[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20201005121813.GA6617@nanopsycho>
Date: Mon, 5 Oct 2020 14:18:13 +0200
From: Jiri Pirko <jiri@...nulli.us>
To: Moshe Shemesh <moshe@...dia.com>
Cc: Moshe Shemesh <moshe@...lanox.com>,
"David S. Miller" <davem@...emloft.net>,
Jakub Kicinski <kuba@...nel.org>, Jiri Pirko <jiri@...dia.com>,
netdev@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH net-next 03/16] devlink: Add devlink reload limit option
Sun, Oct 04, 2020 at 08:42:47AM CEST, moshe@...dia.com wrote:
>
>On 10/3/2020 10:51 AM, Jiri Pirko wrote:
>> Thu, Oct 01, 2020 at 03:59:06PM CEST, moshe@...lanox.com wrote:
>>
>> [...]
>>
>> > enum devlink_attr {
>> > /* don't change the order or add anything between, this is ABI! */
>> > DEVLINK_ATTR_UNSPEC,
>> > @@ -507,6 +524,7 @@ enum devlink_attr {
>> >
>> > DEVLINK_ATTR_RELOAD_ACTION, /* u8 */
>> > DEVLINK_ATTR_RELOAD_ACTIONS_PERFORMED, /* u64 */
>> > + DEVLINK_ATTR_RELOAD_LIMIT, /* u8 */
>> Hmm, why there could be specified only single "limit"? I believe this
>> should be a bitfield. Same for the internal api to the driver.
>
>
>Why bitfield ? Either the user asks for a specific limit or he doesn't ask
>for any (unspecified).
He can ask for multiple limits: No_link_flag , no_something_else. Could
be totally unrelated limitations. Let's just have the UAPI ready for
this once we define it from scratch.
>
>If the user doesn't need limitation he will not specify a limit.
>
>> [...]
Powered by blists - more mailing lists