Claude review: drm/bridge: synopsys: dw-dp: Add audio support

[Claude Code Review Bot <claude-review@example.com>]

Patch Review

The main audio implementation. Uses the `drm_bridge_funcs` DP audio callbacks.

**Issue — `AUDIO_DATA_IN_EN_CHANNEL*` bit definitions overlap with `AUDIO_INF_SELECT`:**
```c
 #define AUDIO_DATA_IN_EN			GENMASK(4, 1)
+#define AUDIO_DATA_IN_EN_CHANNEL12		BIT(0)
+#define AUDIO_DATA_IN_EN_CHANNEL34		BIT(1)
+#define AUDIO_DATA_IN_EN_CHANNEL56		BIT(2)
+#define AUDIO_DATA_IN_EN_CHANNEL78		BIT(3)
 #define AUDIO_INF_SELECT			BIT(0)
```
These `CHANNEL*` defines are *field values* used inside `FIELD_PREP(AUDIO_DATA_IN_EN, ...)`, not raw register bits. This is correct usage, but the naming convention (`BIT(0)` etc. placed right next to raw register bit defines) is confusing. A comment or renaming (e.g., `DW_DP_AUD_CH12_EN = BIT(0)`) would improve readability.

**Issue — both clocks enabled then one disabled:**
```c
+	clk_prepare_enable(dp->spdif_clk);
+	clk_prepare_enable(dp->i2s_clk);
+	/* ... configure registers ... */
+	if (dp->audio_interface == DW_DP_AUDIO_I2S)
+		clk_disable_unprepare(dp->spdif_clk);
+	else if (dp->audio_interface == DW_DP_AUDIO_SPDIF)
+		clk_disable_unprepare(dp->i2s_clk);
```
Both clocks are enabled, registers are configured, then the unused one is disabled. The comment says "Wait for inf switch" with a `usleep_range(20, 40)`. No error handling on `clk_prepare_enable` calls here either.

**Issue — channel count support is limited:**
```c
+	switch (params->cea.channels) {
+	case 1:
+	case 2:
+		audio_data_in_en = AUDIO_DATA_IN_EN_CHANNEL12;
+		break;
+	case 8:
+		audio_data_in_en = ...all channels...;
+		break;
+	default:
+		dev_err(dp->dev, "invalid audio channels %d\n", ...);
+		return -EINVAL;
+	}
```
Only 1/2 and 8 channels are supported. 4 and 6 channel configurations (common for surround sound) are rejected. This is likely intentional for initial implementation but worth noting.

**Audio startup/shutdown PM:** `pm_runtime_get_sync` in `startup` and `pm_runtime_put_autosuspend` in `shutdown` — same lack of error handling as noted in patch 9.

**The `hdmi_audio_dai_port = 1` assignment:** This assumes port@1 is the audio DAI port, which aligns with the DT binding change in patch 11 that defines `#sound-dai-cells = <1>`.

Overall the audio implementation follows established patterns from other DRM bridge audio drivers and looks functionally correct for SST use cases.

---
Generated by Claude Code Patch Reviewer