Re: [PATCH v10 3/6] mm: Introduce Reported pages

["Michael S. Tsirkin" <mst@redhat.com>]

On Mon, Sep 23, 2019 at 05:47:24PM +0200, David Hildenbrand wrote:
> On 23.09.19 17:45, David Hildenbrand wrote:
> > On 23.09.19 17:37, Michael S. Tsirkin wrote:
> >> On Mon, Sep 23, 2019 at 08:28:00AM -0700, Alexander Duyck wrote:
> >>> On Mon, Sep 23, 2019 at 8:00 AM Michael S. Tsirkin <mst@redhat.com> wrote:
> >>>>
> >>>> On Mon, Sep 23, 2019 at 07:50:15AM -0700, Alexander Duyck wrote:
> >>>>>>> +static inline void
> >>>>>>> +page_reporting_reset_boundary(struct zone *zone, unsigned int order, int mt)
> >>>>>>> +{
> >>>>>>> +     int index;
> >>>>>>> +
> >>>>>>> +     if (order < PAGE_REPORTING_MIN_ORDER)
> >>>>>>> +             return;
> >>>>>>> +     if (!test_bit(ZONE_PAGE_REPORTING_ACTIVE, &zone->flags))
> >>>>>>> +             return;
> >>>>>>> +
> >>>>>>> +     index = get_reporting_index(order, mt);
> >>>>>>> +     reported_boundary[index] = &zone->free_area[order].free_list[mt];
> >>>>>>> +}
> >>>>>>
> >>>>>> So this seems to be costly.
> >>>>>> I'm guessing it's the access to flags:
> >>>>>>
> >>>>>>
> >>>>>>         /* zone flags, see below */
> >>>>>>         unsigned long           flags;
> >>>>>>
> >>>>>>         /* Primarily protects free_area */
> >>>>>>         spinlock_t              lock;
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>> which is in the same cache line as the lock.
> >>>>>
> >>>>> I'm not sure what you mean by this being costly?
> >>>>
> >>>> I've just been wondering why does will it scale report a 1.5% regression
> >>>> with this patch.
> >>>
> >>> Are you talking about data you have collected from a test you have
> >>> run, or the data I have run?
> >>
> >> About the kernel test robot auto report that was sent recently.
> > 
> > https://lkml.org/lkml/2019/9/21/112
> > 
> > And if I'm correct, that regression is observable in case reporting is
> > not enabled. (so with this patch applied only, e.g., on a bare-metal system)
> > 
> 
> To be even more precise: # CONFIG_PAGE_REPORTING is not set

Even if it was, I'd hope for 0 overhead when not present runtime.

-- 
MST