Received-SPF: pass (google.com: domain of linux-wireless-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII;
 format=flowed
Content-Transfer-Encoding: 7bit
Date:   Thu, 30 Jul 2020 14:53:40 -0700
From:   Aloka Dixit <alokad@codeaurora.org>
To:     Johannes Berg <johannes@sipsolutions.net>
Cc:     linux-wireless@vger.kernel.org,
        linux-wireless-owner@vger.kernel.org
Subject: Re: [PATCH v4 1/2] nl80211: Add FILS discovery support
In-Reply-To: <54628f052b380e660c37cb9b7c3f224976aa1c83.camel@sipsolutions.net>
References: <20200618050427.5891-1-alokad@codeaurora.org>
 <20200618050427.5891-2-alokad@codeaurora.org>
 <c3dbcc3e4ee2d3596625e8c1226325180444a961.camel@sipsolutions.net>
 <2b1ea1c4baedcb119f4e632b26399071@codeaurora.org>
 <54628f052b380e660c37cb9b7c3f224976aa1c83.camel@sipsolutions.net>
Message-ID: <ca081672e9e3407bec949c963add8530@codeaurora.org>
User-Agent: Roundcube Webmail/1.3.9
Sender: linux-wireless-owner@vger.kernel.org
Precedence: bulk

On 2020-07-30 14:22, Johannes Berg wrote:
> On Thu, 2020-07-30 at 14:17 -0700, Aloka Dixit wrote:
> 
>> > OTOH, if it's with headers, how could it be optional? In fact, either
>> > way, how is it optional?
>> >
>> 
>> Template has management frame headers as well. Will change the wording
>> accordingly.
> 
> OK.
> 
>> I made the template optional because FILS discovery may or may not be
>> offloaded to FW.
> 
> But how would anyone know? Try without it, and then try again if that
> fails? Would it fail? I mean, you also said it was required at least 
> for
> 6 GHz, so wouldn't userspace be better off always giving it - and then
> we should probably make it mandatory so it doesn't fall into the trap?
> 

If the template is not provided, FW keeps sending event to get it.
But as my ath11k driver code is limited to 6GHz, it already throws error 
if template not provided.
Yes, in general it will be better to make it mandatory, I will do it in 
next version.

> However - and here that's my ignorance speaking - can it really be
> offloaded? I mean, is everything in there completely determined by the
> beacon already, and so you have no choice in how to build it? Or how
> does that work?
> 

Yes, the frame parameters are fixed except for the timestamp which FW is 
expected to fill.

>> Yeah, I looked through existing examples for NLA_BINARY, those provide
>> only the higher bound for length.
> 
> Yeah, no way to do anything else right now. But you should have a lower
> bound in the code, I think.
> 

Okay.

>> But I can modify it to range once that is added.
> 
> Later maybe :)
> 
> johannes