diff options
Diffstat (limited to 'Documentation/controllers')
| -rw-r--r-- | Documentation/controllers/memcg_test.txt | 342 | ||||
| -rw-r--r-- | Documentation/controllers/memory.txt | 135 |
2 files changed, 467 insertions, 10 deletions
diff --git a/Documentation/controllers/memcg_test.txt b/Documentation/controllers/memcg_test.txt new file mode 100644 index 000000000000..08d4d3ea0d79 --- /dev/null +++ b/Documentation/controllers/memcg_test.txt @@ -0,0 +1,342 @@ +Memory Resource Controller(Memcg) Implementation Memo. +Last Updated: 2008/12/15 +Base Kernel Version: based on 2.6.28-rc8-mm. + +Because VM is getting complex (one of reasons is memcg...), memcg's behavior +is complex. This is a document for memcg's internal behavior. +Please note that implementation details can be changed. + +(*) Topics on API should be in Documentation/controllers/memory.txt) + +0. How to record usage ? + 2 objects are used. + + page_cgroup ....an object per page. + Allocated at boot or memory hotplug. Freed at memory hot removal. + + swap_cgroup ... an entry per swp_entry. + Allocated at swapon(). Freed at swapoff(). + + The page_cgroup has USED bit and double count against a page_cgroup never + occurs. swap_cgroup is used only when a charged page is swapped-out. + +1. Charge + + a page/swp_entry may be charged (usage += PAGE_SIZE) at + + mem_cgroup_newpage_charge() + Called at new page fault and Copy-On-Write. + + mem_cgroup_try_charge_swapin() + Called at do_swap_page() (page fault on swap entry) and swapoff. + Followed by charge-commit-cancel protocol. (With swap accounting) + At commit, a charge recorded in swap_cgroup is removed. + + mem_cgroup_cache_charge() + Called at add_to_page_cache() + + mem_cgroup_cache_charge_swapin() + Called at shmem's swapin. + + mem_cgroup_prepare_migration() + Called before migration. "extra" charge is done and followed by + charge-commit-cancel protocol. + At commit, charge against oldpage or newpage will be committed. + +2. Uncharge + a page/swp_entry may be uncharged (usage -= PAGE_SIZE) by + + mem_cgroup_uncharge_page() + Called when an anonymous page is fully unmapped. I.e., mapcount goes + to 0. If the page is SwapCache, uncharge is delayed until + mem_cgroup_uncharge_swapcache(). + + mem_cgroup_uncharge_cache_page() + Called when a page-cache is deleted from radix-tree. If the page is + SwapCache, uncharge is delayed until mem_cgroup_uncharge_swapcache(). + + mem_cgroup_uncharge_swapcache() + Called when SwapCache is removed from radix-tree. The charge itself + is moved to swap_cgroup. (If mem+swap controller is disabled, no + charge to swap occurs.) + + mem_cgroup_uncharge_swap() + Called when swp_entry's refcnt goes down to 0. A charge against swap + disappears. + + mem_cgroup_end_migration(old, new) + At success of migration old is uncharged (if necessary), a charge + to new page is committed. At failure, charge to old page is committed. + +3. charge-commit-cancel + In some case, we can't know this "charge" is valid or not at charging + (because of races). + To handle such case, there are charge-commit-cancel functions. + mem_cgroup_try_charge_XXX + mem_cgroup_commit_charge_XXX + mem_cgroup_cancel_charge_XXX + these are used in swap-in and migration. + + At try_charge(), there are no flags to say "this page is charged". + at this point, usage += PAGE_SIZE. + + At commit(), the function checks the page should be charged or not + and set flags or avoid charging.(usage -= PAGE_SIZE) + + At cancel(), simply usage -= PAGE_SIZE. + +Under below explanation, we assume CONFIG_MEM_RES_CTRL_SWAP=y. + +4. Anonymous + Anonymous page is newly allocated at + - page fault into MAP_ANONYMOUS mapping. + - Copy-On-Write. + It is charged right after it's allocated before doing any page table + related operations. Of course, it's uncharged when another page is used + for the fault address. + + At freeing anonymous page (by exit() or munmap()), zap_pte() is called + and pages for ptes are freed one by one.(see mm/memory.c). Uncharges + are done at page_remove_rmap() when page_mapcount() goes down to 0. + + Another page freeing is by page-reclaim (vmscan.c) and anonymous + pages are swapped out. In this case, the page is marked as + PageSwapCache(). uncharge() routine doesn't uncharge the page marked + as SwapCache(). It's delayed until __delete_from_swap_cache(). + + 4.1 Swap-in. + At swap-in, the page is taken from swap-cache. There are 2 cases. + + (a) If the SwapCache is newly allocated and read, it has no charges. + (b) If the SwapCache has been mapped by processes, it has been + charged already. + + This swap-in is one of the most complicated work. In do_swap_page(), + following events occur when pte is unchanged. + + (1) the page (SwapCache) is looked up. + (2) lock_page() + (3) try_charge_swapin() + (4) reuse_swap_page() (may call delete_swap_cache()) + (5) commit_charge_swapin() + (6) swap_free(). + + Considering following situation for example. + + (A) The page has not been charged before (2) and reuse_swap_page() + doesn't call delete_from_swap_cache(). + (B) The page has not been charged before (2) and reuse_swap_page() + calls delete_from_swap_cache(). + (C) The page has been charged before (2) and reuse_swap_page() doesn't + call delete_from_swap_cache(). + (D) The page has been charged before (2) and reuse_swap_page() calls + delete_from_swap_cache(). + + memory.usage/memsw.usage changes to this page/swp_entry will be + Case (A) (B) (C) (D) + Event + Before (2) 0/ 1 0/ 1 1/ 1 1/ 1 + =========================================== + (3) +1/+1 +1/+1 +1/+1 +1/+1 + (4) - 0/ 0 - -1/ 0 + (5) 0/-1 0/ 0 -1/-1 0/ 0 + (6) - 0/-1 - 0/-1 + =========================================== + Result 1/ 1 1/ 1 1/ 1 1/ 1 + + In any cases, charges to this page should be 1/ 1. + + 4.2 Swap-out. + At swap-out, typical state transition is below. + + (a) add to swap cache. (marked as SwapCache) + swp_entry's refcnt += 1. + (b) fully unmapped. + swp_entry's refcnt += # of ptes. + (c) write back to swap. + (d) delete from swap cache. (remove from SwapCache) + swp_entry's refcnt -= 1. + + + At (b), the page is marked as SwapCache and not uncharged. + At (d), the page is removed from SwapCache and a charge in page_cgroup + is moved to swap_cgroup. + + Finally, at task exit, + (e) zap_pte() is called and swp_entry's refcnt -=1 -> 0. + Here, a charge in swap_cgroup disappears. + +5. Page Cache + Page Cache is charged at + - add_to_page_cache_locked(). + + uncharged at + - __remove_from_page_cache(). + + The logic is very clear. (About migration, see below) + Note: __remove_from_page_cache() is called by remove_from_page_cache() + and __remove_mapping(). + +6. Shmem(tmpfs) Page Cache + Memcg's charge/uncharge have special handlers of shmem. The best way + to understand shmem's page state transition is to read mm/shmem.c. + But brief explanation of the behavior of memcg around shmem will be + helpful to understand the logic. + + Shmem's page (just leaf page, not direct/indirect block) can be on + - radix-tree of shmem's inode. + - SwapCache. + - Both on radix-tree and SwapCache. This happens at swap-in + and swap-out, + + It's charged when... + - A new page is added to shmem's radix-tree. + - A swp page is read. (move a charge from swap_cgroup to page_cgroup) + It's uncharged when + - A page is removed from radix-tree and not SwapCache. + - When SwapCache is removed, a charge is moved to swap_cgroup. + - When swp_entry's refcnt goes down to 0, a charge in swap_cgroup + disappears. + +7. Page Migration + One of the most complicated functions is page-migration-handler. + Memcg has 2 routines. Assume that we are migrating a page's contents + from OLDPAGE to NEWPAGE. + + Usual migration logic is.. + (a) remove the page from LRU. + (b) allocate NEWPAGE (migration target) + (c) lock by lock_page(). + (d) unmap all mappings. + (e-1) If necessary, replace entry in radix-tree. + (e-2) move contents of a page. + (f) map all mappings again. + (g) pushback the page to LRU. + (-) OLDPAGE will be freed. + + Before (g), memcg should complete all necessary charge/uncharge to + NEWPAGE/OLDPAGE. + + The point is.... + - If OLDPAGE is anonymous, all charges will be dropped at (d) because + try_to_unmap() drops all mapcount and the page will not be + SwapCache. + + - If OLDPAGE is SwapCache, charges will be kept at (g) because + __delete_from_swap_cache() isn't called at (e-1) + + - If OLDPAGE is page-cache, charges will be kept at (g) because + remove_from_swap_cache() isn't called at (e-1) + + memcg provides following hooks. + + - mem_cgroup_prepare_migration(OLDPAGE) + Called after (b) to account a charge (usage += PAGE_SIZE) against + memcg which OLDPAGE belongs to. + + - mem_cgroup_end_migration(OLDPAGE, NEWPAGE) + Called after (f) before (g). + If OLDPAGE is used, commit OLDPAGE again. If OLDPAGE is already + charged, a charge by prepare_migration() is automatically canceled. + If NEWPAGE is used, commit NEWPAGE and uncharge OLDPAGE. + + But zap_pte() (by exit or munmap) can be called while migration, + we have to check if OLDPAGE/NEWPAGE is a valid page after commit(). + +8. LRU + Each memcg has its own private LRU. Now, it's handling is under global + VM's control (means that it's handled under global zone->lru_lock). + Almost all routines around memcg's LRU is called by global LRU's + list management functions under zone->lru_lock(). + + A special function is mem_cgroup_isolate_pages(). This scans + memcg's private LRU and call __isolate_lru_page() to extract a page + from LRU. + (By __isolate_lru_page(), the page is removed from both of global and + private LRU.) + + +9. Typical Tests. + + Tests for racy cases. + + 9.1 Small limit to memcg. + When you do test to do racy case, it's good test to set memcg's limit + to be very small rather than GB. Many races found in the test under + xKB or xxMB limits. + (Memory behavior under GB and Memory behavior under MB shows very + different situation.) + + 9.2 Shmem + Historically, memcg's shmem handling was poor and we saw some amount + of troubles here. This is because shmem is page-cache but can be + SwapCache. Test with shmem/tmpfs is always good test. + + 9.3 Migration + For NUMA, migration is an another special case. To do easy test, cpuset + is useful. Following is a sample script to do migration. + + mount -t cgroup -o cpuset none /opt/cpuset + + mkdir /opt/cpuset/01 + echo 1 > /opt/cpuset/01/cpuset.cpus + echo 0 > /opt/cpuset/01/cpuset.mems + echo 1 > /opt/cpuset/01/cpuset.memory_migrate + mkdir /opt/cpuset/02 + echo 1 > /opt/cpuset/02/cpuset.cpus + echo 1 > /opt/cpuset/02/cpuset.mems + echo 1 > /opt/cpuset/02/cpuset.memory_migrate + + In above set, when you moves a task from 01 to 02, page migration to + node 0 to node 1 will occur. Following is a script to migrate all + under cpuset. + -- + move_task() + { + for pid in $1 + do + /bin/echo $pid >$2/tasks 2>/dev/null + echo -n $pid + echo -n " " + done + echo END + } + + G1_TASK=`cat ${G1}/tasks` + G2_TASK=`cat ${G2}/tasks` + move_task "${G1_TASK}" ${G2} & + -- + 9.4 Memory hotplug. + memory hotplug test is one of good test. + to offline memory, do following. + # echo offline > /sys/devices/system/memory/memoryXXX/state + (XXX is the place of memory) + This is an easy way to test page migration, too. + + 9.5 mkdir/rmdir + When using hierarchy, mkdir/rmdir test should be done. + Use tests like the following. + + echo 1 >/opt/cgroup/01/memory/use_hierarchy + mkdir /opt/cgroup/01/child_a + mkdir /opt/cgroup/01/child_b + + set limit to 01. + add limit to 01/child_b + run jobs under child_a and child_b + + create/delete following groups at random while jobs are running. + /opt/cgroup/01/child_a/child_aa + /opt/cgroup/01/child_b/child_bb + /opt/cgroup/01/child_c + + running new jobs in new group is also good. + + 9.6 Mount with other subsystems. + Mounting with other subsystems is a good test because there is a + race and lock dependency with other cgroup subsystems. + + example) + # mount -t cgroup none /cgroup -t cpuset,memory,cpu,devices + + and do task move, mkdir, rmdir etc...under this. diff --git a/Documentation/controllers/memory.txt b/Documentation/controllers/memory.txt index 1c07547d3f81..e1501964df1e 100644 --- a/Documentation/controllers/memory.txt +++ b/Documentation/controllers/memory.txt @@ -137,7 +137,32 @@ behind this approach is that a cgroup that aggressively uses a shared page will eventually get charged for it (once it is uncharged from the cgroup that brought it in -- this will happen on memory pressure). -2.4 Reclaim +Exception: If CONFIG_CGROUP_CGROUP_MEM_RES_CTLR_SWAP is not used.. +When you do swapoff and make swapped-out pages of shmem(tmpfs) to +be backed into memory in force, charges for pages are accounted against the +caller of swapoff rather than the users of shmem. + + +2.4 Swap Extension (CONFIG_CGROUP_MEM_RES_CTLR_SWAP) +Swap Extension allows you to record charge for swap. A swapped-in page is +charged back to original page allocator if possible. + +When swap is accounted, following files are added. + - memory.memsw.usage_in_bytes. + - memory.memsw.limit_in_bytes. + +usage of mem+swap is limited by memsw.limit_in_bytes. + +Note: why 'mem+swap' rather than swap. +The global LRU(kswapd) can swap out arbitrary pages. Swap-out means +to move account from memory to swap...there is no change in usage of +mem+swap. + +In other words, when we want to limit the usage of swap without affecting +global LRU, mem+swap limit is better than just limiting swap from OS point +of view. + +2.5 Reclaim Each cgroup maintains a per cgroup LRU that consists of an active and inactive list. When a cgroup goes over its limit, we first try @@ -207,12 +232,6 @@ exceeded. The memory.stat file gives accounting information. Now, the number of caches, RSS and Active pages/Inactive pages are shown. -The memory.force_empty gives an interface to drop *all* charges by force. - -# echo 1 > memory.force_empty - -will drop all charges in cgroup. Currently, this is maintained for test. - 4. Testing Balbir posted lmbench, AIM9, LTP and vmmstress results [10] and [11]. @@ -242,10 +261,106 @@ reclaimed. A cgroup can be removed by rmdir, but as discussed in sections 4.1 and 4.2, a cgroup might have some charge associated with it, even though all -tasks have migrated away from it. Such charges are automatically dropped at -rmdir() if there are no tasks. +tasks have migrated away from it. +Such charges are freed(at default) or moved to its parent. When moved, +both of RSS and CACHES are moved to parent. +If both of them are busy, rmdir() returns -EBUSY. See 5.1 Also. + +Charges recorded in swap information is not updated at removal of cgroup. +Recorded information is discarded and a cgroup which uses swap (swapcache) +will be charged as a new owner of it. + + +5. Misc. interfaces. + +5.1 force_empty + memory.force_empty interface is provided to make cgroup's memory usage empty. + You can use this interface only when the cgroup has no tasks. + When writing anything to this + + # echo 0 > memory.force_empty + + Almost all pages tracked by this memcg will be unmapped and freed. Some of + pages cannot be freed because it's locked or in-use. Such pages are moved + to parent and this cgroup will be empty. But this may return -EBUSY in + some too busy case. + + Typical use case of this interface is that calling this before rmdir(). + Because rmdir() moves all pages to parent, some out-of-use page caches can be + moved to the parent. If you want to avoid that, force_empty will be useful. + +5.2 stat file + memory.stat file includes following statistics (now) + cache - # of pages from page-cache and shmem. + rss - # of pages from anonymous memory. + pgpgin - # of event of charging + pgpgout - # of event of uncharging + active_anon - # of pages on active lru of anon, shmem. + inactive_anon - # of pages on active lru of anon, shmem + active_file - # of pages on active lru of file-cache + inactive_file - # of pages on inactive lru of file cache + unevictable - # of pages cannot be reclaimed.(mlocked etc) + + Below is depend on CONFIG_DEBUG_VM. + inactive_ratio - VM inernal parameter. (see mm/page_alloc.c) + recent_rotated_anon - VM internal parameter. (see mm/vmscan.c) + recent_rotated_file - VM internal parameter. (see mm/vmscan.c) + recent_scanned_anon - VM internal parameter. (see mm/vmscan.c) + recent_scanned_file - VM internal parameter. (see mm/vmscan.c) + + Memo: + recent_rotated means recent frequency of lru rotation. + recent_scanned means recent # of scans to lru. + showing for better debug please see the code for meanings. + + +5.3 swappiness + Similar to /proc/sys/vm/swappiness, but affecting a hierarchy of groups only. + + Following cgroup's swapiness can't be changed. + - root cgroup (uses /proc/sys/vm/swappiness). + - a cgroup which uses hierarchy and it has child cgroup. + - a cgroup which uses hierarchy and not the root of hierarchy. + + +6. Hierarchy support + +The memory controller supports a deep hierarchy and hierarchical accounting. +The hierarchy is created by creating the appropriate cgroups in the +cgroup filesystem. Consider for example, the following cgroup filesystem +hierarchy + + root + / | \ + / | \ + a b c + | \ + | \ + d e + +In the diagram above, with hierarchical accounting enabled, all memory +usage of e, is accounted to its ancestors up until the root (i.e, c and root), +that has memory.use_hierarchy enabled. If one of the ancestors goes over its +limit, the reclaim algorithm reclaims from the tasks in the ancestor and the +children of the ancestor. + +6.1 Enabling hierarchical accounting and reclaim + +The memory controller by default disables the hierarchy feature. Support +can be enabled by writing 1 to memory.use_hierarchy file of the root cgroup + +# echo 1 > memory.use_hierarchy + +The feature can be disabled by + +# echo 0 > memory.use_hierarchy + +NOTE1: Enabling/disabling will fail if the cgroup already has other +cgroups created below it. + +NOTE2: This feature can be enabled/disabled per subtree. -5. TODO +7. TODO 1. Add support for accounting huge pages (as a separate controller) 2. Make per-cgroup scanner reclaim not-shared pages first |