On Thu, 15 Jun 2006, Christoph Lameter wrote: > On Wed, 14 Jun 2006, Mel Gorman wrote: > >> No, it hasn't. It's really a NUMA thing and there are a number of >> important restrictions as far as hotremove is concerned. > > No, page migration is not only a NUMA thing, and yes there are still > restrictions but there is code in Andrews tree to relax some of them. > ok, sorry, I should have phrased that better. > The migrate_pages() function is designed to migrate between arbitrary > lists of pages (or using a allocation function(mm)). You can use that in > the non NUMA case without a problem (Hugh Dickins did that to test the > page migration on SMP as far as I can tell). It was designed after all by > keeping the needs of memory hotplug in mind. > FWIW, I expect to use the page migration functions for addressing fragmentation in the page allocator issues at some point in the future. -- Mel Gorman Part-time Phd Student Linux Technology Center University of Limerick IBM Dublin Software Lab
