MIME-Version: 1.0
References: <1637571856-1191-1-git-send-email-loic.poulain@linaro.org> <8735nobgje.fsf@codeaurora.org>
In-Reply-To: <8735nobgje.fsf@codeaurora.org>
From:   Loic Poulain <loic.poulain@linaro.org>
Date:   Mon, 22 Nov 2021 13:05:30 +0100
Message-ID: <CAMZdPi9xWdJU_-k9mJxTkeyctsEmNCuzvBpdR-FMg=onEim+9w@mail.gmail.com>
Subject: Re: [PATCH] brcmfmac: Configure keep-alive packet on suspend
To:     Kalle Valo <kvalo@codeaurora.org>
Cc:     aspriel@gmail.com, franky.lin@broadcom.com,
        hante.meuleman@broadcom.com, chi-hsien.lin@infineon.com,
        wright.feng@infineon.com, chung-hsien.hsu@infineon.com,
        linux-wireless@vger.kernel.org, brcm80211-dev-list.pdl@broadcom.com
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Hi Kalle,

On Mon, 22 Nov 2021 at 12:01, Kalle Valo <kvalo@codeaurora.org> wrote:
>
> Loic Poulain <loic.poulain@linaro.org> writes:
>
> > When system enter suspend, there is no more wireless traffic, and
> > if there is no incoming data, most of the AP kick-out the client
> > station after few minutes because of inactivity.
> >
> > The usual way to prevent this is to submit a Null function frame
> > periodically as a keep-alive. This is supported by brcm controllers
> > and can be configured via the mkeep_alive IOVAR.
>
> This is with brcmfmac in client mode, right?

Right, it's in client mode.

> Wouldn't it make more sense to disconnect entirely during suspend?
> Nobody is processing the data packets anyway during suspend.

Disconnect is performed automatically when wowlan is not enabled,
otherwise we may want to wake-up on events (disconnect,
4-way-handshake) or data packets (magic, unicast, etc...). Some
devices use suspend aggressively such as Android in which the network
link is expected to be maintained.

Regards,
Loic