[Mlir-commits] [mlir] [MLIR][GPU-LLVM] Add in-pass signature update option for opencl kernels (PR #105664)

[Victor Perez]

================
@@ -325,16 +359,45 @@ struct GPUToLLVMSPVConversionPass final
     LLVMTypeConverter converter(context, options);
     LLVMConversionTarget target(*context);
 
+    if (forceOpenclAddressSpaces) {
+      MemorySpaceToOpenCLMemorySpaceConverter converter;
+      AttrTypeReplacer replacer;
+      replacer.addReplacement([&converter](BaseMemRefType origType)
+                                  -> std::optional<BaseMemRefType> {
+        return converter.convertType<BaseMemRefType>(origType);
+      });
+
+      replacer.recursivelyReplaceElementsIn(getOperation(),
+                                            /*replaceAttrs=*/true,
+                                            /*replaceLocs=*/false,
+                                            /*replaceTypes=*/true);
+    }
+
     target.addIllegalOp<gpu::BarrierOp, gpu::BlockDimOp, gpu::BlockIdOp,
                         gpu::GPUFuncOp, gpu::GlobalIdOp, gpu::GridDimOp,
                         gpu::ReturnOp, gpu::ShuffleOp, gpu::ThreadIdOp>();
 
     populateGpuToLLVMSPVConversionPatterns(converter, patterns);
+    populateFuncToLLVMConversionPatterns(converter, patterns);
+    populateFinalizeMemRefToLLVMConversionPatterns(converter, patterns);
----------------
victor-eds wrote:

Users could create a pass, use patterns exposed by this pass, use more patterns thanks to the convert to LLVM interface and populate memspace conversion with the entrypoint exposed by our pass. I think it's alright to drop this, keep pass simple and let users extend their passes at their will/use more than one pass.

https://github.com/llvm/llvm-project/pull/105664