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: faux: Allow retrieving a bound Device
Date: Fri, 05 Jun 2026 06:03:38 +1000
Message-ID: <review-patch3-20260604192740.659240-4-lyude@redhat.com>
In-Reply-To: <20260604192740.659240-4-lyude@redhat.com>
References: <20260604192740.659240-1-lyude@redhat.com> <20260604192740.659240-4-lyude@redhat.com>
X-Mailer: Claude Code Patch Reviewer
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0

Patch Review

**Correct and well-justified change.** The `AsRef<device::Device>` =E2=86=
=92 `AsRef<device::Device<device::Bound>>` change strengthens the type-leve=
l guarantee.

**Safety comment is thorough:**
```rust
// - faux_match() always returns 1, and probe runs synchronously (PROBE_FOR=
CE_SYNCHRONOUS).
// - suppress_bind_attrs =3D true on faux_driver prevents userspace-trigger=
ed unbind via sysfs
// - mem::forget(Registration) is not a problem; if the Registration is lea=
ked, the faux
//   device stays bound forever.
```
This is exactly the level of reasoning needed. The three points cover the k=
ey attack vectors against the `Bound` invariant.

**Potential downstream breakage:** Any existing code using `faux::Registrat=
ion::as_ref()` where a `&device::Device` (without the `Bound` parameter) wa=
s expected will break. Since this is on drm-rust-next and the faux bindings=
 are relatively new, this should be fine, but the cover letter doesn't ment=
ion this API change explicitly.

---
Generated by Claude Code Patch Reviewer