H2SO4

More Debugger Trickery For The Interested

A minidump contains 3 main items of information:

1) The stack of the thread which directly caused the crash. This can be viewed with the various 'k' (stack unwind) commands:

1: kd> kL
Child-SP RetAddr Call Site
fffffa60`09cd0528 fffff800`01cb60ee nt!KeBugCheckEx
fffffa60`09cd0530 fffff800`01cb5abc nt!KiBugCheckDispatch+0x6e
fffffa60`09cd0670 fffff800`01cc96bd nt!KiSystemServiceHandler+0x7c
fffffa60`09cd06b0 fffff800`01cd0cff nt!RtlpExecuteHandlerForException+0xd
fffffa60`09cd06e0 fffff800`01c8dd83 nt!RtlDispatchException+0x22f
fffffa60`09cd0dd0 fffff800`01cb61a9 nt!KiDispatchException+0xc3
fffffa60`09cd13d0 fffff800`01cb4d8d nt!KiExceptionDispatch+0xa9
fffffa60`09cd15b0 fffff960`0011c947 nt!KiGeneralProtectionFault+0xcd
fffffa60`09cd1740 fffff960`00121e2d win32k!FindTimer+0x57
fffffa60`09cd1790 fffff800`01cb5df3 win32k!NtUserKillTimer+0x5d
fffffa60`09cd17d0 00000000`7790c24a nt!KiSystemServiceCopyEnd+0x13

2) The processor register context of that thread:

1: kd> r
rax=fffffa6009cd0630 rbx=fffffa6009cd17d0 rcx=000000000000003b
rdx=00000000c0000005 rsi=fffff80001cb5df3 rdi=fffff80001e64df4
rip=fffff80001cb6350 rsp=fffffa6009cd0528 rbp=fffffa6009cd1508
r8=fffff9600011c947 r9=fffffa6009cd0ee0 r10=0000000000000000
r11=0000000000000001 r12=fffffa6009cc4000 r13=fffffa6009cd4000
r14=fffff80001c61000 r15=fffff80001daf4ec
iopl=0 nv up ei ng nz na po nc
cs=0010 ss=0018 ds=002b es=002b fs=0053 gs=002b efl=00000286
nt!KeBugCheckEx:
fffff800`01cb6350 48894c2408 mov qword ptr [rsp+8],rcx ss:0018:fffffa60`09cd0530=000000000000003b

In this case, the 64-bit instruction pointer (RIP) was in nt!KeBugCheckEx at the time of the crash, which is not surprising given that's the "bluescreen" function.

3) A list of loaded modules at the time of the crash. The 'lm' (list modules) command can show these:

1: kd> lm
start end module name
fffff800`01c1b000 fffff800`01c61000 hal (deferred)
fffff800`01c61000 fffff800`02179000 nt (pdb symbols) c:\symcache\ntkrnlmp.pdb\149C563625CA49CEA2881CEDF5D55CCF2\ntkrnlmp.pdb
fffff960`00050000 fffff960`00301000 win32k (pdb symbols) c:\symcache\win32k.pdb\97A727330C184A9B9E1BDA0C3293AA142\win32k.pdb
fffff960`00410000 fffff960`0041a000 TSDDD (deferred)
fffff960`00620000 fffff960`00631000 cdd (deferred)
...

In the example above, the debugger has only encountered "nt" (NTOSKRNL itself) and win32k.sys code in the stack, which is why it has downloaded symbols (they have a PDB extension) from the MS symbol server for those two binaries, but not for hal.dll, TSDDD.dll, and cdd.dll. Their symbol status is listed as "deferred".

Even More Debugger Trickery For The Really Interested

Minidumps can also easily reveal basic information about the Windows version, service pack level, time of the crash, and system uptime:

1: kd> vertarget
Windows Server 2008/Windows Vista Kernel Version 6001 (Service Pack 1) MP (4 procs) Free x64
Product: WinNt, suite: TerminalServer SingleUserTS
Built by: 6001.18145.amd64fre.vistasp1_gdr.080917-1612
Machine Name:
Kernel base = 0xfffff800`01c61000 PsLoadedModuleList = 0xfffff800`01e26db0
Debug session time: Thu Apr 2 07:07:55.076 2009 (GMT+11)
System Uptime: 0 days 0:02:00.122

Say you've identified a driver that you believe to be responsible, and now you want to see more particulars about that binary. Use 'lmvm' with the module name:

1: kd> lmvm tdrpm147
start end module name
fffffa60`0140c000 fffffa60`01590000 tdrpm147 (deferred)
Image path: \SystemRoot\system32\DRIVERS\tdrpm147.sys
Image name: tdrpm147.sys
Timestamp: Mon Oct 13 21:14:16 2008 (48F31F78)
CheckSum: 0018472C
ImageSize: 00184000
Translations: 0000.04b0 0000.04e4 0409.04b0 0409.04e4

To get more info on the processor(s):

1: kd> !cpuinfo
CP F/M/S Manufacturer MHz PRCB Signature MSR 8B Signature Features
1 6,15,11 GenuineIntel 2405 000000b600000000 20193ffe
Cached Update Signature 000000b600000000
Initial Update Signature 000000b600000000


To see the writeup for a given bugcheck code:

1: kd> !analyze -show D1
DRIVER_IRQL_NOT_LESS_OR_EQUAL (d1)
An attempt was made to access a pageable (or completely invalid) address at an
interrupt request level (IRQL) that is too high. This is usually
caused by drivers using improper addresses.
If kernel debugger is available get stack backtrace.
Arguments:
Arg1: 0000000000000000, memory referenced
Arg2: 0000000000000000, IRQL
Arg3: 0000000000000000, value 0 = read operation, 1 = write operation
Arg4: 0000000000000000, address which referenced memory

And the most useful command of all (try it):

windbg> .hh

=============================================

Full dumps allow for much more meaningful analysis but they're hundreds of MB or even several GB in size, depending on the dump type, and for a 2-minute look at someone's BSoD problem a minidump is perfectly sufficient. If their system is going up and down like a yo-yo and the dumps all point at the same 3rd-party driver - bingo. Otherwise, if virtually every dump points at a different culprit and they mostly look odd, chances are it's a hardware issue.