[Mlir-commits] [clang] [llvm] [mlir] [OpenMP] Migrate GPU Reductions CodeGen from Clang to OMPIRBuilder (PR #80343)

[Jan Leyonberg]

================
@@ -607,15 +621,17 @@ class OpenMPIRBuilder {
   /// Generator for '#omp barrier'
   ///
   /// \param Loc The location where the barrier directive was encountered.
-  /// \param DK The kind of directive that caused the barrier.
+  /// \param Kind The kind of directive that caused the barrier.
   /// \param ForceSimpleCall Flag to force a simple (=non-cancellation) barrier.
   /// \param CheckCancelFlag Flag to indicate a cancel barrier return value
   ///                        should be checked and acted upon.
+  /// \param ThreadID Optional parameter to pass in any existing ThreadID value.
   ///
   /// \returns The insertion point after the barrier.
-  InsertPointTy createBarrier(const LocationDescription &Loc, omp::Directive DK,
-                              bool ForceSimpleCall = false,
-                              bool CheckCancelFlag = true);
+  InsertPointTy createBarrier(const LocationDescription &Loc,
+                              omp::Directive Kind, bool ForceSimpleCall = false,
+                              bool CheckCancelFlag = true,
+                              Value *ThreadID = nullptr);
----------------
jsjodin wrote:

The ThreadID doesn't seem to be passed in the original code. If there was reuse before perhaps some optimization is not happening? Probably worth looking into the cause of this instead of adding the parameter.

https://github.com/llvm/llvm-project/pull/80343