Re: [patch 3/5] mm: try to distribute dirty pages fairly acrosszones

[KAMEZAWA Hiroyuki]

On Thu, 24 Nov 2011 14:11:55 +0100
Johannes Weiner <hannes@xxxxxxxxxxx> wrote:

> On Thu, Nov 24, 2011 at 10:07:55AM +0900, KAMEZAWA Hiroyuki wrote:
			goto this_zone_full;
> > >  
> > >  		BUILD_BUG_ON(ALLOC_NO_WATERMARKS < NR_WMARK);
> > >  		if (!(alloc_flags & ALLOC_NO_WATERMARKS)) {
> > 
> > This wil call 
> > 
> >                 if (NUMA_BUILD)
> >                         zlc_mark_zone_full(zonelist, z);
> > 
> > And this zone will be marked as full. 
> > 
> > IIUC, zlc_clear_zones_full() is called only when direct reclaim ends.
> > So, if no one calls direct-reclaim, 'full' mark may never be cleared
> > even when number of dirty pages goes down to safe level ?
> > I'm sorry if this is alread discussed.
> 
> It does not remember which zones are marked full for longer than a
> second - see zlc_setup() - and also ignores this information when an
> iteration over the zonelist with the cache enabled came up
> empty-handed.
> 
Ah, thank you for clarification.
I understand how zlc_active/did_zlc_setup/zlc_setup()...complicated ;)

Thanks,
-Kame

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