Arm backend: Replace VkQueueWaitIdle with fences

[wwwind]

Use fences to wait for the submitted VGF command buffer instead of calling vkQueueWaitIdle(). This avoids stalling the whole queue and only waits for the work submitted by the backend.

Create the execution fence once during VGF setup and reuse it across execute_vgf() calls by resetting it before each submission. Keep the one-shot fence helper for temporary setup submissions.

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/20186

• 📄 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, 2 Unclassified Failures

As of commit ca52cb4 with merge base cec0814 ():

NEW FAILURE - The following job has failed:

• trunk / test-qnn-model (fp32, ic3) / linux-job (gh)
  RuntimeError: Command docker exec -t dd90ffbfb3ab9f1d97a0906fe8842b9a56341a58e6ffcf71d702c2a24f6f7c5c /exec failed with exit code 1

UNCLASSIFIED FAILURES - DrCI could not classify the following jobs because the workflow did not run on the merge base. The failures may be pre-existing on trunk or introduced by this PR:

• Check Labels / Check labels (gh) (this job did not run on the merge base, so DrCI cannot tell whether the failure is pre-existing)
  Process completed with exit code 1.
• Labeler / label (gh) (this job did not run on the merge base, so DrCI cannot tell whether the failure is pre-existing)
  Requires authentication

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

[robell]

Approved our side.

[zingo]

OK to merge if tests pass

[wwwind]

the new failure is unrelevant. fails to download llvm