From mboxrd@z Thu Jan 1 00:00:00 1970 From: Joseph Salisbury Subject: Re: [4.5-rc4 Regression] qla2xxx: Add irq affinity notification Date: Mon, 11 Apr 2016 22:10:11 -0400 Message-ID: <570C5903.1000101@canonical.com> References: <570BDE44.5020301@canonical.com> <32AB2218-5D39-4968-A07F-E527B0557F04@qlogic.com> <570BEFE1.6040703@canonical.com> <44AC3861-DFD2-4420-9ACA-E6AE7A3BDE99@qlogic.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <44AC3861-DFD2-4420-9ACA-E6AE7A3BDE99@qlogic.com> Sender: linux-kernel-owner@vger.kernel.org To: Quinn Tran Cc: Dept-Eng QLA2xxx Upstream , "jejb@linux.vnet.ibm.com" , "Martin K. Petersen" , linux-scsi , linux-kernel , "hch@lst.de" , "bart.vanassche@sandisk.com" , Himanshu Madhani , "nab@linux-iscsi.org" List-Id: linux-scsi@vger.kernel.org On 04/11/2016 05:28 PM, Quinn Tran wrote: > Joe, > > I see the crash point. We=E2=80=99re accessing Null pointer. The ad= apter in use is an older 4G adapter, where it does not have MSIX suppor= t. We=E2=80=99re tripping over the same shared code segment. The foll= owing is the propose fix. Let me know if it works. I=E2=80=99ll will = follow up with a patch for upstream submission. > > Thanks. > > diff --git a/drivers/scsi/qla2xxx/qla_isr.c b/drivers/scsi/qla2xxx/ql= a_isr.c > index 4af9547..79469de 100644 > --- a/drivers/scsi/qla2xxx/qla_isr.c > +++ b/drivers/scsi/qla2xxx/qla_isr.c > @@ -2552,7 +2552,7 @@ void qla24xx_process_response_queue(struct scsi= _qla_host *vha, > if (!vha->flags.online) > return; > =20 > - if (rsp->msix->cpuid !=3D smp_processor_id()) { > + if (rsp->msix && (rsp->msix->cpuid !=3D smp_processor_id())) = { > /* if kernel does not notify qla of IRQ's CPU change, > * then set it here. > */ > > > > > Regards, > Quinn Tran Good news, testing with your new patch resolved the bug and did not require any reverts. Thanks for the quick response! Testing results in comment #39: http://pad.lv/1554003 > > > > > > > -----Original Message----- > From: Joseph Salisbury > Date: Monday, April 11, 2016 at 11:41 AM > To: Quinn Tran > Cc: Dept-Eng QLA2xxx Upstream , "jejb@li= nux.vnet.ibm.com" , "Martin K. Petersen" , linux-scsi , linux= -kernel , "hch@lst.de" , "bar= t.vanassche@sandisk.com" , Himanshu Madhani= , Nicholas Bellinger > Subject: Re: [4.5-rc4 Regression] qla2xxx: Add irq affinity notificat= ion > >> On 04/11/2016 01:48 PM, Quinn Tran wrote: >>> Joe, >>> >>> How do I get access to this specific Ubuntu kernel where the bug is= found? Is there stack trace/bug report that you could share? Any dat= a would be helpful. Thanks. >> The git tree for the specific Ubuntu kernel that exhibits this bug c= an >> be found here: >> git://git.launchpad.net/~ubuntu-kernel/ubuntu/+source/linux/+git/xen= ial >> >> The bug can be found here, and has screen shots of the panic as well= as >> dmesg under the "Attachments" header: >> http://pad.lv/1554003 >> >> Just let me know if additional debug information is needed. >> >> >>> In the mean time, I will download 4.5 rc4 to re-verify. >>> >>> Regards, >>> Quinn Tran >>> >>> >>> >>> >>> >>> >>> -----Original Message----- >>> From: Joseph Salisbury >>> Date: Monday, April 11, 2016 at 10:26 AM >>> To: Quinn Tran >>> Cc: Dept-Eng QLA2xxx Upstream , "jejb@= linux.vnet.ibm.com" , "Martin K. Petersen" , linux-scsi , lin= ux-kernel , "hch@lst.de" , "b= art.vanassche@sandisk.com" , Himanshu Madha= ni , Nicholas Bellinger >>> Subject: [4.5-rc4 Regression] qla2xxx: Add irq affinity notificatio= n >>> >>>> Hello Quinn, >>>> >>>> A kernel bug report was opened against Ubuntu [0]. After a kernel >>>> bisect, it was found that reverting the following commit resolved = this bug: >>>> >>>> commit cdb898c52d1dfad4b4800b83a58b3fe5d352edde >>>> Author: Quinn Tran >>>> Date: Thu Dec 17 14:57:05 2015 -0500 >>>> =20 >>>> qla2xxx: Add irq affinity notification >>>> >>>> >>>> However, the prior commit also required the following three commit= s to >>>> also be reverted: >>>> >>>> commit 5327c7dbd1a7fd980608f44789076a636e5ee5fc >>>> Author: Quinn Tran >>>> Date: Wed Feb 10 18:59:14 2016 -0500 >>>> >>>> qla2xxx: use TARGET_SCF_USE_CPUID flag to indiate CPU Affinity >>>> >>>> commit 9095adaab8c1d82707e4e9961b6ad79b62f3361b >>>> Author: Quinn Tran >>>> Date: Wed Feb 10 18:59:13 2016 -0500 >>>> =20 >>>> target/transport: add flag to indicate CPU Affinity is observed >>>> >>>> commit fb3269baf4ecc2ce6d17d4eb537080035bdf6d5b >>>> Author: Quinn Tran >>>> Date: Thu Dec 17 14:57:06 2015 -0500 >>>> =20 >>>> qla2xxx: Add selective command queuing >>>> >>>> >>>> >>>> The regression was introduced as of v4.5-rc4. >>>> =20 >>>> I was hoping to get your feedback, since you are the patch author.= The >>>> dependant reverts all look like they are improving cpu affinity, w= hich >>>> would likely impact performance. Do you thing there is a way forw= ard >>>> instead of the reverts, or would it be best to submit a revert req= uest? >>>> =20 >>>> =20 >>>> Thanks, >>>> =20 >>>> Joe >>>> >>>> >>>> [0] http://pad.lv/1554003 >>>> >>>>