[LSF/MM/BPF TOPIC] [LSF/MM/BPF ATTEND] : Two stage IOMMU DMA mapping operations

Jason Gunthorpe jgg at nvidia.com
Tue Mar 5 05:03:31 PST 2024


On Tue, Feb 27, 2024 at 08:17:27AM +0000, Chaitanya Kulkarni wrote:
> Hi,
> 
> * Problem Statement :-
> -------------------------------------------------------------------------
> The existing IOMMU DMA mapping operation is performed in two steps at the
> same time (one-shot):
> 1. Allocates IOVA space.
> 2. Actually maps DMA pages to that space.
> For example, map scatter-gather list:

For clarity, this has come out of last years topic on the "physr" - we
agreed to a general direction where instead of adding a parallel DMA
API surface for a new scatterlist alternative we'd improve the DMA API
so callers can efficiently bring their own data structure.

https://lore.kernel.org/linux-rdma/Y8v+qVZ8OmodOCQ9@nvidia.com/

Jason



More information about the Linux-nvme mailing list