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;
Message-ID: <dab1abfd-572f-3b4d-4abb-499f61fc91ee@jetfuse.net>
Date:   Fri, 14 Jan 2022 11:22:26 -0600
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.4.0
Subject: Re: iwlwifi: AX201 not using 2 receive streams
Content-Language: en-US
To:     Ben Greear <greearb@candelatech.com>,
        linux-wireless@vger.kernel.org
References: <3cb11500-6828-6ded-bebc-cdb6e1f97bc0@jetfuse.net>
 <521b0dba-5866-ff34-da5a-479bc95cc9d9@candelatech.com>
 <60724ae6-e8e2-183d-575a-0186e45c3fae@jetfuse.net>
 <679864d0-25da-b3c7-b1da-c77deb52fd3a@candelatech.com>
From:   Brandon Nielsen <nielsenb@jetfuse.net>
In-Reply-To: <679864d0-25da-b3c7-b1da-c77deb52fd3a@candelatech.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 1/13/22 11:12 AM, Ben Greear wrote:
> On 1/11/22 11:46 AM, Brandon Nielsen wrote:
>> On 1/8/22 3:22 PM, Ben Greear wrote:
>>> On 1/8/22 12:50 PM, Brandon Nielsen wrote:
>>>> I have a system with Intel AX201 not using 2 spatial streams in the 
>>>> RX (download) direction. TX (upload) functions as expected. I'm 
>>>> seeing similar behavior with an AX200 system. I'm seeing the same 
>>>> behavior on 5.15.12 as I am with 5.16.0-rc8 so I didn't mark it as a 
>>>> regression. Windows works as expected.
>>>>
>>>> I have opened bug 215465[0] with some more detail.
>>>>
>>>> [0] - https://bugzilla.kernel.org/show_bug.cgi?id=215465
>>>>
>>>
>>> We saw something similar with ax210, but only on 6E, in kernel 
>>> 5.15.7+. On 5Ghz, it worked 2x2 as expected.
>>> On a 5.15.0-rc something, it worked OK for us.
>>>
>>> We verified same firmware is used in both cases, but have not fully 
>>> bisected the problem yet but we are
>>> working on it.  In case you find the problem patch before us, please 
>>> let me know!
>>>
>>> Thanks,
>>> Ben
>>>
>>
>> I'm still seeing this as far back as 5.11.12, so I'm not convinced 
>> this is a regression. At least for the 5 GHz / 802.11ac case.
>>
> 
> Our engineer Chuck Rekierre bisected the problem we saw to a patch that 
> came in as part of a stable patch.
> 
> Please see his revert patch: [PATCH] Revert "iwlwifi: mvm: disable 
> RX-diversity in powersave"
> posted yesterday, and if you get a chance to test this, please let us 
> know the results.
> 
> Thanks,
> Ben
> 

Just tried it on one of the AX200 systems exhibiting the issue. No 
change. Rock solid MCS 9, but no engagement of the second RX stream.