mbox series

[rdma-next,0/2] Provide already supported real-time timestamp

Message ID cover.1623829775.git.leonro@nvidia.com (mailing list archive)
Headers show
Series Provide already supported real-time timestamp | expand

Message

Leon Romanovsky June 16, 2021, 7:57 a.m. UTC
From: Leon Romanovsky <leonro@nvidia.com>

In case device supports only real-time timestamp, the kernel will
fail to create QP despite rdma-core requested such timestamp type.

It is because device returns free-running timestamp, and the conversion
from free-running to real-time is performed in the user space.

This series fixes it, by returning real-time timestamp.

Thanks

Aharon Landau (2):
  RDMA/mlx5: Refactor get_ts_format functions to simplify code
  RDMA/mlx5: Support real-time timestamp directly from the device

 drivers/infiniband/hw/mlx5/cq.c               |   6 +-
 drivers/infiniband/hw/mlx5/main.c             |   6 ++
 drivers/infiniband/hw/mlx5/mlx5_ib.h          |   7 ++
 drivers/infiniband/hw/mlx5/qp.c               | 102 ++++++++++--------
 .../net/ethernet/mellanox/mlx5/core/en_main.c |   8 +-
 .../ethernet/mellanox/mlx5/core/lib/clock.h   |  10 +-
 include/linux/mlx5/mlx5_ifc.h                 |  36 ++-----
 include/linux/mlx5/qp.h                       |   4 +-
 include/uapi/rdma/mlx5-abi.h                  |   2 +
 9 files changed, 94 insertions(+), 87 deletions(-)

Comments

Jason Gunthorpe June 21, 2021, 11:37 p.m. UTC | #1
On Wed, Jun 16, 2021 at 10:57:37AM +0300, Leon Romanovsky wrote:
> From: Leon Romanovsky <leonro@nvidia.com>
> 
> In case device supports only real-time timestamp, the kernel will
> fail to create QP despite rdma-core requested such timestamp type.
> 
> It is because device returns free-running timestamp, and the conversion
> from free-running to real-time is performed in the user space.
> 
> This series fixes it, by returning real-time timestamp.
> 
> Thanks
> 
> Aharon Landau (2):
>   RDMA/mlx5: Refactor get_ts_format functions to simplify code
>   RDMA/mlx5: Support real-time timestamp directly from the device

This looks fine, can you update the shared branch please

Jason
Leon Romanovsky June 22, 2021, 6:42 a.m. UTC | #2
On Mon, Jun 21, 2021 at 08:37:34PM -0300, Jason Gunthorpe wrote:
> On Wed, Jun 16, 2021 at 10:57:37AM +0300, Leon Romanovsky wrote:
> > From: Leon Romanovsky <leonro@nvidia.com>
> > 
> > In case device supports only real-time timestamp, the kernel will
> > fail to create QP despite rdma-core requested such timestamp type.
> > 
> > It is because device returns free-running timestamp, and the conversion
> > from free-running to real-time is performed in the user space.
> > 
> > This series fixes it, by returning real-time timestamp.
> > 
> > Thanks
> > 
> > Aharon Landau (2):
> >   RDMA/mlx5: Refactor get_ts_format functions to simplify code
> >   RDMA/mlx5: Support real-time timestamp directly from the device
> 
> This looks fine, can you update the shared branch please

9a1ac95a59d0 RDMA/mlx5: Refactor get_ts_format functions to simplify code

Applied, thanks

> 
> Jason
Jason Gunthorpe June 22, 2021, 6:12 p.m. UTC | #3
On Tue, Jun 22, 2021 at 09:42:07AM +0300, Leon Romanovsky wrote:
> On Mon, Jun 21, 2021 at 08:37:34PM -0300, Jason Gunthorpe wrote:
> > On Wed, Jun 16, 2021 at 10:57:37AM +0300, Leon Romanovsky wrote:
> > > From: Leon Romanovsky <leonro@nvidia.com>
> > > 
> > > In case device supports only real-time timestamp, the kernel will
> > > fail to create QP despite rdma-core requested such timestamp type.
> > > 
> > > It is because device returns free-running timestamp, and the conversion
> > > from free-running to real-time is performed in the user space.
> > > 
> > > This series fixes it, by returning real-time timestamp.
> > > 
> > > Thanks
> > > 
> > > Aharon Landau (2):
> > >   RDMA/mlx5: Refactor get_ts_format functions to simplify code
> > >   RDMA/mlx5: Support real-time timestamp directly from the device
> > 
> > This looks fine, can you update the shared branch please
> 
> 9a1ac95a59d0 RDMA/mlx5: Refactor get_ts_format functions to simplify code
> 
> Applied, thanks

Done, thanks

Jason