On Thu, Oct 02 2008, Arjan van de Ven wrote:
quoted text
> On Thu, 2 Oct 2008 15:27:47 +0200
> Jens Axboe <jens.axboe@oracle.com> wrote:
> 
> > On Thu, Oct 02 2008, Arjan van de Ven wrote:
> > > On Thu, 2 Oct 2008 11:45:37 +0200
> > > Jens Axboe <jens.axboe@oracle.com> wrote:
> > > 
> > > > 
> > > > That's a good idea, just bump the priority a little bit. Arjan,
> > > > did you test that out? I'd suggest just trying prio level 0 and
> > > > still using best-effort scheduling. Probably still need the sync
> > > > marking, would be interesting to experiment with though.
> > > 
> > > I looked at 0 but it appears the 0 is the default for everyone...
> > > if everyone just defaulted to > 0 then yes I would have picked 0.
> > 
> > That's not correct, class BE and value 4 is the default (and the code
> > defaults to that, if you haven't set a value yourself):
> > 
> > #define IOPRIO_NORM     (4)
> > static inline int task_ioprio(struct io_context *ioc)
> > {
> >         if (ioprio_valid(ioc->ioprio))
> >                 return IOPRIO_PRIO_DATA(ioc->ioprio);
> > 
> >         return IOPRIO_NORM;
> > }
> > 
> > static inline int task_ioprio_class(struct io_context *ioc)
> > {
> >         if (ioprio_valid(ioc->ioprio))
> >                 return IOPRIO_PRIO_CLASS(ioc->ioprio);
> > 
> >         return IOPRIO_CLASS_BE;
> > }
> > 
> > So if you use IOPRIO_CLASS_BE and 0 for the ioprio, you will have the
> > highest priority of the default scheduling class.
> 
> ok
> 
> I checked not by looking at the code, but running ionice -p <pid> on a
> bunch of things and they came back as 0 

Yes, it'll report '0' which means 'not set'. The kernel inteprets 'not
set' as the default values, BE/4. There's a big diffence, since '0'
means that we track CPU nice values where as if it returned be/4 then
that is a strict/fixed setting.

-- 
Jens Axboe

--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/