Cortex_m backend: Fix add implementation

[AdrianLundell]

• Call CMSIS-NN kernel with correct argument order and signs
• Change python implementation to reflect CMSIS-NN behaviour
• Fix scale calculations
• Remove broken broadcasting support
• Add pass to lower scalar version ops
• Remove unused definitions/ implementations in operators.py, operators.yaml and op_quantized_add.cpp

Note: arm_elementwise_add_s8 does not natively support broadcasting, so simply resizing the output tensor will not work. Enabling this in an efficient way is not straight forward, so avoid fusing these ops for now to avoid break graphs.

cc @freddan80 @per @zingo @oscarandersson8218 @digantdesai

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/15100

• 📄 Preview Python docs built from this PR

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

❗ 1 Active SEVs

There are 1 currently active SEVs. If your PR is affected, please view them below:

• AWS was down, GHA infrastructure effected / recovering

❌ 7 New Failures, 4 Unrelated Failures

As of commit 8b68170 with merge base 73f7286 ():

NEW FAILURES - The following jobs have failed:

• pull / test-samsung-models-linux / linux-job (gh)
  RuntimeError: Command docker exec -t 7fb1ba35213eb5ced469bf15d867de4105158085f6ac719275ffdecaee775418 /exec failed with exit code 1
• pull / unittest-arm-backend-with-no-fvp (test_pytest_models) / linux-job (gh)
  RuntimeError: Command docker exec -t 94a0befb5f02d0abd72221462061a10018a6b6ec2e281f56fa6c499fc18346a1 /exec failed with exit code 1
• pull / unittest-nxp-neutron / linux-job (gh)
  RuntimeError: Command docker exec -t e018987cc0d4b64cb924ec2957a5108af001be75267c700c9acaaea4d87fdf60 /exec failed with exit code 1
• Test CUDA Builds / export-voxtral-cuda-non-quantized / linux-job (gh)
  RuntimeError: Command docker exec -t 1b317972e0706f7fc6c58f70d04dec47f7da1e8559a4a80bd403213e43580ded /exec failed with exit code 2
• Test CUDA Builds / export-voxtral-cuda-quantized-int4-tile-packed / linux-job (gh)
  RuntimeError: Command docker exec -t 0f5d45ef14092aebaf9d351016446131d31d0fd4688979d3c2da2caf0cc77d10 /exec failed with exit code 2
• Test CUDA Builds / export-voxtral-cuda-quantized-int4-weight-only / linux-job (gh)
  RuntimeError: Command docker exec -t 214274a0a8eae5c33c386da2da6fae01955b18682638931796299f6f10dc1083 /exec failed with exit code 2
• Test Metal Backend / export-voxtral-metal-artifact / macos-job (gh)
  RuntimeError: Command bash /Users/ec2-user/runner/_work/_temp/exec_script failed with exit code 2

BROKEN TRUNK - The following jobs failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• Build documentation / build (buck2) / Build doc (gh) (trunk failure)
• pull / android / run-emulator (gh) (trunk failure)
  The process '/usr/bin/sh' failed with exit code 255
• pull / unittest / linux / linux-job (gh) (trunk failure)
• pull / unittest-editable / linux / linux-job (gh) (trunk failure)

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

[psiddh]

Should we the add an explicit shape check before running the add operation ?
If shapes do not match, return an error and log a clear message (e.g., "Input tensors must have the same shape for CMSIS-NN add operation. Broadcasting is not supported."). (or) something like that..