Re: [PATCH v2 2/2] mm/compaction: avoid rescanning pageblocks in isolate_freepages

[Vlastimil Babka]

On 05/08/2014 07:28 AM, Joonsoo Kim wrote:
> On Wed, May 07, 2014 at 02:09:10PM +0200, Vlastimil Babka wrote:
> > The compaction free scanner in isolate_freepages() currently remembers PFN of
> > the highest pageblock where it successfully isolates, to be used as the
> > starting pageblock for the next invocation. The rationale behind this is that
> > page migration might return free pages to the allocator when migration fails
> > and we don't want to skip them if the compaction continues.
> >
> > Since migration now returns free pages back to compaction code where they can
> > be reused, this is no longer a concern. This patch changes isolate_freepages()
> > so that the PFN for restarting is updated with each pageblock where isolation
> > is attempted. Using stress-highalloc from mmtests, this resulted in 10%
> > reduction of the pages scanned by the free scanner.
>
> Hello,
>
> Although this patch could reduce page scanned, it is possible to skip
> scanning fresh pageblock. If there is zone lock contention and we are on
> asyn compaction, we stop scanning this pageblock immediately. And
> then, we will continue to scan next pageblock. With this patch,
> next_free_pfn is updated in this case, so we never come back again to this
> pageblock. Possibly this makes compaction success rate low, doesn't
> it?

Hm, you're right and thanks for catching that, but I think this is a 
sign of a worse and older issue than skipping a pageblock?
When isolate_freepages_block() breaks loop due to lock contention, then 
isolate_freepages() (which called it) should also immediately quit its 
loop. Trying another pageblock in the same zone with the same zone->lock 
makes no sense here? If this is fixed, then the issue you're pointing 
out will also be fixed as next_free_pfn will still point to the 
pageblock where the break occured.

> Thanks.

--
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/