From mboxrd@z Thu Jan 1 00:00:00 1970 From: James Bottomley Subject: Re: [BUG] linux-next: Tree for March 25 kernel oops, when loading mpt fusion driver - regression Date: Mon, 07 Jul 2008 15:25:21 -0500 Message-ID: <1215462321.3274.34.camel@localhost.localdomain> References: <20080325175224.3ee62b09.sfr@canb.auug.org.au> <47E9F0B0.4060006@linux.vnet.ibm.com> <20080626203857.GB3815@redhat.com> <1215372215.4197.28.camel@localhost.localdomain> <487279C9.2010802@linux.vnet.ibm.com> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <487279C9.2010802@linux.vnet.ibm.com> Sender: linux-next-owner@vger.kernel.org To: Kamalesh Babulal Cc: David Teigland , Stephen Rothwell , linux-next@vger.kernel.org, LKML , linux-scsi@vger.kernel.org, Eric.Moore@lsi.com, Andy Whitcroft List-Id: linux-scsi@vger.kernel.org On Tue, 2008-07-08 at 01:47 +0530, Kamalesh Babulal wrote: > James Bottomley wrote: > > On Thu, 2008-06-26 at 15:38 -0500, David Teigland wrote: > >> On Wed, Mar 26, 2008 at 12:14:00PM +0530, Kamalesh Babulal wrote: > >>> Hi Stephen, > >>> > >>> Kernel bug is hit while booting up the next-20080325 kernel with MPT > >>> Fusion driver built in.This was reported previously for the > >>> next-20080320 kernel > >>> http://marc.info/?l=linux-next&m=120601013920868&w=2 > >> Hi, did you ever get this fixed? I've been having the same problem, > >> http://marc.info/?l=linux-scsi&m=121061780821823&w=4 > >> still exists on 2.6.26-rc8 for me, > >> > >> Loading scsi_transport_spi.ko module > >> Loading mptscsih.ko module > >> Loading mptspi.ko module > >> Fusion MPT SPI Host driver 3.04.06 > >> ACPI: PCI Interrupt 0000:86:01.0[A] -> GSI 32 (level, low) -> IRQ 32 > >> mptbase: ioc0: Initiating bringup > >> ioc0: LSI53C1030 B2: Capabilities={Initiator,Target} > >> mptbase: ioc0: PCI-MSI enabled > > > > I speculate that this might be related to the problem. The fusion > > driver was recently flipped to default to MSI enable (previously it had > > disabled them). Could you boot it with the fusion module parameter > > > > mpt_msi_enable=0 > > > > if it actually boots that will be a good confirmation that MSI is to > > blame. > > > >> mptbase: ioc0: Initiating recovery > > > > This I suspect is because of a lost interrupt. > > > >> BUG: unable to handle kernel NULL pointer dereference at 0000000000000948 > > > > And this one looks to be a second bug: The MSI fault path is actually > > causing this oops. > > > > James > > > Hi James, > > Kernel Bug is still reproducible with 2.6.26-rc9-git1 kernel, I tried booting with mpt_msi_enable=0 > as boot paramenter, but the call trace remains the same > > oading mptbase.ko module[ 6.766853] Fusion MPT base driver 3.04.06 > > [ 6.783482] Copyright (c) 1999-2007 LSI Corporation > Loading mptscsih.ko module > Loading mptspi.ko module > [ 6.813157] Fusion MPT SPI Host driver 3.04.06 > [ 6.827369] ACPI: PCI Interrupt 0000:01:01.0[A] -> GSI 22 (level, low) -> IRQ 22 > [ 6.850624] mptbase: ioc0: Initiating bringup > [ 6.985601] ioc0: LSI53C1030 B2: Capabilities={Initiator} > [ 7.002711] mptbase: ioc0: PCI-MSI enabled ^^^^^^^^^^^^^^^ If you had successfully specified mpt_msi_enable=0 this line wouldn't appear. My best guess is that you incorrectly specified it as a kernel command line parameter rather than a module parameter. James