Re: [Xen-devel] [PATCH] xen/sched_credit: Use delay to control scheduling frequency

[George Dunlap <george.dunlap@xxxxxxxxxx>]

On Sun, 2012-01-08 at 12:03 +0000, Lv, Hui wrote:
> Thanks, George.
> Should I send a revised version?
> Can it be checked in?

Yes, please send a revised version.

It will be checked in if one of the maintainers thinks there's enough
consensus (probably if no one has any more comments in the next day or
two).

Thanks,
 -George

> 
> -----Original Message-----
> From: George Dunlap [mailto:george.dunlap@xxxxxxxxxxxxx] 
> Sent: Saturday, January 07, 2012 3:57 AM
> To: Lv, Hui
> Cc: xen-devel@xxxxxxxxxxxxxxxxxxx; raistlin@xxxxxxxx; JBeulich@xxxxxxxx; Ian 
> Campbell
> Subject: Re: [PATCH] xen/sched_credit: Use delay to control scheduling 
> frequency
> 
> Sorry for the delay; just catching up after the Christmas holidays.
> 
> On 26/12/11 03:46, Hui Lv wrote:
> > @@ -1533,6 +1569,15 @@ csched_init(struct scheduler *ops)
> >       prv->tick_period_us = prv->tslice_ms * 1000 / prv->ticks_per_tslice;
> >       prv->credits_per_tslice = CSCHED_CREDITS_PER_MSEC * prv->tslice_ms;
> >
> > +    if ( MICROSECS(sched_ratelimit_us)>  MILLISECS(sched_credit_tslice_ms) 
> > )
> > +    {
> > +        printk("WARNING: sched_ratelimit_us>"
> > +               "sched_credit_tslice_ms is undefined\n"
> > +               "ratelimit_us is set to 1000 * tslice_ms forcely\n")
> The standard idiom for this kind of message would be:
>   WARNING [what's wrong]
>   [What you're doing about it]
> 
> So the last sentence of the warning should be:
>    Setting ratelimit_us to 1000 * tslice_ms
> 
> (Grammatically, you could say "Forcing ratelimit..." but I think "force" 
> is too strong in this case.)
> 
> Other than that, I'm happy with it, if everyone else is:
> 
> Acked-by: George Dunlap <george.dunlap@xxxxxxxxxxxxx>
> 



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel