Received-SPF: pass (google.com: domain of linux-wireless-owner@vger.kernel.org designates 23.128.96.37 as permitted sender) client-ip=23.128.96.37;
Date:   Wed, 20 Sep 2023 16:36:09 -0700
From:   Brian Norris <briannorris@chromium.org>
To:     Dmitry Antipov <dmantipov@yandex.ru>
Cc:     Kalle Valo <kvalo@kernel.org>, linux-wireless@vger.kernel.org,
        lvc-project@linuxtesting.org
Subject: Re: [PATCH 2/2] wifi: mwifiex: followup PCIE and related cleanups
Message-ID: <ZQuB6aEcMGSsmvLs@google.com>
References: <20230919132804.73340-1-dmantipov@yandex.ru>
 <20230919132804.73340-2-dmantipov@yandex.ru>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20230919132804.73340-2-dmantipov@yandex.ru>
Precedence: bulk

On Tue, Sep 19, 2023 at 04:28:00PM +0300, Dmitry Antipov wrote:
> Introduce a few more (PCIE and generic interface related)
> cleanups which becomes reasonable after the previous patch.
> 
> Signed-off-by: Dmitry Antipov <dmantipov@yandex.ru>
> ---
>  drivers/net/wireless/marvell/mwifiex/main.h |  4 +--
>  drivers/net/wireless/marvell/mwifiex/pcie.c | 28 +++++----------------
>  2 files changed, 8 insertions(+), 24 deletions(-)

...

> --- a/drivers/net/wireless/marvell/mwifiex/pcie.c
> +++ b/drivers/net/wireless/marvell/mwifiex/pcie.c
> @@ -710,18 +710,12 @@ static int mwifiex_pm_wakeup_card_complete(struct mwifiex_adapter *adapter)
>   * The host interrupt mask is read, the disable bit is reset and
>   * written back to the card host interrupt mask register.
>   */
> -static int mwifiex_pcie_disable_host_int(struct mwifiex_adapter *adapter)
> +static void mwifiex_pcie_disable_host_int(struct mwifiex_adapter *adapter)
>  {
>  	if (mwifiex_pcie_ok_to_access_hw(adapter))
>  		mwifiex_write_reg(adapter, PCIE_HOST_INT_MASK, 0x00000000);
>  
>  	atomic_set(&adapter->tx_hw_pending, 0);
> -	return 0;
>  }

I have a nagging feeling that some of this is hitting
partially-implemented features (esp. around error handling for the sleep
cookie / mwifiex_pcie_ok_to_access_hw() stuff), but I also bet that the
only chip using that feature (8766) is in bad enough shape that no one
is getting reliable usage out of it anyway. (I know I'm not; ChromeOS
never picked up PCIe support until 8997, where we worked out most of the
driver/firmware bugs.)

So maybe no one cares, and simplifying the dead logic is a net win.

Acked-by: Brian Norris <briannorris@chromium.org>