AMDGPU: Use constant and externally_initialized for block handle

The runtime initializes this.

diff --git a/llvm/lib/Target/AMDGPU/AMDGPUOpenCLEnqueuedBlockLowering.cpp b/llvm/lib/Target/AMDGPU/AMDGPUOpenCLEnqueuedBlockLowering.cpp
index 111a99f..6f29d29 100644 (file)
--- a/llvm/lib/Target/AMDGPU/AMDGPUOpenCLEnqueuedBlockLowering.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUOpenCLEnqueuedBlockLowering.cpp
@@ -124,11 +124,11 @@ bool AMDGPUOpenCLEnqueuedBlockLowering::runOnModule(Module &M) {
 
       auto *GV = new GlobalVariable(
           M, HandleTy,
-          /*isConstant=*/false, GlobalValue::ExternalLinkage,
+          /*isConstant=*/true, GlobalValue::ExternalLinkage,
           /*Initializer=*/Constant::getNullValue(HandleTy), RuntimeHandle,
           /*InsertBefore=*/nullptr, GlobalValue::NotThreadLocal,
           AMDGPUAS::GLOBAL_ADDRESS,
-          /*isExternallyInitialized=*/false);
+          /*isExternallyInitialized=*/true);
       LLVM_DEBUG(dbgs() << "runtime handle created: " << *GV << '\n');
 
       for (User *U : F.users())

diff --git a/llvm/test/CodeGen/AMDGPU/enqueue-kernel.ll b/llvm/test/CodeGen/AMDGPU/enqueue-kernel.ll
index 71ec864..d0fb84c 100644 (file)
--- a/llvm/test/CodeGen/AMDGPU/enqueue-kernel.ll
+++ b/llvm/test/CodeGen/AMDGPU/enqueue-kernel.ll
@@ -109,11 +109,11 @@ define internal amdgpu_kernel void @1(<{ i32, i32, ptr addrspace(1), i8 }> %arg)
 attributes #0 = { "enqueued-block" }
 ;.
 ; CHECK: @[[KERNEL_ADDRESS_USER:[a-zA-Z0-9_$"\\.-]+]] = global [1 x ptr] [ptr addrspacecast (ptr addrspace(1) @block_has_used_kernel_address.runtime_handle to ptr)]
-; CHECK: @[[__TEST_BLOCK_INVOKE_KERNEL_RUNTIME_HANDLE:[a-zA-Z0-9_$"\\.-]+]] = addrspace(1) global [[BLOCK_RUNTIME_HANDLE_T:%.*]] zeroinitializer
-; CHECK: @[[__TEST_BLOCK_INVOKE_2_KERNEL_RUNTIME_HANDLE:[a-zA-Z0-9_$"\\.-]+]] = addrspace(1) global [[BLOCK_RUNTIME_HANDLE_T:%.*]] zeroinitializer
-; CHECK: @[[BLOCK_HAS_USED_KERNEL_ADDRESS_RUNTIME_HANDLE:[a-zA-Z0-9_$"\\.-]+]] = addrspace(1) global [[BLOCK_RUNTIME_HANDLE_T:%.*]] zeroinitializer
-; CHECK: @[[__AMDGPU_ENQUEUED_KERNEL_RUNTIME_HANDLE:[a-zA-Z0-9_$"\\.-]+]] = addrspace(1) global [[BLOCK_RUNTIME_HANDLE_T:%.*]] zeroinitializer
-; CHECK: @[[__AMDGPU_ENQUEUED_KERNEL_1_RUNTIME_HANDLE:[a-zA-Z0-9_$"\\.-]+]] = addrspace(1) global [[BLOCK_RUNTIME_HANDLE_T:%.*]] zeroinitializer
+; CHECK: @[[__TEST_BLOCK_INVOKE_KERNEL_RUNTIME_HANDLE:[a-zA-Z0-9_$"\\.-]+]] = addrspace(1) externally_initialized constant [[BLOCK_RUNTIME_HANDLE_T:%.*]] zeroinitializer
+; CHECK: @[[__TEST_BLOCK_INVOKE_2_KERNEL_RUNTIME_HANDLE:[a-zA-Z0-9_$"\\.-]+]] = addrspace(1) externally_initialized constant [[BLOCK_RUNTIME_HANDLE_T:%.*]] zeroinitializer
+; CHECK: @[[BLOCK_HAS_USED_KERNEL_ADDRESS_RUNTIME_HANDLE:[a-zA-Z0-9_$"\\.-]+]] = addrspace(1) externally_initialized constant [[BLOCK_RUNTIME_HANDLE_T:%.*]] zeroinitializer
+; CHECK: @[[__AMDGPU_ENQUEUED_KERNEL_RUNTIME_HANDLE:[a-zA-Z0-9_$"\\.-]+]] = addrspace(1) externally_initialized constant [[BLOCK_RUNTIME_HANDLE_T:%.*]] zeroinitializer
+; CHECK: @[[__AMDGPU_ENQUEUED_KERNEL_1_RUNTIME_HANDLE:[a-zA-Z0-9_$"\\.-]+]] = addrspace(1) externally_initialized constant [[BLOCK_RUNTIME_HANDLE_T:%.*]] zeroinitializer
 ;.
 ; CHECK-LABEL: define {{[^@]+}}@non_caller
 ; CHECK-SAME: (ptr addrspace(1) [[A:%.*]], i8 [[B:%.*]], ptr addrspace(1) [[C:%.*]], i64 [[D:%.*]]) {