Re: [PATCH] Gigabit Ethernet driver of Topcliff PCH

[Eric Dumazet <eric.dumazet@gmail.com>]

Le mardi 31 août 2010 à 23:15 +0900, Masayuki Ohtake a écrit :
> Hi Sam, Joe and Stephen
> 
> Thank you for your comments.
> We have modified this driver for your comment.
> 
> ---
> Gigabit Ethernet driver of Topcliff PCH

> +			skb->protocol = eth_type_trans(skb, netdev);
> +			if ((tcp_ip_status & PCH_GBE_RXD_ACC_STAT_TCPIPOK) ==
> +			    PCH_GBE_RXD_ACC_STAT_TCPIPOK) {
> +				skb->ip_summed = CHECKSUM_UNNECESSARY;
> +			} else {
> +				skb->ip_summed = CHECKSUM_NONE;
> +			}
> +
> +			if (netif_receive_skb(skb) == NET_RX_DROP) {
> +				adapter->stats.rx_dropped++;

dont use pr_err() here, dropping a frame is a normal condition.

> +				pr_err("Receive Netif Receive Dropped Error\n");
> +			}
> +			(*work_done)++;
> +			netdev->last_rx = jiffies;

really ? last_rx is a thing of the past.

> +			pr_debug
> +			    ("Receive skb->ip_summed: %d length: %d\n",
> +			     skb->ip_summed, length);
> +		}
> +dorrop:
> +		/* return some buffers to hardware, one at a time is too slow */
> +		if (unlikely(cleaned_count >= PCH_GBE_RX_BUFFER_WRITE)) {
> +			pch_gbe_alloc_rx_buffers(adapter, rx_ring,
> +						 cleaned_count);
> +			cleaned_count = 0;
> +		}
> +		if (++i == rx_ring->count)
> +			i = 0;
> +	}
> +	rx_ring->next_to_clean = i;
> +	if (cleaned_count)
> +		pch_gbe_alloc_rx_buffers(adapter, rx_ring, cleaned_count);
> +	return cleaned;
> +}
> +



I find hard to believe this driver needs to copy all outgoing frames on
pre-allocated skbs.

+       /* [Header:14][payload] ---> [Header:14][paddong:2][payload]    */
+       memcpy(tmp_skb->data, skb->data, ETH_HLEN);
+       tmp_skb->data[ETH_HLEN] = 0x00;
+       tmp_skb->data[ETH_HLEN + 1] = 0x00;
+       tmp_skb->len = skb->len;
+       memcpy(&tmp_skb->data[ETH_HLEN + 2], &skb->data[ETH_HLEN],
+              (skb->len - ETH_HLEN));
+       buffer_info->kernel_skb = skb;
+       skb = tmp_skb;

Whats the deal here please ?