Strange error on embedded Geode processor

[Comark Corp]

Any idea what might cause this error? The .bitcfg was set to run for 12 hours and halt on the first error. It appears that it ran the full 12 hours, then halted with this strange error. I've snipped out the irrelevant SMART data.

Thanks.

PassMark BurnInTest Log file - http://www.passmark.com
================================================== ======

BurnInTest V5.1 Pro 1011
Logging detail level: Normal

******************
SYSTEM INFORMATION
******************
Network Name: LE363-EVAL

Date: 07/21/06 10:24:17
Operating system: Windows XP Professional Service Pack 2 build 2600
Number of CPUs: 1
CPU manufacturer: Geode by NSC
CPU type: 586 Gen
CPU features: MMX 3DNow!
CPU1 speed: 400.0 MHz
CPU L2 Cache: Unknown
RAM: 503 MB
Video card: Advanced Micro Devices Windows Graphics Driver (Resolution: 1280x1024x32)
Disk drive: Model HEJ423020F9AT00 (Size: 18.6GB)
Optical drive: TOSHIBA CD/DVDW SD-R6472

******************
DETAILED EVENT LOG
******************
LOG NOTE: 2006-07-21 10:24:17, Status, PassMark BurnInTest V5.1 Pro 1011
LOG NOTE: 2006-07-21 10:24:18, Status, Main Tests started
LOG NOTE: 2006-07-21 10:24:23, Disk, Drive C: SMART attribute data
LOG NOTE: 2006-07-21 10:24:25, Disk, Model HEJ423020F9AT00, Serial# 5J3722
LOG NOTE: 2006-07-21 10:24:25, Disk, Power on hours count: 13745
LOG NOTE: 2006-07-21 10:24:26, Disk, Power cycle count: 76
LOG NOTE: 2006-07-21 10:24:26, Disk, Temperature: 33C
LOG NOTE: 2006-07-21 10:24:26, Disk, All other SMART attributes within thresholds.
LOG NOTE: 2006-07-21 10:49:16, Disk, C, SMART Reallocated Sector

[SNIP more SMART data]
LOG NOTE: 2006-07-21 22:15:41, Disk, C, SMART Reallocated Sector Count start value: 0, current value 0
WARNING : 2006-07-21 22:24:24, CPU - Maths, No operations reported in timeout period
LOG NOTE: 2006-07-21 22:24:24, Status, Test run stopped

**************
RESULT SUMMARY
**************
Test Start time: Fri Jul 21 10:24:17 2006
Test Stop time: Fri Jul 21 22:24:18 2006
Test Duration: 012h 00m 01s

Test Name Cycles Operations Result Errors Last Error
CPU - Maths 12 -40063254 FAIL 1 No operations reported in timeout period
CPU - SIMD 29 1.008 Billion PASS 0 No errors
Memory (RAM) 64 44.929 Billion PASS 0 No errors
2D Graphics 278 278798 PASS 0 No errors
Disk (A: ) 1883 576 Million PASS 0 No errors
Disk (C: ) 29 6.122 Billion PASS 0 No errors
Network 1 756 6.053 Million PASS 0 No errors
CD/DVD (D 21 14.736 Billion PASS 0 No errors
Sound 283 478 Million PASS 0 No errors
Parallel Port 4 1.252 Million PASS 0 No errors
Serial Port 1 384 22.118 Million PASS 0 No errors
Serial Port 2 380 21.906 Million PASS 0 No errors
TEST RUN FAILED

*******************************************
SERIOUS ERROR SUMMARY FOR THE LAST TEST RUN
*******************************************

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

[Ian (PassMark)]

I believe this is a BurnInTest software problem. We will release a new version of BurnInTest later this week that I beleive will resolve this problem. I will update the post at this time.

Best regagards,
Ian

[Comark Corp]

I believe this is a BurnInTest software problem. We will release a new version of BurnInTest later this week that I beleive will resolve this problem. I will update the post at this time.

Best regagards,
Ian

Thank you Ian. The above testing was done under WinXP SP2.
As an afterthought, could this software problem also be the cause of bugchecks under Win2K SP4? I was unable to test this board at all with win2K and get any meaningful results. The tests would begin, and the system would reboot in 5 seconds or less.

I have the minidumps if they would be of any use to you, please let me know.

Thanks again.

From WinDbg:

************************************************** *****************************
* *
* Bugcheck Analysis *
* *
************************************************** *****************************

KMODE_EXCEPTION_NOT_HANDLED (1e)
This is a very common bugcheck. Usually the exception address pinpoints
the driver/function that caused the problem. Always note this address
as well as the link date of the driver/image that contains this address.
Arguments:
Arg1: c0000005, The exception code that was not handled
Arg2: a0157528, The address that the exception occurred at
Arg3: 00000001, Parameter 0 of the exception
Arg4: 0000000c, Parameter 1 of the exception

Debugging Details:
------------------


EXCEPTION_CODE: (NTSTATUS) 0xc0000005 - The instruction at "0x%08lx" referenced memory at "0x%08lx". The memory could not be "%s".

FAULTING_IP:
win32k!BltLnkSrcCopyMsk1+30
a0157528 ?? ???

EXCEPTION_PARAMETER1: 00000001

EXCEPTION_PARAMETER2: 0000000c

WRITE_ADDRESS: unable to read from 8047edb4
unable to read from 8047e864
unable to read from 8047e738
unable to read from 8046a85c
unable to read from 8047e750
unable to read from 8047e860
unable to read from 80471038
unable to read from 8047e8dc
unable to read from 8047ed80
0000000c

DEFAULT_BUCKET_ID: DRIVER_FAULT

BUGCHECK_STR: 0x1E

LAST_CONTROL_TRANSFER: from 0000001e to 804308ad

STACK_TEXT:
f21ce0a4 0000001e c0000005 a0157528 00000001 nt!KiDeliverApc+0x81


FOLLOWUP_IP:
win32k!BltLnkSrcCopyMsk1+30
a0157528 ?? ???

FOLLOWUP_NAME: MachineOwner

SYMBOL_NAME: win32k!BltLnkSrcCopyMsk1+30

MODULE_NAME: win32k

IMAGE_NAME: win32k.sys

DEBUG_FLR_IMAGE_TIMESTAMP: 3eb1ea3c

STACK_COMMAND: kb

BUCKET_ID: 0x1E_W_win32k!BltLnkSrcCopyMsk1+30

Followup: MachineOwner
---------

[Ian (PassMark)]

I don't know if it is related. The access violation crash is at 0xa0157528 in Win32k.sys in function BltLnkSrcCopyMsk1. After a bit of reading, crashes in the module win32k.sys do not seem that uncommon. Eg
W2000 SP3
http://support.microsoft.com/kb/811363/
WXP
http://support.microsoft.com/?kbid=329052

After a bit more reading, my guess is that this system might have Faulty RAM. Please try replacing it.

If it is reproducible on other like systems, I would next suspect a CPU patch problem.

Best regards,
Ian

[Ian (PassMark)]

We have created a debug version that we believe resolves this problem. Could you please try it out and let me know.

The new debug version can be downloaded from:
http://www.passmark.com/ftp/bitpro511012debug2.exe

Thanks.
Ian

[Comark Corp]

We have created a debug version that we believe resolves this problem. Could you please try it out and let me know.

The new debug version can be downloaded from:
http://www.passmark.com/ftp/bitpro511012debug2.exe

Thanks.
Ian

Hello Ian, This seemed to work. However, I went back to version 1011 and the failure is not consistent - the next two 12 hour burn-ins did not cause a repeat of the error.

Also, FWIW, the bugchecks are due to some kind of timing problem on the primary IDE channel. Removing the CD-RW drive corrected the issue. I am working with the board vendor to resolve the issue.

Thanks again.

[Ian (PassMark)]

Thanks for the feedback. Much appreciated.

Best regards,
Ian

[thanh]

CD Version

We have created a debug version that we believe resolves this problem. Could you please try it out and let me know.

The new debug version can be downloaded from:
http://www.passmark.com/ftp/bitpro511012debug2.exe

Thanks.
Ian

Hi,

Can I have this update for a CD version?

Thanks

Thanh