From mboxrd@z Thu Jan  1 00:00:00 1970
From: Claude Code Review Bot <claude-review@example.com>
To: dri-devel-reviews@example.com
Subject: Claude review: rust: auxiliary: add registration_data_with() for ForLt types
Date: Thu, 04 Jun 2026 12:08:39 +1000
Message-ID: <review-patch3-20260603011020.2073650-4-dakr@kernel.org>
In-Reply-To: <20260603011020.2073650-4-dakr@kernel.org>
References: <20260603011020.2073650-1-dakr@kernel.org> <20260603011020.2073650-4-dakr@kernel.org>
X-Mailer: Claude Code Patch Reviewer
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0

Patch Review

The private helper extraction is clean:

```rust
unsafe fn registration_data_pinned<F: ForLt + 'static>(&self) -> Result<Pin=
<&F::Of<'static>>>
```

The `registration_data_with` closure-based API correctly uses an HRTB closu=
re:

```rust
pub fn registration_data_with<F: ForLt + 'static, R>(
    &self,
    f: impl for<'a> FnOnce(Pin<&'a F::Of<'a>>) -> R,
) -> Result<R>
```

The soundness argument is correct: the `for<'a>` quantification prevents th=
e closure from capturing external references with a concrete short lifetime=
 and smuggling them into the invariant type. The round-trip (`'bound` =E2=
=86=92 `'static` =E2=86=92 `'a`) is safe because `'a` is universally quanti=
fied.

Relaxing `Registration`'s bound from `CovariantForLt` to `ForLt` is the rig=
ht move =E2=80=94 it lets invariant types be registered while `registration=
_data()` (requiring `CovariantForLt`) remains available as a convenience.

**Minor observation:** The three separate `unsafe` blocks in `registration_=
data_with` could arguably share a single safety comment since they're all j=
ustified by the same HRTB argument. The current form with individual commen=
ts is verbose but not wrong.

---
Generated by Claude Code Patch Reviewer