Claude review: drm/panel: Add driver for Novatek NT37705 panel

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

Patch Review

This is the main patch. It's a typical DSI panel driver with DSC support. Several issues:

**Issue 1 (build failure): Missing `DRM_DISPLAY_DSC_HELPER` in Kconfig**

```
+config DRM_PANEL_NOVATEK_NT37705
+	tristate "Novatek NT37705-based DSI panel"
+	depends on OF
+	depends on DRM_MIPI_DSI
+	depends on BACKLIGHT_CLASS_DEVICE
+	select DRM_KMS_HELPER
```

The driver includes `<drm/display/drm_dsc.h>` and `<drm/display/drm_dsc_helper.h>` and calls `drm_dsc_pps_payload_pack()`. This requires `DRM_DISPLAY_DSC_HELPER`. Without it, the build will fail with an undefined symbol for `drm_dsc_pps_payload_pack`. Compare with the NT37801 Kconfig entry which has:
```
	select DRM_DISPLAY_DSC_HELPER
	select DRM_DISPLAY_HELPER
```

This needs to be:
```
	select DRM_DISPLAY_DSC_HELPER
	select DRM_DISPLAY_HELPER
	select DRM_KMS_HELPER
```

**Issue 2 (resource leak): Incomplete error cleanup in `nt37705_prepare()`**

After PPS or compression_mode failure, the function returns without cleaning up:
```c
	ret = mipi_dsi_picture_parameter_set(ctx->dsi, &pps);
	if (ret < 0) {
		dev_err(panel->dev, "failed to transmit PPS: %d\n", ret);
		return ret;
	}

	ret = mipi_dsi_compression_mode(ctx->dsi, true);
	if (ret < 0) {
		dev_err(dev, "failed to enable compression mode: %d\n", ret);
		return ret;
	}
```

At this point regulators are enabled, the reset GPIO is deasserted, and the panel has been initialized via `nt37705_on()`. Returning without cleanup leaks the regulators and leaves the panel in an inconsistent state. The NT37801 driver handles this correctly with a `goto err` pattern that deasserts reset and disables regulators. This should follow the same pattern, using a common error label.

**Minor: Standard DCS commands sent with extra `0x00` parameter**

```c
	mipi_dsi_dcs_write_seq_multi(&dsi_ctx, 0x11, 0x00);
	mipi_dsi_msleep(&dsi_ctx, 120);
	mipi_dsi_dcs_write_seq_multi(&dsi_ctx, 0x29, 0x00);
```

`0x11` is `MIPI_DCS_EXIT_SLEEP_MODE` and `0x29` is `MIPI_DCS_SET_DISPLAY_ON` -- both are defined as zero-parameter commands in the MIPI DCS spec. The extra `0x00` byte is non-standard. Similarly in `nt37705_off()`:
```c
	mipi_dsi_dcs_write_seq_multi(&dsi_ctx, 0x28, 0x00);
	...
	mipi_dsi_dcs_write_seq_multi(&dsi_ctx, 0x10, 0x00);
```

Since this is generated from vendor device tree, the extra byte may be intentional for this IC. But ideally these should use the dedicated multi helpers (`mipi_dsi_dcs_exit_sleep_mode_multi`, `mipi_dsi_dcs_set_display_on_multi`, etc.) or at least use the named `MIPI_DCS_*` constants.

**Minor: Tear enable without named constant**

```c
	mipi_dsi_dcs_write_seq_multi(&dsi_ctx, 0x35);
```

`0x35` is `MIPI_DCS_SET_TEAR_ON`. Consider using the named constant for readability, or the dedicated helper if one exists.

**Minor: Kconfig help text has a doubled "the"**

```
+	  display panels, such as the one found in the The Fairphone (Gen. 6)
```

Should be "the Fairphone (Gen. 6)".

**Observation: TODOs in probe()**

Two TODO comments are left in the code:
```c
	/* TODO: Pass slice_per_pkt = 2 */
	...
	/*
	 * TODO: hdisplay should be read from the selected mode once
	 * it is passed back to drm_panel (in prepare?)
	 */
```

These are pre-existing framework limitations also present in NT37801, so they're acceptable for now, but worth tracking.

---
Generated by Claude Code Patch Reviewer