ACK: [SRU][jammy][PATCH 0/6] Backport RDMA DMABUF

Philip Cox philip.cox at canonical.com
Mon Oct 30 14:27:30 UTC 2023


On Wed, 2023-10-25 at 15:07 -0500, Ian May wrote:
> SRU Justification:
> 
> [Impact]
> 
> * From Nvidia:
> 
> "We are working on a high performance networking solution with real
> customers. That solution is being developed using the Ubuntu 22.04
> LTS distro release and the distro kernel (lowlatency flavour). This
> “dma_buf” patchset consists of upstreamed patches that allow buffers
> to be shared between drivers thus enhancing performance while
> reducing copying of data.
> 
> Our team is currently engaged in the development of a high-
> performance networking solution tailored to meet the demands of real-
> world customers. This cutting-edge solution is being crafted on the
> foundation of Ubuntu 22.04 LTS, utilizing the distribution's kernel,
> specifically the lowlatency flavor.
> 
> At the heart of our innovation lies the transformative "dma_buf"
> patchset, comprising a series of patches that have been integrated
> into the upstream kernel in 5.16 and 5.17. These patches introduce a
> groundbreaking capability: enabling the seamless sharing of buffers
> among various drivers. This not only bolsters the solution's
> performance but also minimizes the need for data copying, effectively
> enhancing efficiency across the board.
> 
> The new functionality is isolated such that existing user will not
> execute these new code paths."
> 
> Upstream Reference
> 
> https://lore.kernel.org/all/20211012120903.96933-1-galpress@amazon.com/
> https://lore.kernel.org/all/0-v1-bd147097458e+ede-umem_dmabuf_jgg@nvidia.com/
> The patch "[PATCH 1/4] net/mlx5: Add IFC bits for mkey ATS" is
> already in Jammy and was included in:
> "2a1e6097e9b9 UBUNTU: SAUCE: RDMA/core: Updated ib_peer_memory"
> 
> [Test Plan]
> 
> * Testing instructions are outlined in the SF case and has been
> tested on in house hardware and externally by Nvidia.
> 
> [Where problems could occur?]
> 
> * This introduces new code paths so regression potential should be
> low.
> 
> [Other Info]
> 
> * SF#00370664
> 
> 
> Gal Pressman (3):
>   dma-buf: Fix pin callback comment
>   RDMA/umem: Allow pinned dmabuf umem usage
>   RDMA/efa: Add support for dmabuf memory regions
> 
> Jason Gunthorpe (3):
>   RDMA/core: Add UVERBS_ATTR_RAW_FD
>   RDMA/mlx5: Add support for dmabuf to devx umem
>   RDMA/mlx5: Enable ATS support for MRs and umems
> 
>  drivers/infiniband/core/umem_dmabuf.c    |  51 +++++++++
>  drivers/infiniband/core/uverbs_ioctl.c   |   8 ++
>  drivers/infiniband/hw/efa/efa.h          |   4 +
>  drivers/infiniband/hw/efa/efa_main.c     |   1 +
>  drivers/infiniband/hw/efa/efa_verbs.c    | 127 +++++++++++++++++----
> --
>  drivers/infiniband/hw/mlx5/devx.c        |  54 +++++++---
>  drivers/infiniband/hw/mlx5/mlx5_ib.h     |  36 +++++++
>  drivers/infiniband/hw/mlx5/mr.c          |   5 +-
>  include/linux/dma-buf.h                  |   4 +-
>  include/rdma/ib_umem.h                   |  11 ++
>  include/rdma/uverbs_ioctl.h              |  13 +++
>  include/uapi/rdma/mlx5_user_ioctl_cmds.h |   1 +
>  12 files changed, 265 insertions(+), 50 deletions(-)
> 
> -- 
> 2.25.1
> 
> 

-- 
Acked-by: Philip Cox <philip.cox at canonical.com>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/kernel-team/attachments/20231030/cc3b1c3a/attachment.html>


More information about the kernel-team mailing list