[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAMZ6RqLvbp8EStaSRFQUimhUMpn75=3pkQZYspnP1gYRsspv-g@mail.gmail.com>
Date: Sat, 19 Aug 2023 22:10:36 +0900
From: Vincent Mailhol <vincent.mailhol@...il.com>
To: Jakub Kicinski <kuba@...nel.org>
Cc: Martin Hundebøll <martin@...nix.com>,
Wolfgang Grandegger <wg@...ndegger.com>,
Marc Kleine-Budde <mkl@...gutronix.de>,
"David S . Miller" <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>,
Paolo Abeni <pabeni@...hat.com>,
Chandrasekar Ramakrishnan <rcsekar@...sung.com>,
linux-can <linux-can@...r.kernel.org>,
netdev <netdev@...r.kernel.org>,
open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 1/2] can: netlink: support setting hardware filters
On Sat. 19 Aug. 2023, 01:19, Jakub Kicinski <kuba@...nel.org> wrote:
>
> On Thu, 17 Aug 2023 12:10:13 +0200 Martin Hundebøll wrote:
> > + int len = nla_len(data[IFLA_CAN_HW_FILTER]);
> > + int num_filter = len / sizeof(struct can_filter);
> > + struct can_filter *filter = nla_data(data[IFLA_CAN_HW_FILTER]);
>
> This will prevent you from ever extending struct can_filter in
> a backward-compatible fashion, right? I obviously know very little
> about CAN but are you confident a more bespoke API to manipulate
> filters individually and allow extensibility is not warranted?
I follow Jakub's point of view.
The current struct can_filter is not sound. Some devices such as the
ES582.1 supports filtering of the CAN frame based on the flags (i.e.
SFF/EFF, RTR, FDF).
I think that each of the fields of the filter should have its own NLA
declaration with the whole thing wrapped within a NLA_NESTED_ARRAY.
I also think that there should then be a method to report the precise
filtering capabilities of the hardware.
Yours sincerely,
Vincent Mailhol
Powered by blists - more mailing lists