[Mlir-commits] [mlir] [MLIR][XeGPU] Remove offsets from create_nd_tdesc & remove update_nd_offset, move offsets to load/store/prefetch ops (PR #193330)

[Jianhui Li]

================
@@ -689,30 +542,15 @@ LogicalResult StoreNdOp::verify() {
   return success();
 }
 
-//===----------------------------------------------------------------------===//
-// XeGPU_UpdateNDOffsetOp
-//===----------------------------------------------------------------------===//
-LogicalResult UpdateNdOffsetOp::verify() {
-  auto ty = getTensorDescType();
-
-  // number of offsets specified must match the rank of the tensor descriptor
-  if (ty.getRank() != (int64_t)getNumOffsets()) {
-    return emitOpError("Invalid number of offsets.");
-  }
-  return success();
-}
-
 //===----------------------------------------------------------------------===//
 // XeGPU_PrefetchOp
 //===----------------------------------------------------------------------===//
 LogicalResult PrefetchOp::verify() {
   auto tdescTy = getTensorDescType();
----------------
Jianhui-Li wrote:

why prefetch still work with getTensorDescType()? 

https://github.com/llvm/llvm-project/pull/193330