Received-SPF: pass (google.com: domain of linux-wireless-owner@vger.kernel.org designates 23.128.96.32 as permitted sender) client-ip=23.128.96.32;
Message-ID: <b6cb54b5-4c37-8c88-beb3-5fc14bb1aaee@candelatech.com>
Date:   Thu, 5 Oct 2023 16:38:58 -0700
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.10.0
Subject: Re: [PATCH] mac80211: Allow STA to connect in AX mode to MLD AP.
Content-Language: en-US
From:   Ben Greear <greearb@candelatech.com>
To:     Johannes Berg <johannes@sipsolutions.net>,
        linux-wireless@vger.kernel.org
References: <20231003164326.857433-1-greearb@candelatech.com>
 <4ae5a8dbe80e0eb0b96e9165892746fc3b3e9197.camel@sipsolutions.net>
 <22462d21-86c0-a2f0-265f-044bfb14719a@candelatech.com>
Organization: Candela Technologies
In-Reply-To: <22462d21-86c0-a2f0-265f-044bfb14719a@candelatech.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 10/4/23 13:36, Ben Greear wrote:
> On 10/4/23 12:38, Johannes Berg wrote:
>> On Tue, 2023-10-03 at 09:43 -0700, greearb@candelatech.com wrote:
>>> From: Ben Greear <greearb@candelatech.com>
>>>
>>> Check if user has configured STA to be AX mode, and if so,
>>> skip the check for MLD elements (as they would not be needed
>>> in AX mode).
>>
>> Eh, no, I think this is wrong.
>>
>>>
>>>           if (ieee80211_vif_is_mld(&sdata->vif)) {
>>
>> If you get past this if, you've (locally) committed to doing EHT
>> already, not just HE (née 11ax), so should have an EHT connection?
>>
>> Though the change is hard to read - but why are you telling the
>> supplicant to connect with MLO if you wanted to not use EHT?
> 
> It is mostly white-space changes, but the diff is ugly.
> 
> I caused the DISABLE_EHT flag to be set in mac80211, and that generally
> seems to work in my kernel, but something is still trying to do MLO I guess.
> 
> I did not specifically tell supplicant to do MLO, and in fact I told
> it to disable_be, but maybe that is not fully implemented.  My supplicant
> is a few months old vs upstream at this point.

I created a patch to supplicant where it will disable MLO if disable_eht
is true, and that seems to have properly associated a non-MLO station
in AX mode.

So I agree that my patch to the kernel is not needed.

Thanks for the patch review and comments.

Thanks,
Ben