[K42-discussion] problem of tracing my event in application code

[Dilma DaSilva]

Hello Tao,

Bob Wisniewski is the expert on k42's tracing. He'll be back from
vacation tomorrow and I'm sure you'll get an answer from him.
In the meantime, let's try to gather more information about what is
going on:

(1) You said you're setting the trace mask to 0x10 through
/ksys/traceMask. After doing that,  verify that the mask is
set by going to the console and typing T|G
This should print in the console:
 test [?]> T|G
T|G
vp 0 trace mask 0x10

You mentioned that you inspected the trace file. 
Were the other User Events in your trace?  If not, than the trace mask 
was not properly set.

(2) I'm assuming that you used "traceTool" to generate a text view of the trace. Make sure
that you're using an up-to-date traceTool. You did a full_snapshot, so 
it should be ok. For people who keep multiple trees, it may happen
that the person rebuilds one of them and is using traceTool from
another, not current, tree.

(3) Any chance that your application ran so long that the trace buffer has 
overflown and your ServicEThread event has been dropped? When you dump 
the trace buffer into a file, msgs are printed in the console; if
there were "dropped" events, something like this is printed:

(4) Remember that if you're running with more than one processor, one
.trc file is output for each processor, and you have to inspect all of 
them to make sure the event has not been generated.

Dilma