|WARNING| [v2] dts: update single core performance test configuration

[dpdklab at iol.unh.edu]

Test-Label: ai-code-review
Test-Status: WARNING
https://dpdk.org/patch/172117

_AI Code Review_

=== Patch Review: series.patch (via Claude) ===
Review date: 2026-10-27

# DPDK DTS Patch Review

## ERRORS

**1. Missing null check before dictionary access (line 168-169)**

```python
params["pass"] = float(params["performance_delta"]) >= -params.get(
    "delta_tolerance", self.delta_tolerance
)
```

The code calls `params.get("delta_tolerance", self.delta_tolerance)` which will return `self.delta_tolerance` if the key is missing. However, `self.delta_tolerance` is declared as `delta_tolerance: float | None = 0.05` (line 65), meaning it can be `None`. If a user sets `delta_tolerance: null` in the YAML config and omits per-parameter `delta_tolerance`, this will cause `TypeError: '>=' not supported between instances of 'float' and 'NoneType'` at runtime.

**Fix:** Either remove `| None` from the type annotation (making 0.05 a true default), or add a null check:
```python
tolerance = params.get("delta_tolerance", self.delta_tolerance)
if tolerance is None:
    tolerance = 0.05  # or raise error
params["pass"] = float(params["performance_delta"]) >= -tolerance
```

**2. Same null-safety issue in verify call (line 176-178)**

```python
f"""Packets forwarded is less than {
    (1 - params.get("delta_tolerance", self.delta_tolerance)) * 100
}%
```

If `self.delta_tolerance` is `None`, this will raise `TypeError: unsupported operand type(s) for -: 'int' and 'NoneType'`.

**Fix:** Use the same tolerance resolution as above before the arithmetic.

## WARNINGS

**1. Configuration validation missing**

The `test_parameters` list items can have optional `delta_tolerance` fields (per line 35, 58), but there is no validation that each parameter dict contains the required keys (`frame_size`, `num_descriptors`, `expected_mpps`). If a user omits one of these in the YAML, the code will raise `KeyError` at runtime rather than failing early with a clear message.

**Suggest:** Add validation in `set_up_suite()` or use Pydantic models for the test parameters.

**2. Release notes missing**

This patch adds user-facing configuration options (`cores`, per-parameter `delta_tolerance`, `traffic_duration`, `test_repetitions`) and changes test behavior. Per the guidelines, "Changes to existing API require release notes" and "New drivers or subsystems must have release notes". While this is test framework configuration rather than library API, user-facing test behavior changes should be documented in release notes so users know the new fields are available.

**Suggest:** Update `doc/guides/rel_notes/release_25_03.rst` (or the current release file) with a note about the new test configuration options.

## INFO

**1. Comment grammar (line 36)**

```python
#   `cores`:
#     The number of core to provide TREX traffic generator.
```

Minor: "core" should be "cores" (plural).

**2. YAML comment formatting inconsistency**

The example config comments in `tests_config.example.yaml` (lines 5-14) use a different indentation style than the rest of the file. This is cosmetic but worth noting for consistency.

---

**Pass 1 (correctness bugs):** The `None` type in `delta_tolerance` can cause arithmetic operations to fail at runtime when the user sets `delta_tolerance: null` or omits it in YAML. This is a correctness bug -- the code will crash in a reachable path.

**Pass 2 (style/process false positives):** No style items were flagged that I concluded were acceptable.

**Pass 3 (evidence verification):** 
- Error 1: Line 168-169 shows `params.get("delta_tolerance", self.delta_tolerance)` where `self.delta_tolerance` can be `None` per line 65, used in arithmetic operation.
- Error 2: Line 176-178 shows the same pattern in the f-string.
- Warning 1: Lines 134-158 show `params["frame_size"]`, `params["num_descriptors"]` access without validation.
- Warning 2: No release notes file modified in the patch.

Review-Result: ERROR