|WARNING| [v2] net/i40e: add IEEE1588 handling to AVX vector paths

[dpdklab at iol.unh.edu]

Test-Label: ai-code-review
Test-Status: WARNING
https://dpdk.org/patch/171705

_AI Code Review_

=== Patch Review: series.patch (via Claude) ===
Review date: 2026-10-19

# Review of DPDK Patch: net/i40e IEEE1588 handling in AVX vector paths

## ERRORS

None identified.

## WARNINGS

### Missing Release Notes

This patch adds IEEE1588 support to i40e AVX2/AVX512 vector Rx paths -- a user-visible feature enhancement that changes when the vector path can be used.
The release notes (`doc/guides/rel_notes/release_*.rst`) should document this change in the "New Features" or "Drivers" section.

Applications that previously fell back to scalar Rx when IEEE1588 was enabled will now get vector Rx on x86 AVX2/AVX512, which may affect performance characteristics.

**Suggested action:** Add a release note documenting that i40e vector Rx (AVX2/AVX512 only) now supports IEEE1588, removing the need to fall back to scalar Rx for PTP workloads on those platforms.

---

### Missing PMD Features Documentation

The patch enables IEEE1588 timesync on the i40e vector Rx paths, but there is no mention of updating the feature matrix in `doc/guides/nics/features/i40e.ini` or the i40e driver documentation.

IEEE1588 support is already in the matrix (it was supported on scalar paths), but if the limitations have changed (now works on AVX2/AVX512 vector, but not other vector paths like SSE/NEON), that should be documented so users understand when they will or won't get vector Rx with PTP enabled.

**Suggested action:** Verify `doc/guides/nics/features/i40e.ini` correctly reflects IEEE1588 support, and if there are architecture-specific limitations (works on x86 AVX2/512 vector but not ARM NEON), document them in `doc/guides/nics/i40e.rst`.

---

## INFO

### Code Duplication Between AVX2 and AVX512

The IEEE1588 handling block is duplicated nearly verbatim between `i40e_rxtx_vec_avx2.c` and `i40e_rxtx_vec_avx512.c` (59 vs 61 lines, with identical logic).

This is consistent with the existing pattern in this driver (FDIR handling is also duplicated), so it is not a defect, but it does increase maintenance burden.
A future refactor could extract the common logic into a shared macro or static inline function in `i40e_rxtx_vec_common.h`.

Not a requirement for this patch, just an observation for future cleanup.

---

Review-Result: WARNING