[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <5e561756-26b4-9a71-8fe2-c876e0e7d1af@hartkopp.net>
Date: Tue, 12 Nov 2019 19:24:54 +0100
From: Oliver Hartkopp <socketcan@...tkopp.net>
To: Uwe Kleine-König <u.kleine-koenig@...gutronix.de>,
Marc Kleine-Budde <mkl@...gutronix.de>
Cc: Oleksij Rempel <o.rempel@...gutronix.de>,
dev.kurt@...dijck-laurijssen.be, wg@...ndegger.com,
netdev@...r.kernel.org, kernel@...gutronix.de,
linux-can@...r.kernel.org
Subject: Re: [PATCH v1 1/9] can: af_can: export can_sock_destruct()
On 12/11/2019 12.45, Uwe Kleine-König wrote:
> Hello Marc,
>
> On Tue, Nov 12, 2019 at 12:39:27PM +0100, Marc Kleine-Budde wrote:
>> On 11/12/19 12:37 PM, Uwe Kleine-König wrote:
>>> On Tue, Nov 12, 2019 at 12:15:52PM +0100, Oleksij Rempel wrote:
>>>> +EXPORT_SYMBOL(can_sock_destruct);
>>>
>>> If the users are only expected to be another can module, it might make
>>> sense to use a namespace here?!
>>
>> How?
>
> Use
>
> EXPORT_SYMBOL_NS(can_sock_destruct, CAN)
>
> instead of the plain EXPORT_SYMBOL, and near the declaration of
> can_sock_destruct or in the source that makes use of the symbol add:
>
> MODULE_IMPORT_NS(CAN);
>
> See https://lwn.net/Articles/760045/ for some details.
Looks nice! Good idea!
But I would tend to introduce the symbol namespaces for this and the
other (existing) symbols via can-next and not within this patch set that
addresses the j1939 fixes.
Best regards,
Oliver
Powered by blists - more mailing lists