linux/include
KAMEZAWA Hiroyuki 7a81b88cb5 memcg: introduce charge-commit-cancel style of functions
There is a small race in do_swap_page().  When the page swapped-in is
charged, the mapcount can be greater than 0.  But, at the same time some
process (shares it ) call unmap and make mapcount 1->0 and the page is
uncharged.

      CPUA 			CPUB
       mapcount == 1.
   (1) charge if mapcount==0     zap_pte_range()
                                (2) mapcount 1 => 0.
			        (3) uncharge(). (success)
   (4) set page's rmap()
       mapcount 0=>1

Then, this swap page's account is leaked.

For fixing this, I added a new interface.
  - charge
   account to res_counter by PAGE_SIZE and try to free pages if necessary.
  - commit
   register page_cgroup and add to LRU if necessary.
  - cancel
   uncharge PAGE_SIZE because of do_swap_page failure.

     CPUA
  (1) charge (always)
  (2) set page's rmap (mapcount > 0)
  (3) commit charge was necessary or not after set_pte().

This protocol uses PCG_USED bit on page_cgroup for avoiding over accounting.
Usual mem_cgroup_charge_common() does charge -> commit at a time.

And this patch also adds following function to clarify all charges.

  - mem_cgroup_newpage_charge() ....replacement for mem_cgroup_charge()
	called against newly allocated anon pages.

  - mem_cgroup_charge_migrate_fixup()
        called only from remove_migration_ptes().
	we'll have to rewrite this later.(this patch just keeps old behavior)
	This function will be removed by additional patch to make migration
	clearer.

Good for clarifying "what we do"

Then, we have 4 following charge points.
  - newpage
  - swap-in
  - add-to-cache.
  - migration.

[akpm@linux-foundation.org: add missing inline directives to stubs]
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Cc: Balbir Singh <balbir@in.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 08:31:04 -08:00
..
acpi trivial: fix an -> a typos in documentation and comments 2009-01-06 11:28:07 +01:00
asm-arm
asm-frv frv: introduce asm/swab.h 2009-01-06 18:10:28 -08:00
asm-generic remove linux/hardirq.h from asm-generic/local.h 2009-01-06 15:59:13 -08:00
asm-h8300
asm-m32r m32r: introduce asm/swab.h 2009-01-06 18:10:28 -08:00
asm-m68k m68k: introduce asm/swab.h 2009-01-06 18:10:27 -08:00
asm-mn10300 mn10300: introduce asm/swab.h 2009-01-06 18:10:29 -08:00
crypto crypto: aes - Precompute tables 2008-12-25 11:05:13 +11:00
drm drm: Add a debug node for vblank state. 2008-12-29 17:47:27 +10:00
keys KEYS: Disperse linux/key_ui.h 2008-11-14 10:39:13 +11:00
linux memcg: introduce charge-commit-cancel style of functions 2009-01-08 08:31:04 -08:00
math-emu
media V4L/DVB (10141): v4l2: debugging API changed to match against driver name instead of ID. 2009-01-02 17:11:52 -02:00
mtd trivial: fix then -> than typos in comments and documentation 2009-01-06 11:28:06 +01:00
net wimax: headers for kernel API and user space interaction 2009-01-07 10:00:16 -08:00
pcmcia
rdma
rxrpc
scsi [SCSI] fcoe: Fibre Channel over Ethernet 2008-12-29 11:24:33 -06:00
sound Merge branch 'topic/asoc' into for-linus 2009-01-06 09:48:51 +01:00
trace sched, trace: update trace_sched_wakeup() 2008-12-25 13:10:21 +01:00
video video: sh_mobile_lcdcfb deferred io support 2008-12-22 18:44:48 +09:00
xen xen: add xenfs to allow usermode <-> Xen interaction 2009-01-08 08:30:59 -08:00
Kbuild