Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20230320213508.2358213-1-martin.blumenstingl@googlemail.com>
 <20230320213508.2358213-3-martin.blumenstingl@googlemail.com> <f7b9dda9d852456caffc3c0572f88947@realtek.com>
In-Reply-To: <f7b9dda9d852456caffc3c0572f88947@realtek.com>
From:   Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Date:   Thu, 23 Mar 2023 12:16:52 +0100
Message-ID: <CAFBinCCspK=GaCMEiHsXi=0H4Sbp2vg_4EK=8bqQLWR8+qg7Sw@mail.gmail.com>
Subject: Re: [PATCH v3 2/9] wifi: rtw88: sdio: Add HCI implementation for SDIO
 based chipsets
To:     Ping-Ke Shih <pkshih@realtek.com>
Cc:     "linux-wireless@vger.kernel.org" <linux-wireless@vger.kernel.org>,
        Yan-Hsuan Chuang <tony0620emma@gmail.com>,
        Kalle Valo <kvalo@kernel.org>,
        Ulf Hansson <ulf.hansson@linaro.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "netdev@vger.kernel.org" <netdev@vger.kernel.org>,
        "linux-mmc@vger.kernel.org" <linux-mmc@vger.kernel.org>,
        Chris Morgan <macroalpha82@gmail.com>,
        Nitin Gupta <nitin.gupta981@gmail.com>,
        Neo Jou <neojou@gmail.com>,
        Jernej Skrabec <jernej.skrabec@gmail.com>,
        Larry Finger <Larry.Finger@lwfinger.net>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Hello Ping-Ke,

On Thu, Mar 23, 2023 at 3:23=E2=80=AFAM Ping-Ke Shih <pkshih@realtek.com> w=
rote:
[...]
> > +       if (direct) {
> > +               addr =3D rtw_sdio_to_bus_offset(rtwdev, addr);
> > +               val =3D rtw_sdio_readl(rtwdev, addr, &ret);
> > +       } else if (addr & 3) {
>
> else if (IS_ALIGNED(addr, 4) {
I'll add these IS_ALIGNED in v4
Also I found an issue with RTW_WCPU_11N devices where indirect read
works differently (those can't use
REG_SDIO_INDIRECT_REG_CFG/REG_SDIO_INDIRECT_REG_DATA but need to go
through the normal path with WLAN_IOREG_OFFSET instead). I'll also
include that fix in v4

[...]
> > +       ret =3D rtw_register_hw(rtwdev, hw);
> > +       if (ret) {
> > +               rtw_err(rtwdev, "failed to register hw");
> > +               goto err_destroy_txwq;
> > +       }
> > +
>
> Today, people reported there is race condition between register netdev an=
d NAPI
> in rtw89 driver. I wonder if there will be in register netdev and request=
 IRQ.
>
> You can add a msleep(10 * 100) here, and then do 'ifconfig up' and 'iw sc=
an'
> quickly right after SDIO probe to see if it can work well. Otherwise, swi=
tching
> the order of rtw_register_hw() and rtw_sdio_request_irq() could be a poss=
ible
> solution.
I tried with 1 second and 10 second delays here and could not find any prob=
lems.
That said, I am still going to swap the order because a) it seems to
be what most drivers do (ath9k for example) and b) SDIO is a slow bus
and especially slow on my Amlogic SM1 SoC which has some SDIO DMA
errata.
Also testing this made me find another bug during module unregister (I
thought I caught all of them by now) and will include this fix in v4
as well.


As always: thank you for all your inputs!

Best regards,
Martin