Re: [net-next v11 4/7] net: mtip: The L2 switch driver for imx287

[Paolo Abeni <pabeni@redhat.com>]

On 5/4/25 4:55 PM, Lukasz Majewski wrote:
> +		/* This does 16 byte alignment, exactly what we need.
> +		 * The packet length includes FCS, but we don't want to
> +		 * include that when passing upstream as it messes up
> +		 * bridging applications.
> +		 */
> +		skb = netdev_alloc_skb(pndev, pkt_len + NET_IP_ALIGN);
> +		if (unlikely(!skb)) {
> +			dev_dbg(&fep->pdev->dev,
> +				"%s: Memory squeeze, dropping packet.\n",
> +				pndev->name);
> +			pndev->stats.rx_dropped++;
> +			goto err_mem;
> +		} else {
> +			skb_reserve(skb, NET_IP_ALIGN);
> +			skb_put(skb, pkt_len);      /* Make room */
> +			skb_copy_to_linear_data(skb, data, pkt_len);
> +			skb->protocol = eth_type_trans(skb, pndev);
> +			napi_gro_receive(&fep->napi, skb);
> +		}
> +
> +		bdp->cbd_bufaddr = dma_map_single(&fep->pdev->dev, data,
> +						  bdp->cbd_datlen,
> +						  DMA_FROM_DEVICE);
> +		if (unlikely(dma_mapping_error(&fep->pdev->dev,
> +					       bdp->cbd_bufaddr))) {
> +			dev_err(&fep->pdev->dev,
> +				"Failed to map descriptor rx buffer\n");
> +			pndev->stats.rx_errors++;
> +			pndev->stats.rx_dropped++;
> +			dev_kfree_skb_any(skb);
> +			goto err_mem;
> +		}

This is doing the mapping and ev. dropping the skb _after_ pushing the
skb up the stack, you must attempt the mapping first.

> +static void mtip_free_buffers(struct net_device *dev)
> +{
> +	struct mtip_ndev_priv *priv = netdev_priv(dev);
> +	struct switch_enet_private *fep = priv->fep;
> +	struct sk_buff *skb;
> +	struct cbd_t *bdp;
> +	int i;
> +
> +	bdp = fep->rx_bd_base;
> +	for (i = 0; i < RX_RING_SIZE; i++) {
> +		skb = fep->rx_skbuff[i];
> +
> +		if (bdp->cbd_bufaddr)
> +			dma_unmap_single(&fep->pdev->dev, bdp->cbd_bufaddr,
> +					 MTIP_SWITCH_RX_FRSIZE,
> +					 DMA_FROM_DEVICE);
> +		if (skb)
> +			dev_kfree_skb(skb);

I suspect that on error paths mtip_free_buffers() can be invoked
multiple consecutive times with any successful allocation in between:
skb will be freed twice. Likely you need to clear fep->rx_skbuff[i] here.

Side note: this patch is way too big for a proper review: you need to
break it in multiple smaller ones, introducing the basic features
separately.

Cheers,

Paolo