A person I work with has been trying to use the remote debugging
capabilites of WinDBG and running into weird problems. In his
words:
> Often, after issuing a .reboot command, when the system comes back and
> breaks due to an int 3 in my code, I see the following output in the
> debugger:
>
> ---------
>
> Kernel base = 0x80a02000 PsLoadedModuleList = 0x80bc2810
> System Uptime: not available
> c0000005 Exception in debugger client ”Ë callback.
> PC: 00000000 VA: 00000000 R/W: 0 Parameter: 0001003f
> P3.sys: Using TimeOutRetry value of 1 minutes (60 seconds)
> KD: Accessing
‘x:\windows\drivers\console\objchk_wnet_x86\i386\egencon.sys’
> (\SystemRoot\system32\DRIVERS\egencon.sys)
> File size 11K…
> MmLoadSystemImage: Pulled \SystemRoot\system32\DRIVERS\egencon.sys from
kd
> EgenTest: Timestamp d:\source\windows\drivers\console\egencon.c - Last
source modification Wed Sep 4 17:31:03 2002 Compiled Sep 4 2002
17:34:12
> Break instruction exception - code 80000003 (first chance)
> egencon!EgenConAddDevice+7:
> f81a0817 cc int 3
> c0000005 Exception in debugger client † callback.
> PC: 5f08458b VA: 5f08458b R/W: 0 Parameter: 0001003f
>
> ----------
>
> Afterwards, if I step (F10) or go (F5), my remote debugger claims the
> debugee is not connected. On the debug server, it claims the debuggee
is
> running. However if I close both versions of the debugger and reopen
them I
> find it is still at the original int 3 in my code and as an added bonus
I
> can no longer get symbols.
Any suggestions? Anyone else seen this problem?
Don Burn
Egenera, Inc
Do you know what version of the debugger this is from? If it is not 6.0.17 he/she should upgrade.
If it is 6.0.17 you can ask the person to attach the debugger to the debugger before the .reboot. When the exception occurs he/she can then do a ".dump /mfh c0000005.dmp and ZIP or CAB it up and send to xxxxx@microsoft.com.
For example, assuming windbg is already running and attached to the target:
tlist | findstr /I windbg
cdb -p
[wait for exception to occur]
.dump /mfh c0000005.dmp
Don’t send an uncompressed dump, though…
-----Original Message-----
From: xxxxx@acm.org [mailto:xxxxx@acm.org]
Sent: Friday, September 06, 2002 8:46 AM
To: Kernel Debugging Interest List
Subject: [windbg] Remote Debugging problem with WinDBG
A person I work with has been trying to use the remote debugging capabilites of WinDBG and running into weird problems. In his
words:
>> Often, after issuing a .reboot command, when the system comes back
>> and breaks due to an int 3 in my code, I see the following output in
>> the
>> debugger:
>>
>> ---------
>>
>> Kernel base = 0x80a02000 PsLoadedModuleList = 0x80bc2810 System
>> Uptime: not available c0000005 Exception in debugger client ??
>> callback.
>> PC: 00000000 VA: 00000000 R/W: 0 Parameter: 0001003f
>> P3.sys: Using TimeOutRetry value of 1 minutes (60 seconds)
>> KD: Accessing
‘x:\windows\drivers\console\objchk_wnet_x86\i386\egencon.sys’
>> (\SystemRoot\system32\DRIVERS\egencon.sys)
>> File size 11K…
>> MmLoadSystemImage: Pulled \SystemRoot\system32\DRIVERS\egencon.sys
>> from
kd
>> EgenTest: Timestamp d:\source\windows\drivers\console\egencon.c -
>> Last
source modification Wed Sep 4 17:31:03 2002 Compiled Sep 4 2002 17:34:12
>> Break instruction exception - code 80000003 (first chance)
>> egencon!EgenConAddDevice+7:
>> f81a0817 cc int 3
>> c0000005 Exception in debugger client ? callback.
>> PC: 5f08458b VA: 5f08458b R/W: 0 Parameter: 0001003f
>>
>> ----------
>>
>> Afterwards, if I step (F10) or go (F5), my remote debugger claims the
>> debugee is not connected. On the debug server, it claims the debuggee
is
>> running. However if I close both versions of the debugger and reopen
them I
>> find it is still at the original int 3 in my code and as an added
>> bonus
I
>> can no longer get symbols.
Any suggestions? Anyone else seen this problem?
Don Burn
Egenera, Inc
—
You are currently subscribed to windbg as: xxxxx@microsoft.com To unsubscribe send a blank email to %%email.unsub%%
> Do you know what version of the debugger this is from? If it is not =
6.0.17 he/she should upgrade.
If it is 6.0.17 you can ask the person to attach the debugger to the =
debugger before the .reboot. When the exception occurs he/she can then =
do a ".dump /mfh c0000005.dmp and ZIP or CAB it up and send to =
xxxxx@microsoft.com.
For example, assuming windbg is already running and attached to the =
target:
tlist | findstr /I windbg
cdb -p
> [wait for exception to occur]
> .dump /mfh c0000005.dmp
>
> Don’t send an uncompressed dump, though…
>
This is 6.0.17 using the debugging server capabilites (i.e. .server and
connect to remote session). The behavior is erratic, of course this is
just what you do not want while trying to debug.
Don Burn
Egenera, Inc.
You could always run with a debugger attached to windbg just in case it
happens again.
-----Original Message-----
From: xxxxx@acm.org [mailto:xxxxx@acm.org]
Sent: Friday, September 06, 2002 11:21 AM
To: Kernel Debugging Interest List
Subject: [windbg] RE: Remote Debugging problem with WinDBG
Do you know what version of the debugger this is from? If it is not =
6.0.17 he/she should upgrade.
If it is 6.0.17 you can ask the person to attach the debugger to the =
debugger before the .reboot. When the exception occurs he/she can
then = do a ".dump /mfh c0000005.dmp and ZIP or CAB it up and send to
= xxxxx@microsoft.com.
For example, assuming windbg is already running and attached to the =
target:
tlist | findstr /I windbg
cdb -p
> [wait for exception to occur]
> .dump /mfh c0000005.dmp
>
> Don’t send an uncompressed dump, though…
>
This is 6.0.17 using the debugging server capabilites (i.e. .server and
connect to remote session). The behavior is erratic, of course this is
just what you do not want while trying to debug.
Don Burn
Egenera, Inc.
—
You are currently subscribed to windbg as: xxxxx@microsoft.com To
unsubscribe send a blank email to %%email.unsub%%