On Mon, Feb 03, 2014 at 05:20:46PM -0800, David Rientjes wrote: > On Tue, 4 Feb 2014, Joonsoo Kim wrote: > > > I think that you need more code to skip this type of page correctly. > > Without page_mapped() check, this code makes migratable pages be skipped, > > since if page_mapped() case, page_count() may be more than zero. > > > > So I think that you need following change. > > > > (!page_mapping(page) && !page_mapped(page) && page_count(page)) > > > > These pages returned by get_user_pages() will have a mapcount of 1 so this > wouldn't actually fix the massive lock contention. page_mapping() is only > going to be NULL for pages off the lru like these are for > PAGE_MAPPING_ANON. Okay. It can't fix your situation. Anyway, *normal* anon pages may be mapped and have positive page_count(), so your code such as '!page_mapping(page) && page_count(page)' makes compaction skip these *normal* anon pages and this is incorrect behaviour. Thanks. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>