From mboxrd@z Thu Jan  1 00:00:00 1970
From: Claude Code Review Bot <claude-review@example.com>
To: dri-devel-reviews@example.com
Subject: Claude review: gpu: nova-core: mm: Add common memory management types
Date: Wed, 01 Apr 2026 07:33:01 +1000
Message-ID: <review-patch8-20260311004008.2208806-9-joelagnelf@nvidia.com>
In-Reply-To: <20260311004008.2208806-9-joelagnelf@nvidia.com>
References: <20260311004008.2208806-1-joelagnelf@nvidia.com> <20260311004008.2208806-9-joelagnelf@nvidia.com>
X-Mailer: Claude Code Patch Reviewer
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit
MIME-Version: 1.0

Patch Review

Clean ownership model. `GpuMm` owns buddy allocator, PRAMIN, and TLB. The initialization in `gpu.rs` properly sets up the PRAMIN region to cover all physical VRAM (including GSP-reserved areas).

**Observation:** In the `Gpu::new` initializer:
```rust
GpuMm::new(devres_bar.clone(), GpuBuddyParams {
    base_offset: usable_vram.start,
    size: usable_vram.end - usable_vram.start,
    chunk_size: Alignment::new::<SZ_4K>(),
}, pramin_vram_region)?
```
The buddy allocator is initialized with the usable VRAM region, while PRAMIN covers all VRAM. This is correct -- you need PRAMIN access to GSP-reserved areas (like the BAR1 PDB) but shouldn't allocate from them.

No issues.

---
Generated by Claude Code Patch Reviewer