[HLSL] Use ExtVector for firstbit intrinsics

[V-FEXrt]

Fixes #142430

firstbit intrinsics were using the wrong vector type which causes some conversions to fail. This PR switches them to ExtVector which resolves the issue

[llvmbot]

@llvm/pr-subscribers-clang

@llvm/pr-subscribers-hlsl

Author: Ashley Coleman (V-FEXrt)

Changes

Fixes #142430

firstbit intrinsics were using the wrong vector type which causes some conversions to fail. This PR switches them to ExtVector which resolves the issue

---

Full diff: https://github.com/llvm/llvm-project/pull/142679.diff

3 Files Affected:

• (modified) clang/lib/Sema/SemaHLSL.cpp (+1-2)
• (modified) clang/test/CodeGenHLSL/builtins/firstbithigh.hlsl (+8)
• (modified) clang/test/CodeGenHLSL/builtins/firstbitlow.hlsl (+8)

diff --git a/clang/lib/Sema/SemaHLSL.cpp b/clang/lib/Sema/SemaHLSL.cpp
index 43db85594de3d..b923d1204d8d6 100644
--- a/clang/lib/Sema/SemaHLSL.cpp
+++ b/clang/lib/Sema/SemaHLSL.cpp
@@ -2485,8 +2485,7 @@ bool SemaHLSL::CheckBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall) {
 
     if (auto *VecTy = EltTy->getAs<VectorType>()) {
       EltTy = VecTy->getElementType();
-      ResTy = SemaRef.Context.getVectorType(ResTy, VecTy->getNumElements(),
-                                            VecTy->getVectorKind());
+      ResTy = SemaRef.Context.getExtVectorType(ResTy, VecTy->getNumElements());
     }
 
     if (!EltTy->isIntegerType()) {
diff --git a/clang/test/CodeGenHLSL/builtins/firstbithigh.hlsl b/clang/test/CodeGenHLSL/builtins/firstbithigh.hlsl
index debf6b6d3e3f5..a71b1878f8b55 100644
--- a/clang/test/CodeGenHLSL/builtins/firstbithigh.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/firstbithigh.hlsl
@@ -151,3 +151,11 @@ uint3 test_firstbithigh_long3(int64_t3 p0) {
 uint4 test_firstbithigh_long4(int64_t4 p0) {
   return firstbithigh(p0);
 }
+
+// CHECK-LABEL: test_firstbithigh_upcast
+// CHECK: [[FBH:%.*]] = call <4 x i32> @llvm.[[TARGET]].firstbituhigh.v4i32(<4 x i32> %{{.*}})
+// CHECK: [[CONV:%.*]] = zext <4 x i32> [[FBH]] to <4 x i64>
+// CHECK: ret <4 x i64> [[CONV]]
+uint64_t4 test_firstbithigh_upcast(uint4 p0) {
+  return firstbithigh(p0);
+}
diff --git a/clang/test/CodeGenHLSL/builtins/firstbitlow.hlsl b/clang/test/CodeGenHLSL/builtins/firstbitlow.hlsl
index 5d490fabc5bc8..007db0c9c2ad5 100644
--- a/clang/test/CodeGenHLSL/builtins/firstbitlow.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/firstbitlow.hlsl
@@ -151,3 +151,11 @@ uint3 test_firstbitlow_long3(int64_t3 p0) {
 uint4 test_firstbitlow_long4(int64_t4 p0) {
   return firstbitlow(p0);
 }
+
+// CHECK-LABEL: test_firstbitlow_upcast
+// CHECK: [[FBL:%.*]] = call <4 x i32> @llvm.[[TARGET]].firstbitlow.v4i32(<4 x i32> %{{.*}})
+// CHECK: [[CONV:%.*]] = zext <4 x i32> [[FBL]] to <4 x i64>
+// CHECK: ret <4 x i64> [[CONV]]
+uint64_t4 test_firstbitlow_upcast(uint4 p0) {
+  return firstbitlow(p0);
+}

[spall]

I don't fully understand why this was a problem? Can you reshare the original crash you saw, not the minimized godbolt example?
This for example uses getVectorType

llvm-project/clang/lib/Sema/SemaHLSL.cpp

Line 2170 in 97686f2

static void SetElementTypeAsReturnType(Sema *S, CallExpr *TheCall,

Should we change it there as well?

[V-FEXrt]

Can you reshare the original crash you saw, not the minimized godbolt example?

https://godbolt.org/z/f4EEWrvac

Should we change it there as well?

Yeah I think so

[spall]

I don't fully understand this but lgtm I guess.