Arm backend: Remove old transform_for_cortex_m_backend and --enable_qdq_fusion_pass (#17740)

Summary:
Remove the transform_for_cortex_m_backend() function and deprecate
--enable_qdq_fusion_pass CLI flag from aot_arm_compiler.py. Instead,
ReplaceQuantNodesPass is now applied directly inside
to_edge_TOSA_delegate() and to_edge_no_delegate(), making each
compilation path self-contained rather than relying on a post-hoc fixup
applied to all targets.
This is a prerequisite for PR #17075, which introduces Cortex-M as a
first-class compilation target with its own dedicated pipeline.


cc @digantdesai @SS-JIA @freddan80 @per @zingo @oscarandersson8218
@mansnils @Sebastian-Larsson @robell

---------

Co-authored-by: Github Executorch <github_executorch@arm.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

Author: 3 people

examples/arm/aot_arm_compiler.py

@@ -38,14 +38,6 @@
 from executorch.backends.arm.vgf import VgfCompileSpec
 
 # To use Cortex-M backend
-from executorch.backends.cortex_m.passes.convert_to_cortex_m_pass import (
-    ConvertToCortexMPass,
-)
-
-from executorch.backends.cortex_m.passes.quantized_op_fusion_pass import (
-    QuantizedOpFusionPass,
-)
-
 from executorch.backends.cortex_m.passes.replace_quant_nodes_pass import (
     ReplaceQuantNodesPass,
 )
@@ -207,6 +199,14 @@ def _load_serialized_model(
     return model, example_inputs
 
 
+def _apply_replace_quant_nodes(edge, args):
+    """Apply the replace_quant_nodes pass to the edge graph module."""
+
+    if args.target != "vgf" and not args.direct_drive:
+        edge = edge.transform([ReplaceQuantNodesPass()])
+    return edge
+
+
 def get_model_and_inputs_from_name(
     model_name: str, model_input: str | None
 ) -> Tuple[torch.nn.Module, Any]:
@@ -606,7 +606,7 @@ def get_args():
     parser.add_argument(
         "--enable_qdq_fusion_pass",
         action="store_true",
-        help="Enable the Quantized qdq fusion Op passes",
+        help="[DEPRECATED] This flag is no longer used and will be removed in a future release.",
     )
     parser.add_argument(
         "--enable_debug_mode",
@@ -787,6 +787,11 @@ def to_edge_TOSA_delegate(
         ),
     )
 
+    # Replace quantized_decomposed::{quantize,dequantize}_per_tensor nodes
+    # with cortex_m:: equivalents for int8 QDQ ops remaining outside the
+    # delegated subgraph.
+    edge = _apply_replace_quant_nodes(edge, args)
+
     return model_quant, edge
 
 
@@ -822,27 +827,12 @@ def to_edge_no_delegate(
         ),
     )
 
-    return model_quant, edge
-
-
-def transform_for_cortex_m_backend(edge_program_manager, args):
-    # Let's make sure we are using optimized Cortex M backend
-    # NB: If we can't find and replace ops those are expected to be replaced,
-    # bad things will happen at runtime, like "missing operator" errors!
+    # Replace quantized_decomposed::{quantize,dequantize}_per_tensor nodes
+    # with cortex_m:: equivalents for int8 QDQ ops remaining outside the
+    # delegated subgraph.
+    edge = _apply_replace_quant_nodes(edge, args)
 
-    # Instantiate the mandatory ReplaceQuantNodesPass
-    passes = [ReplaceQuantNodesPass]
-    if args.enable_qdq_fusion_pass:
-        passes += [ConvertToCortexMPass, QuantizedOpFusionPass]
-    current_edge = edge_program_manager
-    for pass_cls in passes:
-        transform_pass = (
-            pass_cls(current_edge.exported_program())
-            if pass_cls.__name__ == "QuantizedLinearFusionPass"
-            else pass_cls()
-        )
-        current_edge = current_edge.transform([transform_pass])
-    return current_edge
+    return model_quant, edge
 
 
 if __name__ == "__main__":  # noqa: C901
@@ -863,6 +853,13 @@ def transform_for_cortex_m_backend(edge_program_manager, args):
     model = exported_program.module()
     model_fp32 = model
 
+    if args.enable_qdq_fusion_pass:
+        logging.warning(
+            "--enable_qdq_fusion_pass is deprecated and has no effect. "
+            "Quantized node replacement is now handled within the "
+            "respective compilation paths."
+        )
+
     model_name = os.path.basename(os.path.splitext(args.model_name)[0])
     if args.intermediates:
         os.makedirs(args.intermediates, exist_ok=True)
@@ -885,11 +882,6 @@ def transform_for_cortex_m_backend(edge_program_manager, args):
             exported_program, args, model, example_inputs
         )
 
-    # Cortex-m ops are never included in vgf or direct-drive
-    if args.target != "vgf" and not args.direct_drive:
-        # Transform so we can use ops from the Cortex M backend
-        edge = transform_for_cortex_m_backend(edge, args)
-
     dump_delegation_info(edge, args.intermediates)
 
     edge_program_manager_copy = copy.deepcopy(edge)