Server Products
Data Center Products including boards, integrated systems, Intel® Xeon® Processors, RAID Storage, and Intel® Xeon® Processors
4762 Discussions

S1200BTL - resume from S3(sleep) - BSOD if external PCIe card is used ?

idata
Employee
1,767 Views

Hello,

Running v30 bios.

First question is if this board supports S3(sleep) power state ? - If yes we can proceed further.

I'm getting BSOD in windows 7 64bit after S3 resume if external PCIe video card is present.

External video card is Matrox M9128 PCIe x16, also tried nvidia quadro NVS 295 - exacly the same bahaviour - maybe it happens with all external PCIe cards ? It is not necessary to use external card - if we're using KVM video and external card is not connected to any monitor(just plugged in) - the same happens.

To reproduce:

1. Sleep command works fine - system turns off.

2. System resumes from S3 - functionality is ok - external video card works fine, system works fine.

3. BSOD after about 5mins after resume.

After resume:

Mainboard diagnostics leds:

ooooooox (o=on, x=off) - value 0xFE is not documented ?

Windows BSOD:

BMC events (first event is power on for resume):

Any ideas ?

thanks

Vilius Mockunas

0 Kudos
6 Replies
Daniel_O_Intel
Employee
516 Views

S3 is definitely supported in BIOS 30 and later - see http://downloadmirror.intel.com/20476/eng/Release.txt http://downloadmirror.intel.com/20476/eng/Release.txt

The only thing I would think to check is, if you don't use either card, is S3 okay? Does it run forever without a blue screen?

The error does indicate that one of the drivers is failing to come back correctly from S3, but I've never seen one take five minutes to error. It's usually immediate.

0 Kudos
idata
Employee
516 Views

Hello,

If external video card is not plugged in S3 is OK.

Some thoughts:

 

At first I blamed video card but afer trying two cards from different vendors and observing exactly the same behavior..

Anyways I think the clue could be boards diagnostic leds after S3 resume - what state they should be after successful S3 resume ? (what I'm getting I described above).

Another thing I could think of is BMC error - it's "BIOS POST Progress". If we simulate some usual BSOD because of driver fault - what BMC error we should get ?

Time untill error - I never counted exacly but BMC log shows it's about 10mins.

Vilius M

0 Kudos
Daniel_O_Intel
Employee
516 Views

I checked the supported product list at http://serverconfigurator.intel.com/sctthol.aspx?search=Server+Board+S1200BTL http://serverconfigurator.intel.com/sctthol.aspx?search=Server+Board+S1200BTL, but didn't see *any* video cards on the list.

As far as I know, the BMC shouldn't be trapping any software issue, only hardware issues. So a strictly driver-based BSOD wouldn't show in the System Event Log, even if you were to artificially cause one.

The LEDs aren't going to be much clue either, because the BMC stops updating them once the OS hits protected mode. The LED state is only relevant during POST and right as the OS is taking over.

0 Kudos
idata
Employee
516 Views

Hello,

I dug a little bit further on this.

Before going further - the problem is evident using 3 very different unsupported PCIe external cards - so I believe that the same issue should happen with supported external PCIe cards either (don't have one).

OS Windows 7 x64

BSOD is caused by integrated raid module RMS2LL040 (which is installed on my board, board and raid module bioses are up-to-date), if the following conditions are true:

1. At least one disk(I use SATA) is connected to RMS2LL040 - no disk - no "SAS2 2008 Falcon -StorPort" windows storage controller device - no issue.

2. External PCIe card should be connected to slot6 (electrical x8 physical x16) - other slots no problem, empty slo6 - no problem.

What happens:

After S3 resume - 10min passes and BSOD described above.

10min period is stable - some kind of timeout ?????

Some debug info:

0: kd> !analyze -v

 

*******************************************************************************

 

* *

 

* Bugcheck Analysis *

 

* *

 

*******************************************************************************

 

DRIVER_POWER_STATE_FAILURE (9f)

 

A driver is causing an inconsistent power state.

 

Arguments:

 

Arg1: 0000000000000003, A device object has been blocking an Irp for too long a time

 

Arg2: fffffa80076669c0, Physical Device Object of the stack

 

Arg3: fffff80000b9c3d8, Functional Device Object of the stack

 

Arg4: fffffa8008414470, The blocked IRP

 

 

Debugging Details:

 

------------------

 

DRVPOWERSTATE_SUBCODE: 3

 

IMAGE_NAME: lsi_sas2.sys

 

DEBUG_FLR_IMAGE_TIMESTAMP: 4b50c612

 

MODULE_NAME: lsi_sas2

 

FAULTING_MODULE: fffff880010fd000 lsi_sas2

 

DEFAULT_BUCKET_ID: VISTA_DRIVER_FAULT

 

BUGCHECK_STR: 0x9F

 

PROCESS_NAME: System

 

CURRENT_IRQL: 2

 

 

STACK_TEXT:

 

fffff800`00b9c388 fffff800`02706cd2 : 00000000`0000009f 00000000`00000003 fffffa80`076669c0 fffff800`00b9c3d8 : nt!KeBugCheckEx

 

fffff800`00b9c390 fffff800`026a45fc : fffff800`00b9c4c0 fffff800`00b9c4c0 00000000`00000000 00000000`00000003 : nt! ?? ::FNODOBFM::`string'+0x34a90

 

fffff800`00b9c430 fffff800`026a4496 : fffff800`02835f50 00000000`0000d690 00000000`00000000 00000000`00000000 : nt!KiProcessTimerDpcTable+0x6c

 

fffff800`00b9c4a0 fffff800`026a437e : 00000001`febf983c fffff800`00b9cb18 00000000`0000d690 fffff800`02812488 : nt!KiProcessExpiredTimerList+0xc6

 

fffff800`00b9caf0 fffff800`026a4167 : 00000000`6d892fc3 00000000`0000d690 00000000`6d892ff3 00000000`00000090 : nt!KiTimerExpiration+0x1be

 

fffff800`00b9cb90 fffff800`0269096a : fffff800`0280ee80 fffff800`0281ccc0 00000000`00000001 fffff880`00000000 : nt!KiRetireDpcList+0x277

 

fffff800`00b9cc40 00000000`00000000 : fffff800`00b9d000 fffff800`00b97000 fffff800`00b9cc00 00000000`00000000 : nt!KiIdleLoop+0x5a

 

 

STACK_COMMAND: kb

 

FOLLOWUP_NAME: MachineOwner

 

FAILURE_BUCKET_ID: X64_0x9F_3_disk_IMAGE_lsi_sas2.sys

 

BUCKET_ID: X64_0x9F_3_disk_IMAGE_lsi_sas2.sys

 

Followup: MachineOwner

 

---------

 

Any ideas how to resolve this ?

Regards,

Vilius Mockunas

0 Kudos
idata
Employee
516 Views

Hi,

Unfortunately all drivers I tried demonstrate the same behavior.

thanks,

Vilius

0 Kudos
Reply