Received-SPF: softfail (google.com: domain of transitioning linux-kernel-owner@vger.kernel.org does not designate 23.128.96.19 as permitted sender) client-ip=23.128.96.19;
MIME-Version: 1.0
References: <20220518090738.2694556-1-yangyingliang@huawei.com>
In-Reply-To: <20220518090738.2694556-1-yangyingliang@huawei.com>
From:   Loic Poulain <loic.poulain@linaro.org>
Date:   Wed, 18 May 2022 11:13:56 +0200
Message-ID: <CAMZdPi9uvh4E70-AXpGrdzkgh35mfWQbhL8Kxw_o9_DsfL2gbw@mail.gmail.com>
Subject: Re: [PATCH -next] net: wwan: t7xx: use GFP_ATOMIC under spin lock in t7xx_cldma_gpd_set_next_ptr()
To:     Yang Yingliang <yangyingliang@huawei.com>
Cc:     linux-kernel@vger.kernel.org, linux-mediatek@lists.infradead.org,
        linux-arm-kernel@lists.infradead.org, netdev@vger.kernel.org,
        haijun.liu@mediatek.com, chandrashekar.devegowda@intel.com,
        ricardo.martinez@linux.intel.com, davem@davemloft.net,
        kuba@kernel.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Hi Yang,

On Wed, 18 May 2022 at 10:57, Yang Yingliang <yangyingliang@huawei.com> wrote:
>
> Sometimes t7xx_cldma_gpd_set_next_ptr() is called under spin lock,
> so add a parameter in t7xx_cldma_gpd_set_next_ptr() to make if it
> use GFP_ATOMIC flag.
>
> Fixes: 39d439047f1d ("net: wwan: t7xx: Add control DMA interface")
> Reported-by: Hulk Robot <hulkci@huawei.com>
> Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
> ---
>  drivers/net/wwan/t7xx/t7xx_hif_cldma.c | 13 ++++++++-----
>  1 file changed, 8 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/net/wwan/t7xx/t7xx_hif_cldma.c b/drivers/net/wwan/t7xx/t7xx_hif_cldma.c
> index 0c52801ed0de..1fa9bb763831 100644
> --- a/drivers/net/wwan/t7xx/t7xx_hif_cldma.c
> +++ b/drivers/net/wwan/t7xx/t7xx_hif_cldma.c
> @@ -91,9 +91,12 @@ static void t7xx_cldma_gpd_set_next_ptr(struct cldma_gpd *gpd, dma_addr_t next_p
>  }
>
>  static int t7xx_cldma_alloc_and_map_skb(struct cldma_ctrl *md_ctrl, struct cldma_request *req,
> -                                       size_t size)
> +                                       size_t size, bool is_atomic)

Would be simpler to directly pass the gfp_mask as a parameter.


>  {
> -       req->skb = __dev_alloc_skb(size, GFP_KERNEL);
> +       if (is_atomic)
> +               req->skb = __dev_alloc_skb(size, GFP_ATOMIC);
> +       else
> +               req->skb = __dev_alloc_skb(size, GFP_KERNEL);
>         if (!req->skb)
>                 return -ENOMEM;
>
> @@ -174,7 +177,7 @@ static int t7xx_cldma_gpd_rx_from_q(struct cldma_queue *queue, int budget, bool
>                 spin_unlock_irqrestore(&queue->ring_lock, flags);
>                 req = queue->rx_refill;
>
> -               ret = t7xx_cldma_alloc_and_map_skb(md_ctrl, req, queue->tr_ring->pkt_size);
> +               ret = t7xx_cldma_alloc_and_map_skb(md_ctrl, req, queue->tr_ring->pkt_size, false);
>                 if (ret)
>                         return ret;
>
> @@ -402,7 +405,7 @@ static struct cldma_request *t7xx_alloc_rx_request(struct cldma_ctrl *md_ctrl, s
>         if (!req->gpd)
>                 goto err_free_req;
>
> -       val = t7xx_cldma_alloc_and_map_skb(md_ctrl, req, pkt_size);
> +       val = t7xx_cldma_alloc_and_map_skb(md_ctrl, req, pkt_size, false);
>         if (val)
>                 goto err_free_pool;
>
> @@ -801,7 +804,7 @@ static int t7xx_cldma_clear_rxq(struct cldma_ctrl *md_ctrl, int qnum)
>                 if (req->skb)
>                         continue;
>
> -               ret = t7xx_cldma_alloc_and_map_skb(md_ctrl, req, rxq->tr_ring->pkt_size);
> +               ret = t7xx_cldma_alloc_and_map_skb(md_ctrl, req, rxq->tr_ring->pkt_size, true);
>                 if (ret)
>                         break;
>
> --
> 2.25.1
>