Claude review: drm/panel: Add driver for Novatek NT35532

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

Patch Review

This is the core panel driver, auto-generated from vendor device tree, with 779 lines of register init sequences.

**Issue (moderate): Sleep/display-on commands sent via `mipi_dsi_generic_write_seq_multi` instead of DCS helpers**

In `nt35532_on()`:
```c
mipi_dsi_generic_write_seq_multi(&dsi_ctx, MIPI_DCS_EXIT_SLEEP_MODE, 0x00);
mipi_dsi_msleep(&dsi_ctx, 120);
mipi_dsi_generic_write_seq_multi(&dsi_ctx, MIPI_DCS_SET_DISPLAY_ON, 0x00);
```

Using `mipi_dsi_generic_write_seq_multi` with DCS command bytes sends these as **generic write packets** rather than **DCS write packets**. The correct approach is to use the dedicated `_multi` DCS helpers, as is already done in `nt35532_off()`:
```c
mipi_dsi_dcs_set_display_off_multi(&dsi_ctx);
mipi_dsi_dcs_enter_sleep_mode_multi(&dsi_ctx);
```

The `nt35532_on()` function should similarly use:
```c
mipi_dsi_dcs_exit_sleep_mode_multi(&dsi_ctx);
mipi_dsi_msleep(&dsi_ctx, 120);
mipi_dsi_dcs_set_display_on_multi(&dsi_ctx);
```

Additionally, the trailing `0x00` parameter in these calls is suspicious — DCS exit_sleep_mode and set_display_on are zero-parameter commands. The generic_write sends `{0x11, 0x00}` as a 2-byte payload, which is not standard. Some panels may tolerate it but it's technically incorrect.

**Minor: `select DRM_KMS_HELPER` in Kconfig may be unnecessary**
```
select DRM_KMS_HELPER
```
The driver uses `drm_connector_helper_get_modes_fixed()` which comes from `drm_probe_helper.h`. Check whether this is already pulled in transitively via `DRM_MIPI_DSI`. Many panel drivers don't explicitly select `DRM_KMS_HELPER`.

**Style: `container_of_const` usage**
```c
return container_of_const(panel, struct novatek_nt35532, panel);
```
The function returns a non-const `struct novatek_nt35532 *`, but `container_of_const` is designed to propagate const-ness. Since all callers need a mutable pointer, `container_of` would be more conventional here. That said, some newer panel drivers do use `container_of_const` — this is a minor style point.

**No issues with**: regulator bulk handling, reset sequence, probe/remove lifecycle, use of `devm_drm_panel_alloc`, mode definition. The driver is clean for an auto-generated panel driver.

---
Generated by Claude Code Patch Reviewer