|WARNING| [2/2] net/iavf: disable runtime queue setup during queue rate limiting

[dpdklab at iol.unh.edu]

Test-Label: ai-code-review
Test-Status: WARNING
https://dpdk.org/patch/166634

_AI Code Review_

=== Patch Review: series.patch (via Claude) ===
Review date: 2026-07-09

# DPDK Patch Review

## Patch 1/2: doc: fix ice scheduler rate-limiter burst size description

### Info

**Documentation clarity improvement:**
The rewording improves readability by fixing the line-break issue and adds useful information about out-of-range value rejection.

**No issues found.**

---

## Patch 2/2: net/iavf: disable runtime queue setup during queue rate limiting

### Errors

**Resource leak in iavf_hierarchy_commit() (CORRECTED):**
The patch FIXES a pre-existing memory leak by adding `rte_free(vf->qtc_map)` before reassignment. This is correct.

**No errors found.**

### Warnings

**Release notes entry is incomplete:**
The release notes describe the new behavior (runtime queue setup disabled during rate limiting) but do not mention that the patch also **fixes two memory management bugs** (qtc_map leak on re-commit, and stale pointer on VF reset). These are correctness fixes and should be documented.

Suggested addition to `doc/guides/rel_notes/release_26_07.rst`:

```rst
* **Updated Intel iavf driver.**

  * Runtime Rx/Tx queue setup is now automatically disabled while a
    per-queue bandwidth ``rte_tm`` hierarchy is committed, to avoid
    corrupting queue state on E810 VFs.
  * Fixed memory leak when re-committing a ``rte_tm`` hierarchy.
  * Fixed stale pointer to freed memory after VF reset.
```

### Info

**Design approach:**
The solution is well-chosen: dynamically withdrawing the runtime queue setup capability when the race condition exists is cleaner than requiring applications to know about a devarg. The use of `qtc_map` as the signal is appropriate since it directly represents the committed state that causes the race.

**Code clarity:**
The comment in `iavf_dev_info_get()` clearly explains why the capability is conditionally advertised. The comment in `iavf_uninit_vf()` explains why the free is needed. Both are helpful.

**Correctness verification:**
The fix correctly addresses all three issues:
1. Race condition: capability withdrawn when `qtc_map` is set
2. Memory leak: `rte_free(vf->qtc_map)` added before reassignment in `iavf_hierarchy_commit()`
3. Stale pointer: `rte_free(vf->qtc_map)` and `vf->qtc_map = NULL` added in `iavf_uninit_vf()`

---

## Summary

**Patch 1/2:** Clean documentation fix. No issues.

**Patch 2/2:** The code changes are correct and fix real bugs. The only issue is that the release notes focus on the new behavioral restriction but omit the memory management fixes that are also part of this patch. Add a note about the leak and stale pointer fixes to the release notes.