Claude review: drm/bridge: ite-it66121: Set TX mode in the .atomic_enable callback

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

Patch Review

**Issue: Duplicate register writes not removed from `.mode_set`**

The patch adds a new `it66121_set_tx_mode()` called from `it66121_bridge_enable()`, but the existing code in `it66121_bridge_mode_set()` still writes the same registers:

```c
/* Enable AVI infoframe */
if (regmap_write(ctx->regmap, IT66121_AVI_INFO_PKT_REG,
		 IT66121_AVI_INFO_PKT_ON | IT66121_AVI_INFO_PKT_RPT))
	goto unlock;

/* Set TX mode to HDMI */
if (regmap_write(ctx->regmap, IT66121_HDMI_MODE_REG, IT66121_HDMI_MODE_HDMI))
	goto unlock;
```

These lines at the current tree's lines 796-803 are not removed. The diffstat shows `1 file changed, 18 insertions(+)` with zero deletions, confirming the old code remains. The commit message says "it is more correct to do this in .atomic_enable" but doesn't explain why the old writes are kept. These should be removed from `.mode_set` to avoid writing the same registers twice per mode change, and to avoid confusion about which callback is the authoritative one.

**Minor: Error handling style**

```c
if (regmap_write(ctx->regmap, IT66121_AVI_INFO_PKT_REG,
		 IT66121_AVI_INFO_PKT_ON | IT66121_AVI_INFO_PKT_RPT))
	goto unlock;

regmap_write(ctx->regmap, IT66121_HDMI_MODE_REG, IT66121_HDMI_MODE_HDMI);
```

The first `regmap_write` checks for errors and jumps to `unlock`, but the second `regmap_write` result is silently discarded. This is inconsistent — either both should check for errors or neither should (the existing `.mode_set` code checks both). This is a minor nit since the function returns void and there's nothing to do on error except release the lock, but consistency is preferred.

**Ordering observation**: The call to `it66121_set_tx_mode()` is placed *after* `it66121_set_mute(ctx, false)` in `.atomic_enable`. This means the display is unmuted before the TX mode is configured. On a DVI monitor receiving HDMI-mode infoframes during this window, there could be a brief glitch. Consider placing `it66121_set_tx_mode()` before the unmute. This ordering also matters because in the existing `.mode_set`, the TX mode/infoframe setup is done *before* the AFE configuration and *before* unmute — moving it after unmute changes the sequencing.

---
Generated by Claude Code Patch Reviewer