Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Wed, 27 May 2020 00:31:50 -0700
From:   Nathan Chancellor <natechancellor@gmail.com>
To:     Bartosz Golaszewski <brgl@bgdev.pl>
Cc:     Rob Herring <robh+dt@kernel.org>,
        "David S . Miller" <davem@davemloft.net>,
        Matthias Brugger <matthias.bgg@gmail.com>,
        John Crispin <john@phrozen.org>,
        Sean Wang <sean.wang@mediatek.com>,
        Mark Lee <Mark-MC.Lee@mediatek.com>,
        Jakub Kicinski <kuba@kernel.org>,
        Arnd Bergmann <arnd@arndb.de>,
        Fabien Parent <fparent@baylibre.com>,
        Heiner Kallweit <hkallweit1@gmail.com>,
        Edwin Peer <edwin.peer@broadcom.com>,
        devicetree@vger.kernel.org, linux-kernel@vger.kernel.org,
        netdev@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
        linux-mediatek@lists.infradead.org,
        Stephane Le Provost <stephane.leprovost@mediatek.com>,
        Pedro Tsai <pedro.tsai@mediatek.com>,
        Andrew Perepech <andrew.perepech@mediatek.com>,
        Bartosz Golaszewski <bgolaszewski@baylibre.com>,
        clang-built-linux@googlegroups.com
Subject: Re: [PATCH v5 06/11] net: ethernet: mtk-star-emac: new driver
Message-ID: <20200527073150.GA3384158@ubuntu-s3-xlarge-x86>
References: <20200522120700.838-1-brgl@bgdev.pl>
 <20200522120700.838-7-brgl@bgdev.pl>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20200522120700.838-7-brgl@bgdev.pl>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Fri, May 22, 2020 at 02:06:55PM +0200, Bartosz Golaszewski wrote:

<snip>

> diff --git a/drivers/net/ethernet/mediatek/mtk_star_emac.c b/drivers/net/ethernet/mediatek/mtk_star_emac.c
> new file mode 100644
> index 000000000000..789c77af501f
> --- /dev/null
> +++ b/drivers/net/ethernet/mediatek/mtk_star_emac.c
> @@ -0,0 +1,1678 @@

<snip>

I've searched netdev and I cannot find any reports from others but this
function introduces a clang warning:

drivers/net/ethernet/mediatek/mtk_star_emac.c:1296:6: warning: variable 'new_dma_addr' is used uninitialized whenever 'if' condition is true [-Wsometimes-uninitialized]
        if (!new_skb) {
            ^~~~~~~~
drivers/net/ethernet/mediatek/mtk_star_emac.c:1321:23: note: uninitialized use occurs here
        desc_data.dma_addr = new_dma_addr;
                             ^~~~~~~~~~~~
drivers/net/ethernet/mediatek/mtk_star_emac.c:1296:2: note: remove the 'if' if its condition is always false
        if (!new_skb) {
        ^~~~~~~~~~~~~~~
drivers/net/ethernet/mediatek/mtk_star_emac.c:1285:6: warning: variable 'new_dma_addr' is used uninitialized whenever 'if' condition is true [-Wsometimes-uninitialized]
        if ((desc_data.flags & MTK_STAR_DESC_BIT_RX_CRCE) ||
            ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/net/ethernet/mediatek/mtk_star_emac.c:1321:23: note: uninitialized use occurs here
        desc_data.dma_addr = new_dma_addr;
                             ^~~~~~~~~~~~
drivers/net/ethernet/mediatek/mtk_star_emac.c:1285:2: note: remove the 'if' if its condition is always false
        if ((desc_data.flags & MTK_STAR_DESC_BIT_RX_CRCE) ||
        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/net/ethernet/mediatek/mtk_star_emac.c:1285:6: warning: variable 'new_dma_addr' is used uninitialized whenever '||' condition is true [-Wsometimes-uninitialized]
        if ((desc_data.flags & MTK_STAR_DESC_BIT_RX_CRCE) ||
            ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/net/ethernet/mediatek/mtk_star_emac.c:1321:23: note: uninitialized use occurs here
        desc_data.dma_addr = new_dma_addr;
                             ^~~~~~~~~~~~
drivers/net/ethernet/mediatek/mtk_star_emac.c:1285:6: note: remove the '||' if its condition is always false
        if ((desc_data.flags & MTK_STAR_DESC_BIT_RX_CRCE) ||
            ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/net/ethernet/mediatek/mtk_star_emac.c:1274:25: note: initialize the variable 'new_dma_addr' to silence this warning
        dma_addr_t new_dma_addr;
                               ^
                                = 0
3 warnings generated.

> +static int mtk_star_receive_packet(struct mtk_star_priv *priv)
> +{
> +	struct mtk_star_ring *ring = &priv->rx_ring;
> +	struct device *dev = mtk_star_get_dev(priv);
> +	struct mtk_star_ring_desc_data desc_data;
> +	struct net_device *ndev = priv->ndev;
> +	struct sk_buff *curr_skb, *new_skb;
> +	dma_addr_t new_dma_addr;

Uninitialized here

> +	int ret;
> +
> +	spin_lock(&priv->lock);
> +	ret = mtk_star_ring_pop_tail(ring, &desc_data);
> +	spin_unlock(&priv->lock);
> +	if (ret)
> +		return -1;
> +
> +	curr_skb = desc_data.skb;
> +
> +	if ((desc_data.flags & MTK_STAR_DESC_BIT_RX_CRCE) ||
> +	    (desc_data.flags & MTK_STAR_DESC_BIT_RX_OSIZE)) {
> +		/* Error packet -> drop and reuse skb. */
> +		new_skb = curr_skb;
> +		goto push_new_skb;

this goto

> +	}
> +
> +	/* Prepare new skb before receiving the current one. Reuse the current
> +	 * skb if we fail at any point.
> +	 */
> +	new_skb = mtk_star_alloc_skb(ndev);
> +	if (!new_skb) {
> +		ndev->stats.rx_dropped++;
> +		new_skb = curr_skb;
> +		goto push_new_skb;

and this goto

> +	}
> +
> +	new_dma_addr = mtk_star_dma_map_rx(priv, new_skb);
> +	if (dma_mapping_error(dev, new_dma_addr)) {
> +		ndev->stats.rx_dropped++;
> +		dev_kfree_skb(new_skb);
> +		new_skb = curr_skb;
> +		netdev_err(ndev, "DMA mapping error of RX descriptor\n");
> +		goto push_new_skb;
> +	}
> +
> +	/* We can't fail anymore at this point: it's safe to unmap the skb. */
> +	mtk_star_dma_unmap_rx(priv, &desc_data);
> +
> +	skb_put(desc_data.skb, desc_data.len);
> +	desc_data.skb->ip_summed = CHECKSUM_NONE;
> +	desc_data.skb->protocol = eth_type_trans(desc_data.skb, ndev);
> +	desc_data.skb->dev = ndev;
> +	netif_receive_skb(desc_data.skb);
> +
> +push_new_skb:
> +	desc_data.dma_addr = new_dma_addr;

assign it uninitialized here.

> +	desc_data.len = skb_tailroom(new_skb);
> +	desc_data.skb = new_skb;
> +
> +	spin_lock(&priv->lock);
> +	mtk_star_ring_push_head_rx(ring, &desc_data);
> +	spin_unlock(&priv->lock);
> +
> +	return 0;
> +}

I don't know if there should be a new label that excludes that
assignment for those particular gotos or if new_dma_addr should
be initialized to something at the top. Please take a look at
addressing this when you get a chance.

Cheers,
Nathan