Re: [PATCH v8]xen: sched: convert RTDS from time to event driven model

[Dario Faggioli <dario.faggioli@citrix.com>]

[-- Attachment #1.1: Type: text/plain, Size: 2246 bytes --]

On Fri, 2016-03-11 at 23:54 -0500, Meng Xu wrote:
> 
> > @@ -1150,6 +1300,101 @@ rt_dom_cntl(
> >      return rc;
> >  }
> > 
> > +/*
> > + * The replenishment timer handler picks vcpus
> > + * from the replq and does the actual replenishment.
> > + */
> > +static void repl_handler(void *data){
> > +    unsigned long flags;
> > +    s_time_t now = NOW();
> > +    struct scheduler *ops = data;
> > +    struct rt_private *prv = rt_priv(ops);
> > +    struct list_head *replq = rt_replq(ops);
> > +    struct list_head *runq = rt_runq(ops);
> > +    struct timer *repl_timer = prv->repl_timer;
> > +    struct list_head *iter, *tmp;
> > +    struct list_head tmp_replq;
> > +    struct rt_vcpu *svc = NULL;
> > +
> > +    spin_lock_irqsave(&prv->lock, flags);
> Hmm, I haven't thought hard about the choice between
> spin_lock_irqsave() and spin_lock_irq(), before.
> 
> Hi Dario,
> Is it better to use spin_lock_irqsave() or spin_lock_irq() at this
> place?
> 
Just very quickly about this (I'll comment about the rest of the patch
later).

> I'm not quite sure if the handler can be called in an interrupt
> disabled context? Can it?
>
I recommend looking at what happens inside init_timer(), i.e., where a
pointer to this function is stashed. Then, still in xen/common/timer.c,
check where this (and, in general, a timer handling function provided
to timer_init()) is actually invoked.

When you'll find that spot, the answer to whether spin_lock_irq() is
safe or not in here, will appear quite evident. :-)

> When I used the spin_lock_irq(save), I just refered to what credit2
> scheduler does, but didn't think hard enough about which one has
> better performance.
>
I'm not sure what you mean when you talk about Credit2, as there are no
timers in there. In any case, it is indeed the case that, if just
_irq() is safe, we should use it, as it's cheaper.

Regards,
Dario
-- 
<<This happens because I choose it to happen!>> (Raistlin Majere)
-----------------------------------------------------------------
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)


[-- Attachment #1.2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 181 bytes --]

[-- Attachment #2: Type: text/plain, Size: 126 bytes --]

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