Claude review: drm/tyr: Add generic slot manager

[Claude Code Review Bot <claude-review@example.com>]

Patch Review

**Author:** Boris Brezillon

This is the most architecturally significant patch. The `SlotManager` abstraction is clean and well-documented.

A few observations:

1. In `idle_slot()`, the seat is read and then immediately overwritten:

```rust
+    fn idle_slot(&mut self, slot_idx: usize, locked_seat: &LockedSeat<T, MAX_SLOTS>) -> Result {
+        let slot = take(&mut self.slots[slot_idx]);
+
+        if let Slot::Active(slot_info) = slot {
+            self.slots[slot_idx] = Slot::Idle(SlotInfo {
+                slot_data: slot_info.slot_data,
+                seqno: slot_info.seqno,
+            })
+        };
+
+        *locked_seat.access_mut(self) = match locked_seat.access(self) {
```

The `access()` followed immediately by `access_mut()` on the same `LockedBy` works because `access()` returns a reference that is dropped before `access_mut()` is called (due to the `match` expression borrowing temporarily). But this pattern is subtle. It effectively copies the `SeatInfo` fields. This is fine since `SeatInfo` is `Copy`-like (contains `u8` and `u64`), but making `SeatInfo` derive `Clone`/`Copy` and simplifying would be cleaner.

2. `use_seqno` starts at 1 and increments on each activation. With `u64`, overflow is practically impossible (would take ~584 years at 1 billion activations/second), so no concern there.

3. The `Deref`/`DerefMut` implementations on `SlotManager` exposing the inner `T` are convenient but somewhat unusual — they allow calling `SlotOperations` methods directly on the manager. This is intentional for the `AsSlotManager` use case where the manager needs to call `as_enable()` etc. directly.

4. `record_active_slot` returns `Result` but can never fail — it always returns `Ok(())`. Minor cleanup opportunity.

---
Generated by Claude Code Patch Reviewer