Unsolved
6 Posts
0
699
January 2nd, 2023 01:00
XPS 9520, random BSOD, memory corruption caused by USB storage
I have issues with BSODs on DELL XPS 9520 and external usb storage devices. My computer has 64 Gb of system memory, Windows 11 22H1 22000.1335, BIOS 1.7.0, latest drivers. Memtest86 for 8 hours, and Windows memory test for 2 hours - both are OK.
I have noticed then when I do read 30-50 gigabytes of data from usb external drive or card reader, then system memory may get corrupted. Computer does not crash immediately every time after memory got corrupted. Browser or other programs may begin to crash firstly. Then in minute or two after this, if you continue to use computer, there will be obligatory BSOD with random error codes: PFN_LIST_CORRUPT, CACHE_MANAGER, MEMORY_MANAGEMENT, DPC WATCHDOG VIOLATION.
I have reproduced this problem deliberately many times. Key factor is reading/writing a lot of data via usb3. It seems that random drivers are crashing on memory/cache operations, depending on what memory region has been corrupted. Does anyone have the same problems? Any suggestions?
lexa-a128
6 Posts
0
January 2nd, 2023 01:00
I have tried to debug BSOD crash dumps and got following stacks: (for 4 different crashes)
nt!KeBugCheckEx
nt!MiMakeProtoTransition+0x245fdc
nt!MiInsertPagesInList+0x33d
nt!MmUnmapViewInSystemCache+0x429
nt!CcUnmapVacb+0x63
nt!CcUnmapVacbArray+0x1d4
nt!CcUnmapAndPurge+0x36
nt!CcDeleteSharedCacheMap+0x1a0
nt!CcWriteBehindInternal+0x48c
nt!CcWriteBehind+0xb0
nt!CcCachemapUninitWorkerThread+0x112
nt!ExpWorkerThread+0x14f
nt!PspSystemThreadStartup+0x55
nt!KiStartSystemThread+0x34
.
.
nt!KeBugCheckEx
nt!KeAccumulateTicks+0x20f
nt!KiUpdateRunTime+0x61
nt!KeClockInterruptNotify+0x11a
nt!HalpTimerClockIpiRoutine+0x16
nt!KiCallInterruptServiceRoutine+0xa0
nt!KiInterruptSubDispatchNoLockNoEtw+0xfa
nt!KiInterruptDispatchNoLockNoEtw+0x37
nt!MiFlushSectionInternal+0x293
nt!MmFlushSection+0x15a
nt!CcFlushCachePriv+0x471
nt!CcCoherencyFlushAndPurgeCache+0x7f
Ntfs!NtfsCoherencyFlushAndPurgeCache+0x55
Ntfs!NtfsFlushUserStream+0xdf
Ntfs!NtfsFlushVolumeFlushSingleFcb+0x4f8
Ntfs!NtfsFlushVolumeMultiThreadFcbWalkerWorker+0x2a8
nt!ExpWorkerThread+0x14f
nt!PspSystemThreadStartup+0x55
nt!KiStartSystemThread+0x34
.
.
nt!MiGetPage+0x6d
nt!MiMakePageAvoidRead+0x179
nt!MmCopyToCachedPage+0x219
nt!CcMapAndCopyInToCache+0x400
nt!CcCopyWriteEx+0x168
Ntfs!NtfsCopyWriteA+0x5ed
FLTMGR!FltpPerformFastIoCall+0x163
FLTMGR!FltpPassThroughFastIo+0x10a
FLTMGR!FltpFastIoWrite+0x15c
nt!NtWriteFile+0x424
nt!KiSystemServiceCopyEnd+0x25
0x00007ff9`bd4e4134
.
.
nt!KeBugCheckEx
nt!MiUnlinkPageFromListEx+0x21b7b7
nt!MmCheckCachedPageStates+0xa46
nt!CcFetchDataForRead+0x95
nt!CcMapAndCopyFromCache+0xf2
nt!CcCopyReadEx+0x1bf
Ntfs!NtfsCopyReadA+0x40a
FLTMGR!FltpPerformFastIoCall+0x163
FLTMGR!FltpPassThroughFastIo+0x10a
FLTMGR!FltpFastIoRead+0x15c
nt!IopReadFile+0x2d4
nt!NtReadFile+0xc9
nt!KiSystemServiceCopyEnd+0x25
0x00007ffa`518240f4
John harper
2 Intern
•
346 Posts
0
January 8th, 2023 22:00
Application, operating system, and driver bugs can all cause flash drive corruption by preventing files from being saved correctly or damaging the integrity of the file system. Files that have disappeared from a flash drive corrupted by software bugs can sometimes be recovered using the ATTRIB command.
lexa-a128
6 Posts
0
January 16th, 2023 05:00
I've performed further investigation, and it turns out that this problem is not related to usb3.
I've found the way to reproduce the problem without reading data via usb:
1) Power off computer for several hours. (One hour off was not enough to reproduce problem.)
2) Turn on computer and boot Windows 11.
3) Run program that allocates 60 Gb of memory, which I wrote it myself.
4) There will be one of same BSODs instantly.
Then I've run Memtest86 right after cold start of laptop (which was powered off for several hours),
Only then tests (both BIOS diagnostic and memtest86) showed errors in memory.
It was useless to run tests for many hours - they did not show errors. Key was to run it just right
after cold start, after computer is off for several hours.
Now I've changed memory modules to new ones and testing (same Samsung DDR5-4800 2x32Gb)
lexa-a128
6 Posts
0
January 16th, 2023 06:00
Random only test of memtest86 showing errors with old modules: