[MLIR] `--affine-loop-fusion` crashes during dependence analysis

### Description

I encountered a crash when running `--affine-loop-fusion` on the following MLIR input.
The crash is triggered during affine dependence analysis and results in an `UNREACHABLE` assertion in mlir/lib/Dialect/Affine/Analysis/Utils.cpp:256.

### Tested commit

[3b8ee89](https://github.com/llvm/llvm-project/commit/3b8ee89006f5a228fce75528f8dd80b4229abcc1)

### Test case

```mlir
#map = affine_map<(d0) -> (d0)>
#map1 = affine_map<(d0) -> (d0 + 1)>
module {
func.func @main(%arg0: tensor<4xf16>) {
%0 = ub.poison : f16
%cst = arith.constant 1.000000e+00 : f16
%1 = bufferization.to_buffer %arg0 : tensor<4xf16> to memref<4xf16, strided<[?], offset: ?>>
%alloc = memref.alloc() : memref<4xf16, 1>
%alloc_0 = memref.alloc() : memref<4xf16>
affine.for %arg1 = 0 to 4 {
affine.for %arg2 = #map(%arg1) to #map1(%arg1) {
%2 = affine.load %1[%arg2] : memref<4xf16, strided<[?], offset: ?>>
%3 = arith.negf %2 : f16
%4 = math.exp %3 : f16
%5 = arith.addf %4, %cst : f16
%6 = arith.divf %cst, %5 : f16
affine.store %6, %alloc[%arg2] : memref<4xf16, 1>
}
}
affine.for %arg1 = 0 to 4 step 128 {
%2 = vector.transfer_read %alloc[%arg1], %0 : memref<4xf16, 1>, vector<128xf16>
}
return
}
}
```

### Steps to reproduce

```
mlir-opt test.mlir --affine-loop-fusion
```

### stack trace

```
unexpected op
UNREACHABLE executed at /home/workdir/llvm-project-latest/mlir/lib/Dialect/Affine/Analysis/Utils.cpp:256!
PLEASE submit a bug report to https://github.com/llvm/llvm-project/issues/ and include the crash backtrace and instructions to reproduce the bug.
Stack dump:
0. Program arguments: mlir-opt test.mlir --affine-loop-fusion
#0 0x0000555a2cb1cd0b llvm::sys::PrintStackTrace(llvm::raw_ostream&, int) (../llvm-project-latest/build/bin/mlir-opt+0x1fc2d0b)
#1 0x0000555a2cb19b11 llvm::sys::RunSignalHandlers() (../llvm-project-latest/build/bin/mlir-opt+0x1fbfb11)
#2 0x0000555a2cb1da3b SignalHandler(int, siginfo_t*, void*) Signals.cpp:0:0
#3 0x00007f0704ff6520 (/usr/lib/x86_64-linux-gnu/libc.so.6+0x42520)
#4 0x00007f070504a9fc pthread_kill (/usr/lib/x86_64-linux-gnu/libc.so.6+0x969fc)
#5 0x00007f0704ff6476 gsignal (/usr/lib/x86_64-linux-gnu/libc.so.6+0x42476)
#6 0x00007f0704fdc7f3 abort (/usr/lib/x86_64-linux-gnu/libc.so.6+0x287f3)
#7 0x0000555a2cb0313f (../llvm-project-latest/build/bin/mlir-opt+0x1fa913f)
#8 0x0000555a332e0c1c (../llvm-project-latest/build/bin/mlir-opt+0x8786c1c)
#9 0x0000555a332e098e bool __gnu_cxx::__ops::_Iter_pred, llvm::ArrayRef) const::'lambda'(mlir::Operation*)>::operator()(mlir::Operation* const*) Utils.cpp:0:0
#10 0x0000555a332cb687 mlir::affine::MemRefDependenceGraph::init(bool) (../llvm-project-latest/build/bin/mlir-opt+0x8771687)
#11 0x0000555a33029e56 void llvm::function_ref::callback_fn<(anonymous namespace)::LoopFusion::runOnOperation()::$_0>(long, mlir::Operation*) LoopFusion.cpp:0:0
#12 0x0000555a2cbdb47e void mlir::detail::walk(mlir::Operation*, llvm::function_ref, mlir::WalkOrder) (../llvm-project-latest/build/bin/mlir-opt+0x208147e)
#13 0x0000555a33027c16 (anonymous namespace)::LoopFusion::runOnOperation() LoopFusion.cpp:0:0
#14 0x0000555a34d1c3ee mlir::detail::OpToOpPassAdaptor::run(mlir::Pass*, mlir::Operation*, mlir::AnalysisManager, bool, unsigned int) (../llvm-project-latest/build/bin/mlir-opt+0xa1c23ee)
#15 0x0000555a34d1d364 mlir::detail::OpToOpPassAdaptor::runPipeline(mlir::OpPassManager&, mlir::Operation*, mlir::AnalysisManager, bool, unsigned int, mlir::PassInstrumentor*, mlir::PassInstrumentation::PipelineParentInfo const*) (../llvm-project-latest/build/bin/mlir-opt+0xa1c3364)
#16 0x0000555a34d240fa mlir::PassManager::runPasses(mlir::Operation*, mlir::AnalysisManager) (../llvm-project-latest/build/bin/mlir-opt+0xa1ca0fa)
#17 0x0000555a34d23a5d mlir::PassManager::run(mlir::Operation*) (../llvm-project-latest/build/bin/mlir-opt+0xa1c9a5d)
#18 0x0000555a2cbc858f performActions(llvm::raw_ostream&, std::shared_ptr const&, mlir::MLIRContext*, mlir::MlirOptMainConfig const&) MlirOptMain.cpp:0:0
#19 0x0000555a2cbc7761 llvm::LogicalResult llvm::function_ref>, llvm::MemoryBufferRef const&, llvm::raw_ostream&)>::callback_fn>, mlir::DialectRegistry&, mlir::MlirOptMainConfig const&)::$_0>(long, std::unique_ptr>, llvm::MemoryBufferRef const&, llvm::raw_ostream&) MlirOptMain.cpp:0:0
#20 0x0000555a35076ec6 mlir::splitAndProcessBuffer(std::unique_ptr>, llvm::function_ref>, llvm::MemoryBufferRef const&, llvm::raw_ostream&)>, llvm::raw_ostream&, llvm::StringRef, llvm::StringRef) (../llvm-project-latest/build/bin/mlir-opt+0xa51cec6)
#21 0x0000555a2cbbdba6 mlir::MlirOptMain(llvm::raw_ostream&, std::unique_ptr>, mlir::DialectRegistry&, mlir::MlirOptMainConfig const&) (../llvm-project-latest/build/bin/mlir-opt+0x2063ba6)
#22 0x0000555a2cbbdf7a mlir::MlirOptMain(int, char**, llvm::StringRef, llvm::StringRef, mlir::DialectRegistry&) (../llvm-project-latest/build/bin/mlir-opt+0x2063f7a)
#23 0x0000555a2cbbe1eb mlir::MlirOptMain(int, char**, llvm::StringRef, mlir::DialectRegistry&) (../llvm-project-latest/build/bin/mlir-opt+0x20641eb)
#24 0x0000555a2cb027cd main (../llvm-project-latest/build/bin/mlir-opt+0x1fa87cd)
#25 0x00007f0704fddd90 (/usr/lib/x86_64-linux-gnu/libc.so.6+0x29d90)
#26 0x00007f0704fdde40 __libc_start_main (/usr/lib/x86_64-linux-gnu/libc.so.6+0x29e40)
#27 0x0000555a2cb02325 _start (../llvm-project-latest/build/bin/mlir-opt+0x1fa8325)
Aborted (core dumped)
```

Start by running `mlir-opt test.mlir --affine-loop-fusion` with the supplied MLIR input and inspect `mlir/lib/Dialect/Affine/Analysis/Utils.cpp:256`, following the stack through `MemRefDependenceGraph::init` and `LoopFusion.cpp`. Add a regression test using this reproducer and verify that affine loop fusion no longer aborts during dependence analysis.

Written by the indexing model from the issue text.