Received-SPF: pass (google.com: domain of linux-wireless-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Kalle Valo <kvalo@kernel.org>
To:     Baochen Qiang <quic_bqiang@quicinc.com>
Cc:     <ath12k@lists.infradead.org>, <linux-wireless@vger.kernel.org>
Subject: Re: [PATCH v2] wifi: ath12k: Use msdu_end to check MCBC
References: <20230509033638.3228-1-quic_bqiang@quicinc.com>
Date:   Fri, 26 May 2023 12:32:00 +0300
In-Reply-To: <20230509033638.3228-1-quic_bqiang@quicinc.com> (Baochen Qiang's
        message of "Tue, 9 May 2023 11:36:38 +0800")
Message-ID: <878rdbl9pr.fsf@kernel.org>
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/26.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain
Precedence: bulk

Baochen Qiang <quic_bqiang@quicinc.com> writes:

> We are seeing a very low TCP throughput testing with some specific
> tools. This is because for sub-frames of an AMSDU, MCBC flag in
> mpdu_start may be not valid, and as a result those frames would be
> dropped by kernel.
>
> Add a new helper to get it from msdu_end.
>
> Since original helper is not used for now, add __maybe_unused
> attribute to make GCC happy.
>
> Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.0-03427-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.15378.4
>
> Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com>
> ---
> v2:
>  1. add a new helper according to Vasanth's comment.
>  2. change to use __le16_to_cpu instead of __le32_to_cpu
>
>  drivers/net/wireless/ath/ath12k/dp_rx.c | 13 ++++++++++---
>  drivers/net/wireless/ath/ath12k/hal.c   | 14 ++++++++++++++
>  drivers/net/wireless/ath/ath12k/hal.h   |  1 +
>  3 files changed, 25 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/net/wireless/ath/ath12k/dp_rx.c b/drivers/net/wireless/ath/ath12k/dp_rx.c
> index 8c8162fbe5c6..972034b7f159 100644
> --- a/drivers/net/wireless/ath/ath12k/dp_rx.c
> +++ b/drivers/net/wireless/ath/ath12k/dp_rx.c
> @@ -193,13 +193,20 @@ static void ath12k_dp_rxdesc_set_msdu_len(struct ath12k_base *ab,
>  	ab->hw_params->hal_ops->rx_desc_set_msdu_len(desc, len);
>  }
>  
> -static bool ath12k_dp_rx_h_is_mcbc(struct ath12k_base *ab,
> -				   struct hal_rx_desc *desc)
> +static __maybe_unused bool ath12k_dp_rx_h_is_mcbc(struct ath12k_base *ab,
> +						  struct hal_rx_desc *desc)
>  {
>  	return (ath12k_dp_rx_h_first_msdu(ab, desc) &&
>  		ab->hw_params->hal_ops->rx_desc_is_mcbc(desc));
>  }

Using __maybe_unused is usually a bad idea, it should be used only on
very special cases.

But why do you leave struct hal_ops::rx_desc_is_mcbc? Nobody is using it
and we should not have dead code lying around.

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

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