Claude review: arm64: dts: ti: k3-j721e: Add GPU node

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

Patch Review

**Author:** Antonios Christidis

This adds the Series 8XE (GE8430) GPU node for the J721E SoC.

**Ordering issue:** This patch uses the `"ti,j721e-gpu"` compatible string, but the dt-binding adding this compatible doesn't come until patch 2. Per kernel convention, the binding documentation should come first so that `dt_binding_check` can validate the DTS. These two patches should be reordered.

**The DTS node itself looks reasonable:**

```dts
gpu: gpu@4e20000000 {
    compatible = "ti,j721e-gpu", "img,img-ge8430", "img,img-rogue";
    reg = <0x4e 0x20000000 0x00 0x80000>;
    clocks = <&k3_clks 125 0>;
    clock-names = "core";
    assigned-clocks = <&k3_clks 125 0>;
    assigned-clock-rates = <750000000>;
    interrupts = <GIC_SPI 24 IRQ_TYPE_LEVEL_HIGH>;
    power-domains = <&k3_pds 125 TI_SCI_PD_EXCLUSIVE>,
            <&k3_pds 126 TI_SCI_PD_EXCLUSIVE>;
    power-domain-names = "a", "b";
    dma-coherent;
};
```

- The compatible string triple (`ti,j721e-gpu`, `img,img-ge8430`, `img,img-rogue`) correctly follows the fallback chain pattern.
- `power-domain-names = "a", "b"` -- these names are very generic. Is this the established pattern for TI GPU power domains? If so, fine, but more descriptive names would be preferable.
- **Missing `status = "disabled"`**: Many TI DTS main dtsi files use `status = "disabled"` for peripheral nodes, with board-level dts files enabling them. Is the GPU always expected to be enabled? If not, this should be disabled by default.

**Commit message** is terse ("Add the series 8XE GPU node for j721e device tree.") but adequate for a DTS patch.

---

---
Generated by Claude Code Patch Reviewer