NAK: [PATCH 0/2][SRU I/H] UBUNTU: SAUCE: RDMA/core: Updated ib_peer_memory

Tim Gardner tim.gardner at canonical.com
Mon Oct 18 12:02:18 UTC 2021


No upstream review whatsoever ? That is pretty weak provenance. We 
generally require SRU patches to have at least made it into linux-next.

The patches to infiniband core affect not only mlx5, but more then 10 
other drivers. What testing has been performed for those ? (I suspect none).

This also looks like it should be split into 3 patches with a bit more 
explanation for each.

rtg

On 10/14/21 4:44 PM, dann frazier wrote:
> BugLink: https://launchpad.net/bugs/1947206
> 
> Nvidia have discovered some issues with the PeerDirect feature we are
> carrying in impish and hirsute. They've provided a patch for each tree
> to address them, which has been validated in their internal testing.
> I've also verified there are no issues w/ our regression test for
> this feature (attached to the above bug).
> 
> Jason Gunthorpe (1):
>    [SRU Impish] UBUNTU: SAUCE: RDMA/core: Updated ib_peer_memory
> 
>   drivers/infiniband/core/ib_peer_mem.h |   8 +-
>   drivers/infiniband/core/peer_mem.c    | 211 +++++++++++++++++++++-----
>   drivers/infiniband/hw/mlx5/devx.c     |   2 +
>   drivers/infiniband/hw/mlx5/mr.c       |  47 ++++--
>   include/linux/mlx5/mlx5_ifc.h         |  11 +-
>   include/rdma/ib_umem.h                |   4 +
>   include/rdma/peer_mem.h               |  10 ++
>   7 files changed, 235 insertions(+), 58 deletions(-)
> 
> Jason Gunthorpe (1):
>    [SRU Hirsute] UBUNTU: SAUCE: RDMA/core: Updated ib_peer_memory
> 
>   drivers/infiniband/core/ib_peer_mem.h |    8 +
>   drivers/infiniband/core/peer_mem.c    |  211 +++++++++++++++++++++++++++-------
>   drivers/infiniband/hw/mlx5/devx.c     |    2
>   drivers/infiniband/hw/mlx5/mr.c       |   37 ++++-
>   include/linux/mlx5/mlx5_ifc.h         |   11 +
>   include/rdma/ib_umem.h                |    5
>   include/rdma/peer_mem.h               |   10 +
>   7 files changed, 231 insertions(+), 53 deletions(-)
> 

-- 
-----------
Tim Gardner
Canonical, Inc



More information about the kernel-team mailing list