[Mlir-commits] [llvm] [mlir] [mlir][spirv] Support `gpu` in `convert-to-spirv` pass (PR #105010)

[Angel Zhang]

================
@@ -79,6 +80,7 @@ struct ConvertToSPIRVPass final
     arith::populateArithToSPIRVPatterns(typeConverter, patterns);
     populateBuiltinFuncToSPIRVPatterns(typeConverter, patterns);
     populateFuncToSPIRVPatterns(typeConverter, patterns);
+    populateGPUToSPIRVPatterns(typeConverter, patterns);
----------------
angelz913 wrote:

The `populateGPUToSPIRVPatterns` only includes lowering from GPU ops to SPIR-V

https://github.com/llvm/llvm-project/pull/105010