[llvm] [mlir] [MLIR][AMDGPU] Adding dynamic size check to avoid subword buffer load (PR #135014)

[Krzysztof Drewniak via llvm-commits]

================
@@ -141,6 +243,7 @@ struct TransferReadLowering final : OpRewritePattern<vector::TransferReadOp> {
 void mlir::amdgpu::populateAmdgpuTransferReadToLoadPatterns(
     RewritePatternSet &patterns) {
   patterns.add<TransferReadLowering>(patterns.getContext());
+  vector::populateVectorTransferLoweringPatterns(patterns);
----------------
krzysz00 wrote:

I think these sorts of limited-applicability patterns don't often populate their next step - I say keep the attribute

https://github.com/llvm/llvm-project/pull/135014