Cortex-M backend: Refactor quantized_op_fusion_pass

[AdrianLundell]

Move to use the AtenToCortexMPass instead.

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

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/20179

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 1 New Failure, 78 Pending, 4 Unrelated Failures

As of commit d107ddb with merge base 2b9e9bf ():

NEW FAILURE - The following job has failed:

• pull / test-multimodal-linux (gemma3-4b) / linux-job (gh)
  RuntimeError: Command docker exec -t f418c26acd9e63cd25a438a1c5d7535f406d1a6afd72682dcdb137ba7d75adb1 /exec failed with exit code 139

FLAKY - The following jobs failed but were likely due to flakiness present on trunk:

• periodic / test-models-linux (cmake, mv3, portable, linux.2xlarge, 90) / linux-job (gh) (matched linux rule in flaky-rules.json)
  The process '/usr/bin/git' failed with exit code 128
• pull / test-arm-backend-no-driver (test_run_tosa) / linux-job (gh) (detected as infra flaky with no log or failing log classifier)
• trunk / test-arm-backend-vkml (test_pytest_models_vkml) / linux-job (gh) (matched linux rule in flaky-rules.json)
  The process '/usr/bin/git' failed with exit code 128
• trunk / test-torchao-huggingface-checkpoints (phi_4_mini, linux.2xlarge, executorch-ubuntu-22.04-clang12,... / linux-job (gh) (matched linux rule in flaky-rules.json)
  The process '/usr/bin/git' failed with exit code 128

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[Copilot]

Pull request overview

This PR refactors the Cortex-M backend pass pipeline by removing the standalone QuantizedOpFusionPass and moving its operator replacement/fusion responsibilities into AtenToCortexMPass, simplifying pass management and consolidation of Cortex-M dialect substitutions.

Changes:

• Deleted quantized_op_fusion_pass.py and removed it from imports/build rules and the Cortex-M pass list.
• Extended AtenToCortexMPass with additional dialect substitutions for quantized add/mul/softmax/maxpool/min/max/permute/pad (and quantize/dequantize per-tensor).
• Updated BUCK targets and __init__.py exports to reflect the refactor.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
backends/cortex_m/passes/quantized_op_fusion_pass.py	Removed the old ExportPass-based fusion/replacement implementation.
backends/cortex_m/passes/cortex_m_pass_manager.py	Dropped QuantizedOpFusionPass from the pass list and imports.
backends/cortex_m/passes/BUCK	Removed the deleted file from build sources.
backends/cortex_m/passes/aten_to_cortex_m_pass.py	Added/ported substitution logic that previously lived in QuantizedOpFusionPass.
backends/cortex_m/passes/__init__.py	Removed export/import of QuantizedOpFusionPass.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[rascani]

LGTM, just a couple minor nits.

[rascani]

Should we also remove the ReplaceQuantNodePass?

[AdrianLundell]

It is used in aot_arm_compiler so it would be a slightly larger fix and should perhaps be treated as an API which should be deprecated properly, so maybe in another PR?

[Copilot]

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.