All compiler fixes to get ir branch work with falcor feature demo.

- support overloaded generic function. this involves adding a new expression type, `OverloadedExpr2` to hold the candidate expressions for the generic function decl being referenced.
- make BitNot a normal IROp instead of an IRPseudoOp
- make sure we clone the decorations of parameters when cloning ir functions
- propagate geometry shader entry point attributes (`[maxvertexcount]` and `[instance]`) through HLSL emit
- IR emit: handle geometry shader entry-point parameter decorations, such as 'triangle'.
- IR emit: treat geometry shader stream output typed ir value as `should fold into use`.

12 files changed, 186 insertions, 16 deletions

diff --git a/source/slang/ast-legalize.cpp b/source/slang/ast-legalize.cpp
index 996a66eac..5df6b80f5 100644
--- a/source/slang/ast-legalize.cpp
+++ b/source/slang/ast-legalize.cpp
@@ -5091,6 +5091,14 @@ struct FindIRDeclUsedByASTVisitor
         // TODO: need to walk the lookup result too
     }
 
+    void visitOverloadedExpr2(OverloadedExpr2* expr)
+    {
+        walkExpr(expr->base);
+        for (auto & candidate : expr->candidiateExprs)
+            walkExpr(candidate);
+    }
+
+
     void visitConstantExpr(ConstantExpr*)
     {}
 
diff --git a/source/slang/check.cpp b/source/slang/check.cpp
index 3fe4f8c5d..dfc09c485 100644
--- a/source/slang/check.cpp
+++ b/source/slang/check.cpp
@@ -1466,7 +1466,9 @@ namespace Slang
                 // For now we will do this in a completely ad hoc fashion,
                 // but it would be nice to have some generic routine to
                 // do the needed type checking/coercion.
-                if(getText(hlslUncheckedAttribute->getName()) == "numthreads")
+                auto attribText = getText(hlslUncheckedAttribute->getName());
+                
+                if(attribText == "numthreads")
                 {
                     if(hlslUncheckedAttribute->args.Count() != 3)
                         return m;
@@ -1490,8 +1492,33 @@ namespace Slang
 
                     return hlslNumThreadsAttribute;
                 }
-            }
+                else if (attribText == "maxvertexcount")
+                {
+                    if (hlslUncheckedAttribute->args.Count() != 1)
+                        return m;
+                    auto val = checkConstantIntVal(hlslUncheckedAttribute->args[0]);
+                    auto hlslMaxVertexCountAttrib = new HLSLMaxVertexCountAttribute();
+
+                    hlslMaxVertexCountAttrib->loc = hlslUncheckedAttribute->loc;
+                    hlslMaxVertexCountAttrib->name = hlslUncheckedAttribute->getName();
+                    hlslMaxVertexCountAttrib->args = hlslUncheckedAttribute->args;
+                    hlslMaxVertexCountAttrib->value = (int32_t)val->value;
+                    return hlslMaxVertexCountAttrib;
+                }
+                else if (attribText == "instance")
+                {
+                    if (hlslUncheckedAttribute->args.Count() != 1)
+                        return m;
+                    auto val = checkConstantIntVal(hlslUncheckedAttribute->args[0]);
+                    auto attrib = new HLSLInstanceAttribute();
 
+                    attrib->loc = hlslUncheckedAttribute->loc;
+                    attrib->name = hlslUncheckedAttribute->getName();
+                    attrib->args = hlslUncheckedAttribute->args;
+                    attrib->value = (int32_t)val->value;
+                    return attrib;
+                }
+            }
             // Default behavior is to leave things as they are,
             // and assume that modifiers are mostly already checked.
             //
@@ -2202,9 +2229,10 @@ namespace Slang
                 // to avoid recursion here.
                 if (functionNode->Body)
                 {
+                    auto oldFunc = function;
                     this->function = functionNode;
                     checkStmt(functionNode->Body);
-                    this->function = nullptr;
+                    this->function = oldFunc;
                 }
             }
         }
@@ -2630,6 +2658,7 @@ namespace Slang
         {
             if (functionNode->IsChecked(DeclCheckState::CheckedHeader)) return;
             functionNode->SetCheckState(DeclCheckState::CheckingHeader);
+            auto oldFunc = this->function;
             this->function = functionNode;
             auto returnType = CheckProperType(functionNode->ReturnType);
             functionNode->ReturnType = returnType;
@@ -2648,7 +2677,7 @@ namespace Slang
                 else
                     paraNames.Add(para->getName());
             }
-            this->function = NULL;
+            this->function = oldFunc;
             functionNode->SetCheckState(DeclCheckState::CheckedHeader);
 
             // One last bit of validation: check if we are redeclaring an existing function
@@ -5745,6 +5774,13 @@ namespace Slang
                     AddDeclRefOverloadCandidates(item, context);
                 }
             }
+            else if (auto overloadedExpr2 = funcExpr.As<OverloadedExpr2>())
+            {
+                for (auto item : overloadedExpr2->candidiateExprs)
+                {
+                    AddOverloadCandidates(item, context);
+                }
+            }
             else if (auto typeType = funcExprType->As<TypeType>())
             {
                 // If none of the above cases matched, but we are
@@ -5944,6 +5980,10 @@ namespace Slang
             {
                 context.baseExpr = funcOverloadExpr->base;
             }
+            else if (auto funcOverloadExpr2 = funcExpr.As<OverloadedExpr2>())
+            {
+                context.baseExpr = funcOverloadExpr2->base;
+            }
             AddOverloadCandidates(funcExpr, context);
 
             if (context.bestCandidates.Count() > 0)
@@ -6172,15 +6212,14 @@ namespace Slang
                     // There were multiple viable candidates, but that isn't an error: we just need
                     // to complete all of them and create an overloaded expression as a result.
 
-                    LookupResult result;
+                    auto overloadedExpr = new OverloadedExpr2();
+                    overloadedExpr->base = context.baseExpr;
                     for (auto candidate : context.bestCandidates)
                     {
                         auto candidateExpr = CompleteOverloadCandidate(context, candidate);
+                        overloadedExpr->candidiateExprs.Add(candidateExpr);
                     }
-
-                    throw "what now?";
-//                        auto overloadedExpr = new OverloadedExpr();
-//                        return overloadedExpr;
+                    return overloadedExpr;
                 }
             }
             else if (context.bestCandidate)
@@ -6400,6 +6439,13 @@ namespace Slang
             return expr;
         }
 
+        RefPtr<Expr> visitOverloadedExpr2(OverloadedExpr2* expr)
+        {
+            SLANG_DIAGNOSE_UNEXPECTED(getSink(), expr, "should not appear in input syntax");
+            return expr;
+        }
+
+
         RefPtr<Expr> visitAggTypeCtorExpr(AggTypeCtorExpr* expr)
         {
             SLANG_DIAGNOSE_UNEXPECTED(getSink(), expr, "should not appear in input syntax");
diff --git a/source/slang/emit.cpp b/source/slang/emit.cpp
index 33328fbb1..18a10460f 100644
--- a/source/slang/emit.cpp
+++ b/source/slang/emit.cpp
@@ -2315,7 +2315,7 @@ struct EmitVisitor
                     CASE(kIRPseudoOp_Pos, +);
                     CASE(kIROp_Neg, -);
                     CASE(kIROp_Not, !);
-                    CASE(kIRPseudoOp_BitNot, ~);
+                    CASE(kIROp_BitNot, ~);
 #undef CASE
 
 #define CASE(NAME, OP) case kIRPseudoOp_##NAME: EmitUnaryAssignExpr(outerPrec, kEOp_Prefix, #OP, "", callExpr); return
@@ -2554,6 +2554,11 @@ struct EmitVisitor
         emitName(expr->lookupResult2.getName());
     }
 
+    void visitOverloadedExpr2(OverloadedExpr2* expr, ExprEmitArg const& arg)
+    {
+        ExprVisitorWithArg<Slang::EmitVisitor, Slang::ExprEmitArg>::dispatch(expr->candidiateExprs[0].Ptr(), arg);
+    }
+
     void setSampleRateFlag()
     {
         context->shared->entryPointLayout->flags |= EntryPointLayout::Flag::usesAnySampleRateInput;
@@ -4928,6 +4933,10 @@ emitDeclImpl(decl, nullptr);
             // types.
             return true;
         }
+        else if (type->As<HLSLStreamOutputType>())
+        {
+            return true;
+        }
         else if(type->As<TextureTypeBase>())
         {
             // GLSL doesn't allow texture/resource types to
@@ -5512,7 +5521,12 @@ emitDeclImpl(decl, nullptr);
                 emitIROperand(ctx, inst->getArg(0));
             }
             break;
-
+        case kIROp_BitNot:
+            {
+                emit("~");
+                emitIROperand(ctx, inst->getArg(0));
+            }
+            break;
         case kIROp_Sample:
             emitIROperand(ctx, inst->getArg(0));
             emit(".Sample(");
@@ -6266,7 +6280,22 @@ emitDeclImpl(decl, nullptr);
                     emit(")]\n");
                 }
                 break;
-
+            case Stage::Geometry:
+            {
+                if (auto attrib = entryPointLayout->entryPoint->FindModifier<HLSLMaxVertexCountAttribute>())
+                {
+                    emit("[maxvertexcount(");
+                    Emit(attrib->value);
+                    emit(")]\n");
+                }
+                if (auto attrib = entryPointLayout->entryPoint->FindModifier<HLSLInstanceAttribute>())
+                {
+                    emit("[instance(");
+                    Emit(attrib->value);
+                    emit(")]\n");
+                }
+            }
+            break;
             // TODO: There are other stages that will need this kind of handling.
             default:
                 break;
@@ -6279,6 +6308,7 @@ emitDeclImpl(decl, nullptr);
 
         emit("(");
         auto firstParam = func->getFirstParam();
+        int pIdx = 0;
         for( auto pp = firstParam; pp; pp = pp->getNextParam() )
         {
             if(pp != firstParam)
@@ -6286,9 +6316,27 @@ emitDeclImpl(decl, nullptr);
 
             auto paramName = getIRName(pp);
             auto paramType = pp->getType();
+            if (auto decor = pp->findDecoration<IRHighLevelDeclDecoration>())
+            {
+                if (decor->decl)
+                {
+                    auto primType = decor->decl->FindModifier<HLSLGeometryShaderInputPrimitiveTypeModifier>();
+                    if (dynamic_cast<HLSLTriangleModifier*>(primType))
+                        emit("triangle ");
+                    else if (dynamic_cast<HLSLPointModifier*>(primType))
+                        emit("point ");
+                    else if (dynamic_cast<HLSLLineModifier*>(primType))
+                        emit("line ");
+                    else if (dynamic_cast<HLSLLineAdjModifier*>(primType))
+                        emit("lineadj ");
+                    else if (dynamic_cast<HLSLTriangleAdjModifier*>(primType))
+                        emit("triangleadj ");
+                }
+            }
             emitIRParamType(ctx, paramType, paramName);
 
             emitIRSemantics(ctx, pp);
+            pIdx++;
         }
         emit(")");
 
diff --git a/source/slang/expr-defs.h b/source/slang/expr-defs.h
index e3c16a674..4b47cc883 100644
--- a/source/slang/expr-defs.h
+++ b/source/slang/expr-defs.h
@@ -30,6 +30,18 @@ SYNTAX_CLASS(OverloadedExpr, Expr)
     FIELD(LookupResult, lookupResult2)
 END_SYNTAX_CLASS()
 
+// An expression that references an overloaded set of declarations
+// having the same name.
+SYNTAX_CLASS(OverloadedExpr2, Expr)
+
+    // Optional: the base expression is this overloaded result
+    // arose from a member-reference expression.
+    SYNTAX_FIELD(RefPtr<Expr>, base)
+
+    // The lookup result that was ambiguous
+    FIELD(List<RefPtr<Expr>>, candidiateExprs)
+END_SYNTAX_CLASS()
+
 SYNTAX_CLASS(ConstantExpr, Expr)
     FIELD(Token, token)
 
diff --git a/source/slang/ir-inst-defs.h b/source/slang/ir-inst-defs.h
index cc5474842..fa02da683 100644
--- a/source/slang/ir-inst-defs.h
+++ b/source/slang/ir-inst-defs.h
@@ -231,6 +231,7 @@ INTRINSIC(Pos)
 
 INST(Neg, neg, 1, 0)
 INST(Not, not, 1, 0)
+INST(BitNot, bitnot, 1, 0)
 
 #if 0
 INTRINSIC(PreInc)
@@ -286,7 +287,6 @@ PSEUDO_INST(XorAssign )
 PSEUDO_INST(LshAssign)
 PSEUDO_INST(RshAssign)
 PSEUDO_INST(Assign)
-PSEUDO_INST(BitNot)
 PSEUDO_INST(And)
 PSEUDO_INST(Or)
 
diff --git a/source/slang/ir.cpp b/source/slang/ir.cpp
index f65af2ba4..994ac82ff 100644
--- a/source/slang/ir.cpp
+++ b/source/slang/ir.cpp
@@ -3591,6 +3591,7 @@ namespace Slang
                 IRParam* clonedParam = builder->emitParam(
                     context->maybeCloneType(
                         originalParam->getType()));
+                cloneDecorations(context, clonedParam, originalParam);
                 registerClonedValue(context, clonedParam, originalParam);
             }
         }
diff --git a/source/slang/lower-to-ir.cpp b/source/slang/lower-to-ir.cpp
index c8e010f7d..5e7e05a23 100644
--- a/source/slang/lower-to-ir.cpp
+++ b/source/slang/lower-to-ir.cpp
@@ -705,7 +705,6 @@ LoweredValInfo emitCallToDeclRef(
             CASE(kIRPseudoOp_PostInc, kIROp_Add);
             CASE(kIRPseudoOp_PostDec, kIROp_Sub);
 #undef CASE
-
             default:
                 SLANG_UNIMPLEMENTED_X("IR pseudo-op");
                 break;
@@ -1197,6 +1196,11 @@ struct ExprLoweringVisitorBase : ExprVisitor<Derived, LoweredValInfo>
         SLANG_UNEXPECTED("overloaded expressions should not occur in checked AST");
     }
 
+    LoweredValInfo visitOverloadedExpr2(OverloadedExpr2* /*expr*/)
+    {
+        SLANG_UNEXPECTED("overloaded expressions should not occur in checked AST");
+    }
+
     LoweredValInfo visitIndexExpr(IndexExpr* expr)
     {
         auto type = lowerType(context, expr->type);
@@ -3521,7 +3525,6 @@ struct DeclLoweringVisitor : DeclVisitor<DeclLoweringVisitor, LoweredValInfo>
         for( auto paramInfo : parameterLists.params )
         {
             RefPtr<Type> irParamType = lowerSimpleType(context, paramInfo.type);
-
             switch( paramInfo.direction )
             {
             case kParameterDirection_In:
diff --git a/source/slang/mangle.cpp b/source/slang/mangle.cpp
index 1a412d041..29446d942 100644
--- a/source/slang/mangle.cpp
+++ b/source/slang/mangle.cpp
@@ -124,6 +124,12 @@ namespace Slang
         {
             emitQualifiedName(context, declRefType->declRef);
         }
+        else if (auto arrType = dynamic_cast<ArrayExpressionType*>(type))
+        {
+            emitRaw(context, "a");
+            emitSimpleIntVal(context, arrType->ArrayLength);
+            emitType(context, arrType->baseType);
+        }
         else
         {
             SLANG_UNEXPECTED("unimplemented case in mangling");
diff --git a/source/slang/modifier-defs.h b/source/slang/modifier-defs.h
index cd8f2524b..124dd2882 100644
--- a/source/slang/modifier-defs.h
+++ b/source/slang/modifier-defs.h
@@ -305,6 +305,16 @@ SYNTAX_CLASS(HLSLNumThreadsAttribute, HLSLAttribute)
     FIELD(int32_t, z)
 END_SYNTAX_CLASS()
 
+SYNTAX_CLASS(HLSLMaxVertexCountAttribute, HLSLAttribute)
+    // The number of max vertex count for geometry shader
+    FIELD(int32_t, value)
+END_SYNTAX_CLASS()
+
+SYNTAX_CLASS(HLSLInstanceAttribute, HLSLAttribute)
+    // The number of instances to run for geometry shader
+    FIELD(int32_t, value)
+END_SYNTAX_CLASS()
+
 // HLSL modifiers for geometry shader input topology
 SIMPLE_SYNTAX_CLASS(HLSLGeometryShaderInputPrimitiveTypeModifier, Modifier)
 SIMPLE_SYNTAX_CLASS(HLSLPointModifier        , HLSLGeometryShaderInputPrimitiveTypeModifier)
diff --git a/source/slang/slang-stdlib.cpp b/source/slang/slang-stdlib.cpp
index 3e526e96c..bee823d69 100644
--- a/source/slang/slang-stdlib.cpp
+++ b/source/slang/slang-stdlib.cpp
@@ -199,7 +199,7 @@ namespace Slang
         { kIRPseudoOp_Pos,     "+",    ARITHMETIC_MASK },
         { kIROp_Neg,     "-",    ARITHMETIC_MASK },
         { kIROp_Not,     "!",    ANY_MASK        },
-        { kIRPseudoOp_BitNot,    "~",    INT_MASK        },
+        { kIROp_BitNot,    "~",    INT_MASK        },
         { kIRPseudoOp_PreInc,  "++",   ARITHMETIC_MASK | ASSIGNMENT },
         { kIRPseudoOp_PreDec,  "--",   ARITHMETIC_MASK | ASSIGNMENT },
         { kIRPseudoOp_PostInc, "++",   ARITHMETIC_MASK | ASSIGNMENT | POSTFIX },
diff --git a/tests/compute/generics-overload.expected.txt b/tests/compute/generics-overload.expected.txt
new file mode 100644
index 000000000..98798bd61
--- /dev/null
+++ b/tests/compute/generics-overload.expected.txt
@@ -0,0 +1,4 @@
+0
+3F800000
+40000000
+40400000
\ No newline at end of file
diff --git a/tests/compute/generics-overload.slang b/tests/compute/generics-overload.slang
new file mode 100644
index 000000000..e530acb59
--- /dev/null
+++ b/tests/compute/generics-overload.slang
@@ -0,0 +1,32 @@
+//TEST(smoke,compute):COMPARE_COMPUTE:-xslang -use-ir
+//TEST_INPUT:ubuffer(data=[0 0 0 0], stride=4):dxbinding(0),glbinding(0),out
+// Confirm that generics syntax can be used in user
+// code and generates valid output.
+
+RWStructuredBuffer<float> outputBuffer;
+
+
+__generic<T>
+T test(T val)
+{
+	return val;
+}
+
+__generic<T>
+T test(T val, int a)
+{
+	return val;
+}
+
+
+[numthreads(4, 1, 1)]
+void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)
+{
+	uint tid = dispatchThreadID.x;
+
+	float inVal = float(tid);
+
+	float outVal = test<float>(inVal, 0);
+
+	outputBuffer[tid] = outVal;
+}
\ No newline at end of file