Sender: Larry Finger <larry.finger@gmail.com>
Message-ID: <121c0039-5f0c-7c4e-5b07-9193ed547079@lwfinger.net>
Date:   Sat, 25 Feb 2023 20:09:51 -0600
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.8.0
From:   Larry Finger <Larry.Finger@lwfinger.net>
Subject: Re: [PATCH] wifi: wext: warn about usage only once
To:     Linus Torvalds <torvalds@linux-foundation.org>,
        Kalle Valo <kvalo@kernel.org>,
        "Berg, Johannes" <johannes.berg@intel.com>
Cc:     linux-wireless@vger.kernel.org, netdev@vger.kernel.org,
        Nicolas Cavallari <Nicolas.Cavallari@green-communications.fr>,
        Johannes Berg <johannes.berg@intel.com>,
        Jan Engelhardt <jengelh@inai.de>
References: <20230224135933.94104aeda1a0.Ie771c6a66d7d6c3cf67da5f3b0c66cea66fd514c@changeid>
 <dff29d82-9c4b-1933-c1c1-a3becf2a0f1f@lwfinger.net>
 <CAHk-=whwDRefJJq0K8bXXSNY3-Zy8=Z3ZiKYh2mOOvfT-MqNhA@mail.gmail.com>
Content-Language: en-US
In-Reply-To: <CAHk-=whwDRefJJq0K8bXXSNY3-Zy8=Z3ZiKYh2mOOvfT-MqNhA@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 2/24/23 13:44, Linus Torvalds wrote:
> You *could* improve on it further by having some kind of timed
> rate-limiting, where every 24 hours you'd clear the warning mask, so
> that you'd warn about these things once a day. That *can* be useful
> for when people just don't notice the warning the first time around,
> and "once a day" is not a horribly problem that fills up the logs like
> the current situation does.
> 
> But again - I personally think even just a pr_warn_once() is likely
> good enough. Because all I want is to not have that horrible
> log-flushing behavior.

To all,

I posted my list of 8 different tasks that generated this warning to the 
openSUSE developers mailing list, and got back a reply from Jan Engelhardt 
pointed me toward the libqt5-qtbase project, which contains the following snippet:

     case ARPHRD_ETHER:
         // check if it's a WiFi interface
         if (qt_safe_ioctl(socket, SIOCGIWMODE, req) >= 0)
             return QNetworkInterface::Wifi;
         return QNetworkInterface::Ethernet;

I am not entirely sure why Qt needs to know what type of device the network is 
using. I tested by replacing this with

     case ARPHRD_ETHER:
         return QNetworkInterface::Ethernet;


After rebuilding the entire project, and reinstalling all 31 packages generated 
in a new build, my system now displays only 3 remaining warnings, namely

warning: `nspr-2' uses wireless extensions that are deprecated for modern 
drivers; use nl80211
warning: `ThreadPoolForeg' uses wireless extensions that are deprecated for 
modern drivers; use nl80211
warning: `nspr-8' uses wireless extensions that are deprecated for modern 
drivers; use nl80211

To answer Kalle's question, libQt is responsible for most of the warnings that 
were reported here.

In case Qt really needs to know what network it is on, what is a better way to 
detect if the network is on a Wifi device?

Larry