|WARNING| [v3] net/iavf: add PTP clock read support

[dpdklab at iol.unh.edu]

Test-Label: ai-code-review
Test-Status: WARNING
https://dpdk.org/patch/168930

_AI Code Review_

=== Patch Review: series.patch (via Claude) ===
Review date: 2026-09-02

# DPDK Patch Review

## Summary
The patch adds PTP clock read support to the Intel IAVF driver by implementing the `timesync_read_time` ethdev API. The implementation fetches PHC time from the PF via virtchnl.

---

## Errors

### 1. MTU-related errors do not apply
After reviewing the patch content, there are no MTU, scatter Rx, or frame length calculations present. The MTU-related checks from the guidelines are not applicable to this patch.

### 2. No correctness bugs identified
The error paths properly handle capability checks and return appropriate error codes. The `iavf_phc_get_time` function uses an existing lock (`phc_time_aq_lock`) which prevents races. Resource management appears correct - no allocations are made that could leak.

---

## Warnings

### 1. Missing release notes update
**Issue:** The patch adds new functionality (`timesync_read_time` API implementation) but does not update the release notes.

**Why it matters:** Per the guidelines, changes that add API implementations to drivers should be documented in release notes.

**Suggested fix:** Add an entry to `doc/guides/rel_notes/release_26_03.rst` (or current release) under the "New Features" section:
```rst
* **Added PTP clock read support to iavf driver.**

  Added support for reading the PTP Hardware Clock (PHC) via the
  ``timesync_read_time`` ethdev API when the PF grants the capability.
```

### 2. Missing feature documentation update
**Issue:** The patch adds PTP-related functionality but there's no indication that the PMD features matrix was updated.

**Why it matters:** PMD features must match the features matrix in `doc/guides/nics/features/`.

**Suggested fix:** Verify that `doc/guides/nics/features/iavf.ini` is updated with the timesync capability if appropriate, or confirm in commit message that the feature matrix doesn't cover this specific API.

---

## Info

### 1. Function signature change could be simplified
**Observation:** The refactoring splits `iavf_get_phc_time` into two functions - one that takes `adapter` and `time` pointer, and a wrapper that extracts `phc_time` from `rxq`. This is reasonable for code reuse.

**Alternative:** The naming is slightly confusing since both functions have similar names. Consider:
- `iavf_phc_get_time` (new, adapter-based)
- `iavf_get_phc_time` (existing, rxq-based)

A clearer distinction might be `iavf_adapter_read_phc` and `iavf_rxq_update_phc_time`, but the current naming is acceptable.

### 2. Error code consistency
**Observation:** The `iavf_timesync_read_time` function returns `-EIO` for both closed adapter and virtchnl command failure cases. This is acceptable but could be more specific:
- `-ENODEV` for closed adapter
- `-EIO` for communication failure

Current implementation is within guidelines.

---

## Summary of Required Changes

**Must fix:**
- None - no correctness bugs found

**Should fix:**
1. Add release notes entry for the new feature
2. Verify/update PMD features documentation

**Consider:**
- (Optional) More specific error codes in `iavf_timesync_read_time`

---

## Code Quality Assessment

 **Correctness:** No bugs, race conditions, or resource leaks identified  
 **Style:** Follows DPDK C coding style  
 **API usage:** Proper use of `rte_ns_to_timespec`, capability checks  
 **Documentation:** Missing release notes (Warning level)

The code changes are functionally sound. The main gap is documentation - release notes should be added to document this new capability.