[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH v3] Xen sched: Fix multiple runqueues in credit2



>>> On 10.02.14 at 10:52, Dario Faggioli <dario.faggioli@xxxxxxxxxx> wrote:
> On Mon, 2014-02-10 at 08:52 +0000, Jan Beulich wrote:
>> >>> On 09.02.14 at 02:57, Justin Weaver <jtweaver@xxxxxxxxxx> wrote:
>> > @@ -1959,15 +1961,25 @@ static void init_pcpu(const struct scheduler *ops, 
> int cpu)
>> >          return;
>> >      }
>> >  
>> > -    /* Figure out which runqueue to put it in */
>> > +    /*
>> > +     * Choose which run queue to add cpu to based on its socket.
>> > +     * If it's CPU 0, hard code it to run queue 0 (it doesn't get a 
>> > STARTING
>> > +     * callback and socket information is not yet available for it).
>> 
>> Did you verify that last part to be the case? Because if so, we would
>> probably be better off fixing the initialization ordering.
>> 
> Last part == "socket information is not yet available" ? If yes, yes, at
> least on my system, cpu_to_socket() always return 0 (or, if I statically
> initialize the array to -1, it always return -1) at that time, and I
> have CPU0 on socket 1, so I'm quite sure that is the case.

Okay.

> By fixing the init order, do you mean moving whatever does the
> cpu-to-socket mapping before scheduler's initialization?

Yes (or vice versa), if reasonably possible.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.