Blocked thread, historical information, execution residue, hidden exception, dynamic memory corruption, incorrect stack trace and not my version: pattern cooperation

One application was hanging after printing and its user dump was taken. Inspection of threads reveals only one blocked thread:

0:000> ~*kL

.  0  Id: 486c.5704 Suspend: 1 Teb: 7ffa9000 Unfrozen
ChildEBP RetAddr 
01c0fb68 7c827cfb ntdll!KiFastSystemCallRet
01c0fb6c 77e6202c ntdll!NtWaitForMultipleObjects+0xc
01c0fc14 77e62fbe kernel32!WaitForMultipleObjectsEx+0x11a
01c0fc30 00439f16 kernel32!WaitForMultipleObjects+0×18
01c0fd94 77bcaed6 App!~MyPrinter+0xe
01c0fdc0 77bcaf82 msvcrt!_cinit+0xad
01c0fdd0 77baf582 msvcrt!_cexit+0xb
01c0fde0 7c81a352 msvcrt!__CRTDLL_INIT+0×15c
01c0fe00 7c830e70 ntdll!LdrpCallInitRoutine+0×14
01c0feb8 77e668a3 ntdll!LdrShutdownProcess+0×182

We see that the process was being terminated and the last thread was blocked in a destructor of MyPrinter class. However we see some historical information, faultrep.dll unloaded, and this means that some exception happened prior to the process shutdown:

0:000> lm
Unloaded modules:
69500000 69517000   faultrep.dll

Then we look at the thread raw stack to see any signs of execution residue and hidden exceptions in particular:

0:000> !teb
TEB at 7ffa9000
    ExceptionList:        01c0fc04
    StackBase:            01c10000
    StackLimit:           01c08000
    SubSystemTib:         00000000
    FiberData:            00001e00
    ArbitraryUserPointer: 00000000
    Self:                 7ffa9000
    EnvironmentPointer:   00000000
    ClientId:             0000486c . 00005704
    RpcHandle:            00000000
    Tls Storage:          00000000
    PEB Address:          7ffd9000
    LastErrorValue:       0
    LastStatusValue:      c0000010
    Count Owned Locks:    0
    HardErrorMode:        0

0:000> dds 01c08000 01c10000
01c08000  00000000
01c08004  00000000
01c08008  00000000
01c0e704  00001000
01c0e708  01c0ea14
01c0e70c  7c82855e ntdll!KiUserExceptionDispatcher+0xe
01c0e710  01c0a000
01c0e714  01c0e73c
01c0e718  01c0e720
01c0e71c  01c0e73c
01c0e720  c0000005
01c0e724  00000000
01c0e728  00000000
01c0e72c  7c82a756 ntdll!RtlpCoalesceFreeBlocks+0×370
01c0e730  00000002
01c0e734  00000000

Applying exception context we see heap corruption:

0:000> .cxr 01c0e73c
eax=01998db0 ebx=00370000 ecx=01998db0 edx=4d445443 esi=01998da8 edi=01998010
eip=7c82a756 esp=01c0ea08 ebp=01c0ea14 iopl=0 nv up ei pl zr na pe nc
cs=001b ss=0023 ds=0023 es=0023 fs=003b gs=0000    efl=00010246
7c82a756 3b4a04  cmp    ecx,dword ptr [edx+4] ds:0023:4d445447=????????

However stack trace a bit incorrect at the end (shown in magenta) but this can be explained that it is a residual stack trace that happened in the past with some data overwritten by exception processing:

0:000> kL
  *** Stack trace for last set context - .thread/.cxr resets it
ChildEBP RetAddr 
01c0ea14 7c82a82b ntdll!RtlpCoalesceFreeBlocks+0x370
01c0eafc 77bbcef6 ntdll!RtlFreeHeap+0x38e
01c0eb44 00421fdc msvcrt!free+0xc3
01c0eb6c 00422395 App!SetData+0x3a5
01c0f81c 7c829f60 App!CreatePrinter+0x6a
01c0f824 7c829f59 ntdll!CheckHeapFillPattern+0×64
01c0f948 7c82770b ntdll!RtlFreeHeap+0×70f
01c0f96c 77c27631 ntdll!NtQueryValueKey+0xc
01c0fab8 00000000 gdi32!ReadDisableMetaFilesRegKey+0xfc

Because the hang happens after printing we detect old printer drivers in the module list (certainly not my version of the environment used for application testing):

0:000> lmt
start    end        module name
6b900000 6b925000   PrinterDriverA  [...] 2002
6c900000 6c947000   PrinterDriverB  […] 2002

- Dmitry Vostokov @ -

Leave a Reply

You must be logged in to post a comment.