PCIE Device causing System Halt on Shutdown

[ck1111]

Hi all--
I am having problems with
A)System Halt on shutdowns or
B)Machine Check Exception Error.
If B) happens the system auto re-starts (ASR) I assume.
I do not get either of these errors when doing a normal restart.

The PCIE device also works as expected when the system is running.

The Errors happen regardless of the PCIE device driver, is installed or is not installed.

System Info:
HP Proliant DL380G6
Power Supply: 460W
Memory(RAM): 6.00GB
Processor: 1 Intel Xeon E5520 @ 2.27GHz (4 cores)
Storage Controller: Smart Array P410i
BIOS: P62 (5/14/2010)
OS: Windows 2008 Enterprise
Service Pack 2
64 Bit
Microsoft Windows Updates are current.

PCIE Device--Modem MT9234ZPX PCIE

Problem:
(Blue Screen) System Halt on Shutdown with the following:
*** Hardware Malfunction
Call your hardware vendor for support
*** The system has halted ***
or
NMI: Memory/Parity Error, Etc. Etc. Etc.


Integrated Management Log (.IML) Reports:

Uncorrectable PCI Express Error (Embedded Device, Bus 0, Device 8, Function 0, Error Status 0x00004000
Severity: Critical
Class: PCI Bus

Or I sometimes get the following in which the system will reboot instead of Halt:

Uncorrectable Machine Check Exception(Board 0,Processor1,APIC ID 0x00000001,Bank 0x00000005,Status 0xBE000000'00800400,Address 0x00000000'000091FB,Misc 0x00000000'00000000)

Steps Taken:
I have tried both riser positions on the MB and have tried various PCIe expansion slots, all produce the same errors.

I know that the PCIe device is working as it works when the system is booted and I have also tested it in another PC running Windows 2008 Enterprise 64 without the errors.

HP Diagnostics does not report any errors.
I have also tried pulling/swapping ram, this also does not make a difference.


HP BIOS and Driver Updates:

BIOS Update From (P62 3/30/2010) - to - (P62 5/14/2010)
cp012068 HP Online Flash Component for Windows-System Programmable Logic Device (HP Proliant DL380 G6)
cp011328 HP ProLiant Smart Array SAS/SATA Controller Driver for Windows Server 2008 x64 Edition
cp011944 HP Online ROM Flash Component for Windows - HP Integrated Lights-Out 2

cp011799 HP Proliant iLO 2 Management Controller Driver 1.13.0.0 Server 2008 x64

I have noticed that if I make a BIOS advanced power change in PCIE Force generation 1 or generation 2 or even switch it back to Auto that the very next shutdown will succeed, however on the next bootup/shutdown the Halt (bluescreen) or Machine check exception will happen thereafter.

I have also tried disabling the HP Proliant System Shutdown Service and while this did allow me to shut the system down a few times (under 5 times)without the Halt or Exception the Errors are now happening again.

Is anyone else out there experiencing this on these systems?

Any help is very much appreciated.

Thanks

[Jan Soska]

hello,
did not see such error yet.
My sugestion:
1) install latest HP PSP (8.5 I think is latest.)
2) use third party memtest x86 - maybe it will find something.
3) check P410 firmware - I noticed some critical updates.

let us know.

Jan

[Nicolai Rasmussen]

Hi, did you find a solution to this issue? I'm having a similar problem.

[ck1111]

Hello,
Yes, it was found that the PCIe bridge chipset on the device is PCIe 1.0a compliant device.

As explained above (system specs) in Windows 2008 Server I would get a Blue Screen on Shutdown (NMI Parity Error) and or System Restart on Shutdown (Machine Check Exception Error)

The Problem:
There are times when attempting to resume from S1 in a Windows Environment, that the system does not wake up if a PCIe 1.0a-compliant device is in the system. This problem is not caused by the device, but is instead due to a problem with the PCI Express Base 1.0a specification with regards to exiting L1 state. The PCI Express Base 1.1 specification resolved the problem; however, because the chip set comply's with the PCIe Base 1.0a specification, they adhere to the PCIe Base1.0a rules for exiting L1.

Fix:
The issue can be resolved by preventing the Windows operating system from placing the device into a low-power device state (D1, D2, or D3 hot) when suspending the system into S1.
The following instructions tell the operating system to leave the device in D0 when suspending the system.
Copy the lines below starting with "Windows Registry Editor" into a text file ending in .reg
(NoPmCaps.reg, for example).
You can right-click on this file in the Explorer Shell and merge its contents into the registry.
This sets PCI_HACK_NO_PM_CAPS (0x0000000020000000) for VID
104C and DID 8231.

Windows Registry Editor Version 5.00
[HKEY_LOCAL_MACHINESYSTEMCurrentControlSetServicespciParameters]
"104C8231"=hex:00,00,00,20,00,00,00,00

This resolved the issue for us.
So you may want to check the chip vendor specs.

Hope this helps.