[clang] [Clang][OpenCL][AMDGPU] Allow a kernel to call another kernel (PR #115821)

[Aniket Lal via cfe-commits]

================
@@ -196,13 +203,21 @@ class GlobalDecl {
   }
 
   GlobalDecl getWithKernelReferenceKind(KernelReferenceKind Kind) {
-    assert(isa<FunctionDecl>(getDecl()) &&
-           cast<FunctionDecl>(getDecl())->hasAttr<CUDAGlobalAttr>() &&
-           "Decl is not a GPU kernel!");
+    assert((isa<FunctionDecl>(getDecl()) &&
+            cast<FunctionDecl>(getDecl())->hasAttr<CUDAGlobalAttr>() &&
+            "Decl is not a GPU kernel!") ||
+           (isDeclOpenCLKernel() && "Decl is not a OpenCL kernel!"));
     GlobalDecl Result(*this);
     Result.Value.setInt(unsigned(Kind));
     return Result;
   }
+
+  bool isDeclOpenCLKernel() const {
+    auto FD = dyn_cast<FunctionDecl>(getDecl());
+    if (FD)
+      return FD->hasAttr<OpenCLKernelAttr>();
+    return FD;
+  }
----------------
lalaniket8 wrote:

Commited suggestion in 013801b1fc82ec6806b876d93da24d4f0f2ed098

> 



https://github.com/llvm/llvm-project/pull/115821
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits