|WARNING| net/iavf: fix AdminQ polling after interrupt setup failure

[dpdklab at iol.unh.edu]

Test-Label: ai-code-review
Test-Status: WARNING
https://dpdk.org/patch/172329

_AI Code Review_

=== Patch Review: series.patch (via Claude) ===
Review date: 2026-11-02

# DPDK Patch Review

## ERRORS

None identified.

## WARNINGS

**W1: Missing release notes update**
The patch fixes a significant functional bug (configuration failure on FreeBSD) but does not update the release notes. API changes and driver bug fixes that affect user-facing behavior should be documented in `doc/guides/rel_notes/release_*.rst`.

Suggested fix: Add an entry to the "Fixes" section of the current release notes documenting that AdminQ polling now works correctly when interrupt registration fails (e.g., on FreeBSD with nic_uio).

## INFO

None.

---

## Review Summary

The patch addresses a real correctness issue where the driver incorrectly determines whether to use alarm polling for AdminQ events. The fix is sound:

1. **Correctness**: The new `adminq_polling` flag correctly tracks the actual polling mode instead of deriving it from the PF capability alone. The initialization sequence properly handles the interrupt registration failure case.

2. **Resource management**: The error path correctly unregisters the interrupt callback with `rte_intr_callback_unregister_sync()` when `rte_intr_enable()` fails, preventing a callback leak.

3. **Error propagation**: Return values from `rte_intr_callback_register()` and `rte_intr_enable()` are properly checked. The logic correctly falls back to polling when either fails.

4. **Code style**: Follows DPDK conventions. Variable naming is appropriate. Comment placement is correct.

5. **Logic**: The condition inversion from checking the PF capability flag to checking the state flag is correct and makes the code's intent clearer.

The mailmap update and commit metadata are appropriate. The only issue is the missing release notes update for a user-visible behavioral fix.

Review-Result: WARNING