[MLIR][Python] fix class name of powf and negf in linalg (#97696)

The following logic can lead to a class name mismatch when using
`linalg.powf` in Python. This PR fixed the issue and also renamed
`NegfOp` to `NegFOp` in linalg to adhere to the naming convention, as
exemplified by `arith::NegFOp`.


https://github.com/llvm/llvm-project/blob/173514d58ec4e6166670f1e37a038df3865c8b96/mlir/python/mlir/dialects/linalg/opdsl/lang/dsl.py#L140-L143
```
# linalg.powf(arg0, arg1, outs=[init_result.result])
NotImplementedError: Unknown named op_name / op_class_name: powf / PowfOp
```

Author: xurui1995

mlir/include/mlir/Dialect/Linalg/IR/LinalgNamedStructuredOps.yaml

@@ -271,7 +271,7 @@ structured_op: !LinalgStructuredOpConfig
 --- !LinalgOpConfig
 metadata: !LinalgOpMetadata
   name: negf
-  cpp_class_name: NegfOp
+  cpp_class_name: NegFOp
   doc: |-
     Applies negf(x) elementwise.
 

mlir/python/mlir/dialects/linalg/opdsl/ops/core_named_ops.py

@@ -96,7 +96,7 @@ def floor(
     O[None] = UnaryFn.floor(I[None])
 
 
-@linalg_structured_op
+@linalg_structured_op(op_class_name="NegFOp")
 def negf(
     I=TensorDef(T1),
     O=TensorDef(T1, output=True),
@@ -330,7 +330,7 @@ def min(
     O[None] = BinaryFn.min_signed(lhs[None], rhs[None])
 
 
-@linalg_structured_op
+@linalg_structured_op(op_class_name="PowFOp")
 def powf(
     lhs=TensorDef(T1),
     rhs=TensorDef(T1),

mlir/test/python/dialects/linalg/opdsl/test_core_named_ops.py

@@ -2,3 +2,7 @@
 
 # Just verify that at least one known op is generated.
 # CHECK: name: matmul
+
+# verify some special cases: negf->NegFOp, powf->PowFOp
+# CHECK cpp_class_name: NegFOp
+# CHECK cpp_class_name: PowFOp