[Mlir-commits] [mlir] [mlir][vector] Fix invalid `LoadOp` indices being created (PR #75519)

[Rik Huijzer]

rikhuijzer wrote:

I'm working on reverting this PR. It crashes on the clang-aarch64-sve-vls builder:
```
******************** TEST 'MLIR :: Integration/Dialect/Vector/CPU/ArmSVE/arrays-of-scalable-vectors.mlir' FAILED ********************
Exit Code: 2
Command Output (stdout):
--
# RUN: at line 1
/home/tcwg-buildbot/worker/clang-aarch64-sve-vls/stage1/bin/mlir-opt /home/tcwg-buildbot/worker/clang-aarch64-sve-vls/llvm/mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/arrays-of-scalable-vectors.mlir -convert-vector-to-scf -arm-sve-legalize-vector-storage -convert-vector-to-llvm="enable-arm-sve" -test-lower-to-llvm |  mlir-cpu-runner -e=entry -entry-point-result=void --march=aarch64 --mattr="+sve" -shared-libs=/home/tcwg-buildbot/worker/clang-aarch64-sve-vls/stage1/./lib/libmlir_c_runner_utils.so |  /home/tcwg-buildbot/worker/clang-aarch64-sve-vls/stage1/bin/FileCheck /home/tcwg-buildbot/worker/clang-aarch64-sve-vls/llvm/mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/arrays-of-scalable-vectors.mlir
# executed command: /home/tcwg-buildbot/worker/clang-aarch64-sve-vls/stage1/bin/mlir-opt /home/tcwg-buildbot/worker/clang-aarch64-sve-vls/llvm/mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/arrays-of-scalable-vectors.mlir -convert-vector-to-scf -arm-sve-legalize-vector-storage -convert-vector-to-llvm=enable-arm-sve -test-lower-to-llvm
# .---command stderr------------
# | /home/tcwg-buildbot/worker/clang-aarch64-sve-vls/llvm/mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/arrays-of-scalable-vectors.mlir:66:15: error: 'memref.load' op incorrect number of indices for load, expected 2 but got 1
# |   %vector_a = vector.transfer_read %a[%c0, %c0, %c0], %cst, %mask_a {in_bounds = [true, true, true]} : memref<1x2x?xf32>, vector<1x2x[4]xf32>
# |               ^
# | /home/tcwg-buildbot/worker/clang-aarch64-sve-vls/llvm/mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/arrays-of-scalable-vectors.mlir:66:15: note: see current operation: %21 = "memref.load"(%20, %arg3) <{nontemporal = false}> : (memref<1x2xvector<[4]xi1>>, index) -> vector<[4]xi1>
# `-----------------------------
# error: command failed with exit status: 1
# executed command: mlir-cpu-runner -e=entry -entry-point-result=void --march=aarch64 --mattr=+sve -shared-libs=/home/tcwg-buildbot/worker/clang-aarch64-sve-vls/stage1/./lib/libmlir_c_runner_utils.so
# .---command stderr------------
# | Error: entry point not found
# `-----------------------------
# error: command failed with exit status: 1
# executed command: /home/tcwg-buildbot/worker/clang-aarch64-sve-vls/stage1/bin/FileCheck /home/tcwg-buildbot/worker/clang-aarch64-sve-vls/llvm/mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/arrays-of-scalable-vectors.mlir
# .---command stderr------------
# | FileCheck error: '<stdin>' is empty.
# | FileCheck command line:  /home/tcwg-buildbot/worker/clang-aarch64-sve-vls/stage1/bin/FileCheck /home/tcwg-buildbot/worker/clang-aarch64-sve-vls/llvm/mlir/test/Integration/Dialect/Vector/CPU/ArmSVE/arrays-of-scalable-vectors.mlir
# `-----------------------------
# error: command failed with exit status: 2
--
********************
```

https://github.com/llvm/llvm-project/pull/75519