Claude review: drm/exynos/gem: Fix printk specifier for gem_object.size and dma_addr

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

Patch Review

**Verdict: Looks good.**

Two changes are made in two call sites (`exynos_drm_alloc_buf` and `exynos_drm_free_buf`):

1. **`size` format: `%lx` → `%zx`** — Correct. `drm_gem_object.size` is `size_t` (confirmed in `include/drm/drm_gem.h:341`), and `%zx` is the proper format specifier for `size_t`. This eliminates the warning on architectures where `size_t != unsigned long` (e.g., 32-bit ARM).

2. **`dma_addr` format: `%lx` with cast → `%pad` by reference** — Correct and an improvement. Per `Documentation/core-api/printk-formats.rst:257`, `%pad` is the canonical kernel format for `dma_addr_t`, and it is passed by reference (hence the `&exynos_gem->dma_addr`). This removes the lossy `(unsigned long)` cast that would truncate the address on 32-bit platforms with 64-bit DMA addresses (e.g., `CONFIG_ARCH_DMA_ADDR_T_64BIT`).

```diff
-	DRM_DEV_DEBUG_KMS(drm_dev_dma_dev(dev), "dma_addr(0x%lx), size(0x%lx)\n",
-			(unsigned long)exynos_gem->dma_addr, exynos_gem->base.size);
+	DRM_DEV_DEBUG_KMS(drm_dev_dma_dev(dev), "dma_addr(%pad), size(0x%zx)\n",
+			  &exynos_gem->dma_addr, exynos_gem->base.size);
```

Minor note: `%pad` already prints the `0x` prefix (the kernel documentation shows `0x01234567`), so the output format changes slightly from `dma_addr(0x...)` to `dma_addr(0x...)` — this is fine and arguably more correct since it avoids a double `0x` if one was ever added.

The alignment fix (extra space to align under `drm_dev_dma_dev`) is also a nice touch.

**Reviewed-by worthy.** The commit message is clear, references the Fixes tag and the build robot report, and the code change is mechanical and correct.

---
Generated by Claude Code Patch Reviewer