Bug 12733

Summary: mpt fusion driver crashes with LSI22320SE installed and BIOS ACPI turned off
Product: Platform Specific/Hardware Reporter: Maurice Volaski (mvolaski)
Component: i386Assignee: platform_i386
Status: CLOSED INSUFFICIENT_DATA    
Severity: high CC: alan
Priority: P1    
Hardware: All   
OS: Linux   
Kernel Version: 2.6.27.5-117 Subsystem:
Regression: No Bisected commit-id:

Description Maurice Volaski 2009-02-17 15:04:16 UTC
On a system with a Supermicro X7DWN+ with ACPI turned off, and an LSI22320SE card installed, the mptdriver will crash. (The problem appears to affect other boards as well).

Here is the relevant log:

Oct  5 21:19:54 localhost kernel: Fusion MPT SPI Host driver 3.04.07
Oct  5 21:19:54 localhost kernel: mptbase: ioc0: Initiating bringup
Oct  5 21:19:54 localhost kernel: [drm] Initialized radeon 1.29.0 20080528 on minor 0
Oct  5 21:19:54 localhost kernel: ioc0: LSI53C1030 C0: Capabilities={Initiator,Target}
Oct  5 21:19:54 localhost kernel: scsi2 : ioc0: LSI53C1030 C0, FwRev=01033400h, Ports=1, MaxQ=255, IRQ=52
Oct  5 21:19:54 localhost kernel: mptbase: ioc1: Initiating bringup
Oct  5 21:19:54 localhost kernel: ioc1: LSI53C1030 C0: Capabilities={Initiator,Target}
Oct  5 21:19:54 localhost kernel: scsi3 : ioc1: LSI53C1030 C0, FwRev=01033400h, Ports=1, MaxQ=255, IRQ=52
Oct  5 21:19:54 localhost kernel: device-mapper: multipath: version 1.0.5 loaded
Oct  5 21:19:54 localhost kernel: EXT3 FS on dm-0, internal journal
Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: attempting task abort! (sc=f5aca500)
Oct  5 21:19:54 localhost kernel: scsi 3:0:4:0: CDB: Inquiry: 12 00 00 00 24 00
Oct  5 21:19:54 localhost kernel: mptbase: ioc1: Initiating recovery
Oct  5 21:19:54 localhost kernel: scsi 3:0:4:0: mptscsih: ioc1: completing cmds: fw_channel 0, fw_id 4, sc=f5aca500, mf = f48e3160, idx=19
Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: Issue of TaskMgmt failed!
Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: task abort: FAILED (sc=f5aca500)
Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: attempting target reset! (sc=f5aca500)
Oct  5 21:19:54 localhost kernel: scsi 3:0:4:0: CDB: Inquiry: 12 00 00 00 24 00
Oct  5 21:19:54 localhost kernel: mptbase: ioc1: Initiating recovery
Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: Issue of TaskMgmt failed!
Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: target reset: FAILED (sc=f5aca500)
Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: attempting bus reset! (sc=f5aca500)
Oct  5 21:19:54 localhost kernel: scsi 3:0:4:0: CDB: Inquiry: 12 00 00 00 24 00
Oct  5 21:19:54 localhost kernel: scsi target3:0:4: mptspi: ioc1: mpt_config failed
Oct  5 21:19:54 localhost kernel: scsi target3:0:4: FAST-5 WIDE SCSI 3.4 MB/s ST QAS RDSTRM WRFLOW (572 ns, offset 133)
Oct  5 21:19:54 localhost kernel: mptbase: ioc1: Initiating recovery

and

Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: Issue of TaskMgmt failed!
Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: bus reset: FAILED (sc=f5aca500)
Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: attempting host reset! (sc=f5aca500)
Oct  5 21:19:54 localhost kernel: mptbase: ioc1: Initiating recovery
Oct  5 21:19:54 localhost kernel: scsi target3:0:4: mptspi: ioc1: mpt_config failed
Oct  5 21:19:54 localhost kernel: scsi target3:0:4: mptspi: ioc1: mpt_config failed
Oct  5 21:19:54 localhost kernel: scsi target3:0:4: FAST-5 WIDE SCSI 3.4 MB/s ST QAS RDSTRM WRFLOW (572 ns, offset 133)
Oct  5 21:19:54 localhost kernel: scsi target3:0:4: mptspi: ioc1: mpt_config failed
Oct  5 21:19:54 localhost kernel: scsi target3:0:4: mptspi: ioc1: mpt_config failed
Oct  5 21:19:54 localhost kernel: scsi target3:0:4: FAST-5 WIDE SCSI 3.4 MB/s ST QAS RDSTRM WRFLOW (572 ns, offset 133)
Oct  5 21:19:54 localhost kernel: scsi target3:0:4: mptspi: ioc1: mpt_config failed
Comment 1 Anonymous Emailer 2009-02-17 15:16:58 UTC
Reply-To: James.Bottomley@HansenPartnership.com

On Tue, 2009-02-17 at 15:04 -0800, bugme-daemon@bugzilla.kernel.org
wrote:
> http://bugzilla.kernel.org/show_bug.cgi?id=12733
> 
>            Summary: mpt fusion driver crashes with LSI22320SE installed and
>                     BIOS ACPI turned off
>            Product: SCSI Drivers
>            Version: 2.5
>      KernelVersion: 2.6.27.5-117
>           Platform: All
>         OS/Version: Linux
>               Tree: Mainline
>             Status: NEW
>           Severity: high
>           Priority: P1
>          Component: Other
>         AssignedTo: scsi_drivers-other@kernel-bugs.osdl.org
>         ReportedBy: mvolaski@aecom.yu.edu
> 
> 
> On a system with a Supermicro X7DWN+ with ACPI turned off, and an LSI22320SE
> card installed, the mptdriver will crash. (The problem appears to affect
> other
> boards as well).

The first question is can this board operate without ACPI?  It's used to
get the interrupt routings and your log below indicates that the driver
isn't receiving the card interrupt.

> Here is the relevant log:
> 
> Oct  5 21:19:54 localhost kernel: Fusion MPT SPI Host driver 3.04.07
> Oct  5 21:19:54 localhost kernel: mptbase: ioc0: Initiating bringup
> Oct  5 21:19:54 localhost kernel: [drm] Initialized radeon 1.29.0 20080528 on
> minor 0
> Oct  5 21:19:54 localhost kernel: ioc0: LSI53C1030 C0:
> Capabilities={Initiator,Target}
> Oct  5 21:19:54 localhost kernel: scsi2 : ioc0: LSI53C1030 C0,
> FwRev=01033400h,
> Ports=1, MaxQ=255, IRQ=52
> Oct  5 21:19:54 localhost kernel: mptbase: ioc1: Initiating bringup
> Oct  5 21:19:54 localhost kernel: ioc1: LSI53C1030 C0:
> Capabilities={Initiator,Target}
> Oct  5 21:19:54 localhost kernel: scsi3 : ioc1: LSI53C1030 C0,
> FwRev=01033400h,
> Ports=1, MaxQ=255, IRQ=52

Is 52 the right interrupt number?

> Oct  5 21:19:54 localhost kernel: device-mapper: multipath: version 1.0.5
> loaded
> Oct  5 21:19:54 localhost kernel: EXT3 FS on dm-0, internal journal
> Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: attempting task abort!
> (sc=f5aca500)
> Oct  5 21:19:54 localhost kernel: scsi 3:0:4:0: CDB: Inquiry: 12 00 00 00 24
> 00
> Oct  5 21:19:54 localhost kernel: mptbase: ioc1: Initiating recovery
> Oct  5 21:19:54 localhost kernel: scsi 3:0:4:0: mptscsih: ioc1: completing
> cmds: fw_channel 0, fw_id 4, sc=f5aca500, mf = f48e3160, idx=19
> Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: Issue of TaskMgmt failed!
> Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: task abort: FAILED
> (sc=f5aca500)
> Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: attempting target reset!
> (sc=f5aca500)
> Oct  5 21:19:54 localhost kernel: scsi 3:0:4:0: CDB: Inquiry: 12 00 00 00 24
> 00
> Oct  5 21:19:54 localhost kernel: mptbase: ioc1: Initiating recovery
> Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: Issue of TaskMgmt failed!
> Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: target reset: FAILED
> (sc=f5aca500)
> Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: attempting bus reset!
> (sc=f5aca500)
> Oct  5 21:19:54 localhost kernel: scsi 3:0:4:0: CDB: Inquiry: 12 00 00 00 24
> 00
> Oct  5 21:19:54 localhost kernel: scsi target3:0:4: mptspi: ioc1: mpt_config
> failed
> Oct  5 21:19:54 localhost kernel: scsi target3:0:4: FAST-5 WIDE SCSI 3.4 MB/s
> ST QAS RDSTRM WRFLOW (572 ns, offset 133)
> Oct  5 21:19:54 localhost kernel: mptbase: ioc1: Initiating recovery
> 
> and
> 
> Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: Issue of TaskMgmt failed!
> Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: bus reset: FAILED
> (sc=f5aca500)
> Oct  5 21:19:54 localhost kernel: mptscsih: ioc1: attempting host reset!
> (sc=f5aca500)
> Oct  5 21:19:54 localhost kernel: mptbase: ioc1: Initiating recovery
> Oct  5 21:19:54 localhost kernel: scsi target3:0:4: mptspi: ioc1: mpt_config
> failed
> Oct  5 21:19:54 localhost kernel: scsi target3:0:4: mptspi: ioc1: mpt_config
> failed
> Oct  5 21:19:54 localhost kernel: scsi target3:0:4: FAST-5 WIDE SCSI 3.4 MB/s
> ST QAS RDSTRM WRFLOW (572 ns, offset 133)
> Oct  5 21:19:54 localhost kernel: scsi target3:0:4: mptspi: ioc1: mpt_config
> failed
> Oct  5 21:19:54 localhost kernel: scsi target3:0:4: mptspi: ioc1: mpt_config
> failed
> Oct  5 21:19:54 localhost kernel: scsi target3:0:4: FAST-5 WIDE SCSI 3.4 MB/s
> ST QAS RDSTRM WRFLOW (572 ns, offset 133)
> Oct  5 21:19:54 localhost kernel: scsi target3:0:4: mptspi: ioc1: mpt_config
> failed

James
Comment 2 Maurice Volaski 2009-02-17 15:40:23 UTC
With an LSI 8704ELP and LSI SAS3081E-R present and ACPI off, the board works fine. Only the presence of an LSI 22320 causes a crash. 

I don't know about the interrupt number. When I receive the cards back, I can can try to find out.

Our vendor has tested at least one other manufacturer's board with the LSI 22320 and it, too, shows the same crashing behavior when ACPI is turned off.
Comment 3 Anonymous Emailer 2009-02-18 07:32:08 UTC
Reply-To: James.Bottomley@HansenPartnership.com

On Tue, 2009-02-17 at 15:40 -0800, bugme-daemon@bugzilla.kernel.org
wrote:
> http://bugzilla.kernel.org/show_bug.cgi?id=12733
> 
> 
> 
> 
> 
> ------- Comment #2 from mvolaski@aecom.yu.edu  2009-02-17 15:40 -------
> With an LSI 8704ELP and LSI SAS3081E-R present and ACPI off, the board works
> fine. Only the presence of an LSI 22320 causes a crash. 
> 
> I don't know about the interrupt number. When I receive the cards back, I can
> can try to find out.
> 
> Our vendor has tested at least one other manufacturer's board with the LSI
> 22320 and it, too, shows the same crashing behavior when ACPI is turned off.

But the point is that I don't think this is a MPT bug ... it's using the
IRQ line the platform code assigns to it.  If that IRQ line is wrong,
you get the symptoms you're displaying (misrouted IRQ).

I'm not even sure this is a bug:  some modern motherboards simply can't
be configured without ACPI because there's no way to get the IRQ routing
information otherwise.

If your motherboard isn't one of these, then it could be a failure of
the non-ACPI IRQ routing platform code.  None of this can be diagnosed
from the snippet of the log you posted.

James