BeeEngine/glslang-zig
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Reserve unused std140 uniform block in reflection, and fix uniform block matrix layout (#2041)

Browse source 
According to the spec glsl4.60.7:
4.4.5. Uniform and Shader Storage Block Layout Qualifiers:
"The packed qualifier overrides only std140, std430, and shared; other qualifiers are inherited.
When packed is used, no shareable layout is guaranteed. The compiler and linker can optimize
memory use based on what variables actively get used and on other criteria. Offsets must be
queried, as there is no other way of guaranteeing where (and which) variables reside within the
block"

we should reserve std140 block and shared block in reflection.

According to the spec glsl4.60.7:
4.4.5. Uniform and Shader Storage Block Layout Qualifiers:
"The row_major and column_major qualifiers only affect the layout of matrices, including all
matrices contained in structures and arrays they are applied to, to all depths of nesting. These
qualifiers can be applied to other types, but will have no effect."

We need ensure all matrix block member been effect.

Support EShMsgKeepUncalled in reflection

EShMsgKeepUncalled  is a link message for link program.
We need only one option to control uncalled function optimization.
If we set EShMsgKeepUncalled as false in link time, linker won't be keep the uncall function sequence in AST,  and if we set EShMsgKeepUncalled as true in link time, linker will keep all uncalled function sequence in AST.
So, in reflecte time, we just only travers all function sequence. It make EShMsgKeepUncalled  only work at linker, and can effect reflection.

Recursively layout packing to "block member"

layout packing isn't set recursively, it causes TReflection::getOffsets doesn't work correctly.
This commit is contained in:
Roy.li 2020-05-15 16:26:48 +08:00 • committed by GitHub
parent c6874320de
commit 24dcbd1b1f
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
9 changed files with 315 additions and 101 deletions
Download patch file Download diff file Expand all files Collapse all files

106
glslang/MachineIndependent/reflection.cpp
View file

@ -77,10 +77,10 @@ namespace glslang {
// This is in the glslang namespace directly so it can be a friend of TReflection.
//
class TReflectionTraverser : public TLiveTraverser {
class TReflectionTraverser : public TIntermTraverser {
public:
TReflectionTraverser(const TIntermediate& i, TReflection& r) :
TLiveTraverser(i), reflection(r) { }
TIntermTraverser(), intermediate(i), reflection(r), updateStageMasks(true) { }
virtual bool visitBinary(TVisit, TIntermBinary* node);
virtual void visitSymbol(TIntermSymbol* base);
@ -92,11 +92,37 @@ public:
if (processedDerefs.find(&base) == processedDerefs.end()) {
processedDerefs.insert(&base);
uint32_t blockIndex = -1;
uint32_t offset = -1;
TList<TIntermBinary*> derefs;
TString baseName = base.getName();
if (base.getType().getBasicType() == EbtBlock) {
offset = 0;
bool anonymous = IsAnonymous(baseName);
const TString& blockName = base.getType().getTypeName();
if (!anonymous)
baseName = blockName;
else
baseName = "";
if (base.getType().isArray()) {
TType derefType(base.getType(), 0);
assert(!anonymous);
for (int e = 0; e < base.getType().getCumulativeArraySize(); ++e)
blockIndex = addBlockName(blockName + "[" + String(e) + "]", derefType,
intermediate.getBlockSize(base.getType()));
}
else
blockIndex = addBlockName(blockName, base.getType(), intermediate.getBlockSize(base.getType()));
}
// Use a degenerate (empty) set of dereferences to immediately put as at the end of
// the dereference change expected by blowUpActiveAggregate.
TList<TIntermBinary*> derefs;
blowUpActiveAggregate(base.getType(), base.getName(), derefs, derefs.end(), -1, -1, 0, 0,
base.getQualifier().storage, true);
blowUpActiveAggregate(base.getType(), baseName, derefs, derefs.end(), offset, blockIndex, 0, 0,
base.getQualifier().storage, updateStageMasks);
}
}
@ -155,9 +181,9 @@ public:
void getOffsets(const TType& type, TVector<int>& offsets)
{
const TTypeList& memberList = *type.getStruct();
int memberSize = 0;
int offset = 0;
for (size_t m = 0; m < offsets.size(); ++m) {
// if the user supplied an offset, snap to it now
if (memberList[m].type->getQualifier().hasOffset())
@ -334,7 +360,8 @@ public:
for (int i = 0; i < arrayIterateSize; ++i) {
TString newBaseName = name;
newBaseName.append(TString("[") + String(i) + "]");
if (terminalType->getBasicType() != EbtBlock)
newBaseName.append(TString("[") + String(i) + "]");
TType derefType(*terminalType, 0);
if (offset >= 0)
offset = baseOffset + stride * i;
@ -643,13 +670,17 @@ public:
blocks.back().numMembers = countAggregateMembers(type);
EShLanguageMask& stages = blocks.back().stages;
stages = static_cast<EShLanguageMask>(stages | 1 << intermediate.getStage());
if (updateStageMasks) {
EShLanguageMask& stages = blocks.back().stages;
stages = static_cast<EShLanguageMask>(stages | 1 << intermediate.getStage());
}
} else {
blockIndex = it->second;
EShLanguageMask& stages = blocks[blockIndex].stages;
stages = static_cast<EShLanguageMask>(stages | 1 << intermediate.getStage());
if (updateStageMasks) {
EShLanguageMask& stages = blocks[blockIndex].stages;
stages = static_cast<EShLanguageMask>(stages | 1 << intermediate.getStage());
}
}
return blockIndex;
@ -995,8 +1026,10 @@ public:
return type.isArray() ? type.getOuterArraySize() : 1;
}
const TIntermediate& intermediate;
TReflection& reflection;
std::set<const TIntermNode*> processedDerefs;
bool updateStageMasks;
protected:
TReflectionTraverser(TReflectionTraverser&);
@ -1029,8 +1062,15 @@ bool TReflectionTraverser::visitBinary(TVisit /* visit */, TIntermBinary* node)
// To reflect non-dereferenced objects.
void TReflectionTraverser::visitSymbol(TIntermSymbol* base)
{
if (base->getQualifier().storage == EvqUniform)
addUniform(*base);
if (base->getQualifier().storage == EvqUniform) {
if (base->getBasicType() == EbtBlock) {
if (reflection.options & EShReflectionSharedStd140Blocks) {
addUniform(*base);
}
} else {
addUniform(*base);
}
}
if ((intermediate.getStage() == reflection.firstStage && base->getQualifier().isPipeInput()) ||
(intermediate.getStage() == reflection.lastStage && base->getQualifier().isPipeOutput()))
@ -1135,15 +1175,39 @@ bool TReflection::addStage(EShLanguage stage, const TIntermediate& intermediate)
TReflectionTraverser it(intermediate, *this);
// put the entry point on the list of functions to process
it.pushFunction(intermediate.getEntryPointMangledName().c_str());
// process all the functions
while (! it.functions.empty()) {
TIntermNode* function = it.functions.back();
it.functions.pop_back();
function->traverse(&it);
for (auto& sequnence : intermediate.getTreeRoot()->getAsAggregate()->getSequence()) {
if (sequnence->getAsAggregate() != nullptr) {
if (sequnence->getAsAggregate()->getOp() == glslang::EOpLinkerObjects) {
it.updateStageMasks = false;
TIntermAggregate* linkerObjects = sequnence->getAsAggregate();
for (auto& sequnence : linkerObjects->getSequence()) {
auto pNode = sequnence->getAsSymbolNode();
if (pNode != nullptr && pNode->getQualifier().storage == EvqUniform &&
(options & EShReflectionSharedStd140Blocks)) {
if (pNode->getBasicType() == EbtBlock) {
// collect std140 and shared uniform block form AST
if (pNode->getQualifier().layoutPacking == ElpStd140 ||
pNode->getQualifier().layoutPacking == ElpShared) {
pNode->traverse(&it);
}
}
}
}
} else {
// This traverser will travers all function in AST.
// If we want reflect uncalled function, we need set linke message EShMsgKeepUncalled.
// When EShMsgKeepUncalled been set to true, all function will be keep in AST, even it is a uncalled function.
// This will keep some uniform variables in reflection, if those uniform variables is used in these uncalled function.
//
// If we just want reflect only live node, we can use a default link message or set EShMsgKeepUncalled false.
// When linke message not been set EShMsgKeepUncalled, linker won't keep uncalled function in AST.
// So, travers all function node can equivalent to travers live function.
it.updateStageMasks = true;
sequnence->getAsAggregate()->traverse(&it);
}
}
}
it.updateStageMasks = true;
buildCounterIndices(intermediate);
buildUniformStageMask(intermediate);