Arm: fix size_t log format specifiers

[ricardoasouz]

Fixes #13038.

Summary

Replace the remaining %zu / %zd format specifiers in the Arm executor runner with %lu using the existing printf_size_t logging type.

The runner already documents this convention because newlib-nano printf does not reliably handle the z length modifier on the bare-metal/FVP path. A few logging sites still used z, which could result in malformed output on FVP.

Validation

• git diff --check
• ExecuTorch lintrunner: no lint issues
• Built arm_executor_runner for ethos-u55-128 with Arm GNU Toolchain 15.2.1
• Ran on Corstone-300 FVP with Ethos-U55-128
• Enabled:
  • ET_ARM_ETHOSU_PER_DELEGATE_PROFILING=ON
  • ET_ARM_ETHOSU_PROFILE_IO_COPIES=ON
• Verified one Ethos-U delegated partition executed
• Verified per-delegate PMU and I/O-copy logs render numeric indices correctly
• FVP completed with exit status 0 and reported no problems

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/22946

• 📄 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:

• linux.dgx.b200 runner queuing exceeds 4 hours and dgxb200-05 "CUDA not available"

❌ 1 New Failure, 1 Pending

As of commit f27edb6 with merge base 6cbdc03 ():

NEW FAILURE - The following job has failed:

• pull / unittest-buck / macos / macos-job (gh)
  torchaudio

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

[pytorch-bot]

Workflows were awaiting approval. CI has now been triggered for the ciflow labels on this PR.

[ricardoasouz]

@pytorchbot label "release notes: none"

[Erik-Lundell]

Hey, thanks for the fix! The runner defines printf_size_t, could you use that instead of unsigned long?

[ricardoasouz]

Hey, thanks for the fix! The runner defines printf_size_t, could you use that instead of unsigned long?

Thanks! Updated the casts introduced by this PR to use printf_size_t. arm_perf_monitor.cpp defines the alias locally, consistent with the other Arm translation units. Re-validated on Corstone-300 FVP with per-delegate and I/O-copy profiling enabled.