From:   Kalle Valo <kvalo@kernel.org>
To:     David Mosberger-Tang <davidm@egauge.net>
Cc:     Ajay Singh <ajay.kathat@microchip.com>,
        Claudiu Beznea <claudiu.beznea@microchip.com>,
        linux-wireless <linux-wireless@vger.kernel.org>,
        "netdev\@vger.kernel.org" <netdev@vger.kernel.org>
Subject: Re: RFC: wilc1000: refactor TX path to use sk_buff queue
References: <e3502ecffe0c4c01b263ada8deed814d5135c24c.camel@egauge.net>
Date:   Tue, 14 Dec 2021 19:36:47 +0200
In-Reply-To: <e3502ecffe0c4c01b263ada8deed814d5135c24c.camel@egauge.net>
        (David Mosberger-Tang's message of "Sat, 11 Dec 2021 13:32:09 -0700")
Message-ID: <8735mvhyvk.fsf@codeaurora.org>
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/26.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain
Precedence: bulk

David Mosberger-Tang <davidm@egauge.net> writes:

> I'd like to propose to restructure the wilc1000 TX path to take
> advantage of the existing sk_buff queuing and buffer operations rather
> than using a driver-specific solution.  To me, the resulting code looks
> simpler and the diffstat shows a fair amount of code-reduction:
>
>  cfg80211.c |   35 ----
>  mon.c      |   36 ----
>  netdev.c   |   28 ---
>  netdev.h   |   10 -
>  wlan.c     |  499 ++++++++++++++++++++++++++-----------------------------------
>  wlan.h     |   51 ++----
>  6 files changed, 255 insertions(+), 404 deletions(-)

Looks like a very good cleanup.

> +static void wilc_wlan_txq_drop_net_pkt(struct sk_buff *skb)
> +{
> +	struct wilc_vif *vif = netdev_priv(skb->dev);
> +	struct wilc *wilc = vif->wilc;
> +	struct wilc_skb_tx_cb *tx_cb = WILC_SKB_TX_CB(skb);
> +
> +	if ((u8)tx_cb->q_num >= NQUEUES) {
> +		netdev_err(vif->ndev, "Invalid AC queue number %d",
> +			   tx_cb->q_num);
> +		return;
> +	}

But why the cast here? Casting should be avoided as much as possible,
they just create so many problems if used badly.

-- 
https://patchwork.kernel.org/project/linux-wireless/list/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches