From mboxrd@z Thu Jan 1 00:00:00 1970 From: James Bottomley Subject: Re: [PATCH] pci: enable MSI on 8132 Date: Wed, 18 Feb 2009 14:18:33 -0600 Message-ID: <1234988313.22084.43.camel@localhost.localdomain> References: <499B46B2.5040601@kernel.org> <499B6BF7.9090300@kernel.org> <499B724A.2040408@kernel.org> <499B774C.5010705@kernel.org> <499B9129.50104@kernel.org> <20090218122137.GJ16841@parisc-linux.org> <20090218100448.7f7c5b86.akpm@linux-foundation.org> <1234982285.3225.42.camel@localhost.localdomain> <499C5CAD.1070404@kernel.org> <1234984495.3225.48.camel@localhost.localdomain> <499C6BC9.6010706@garzik.org> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <499C6BC9.6010706@garzik.org> Sender: linux-pci-owner@vger.kernel.org To: Jeff Garzik Cc: Yinghai Lu , Andrew Morton , david@lang.hm, Matthew Wilcox , Jesse Barnes , linux-kernel , linux-scsi@vger.kernel.org, DL-MPTFusionLinux@lsi.com, linux-pci@vger.kernel.org List-Id: linux-scsi@vger.kernel.org On Wed, 2009-02-18 at 15:12 -0500, Jeff Garzik wrote: > James Bottomley wrote: > > Like I said, I'm happy to have MSI completely disabled until LSI wants > > to comment, so no ... only the PCI quirk fix. The true fix is to have > > the drivers participate in dynamic testing of MSI IRQ routing, but I've > > somewhat lost sight of that. > > That's not really a fix at all, just additional, unneeded overhead for > 99% of users. We don't need our drivers bloated with self-check code > for all the components in our various computer chips... So you think the current per driver variably named parameter to turn off MSI is better? Especially as most of the users who trip across a platform having MSI problems don't even know that the storage failure is caused by MSI, let alone how to find the parameter to fix the problem. James