From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH] net: ucc_geth, increase no. of HW RX descriptors Date: Wed, 02 Nov 2011 00:56:50 -0400 (EDT) Message-ID: <20111102.005650.1042174729471815202.davem@davemloft.net> References: <1319548630-6561-1-git-send-email-Joakim.Tjernlund@transmode.se> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: avorontsov@ru.mvista.com, netdev@vger.kernel.org To: Joakim.Tjernlund@transmode.se Return-path: Received: from shards.monkeyblade.net ([198.137.202.13]:46070 "EHLO shards.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750885Ab1KBE45 (ORCPT ); Wed, 2 Nov 2011 00:56:57 -0400 In-Reply-To: <1319548630-6561-1-git-send-email-Joakim.Tjernlund@transmode.se> Sender: netdev-owner@vger.kernel.org List-ID: From: Joakim Tjernlund Date: Tue, 25 Oct 2011 15:17:10 +0200 > In a busy network we see ucc_geth is dropping RX pkgs every now > and then. Increase the first RX queues HW descriptors from > 16 to 32 to deal with this. > > Signed-off-by: Joakim Tjernlund 1) This patch doesn't apply, because in the current tree this dirver lives under drivers/net/ethernet/freescale/ not drivers/net 2) Why not increase RX_BD_RING_LEN and thus all the ring sizes, is there a special reason only the first ring's size should be increased? Because the way you've made this change is confusing, someone will see RX_BD_RING_LEN's definition and believe that this is the size of the primary ring unless they happen to notice how .bdRingLenRx is initialized in this special way. Thus, this patch needs changes before it can even be considered for inclusion.