|WARNING| [3/3] test/fib6: reproduce tbl8 leak on failed trie insert

[dpdklab at iol.unh.edu]

Test-Label: ai-code-review
Test-Status: WARNING
https://dpdk.org/patch/171759

_AI Code Review_

=== Patch Review: series.patch (via Claude) ===
Review date: 2026-10-20

# DPDK Patch Review

## PATCH 1/3: fib6: fix temporary tbl8 for byte-aligned prefixes

### Errors

None identified.

### Warnings

None identified.

### Info

The logic in both new conditionals uses the pattern `i >= first_tbl8_byte + !(common_bytes < 3)`.
This is correct: when `common_bytes < 3` is true, `!` yields 0, so the bound is `first_tbl8_byte`;
when false, `!` yields 0 from 1, which produces `first_tbl8_byte + 1`.
However, the expression could be clearer:

```c
/* Current */
for (i = IPV6_MAX_IDX; i >= first_tbl8_byte + !(common_bytes < 3); i--) {

/* Equivalent, more explicit */
uint8_t edge_tbl8_byte = (common_bytes < 3) ? first_tbl8_byte : first_tbl8_byte + 1;
for (i = IPV6_MAX_IDX; i >= edge_tbl8_byte; i--) {
```

This is a style preference, not a defect.

---

## PATCH 2/3: fib6: fix tbl8 reservation with RCU defer queue

### Errors

None identified.

### Warnings

**Missing release notes**: This patch fixes a bug (tbl8 accounting when using QSBR defer mode) affecting user-visible behavior.
The fix should be documented in the release notes for the current release.

**Reclaim parameters**: The `rte_rcu_qsbr_dq_reclaim()` call passes `dp->tbl8_pool_pos - dp->cur_tbl8s` as the `n` parameter,
which is the number of entries to attempt to reclaim.
After reclamation, `tbl8s_in_defer_queue` is recalculated as the same expression.
If `rte_rcu_qsbr_dq_reclaim()` reclaimed fewer entries than requested (partial reclaim),
the second calculation still reflects the actual state (deferred entries that could not be reclaimed remain in the queue).
This is correct behavior.

---

## PATCH 3/3: test/fib6: reproduce tbl8 leak on failed trie insert

### Errors

None identified.

### Warnings

None identified.

### Info

**Test coverage**: The three new test cases effectively exercise the bugs fixed in patches 1 and 2:
- `test_tbl8_install_peak`: validates that byte-aligned prefixes (/32, /24, /16) do not consume temporary tbl8s
- `test_tbl8_install_peak_dq`: verifies that deferred tbl8s are accounted for in reservation checks
- `test_tbl8_dq_modify`: confirms that modifying a prefix with deferred child groups is handled correctly

All three tests use `RTE_TEST_ASSERT` macros and integrate with the DPDK test framework correctly.

---

## Summary

The patch series correctly addresses two related resource accounting bugs in the IPv6 trie implementation:
1. Temporary tbl8 allocation for byte-aligned prefixes
2. Missing accounting for deferred tbl8s in QSBR defer mode

The fixes are localized, the test coverage is comprehensive, and the code follows DPDK conventions.
One warning applies: patch 2/3 should update the release notes.

Review-Result: WARNING