Andrew, These are writeback patches intended for 2.6.36. It's combined from 2 previous patchsets: writeback cleanups and trivial fixes <http://lkml.org/lkml/2010/7/10/153> writeback: try to write older pages first <http://lkml.org/lkml/2010/7/22/47> changelog: - removed patch "writeback: take account of NR_WRITEBACK_TEMP in balance_dirty_pages()" - added patch "writeback: explicit low bound for vm.dirty_ratio" - use "if (list_empty(&wb->b_io))" directly in "writeback: sync expired inodes first in background writeback" - fix misplaced chunk for removing more_io in include/trace/events/ext4.h - update comments in "writeback: fix queue_io() ordering" - update comments in "writeback: add comment to the dirty limits functions" - patch "writeback: try more writeback as long as something was written" will no longer livelock sync() with Jan's sync() livelock avoidance patches [PATCH 01/13] writeback: reduce calls to global_page_state in balance_dirty_pages() [PATCH 02/13] writeback: avoid unnecessary calculation of bdi dirty thresholds [PATCH 03/13] writeback: add comment to the dirty limits functions [PATCH 04/13] writeback: dont redirty tail an inode with dirty pages [PATCH 05/13] writeback: fix queue_io() ordering [PATCH 06/13] writeback: merge for_kupdate and !for_kupdate cases [PATCH 07/13] writeback: explicit low bound for vm.dirty_ratio [PATCH 08/13] writeback: pass writeback_control down to move_expired_inodes() [PATCH 09/13] writeback: the kupdate expire timestamp should be a moving target [PATCH 10/13] writeback: kill writeback_control.more_io [PATCH 11/13] writeback: sync expired inodes first in background writeback [PATCH 12/13] writeback: try more writeback as long as something was written [PATCH 13/13] writeback: introduce writeback_control.inodes_written Thanks, Fengguang -- 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>