[RFC][mmotm] Documentation update

From: KAMEZAWA Hiroyuki
Date: Wed Dec 03 2008 - 23:22:21 EST

mmotm-2008-12-03 includes memcg-explain-details-and-test-document.patch

But it's still rough and not complete.
I'd like to update it to be readable becasue memcg is a blackbox for the most
of developpers but it has many hooks to global VM.

If you have reuqest as "explain xxx in detail", please tell me.
I'll not send this out for a while.

Of course, your own patch is welcome.


Documentation/controllers/memcg_test.txt | 216 ++++++++++++++++++++++++++-----
mm/memcontrol.c | 4
2 files changed, 189 insertions(+), 31 deletions(-)

Index: mmotm-2.6.28-Dec03/Documentation/controllers/memcg_test.txt
--- mmotm-2.6.28-Dec03.orig/Documentation/controllers/memcg_test.txt
+++ mmotm-2.6.28-Dec03/Documentation/controllers/memcg_test.txt
@@ -1,59 +1,74 @@
Memory Resource Controller(Memcg) Implementation Memo.
-Last Updated: 2009/12/03
+Last Updated: 2009/12/04
+Base Kernel Version: 2.6.28-rc7-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 and some test
-patterns tend to be racy.
+patterns tend to be racy. Please note that explanation about implementation
+details can be very old.

-1. charges
+(*) 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.
+ swap_cgroup ... an entry per swp_entry
+ Both of them are allocated at boot.
+ 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

- called at new page fault and COW.
+ Called at new page fault and Copy-On-Write.

- called at do_swap_page() and swapoff.
- followed by charge-commit-cancel protocol.
- (With swap accounting) at commit, charges recorded in swap is removed.
+ Called at do_swap_page() (page fault on swap entry) and swapoff.
+ Followed by charge-commit-cancel protocol. (With swap accounting)
+ At commit, charges recorded in swap_cgroup is removed.

- called at add_to_page_cache()
+ Called at add_to_page_cache()

- mem_cgroup_cache_charge_swapin)()
- called by shmem's swapin processing.
+ mem_cgroup_cache_charge_swapin()
+ Called at shmem's swapin.

- called before migration. "extra" charge is done
- followed by charge-commit-cancel protocol.
+ 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
+2. Uncharge
a page/swp_entry may be uncharged (usage -= PAGE_SIZE) by

- called when an anonymous page is unmapped. If the page is SwapCache
- uncharge is delayed until mem_cgroup_uncharge_swapcache().
+ 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().

- called when a page-cache is deleted from radix-tree. If the page is
+ Called when a page-cache is deleted from radix-tree. If the page is
SwapCache, uncharge is delayed until mem_cgroup_uncharge_swapcache()

- called when SwapCache is removed from radix-tree. the charge itself
+ 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.)

- called when swp_entry's refcnt goes down to be 0. charge against swap
+ Called when swp_entry's refcnt goes down to be 0. Charge against swap

mem_cgroup_end_migration(old, new)
- at success of migration -> old is uncharged (if necessary), charge
- to new is committed. at failure, charge to old is committed.
+ 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 charge.
+ 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.
@@ -68,23 +83,153 @@ patterns tend to be racy.

At cancel(), simply usage -= PAGE_SIZE.

-4. Typical Tests.
+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.
+ Yet 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 have been mapped by some process, it has been
+ charged already.
+ In case (a), we charge it. In case (b), we don't charge it.
+ (But racy state between (a) and (b) exists. We do check it.)
+ At charging, a charge recorded in swap_cgroup is moved to page_cgroup.
+ 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 exits,
+ (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 logic is page-migration-handler.
+ memcg have 2 routine. Assume migrate page's contents from OLDPAGE
+ 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 the all works.
+ The point is....
+ - If OLDPAGE is anonymous, all charges will be dropped at (d)
+ - If OLDPAGE is SwapCache, charges will be kept at (g)
+ - If OLDPAGE is page-cache, charges will be kept at (g)
+ At (e-1)(e-2) and (f), there are no hooks of memcg.
+ memcg provides following hooks.
+ - mem_cgroup_prepare_migration(OLDPAGE)
+ Called at (b) and 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 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().
+ One special function is mem_cgroup_isolate_pages(). This scans
+ memcg's private LRU and call __isolate_lru_page().
+ (By __isolate_lru_page(), the page is removed from private LRU, too)

- Tests for racy cases.

- 4.1 small limit to memcg.
+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.)

- 4.2 shmem
+ 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.

- 4.3 migration
+ 9.3 Migration
For NUMA, migration is an another special. To do easy test, cpuset
is useful. Following is a sample script to do migration.

@@ -118,20 +263,20 @@ patterns tend to be racy.
G2_TASK=`cat ${G2}/tasks`
move_task "${G1_TASK}" ${G2} &
- 4.4 memory hotplug.
+ 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.

- 4.5 mkdir/rmdir
+ 9.5 mkdir/rmdir
When using hierarchy, mkdir/rmdir test should be done.
tests like following.

- #echo 1 >/opt/cgroup/01/memory/use_hierarchy
- #mkdir /opt/cgroup/01/child_a
- #mkdir /opt/cgroup/01/child_b
+ 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
@@ -143,3 +288,12 @@ patterns tend to be racy.

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 ia 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.
Index: mmotm-2.6.28-Dec03/mm/memcontrol.c
--- mmotm-2.6.28-Dec03.orig/mm/memcontrol.c
+++ mmotm-2.6.28-Dec03/mm/memcontrol.c
@@ -6,6 +6,10 @@
* Copyright 2007 OpenVZ SWsoft Inc
* Author: Pavel Emelianov <xemul@xxxxxxxxxx>
+ * Documentations are available at
+ * Documentation/controllers/memory.txt
+ * Documentation/controllers/memcg_test.txt
+ *
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/