From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263622AbVHFWIF (ORCPT ); Sat, 6 Aug 2005 18:08:05 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263605AbVHFWIF (ORCPT ); Sat, 6 Aug 2005 18:08:05 -0400 Received: from stat16.steeleye.com ([209.192.50.48]:62682 "EHLO hancock.sc.steeleye.com") by vger.kernel.org with ESMTP id S261310AbVHFWID (ORCPT ); Sat, 6 Aug 2005 18:08:03 -0400 Subject: RE: As of 2.6.13-rc1 Fusion-MPT very slow From: James Bottomley To: Holger Kiehl Cc: "Moore, Eric Dean" , Andrew Morton , linux-kernel , SCSI Mailing List In-Reply-To: References: <91888D455306F94EBD4D168954A9457C035CB64A@nacos172.co.lsil.com> <1123350790.5092.2.camel@mulgrave> Content-Type: text/plain; charset=ISO-8859-1 Date: Sat, 06 Aug 2005 17:07:43 -0500 Message-Id: <1123366064.5102.3.camel@mulgrave> Mime-Version: 1.0 X-Mailer: Evolution 2.0.4 (2.0.4-4) Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Sat, 2005-08-06 at 21:12 +0000, Holger Kiehl wrote: > I tried from 2.6.13-rc2-mm2 up to 2.6.13-rc4-mm1 and always get the following > error when applying this patch: > > CC drivers/message/fusion/mptbase.o > CC drivers/message/fusion/mptscsih.o > CC drivers/message/fusion/mptspi.o > drivers/message/fusion/mptspi.c: In function â..mptspi_target_allocâ..: > drivers/message/fusion/mptspi.c:113: error: invalid storage class for function â..mptspi_write_offsetâ.. > drivers/message/fusion/mptspi.c:114: error: invalid storage class for function â..mptspi_write_widthâ.. > drivers/message/fusion/mptspi.c:131: warning: implicit declaration of function â..mptspi_write_widthâ.. > drivers/message/fusion/mptspi.c: At top level: > drivers/message/fusion/mptspi.c:453: warning: conflicting types for â..mptspi_write_widthâ.. > drivers/message/fusion/mptspi.c:453: error: static declaration of â..mptspi_write_widthâ.. follows non-static declaration > drivers/message/fusion/mptspi.c:131: error: previous implicit declaration of â..mptspi_write_widthâ.. was here This lot are all gcc-4 being silly about a declaration, as you noticed. Still, there's no reason not to make the static functions declared at the top of the file. > drivers/message/fusion/mptspi.c:505: error: unknown field â..get_hold_mcsâ.. specified in initializer > drivers/message/fusion/mptspi.c:505: warning: excess elements in struct initializer > drivers/message/fusion/mptspi.c:505: warning: (near initialization for â..mptspi_transport_functionsâ..) > drivers/message/fusion/mptspi.c:506: error: unknown field â..set_hold_mcsâ.. specified in initializer > drivers/message/fusion/mptspi.c:506: warning: excess elements in struct initializer > drivers/message/fusion/mptspi.c:506: warning: (near initialization for â..mptspi_transport_functionsâ..) > drivers/message/fusion/mptspi.c:507: error: unknown field â..show_hold_mcsâ.. specified in initializer > drivers/message/fusion/mptspi.c:507: warning: excess elements in struct initializer > drivers/message/fusion/mptspi.c:507: warning: (near initialization for â..mptspi_transport_functionsâ..) This is actually because -mm is slightly behind the scsi-misc tree. It looks like the hold_mcs parameters haven't propagated into the -mm tree yet. You should be able to correct this by cutting these three lines: .get_hold_mcs = mptspi_read_parameters, .set_hold_mcs = mptspi_write_hold_mcs, .show_hold_mcs = 1, Out of the code at lines 505-507. You'll get a warning about mptspi_write_hold_mcs() being defined but not used which you can ignore. James