Re: [PATCH 1/3] Reintroduce zone_reclaim_interval for whenzone_reclaim() scans and fails to avoid CPU spinning at 100% on NUMA

[Mel Gorman]

On Mon, Jun 08, 2009 at 09:31:09AM -0400, Rik van Riel wrote:
> Mel Gorman wrote:
>
>> The scanning occurs because zone_reclaim() cannot tell
>> in advance the scan is pointless because the counters do not distinguish
>> between pagecache pages backed by disk and by RAM. 
>
> Yes it can.  Since 2.6.27, filesystem backed and swap/ram backed
> pages have been living on separate LRU lists. 

Yes, they're on separate LRU lists but they are not the only pages on those
lists. The tmpfs pages are mixed in together with anonymous pages so we
cannot use NR_*_ANON.

Look at patch 2 and where I introduced;

       /*
        * Work out how many page cache pages we can reclaim in this mode.
        *
        * NOTE: Ideally, tmpfs pages would be accounted as if they were
        *       NR_FILE_MAPPED as swap is required to discard those
        *       pages even when they are clean. However, there is no
        *       way of quickly identifying the number of tmpfs pages
        */
       pagecache_reclaimable = zone_page_state(zone, NR_FILE_PAGES);
       if (!(zone_reclaim_mode & RECLAIM_WRITE))
               pagecache_reclaimable -= zone_page_state(zone, NR_FILE_DIRTY);
       if (!(zone_reclaim_mode & RECLAIM_SWAP))
               pagecache_reclaimable -= zone_page_state(zone, NR_FILE_MAPPED);

If the tmpfs pages can be accounted for there, then chances are that patch
1 goes away - at least until some other situation is encountered where
we scan erroneously.

> This allows you to
> fix the underlying problem, instead of having to add a retry
> interval.
>

Which is obviously my preference but after looking around for a bit, I
didn't spot an obvious answer.

-- 
Mel Gorman
Part-time Phd Student                          Linux Technology Center
University of Limerick                         IBM Dublin Software Lab
--
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/