|WARNING| drivers/bus: remove blocklist evaluation when probing a device

[dpdklab at iol.unh.edu]

Test-Label: ai-code-review
Test-Status: WARNING
https://dpdk.org/patch/166582

_AI Code Review_

=== Patch Review: series.patch (via Claude) ===
Review date: 2026-07-06

# DPDK Patch Review

## Summary
This patch removes redundant blocklist evaluation from bus driver probe functions. The EAL already performs this check before calling `.probe_device`, making the bus-level checks unnecessary.

## Findings

### Errors
None.

### Warnings

**1. Missing release notes update**

API/behavioral change without release notes. While this is an internal implementation change (removal of redundant logic), it alters the behavior visible to debug logs and potentially affects return codes from probe functions. A brief note in the release notes would document this cleanup.

**Suggested action:**
Add a line to `doc/guides/rel_notes/release_26_07.rst` (or current release) under a "Bus Drivers" or "Internal Changes" section:
```rst
* **Removed redundant blocklist checks from bus probe functions.**

  Bus drivers (PCI, auxiliary, FSLMC, VMBUS) no longer evaluate device
  blocklist policy during probe. EAL performs this check before calling
  the bus probe function, making the bus-level checks redundant.
```

**2. Inconsistent return values removed**

The removed code blocks returned inconsistent error codes:
- `auxiliary`: returns `-1` for blocked devices
- `fslmc`: returns `0` for blocked devices  
- `pci`: returns `1` for blocked devices
- `vmbus`: returns `1` for blocked devices

While removing this inconsistency is good, the patch does not document whether callers (EAL probe logic) ever observe these return values or if they are always short-circuited by the prior EAL blocklist check.

**Suggested action:**
Add a comment in the commit message explaining that EAL's blocklist check prevents these code paths from ever being reached, confirming that the inconsistent return values were dead code.

**3. Log message removal not mentioned**

The patch silently removes informational log messages ("Device is blocked, not initializing" / "Device %s Blocked, skipping"). Applications or test frameworks may have depended on these log messages to confirm blocklist behavior. While the removal is correct (EAL logs this information), it's a user-visible change.

**Suggested action:**
Mention in the commit message or release notes that the per-bus "device blocked" log messages are removed since EAL already logs blocklist actions.

### Info

**1. Verify EAL blocklist check coverage**

The patch assumes EAL always checks `devargs->policy == RTE_DEV_BLOCKED` before calling `.probe_device`. This should be verified in the EAL probe path (`lib/eal/common/eal_common_dev.c` or equivalent). If the EAL check is conditional or has exceptions, removing these guards could expose devices to probing when they should be blocked.

**Suggested verification:**
Confirm that `local_dev_probe()` or equivalent EAL probe function unconditionally checks `devargs->policy` before invoking the bus `.probe_device` callback.

---

## Code Style and Correctness
No style violations or correctness bugs identified. The removed code blocks are straightforward conditionals with no resource management or complex logic.

## Testing
No test additions needed for a removal of redundant logic. Existing blocklist tests should continue to pass if EAL correctly performs the check.

---

## Conclusion
The patch is fundamentally sound--it removes dead code that duplicates EAL functionality. The main concern is ensuring that the change is documented (release notes, commit message clarification) and that the assumption about EAL's blocklist enforcement is verified to be universally true.