Hubertus Franke wrote:


Marc E. Fiuczynski wrote:

http://kerneltrap.org/node/view/4111

Con Kovalis' work appears relevant to CKRM. I say this because I do not
believe the current CKRM cpu scheduler will be the final/only one that will
work with CKRM.


Hubertus & Haoqiang, what are your thoughts on this?

Marc



Marc, I just looked at the patch. In principle, I have no objection in going with something like that way.

There is some minor overhead ( function call overheads ), but
they should be negligable.

If something like this gets accepted into the mainline, we
could move the scheduler over.

In our scheduling effort we have tried to stay close to the
the O(1) scheduling behavior and calling interface.
Thus I would not expect that to take a tremendously long time
since the O(1) scheduler was integrated as u.ingosched.

However, we can't fight fires at all fronts. Right now we need
to fix the load_balancing/weight adjustment of the current
ckrm scheduler.
If other schedulers work out better, so be it. I am for progression.

I'm in the process of merging the PlugSched patches with your latest CKRM patches for 2.6.10 and I have a question regarding one of your macros which I believe may not be doing what it's name implies:


#define task_is_running(p)      (this_rq() == task_rq(p))

I.e. the name implies that the macro returns true if task p is running BUT it will actually return true if task p is assigned to the same cpu as the currently running task regardless of whether p is running (i.e. on a run queue) or not.

Is that what was intended?

Peter
--
Peter Williams                                   [EMAIL PROTECTED]

"Learning, n. The kind of ignorance distinguishing the studious."
 -- Ambrose Bierce


------------------------------------------------------- This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting Tool for open source databases. Create drag-&-drop reports. Save time by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc. Download a FREE copy at http://www.intelliview.com/go/osdn_nl _______________________________________________ ckrm-tech mailing list https://lists.sourceforge.net/lists/listinfo/ckrm-tech

Reply via email to