Re: [for-next 09/14] {net,IB}/mlx5: Add has_tag to mlx5_flow_act

[Saeed Mahameed <saeedm@mellanox.com>]

On Mon, 2018-03-05 at 14:07 -0700, Jason Gunthorpe wrote:
> On Mon, Mar 05, 2018 at 12:46:32PM -0800, Saeed Mahameed wrote:
> > From: Matan Barak <matanb@mellanox.com>
> > 
> > The has_tag member will indicate whether a tag action was specified
> > in flow specification.
> 
> It would be good to describe in the commit message why
> 
>  flow_act.flow_tag != MLX5_FS_DEFAULT_FLOW_TAG
> 
> isn't good enough anymore.

A flow tag 0 = MLX5_FS_DEFAULT_FLOW_TAG is assumed a valid flow tag
that is currently used by RDMA driver, whereas in HW flow_tag = 0 means
that the user doesn't care about flow_tag. HW always provide flow_tag =
0 if all flow_tags requested on a specific flow are 0.

So we need a way (in the driver) to differentiate between a user really
requesting flow_tag = 0 and a user who does not care, in order to be
able to report conflicting flow tags on a specific flow.

We will add this to commit message.

> 
> > Signed-off-by: Matan Barak <matanb@mellanox.com>
> > Reviewed-by: Aviad Yehezkel <aviadye@mellanox.com>
> > Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
> >  drivers/infiniband/hw/mlx5/main.c                 | 3 ++-
> >  drivers/net/ethernet/mellanox/mlx5/core/en_tc.c   | 1 +
> >  drivers/net/ethernet/mellanox/mlx5/core/fs_core.c | 2 +-
> >  include/linux/mlx5/fs.h                           | 1 +
> >  4 files changed, 5 insertions(+), 2 deletions(-)
> 
> Assuming there is a good reason to do this:
> 
> Acked-by: Jason Gunthorpe <jgg@mellanox.com>
> 
> Jason