From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ben Hutchings Subject: Re: [PATCH v4] lpc32xx: Added ethernet driver Date: Tue, 6 Mar 2012 01:26:17 +0000 Message-ID: <1330997177.3022.51.camel@deadeye> References: <1330983641-32622-1-git-send-email-stigge@antcom.de> <1330987524.2538.61.camel@bwh-desktop> <1330994958.9504.2.camel@edumazet-glaptop> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: Roland Stigge , , , , , , , , , , , , , To: Eric Dumazet Return-path: In-Reply-To: <1330994958.9504.2.camel@edumazet-glaptop> Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On Mon, 2012-03-05 at 16:49 -0800, Eric Dumazet wrote: > On Mon, 2012-03-05 at 22:45 +0000, Ben Hutchings wrote: > > On Mon, 2012-03-05 at 22:40 +0100, Roland Stigge wrote: > > [...] > > ... > > > > + /* Copy packet from buffer */ > > > + memcpy(prdbuf, > > > + pldat->rx_buff_v[rxconsidx], len); > > > + > > > + /* Pass to upper layer */ > > > + skb->protocol = eth_type_trans(skb, ndev); > > > + netif_rx(skb); > > > + ndev->last_rx = jiffies; > > > > Drivers don't need to set last_rx any more. > > Also, a NAPI driver should call netif_receive_skb() instead of > netif_rx() > > netif_rx() is more expensive because of additional queue, while > netif_receive_skb() directly calls the upper stacks. Yes, but not until the poll function is changed to enable IRQs. Ben. -- Ben Hutchings, Staff Engineer, Solarflare Not speaking for my employer; that's the marketing department's job. They asked us to note that Solarflare product names are trademarked.