[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <87il9y63ih.fsf@intel.com>
Date: Tue, 01 Aug 2023 11:06:46 -0700
From: Vinicius Costa Gomes <vinicius.gomes@...el.com>
To: Vladimir Oltean <vladimir.oltean@....com>
Cc: netdev@...r.kernel.org, "David S. Miller" <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>,
Jakub Kicinski <kuba@...nel.org>,
Paolo Abeni <pabeni@...hat.com>,
Jamal Hadi Salim <jhs@...atatu.com>,
Cong Wang <xiyou.wangcong@...il.com>,
Jiri Pirko <jiri@...nulli.us>, linux-kernel@...r.kernel.org,
intel-wired-lan@...ts.osuosl.org,
Muhammad Husaini Zulkifli <muhammad.husaini.zulkifli@...el.com>,
Peilin Ye <yepeilin.cs@...il.com>,
Pedro Tammela <pctammela@...atatu.com>,
Richard Cochran <richardcochran@...il.com>,
Zhengchao Shao <shaozhengchao@...wei.com>,
Maxim Georgiev <glipus@...il.com>
Subject: Re: [PATCH v2 net-next 7/9] net: netdevsim: mimic tc-taprio offload
Vladimir Oltean <vladimir.oltean@....com> writes:
> On Tue, Aug 01, 2023 at 10:39:23AM -0700, Vinicius Costa Gomes wrote:
>> Hi Vladimir,
>>
>> Vladimir Oltean <vladimir.oltean@....com> writes:
>>
>> > On Wed, Jun 14, 2023 at 05:06:24PM -0700, Vinicius Costa Gomes wrote:
>> >> > +static int nsim_setup_tc_taprio(struct net_device *dev,
>> >> > + struct tc_taprio_qopt_offload *offload)
>> >> > +{
>> >> > + int err = 0;
>> >> > +
>> >> > + switch (offload->cmd) {
>> >> > + case TAPRIO_CMD_REPLACE:
>> >> > + case TAPRIO_CMD_DESTROY:
>> >> > + break;
>> >>
>> >> I was thinking about how useful would proper validation of the
>> >> parameters be? Thinking that we could detect "driver API" breakages
>> >> earlier, and we want it documented that the drivers should check for the
>> >> things that it supports.
>> >>
>> >> Makes sense?
>> >
>> > Sorry, I lack imagination as to what the netdevsim driver may check for.
>> > The taprio offload parameters should always be valid, properly speaking,
>> > otherwise the Qdisc wouldn't be passing them on to the driver. At least
>> > that would be the intention. The rest are hardware specific checks for
>> > hardware specific limitations. Here there is no hardware.
>> >
>>
>> Trying to remember what was going through my mind when I said that.
>>
>> What I seem to recall is something that would help us "keep honest":
>> I was worrying about someone (perhaps myself ;-) sneaking a new feature
>> in taprio and forgetting to update other drivers.
>>
>> I thought that adding a check for the existing parameters would help
>> detect those kind of things. If anything unknown was there in the
>> offload struct, netdevsim would complain loudly.
>>
>> Perhaps I was worrying too much. And the way to solve that is to keep
>> active attention against that during review.
>
> Ok, so I'm not making any change to the patch set as a result of this
> comment, would you agree?
Agreed.
Cheers,
--
Vinicius
Powered by blists - more mailing lists