On Thu, Mar 04, 2010 at 10:41:43PM +0530, Balbir Singh wrote: > * Andrea Righi <arighi@xxxxxxxxxxx> [2010-03-04 11:40:11]: > > > Control the maximum amount of dirty pages a cgroup can have at any given time. > > > > Per cgroup dirty limit is like fixing the max amount of dirty (hard to reclaim) > > page cache used by any cgroup. So, in case of multiple cgroup writers, they > > will not be able to consume more than their designated share of dirty pages and > > will be forced to perform write-out if they cross that limit. > > > > The overall design is the following: > > > > - account dirty pages per cgroup > > - limit the number of dirty pages via memory.dirty_ratio / memory.dirty_bytes > > and memory.dirty_background_ratio / memory.dirty_background_bytes in > > cgroupfs > > - start to write-out (background or actively) when the cgroup limits are > > exceeded > > > > This feature is supposed to be strictly connected to any underlying IO > > controller implementation, so we can stop increasing dirty pages in VM layer > > and enforce a write-out before any cgroup will consume the global amount of > > dirty pages defined by the /proc/sys/vm/dirty_ratio|dirty_bytes and > > /proc/sys/vm/dirty_background_ratio|dirty_background_bytes limits. > > > > Changelog (v3 -> v4) > > ~~~~~~~~~~~~~~~~~~~~~~ > > * handle the migration of tasks across different cgroups > > NOTE: at the moment we don't move charges of file cache pages, so this > > functionality is not immediately necessary. However, since the migration of > > file cache pages is in plan, it is better to start handling file pages > > anyway. > > * properly account dirty pages in nilfs2 > > (thanks to Kirill A. Shutemov <kirill@xxxxxxxxxxxxx>) > > * lockless access to dirty memory parameters > > * fix: page_cgroup lock must not be acquired under mapping->tree_lock > > (thanks to Daisuke Nishimura <nishimura@xxxxxxxxxxxxxxxxx> and > > KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>) > > * code restyling > > > > This seems to be converging, what sort of tests are you running on > this patchset? A very simple test at the moment, just some parallel dd's running in different cgroups. For example: - cgroup A: low dirty limits (writes are almost sync) echo 1000 > /cgroups/A/memory.dirty_bytes echo 1000 > /cgroups/A/memory.dirty_background_bytes - cgroup B: high dirty limits (writes are all buffered in page cache) echo 100 > /cgroups/B/memory.dirty_ratio echo 50 > /cgroups/B/memory.dirty_background_ratio Then run the dd's and look at memory.stat: - cgroup A: # dd if=/dev/zero of=A bs=1M count=1000 - cgroup B: # dd if=/dev/zero of=B bs=1M count=1000 A random snapshot during the writes: # grep "dirty\|writeback" /cgroups/[AB]/memory.stat /cgroups/A/memory.stat:filedirty 0 /cgroups/A/memory.stat:writeback 0 /cgroups/A/memory.stat:writeback_tmp 0 /cgroups/A/memory.stat:dirty_pages 0 /cgroups/A/memory.stat:writeback_pages 0 /cgroups/A/memory.stat:writeback_temp_pages 0 /cgroups/B/memory.stat:filedirty 67226 /cgroups/B/memory.stat:writeback 136 /cgroups/B/memory.stat:writeback_tmp 0 /cgroups/B/memory.stat:dirty_pages 67226 /cgroups/B/memory.stat:writeback_pages 136 /cgroups/B/memory.stat:writeback_temp_pages 0 I plan to run more detailed IO benchmark soon. -Andrea -- 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>