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;
To:     linux-wireless@vger.kernel.org
From:   Mark Nelson <mark.a.nelson@gmail.com>
Subject: iwlwifi: understanding potential firmware regression
Message-ID: <de4c5734-f9f6-b492-0e3d-4ff814a4b6cf@gmail.com>
Date:   Mon, 24 May 2021 19:02:26 -0500
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101
 Thunderbird/68.10.0
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Precedence: bulk

Hi Folks,

I've been fighting with an Intel branded ax200 160MHz Rev=0x340 for 
several hours today.  This is on Ubuntu 20.04 setup with the standard 
ubuntu 5.8.0-53-generic kernel.  Out of the gate I was seeing timeout 
(-110) errors in various places in dmesg (The exact error depends on the 
firmware version and module options, but always seemed to be some 
manifestation of the same general problem).  One example with the 
default iwlwifi-cc-a0-55.ucode firmware:

iwlwifi 0000:03:00.0: Failed to set soc latency: -110

but potentially also:

iwlwifi 0000:03:00.0: PHY ctxt cmd error. ret=-110

After many failed attempts and a ton of trial and error, I was able to 
get the driver loaded but only under very specific conditions:

1) The firmware must be iwlwifi-cc-a0-48.ucode (48.4fa0041f.0) or older. 
  All more recent versions I tried beginning with iwlwifi-cc-a0-50.ucode 
would result in the same errors (including on the ubuntu 5.10 oem kernels).

2) iwlwifi-cc-a0-48 and older would only load with iwlmvm power_scheme=1 
set as an option in /etc/modprobe.d/iwlwifi.conf.

That was enough to get the driver to load, but several other iwlwifi 
options appear to be required to avoid instability and tainted kernels: 
swcrypto=0 power_save=0.  At one point I was still having issues with tx 
fifo queue flushing, but that may have been related to using 
backport-iwlwifi-dkms which I've since removed.

For now, I wanted to document the progress and fix for posterity and 
also ask a question:  Is a changelog for the iwlwifi firmware revisions 
posted anywhere?  The primary issue I was facing today seems to be 
related to some change that occurred between iwlwifi-cc-a0-48 and 
iwlwifi-cc-a0-50.  So far I have had very little luck tracking down 
what's different.

Thanks,
Mark