From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Wed, 25 Mar 2009 14:35:58 -0700 (PDT) Message-Id: <20090325.143558.35670359.davem@davemloft.net> From: David Miller In-Reply-To: <2a27d3730903250143u58a3c67dp132a8a5755d50d93@mail.gmail.com> References: <20090323085122.4c9d21f2@nehalam> <20090323.152028.205335400.davem@davemloft.net> <2a27d3730903250143u58a3c67dp132a8a5755d50d93@mail.gmail.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Subject: Re: [Bridge] [PATCH] net/bridge: use the maximum hard_header_len of ports for bridging device List-Id: Linux Ethernet Bridging List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: leoli@freescale.com Cc: netdev@vger.kernel.org, bridge@lists.linux-foundation.org From: Li Yang Date: Wed, 25 Mar 2009 16:43:53 +0800 > Dynamically adjusting is a good idea, but the rx_alloc_extra can only > go up not the other way down in your code. That's not a problem. > Another thought is that if you re-allocate skb here the driver would > be saved from checking the headroom in the fastpath, am I right? You're going to need it for other paths. There are other kinds of tunnels et al. that eat that headroom space on you.