|WARNING| net/bnxt: fix flow db lock-free issue during shutdown

[dpdklab at iol.unh.edu]

Test-Label: ai-code-review
Test-Status: WARNING
https://dpdk.org/patch/170563

_AI Code Review_

=== Patch Review: series.patch (via Claude) ===
Review date: 2026-10-05

# DPDK Patch Review

## Summary
The patch fixes a lock-use-after-destroy bug in the bnxt driver shutdown path by reordering `ulp_tfc_tbl_scope_deinit()` to execute before `pthread_mutex_destroy()`.

---

## Errors

### 1. Use-after-free / Lock-use-after-destroy

**Original code:**
```c
pthread_mutex_destroy(&bp->ulp_ctx->cfg_data->flow_db_lock);
ulp_tfc_tbl_scope_deinit(bp);
```

**Problem:** `ulp_tfc_tbl_scope_deinit()` likely accesses `flow_db_lock` during table scope cleanup. Calling it after `pthread_mutex_destroy()` results in undefined behavior - the function would attempt to lock/unlock a destroyed mutex.

**Fix:** The patch correctly reorders the calls:
```c
ulp_tfc_tbl_scope_deinit(bp);
pthread_mutex_destroy(&bp->ulp_ctx->cfg_data->flow_db_lock);
```

This is the correct fix for the bug described.

---

## Warnings

None.

---

## Info

None.

---

## Conclusion

The patch is **correct** and properly addresses a real correctness bug. The fix follows the standard pattern for resource cleanup: use resources first, then destroy synchronization primitives. The patch makes no other changes and is appropriately scoped to the single bug being fixed.