From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH] net: introduce alloc_skb_order0 Date: Sat, 16 Oct 2010 11:53:42 -0700 (PDT) Message-ID: <20101016.115342.71098036.davem@davemloft.net> References: <20101011155556.GA2431@redhat.com> <1286813133.2737.36.camel@edumazet-laptop> <1286831867.30423.80.camel@edumazet-laptop> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: sgruszka@redhat.com, romieu@fr.zoreil.com, netdev@vger.kernel.org To: eric.dumazet@gmail.com Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:54698 "EHLO sunset.davemloft.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753547Ab0JPSxT (ORCPT ); Sat, 16 Oct 2010 14:53:19 -0400 In-Reply-To: <1286831867.30423.80.camel@edumazet-laptop> Sender: netdev-owner@vger.kernel.org List-ID: From: Eric Dumazet Date: Mon, 11 Oct 2010 23:17:47 +0200 > [PATCH net-next] r8169: use 50% less ram for RX ring > > Using standard skb allocations in r8169 leads to order-3 allocations (if > PAGE_SIZE=4096), because NIC needs 16383 bytes, and skb overhead makes > this bigger than 16384 -> 32768 bytes per "skb" > > Using kmalloc() permits to reduce memory requirements of one r8169 nic > by 4Mbytes. (256 frames * 16Kbytes). This is fine since a hardware bug > requires us to copy incoming frames, so we build real skb when doing > this copy. > > Signed-off-by: Eric Dumazet Applied.