Claude review: drm/amd: treat max_tmds_clock==0 as unconstrained

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

Patch Review

**This is the most straightforward and correct patch in the series.** When an EDID lacks an HDMI VSDB, `max_tmds_clock` will be 0. Treating 0 as "unlimited" rather than "no clock allowed" is the right fix, and matches how other drivers handle this (e.g. the DRM core's `drm_hdmi_compute_mode_clock()` treats 0 as unconstrained).

```c
if (info->max_tmds_clock == 0 ||
    normalized_clk <= info->max_tmds_clock) {
```

This is clean and correct. However, should this also be gated on `edid_overridden`? Without the override flag check, this changes behavior for any display that genuinely reports `max_tmds_clock == 0` through its real EDID, which could potentially allow modes that exceed the display's capabilities. That said, the HDMI spec says a conformant EDID should always include a VSDB with a valid max_tmds_clock if it supports HDMI, so 0 generally means "not specified" rather than "zero limit" — making this change safe regardless of override status.

---
Generated by Claude Code Patch Reviewer