> I can send a patch for this if interested. Has anyone else implemented
> this?
>
At some point we had in the tree where when a thread called the exit system
call, the system would check the threadcontexts in the system and only
called "exitSimLoop" if there was only 1 thread active in the system.

I think Steve (?) removed it on a code update given that it's not
necessarily clear that counting the number of threads that are active is a
good indicator of if the workload calling exit should really exit or not.
I'm not sure about the reason, so maybe Steve could comment or a double
check of the archives is in order because I do remember a similar
functionality to this inside of M5 previously.

I dont think there was anything about "freezing" stats though. So with
regards to that patch, I guess it wouldnt hurt to post it on reviewboard
though so that people can comment on it and potentially get it into the
tree.
-- 
- Korey
_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users

Reply via email to