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: drm/sti: hdmi: remove bridge when component_add fails
Date: Sun, 12 Apr 2026 09:29:49 +1000
Message-ID: <review-patch1-20260410194141.275998-1-osama.abdelkader@gmail.com>
In-Reply-To: <20260410194141.275998-1-osama.abdelkader@gmail.com>
References: <20260410194141.275998-1-osama.abdelkader@gmail.com> <20260410194141.275998-1-osama.abdelkader@gmail.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

**The bug is real.** Before this patch, lines 1462-1463 (post-patch numberi=
ng) were:

```c
drm_bridge_add(&hdmi->bridge);
return component_add(&pdev->dev, &sti_hdmi_ops);
```

If `component_add()` fails, execution returns directly with the error code,=
 leaking both the `drm_bridge_add()` registration and the `i2c_put_adapter(=
)` reference acquired at line 1392.

**The fix is correct.** The new `remove_bridge` label is placed between the=
 `return 0` and the existing `release_adapter` label, so it properly falls =
through to also call `i2c_put_adapter()`:

```c
 remove_bridge:
	drm_bridge_remove(&hdmi->bridge);
 release_adapter:
	i2c_put_adapter(hdmi->ddc_adapt);
```

This matches the reverse order of acquisition (`drm_bridge_add` after adapt=
er, so `drm_bridge_remove` before `i2c_put_adapter`).

**Consistency with `sti_hdmi_remove()`.** The remove function at line 1476 =
calls `drm_bridge_remove()` and `i2c_put_adapter()`, confirming both resour=
ces need cleanup =E2=80=94 consistent with what this patch does on the erro=
r path.

**Minor nit =E2=80=94 commit message style.** The commit message body ("whe=
n component_add fails in sti_hdmi_probe remove the drm bridge and put i2c a=
dapter before return") is a run-on sentence. A slightly improved version:

> When component_add() fails in sti_hdmi_probe(), the DRM bridge and I2C ad=
apter reference are leaked. Add a cleanup path to call drm_bridge_remove() =
and i2c_put_adapter() on failure.

This is a style-only suggestion and shouldn't block the patch.

**Reviewed-by worthy:** Yes. The logic is sound, the error unwind order is =
correct, and the fix is minimal. This is a straightforward resource-leak fi=
x.

---
Generated by Claude Code Patch Reviewer