Claude review: drm/pagemap: Add helper to access zone_device_data

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

Patch Review

The helper `drm_pagemap_page_zone_device_data()` wraps the page→folio conversion and calls `folio_zone_device_data()`. This is good for consistency, but there is a concern:

`folio_zone_device_data()` in the kernel (memremap.h) has:
```c
static inline void *folio_zone_device_data(const struct folio *folio)
{
	VM_WARN_ON_FOLIO(!folio_is_device_private(folio), folio);
	return folio->page.zone_device_data;
}
```

Several call sites in `drm_pagemap.c` access `zone_device_data` on pages that could be **device-coherent**, not just device-private:

- `drm_pagemap_migrate_unmap_pages` (guards with `is_zone_device_page()`, which includes coherent pages)
- `drm_pagemap_migrate_populate_ram_pfn` (compares `zone_device_data` of src/fault pages without checking device_private)
- `__drm_pagemap_migrate_to_ram`, `drm_pagemap_migrate_to_ram`, `drm_pagemap_page_to_dpagemap` — the fault/source page could be from coherent memory since `MIGRATE_VMA_SELECT_DEVICE_COHERENT` is used

Under `CONFIG_DEBUG_VM`, these paths would produce `VM_WARN_ON_FOLIO` splats for device-coherent pages. This isn't a functional bug (the data is still read correctly), but it would be noisy in debug/development builds.

**Suggestion**: Either adjust `folio_zone_device_data()` upstream to accept both private and coherent pages, or have the drm_pagemap helper access `folio->page.zone_device_data` directly (bypassing the debug check). Alternatively, if drm_pagemap is guaranteed to never see device-coherent pages at these call sites, a comment explaining that would be helpful.

**Verdict: Potential issue with device-coherent pages under CONFIG_DEBUG_VM.**

---

---
Generated by Claude Code Patch Reviewer