From mboxrd@z Thu Jan 1 00:00:00 1970 From: Matt Domsch Subject: Re: Aacraid - terrible performance on RHES3 pedge 2450 Date: Mon, 25 Oct 2004 08:01:57 -0500 Sender: linux-raid-owner@vger.kernel.org Message-ID: <20041025130157.GC25232@lists.us.dell.com> References: <60807403EABEB443939A5A7AA8A7458B40B9C9@otce2k01.adaptec.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <60807403EABEB443939A5A7AA8A7458B40B9C9@otce2k01.adaptec.com> To: "Salyzyn, Mark" Cc: Zoe May , linux-raid@vger.kernel.org List-Id: linux-raid.ids On Mon, Oct 25, 2004 at 08:28:47AM -0400, Salyzyn, Mark wrote: > Vary_io may be the reason for this. > > The driver needs to have .vary_io set to 1. > > There are no kernel clues to its existence (that I know of), but the > driver has an SCSI_HAS_VARY_IO to enable. Our build system will produce > correct binaries, but the DKMS tool does not differentiate. > > There is a need to ship a different driver source in the DKMS package > for RHEL3 as a result (Add a #define SCSI_HAS_VARY_IO in the aacraid.h > file). We could do this easily in DKMS with its PATCH_MATCH mechanism, making it match RHEL kernel names and apply the patch only then which sets .vary_io=1. -- Matt Domsch Sr. Software Engineer, Lead Engineer Dell Linux Solutions linux.dell.com & www.dell.com/linux Linux on Dell mailing lists @ http://lists.us.dell.com