[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <066d5ac8-e95a-f4ec-fda7-31716242cc05@mellanox.com>
Date: Wed, 20 Dec 2017 21:21:51 +0200
From: Mark Bloch <markb@...lanox.com>
To: David Miller <davem@...emloft.net>, saeedm@...lanox.com
Cc: dledford@...hat.com, netdev@...r.kernel.org,
linux-rdma@...r.kernel.org, leonro@...lanox.com
Subject: Re: [for-next 01/11] net/mlx5: E-Switch, Refactor vport representors
initialization
On 20/12/2017 19:57, David Miller wrote:
> From: Saeed Mahameed <saeedm@...lanox.com>
> Date: Tue, 19 Dec 2017 12:33:30 -0800
>
>> +int esw_offloads_init_reps(struct mlx5_eswitch *esw)
>> +{
>> + struct mlx5_core_dev *dev = esw->dev;
>> + struct mlx5_esw_offload *offloads;
>> + struct mlx5_eswitch_rep *rep;
>> + int total_vfs = MLX5_TOTAL_VPORTS(dev);
>> + u8 hw_id[ETH_ALEN];
>> + int vport;
>
> Reverse christmas-tree please.
we need dev for MLX5_TOTAL_VPORTS, so we can't have proper reverse christms-tree,
I'll just remove the initialization of total_vfs and do it after the declarations,
and switch between total_vfs and hw_id.
>
>> + esw->offloads.vport_reps =
>> + kcalloc(total_vfs, sizeof(struct mlx5_eswitch_rep),
>> + GFP_KERNEL);
>
> That looks really unpleasant:
>
> x = kcalloc(y,
> z, GFP_KERNEL);
>
> would look so much nicer.
>
No problem.
Thanks,
Mark.
Powered by blists - more mailing lists