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: pass down IRQ_HPD to the drivers
Date: Thu, 23 Apr 2026 09:32:09 +1000
Message-ID: <review-patch2-20260420-hpd-irq-events-v2-2-402ffe27e9e9@oss.qualcomm.com>
In-Reply-To: <20260420-hpd-irq-events-v2-2-402ffe27e9e9@oss.qualcomm.com>
References: <20260420-hpd-irq-events-v2-0-402ffe27e9e9@oss.qualcomm.com> <20260420-hpd-irq-events-v2-2-402ffe27e9e9@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

This is the large mechanical patch extending all `hpd_notify`, `hpd_cb`, `oob_hotplug_event` callbacks with the extra `extra_status` parameter.

**Concern: spurious blank line removal in ti-tfp410.c:**

```c
-
 static inline struct tfp410 *
```

This hunk removes a blank line between two functions in `ti-tfp410.c` that is unrelated to the change. Should be reverted.

**Design concern in `drm_bridge_connector_handle_hpd()`:** The function now receives `extra_status` and passes it to `drm_bridge_connector_hpd_notify()`, but it also unconditionally calls `drm_kms_helper_connector_hotplug_event(connector)` afterward. For a pure IRQ_HPD event (where status hasn't actually changed), this triggers a full connector hotplug event to userspace, which causes unnecessary reprobing. The function probably should skip calling `drm_kms_helper_connector_hotplug_event()` when the status hasn't changed and only `extra_status` was set, or use a more targeted notification path. That said, the current DP drivers handle IRQ_HPD internally (e.g., MSM adds `EV_IRQ_HPD_INT`), so this may be acceptable for now if the cost of the extra userspace notification is low.

**In `drm_bridge_hpd_notify()` (drm_bridge.c):** The patch hardcodes `DRM_CONNECTOR_NO_EXTRA_STATUS`:

```c
	if (bridge->hpd_cb)
		bridge->hpd_cb(bridge->hpd_data, status, DRM_CONNECTOR_NO_EXTRA_STATUS);
```

This is fine as a temporary measure since patch 3 introduces `drm_bridge_hpd_notify_extra()` that passes the real value, and this function becomes the legacy wrapper. The inter-patch dependency is clean.

---
Generated by Claude Code Patch Reviewer