Claude review: drm: renesas: rz-du: Add support for RZ/T2H SoC

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

Patch Review

**Status: Good**

Adds the actual RZ/T2H support. Key additions:

1. **Feature flag infrastructure** — `RZG2L_DU_FEATURE_DPIO_OE` and the `rzg2l_du_has()` helper. The helper is clean:
```c
static inline bool rzg2l_du_has(struct rzg2l_du_device *rcdu,
                unsigned int feature)
{
    return rcdu->info->features & feature;
}
```

2. **DPI output enable** — In `rzg2l_du_start_stop()`, the RZ/T2H variant additionally asserts `DU_MCR0_DPI_EN` (BIT(0)):
```c
u32 val = DU_MCR0_DI_EN;

if (start && rzg2l_du_has(rcdu, RZG2L_DU_FEATURE_DPIO_OE))
    val |= DU_MCR0_DPI_EN;

writel(start ? val : 0, rcdu->mmio + DU_MCR0);
```

This correctly only sets `DPI_EN` when starting, and writes 0 when stopping (clearing both bits).

3. **Device info** for the new SoC with clock limits 5000–100000 kHz (wider than RZ/G2UL/G2L).

The `features` field is added to the struct doc comment in `rzg2l_du_drv.h`, which is good.

One thing to note: the patch series does not add the `r9a09g087` (RZ/N2H) compatible to the `of_device_id` table in the driver. This is fine since the DT binding specifies `renesas,r9a09g077-du` as the fallback for RZ/N2H, so the kernel will match on the RZ/T2H compatible. This is the standard Renesas pattern for compatible SoC variants.

**Overall verdict: The series looks ready to merge.** It's well-reviewed, follows established patterns, and the changes are correct and minimal.

---
Generated by Claude Code Patch Reviewer