[all-commits] [llvm/llvm-project] fd8f69: [mlir][Bufferization] Fix to_buffer(tensor.cast) f...

[Quinn Dawkins via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: fd8f69d3eb9be3a987b4044fa93dd9ed0aafe094
      https://github.com/llvm/llvm-project/commit/fd8f69d3eb9be3a987b4044fa93dd9ed0aafe094
  Author: Quinn Dawkins <quinn.dawkins at gmail.com>
  Date:   2025-07-24 (Thu, 24 Jul 2025)

  Changed paths:
    M mlir/lib/Dialect/Bufferization/IR/BufferizationOps.cpp
    M mlir/test/Dialect/Bufferization/canonicalize.mlir

  Log Message:
  -----------
  [mlir][Bufferization] Fix to_buffer(tensor.cast) folder (#150511)

Previously this folder would ignore the layout and memory space on the
to_buffer op and set it as default. This changes the pattern to retain
both fields from the existing memref type but incorporate the static
shape information from the tensor cast.
The `read_only` attribute was also dropped by the pattern and is
retained now as well.



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications