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/bridge: aux-hpd: let drivers pass IRQ_HPD events
Date: Thu, 23 Apr 2026 09:32:09 +1000
Message-ID: <review-patch3-20260420-hpd-irq-events-v2-3-402ffe27e9e9@oss.qualcomm.com>
In-Reply-To: <20260420-hpd-irq-events-v2-3-402ffe27e9e9@oss.qualcomm.com>
References: <20260420-hpd-irq-events-v2-0-402ffe27e9e9@oss.qualcomm.com> <20260420-hpd-irq-events-v2-3-402ffe27e9e9@oss.qualcomm.com>
X-Mailer: Claude Code Patch Reviewer
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit
MIME-Version: 1.0

Patch Review

**BUILD BUG:** In `include/drm/bridge/aux-bridge.h`, the `#else` stub for `drm_aux_hpd_bridge_notify_extra()` has a semicolon after the declaration instead of the function body:

```c
static inline void drm_aux_hpd_bridge_notify_extra(struct device *dev,
						   enum drm_connector_status status,
						   enum drm_connector_status_extra extra_status);
{
}
```

That trailing `;` after the parameter list makes this a declaration followed by an empty compound statement at file scope. This will fail to compile when `CONFIG_DRM_AUX_HPD_BRIDGE` is not enabled. Should be:

```c
static inline void drm_aux_hpd_bridge_notify_extra(struct device *dev,
						   enum drm_connector_status status,
						   enum drm_connector_status_extra extra_status)
{
}
```

**Also in the same header:** The `drm_aux_hpd_bridge_notify()` inline wrapper is placed **outside** the `#endif` for `CONFIG_DRM_AUX_HPD_BRIDGE`:

```c
#endif

static inline void drm_aux_hpd_bridge_notify(struct device *dev, enum drm_connector_status status)
{
	drm_aux_hpd_bridge_notify_extra(dev, status, DRM_CONNECTOR_NO_EXTRA_STATUS);
}

#endif
```

This means when `CONFIG_DRM_AUX_HPD_BRIDGE` **is** enabled, there will be **two** definitions of `drm_aux_hpd_bridge_notify()` -- one as a non-static exported function from `aux-hpd-bridge.c` (which wasn't removed in this patch; the `EXPORT_SYMBOL_GPL(drm_aux_hpd_bridge_notify)` was only renamed to `_extra`) and one as this `static inline`. Wait -- actually the exported symbol *was* renamed, and the original function body was renamed to `drm_aux_hpd_bridge_notify_extra()`. So the old `drm_aux_hpd_bridge_notify()` no longer exists as an exported function, and this inline wrapper replaces it. That part is fine.

But the placement is still questionable: the inline wrapper calls `drm_aux_hpd_bridge_notify_extra()`, which in the `!CONFIG_DRM_AUX_HPD_BRIDGE` case is the buggy stub (see above). And the wrapper sits between the two `#endif` guards. This works structurally (it's available in all configs), but the intent would be clearer if it were inside the `#if` block as well, with a corresponding stub in the `#else`.

The commit message says "The drm_aux_hpd_bridge_notify() is keps to ease merging" -- "keps" is a typo for "kept".

---
Generated by Claude Code Patch Reviewer