[OPENMP][NVPTX]Fixed emission of the critical region.

Critical regions in NVPTX are the constructs, which, generally speaking,
are not supported by the NVPTX target. Instead we're using special
technique to handle the critical regions. Currently they are supported
only within the loop and all the threads in the loop must execute the
same critical region.
Inside of this special regions the regions still must be emitted as
critical, to avoid possible data races between the teams +
synchronization must use __kmpc_barrier functions.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@348272 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/CodeGen/CGOpenMPRuntimeNVPTX.cpp b/lib/CodeGen/CGOpenMPRuntimeNVPTX.cpp
index cf814a4b203b4cde7d91a365a168e7da07667b98..5f7122e4e8f8c3a7c34e14ddc19bd89aaf0d9453 100644 (file)
--- a/lib/CodeGen/CGOpenMPRuntimeNVPTX.cpp
+++ b/lib/CodeGen/CGOpenMPRuntimeNVPTX.cpp
@@ -2743,14 +2743,16 @@ void CGOpenMPRuntimeNVPTX::emitCriticalRegion(
   CGF.EmitBlock(BodyBB);
 
   // Output the critical statement.
-  CriticalOpGen(CGF);
+  CGOpenMPRuntime::emitCriticalRegion(CGF, CriticalName, CriticalOpGen, Loc,
+                                      Hint);
 
   // After the body surrounded by the critical region, the single executing
   // thread will jump to the synchronisation point.
   // Block waits for all threads in current team to finish then increments the
   // counter variable and returns to the loop.
   CGF.EmitBlock(SyncBB);
-  getNVPTXCTABarrier(CGF);
+  emitBarrierCall(CGF, Loc, OMPD_unknown, /*EmitChecks=*/false,
+                  /*ForceSimpleCall=*/true);
 
   llvm::Value *IncCounterVal =
       CGF.Builder.CreateNSWAdd(CounterVal, CGF.Builder.getInt32(1));

diff --git a/test/OpenMP/nvptx_parallel_codegen.cpp b/test/OpenMP/nvptx_parallel_codegen.cpp
index 08431fccc009a702d8dfe5d1f99216086f51ed64..3dcf33017984816b7df9a00ee6bc4ab636eaec53 100644 (file)
--- a/test/OpenMP/nvptx_parallel_codegen.cpp
+++ b/test/OpenMP/nvptx_parallel_codegen.cpp
@@ -356,7 +356,13 @@ int bar(int n){
 // CHECK:  [[RES:%.+]] = icmp eq i32 [[TID]], [[CC_VAL]]
 // CHECK:  br i1 [[RES]], label
 
-// CHECK:  call void @llvm.nvvm.barrier0()
+// CHECK:  call void @__kmpc_critical(
+// CHECK:  load i32, i32*
+// CHECK:  add nsw i32
+// CHECK:  store i32
+// CHECK:  call void @__kmpc_end_critical(
+
+// CHECK:  call void @__kmpc_barrier(%struct.ident_t* @{{.+}}, i32 %{{.+}})
 // CHECK:  [[NEW_CC_VAL:%.+]] = add nsw i32 [[CC_VAL]], 1
 // CHECK:  store i32 [[NEW_CC_VAL]], i32* [[CC]],
 // CHECK:  br label