On Wed, Sep 29, 2010 at 09:12:25AM -0500, Christoph Lameter wrote: > On Wed, 29 Sep 2010, Mel Gorman wrote: > > > Alternatively we could revisit Christoph's suggestion of modifying > > stat_threshold when under pressure instead of zone_page_state_snapshot. Maybe > > by temporarily stat_threshold when kswapd is awake to a per-zone value > > such that > > > > zone->low + threshold*nr_online_cpus < high > > Updating the threshold also is expensive. Even if it's moved to a read-mostly part of the zone such as after lowmem_reserve? > I thought more along the lines > of reducing the threshold for good if the VM runs into reclaim trouble > because of too high fuzziness in the counters. > That would be unfortunate as it would only take trouble to happen once for performance to be impaired for the remaining uptime of the machine. -- Mel Gorman Part-time Phd Student Linux Technology Center University of Limerick IBM Dublin Software Lab -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxxx For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>