Received-SPF: pass (google.com: best guess record for domain of linux-wireless-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <20200204120614.28861-1-yhchuang@realtek.com> <CAB4CAwdFez_WbRQb2jZQtqspSkO5ujL31ZLt4XshNcWiSqHByA@mail.gmail.com>
 <CAMrEMU_KJbwAK-onH0sRUgUPxDuZtD5bPJN6XmF=a9NqpfeJ+g@mail.gmail.com>
 <CA+ASDXPHWNN+TvrJFi_6w+ep_TBtLJ0refKenNb0tc8Vs+YjgQ@mail.gmail.com>
 <CAMrEMU93LScySw4mpidAC5pVHV_NOShP1_GMMsvsAk1QBhdJjQ@mail.gmail.com>
 <CA+ASDXPHXFquF2iXvdxZXhkkLPndkACgenLJwMR_9ON7zZZM3w@mail.gmail.com>
 <87wo8xxueg.fsf@kamboji.qca.qualcomm.com> <CAMrEMU-nM1O_iJPVgGg2pL6JYWMdRKdPGe5N2rkfOihdmTeMaw@mail.gmail.com>
 <2c91e4e7b36d42a8abad6ae356c2869c@realtek.com>
In-Reply-To: <2c91e4e7b36d42a8abad6ae356c2869c@realtek.com>
From:   Brian Norris <briannorris@chromium.org>
Date:   Thu, 20 Feb 2020 11:04:31 -0800
Message-ID: <CA+ASDXMjx8oQzK61rNHwpkKykgS2v_o+HTUyXujY9VXYNiNfxQ@mail.gmail.com>
Subject: Re: [PATCH] rtw88: disable TX-AMSDU on 2.4G band
To:     Tony Chuang <yhchuang@realtek.com>
Cc:     Justin Capella <justincapella@gmail.com>,
        Kalle Valo <kvalo@codeaurora.org>,
        Chris Chiu <chiu@endlessm.com>,
        linux-wireless <linux-wireless@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-wireless-owner@vger.kernel.org
Precedence: bulk

On Tue, Feb 11, 2020 at 6:56 PM Tony Chuang <yhchuang@realtek.com> wrote:
> Module parameters are really good for me, too. But we've had discussion
> before with Kalle and Brian, they both were trying hard to avoid module
> parameters.

My personal preference is to avoid module parameters when you can fix
the defaults, and that module parameters should never be a workaround
for fixing the default behavior.

I don't think my above preference precludes module parameters: they
can be useful as "extra debug knobs."

But Kalle had a little more nuanced opinion here -- he didn't even
want "debug knobs" for core 802.11 functionality, because (I may be
paraphrasing) one person's "debug knob" easily becomes the next
person's "required knob." Additionally, a mess of disorganized knobs
can make maintenance difficult -- one can't really expect the average
distribution to make a good selection on 100 different parameters; and
for those that do tweak the parameters, it now creates a combinatoric
mess to debug and triage user reports of "it's broken". I can respect
all of those reasons too.

Regards,
Brian