Claude review: drm/vmwgfx: Check vrefresh in drm_mode_setcrtc.

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

Patch Review

**Purpose:** Add `mode_valid` callback to `drm_mode_config_funcs` to reject modes with zero vrefresh during `drm_mode_setcrtc`, and add vrefresh checks to STDU connector mode validation. Fixes IGT test `kms_invalid_mode@overflow-vrefresh`.

**Issues:**

1. **Config `mode_valid` checks `max_primary_mem` but connector checks different limits:** The new `vmw_kms_config_mode_valid()` validates against `dev_priv->max_primary_mem`, while the STDU connector's `vmw_stdu_connector_mode_valid()` also checks `max_mob_pages` and `max_mob_size` plus STDU-specific width/height limits. This is correct — the config-level check is necessarily more permissive since it's display-unit agnostic — but it's worth noting that the config-level check could pass modes that the connector level rejects.

2. **`GPU_TILE_SIZE` moved to header — good refactoring:** Moving the `GPU_TILE_SIZE` define from `vmwgfx_stdu.c` to `vmwgfx_kms.h` is the right thing since it's now used in both files.

3. **Missing Fixes tag:** Unlike patches 1 and 4, this patch has no `Fixes:` tag. The cover letter mentions adding Fixes messages, but this patch doesn't have one. If this is a regression fix, it should have one.

4. **Overall approach looks correct:** Using `drm_mode_config_funcs.mode_valid` is the correct hook for validating modes during `drm_mode_setcrtc` for all display types (SOU, LDU, STDU). The vrefresh check with `drm_mode_vrefresh()` returning 0 is a sensible rejection criterion.

---
Generated by Claude Code Patch Reviewer