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/rockchip: cdn-dp: Add multiple bridges to support PHY port selection
Date: Thu, 05 Mar 2026 13:38:01 +1000
Message-ID: <review-patch7-20260304094152.92-8-kernel@airkyi.com>
In-Reply-To: <20260304094152.92-8-kernel@airkyi.com>
References: <20260304094152.92-1-kernel@airkyi.com> <20260304094152.92-8-kernel@airkyi.com>
X-Mailer: Claude Code Patch Reviewer
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit
MIME-Version: 1.0

Patch Review

The most complex patch. Introduces `cdn_dp_bridge` wrapper, multi-bridge registration, port switching, HPD notification.

**Bug: `prev_port` is used uninitialized in `cdn_dp_bridge_edid_read()`:**

```c
struct cdn_dp_port *prev_port;
...
if (dp->bridge_count > 1 && !port->phy_enabled) {
    for (i = 0; i < dp->bridge_count; i++) {
        if (dp->bridge_list[i] != dp_bridge && dp->bridge_list[i]->enabled)
            goto get_cache;
        if (dp->port[i]->phy_enabled)
            prev_port = dp->port[i];
    }

    if (prev_port) {  // <--- prev_port may be uninitialized!
```

If no port has `phy_enabled` set, `prev_port` is never assigned but is still checked. The compiler may or may not zero-initialize it. This must be initialized to `NULL`.

**Issue (medium): `cdn_dp_connected_port()` has a logic issue with the extcon (non-next_bridge) path:**

```c
for (i = 0; i < dp->ports; i++) {
    port = dp->port[i];
    lanes[i] = cdn_dp_get_port_lanes(port);
    if (!dp->next_bridge_valid)
        return port;
}
```

In the extcon path (`!next_bridge_valid`), this returns the first port unconditionally regardless of whether it has lanes. The old code was:
```c
lanes = cdn_dp_get_port_lanes(port);
if (lanes)
    return port;
```

The new code removed the `lanes` check for the extcon path, so it returns the first port even if it has 0 lanes. This is a regression.

**Issue (minor): `dp->active` check in `cdn_dp_switch_port()` is confusing:**

```c
static int cdn_dp_switch_port(...) {
    if (dp->active)
        return 0;
```

Returning 0 (success) when `dp->active` is true seems wrong for a "switch port" operation. The caller expects to be able to switch, but if the DP is actively streaming, this silently does nothing. A return of `-EBUSY` or at least a log message would be more appropriate.

**Issue (minor): In the HPD notify path, `bridge_list[i]->connected` is set but `dp->connected` is the main connection state.** The interaction between per-bridge `connected` and the global `dp->connected` is subtle and deserves a comment.

---
Generated by Claude Code Patch Reviewer