[HLSL] Fix casting asserts

[farzonl]

There are two issues here. first ICK_Floating_Integral were always defaulting to CK_FloatingToIntegral for vectors regardless of direction of cast. Check was scalar only so added a vec float check to the conditional.

Second issue was float to int casts were resolving to ICK_Integral_Promotion when they need to be resolving to CK_FloatingToIntegral. This was fixed by changing the ordering of conversion checks.

This fixes #82826

[llvmbot]

@llvm/pr-subscribers-hlsl

@llvm/pr-subscribers-clang

Author: Farzon Lotfi (farzonl)

Changes

There are two issues here. first ICK_Floating_Integral were always defaulting to CK_FloatingToIntegral for vectors regardless of direction of cast. Check was scalar only so added a vec float check to the conditional.

Second issue was float to int casts were resolving to ICK_Integral_Promotion when they need to be resolving to CK_FloatingToIntegral. This was fixed by changing the ordering of conversion checks.

This fixes #82826

---

Full diff: https://github.com/llvm/llvm-project/pull/82827.diff

3 Files Affected:

• (modified) clang/lib/Sema/SemaExprCXX.cpp (+1-1)
• (modified) clang/lib/Sema/SemaOverload.cpp (+8-8)
• (modified) clang/test/SemaHLSL/VectorOverloadResolution.hlsl (+44)

diff --git a/clang/lib/Sema/SemaExprCXX.cpp b/clang/lib/Sema/SemaExprCXX.cpp
index 322bd1c87b1d71..99f8abf77c0cb6 100644
--- a/clang/lib/Sema/SemaExprCXX.cpp
+++ b/clang/lib/Sema/SemaExprCXX.cpp
@@ -4843,7 +4843,7 @@ Sema::PerformImplicitConversion(Expr *From, QualType ToType,
                  .get();
       break;
     case ICK_Floating_Integral:
-      if (ToType->isRealFloatingType())
+      if (ToType->isRealFloatingType() || ToType->hasFloatingRepresentation())
         From =
             ImpCastExprToType(From, ToType, CK_IntegralToFloating, VK_PRValue,
                               /*BasePath=*/nullptr, CCK)
diff --git a/clang/lib/Sema/SemaOverload.cpp b/clang/lib/Sema/SemaOverload.cpp
index f7645422348b65..324a7b20113fcd 100644
--- a/clang/lib/Sema/SemaOverload.cpp
+++ b/clang/lib/Sema/SemaOverload.cpp
@@ -1883,9 +1883,16 @@ static bool IsVectorElementConversion(Sema &S, QualType FromType,
     ICK = ICK_Boolean_Conversion;
     return true;
   }
+  
+  if ((FromType->isRealFloatingType() && ToType->isIntegralType(S.Context)) ||
+      (FromType->isIntegralOrUnscopedEnumerationType() &&
+       ToType->isRealFloatingType())) {
+    ICK = ICK_Floating_Integral;
+    return true;
+  }
 
   if (S.IsIntegralPromotion(From, FromType, ToType)) {
-    ICK = ICK_Integral_Promotion;
+    ICK =  ICK_Integral_Promotion;
     return true;
   }
 
@@ -1895,13 +1902,6 @@ static bool IsVectorElementConversion(Sema &S, QualType FromType,
     return true;
   }
 
-  if ((FromType->isRealFloatingType() && ToType->isIntegralType(S.Context)) ||
-      (FromType->isIntegralOrUnscopedEnumerationType() &&
-       ToType->isRealFloatingType())) {
-    ICK = ICK_Floating_Integral;
-    return true;
-  }
-
   return false;
 }
 
diff --git a/clang/test/SemaHLSL/VectorOverloadResolution.hlsl b/clang/test/SemaHLSL/VectorOverloadResolution.hlsl
index e07391f803f899..d8794ef675768b 100644
--- a/clang/test/SemaHLSL/VectorOverloadResolution.hlsl
+++ b/clang/test/SemaHLSL/VectorOverloadResolution.hlsl
@@ -1,4 +1,8 @@
 // RUN: %clang_cc1 -triple dxil-unknown-shadermodel6.6-library -S -fnative-half-type -finclude-default-header -o - -ast-dump %s | FileCheck %s
+// RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -x hlsl -triple \
+// RUN:   dxil-pc-shadermodel6.3-library %s -fnative-half-type \
+// RUN:   -emit-llvm -disable-llvm-passes -O3 -o - | FileCheck %s \ 
+// RUN:   --check-prefixes=CHECKIR
 void Fn(double2 D);
 void Fn(half2 H);
 
@@ -28,3 +32,43 @@ void Fn2(int16_t2 S);
 void Call2(int2 I) {
   Fn2(I);
 }
+
+void Fn3( int64_t2 p0);
+
+// CHECK: FunctionDecl {{.*}} Call3 'void (half2)'
+// CHECK: CallExpr {{.*}} 'void'
+// CHECK-NEXT: ImplicitCastExpr {{.*}} 'void (*)(int64_t2)' <FunctionToPointerDecay>
+// CHECK-NEXT: DeclRefExpr {{.*}} 'void (int64_t2)' lvalue Function {{.*}} 'Fn3' 'void (int64_t2)'
+// CHECK-NEXT: ImplicitCastExpr {{.*}} 'int64_t2':'long __attribute__((ext_vector_type(2)))' <FloatingToIntegral>
+// CHECK-NEXT: ImplicitCastExpr {{.*}} 'half2':'half __attribute__((ext_vector_type(2)))' <LValueToRValue>
+// CHECK-NEXT: DeclRefExpr {{.*}} 'half2':'half __attribute__((ext_vector_type(2)))' lvalue ParmVar {{.*}} 'p0' 'half2':'half __attribute__((ext_vector_type(2)))'
+// CHECKIR: %conv = fptosi <2 x half> %0 to <2 x i64>
+void Call3(half2 p0) {
+  Fn3(p0);
+}
+
+// CHECK: FunctionDecl {{.*}} Call4 'void (float2)'
+// CHECK: CallExpr {{.*}} 'void'
+// CHECK-NEXT: ImplicitCastExpr {{.*}} 'void (*)(int64_t2)' <FunctionToPointerDecay>
+// CHECK-NEXT: DeclRefExpr {{.*}} 'void (int64_t2)' lvalue Function {{.*}} 'Fn3' 'void (int64_t2)'
+// CHECK-NEXT: ImplicitCastExpr {{.*}} 'int64_t2':'long __attribute__((ext_vector_type(2)))' <FloatingToIntegral>
+// CHECK-NEXT: ImplicitCastExpr {{.*}} 'float2':'float __attribute__((ext_vector_type(2)))' <LValueToRValue>
+// CHECK-NEXT: DeclRefExpr {{.*}} 'float2':'float __attribute__((ext_vector_type(2)))' lvalue ParmVar {{.*}} 'p0' 'float2':'float __attribute__((ext_vector_type(2)))'
+// CHECKIR: %conv = fptosi <2 x float> %0 to <2 x i64>
+void Call4(float2 p0) {
+  Fn3(p0);
+}
+
+void Fn4( float2 p0);
+
+// CHECK: FunctionDecl {{.*}} Call5 'void (int64_t2)'
+// CHECK: CallExpr {{.*}} 'void'
+// CHECK-NEXT: ImplicitCastExpr {{.*}} 'void (*)(float2)' <FunctionToPointerDecay>
+// CHECK-NEXT: DeclRefExpr {{.*}} 'void (float2)' lvalue Function {{.*}} 'Fn4' 'void (float2)'
+// CHECK-NEXT: ImplicitCastExpr {{.*}} 'float2':'float __attribute__((ext_vector_type(2)))' <IntegralToFloating>
+// CHECK-NEXT: ImplicitCastExpr {{.*}} 'int64_t2':'long __attribute__((ext_vector_type(2)))' <LValueToRValue>
+// CHECK-NEXT: DeclRefExpr {{.*}} 'int64_t2':'long __attribute__((ext_vector_type(2)))' lvalue ParmVar {{.*}} 'p0' 'int64_t2':'long __attribute__((ext_vector_type(2)))'
+// CHECKIR: %conv = sitofp <2 x i64> %0 to <2 x float>
+void Call5(int64_t2 p0) {
+  Fn4(p0);
+}

[github-actions]

✅ With the latest revision this PR passed the C/C++ code formatter.