[clang] 31d8dbd - [CUDA/SPIR-V] Force passing aggregate type byval

[Shangwu Yao via cfe-commits]

Author: Shangwu Yao
Date: 2022-07-22T20:30:15Z
New Revision: 31d8dbd1e5b4ee0fd04bfeb3a64d8f9f33260905

URL: https://github.com/llvm/llvm-project/commit/31d8dbd1e5b4ee0fd04bfeb3a64d8f9f33260905
DIFF: https://github.com/llvm/llvm-project/commit/31d8dbd1e5b4ee0fd04bfeb3a64d8f9f33260905.diff

LOG: [CUDA/SPIR-V] Force passing aggregate type byval

This patch forces copying aggregate type in kernel arguments by value when
compiling CUDA targeting SPIR-V. The original behavior is not passing by value
when there is any of destructor, copy constructor and move constructor defined
by user. This patch makes the behavior of SPIR-V generated from CUDA follow
the CUDA spec
(https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#global-function-argument-processing),
and matches the NVPTX
implementation (
https://github.com/llvm/llvm-project/blob/41958f76d8a2c47484fa176cba1de565cfe84de7/clang/lib/CodeGen/TargetInfo.cpp#L7241).

Differential Revision: https://reviews.llvm.org/D130387

Added: 
    clang/test/CodeGenCUDASPIRV/copy-aggregate-byval.cu

Modified: 
    clang/lib/CodeGen/TargetInfo.cpp

Removed: 
    


################################################################################
diff  --git a/clang/lib/CodeGen/TargetInfo.cpp b/clang/lib/CodeGen/TargetInfo.cpp
index e8ee5533104ca..fc0952e68a667 100644
--- a/clang/lib/CodeGen/TargetInfo.cpp
+++ b/clang/lib/CodeGen/TargetInfo.cpp
@@ -10449,6 +10449,15 @@ ABIArgInfo SPIRVABIInfo::classifyKernelArgumentType(QualType Ty) const {
       LTy = llvm::PointerType::getWithSamePointeeType(PtrTy, GlobalAS);
       return ABIArgInfo::getDirect(LTy, 0, nullptr, false);
     }
+
+    // Force copying aggregate type in kernel arguments by value when
+    // compiling CUDA targeting SPIR-V. This is required for the object
+    // copied to be valid on the device.
+    // This behavior follows the CUDA spec
+    // https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#global-function-argument-processing,
+    // and matches the NVPTX implementation.
+    if (isAggregateTypeForABI(Ty))
+      return getNaturalAlignIndirect(Ty, /* byval */ true);
   }
   return classifyArgumentType(Ty);
 }

diff  --git a/clang/test/CodeGenCUDASPIRV/copy-aggregate-byval.cu b/clang/test/CodeGenCUDASPIRV/copy-aggregate-byval.cu
new file mode 100644
index 0000000000000..bceca4d4ee5d6
--- /dev/null
+++ b/clang/test/CodeGenCUDASPIRV/copy-aggregate-byval.cu
@@ -0,0 +1,25 @@
+// Tests CUDA kernel arguments get copied by value when targeting SPIR-V, even with
+// destructor, copy constructor or move constructor defined by user.
+
+// RUN: %clang -Xclang -no-opaque-pointers -emit-llvm --cuda-device-only --offload=spirv32 \
+// RUN:   -nocudalib -nocudainc %s -o %t.bc -c 2>&1
+// RUN: llvm-dis %t.bc -o %t.ll
+// RUN: FileCheck %s --input-file=%t.ll
+
+// RUN: %clang -Xclang -no-opaque-pointers -emit-llvm --cuda-device-only --offload=spirv64 \
+// RUN:   -nocudalib -nocudainc %s -o %t.bc -c 2>&1
+// RUN: llvm-dis %t.bc -o %t.ll
+// RUN: FileCheck %s --input-file=%t.ll
+
+class GpuData {
+ public:
+  __attribute__((host)) __attribute__((device)) GpuData(int* src) {}
+  __attribute__((host)) __attribute__((device)) ~GpuData() {}
+  __attribute__((host)) __attribute__((device)) GpuData(const GpuData& other) {}
+  __attribute__((host)) __attribute__((device)) GpuData(GpuData&& other) {}
+};
+
+// CHECK: define
+// CHECK-SAME: spir_kernel void @_Z6kernel7GpuData(%class.GpuData* noundef byval(%class.GpuData) align
+
+__attribute__((global)) void kernel(GpuData output) {}