[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1521158282.3219.53.camel@mellanox.com>
Date: Thu, 15 Mar 2018 23:58:07 +0000
From: Saeed Mahameed <saeedm@...lanox.com>
To: Matan Barak <matanb@...lanox.com>,
Maor Gottlieb <maorg@...lanox.com>,
"leon@...nel.org" <leon@...nel.org>,
"willy@...radead.org" <willy@...radead.org>
CC: "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
"linux-rdma@...r.kernel.org" <linux-rdma@...r.kernel.org>
Subject: Re: [PATCH] mlx5: Remove call to ida_pre_get
On Wed, 2018-03-14 at 19:57 -0700, Matthew Wilcox wrote:
> From: Matthew Wilcox <mawilcox@...rosoft.com>
>
> The mlx5 driver calls ida_pre_get() in a loop for no readily apparent
> reason. The driver uses ida_simple_get() which will call
> ida_pre_get()
> by itself and there's no need to use ida_pre_get() unless using
> ida_get_new().
>
Hi Matthew,
Is this is causing any issues ? or just a simple cleanup ?
Adding Maor, the author of this change,
I believe the idea is to speed up insert_fte (which calls
ida_simple_get) since insert_fte runs under the FTE write semaphore,
in this case if ida_pre_get was successful before taking the semaphore
for all the FTE nodes in the loop, this will be a huge win for
ida_simple_get which will immediately return success without even
trying to allocate.
so it is a best effort to speed up critical path.
Maor, if this is really the case and this is not causing any issues,
then we need to consider adding a comment.
> Signed-off-by: Matthew Wilcox <mawilcox@...rosoft.com>
>
> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
> b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
> index 10e16381f20a..3ba07c7096ef 100644
> --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
> +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
> @@ -1647,7 +1647,6 @@ try_add_to_existing_fg(struct mlx5_flow_table
> *ft,
>
> list_for_each_entry(iter, match_head, list) {
> nested_down_read_ref_node(&iter->g->node,
> FS_LOCK_PARENT);
> - ida_pre_get(&iter->g->fte_allocator, GFP_KERNEL);
> }
>
> search_again_locked:
>
Powered by blists - more mailing lists