Debugger Deadlock with Win2000 
Author Message
 Debugger Deadlock with Win2000

Whilst debugging some C++ code in visual studio, and trying to interrupt it,
my de{*filter*} pauses a few seconds and then in the debug window prints
DBG: Break command failed within 3 seconds.
DBG: Potential deadlock. Soft broken.

At this point, attempts to continue or single step result in nothing
happening. The de{*filter*} thinks the program is running, but Task manager
shows no CPU consumption. Pressing "Break Execution" again, results in the
same message, with the instuction pointer in the same place. I can restart
debugging, and it all works again, until I hit "Break Execution."

Further details:
I can't find anything on MSDN.

The code in question is not multi-threaded, or using any obvious resources,
or doing anything that might cause deadlock.

This problem only appeared when I moved to a new faster computer running
Windows2000.   In all other respects the setups of the old and new PC were
similar.
The problem occurs even if I build a new program
while(true) i++;
from scratch.

I've upgraded to VC6 Service Pack 4 (from SP3), without success.

So the questions.
Does anybody know what the debug message (about deadlock) actually means?
Has anyone else had a similar problem?
Is there a fix?



Sun, 29 Dec 2002 03:00:00 GMT  
 Debugger Deadlock with Win2000
I have experienced the same problem. But for me, Break Execution sometimes
works correctly, but sometimes behaves as you describe.


Quote:
> Whilst debugging some C++ code in visual studio, and trying to interrupt
it,
> my de{*filter*} pauses a few seconds and then in the debug window prints
> DBG: Break command failed within 3 seconds.
> DBG: Potential deadlock. Soft broken.

> At this point, attempts to continue or single step result in nothing
> happening. The de{*filter*} thinks the program is running, but Task manager
> shows no CPU consumption. Pressing "Break Execution" again, results in the
> same message, with the instuction pointer in the same place. I can restart
> debugging, and it all works again, until I hit "Break Execution."

> Further details:
> I can't find anything on MSDN.

> The code in question is not multi-threaded, or using any obvious
resources,
> or doing anything that might cause deadlock.

> This problem only appeared when I moved to a new faster computer running
> Windows2000.   In all other respects the setups of the old and new PC were
> similar.
> The problem occurs even if I build a new program
> while(true) i++;
> from scratch.

> I've upgraded to VC6 Service Pack 4 (from SP3), without success.

> So the questions.
> Does anybody know what the debug message (about deadlock) actually means?
> Has anyone else had a similar problem?
> Is there a fix?



Mon, 30 Dec 2002 03:00:00 GMT  
 Debugger Deadlock with Win2000
I think this also happens on NT. You may want to set int 3 or DebugBreak in
your threads instead of VC++ trying to break and timing out.


Mon, 30 Dec 2002 03:00:00 GMT  
 Debugger Deadlock with Win2000
Same scenario (VC SP4, Win2K), same problem.  And I haven't found anything
on MSDN or the knowledge base either.

It does work sometimes, but the problem happens quite frequently.

Some word from the MS gurus would be much appreciated...

--
Dave



Sat, 04 Jan 2003 03:00:00 GMT  
 
 [ 4 post ] 

 Relevant Pages 

1. Debugger Deadlock with Win2000

2. VC 6.0-Debugger newer find the Win2000 DBG-Files

3. VC 6.0-Debugger never finds the Win2000 DBG-Files

4. Some local variables hidden from debugger (Win2000 Kernel debugging)

5. Debugger sometimes unresponsive under WIN2000

6. "potential deadlock" in debugger trace.

7. PLEASE HELP: I AM NEW TO DEBUGGER (Turbo Debugger 3.0)

8. Debugger vs. not debugger difference

9. Debugger question - debugger came up in assembly

10. How do I debug aspnet_wp.exe deadlocks?

11. GDI+ CImage deadlock

12. multiserver deadlock detection problem

 

 
Powered by phpBB® Forum Software