Add spirv::GlobalVariableOp that allows module level definition of variables

FuncOps in MLIR use explicit capture. So global variables defined in
module scope need to have a symbol name and this should be used to
refer to the variable within the function. This deviates from SPIR-V
spec, which assigns an SSA value to variables at all scopes that can
be used to refer to the variable, which requires SPIR-V functions to
allow implicit capture. To handle this add a new op,
spirv::GlobalVariableOp that can be used to define module scope
variables.
Since instructions need an SSA value, an new spirv::AddressOfOp is
added to convert a symbol reference to an SSA value for use with other
instructions.
This also means the spirv::EntryPointOp instruction needs to change to
allow initializers to be specified using symbol reference instead of
SSA value
The current spirv::VariableOp which returns an SSA value (as defined
by SPIR-V spec) can still be used to define function-scope variables.
PiperOrigin-RevId: 263951109

1 files changed, 2 insertions, 2 deletions

diff --git a/mlir/test/Conversion/GPUToSPIRV/simple.mlir b/mlir/test/Conversion/GPUToSPIRV/simple.mlir
index 671a38bd1fa..dc7964d0c95 100644
--- a/mlir/test/Conversion/GPUToSPIRV/simple.mlir
+++ b/mlir/test/Conversion/GPUToSPIRV/simple.mlir
@@ -1,8 +1,8 @@
 // RUN: mlir-opt -convert-gpu-to-spirv %s -o - | FileCheck %s
 
 // CHECK:       spv.module "Logical" "VulkanKHR" {
-// CHECK-NEXT:    [[VAR1:%.*]] = spv.Variable bind(0, 0) : !spv.ptr<f32, StorageBuffer>
-// CHECK-NEXT:    [[VAR2:%.*]] = spv.Variable bind(0, 1) : !spv.ptr<!spv.array<12 x f32>, StorageBuffer>
+// CHECK-NEXT:    spv.globalVariable !spv.ptr<f32, StorageBuffer> [[VAR1:@.*]] bind(0, 0)
+// CHECK-NEXT:    spv.globalVariable !spv.ptr<!spv.array<12 x f32>, StorageBuffer> [[VAR2:@.*]] bind(0, 1)
 // CHECK-NEXT:    func @kernel_1
 // CHECK-NEXT:      spv.Return
 // CHECK:       spv.EntryPoint "GLCompute" @kernel_1, [[VAR1]], [[VAR2]]