[PATCH] nvme-pci: fix dma unmapping when using PRPs and not using the IOVA mapping

[Keith Busch]

On Mon, Jul 07, 2025 at 05:01:16PM +0200, Christoph Hellwig wrote:
> This looks a lot less complex than I feared.  I still don't think having
> to allocate DMA memory for storing the ranges is all that great, mostly
> because this is exactly the path we're going to hit for non-coherent
> attachment where the DMA coherent memory will have a performance impact
> because it is marked uncachable.

Okay. I was hoping to avoid bringing back a mempool, but yes, I can see
how abusing the dma pool allocation to hold the driver's sgl (that won't
be DMA'd) may be harmful on non-coherent machines. Let's go with your
patch now; we can always continue improving on this if needed later.