[技巧]如何获得某个callstack所在线程的线程号?
检查dump文件的时候, 我们经常会使用下面的命令来获得所有线程上的调用栈
~*e !clrstack
输出结果举例如下:
OS Thread Id: 0x1b30 (15)
Child-SP RetAddr Call Site
0000000034a9ed00 000006424e611fc0 System.Threading.WaitHandle.WaitOne(Int64, Boolean)
0000000034a9ed40 00000642782f174b System.DirectoryServices.Protocols.PartialResultsRetriever.ThreadRoutine()
0000000034a9ed90 000006427838956d System.Threading.ExecutionContext.Run(System.Threading.ExecutionContext, System.Threading.ContextCallback, System.Object)
0000000034a9ede0 000006427f602322 System.Threading.ThreadHelper.ThreadStart()
OS Thread Id: 0x10fc (16)
Unable to walk the managed stack. The current thread is likely not a
managed thread. You can run !threads to get a list of managed threads in
the process
OS Thread Id: 0x26f8 (17)
Child-SP RetAddr Call Site
00000000431bf060 0000064274a56477 System.Threading.WaitHandle.WaitAny(System.Threading.WaitHandle[], Int32, Boolean)
00000000431bf0c0 00000642782f174b System.Net.TimerThread.ThreadProc()
00000000431bf190 000006427838956d System.Threading.ExecutionContext.Run(System.Threading.ExecutionContext, System.Threading.ContextCallback, System.Object)
00000000431bf1e0 000006427f602322 System.Threading.ThreadHelper.ThreadStart()
在这个结果输出里, 可以很轻易地看出线程号15, 16, 17.
然而, 当检查非托管调用栈的时候, 这个线程号却很难看到. 命令:
~*e k
输出片段如下:
Child-SP RetAddr Call Site
00000000`34f1f6d8 00000000`77efb416 ntdll!ZwWaitForSingleObject+0xa [d:\nt\base\…\usrstubs.asm @ 180]
00000000`34f1f6e0 00000000`77efb480 ntdll!RtlpWaitOnCriticalSection+0x240 [d:\nt\….c @ 617]
00000000`34f1f760 00000000`35f7da11 ntdll!RtlEnterCriticalSection+0xa9 [d:\nt\…\resource.c @ 876]
00000000`34f1f790 00000000`35f7dc8b ONETUTIL!COWSHeap::RunLeakDetectionInternal+0x281
00000000`34f1f7e0 00000000`35f7e081 ONETUTIL!COWSMemmgr::AcquireLocalHeap+0x13
00000000`34f1f820 00000000`3561c500 ONETUTIL!COWSMemmgr::InitMemmgrForThread+0x4d
00000000`34f1f880 00000000`3000670d OWSSVR!GetExtensionVersion+0x5b14
00000000`34f1f900 00000000`35febb0d OWSTIMER!VTimerProcessRequest+0xa1 [d:\….cpp @ 225]
00000000`34f1fc30 00000000`35fece4e ONETUTIL!COWSThreadWithHeap::WalkHeap+0x1fd
00000000`34f1fc80 00000000`781337d7 ONETUTIL!COWSThreadWithHeap::Uninitialize+0x97e
00000000`34f1ff20 00000000`78133894 msvcr80!_callthreadstartex+0x17 [f:\dd\vctools\crt_…\crt\src\threadex.c @ 348]
00000000`34f1ff50 00000000`77d6b71a msvcr80!_threadstartex+0x84 [f:\dd\vctools…\crt\src\threadex.c @ 326]
00000000`34f1ff80 00000000`00000000 kernel32!BaseThreadStart+0x3a [d:\nt\…\support.c @ 795]
Child-SP RetAddr Call Site
00000000`35c1f6d8 00000000`77efb416 ntdll!ZwWaitForSingleObject+0xa [d:\nt\base\…\usrstubs.asm @ 180]
00000000`35c1f6e0 00000000`77efb480 ntdll!RtlpWaitOnCriticalSection+0x240 [d:\nt\….c @ 617]
00000000`35c1f760 00000000`35f7da11 ntdll!RtlEnterCriticalSection+0xa9 [d:\nt\base….c @ 876]
00000000`35c1f790 00000000`35f7dc8b ONETUTIL!COWSHeap::RunLeakDetectionInternal+0x281
00000000`35c1f7e0 00000000`35f7e081 ONETUTIL!COWSMemmgr::AcquireLocalHeap+0x13
00000000`35c1f820 00000000`3561c500 ONETUTIL!COWSMemmgr::InitMemmgrForThread+0x4d
00000000`35c1f880 00000000`3000670d OWSSVR!GetExtensionVersion+0x5b14
00000000`35c1f900 00000000`35febb0d OWSTIMER!VTimerProcessRequest+0xa1 [d:\….cpp @ 225]
00000000`35c1fc30 00000000`35fece4e ONETUTIL!COWSThreadWithHeap::WalkHeap+0x1fd
00000000`35c1fc80 00000000`781337d7 ONETUTIL!COWSThreadWithHeap::Uninitialize+0x97e
00000000`35c1ff20 00000000`78133894 msvcr80!_callthreadstartex+0x17 [f:\dd…\crt…threadex.c @ 348]
00000000`35c1ff50 00000000`77d6b71a msvcr80!_threadstartex+0x84 [f:\dd\…\crt\src\t…adex.c @ 326]
00000000`35c1ff80 00000000`00000000 kernel32!BaseThreadStart+0x3a [d:\…base\…support.c @ 795]
那么如何才能得到这个这两个调用栈的线程号码呢?
答案的命令如下:
~*e !teb;k
这个命令的意思是遍历所有的线程, 先输出线程的TEB(Thread Environment Block)数据结构, 再输出调用栈.
根据Mark E. Russinvoich的Windows Internals 5th Edition的P378页的讲解, TEB结构中是存有Thread ID这个信息的, 位置在Active RPC handle的正上方. 书中截图如下:
让我们检查一下在我们的WinDBG中的输出结果吧.
TEB at 000007ffffe66000
ExceptionList: 0000000000000000
StackBase: 0000000034f20000
StackLimit: 0000000034f1c000
SubSystemTib: 0000000000000000
FiberData: 0000000000001e00
ArbitraryUserPointer: 0000000000000000
Self: 000007ffffe66000
EnvironmentPointer: 0000000000000000
ClientId: 0000000000000490 . 00000000000023e8
RpcHandle: 0000000000000000
Tls Storage: 0000000000000000
PEB Address: 000007fffffde000
LastErrorValue: 0
LastStatusValue: c0150008
Count Owned Locks: 0
HardErrorMode: 0
Child-SP RetAddr Call Site
00000000`34f1f6d8 00000000`77efb416 ntdll!ZwWaitForSingleObject+0xa [d:\nt\….asm @ 180]
00000000`34f1f6e0 00000000`77efb480 ntdll!RtlpWaitOnCriticalSection+0x240 [d:\nt\base….c @ 617]
00000000`34f1f760 00000000`35f7da11 ntdll!RtlEnterCriticalSection+0xa9 [d:\nt\base\….c @ 876]
00000000`34f1f790 00000000`35f7dc8b ONETUTIL!COWSHeap::RunLeakDetectionInternal+0x281
00000000`34f1f7e0 00000000`35f7e081 ONETUTIL!COWSMemmgr::AcquireLocalHeap+0x13
00000000`34f1f820 00000000`3561c500 ONETUTIL!COWSMemmgr::InitMemmgrForThread+0x4d
00000000`34f1f880 00000000`3000670d OWSSVR!GetExtensionVersion+0x5b14
00000000`34f1f900 00000000`35febb0d OWSTIMER!VTimerProcessRequest+0xa1 [d:\….cpp @ 225]
00000000`34f1fc30 00000000`35fece4e ONETUTIL!COWSThreadWithHeap::WalkHeap+0x1fd
00000000`34f1fc80 00000000`781337d7 ONETUTIL!COWSThreadWithHeap::Uninitialize+0x97e
00000000`34f1ff20 00000000`78133894 msvcr80!_callthreadstartex+0x17 [f:\dd\….c @ 348]
00000000`34f1ff50 00000000`77d6b71a msvcr80!_threadstartex+0x84 [f:\dd\…\crt\src….c @ 326]
00000000`34f1ff80 00000000`00000000 kernel32!BaseThreadStart+0x3a [d:\nt\base\……c @ 795]
TEB at 000007ffffe64000
ExceptionList: 0000000000000000
StackBase: 0000000035c20000
StackLimit: 0000000035c1c000
SubSystemTib: 0000000000000000
FiberData: 0000000000001e00
ArbitraryUserPointer: 0000000000000000
Self: 000007ffffe64000
EnvironmentPointer: 0000000000000000
ClientId: 0000000000000490 . 000000000000107c
RpcHandle: 0000000000000000
Tls Storage: 0000000000000000
PEB Address: 000007fffffde000
LastErrorValue: 0
LastStatusValue: c0150008
Count Owned Locks: 0
HardErrorMode: 0
Child-SP RetAddr Call Site
00000000`35c1f6d8 00000000`77efb416 ntdll!ZwWaitForSingleObject+0xa [d:\….asm @ 180]
00000000`35c1f6e0 00000000`77efb480 ntdll!RtlpWaitOnCriticalSection+0x240 [d:\nt\…\reso….c @ 617]
00000000`35c1f760 00000000`35f7da11 ntdll!RtlEnterCriticalSection+0xa9 [d:\nt\…\ntdll\…rce.c @ 876]
00000000`35c1f790 00000000`35f7dc8b ONETUTIL!COWSHeap::RunLeakDetectionInternal+0x281
00000000`35c1f7e0 00000000`35f7e081 ONETUTIL!COWSMemmgr::AcquireLocalHeap+0x13
00000000`35c1f820 00000000`3561c500 ONETUTIL!COWSMemmgr::InitMemmgrForThread+0x4d
00000000`35c1f880 00000000`3000670d OWSSVR!GetExtensionVersion+0x5b14
00000000`35c1f900 00000000`35febb0d OWSTIMER!VTimerProcessRequest+0xa1 [d:\…\timerecb…@ 225]
00000000`35c1fc30 00000000`35fece4e ONETUTIL!COWSThreadWithHeap::WalkHeap+0x1fd
00000000`35c1fc80 00000000`781337d7 ONETUTIL!COWSThreadWithHeap::Uninitialize+0x97e
00000000`35c1ff20 00000000`78133894 msvcr80!_callthreadstartex+0x17 [f:\dd\vctools\…\crt\src\t…ex.c @ 348]
00000000`35c1ff50 00000000`77d6b71a msvcr80!_threadstartex+0x84 [f:\dd\vctools\crt_bld\…\crt\src\threadex.c @ 326]
00000000`35c1ff80 00000000`00000000 kernel32!BaseThreadStart+0x3a [d:\nt\…\…c @ 795]
这里的ClientId就应该是线程号了, 长得很奇怪是么? 呵呵, 让我们来找到它们的数字号码吧.
0000000000000490 . 00000000000023e8 = 490.23e8
0000000000000490 . 000000000000107c = 490.107c
输入命令
0:039> ~
0 Id: 490.16ec Suspend: 1 Teb: 000007ff`fffdc000 Unfrozen
1 Id: 490.1884 Suspend: 1 Teb: 000007ff`fffd8000 Unfrozen
2 Id: 490.1200 Suspend: 1 Teb: 000007ff`fffa2000 Unfrozen
3 Id: 490.a10 Suspend: 1 Teb: 000007ff`fff9e000 Unfrozen
…………
26 Id: 490.1c4c Suspend: 1 Teb: 000007ff`ffe90000 Unfrozen
27 Id: 490.1b88 Suspend: 1 Teb: 000007ff`ffe86000 Unfrozen
28 Id: 490.290c Suspend: 1 Teb: 000007ff`ffe72000 Unfrozen
…………34 Id: 490.2b64 Suspend: 1 Teb: 000007ff`ffeec000 Unfrozen
35 Id: 490.2360 Suspend: 1 Teb: 000007ff`ffed8000 Unfrozen
36 Id: 490.a2c Suspend: 1 Teb: 000007ff`fffd6000 Unfrozen
…………62 Id: 490.2b04 Suspend: 0 Teb: 000007ff`ffebc000 Unfrozen
63 Id: 490.2004 Suspend: 1 Teb: 000007ff`ffeba000 Unfrozen
64 Id: 490.24a8 Suspend: 0 Teb: 000007ff`ffeb8000 Unfrozen
…………72 Id: 490.9bc Suspend: 1 Teb: 000007ff`ffea6000 Unfrozen
73 Id: 490.2494 Suspend: 1 Teb: 000007ff`ffea4000 Unfrozen
74 Id: 490.2724 Suspend: 1 Teb: 000007ff`ffea2000 Unfrozen
…………
92 Id: 490.29c8 Suspend: 1 Teb: 000007ff`ffe78000 Unfrozen
93 Id: 490.1da8 Suspend: 1 Teb: 000007ff`ffe76000 Unfrozen
…………98 Id: 490.2650 Suspend: 1 Teb: 000007ff`ffe6a000 Unfrozen
99 Id: 490.23e8 Suspend: 1 Teb: 000007ff`ffe66000 Unfrozen
100 Id: 490.107c Suspend: 1 Teb: 000007ff`ffe64000 Unfrozen
101 Id: 490.1b98 Suspend: 1 Teb: 000007ff`ffe62000 Unfrozen
102 Id: 490.2bd4 Suspend: 1 Teb: 000007ff`ffe60000 Unfrozen
……………108 Id: 490.a5c Suspend: 0 Teb: 000007ff`ffe4e000 Unfrozen
109 Id: 490.980 Suspend: 1 Teb: 000007ff`ffe4a000 Unfrozen
我们终于找到了, 线程号码为99和100.
验证一下吧
0:099> ~99s;k
ntdll!ZwWaitForSingleObject+0xa:
00000000`77ef047a c3 ret
Child-SP RetAddr Call Site
00000000`34f1f6d8 00000000`77efb416 ntdll!ZwWaitForSingleObject+0xa [d:…asm @ 180]
00000000`34f1f6e0 00000000`77efb480 ntdll!RtlpWaitOnCriticalSection+0x240 [d:….c @ 617]
00000000`34f1f760 00000000`35f7da11 ntdll!RtlEnterCriticalSection+0xa9 [d:\nt\….c @ 876]
00000000`34f1f790 00000000`35f7dc8b ONETUTIL!COWSHeap::RunLeakDetectionInternal+0x281
00000000`34f1f7e0 00000000`35f7e081 ONETUTIL!COWSMemmgr::AcquireLocalHeap+0x13
00000000`34f1f820 00000000`3561c500 ONETUTIL!COWSMemmgr::InitMemmgrForThread+0x4d
00000000`34f1f880 00000000`3000670d OWSSVR!GetExtensionVersion+0x5b14
00000000`34f1f900 00000000`35febb0d OWSTIMER!VTimerProcessRequest+0xa1 [d:…@ 225]
00000000`34f1fc30 00000000`35fece4e ONETUTIL!COWSThreadWithHeap::WalkHeap+0x1fd
00000000`34f1fc80 00000000`781337d7 ONETUTIL!COWSThreadWithHeap::Uninitialize+0x97e
00000000`34f1ff20 00000000`78133894 msvcr80!_callthreadstartex+0x17 [f:\…\…@ 348]
00000000`34f1ff50 00000000`77d6b71a msvcr80!_threadstartex+0x84 [f:\dd….c @ 326]
00000000`34f1ff80 00000000`00000000 kernel32!BaseThreadStart+0x3a [d:\….c @ 795]
