}
programPipeline->activeShaderProgram(vertexProgramID);
GLU_EXPECT_NO_ERROR(gl.getError(), "setup pipeline");
}
else
{
// Start using program
gl.useProgram(vertexProgramID);
GLU_EXPECT_NO_ERROR(gl.getError(), "glUseProgram()");
}
// Fetch location for positions positions.
int positionLoc = gl.getAttribLocation(vertexProgramID, "dEQP_Position");
if (positionLoc == -1)
{
string errStr = string("no location found for attribute 'dEQP_Position'");
TCU_FAIL(errStr.c_str());
}
// Iterate all value blocks.
{
const ValueBlock& valueBlock = m_spec.values;
// always render at least one pass even if there is no input/output data
const int numRenderPasses = valueBlock.outputs.empty() ? 1 : (int)valueBlock.outputs[0].elements.size() / valueBlock.outputs[0].type.getScalarSize();
// Iterate all array sub-cases.
for (int arrayNdx = 0; arrayNdx < numRenderPasses; arrayNdx++)
{
vector<VertexArrayBinding> vertexArrays;
int attribValueNdx = 0;
vector<vector<float> > attribValues (valueBlock.inputs.size());
glw::GLenum postDrawError;
BeforeDrawValidator beforeDrawValidator (gl,
(separablePrograms) ? (programPipeline->getPipeline()) : (vertexProgramID),
(separablePrograms) ? (BeforeDrawValidator::TARGETTYPE_PIPELINE) : (BeforeDrawValidator::TARGETTYPE_PROGRAM));
vertexArrays.push_back(va::Float(positionLoc, 4, numVerticesPerDraw, 0, &s_positions[0]));
// Collect VA pointer for inputs
for (size_t valNdx = 0; valNdx < valueBlock.inputs.size(); valNdx++)
{
const Value& val = valueBlock.inputs[valNdx];
const char* const valueName = val.name.c_str();
const DataType dataType = val.type.getBasicType();
const int scalarSize = getDataTypeScalarSize(dataType);
// Replicate values four times.
std::vector<float>& scalars = attribValues[attribValueNdx++];
scalars.resize(numVerticesPerDraw * scalarSize);
if (isDataTypeFloatOrVec(dataType) || isDataTypeMatrix(dataType))
{
for (int repNdx = 0; repNdx < numVerticesPerDraw; repNdx++)
for (int ndx = 0; ndx < scalarSize; ndx++)
scalars[repNdx*scalarSize + ndx] = val.elements[arrayNdx*scalarSize + ndx].float32;
}
else
{
// convert to floats.
for (int repNdx = 0; repNdx < numVerticesPerDraw; repNdx++)
{
for (int ndx = 0; ndx < scalarSize; ndx++)
{
float v = (float)val.elements[arrayNdx*scalarSize + ndx].int32;
DE_ASSERT(val.elements[arrayNdx*scalarSize + ndx].int32 == (int)v);
scalars[repNdx*scalarSize + ndx] = v;
}
}
}
// Attribute name prefix.
string attribPrefix = "";
// \todo [2010-05-27 petri] Should latter condition only apply for vertex cases (or actually non-fragment cases)?
if ((m_spec.caseType == CASETYPE_FRAGMENT_ONLY) || (getDataTypeScalarType(dataType) != TYPE_FLOAT))
attribPrefix = "a_";
// Input always given as attribute.
string attribName = attribPrefix + valueName;
int attribLoc = gl.getAttribLocation(vertexProgramID, attribName.c_str());
if (attribLoc == -1)
{
log << TestLog::Message << "Warning: no location found for attribute '" << attribName << "'" << TestLog::EndMessage;
continue;
}
if (isDataTypeMatrix(dataType))
{
int numCols = getDataTypeMatrixNumColumns(dataType);
int numRows = getDataTypeMatrixNumRows(dataType);
DE_ASSERT(scalarSize == numCols*numRows);
for (int i = 0; i < numCols; i++)
vertexArrays.push_back(va::Float(attribLoc + i, numRows, numVerticesPerDraw, scalarSize*(int)sizeof(float), &scalars[i * numRows]));
}
else
{
DE_ASSERT(isDataTypeFloatOrVec(dataType) || isDataTypeIntOrIVec(dataType) || isDataTypeUintOrUVec(dataType) || isDataTypeBoolOrBVec(dataType));
vertexArrays.push_back(va::Float(attribLoc, scalarSize, numVerticesPerDraw, 0, &scalars[0]));
}
GLU_EXPECT_NO_ERROR(gl.getError(), "set vertex attrib array");
}
GLU_EXPECT_NO_ERROR(gl.getError(), "before set uniforms");
// set reference values for outputs.
for (size_t valNdx = 0; valNdx < valueBlock.outputs.size(); valNdx++)
{
const Value& val = valueBlock.outputs[valNdx];
const char* const valueName = val.name.c_str();
// Set reference value.
string refName = string("ref_") + valueName;
setUniformValue(gl, pipelineProgramIDs, refName, val, arrayNdx, m_testCtx.getLog());
GLU_EXPECT_NO_ERROR(gl.getError(), "set reference uniforms");
}
// set uniform values
for (size_t valNdx = 0; valNdx < valueBlock.uniforms.size(); valNdx++)
{
const Value& val = valueBlock.uniforms[valNdx];
const char* const valueName = val.name.c_str();
setUniformValue(gl, pipelineProgramIDs, valueName, val, arrayNdx, m_testCtx.getLog());
GLU_EXPECT_NO_ERROR(gl.getError(), "set uniforms");
}
// Clear.
gl.clearColor(0.125f, 0.25f, 0.5f, 1.0f);
gl.clear(GL_COLOR_BUFFER_BIT);
GLU_EXPECT_NO_ERROR(gl.getError(), "clear buffer");
// Use program or pipeline
if (separablePrograms)
gl.useProgram(0);
else
gl.useProgram(vertexProgramID);
// Draw.
if (tessellationPresent)
{
gl.patchParameteri(GL_PATCH_VERTICES, 3);
GLU_EXPECT_NO_ERROR(gl.getError(), "set patchParameteri(PATCH_VERTICES, 3)");
}
draw(m_renderCtx,
vertexProgramID,
(int)vertexArrays.size(),
&vertexArrays[0],
(tessellationPresent) ?
(pr::Patches(DE_LENGTH_OF_ARRAY(s_indices), &s_indices[0])) :
(pr::Triangles(DE_LENGTH_OF_ARRAY(s_indices), &s_indices[0])),
(m_spec.expectResult == EXPECT_VALIDATION_FAIL) ?
(&beforeDrawValidator) :
(DE_NULL));
postDrawError = gl.getError();
if (m_spec.expectResult == EXPECT_PASS)
{
// Read back results.
Surface surface (width, height);
const float w = s_positions[3];
const int minY = deCeilFloatToInt32 (((-quadSize / w) * 0.5f + 0.5f) * (float)height + 1.0f);
const int maxY = deFloorFloatToInt32(((+quadSize / w) * 0.5f + 0.5f) * (float)height - 0.5f);
const int minX = deCeilFloatToInt32 (((-quadSize / w) * 0.5f + 0.5f) * (float)width + 1.0f);
const int maxX = deFloorFloatToInt32(((+quadSize / w) * 0.5f + 0.5f) * (float)width - 0.5f);
GLU_EXPECT_NO_ERROR(postDrawError, "draw");
glu::readPixels(m_renderCtx, viewportX, viewportY, surface.getAccess());
GLU_EXPECT_NO_ERROR(gl.getError(), "read pixels");
if (!checkPixels(log, tcu::getSubregion(surface.getAccess(), minX, minY, maxX-minX+1, maxY-minY+1)))
{
log << TestLog::Message << "INCORRECT RESULT for sub-case " << arrayNdx+1 << " of " << numRenderPasses << "):"
<< TestLog::EndMessage;
log << TestLog::Message << "Failing shader input/output values:" << TestLog::EndMessage;
dumpValues(log, valueBlock, arrayNdx);
// Dump image on failure.
log << TestLog::Image("Result", "Rendered result image", surface);
gl.useProgram(0);
m_testCtx.setTestResult(QP_TEST_RESULT_FAIL, "Image comparison failed");
return false;
}
}
else if (m_spec.expectResult == EXPECT_VALIDATION_FAIL)
{
log << TestLog::Message
<< "Draw call generated error: "
<< glu::getErrorStr(postDrawError) << " "
<< ((postDrawError == GL_INVALID_OPERATION) ? ("(expected)") : ("(unexpected)")) << "\n"
<< "Validate status: "
<< glu::getBooleanStr(beforeDrawValidator.getValidateStatus()) << " "
<< ((beforeDrawValidator.getValidateStatus() == GL_FALSE) ? ("(expected)") : ("(unexpected)")) << "\n"
<< "Info log: "
<< ((beforeDrawValidator.getInfoLog().empty()) ? ("[empty string]") : (beforeDrawValidator.getInfoLog())) << "\n"
<< TestLog::EndMessage;
// test result
if (postDrawError != GL_NO_ERROR && postDrawError != GL_INVALID_OPERATION)
{
m_testCtx.setTestResult(QP_TEST_RESULT_FAIL, ("Draw: got unexpected error: " + de::toString(glu::getErrorStr(postDrawError))).c_str());
return false;
}
if (beforeDrawValidator.getValidateStatus() == GL_TRUE)
{
if (postDrawError == GL_NO_ERROR)
m_testCtx.setTestResult(QP_TEST_RESULT_FAIL, "expected validation and rendering to fail but validation and rendering succeeded");
else if (postDrawError == GL_INVALID_OPERATION)
m_testCtx.setTestResult(QP_TEST_RESULT_FAIL, "expected validation and rendering to fail but validation succeeded (rendering failed as expected)");
else
DE_ASSERT(false);
return false;
}
else if (beforeDrawValidator.getValidateStatus() == GL_FALSE && postDrawError == GL_NO_ERROR)
{
m_testCtx.setTestResult(QP_TEST_RESULT_FAIL, "expected validation and rendering to fail but rendering succeeded (validation failed as expected)");
return false;
}
else if (beforeDrawValidator.getValidateStatus() == GL_FALSE && postDrawError == GL_INVALID_OPERATION)
{
// Validation does not depend on input values, no need to test all values
return true;
}
else
DE_ASSERT(false);
}
else
DE_ASSERT(false);
}
}
gl.useProgram(0);
if (separablePrograms)
gl.bindProgramPipeline(0);
GLU_EXPECT_NO_ERROR(gl.getError(), "ShaderCase::execute(): end");
return true;
}
TestCase::IterateResult ShaderLibraryCase::iterate (void)
{
// Initialize state to pass.
m_testCtx.setTestResult(QP_TEST_RESULT_PASS, "Pass");
bool executeOk = execute();
DE_ASSERT(executeOk ? m_testCtx.getTestResult() == QP_TEST_RESULT_PASS : m_testCtx.getTestResult() != QP_TEST_RESULT_PASS);
DE_UNREF(executeOk);
return TestCase::STOP;
}
} // gls
} // deqp