[ofw] svn.2079 checkins and hung dapltest.exe server process

Smith, Stan stan.smith at intel.com
Fri Apr 17 13:10:01 PDT 2009


Fab Tillier wrote:
>> Please note that once the process can't die you can break in with a
>> kernel debugger (local or remote) and dump the zombie process. This
>> will give us a much better understanding of where the problem is.
>
> Note also that if you run with debug drivers, and there's a ref
> counting issue, IBAL will eventually assert that destruction is
> taking too long.
>
> In the release drivers, synchronous destruction blocks forever.
>
> -Fab

Running with debug ibbus/al + windbg there are no timeouts after 15 minutes of waiting for the hung process to do something; process state was left unaltered (no Control-C, etc.).
Using windbg to attach to the hung dapl2test server process, it shows all thread suspended.

Using dapl2test server with a socket CM interface instead ibal-CM, the dapl2test server process exits cleanly as it did in the past.

SWAG........lost IRP when releasing ibal-CM CEPs?
How to verify?

Your thoughts?

Thanks,

Stan.



More information about the ofw mailing list