)]}' { "log": [ { "commit": "50de1dd967d4ba3b8a90ebe7a4f5feca24191317", "tree": "ed2aa4352ae3c2887a90ef6279588a7048397d1b", "parents": [ "17295c88a160c6eea3fcf46cec9d08a0fcb02db9" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Thu Jan 13 15:47:43 2011 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 13 17:32:51 2011 -0800" }, "message": "memcg: fix memory migration of shmem swapcache\n\nIn the current implementation mem_cgroup_end_migration() decides whether\nthe page migration has succeeded or not by checking \"oldpage-\u003emapping\".\n\nBut if we are tring to migrate a shmem swapcache, the page-\u003emapping of it\nis NULL from the begining, so the check would be invalid. As a result,\nmem_cgroup_end_migration() assumes the migration has succeeded even if\nit\u0027s not, so \"newpage\" would be freed while it\u0027s not uncharged.\n\nThis patch fixes it by passing mem_cgroup_end_migration() the result of\nthe page migration.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nReviewed-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nReviewed-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nCc: Hugh Dickins \u003chughd@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "17295c88a160c6eea3fcf46cec9d08a0fcb02db9", "tree": "35fe45824b5ddd6d406d2d7d7562438a4725e57e", "parents": [ "dfe076b0971a783469bc2066e85d46e23c8acb1c" ], "author": { "name": "Jesper Juhl", "email": "jj@chaosbits.net", "time": "Thu Jan 13 15:47:42 2011 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 13 17:32:51 2011 -0800" }, "message": "memcg: use [kv]zalloc[_node] rather than [kv]malloc+memset\n\nIn mem_cgroup_alloc() we currently do either kmalloc() or vmalloc() then\nfollowed by memset() to zero the memory. This can be more efficiently\nachieved by using kzalloc() and vzalloc(). There\u0027s also one situation\nwhere we can use kzalloc_node() - this is what\u0027s new in this version of\nthe patch.\n\nSigned-off-by: Jesper Juhl \u003cjj@chaosbits.net\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nCc: Wu Fengguang \u003cfengguang.wu@intel.com\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "dfe076b0971a783469bc2066e85d46e23c8acb1c", "tree": "ee0856c119ed898212da2555f936fb3284c5ff59", "parents": [ "043d18b1e5bdfc4870b8a19d00f0d5c636a5c231" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Thu Jan 13 15:47:41 2011 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 13 17:32:51 2011 -0800" }, "message": "memcg: fix deadlock between cpuset and memcg\n\nCommit b1dd693e (\"memcg: avoid deadlock between move charge and\ntry_charge()\") can cause another deadlock about mmap_sem on task migration\nif cpuset and memcg are mounted onto the same mount point.\n\nAfter the commit, cgroup_attach_task() has sequence like:\n\ncgroup_attach_task()\n ss-\u003ecan_attach()\n cpuset_can_attach()\n mem_cgroup_can_attach()\n down_read(\u0026mmap_sem) (1)\n ss-\u003eattach()\n cpuset_attach()\n mpol_rebind_mm()\n down_write(\u0026mmap_sem) (2)\n up_write(\u0026mmap_sem)\n cpuset_migrate_mm()\n do_migrate_pages()\n down_read(\u0026mmap_sem)\n up_read(\u0026mmap_sem)\n mem_cgroup_move_task()\n mem_cgroup_clear_mc()\n up_read(\u0026mmap_sem)\n\nWe can cause deadlock at (2) because we\u0027ve already aquire the mmap_sem at (1).\n\nBut the commit itself is necessary to fix deadlocks which have existed\nbefore the commit like:\n\nEx.1)\n move charge | try charge\n --------------------------------------+------------------------------\n mem_cgroup_can_attach() | down_write(\u0026mmap_sem)\n mc.moving_task \u003d current | ..\n mem_cgroup_precharge_mc() | __mem_cgroup_try_charge()\n mem_cgroup_count_precharge() | prepare_to_wait()\n down_read(\u0026mmap_sem) | if (mc.moving_task)\n -\u003e cannot aquire the lock | -\u003e true\n | schedule()\n | -\u003e move charge should wake it up\n\nEx.2)\n move charge | try charge\n --------------------------------------+------------------------------\n mem_cgroup_can_attach() |\n mc.moving_task \u003d current |\n mem_cgroup_precharge_mc() |\n mem_cgroup_count_precharge() |\n down_read(\u0026mmap_sem) |\n .. |\n up_read(\u0026mmap_sem) |\n | down_write(\u0026mmap_sem)\n mem_cgroup_move_task() | ..\n mem_cgroup_move_charge() | __mem_cgroup_try_charge()\n down_read(\u0026mmap_sem) | prepare_to_wait()\n -\u003e cannot aquire the lock | if (mc.moving_task)\n | -\u003e true\n | schedule()\n | -\u003e move charge should wake it up\n\nThis patch fixes all of these problems by:\n1. revert the commit.\n2. To fix the Ex.1, we set mc.moving_task after mem_cgroup_count_precharge()\n has released the mmap_sem.\n3. To fix the Ex.2, we use down_read_trylock() instead of down_read() in\n mem_cgroup_move_charge() and, if it has failed to aquire the lock, cancel\n all extra charges, wake up all waiters, and retry trylock.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nReported-by: Ben Blum \u003cbblum@andrew.cmu.edu\u003e\nCc: Miao Xie \u003cmiaox@cn.fujitsu.com\u003e\nCc: David Rientjes \u003crientjes@google.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Hiroyuki Kamezawa \u003ckamezawa.hiroyuki@gmail.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "043d18b1e5bdfc4870b8a19d00f0d5c636a5c231", "tree": "c2392e1fb2c1f7b23446ae4c59aa2d35134ab3d3", "parents": [ "f3e8eb70b1807d1b30aa6972af0cf30077c40112" ], "author": { "name": "Minchan Kim", "email": "minchan.kim@gmail.com", "time": "Thu Jan 13 15:47:40 2011 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 13 17:32:50 2011 -0800" }, "message": "memcg: remove unnecessary return from void-returning mem_cgroup_del_lru_list()\n\nSigned-off-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nAcked-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "f3e8eb70b1807d1b30aa6972af0cf30077c40112", "tree": "5b541e4850fa6a9922b288e9efd7c632021657f6", "parents": [ "dbd4ea78f002df283c95d9774837041735fa1bf9" ], "author": { "name": "Johannes Weiner", "email": "hannes@cmpxchg.org", "time": "Thu Jan 13 15:47:39 2011 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 13 17:32:50 2011 -0800" }, "message": "memcg: fix unit mismatch in memcg oom limit calculation\n\nAdding the number of swap pages to the byte limit of a memory control\ngroup makes no sense. Convert the pages to bytes before adding them.\n\nThe only user of this code is the OOM killer, and the way it is used means\nthat the error results in a higher OOM badness value. Since the cgroup\nlimit is the same for all tasks in the cgroup, the error should have no\npractical impact at the moment.\n\nBut let\u0027s not wait for future or changing users to trip over it.\n\nSigned-off-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nCc: Greg Thelen \u003cgthelen@google.com\u003e\nCc: David Rientjes \u003crientjes@google.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "dbd4ea78f002df283c95d9774837041735fa1bf9", "tree": "e709c6c5c026b1c230bb87ddcfe8415aaf255820", "parents": [ "2a7106f2cb0768d00fe8c1eb42a754a7d8518f08" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Thu Jan 13 15:47:38 2011 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 13 17:32:50 2011 -0800" }, "message": "memcg: add lock to synchronize page accounting and migration\n\nIntroduce a new bit spin lock, PCG_MOVE_LOCK, to synchronize the page\naccounting and migration code. This reworks the locking scheme of\n_update_stat() and _move_account() by adding new lock bit PCG_MOVE_LOCK,\nwhich is always taken under IRQ disable.\n\n1. If pages are being migrated from a memcg, then updates to that\n memcg page statistics are protected by grabbing PCG_MOVE_LOCK using\n move_lock_page_cgroup(). In an upcoming commit, memcg dirty page\n accounting will be updating memcg page accounting (specifically: num\n writeback pages) from IRQ context (softirq). Avoid a deadlocking\n nested spin lock attempt by disabling irq on the local processor when\n grabbing the PCG_MOVE_LOCK.\n\n2. lock for update_page_stat is used only for avoiding race with\n move_account(). So, IRQ awareness of lock_page_cgroup() itself is not\n a problem. The problem is between mem_cgroup_update_page_stat() and\n mem_cgroup_move_account_page().\n\nTrade-off:\n * Changing lock_page_cgroup() to always disable IRQ (or\n local_bh) has some impacts on performance and I think\n it\u0027s bad to disable IRQ when it\u0027s not necessary.\n * adding a new lock makes move_account() slower. Score is\n here.\n\nPerformance Impact: moving a 8G anon process.\n\nBefore:\n\treal 0m0.792s\n\tuser 0m0.000s\n\tsys 0m0.780s\n\nAfter:\n\treal 0m0.854s\n\tuser 0m0.000s\n\tsys 0m0.842s\n\nThis score is bad but planned patches for optimization can reduce\nthis impact.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Greg Thelen \u003cgthelen@google.com\u003e\nReviewed-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nAcked-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Andrea Righi \u003carighi@develer.com\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Wu Fengguang \u003cfengguang.wu@intel.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "2a7106f2cb0768d00fe8c1eb42a754a7d8518f08", "tree": "730bef06e752c1edcb2d475fd193f94bea00bf6a", "parents": [ "ece72400c2a27a3d726cb0854449f991d9fcd2da" ], "author": { "name": "Greg Thelen", "email": "gthelen@google.com", "time": "Thu Jan 13 15:47:37 2011 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 13 17:32:50 2011 -0800" }, "message": "memcg: create extensible page stat update routines\n\nReplace usage of the mem_cgroup_update_file_mapped() memcg\nstatistic update routine with two new routines:\n* mem_cgroup_inc_page_stat()\n* mem_cgroup_dec_page_stat()\n\nAs before, only the file_mapped statistic is managed. However, these more\ngeneral interfaces allow for new statistics to be more easily added. New\nstatistics are added with memcg dirty page accounting.\n\nSigned-off-by: Greg Thelen \u003cgthelen@google.com\u003e\nSigned-off-by: Andrea Righi \u003carighi@develer.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nCc: Wu Fengguang \u003cfengguang.wu@intel.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "37c2ac7872a9387542616f658d20ac25f5bdb32e", "tree": "d115915db4a61e261012bf0f9c4cf14630243d71", "parents": [ "91600e9e592e48736e630851c83da2ad6bf0e91f" ], "author": { "name": "Andrea Arcangeli", "email": "aarcange@redhat.com", "time": "Thu Jan 13 15:47:16 2011 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 13 17:32:47 2011 -0800" }, "message": "thp: compound_trans_order\n\nRead compound_trans_order safe. Noop for CONFIG_TRANSPARENT_HUGEPAGE\u003dn.\n\nSigned-off-by: Andrea Arcangeli \u003caarcange@redhat.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "2c888cfbc1b45508a44763d85ba2e8ac43faff5f", "tree": "9a7f2214e5d6a01d5724ae63d4d50cddeb2293ff", "parents": [ "97562cd243298acf573620c764a1037bd545c9bc" ], "author": { "name": "Rik van Riel", "email": "riel@redhat.com", "time": "Thu Jan 13 15:47:13 2011 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 13 17:32:46 2011 -0800" }, "message": "thp: fix anon memory statistics with transparent hugepages\n\nCount each transparent hugepage as HPAGE_PMD_NR pages in the LRU\nstatistics, so the Active(anon) and Inactive(anon) statistics in\n/proc/meminfo are correct.\n\nSigned-off-by: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: Andrea Arcangeli \u003caarcange@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "152c9ccb75548c027fa3103efa4fa4e19a345449", "tree": "d22e8f71b3e0689fc0e3ae21609ae1abe5bc4d2f", "parents": [ "ec1685109f1314a30919489ef2800ed626a38c1e" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Thu Jan 13 15:46:56 2011 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 13 17:32:43 2011 -0800" }, "message": "thp: transhuge-memcg: commit tail pages at charge\n\nBy this patch, when a transparent hugepage is charged, not only the head\npage but also all the tail pages are committed, IOW pc-\u003emem_cgroup and\npc-\u003eflags of tail pages are set.\n\nWithout this patch:\n\n- Tail pages are not linked to any memcg\u0027s LRU at splitting. This causes many\n problems, for example, the charged memcg\u0027s directory can never be rmdir\u0027ed\n because it doesn\u0027t have enough pages to scan to make the usage decrease to 0.\n- \"rss\" field in memory.stat would be incorrect. Moreover, usage_in_bytes in\n root cgroup is calculated by the stat not by res_counter(since 2.6.32),\n it would be incorrect too.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrea Arcangeli \u003caarcange@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "ec1685109f1314a30919489ef2800ed626a38c1e", "tree": "48ac32238182623af24b6c25be456c0ca432047b", "parents": [ "500d65d471018d9a13b0d51b7e141ed2a3555c1d" ], "author": { "name": "Andrea Arcangeli", "email": "aarcange@redhat.com", "time": "Thu Jan 13 15:46:56 2011 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 13 17:32:43 2011 -0800" }, "message": "thp: memcg compound\n\nTeach memcg to charge/uncharge compound pages.\n\nSigned-off-by: Andrea Arcangeli \u003caarcange@redhat.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "ebb76ce16daf6908dc030dec1c00827d37129fe5", "tree": "28abdc67466bb5d8d4566e36c914518b0533b4bf", "parents": [ "b83be6f20a0e468f715b14225c9f897538dfe5ad" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed Dec 29 14:07:11 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Dec 30 10:07:06 2010 -0800" }, "message": "memcg: fix wrong VM_BUG_ON() in try_charge()\u0027s mm-\u003eowner check\n\nAt __mem_cgroup_try_charge(), VM_BUG_ON(!mm-\u003eowner) is checked.\nBut as commented in mem_cgroup_from_task(), mm-\u003eowner can be NULL\nin some racy case. This check of VM_BUG_ON() is bad.\n\nA possible story to hit this is at swapoff()-\u003etry_to_unuse(). It passes\nmm_struct to mem_cgroup_try_charge_swapin() while mm-\u003eowner is NULL. If we\ncan\u0027t get proper mem_cgroup from swap_cgroup information, mm-\u003eowner is used\nas charge target and we see NULL.\n\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nReported-by: Hugh Dickins \u003chughd@google.com\u003e\nReported-by: Thomas Meyer \u003cthomas@m3y3r.de\u003e\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nReviewed-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nSigned-off-by: Hugh Dickins \u003chughd@google.com\u003e\nCc: stable@kernel.org\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "a42c390cfa0c2612459d7226ba11612847ca3a64", "tree": "1dc06659b4f232461eb6852b8c4d17f8305aca91", "parents": [ "b1dd693e5b9348bd68a80e679e03cf9c0973b01b" ], "author": { "name": "Michal Hocko", "email": "mhocko@suse.cz", "time": "Wed Nov 24 12:57:08 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Nov 25 06:50:45 2010 +0900" }, "message": "cgroups: make swap accounting default behavior configurable\n\nSwap accounting can be configured by CONFIG_CGROUP_MEM_RES_CTLR_SWAP\nconfiguration option and then it is turned on by default. There is a boot\noption (noswapaccount) which can disable this feature.\n\nThis makes it hard for distributors to enable the configuration option as\nthis feature leads to a bigger memory consumption and this is a no-go for\ngeneral purpose distribution kernel. On the other hand swap accounting\nmay be very usuful for some workloads.\n\nThis patch adds a new configuration option which controls the default\nbehavior (CGROUP_MEM_RES_CTLR_SWAP_ENABLED). If the option is selected\nthen the feature is turned on by default.\n\nIt also adds a new boot parameter swapaccount[\u003d1|0] which enhances the\noriginal noswapaccount parameter semantic by means of enable/disable logic\n(defaults to 1 if no value is provided to be still consistent with\nnoswapaccount).\n\nThe default behavior is unchanged (if CONFIG_CGROUP_MEM_RES_CTLR_SWAP is\nenabled then CONFIG_CGROUP_MEM_RES_CTLR_SWAP_ENABLED is enabled as well)\n\nSigned-off-by: Michal Hocko \u003cmhocko@suse.cz\u003e\nAcked-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "b1dd693e5b9348bd68a80e679e03cf9c0973b01b", "tree": "557c8e0634fba47ea61ed9fbb8742d907c567944", "parents": [ "11e7946f196e5fdde20584e3e58c60335ee3b3bc" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Wed Nov 24 12:57:06 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Nov 25 06:50:44 2010 +0900" }, "message": "memcg: avoid deadlock between move charge and try_charge()\n\n__mem_cgroup_try_charge() can be called under down_write(\u0026mmap_sem)(e.g.\nmlock does it). This means it can cause deadlock if it races with move charge:\n\nEx.1)\n move charge | try charge\n --------------------------------------+------------------------------\n mem_cgroup_can_attach() | down_write(\u0026mmap_sem)\n mc.moving_task \u003d current | ..\n mem_cgroup_precharge_mc() | __mem_cgroup_try_charge()\n mem_cgroup_count_precharge() | prepare_to_wait()\n down_read(\u0026mmap_sem) | if (mc.moving_task)\n -\u003e cannot aquire the lock | -\u003e true\n | schedule()\n\nEx.2)\n move charge | try charge\n --------------------------------------+------------------------------\n mem_cgroup_can_attach() |\n mc.moving_task \u003d current |\n mem_cgroup_precharge_mc() |\n mem_cgroup_count_precharge() |\n down_read(\u0026mmap_sem) |\n .. |\n up_read(\u0026mmap_sem) |\n | down_write(\u0026mmap_sem)\n mem_cgroup_move_task() | ..\n mem_cgroup_move_charge() | __mem_cgroup_try_charge()\n down_read(\u0026mmap_sem) | prepare_to_wait()\n -\u003e cannot aquire the lock | if (mc.moving_task)\n | -\u003e true\n | schedule()\n\nTo avoid this deadlock, we do all the move charge works (both can_attach() and\nattach()) under one mmap_sem section.\nAnd after this patch, we set/clear mc.moving_task outside mc.lock, because we\nuse the lock only to check mc.from/to.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: \u003cstable@kernel.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "112bc2e120a94a511858918d6866a4978f9c500e", "tree": "032f59b4eca42074de6a30e0d4c13426cbbf86c8", "parents": [ "69e83dad5207f8f03c9699e57e1febb114383cb8" ], "author": { "name": "Kirill A. Shutemov", "email": "kirill@shutemov.name", "time": "Wed Nov 24 12:56:58 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Nov 25 06:50:40 2010 +0900" }, "message": "memcg: fix false positive VM_BUG on non-SMP\n\nFix this:\n\n kernel BUG at mm/memcontrol.c:2155!\n invalid opcode: 0000 [#1]\n last sysfs file:\n\n Pid: 18, comm: sh Not tainted 2.6.37-rc3 #3 /Bochs\n EIP: 0060:[\u003cc10731b2\u003e] EFLAGS: 00000246 CPU: 0\n EIP is at mem_cgroup_move_account+0xe2/0xf0\n EAX: 00000004 EBX: c6f931d4 ECX: c681c300 EDX: c681c000\n ESI: c681c300 EDI: ffffffea EBP: c681c000 ESP: c46f3e30\n DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 0068\n Process sh (pid: 18, ti\u003dc46f2000 task\u003dc6826e60 task.ti\u003dc46f2000)\n Stack:\n 00000155 c681c000 0805f000 c46ee180 c46f3e5c c7058820 c1074d37 00000000\n 08060000 c46db9a0 c46ec080 c7058820 0805f000 08060000 c46f3e98 c1074c50\n c106c75e c46f3e98 c46ec080 08060000 0805ffff c46db9a0 c46f3e98 c46e0340\n Call Trace:\n [\u003cc1074d37\u003e] ? mem_cgroup_move_charge_pte_range+0xe7/0x130\n [\u003cc1074c50\u003e] ? mem_cgroup_move_charge_pte_range+0x0/0x130\n [\u003cc106c75e\u003e] ? walk_page_range+0xee/0x1d0\n [\u003cc10725d6\u003e] ? mem_cgroup_move_task+0x66/0x90\n [\u003cc1074c50\u003e] ? mem_cgroup_move_charge_pte_range+0x0/0x130\n [\u003cc1072570\u003e] ? mem_cgroup_move_task+0x0/0x90\n [\u003cc1042616\u003e] ? cgroup_attach_task+0x136/0x200\n [\u003cc1042878\u003e] ? cgroup_tasks_write+0x48/0xc0\n [\u003cc1041e9e\u003e] ? cgroup_file_write+0xde/0x220\n [\u003cc101398d\u003e] ? do_page_fault+0x17d/0x3f0\n [\u003cc108a79d\u003e] ? alloc_fd+0x2d/0xd0\n [\u003cc1041dc0\u003e] ? cgroup_file_write+0x0/0x220\n [\u003cc1077ba2\u003e] ? vfs_write+0x92/0xc0\n [\u003cc1077c81\u003e] ? sys_write+0x41/0x70\n [\u003cc1140e3d\u003e] ? syscall_call+0x7/0xb\n Code: 03 00 74 09 8b 44 24 04 e8 1c f1 ff ff 89 73 04 8d 86 b0 00 00 00 b9 01 00 00 00 89 da 31 ff e8 65 f5 ff ff e9 4d ff ff ff 0f 0b \u003c0f\u003e 0b 0f 0b 0f 0b 90 8d b4 26 00 00 00 00 83 ec 10 8b 0d f4 e3\n EIP: [\u003cc10731b2\u003e] mem_cgroup_move_account+0xe2/0xf0 SS:ESP 0068:c46f3e30\n ---[ end trace 7daa1582159b6532 ]---\n\nlock_page_cgroup and unlock_page_cgroup are implemented using\nbit_spinlock. bit_spinlock doesn\u0027t touch the bit if we are on non-SMP\nmachine, so we can\u0027t use the bit to check whether the lock was taken.\n\nLet\u0027s introduce is_page_cgroup_locked based on bit_spin_is_locked instead\nof PageCgroupLocked to fix it.\n\n[akpm@linux-foundation.org: s/is_page_cgroup_locked/page_is_cgroup_locked/]\nSigned-off-by: Kirill A. Shutemov \u003ckirill@shutemov.name\u003e\nReviewed-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujtisu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "d2e61b8dc99fdb36e0fd176e25365f69afda4ff9", "tree": "580325e7a44fc02d985bca17a1d7601e6ed81cb9", "parents": [ "1093736b3c34319b8f1825a4423414d9cf397d73" ], "author": { "name": "Dan Carpenter", "email": "error27@gmail.com", "time": "Thu Nov 11 14:05:12 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Nov 12 07:55:31 2010 -0800" }, "message": "memcg: null dereference on allocation failure\n\nThe original code had a null dereference if alloc_percpu() failed. This\nwas introduced in commit 711d3d2c9bc3 (\"memcg: cpu hotplug aware percpu\ncount updates\")\n\nSigned-off-by: Dan Carpenter \u003cerror27@gmail.com\u003e\nReviewed-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "26174efd42100eefac67732c0c12f41a205fa335", "tree": "81a7d758a3fcfd755e9e94ee7f66854133c281cf", "parents": [ "1489ebad8b5b20300562f634f279cb9c435fd90b" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed Oct 27 15:33:43 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Oct 27 18:03:10 2010 -0700" }, "message": "memcg: generic filestat update interface\n\nThis patch extracts the core logic from mem_cgroup_update_file_mapped() as\nmem_cgroup_update_file_stat() and adds a wrapper.\n\nAs a planned future update, memory cgroup has to count dirty pages to\nimplement dirty_ratio/limit. And more, the number of dirty pages is\nrequired to kick flusher thread to start writeback. (Now, no kick.)\n\nThis patch is preparation for it and makes other statistics implementation\nclearer. Just a clean up.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nReviewed-by: Greg Thelen \u003cgthelen@google.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "1489ebad8b5b20300562f634f279cb9c435fd90b", "tree": "4bfa738b0733c11120705aaa37a45d87d1dd5534", "parents": [ "711d3d2c9bc3fb7cb5116352fecdb5b4adb6db6e" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed Oct 27 15:33:42 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Oct 27 18:03:09 2010 -0700" }, "message": "memcg: cpu hotplug aware quick acount_move detection\n\nAn event counter MEM_CGROUP_ON_MOVE is used for quick check whether file\nstat update can be done in async manner or not. Now, it use percpu\ncounter and for_each_possible_cpu to update.\n\nThis patch replaces for_each_possible_cpu to for_each_online_cpu and adds\nnecessary synchronization logic at CPU HOTPLUG.\n\n[akpm@linux-foundation.org: coding-style fixes]\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "711d3d2c9bc3fb7cb5116352fecdb5b4adb6db6e", "tree": "09979a5e3d7b2e1fe7b7de193d11d6f468a67e76", "parents": [ "7d74b06f240f1bd1b4b68dd6fe84164d8bf4e315" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed Oct 27 15:33:42 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Oct 27 18:03:09 2010 -0700" }, "message": "memcg: cpu hotplug aware percpu count updates\n\nNow, memcgroup\u0027s per cpu coutner uses for_each_possible_cpu() to get the\nvalue. It\u0027s better to use for_each_online_cpu() and a cpu hotplug\nhandler.\n\nThis patch only handles statistics counter. MEM_CGROUP_ON_MOVE will be\nhandled in another patch.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "7d74b06f240f1bd1b4b68dd6fe84164d8bf4e315", "tree": "3b52304b322390a9b5782b4d498e5e0a14e3266b", "parents": [ "32047e2a85f06633ee4c53e2d0346fbcd34e480b" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed Oct 27 15:33:41 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Oct 27 18:03:09 2010 -0700" }, "message": "memcg: use for_each_mem_cgroup\n\nIn memory cgroup management, we sometimes have to walk through\nsubhierarchy of cgroup to gather informaiton, or lock something, etc.\n\nNow, to do that, mem_cgroup_walk_tree() function is provided. It calls\ngiven callback function per cgroup found. But the bad thing is that it\nhas to pass a fixed style function and argument, \"void*\" and it adds much\ntype casting to memcontrol.c.\n\nTo make the code clean, this patch replaces walk_tree() with\n\n for_each_mem_cgroup_tree(iter, root)\n\nAn iterator style call. The good point is that iterator call doesn\u0027t have\nto assume what kind of function is called under it. A bad point is that\nit may cause reference-count leak if a caller use \"break\" from the loop by\nmistake.\n\nI think the benefit is larger. The modified code seems straigtforward and\neasy to read because we don\u0027t have misterious callbacks and pointer cast.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "32047e2a85f06633ee4c53e2d0346fbcd34e480b", "tree": "25ed1e04bf60a46951581b0ad28a45e51b1602a2", "parents": [ "0c270f8f9988fb0d93ea214fdcff7ab90eb3d894" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed Oct 27 15:33:40 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Oct 27 18:03:09 2010 -0700" }, "message": "memcg: avoid lock in updating file_mapped (Was fix race in file_mapped accouting flag management\n\nAt accounting file events per memory cgroup, we need to find memory cgroup\nvia page_cgroup-\u003emem_cgroup. Now, we use lock_page_cgroup() for guarantee\npc-\u003emem_cgroup is not overwritten while we make use of it.\n\nBut, considering the context which page-cgroup for files are accessed,\nwe can use alternative light-weight mutual execusion in the most case.\n\nAt handling file-caches, the only race we have to take care of is \"moving\"\naccount, IOW, overwriting page_cgroup-\u003emem_cgroup. (See comment in the\npatch)\n\nUnlike charge/uncharge, \"move\" happens not so frequently. It happens only when\nrmdir() and task-moving (with a special settings.)\nThis patch adds a race-checker for file-cache-status accounting v.s. account\nmoving. The new per-cpu-per-memcg counter MEM_CGROUP_ON_MOVE is added.\nThe routine for account move\n 1. Increment it before start moving\n 2. Call synchronize_rcu()\n 3. Decrement it after the end of moving.\nBy this, file-status-counting routine can check it needs to call\nlock_page_cgroup(). In most case, I doesn\u0027t need to call it.\n\nFollowing is a perf data of a process which mmap()/munmap 32MB of file cache\nin a minute.\n\nBefore patch:\n 28.25% mmap mmap [.] main\n 22.64% mmap [kernel.kallsyms] [k] page_fault\n 9.96% mmap [kernel.kallsyms] [k] mem_cgroup_update_file_mapped\n 3.67% mmap [kernel.kallsyms] [k] filemap_fault\n 3.50% mmap [kernel.kallsyms] [k] unmap_vmas\n 2.99% mmap [kernel.kallsyms] [k] __do_fault\n 2.76% mmap [kernel.kallsyms] [k] find_get_page\n\nAfter patch:\n 30.00% mmap mmap [.] main\n 23.78% mmap [kernel.kallsyms] [k] page_fault\n 5.52% mmap [kernel.kallsyms] [k] mem_cgroup_update_file_mapped\n 3.81% mmap [kernel.kallsyms] [k] unmap_vmas\n 3.26% mmap [kernel.kallsyms] [k] find_get_page\n 3.18% mmap [kernel.kallsyms] [k] __do_fault\n 3.03% mmap [kernel.kallsyms] [k] filemap_fault\n 2.40% mmap [kernel.kallsyms] [k] handle_mm_fault\n 2.40% mmap [kernel.kallsyms] [k] do_page_fault\n\nThis patch reduces memcg\u0027s cost to some extent.\n(mem_cgroup_update_file_mapped is called by both of map/unmap)\n\nNote: It seems some more improvements are required..but no idea.\n maybe removing set/unset flag is required.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nReviewed-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Greg Thelen \u003cgthelen@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "0c270f8f9988fb0d93ea214fdcff7ab90eb3d894", "tree": "b4f77a6714d245841ecec535ccbeba535c0d29a4", "parents": [ "45531757b45cae0ce64c5aff08c2534d5a0fa3e7" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed Oct 27 15:33:39 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Oct 27 18:03:09 2010 -0700" }, "message": "memcg: fix race in file_mapped accouting flag management\n\nPresently memory cgroup accounts file-mapped by counter and flag. counter\nis working in the same way with zone_stat but FileMapped flag only exists\nin memcg (for helping move_account).\n\nThis flag can be updated wrongly in a case. Assume CPU0 and CPU1 and a\nthread mapping a page on CPU0, another thread unmapping it on CPU1.\n\n CPU0 \t\tCPU1\n\t\t\t\trmv rmap (mapcount 1-\u003e0)\n add rmap (mapcount 0-\u003e1)\n lock_page_cgroup()\n memcg counter+1\t\t(some delay)\n set MAPPED FLAG.\n unlock_page_cgroup()\n\t\t\t\tlock_page_cgroup()\n\t\t\t\tmemcg counter-1\n\t\t\t\tclear MAPPED flag\n\nIn the above sequence counter is properly updated but FLAG is not. This\nmeans that representing a state by a flag which is maintained by counter\nneeds some special care.\n\nTo handle this, when clearing a flag, this patch check mapcount directly\nand clear the flag only when mapcount \u003d\u003d 0. (if mapcount \u003e0, someone will\nmake it to zero later and flag will be cleared.)\n\nReverse case, dec-after-inc cannot be a problem because page_table_lock()\nworks well for it. (IOW, to make above sequence, 2 processes should touch\nthe same page at once with map/unmap.)\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Greg Thelen \u003cgthelen@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "ad4ca5f4b70236dab5e457ff6567d36f75d2e7c5", "tree": "b3c5f477a79516389201d342dda2a3591cfa4332", "parents": [ "f241e6607b5a5aefa23c652dbe947b7465633984" ], "author": { "name": "Kirill A. Shutemov", "email": "kirill@shutemov.name", "time": "Thu Oct 07 12:59:27 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Oct 07 13:31:21 2010 -0700" }, "message": "memcg: fix thresholds with use_hierarchy \u003d\u003d 1\n\nWe need to check parent\u0027s thresholds if parent has use_hierarchy \u003d\u003d 1 to\nbe sure that parent\u0027s threshold events will be triggered even if parent\nitself is not active (no MEM_CGROUP_EVENTS).\n\nSigned-off-by: Kirill A. Shutemov \u003ckirill@shutemov.name\u003e\nReviewed-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "13d7e3a2dba6a79589ed34dc0b9114d7b5ff9eab", "tree": "b22db0b9ab3af864c7517ca38887048423194e41", "parents": [ "00918b6ab89df8984ca06397cb77994dabd73f9b" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Tue Aug 10 18:03:06 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 11 08:59:19 2010 -0700" }, "message": "memcg: convert to use zone_to_nid() from bare zone-\u003ezone_pgdat-\u003enode_id\n\nWe have zone_to_nid(). this patch convert all existing users of\nzone-\u003ezone_pgdat-\u003enode_id.\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nAcked-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Mel Gorman \u003cmel@csn.ul.ie\u003e\nCc: Nishimura Daisuke \u003cd-nishimura@mtf.biglobe.ne.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "00918b6ab89df8984ca06397cb77994dabd73f9b", "tree": "2ca2f0f0e7f3ca235c254f05759f96f160e3c0ab", "parents": [ "14fec79680f7cc4617d6ba69324e63d4a732986c" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Tue Aug 10 18:03:05 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 11 08:59:19 2010 -0700" }, "message": "memcg: remove nid and zid argument from mem_cgroup_soft_limit_reclaim()\n\nmem_cgroup_soft_limit_reclaim() has zone, nid and zid argument. but nid\nand zid can be calculated from zone. So remove it.\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Mel Gorman \u003cmel@csn.ul.ie\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Nishimura Daisuke \u003cd-nishimura@mtf.biglobe.ne.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "14fec79680f7cc4617d6ba69324e63d4a732986c", "tree": "f8a9b627a03d04ec7c76fb67f8ea66c81c57a92f", "parents": [ "da280d636b83f0f5d92921c99ef5c7d7c3e751cc" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Tue Aug 10 18:03:05 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 11 08:59:19 2010 -0700" }, "message": "memcg: mem_cgroup_shrink_node_zone() doesn\u0027t need sc.nodemask\n\nCurrently mem_cgroup_shrink_node_zone() call shrink_zone() directly. thus\nit doesn\u0027t need to initialize sc.nodemask because shrink_zone() doesn\u0027t\nuse it at all.\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Mel Gorman \u003cmel@csn.ul.ie\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Nishimura Daisuke \u003cd-nishimura@mtf.biglobe.ne.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "f75ca962037ffd639a44fd88933cd9b84c4c4411", "tree": "4773cd57ae45831c11783cd355f9dcb516eb66f0", "parents": [ "158e0a2d1b3cffed8b46cbc56393a1394672ef79" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Tue Aug 10 18:03:02 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 11 08:59:19 2010 -0700" }, "message": "memcg: avoid css_get()\n\nNow, memory cgroup increments css(cgroup subsys state)\u0027s reference count\nper a charged page. And the reference count is kept until the page is\nuncharged. But this has 2 bad effect.\n\n 1. Because css_get/put calls atomic_inc()/dec, heavy call of them\n on large smp will not scale well.\n 2. Because css\u0027s refcnt cannot be in a state as \"ready-to-release\",\n cgroup\u0027s notify_on_release handler can\u0027t work with memcg.\n 3. css\u0027s refcnt is atomic_t, it means smaller than 32bit. Maybe too small.\n\nThis has been a problem since the 1st merge of memcg.\n\nThis is a trial to remove css\u0027s refcnt per a page. Even if we remove\nrefcnt, pre_destroy() does enough synchronization as\n - check res-\u003eusage \u003d\u003d 0.\n - check no pages on LRU.\n\nThis patch removes css\u0027s refcnt per page. Even after this patch, at the\n1st look, it seems css_get() is still called in try_charge().\n\nBut the logic is.\n\n - If a memcg of mm-\u003eowner is cached one, consume_stock() will work.\n At success, return immediately.\n - If consume_stock returns false, css_get() is called and go to\n slow path which may be blocked. At the end of slow path,\n css_put() is called and restart from the start if necessary.\n\nSo, in the fast path, we don\u0027t call css_get() and can avoid access to\nshared counter. This patch can make the most possible case fast.\n\nHere is a result of multi-threaded page fault benchmark.\n\n[Before]\n 25.32% multi-fault-all [kernel.kallsyms] [k] clear_page_c\n 9.30% multi-fault-all [kernel.kallsyms] [k] _raw_spin_lock_irqsave\n 8.02% multi-fault-all [kernel.kallsyms] [k] try_get_mem_cgroup_from_mm \u003c\u003d\u003d\u003d\u003d\u003d(*)\n 7.83% multi-fault-all [kernel.kallsyms] [k] down_read_trylock\n 5.38% multi-fault-all [kernel.kallsyms] [k] __css_put\n 5.29% multi-fault-all [kernel.kallsyms] [k] __alloc_pages_nodemask\n 4.92% multi-fault-all [kernel.kallsyms] [k] _raw_spin_lock_irq\n 4.24% multi-fault-all [kernel.kallsyms] [k] up_read\n 3.53% multi-fault-all [kernel.kallsyms] [k] css_put\n 2.11% multi-fault-all [kernel.kallsyms] [k] handle_mm_fault\n 1.76% multi-fault-all [kernel.kallsyms] [k] __rmqueue\n 1.64% multi-fault-all [kernel.kallsyms] [k] __mem_cgroup_commit_charge\n\n[After]\n 28.41% multi-fault-all [kernel.kallsyms] [k] clear_page_c\n 10.08% multi-fault-all [kernel.kallsyms] [k] _raw_spin_lock_irq\n 9.58% multi-fault-all [kernel.kallsyms] [k] down_read_trylock\n 9.38% multi-fault-all [kernel.kallsyms] [k] _raw_spin_lock_irqsave\n 5.86% multi-fault-all [kernel.kallsyms] [k] __alloc_pages_nodemask\n 5.65% multi-fault-all [kernel.kallsyms] [k] up_read\n 2.82% multi-fault-all [kernel.kallsyms] [k] handle_mm_fault\n 2.64% multi-fault-all [kernel.kallsyms] [k] mem_cgroup_add_lru_list\n 2.48% multi-fault-all [kernel.kallsyms] [k] __mem_cgroup_commit_charge\n\nThen, 8.02% of try_get_mem_cgroup_from_mm() disappears because this patch\nremoves css_tryget() in it. (But yes, this is an extreme case.)\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "158e0a2d1b3cffed8b46cbc56393a1394672ef79", "tree": "a044d53137609fdfc8c1afdde366b3e205c96eda", "parents": [ "73045c47b6facbdf4656e6763c8cb469de4337e2" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Tue Aug 10 18:03:00 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 11 08:59:19 2010 -0700" }, "message": "memcg: use find_lock_task_mm() in memory cgroups oom\n\nWhen the OOM killer scans task, it check a task is under memcg or\nnot when it\u0027s called via memcg\u0027s context.\n\nBut, as Oleg pointed out, a thread group leader may have NULL -\u003emm\nand task_in_mem_cgroup() may do wrong decision. We have to use\nfind_lock_task_mm() in memcg as generic OOM-Killer does.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Oleg Nesterov \u003coleg@redhat.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nReviewed-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "73045c47b6facbdf4656e6763c8cb469de4337e2", "tree": "e791667803261eb9ae5325dcd8c70acefeee9196", "parents": [ "bd0d24bfe8a8f8d2400569740874a67d164d40a9" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Tue Aug 10 18:02:59 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 11 08:59:19 2010 -0700" }, "message": "memcg: remove mem from arg of charge_common\n\nmem_cgroup_charge_common() is always called with @mem \u003d NULL, so it\u0027s\nmeaningless. This patch removes it.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "bd0d24bfe8a8f8d2400569740874a67d164d40a9", "tree": "9d592f6566adacc3e8da3d2c06411a16b2d45763", "parents": [ "2bd9bb206b338888b226e70139a25a67d10007f0" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Tue Aug 10 18:02:59 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 11 08:59:18 2010 -0700" }, "message": "memcg: remove redundant code\n\n- try_get_mem_cgroup_from_mm() calls rcu_read_lock/unlock by itself, so we\n don\u0027t have to call them in task_in_mem_cgroup().\n- *mz is not used in __mem_cgroup_uncharge_common().\n- we don\u0027t have to call lookup_page_cgroup() in mem_cgroup_end_migration()\n after we\u0027ve cleared PCG_MIGRATION of @oldpage.\n- remove empty comment.\n- remove redundant empty line in mem_cgroup_cache_charge().\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nAcked-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "2bd9bb206b338888b226e70139a25a67d10007f0", "tree": "ac556b1bfa52f06fe2998371d10edec38d5cf2a9", "parents": [ "4b53433468c87794b523e4683fbd4e8e8aca1f63" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Tue Aug 10 18:02:58 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 11 08:59:18 2010 -0700" }, "message": "memcg: clean up waiting move acct\n\nNow, for checking a memcg is under task-account-moving, we do css_tryget()\nagainst mc.to and mc.from. But this is just complicating things. This\npatch makes the check easier.\n\nThis patch adds a spinlock to move_charge_struct and guard modification of\nmc.to and mc.from. By this, we don\u0027t have to think about complicated\nraces arount this not-critical path.\n\n[balbir@linux.vnet.ibm.com: don\u0027t crash on a null memcg being passed]\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "4b53433468c87794b523e4683fbd4e8e8aca1f63", "tree": "08f80fd5ec5d824a28aa4e0b15c7e2e83eed9fec", "parents": [ "65e0e811667dedd4f19b268df9d856ecacb629de" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Tue Aug 10 18:02:57 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 11 08:59:18 2010 -0700" }, "message": "memcg: clean up try_charge main loop\n\nmem_cgroup_try_charge() has a big loop in it and seems to be hard to read.\n Most of routines are for slow path. This patch moves codes out from the\nloop and make it clear what\u0027s done.\n\nSummary:\n - refactoring a function to detect a memcg is under acccount move or not.\n - refactoring a function to wait for the end of moving task acct.\n - refactoring a main loop(\u0027s slow path) as a function and make it clear\n why we retry or quit by return code.\n - add fatal_signal_pending() check for bypassing charge loops.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "cc8e970c3ce4d98afa8eb02dbd2526ce57f7611a", "tree": "621ce988444b6df6eb67058b05ca5ddef23f6103", "parents": [ "e17613c39b8894c164df782d0508c27ca559c24b" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Mon Aug 09 17:19:57 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 09 20:45:03 2010 -0700" }, "message": "memcg: add mm_vmscan_memcg_isolate tracepoint\n\nMemcg also need to trace page isolation information as global reclaim.\nThis patch does it.\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nReviewed-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Mel Gorman \u003cmel@csn.ul.ie\u003e\nAcked-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "a63d83f427fbce97a6cea0db2e64b0eb8435cd10", "tree": "8ac229cdf6e2289d97e82e35774057106fe7f4a2", "parents": [ "74bcbf40546bb7500f2a7ba4ff3cc056a6bd004a" ], "author": { "name": "David Rientjes", "email": "rientjes@google.com", "time": "Mon Aug 09 17:19:46 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 09 20:45:02 2010 -0700" }, "message": "oom: badness heuristic rewrite\n\nThis a complete rewrite of the oom killer\u0027s badness() heuristic which is\nused to determine which task to kill in oom conditions. The goal is to\nmake it as simple and predictable as possible so the results are better\nunderstood and we end up killing the task which will lead to the most\nmemory freeing while still respecting the fine-tuning from userspace.\n\nInstead of basing the heuristic on mm-\u003etotal_vm for each task, the task\u0027s\nrss and swap space is used instead. This is a better indication of the\namount of memory that will be freeable if the oom killed task is chosen\nand subsequently exits. This helps specifically in cases where KDE or\nGNOME is chosen for oom kill on desktop systems instead of a memory\nhogging task.\n\nThe baseline for the heuristic is a proportion of memory that each task is\ncurrently using in memory plus swap compared to the amount of \"allowable\"\nmemory. \"Allowable,\" in this sense, means the system-wide resources for\nunconstrained oom conditions, the set of mempolicy nodes, the mems\nattached to current\u0027s cpuset, or a memory controller\u0027s limit. The\nproportion is given on a scale of 0 (never kill) to 1000 (always kill),\nroughly meaning that if a task has a badness() score of 500 that the task\nconsumes approximately 50% of allowable memory resident in RAM or in swap\nspace.\n\nThe proportion is always relative to the amount of \"allowable\" memory and\nnot the total amount of RAM systemwide so that mempolicies and cpusets may\noperate in isolation; they shall not need to know the true size of the\nmachine on which they are running if they are bound to a specific set of\nnodes or mems, respectively.\n\nRoot tasks are given 3% extra memory just like __vm_enough_memory()\nprovides in LSMs. In the event of two tasks consuming similar amounts of\nmemory, it is generally better to save root\u0027s task.\n\nBecause of the change in the badness() heuristic\u0027s baseline, it is also\nnecessary to introduce a new user interface to tune it. It\u0027s not possible\nto redefine the meaning of /proc/pid/oom_adj with a new scale since the\nABI cannot be changed for backward compatability. Instead, a new tunable,\n/proc/pid/oom_score_adj, is added that ranges from -1000 to +1000. It may\nbe used to polarize the heuristic such that certain tasks are never\nconsidered for oom kill while others may always be considered. The value\nis added directly into the badness() score so a value of -500, for\nexample, means to discount 50% of its memory consumption in comparison to\nother tasks either on the system, bound to the mempolicy, in the cpuset,\nor sharing the same memory controller.\n\n/proc/pid/oom_adj is changed so that its meaning is rescaled into the\nunits used by /proc/pid/oom_score_adj, and vice versa. Changing one of\nthese per-task tunables will rescale the value of the other to an\nequivalent meaning. Although /proc/pid/oom_adj was originally defined as\na bitshift on the badness score, it now shares the same linear growth as\n/proc/pid/oom_score_adj but with different granularity. This is required\nso the ABI is not broken with userspace applications and allows oom_adj to\nbe deprecated for future removal.\n\nSigned-off-by: David Rientjes \u003crientjes@google.com\u003e\nCc: Nick Piggin \u003cnpiggin@suse.de\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Oleg Nesterov \u003coleg@redhat.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "25edde0332916ae706ccf83de688be57bcc844b7", "tree": "35a5b0e651f9cdb48d9a55a748970339c4f681bc", "parents": [ "b898cc70019ce1835bbf6c47bdf978adc36faa42" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Mon Aug 09 17:19:27 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Aug 09 20:45:00 2010 -0700" }, "message": "vmscan: kill prev_priority completely\n\nSince 2.6.28 zone-\u003eprev_priority is unused. Then it can be removed\nsafely. It reduce stack usage slightly.\n\nNow I have to say that I\u0027m sorry. 2 years ago, I thought prev_priority\ncan be integrate again, it\u0027s useful. but four (or more) times trying\nhaven\u0027t got good performance number. Thus I give up such approach.\n\nThe rest of this changelog is notes on prev_priority and why it existed in\nthe first place and why it might be not necessary any more. This information\nis based heavily on discussions between Andrew Morton, Rik van Riel and\nKosaki Motohiro who is heavily quotes from.\n\nHistorically prev_priority was important because it determined when the VM\nwould start unmapping PTE pages. i.e. there are no balances of note within\nthe VM, Anon vs File and Mapped vs Unmapped. Without prev_priority, there\nis a potential risk of unnecessarily increasing minor faults as a large\namount of read activity of use-once pages could push mapped pages to the\nend of the LRU and get unmapped.\n\nThere is no proof this is still a problem but currently it is not considered\nto be. Active files are not deactivated if the active file list is smaller\nthan the inactive list reducing the liklihood that file-mapped pages are\nbeing pushed off the LRU and referenced executable pages are kept on the\nactive list to avoid them getting pushed out by read activity.\n\nEven if it is a problem, prev_priority prev_priority wouldn\u0027t works\nnowadays. First of all, current vmscan still a lot of UP centric code. it\nexpose some weakness on some dozens CPUs machine. I think we need more and\nmore improvement.\n\nThe problem is, current vmscan mix up per-system-pressure, per-zone-pressure\nand per-task-pressure a bit. example, prev_priority try to boost priority to\nother concurrent priority. but if the another task have mempolicy restriction,\nit is unnecessary, but also makes wrong big latency and exceeding reclaim.\nper-task based priority + prev_priority adjustment make the emulation of\nper-system pressure. but it have two issue 1) too rough and brutal emulation\n2) we need per-zone pressure, not per-system.\n\nAnother example, currently DEF_PRIORITY is 12. it mean the lru rotate about\n2 cycle (1/4096 + 1/2048 + 1/1024 + .. + 1) before invoking OOM-Killer.\nbut if 10,0000 thrreads enter DEF_PRIORITY reclaim at the same time, the\nsystem have higher memory pressure than priority\u003d\u003d0 (1/4096*10,000 \u003e 2).\nprev_priority can\u0027t solve such multithreads workload issue. In other word,\nprev_priority concept assume the sysmtem don\u0027t have lots threads.\"\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nSigned-off-by: Mel Gorman \u003cmel@csn.ul.ie\u003e\nReviewed-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nReviewed-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: Dave Chinner \u003cdavid@fromorbit.com\u003e\nCc: Chris Mason \u003cchris.mason@oracle.com\u003e\nCc: Nick Piggin \u003cnpiggin@suse.de\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Johannes Weiner \u003channes@cmpxchg.org\u003e\nCc: Christoph Hellwig \u003chch@infradead.org\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Andrea Arcangeli \u003caarcange@redhat.com\u003e\nCc: Michael Rubin \u003cmrubin@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "4d845ebf4cf9e985b1704b1f08b37f744b4ede13", "tree": "d137f9220a1572c2e3b53ce38950314f5aeb82e4", "parents": [ "3c26c9d9597f982973b9b3a32364230096ab0d78" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Tue Jun 29 15:05:18 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 29 15:29:30 2010 -0700" }, "message": "memcg: fix wake up in oom wait queue\n\nOOM-waitqueue should be waken up when oom_disable is canceled. This is a\nfix for 3c11ecf448eff8f1 (\"memcg: oom kill disable and oom status\").\n\nHow to test:\n Create a cgroup A...\n 1. set memory.limit and memory.memsw.limit to be small value\n 2. echo 1 \u003e /cgroup/A/memory.oom_control, this disables oom-kill.\n 3. run a program which must cause OOM.\n\nA program executed in 3 will sleep by oom_waiqueue in memcg. Then, how to\nwake it up is problem.\n\n 1. echo 0 \u003e /cgroup/A/memory.oom_control (enable OOM-killer)\n 2. echo big mem \u003e /cgroup/A/memory.memsw.limit_in_bytes(allow more swap)\n\netc..\n\nWithout the patch, a task in slept can not be waken up.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "2c488db27b614816024e7994117f599337de0f34", "tree": "1ccfe14f5b2bede321004427babdad111a3b0c8e", "parents": [ "907860ed381a31b0102f362df67c1c5cae6ef050" ], "author": { "name": "Kirill A. Shutemov", "email": "kirill@shutemov.name", "time": "Wed May 26 14:42:47 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 27 09:12:44 2010 -0700" }, "message": "memcg: clean up memory thresholds\n\nIntroduce struct mem_cgroup_thresholds. It helps to reduce number of\nchecks of thresholds type (memory or mem+swap).\n\n[akpm@linux-foundation.org: repair comment]\nSigned-off-by: Kirill A. Shutemov \u003ckirill@shutemov.name\u003e\nCc: Phil Carmody \u003cext-phil.2.carmody@nokia.com\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "907860ed381a31b0102f362df67c1c5cae6ef050", "tree": "4f47a6fe898b1f45da505fc0c27d98e66d42aa46", "parents": [ "ac39cf8cb86c45eeac6a592ce0d58f9021a97235" ], "author": { "name": "Kirill A. Shutemov", "email": "kirill@shutemov.name", "time": "Wed May 26 14:42:46 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 27 09:12:44 2010 -0700" }, "message": "cgroups: make cftype.unregister_event() void-returning\n\nSince we are unable to handle an error returned by\ncftype.unregister_event() properly, let\u0027s make the callback\nvoid-returning.\n\nmem_cgroup_unregister_event() has been rewritten to be a \"never fail\"\nfunction. On mem_cgroup_usage_register_event() we save old buffer for\nthresholds array and reuse it in mem_cgroup_usage_unregister_event() to\navoid allocation.\n\nSigned-off-by: Kirill A. Shutemov \u003ckirill@shutemov.name\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Phil Carmody \u003cext-phil.2.carmody@nokia.com\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "ac39cf8cb86c45eeac6a592ce0d58f9021a97235", "tree": "7321cafb0a1f8f2727c86f9d29159751df856c59", "parents": [ "315c1998e10527ff364a9883048455e609bc7232" ], "author": { "name": "akpm@linux-foundation.org", "email": "akpm@linux-foundation.org", "time": "Wed May 26 14:42:46 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 27 09:12:44 2010 -0700" }, "message": "memcg: fix mis-accounting of file mapped racy with migration\n\nFILE_MAPPED per memcg of migrated file cache is not properly updated,\nbecause our hook in page_add_file_rmap() can\u0027t know to which memcg\nFILE_MAPPED should be counted.\n\nBasically, this patch is for fixing the bug but includes some big changes\nto fix up other messes.\n\nNow, at migrating mapped file, events happen in following sequence.\n\n 1. allocate a new page.\n 2. get memcg of an old page.\n 3. charge ageinst a new page before migration. But at this point,\n no changes to new page\u0027s page_cgroup, no commit for the charge.\n (IOW, PCG_USED bit is not set.)\n 4. page migration replaces radix-tree, old-page and new-page.\n 5. page migration remaps the new page if the old page was mapped.\n 6. Here, the new page is unlocked.\n 7. memcg commits the charge for newpage, Mark the new page\u0027s page_cgroup\n as PCG_USED.\n\nBecause \"commit\" happens after page-remap, we can count FILE_MAPPED\nat \"5\", because we should avoid to trust page_cgroup-\u003emem_cgroup.\nif PCG_USED bit is unset.\n(Note: memcg\u0027s LRU removal code does that but LRU-isolation logic is used\n for helping it. When we overwrite page_cgroup-\u003emem_cgroup, page_cgroup is\n not on LRU or page_cgroup-\u003emem_cgroup is NULL.)\n\nWe can lose file_mapped accounting information at 5 because FILE_MAPPED\nis updated only when mapcount changes 0-\u003e1. So we should catch it.\n\nBTW, historically, above implemntation comes from migration-failure\nof anonymous page. Because we charge both of old page and new page\nwith mapcount\u003d0, we can\u0027t catch\n - the page is really freed before remap.\n - migration fails but it\u0027s freed before remap\nor .....corner cases.\n\nNew migration sequence with memcg is:\n\n 1. allocate a new page.\n 2. mark PageCgroupMigration to the old page.\n 3. charge against a new page onto the old page\u0027s memcg. (here, new page\u0027s pc\n is marked as PageCgroupUsed.)\n 4. page migration replaces radix-tree, page table, etc...\n 5. At remapping, new page\u0027s page_cgroup is now makrked as \"USED\"\n We can catch 0-\u003e1 event and FILE_MAPPED will be properly updated.\n\n And we can catch SWAPOUT event after unlock this and freeing this\n page by unmap() can be caught.\n\n 7. Clear PageCgroupMigration of the old page.\n\nSo, FILE_MAPPED will be correctly updated.\n\nThen, for what MIGRATION flag is ?\n Without it, at migration failure, we may have to charge old page again\n because it may be fully unmapped. \"charge\" means that we have to dive into\n memory reclaim or something complated. So, it\u0027s better to avoid\n charge it again. Before this patch, __commit_charge() was working for\n both of the old/new page and fixed up all. But this technique has some\n racy condtion around FILE_MAPPED and SWAPOUT etc...\n Now, the kernel use MIGRATION flag and don\u0027t uncharge old page until\n the end of migration.\n\nI hope this change will make memcg\u0027s page migration much simpler. This\npage migration has caused several troubles. Worth to add a flag for\nsimplification.\n\nReviewed-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nTested-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nReported-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Christoph Lameter \u003ccl@linux-foundation.org\u003e\nCc: \"Kirill A. Shutemov\" \u003ckirill@shutemov.name\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "315c1998e10527ff364a9883048455e609bc7232", "tree": "081c072b7e7a48df369fd80142e611c3f36d3cfd", "parents": [ "5407a56257b6ade44fd9bcac972c99845b7413cd" ], "author": { "name": "Phil Carmody", "email": "ext-phil.2.carmody@nokia.com", "time": "Wed May 26 14:42:43 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 27 09:12:44 2010 -0700" }, "message": "mm: memcontrol - uninitialised return value\n\nOnly an out of memory error will cause ret to be set.\n\nSigned-off-by: Phil Carmody \u003cext-phil.2.carmody@nokia.com\u003e\nAcked-by: Kirill A. Shutemov \u003ckirill@shutemov.name\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "5407a56257b6ade44fd9bcac972c99845b7413cd", "tree": "833a55856af25fe9be5e8a871d89f43677eac512", "parents": [ "df64f81bb1e01cbef967a96642dacf208acb7e72" ], "author": { "name": "Phil Carmody", "email": "ext-phil.2.carmody@nokia.com", "time": "Wed May 26 14:42:42 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 27 09:12:43 2010 -0700" }, "message": "mm: remove unnecessary use of atomic\n\nThe bottom 4 hunks are atomically changing memory to which there are no\naliases as it\u0027s freshly allocated, so there\u0027s no need to use atomic\noperations.\n\nThe other hunks are just atomic_read and atomic_set, and do not involve\nany read-modify-write. The use of atomic_{read,set} doesn\u0027t prevent a\nread/write or write/write race, so if a race were possible (I\u0027m not saying\none is), then it would still be there even with atomic_set.\n\nSee:\nhttp://digitalvampire.org/blog/index.php/2007/05/13/atomic-cargo-cults/\n\nSigned-off-by: Phil Carmody \u003cext-phil.2.carmody@nokia.com\u003e\nAcked-by: Kirill A. Shutemov \u003ckirill@shutemov.name\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "87946a72283be3de936adc754b7007df7d3e6aeb", "tree": "0593c87ba36bae13d6a6d5dda65ebb41354954f8", "parents": [ "90254a65833b67502d14736410b3857a15535c67" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Wed May 26 14:42:39 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 27 09:12:43 2010 -0700" }, "message": "memcg: move charge of file pages\n\nThis patch adds support for moving charge of file pages, which include\nnormal file, tmpfs file and swaps of tmpfs file. It\u0027s enabled by setting\nbit 1 of \u003ctarget cgroup\u003e/memory.move_charge_at_immigrate.\n\nUnlike the case of anonymous pages, file pages(and swaps) in the range\nmmapped by the task will be moved even if the task hasn\u0027t done page fault,\ni.e. they might not be the task\u0027s \"RSS\", but other task\u0027s \"RSS\" that maps\nthe same file. And mapcount of the page is ignored(the page can be moved\neven if page_mapcount(page) \u003e 1). So, conditions that the page/swap\nshould be met to be moved is that it must be in the range mmapped by the\ntarget task and it must be charged to the old cgroup.\n\n[akpm@linux-foundation.org: coding-style fixes]\n[akpm@linux-foundation.org: fix warning]\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "90254a65833b67502d14736410b3857a15535c67", "tree": "e03ec501b4a585c0d112469f36100ec41a715345", "parents": [ "3c11ecf448eff8f12922c498b8274ce98587eb74" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Wed May 26 14:42:38 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 27 09:12:43 2010 -0700" }, "message": "memcg: clean up move charge\n\nThis patch cleans up move charge code by:\n\n- define functions to handle pte for each types, and make\n is_target_pte_for_mc() cleaner.\n\n- instead of checking the MOVE_CHARGE_TYPE_ANON bit, define a function\n that checks the bit.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "3c11ecf448eff8f12922c498b8274ce98587eb74", "tree": "0a28aeef7759ead93ff58fd7db2fe61797523f7e", "parents": [ "9490ff275606da012d5b373342a49610ad61cb81" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed May 26 14:42:37 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 27 09:12:43 2010 -0700" }, "message": "memcg: oom kill disable and oom status\n\nThis adds a feature to disable oom-killer for memcg, if disabled, of\ncourse, tasks under memcg will stop.\n\nBut now, we have oom-notifier for memcg. And the world around memcg is\nnot under out-of-memory. memcg\u0027s out-of-memory just shows memcg hits\nlimit. Then, administrator or management daemon can recover the situation\nby\n\n\t- kill some process\n\t- enlarge limit, add more swap.\n\t- migrate some tasks\n\t- remove file cache on tmps (difficult ?)\n\nUnlike oom-killer, you can take enough information before killing tasks.\n(by gcore, or, ps etc.)\n\n[akpm@linux-foundation.org: coding-style fixes]\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: David Rientjes \u003crientjes@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "9490ff275606da012d5b373342a49610ad61cb81", "tree": "037993e807654da633776066129b001f8bae7d1d", "parents": [ "dc98df5a1b7be402a0e1c71f1b89ccf249ac15ee" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed May 26 14:42:36 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 27 09:12:43 2010 -0700" }, "message": "memcg: oom notifier\n\nConsidering containers or other resource management softwares in userland,\nevent notification of OOM in memcg should be implemented. Now, memcg has\n\"threshold\" notifier which uses eventfd, we can make use of it for oom\nnotification.\n\nThis patch adds oom notification eventfd callback for memcg. The usage is\nvery similar to threshold notifier, but control file is memory.oom_control\nand no arguments other than eventfd is required.\n\n\t% cgroup_event_notifier /cgroup/A/memory.oom_control dummy\n\t(About cgroup_event_notifier, see Documentation/cgroup/)\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: David Rientjes \u003crientjes@google.com\u003e\nCc: Davide Libenzi \u003cdavidel@xmailserver.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "dc98df5a1b7be402a0e1c71f1b89ccf249ac15ee", "tree": "d54f48e26563eadd11114c787b5c27b4e44ebb19", "parents": [ "595f4b694c2424f2af6e1de74c955bbe9cd7b73d" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed May 26 14:42:36 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 27 09:12:43 2010 -0700" }, "message": "memcg: oom wakeup filter\n\nmemcg\u0027s oom waitqueue is a system-wide wait_queue (for handling\nhierarchy.) So, it\u0027s better to add custom wake function and do filtering\nin wake up path.\n\nThis patch adds a filtering feature for waking up oom-waiters. Hierarchy\nis properly handled.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nReviewed-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: David Rientjes \u003crientjes@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "f39d01be4c59a61a08d0cb53f615e7016b85d339", "tree": "6777590e3ff2ddf4df1d38444ba7d692cd463b7b", "parents": [ "54291263519ac2c9bdda68b23b02fef3808deed4", "7db82437cfcac4bdfe79a6323eb554fdfa271623" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 20 09:20:59 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu May 20 09:20:59 2010 -0700" }, "message": "Merge branch \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial\n\n* \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial: (44 commits)\n vlynq: make whole Kconfig-menu dependant on architecture\n add descriptive comment for TIF_MEMDIE task flag declaration.\n EEPROM: max6875: Header file cleanup\n EEPROM: 93cx6: Header file cleanup\n EEPROM: Header file cleanup\n agp: use NULL instead of 0 when pointer is needed\n rtc-v3020: make bitfield unsigned\n PCI: make bitfield unsigned\n jbd2: use NULL instead of 0 when pointer is needed\n cciss: fix shadows sparse warning\n doc: inode uses a mutex instead of a semaphore.\n uml: i386: Avoid redefinition of NR_syscalls\n fix \"seperate\" typos in comments\n cocbalt_lcdfb: correct sections\n doc: Change urls for sparse\n Powerpc: wii: Fix typo in comment\n i2o: cleanup some exit paths\n Documentation/: it\u0027s -\u003e its where appropriate\n UML: Fix compiler warning due to missing task_struct declaration\n UML: add kernel.h include to signal.c\n ...\n" }, { "commit": "747388d78a0ae768fd82b55c4ed38aa646a72364", "tree": "03efabf9761209cffef4e8da9b6989887f76428e", "parents": [ "7f0f15464185a92f9d8791ad231bcd7bf6df54e4" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Tue May 11 14:06:59 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue May 11 17:33:42 2010 -0700" }, "message": "memcg: fix css_is_ancestor() RCU locking\n\nSome callers (in memcontrol.c) calls css_is_ancestor() without\nrcu_read_lock. Because css_is_ancestor() has to access RCU protected\ndata, it should be under rcu_read_lock().\n\nThis makes css_is_ancestor() itself does safe access to RCU protected\narea. (At least, \"root\" can have refcnt\u003d\u003d0 if it\u0027s not an ancestor of\n\"child\". So, we need rcu_read_lock().)\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: \"Paul E. McKenney\" \u003cpaulmck@linux.vnet.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "7f0f15464185a92f9d8791ad231bcd7bf6df54e4", "tree": "a1d4f4c39632659497963b0ccf62828237478d72", "parents": [ "11cad320a4f4bc53d3585c85600c782faa12b99e" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Tue May 11 14:06:58 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue May 11 17:33:42 2010 -0700" }, "message": "memcg: fix css_id() RCU locking for real\n\nCommit ad4ba375373937817404fd92239ef4cadbded23b (\"memcg: css_id() must be\ncalled under rcu_read_lock()\") modifies memcontol.c for fixing RCU check\nmessage. But Andrew Morton pointed out that the fix doesn\u0027t seems sane\nand it was just for hidining lockdep messages.\n\nThis is a patch for do proper things. Checking again, all places,\naccessing without rcu_read_lock, that commit fixies was intentional....\nall callers of css_id() has reference count on it. So, it\u0027s not necessary\nto be under rcu_read_lock().\n\nConsidering again, we can use rcu_dereference_check for css_id(). We know\ncss-\u003eid is valid if css-\u003erefcnt \u003e 0. (css-\u003eid never changes and freed\nafter css-\u003erefcnt going to be 0.)\n\nThis patch makes use of rcu_dereference_check() in css_id/depth and remove\nunnecessary rcu-read-lock added by the commit.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: \"Paul E. McKenney\" \u003cpaulmck@linux.vnet.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "91bc482ec5a615e8ecebc106aaf7d0c267d511de", "tree": "ad75789a5825fb2da75b2dbdc02d1969bad9b338", "parents": [ "e8e8fade7b351f131433fa44ca5d4e960e1c0f9c", "ee84b8243b07c33a5c8aed42b4b2da60cb16d1d2" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 07 13:58:21 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 07 13:58:21 2010 -0700" }, "message": "Merge branch \u0027core-fixes-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip\n\n* \u0027core-fixes-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:\n rcu: create rcu_my_thread_group_empty() wrapper\n memcg: css_id() must be called under rcu_read_lock()\n cgroup: Check task_lock in task_subsys_state()\n sched: Fix an RCU warning in print_task()\n cgroup: Fix an RCU warning in alloc_css_id()\n cgroup: Fix an RCU warning in cgroup_path()\n KEYS: Fix an RCU warning in the reading of user keys\n KEYS: Fix an RCU warning\n" }, { "commit": "ad4ba375373937817404fd92239ef4cadbded23b", "tree": "1c59ebd7a1e3ace3539396dd35ecff0f06132fe8", "parents": [ "1ce7e4ff24fe338438bc7837e02780f202bf202b" ], "author": { "name": "Paul E. McKenney", "email": "paulmck@linux.vnet.ibm.com", "time": "Fri Apr 23 12:26:38 2010 -0700" }, "committer": { "name": "Paul E. McKenney", "email": "paulmck@linux.vnet.ibm.com", "time": "Tue May 04 09:25:03 2010 -0700" }, "message": "memcg: css_id() must be called under rcu_read_lock()\n\nThis patch fixes task_in_mem_cgroup(), mem_cgroup_uncharge_swapcache(),\nmem_cgroup_move_swap_account(), and is_target_pte_for_mc() to protect\ncalls to css_id(). An additional RCU lockdep splat was reported for\nmemcg_oom_wake_function(), however, this function is not yet in\nmainline as of 2.6.34-rc5.\n\nReported-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nTested-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nSigned-off-by: Paul E. McKenney \u003cpaulmck@linux.vnet.ibm.com\u003e\nCc: Andrew Morton \u003cakpm@linux-foundation.org\u003e\n" }, { "commit": "93d5c9be1ddd57d4063ce463c9ac2be1e5ee14f1", "tree": "65455c40ec8c701789d8f16b825e5c290633c910", "parents": [ "cac36f707119b792b2396aed371d6b5cdc194890" ], "author": { "name": "Andrea Arcangeli", "email": "aarcange@redhat.com", "time": "Fri Apr 23 13:17:39 2010 -0400" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Apr 24 11:31:24 2010 -0700" }, "message": "memcg: fix prepare migration\n\nIf a signal is pending (task being killed by sigkill)\n__mem_cgroup_try_charge will write NULL into \u0026mem, and css_put will oops\non null pointer dereference.\n\n BUG: unable to handle kernel NULL pointer dereference at 0000000000000010\n IP: [\u003cffffffff810fc6cc\u003e] mem_cgroup_prepare_migration+0x7c/0xc0\n PGD a5d89067 PUD a5d8a067 PMD 0\n Oops: 0000 [#1] SMP\n last sysfs file: /sys/devices/platform/microcode/firmware/microcode/loading\n CPU 0\n Modules linked in: nfs lockd nfs_acl auth_rpcgss sunrpc acpi_cpufreq pcspkr sg [last unloaded: microcode]\n\n Pid: 5299, comm: largepages Tainted: G W 2.6.34-rc3 #3 Penryn1600SLI-110dB/To Be Filled By O.E.M.\n RIP: 0010:[\u003cffffffff810fc6cc\u003e] [\u003cffffffff810fc6cc\u003e] mem_cgroup_prepare_migration+0x7c/0xc0\n\n[nishimura@mxp.nes.nec.co.jp: fix merge issues]\nSigned-off-by: Andrea Arcangeli \u003caarcange@redhat.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: \u003cstable@kernel.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "6c9468e9eb1252eaefd94ce7f06e1be9b0b641b1", "tree": "797676a336b050bfa1ef879377c07e541b9075d6", "parents": [ "4cb3ca7cd7e2cae8d1daf5345ec99a1e8502cf3f", "c81eddb0e3728661d1585fbc564449c94165cc36" ], "author": { "name": "Jiri Kosina", "email": "jkosina@suse.cz", "time": "Fri Apr 23 02:08:44 2010 +0200" }, "committer": { "name": "Jiri Kosina", "email": "jkosina@suse.cz", "time": "Fri Apr 23 02:08:44 2010 +0200" }, "message": "Merge branch \u0027master\u0027 into for-next\n" }, { "commit": "8725d5416213a145ccc9c236dbd26830ba409e00", "tree": "aff10fc98874adfcb7755ec8bc1f2627036e8dc2", "parents": [ "116354d177ba2da37e91cf884e3d11e67f825efd" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Tue Apr 06 14:35:05 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Apr 07 08:38:05 2010 -0700" }, "message": "memcg: fix race in file_mapped accounting\n\nPresently, memcg\u0027s FILE_MAPPED accounting has following race with\nmove_account (happens at rmdir()).\n\n increment page-\u003emapcount (rmap.c)\n mem_cgroup_update_file_mapped() move_account()\n\t\t\t\t\t lock_page_cgroup()\n\t\t\t\t\t check page_mapped() if\n\t\t\t\t\t page_mapped(page)\u003e1 {\n\t\t\t\t\t\tFILE_MAPPED -1 from old memcg\n\t\t\t\t\t\tFILE_MAPPED +1 to old memcg\n\t\t\t\t\t }\n\t\t\t\t\t .....\n\t\t\t\t\t overwrite pc-\u003emem_cgroup\n\t\t\t\t\t unlock_page_cgroup()\n lock_page_cgroup()\n FILE_MAPPED + 1 to pc-\u003emem_cgroup\n unlock_page_cgroup()\n\nThen,\n\told memcg (-1 file mapped)\n\tnew memcg (+2 file mapped)\n\nThis happens because move_account see page_mapped() which is not guarded\nby lock_page_cgroup(). This patch adds FILE_MAPPED flag to page_cgroup\nand move account information based on it. Now, all checks are synchronous\nwith lock_page_cgroup().\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nReviewed-by: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nReviewed-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Andrea Righi \u003carighi@develer.com\u003e\nCc: Andrea Arcangeli \u003caarcange@redhat.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "e7bbcdf3747e3919c31cfa87853c69d178bce548", "tree": "42493b615337a170f3a019fc41dd90c4d4d9b342", "parents": [ "9d34706f42f9b8c15185423d9af98d37ba21d011" ], "author": { "name": "Dan Carpenter", "email": "error27@gmail.com", "time": "Tue Mar 23 13:35:12 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Mar 24 16:31:19 2010 -0700" }, "message": "memcontrol: fix potential null deref\n\nThere was a potential null deref introduced in c62b1a3b31b5 (\"memcg: use\ngeneric percpu instead of private implementation\").\n\nSigned-off-by: Dan Carpenter \u003cerror27@gmail.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "5cfb80a73b5a52fb19d8b0611203e4dd58e8e9a2", "tree": "bb4db72652827235f1e548c6c9db1a04406c26ad", "parents": [ "ae6be51ed01d6c4aaf249a207b4434bc7785853b" ], "author": { "name": "Daisuke Nishimura", "email": "d-nishimura@mtf.biglobe.ne.jp", "time": "Tue Mar 23 13:35:11 2010 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Mar 24 16:31:19 2010 -0700" }, "message": "memcg: disable move charge in no mmu case\n\nIn commit 02491447 (\"memcg: move charges of anonymous swap\"), I tried to\ndisable move charge feature in no mmu case by enclosing all the related\nfunctions with \"#ifdef CONFIG_MMU\", but the commit places these ifdefs in\nwrong place. (it seems that it\u0027s mangled while handling some fixes...)\n\nThis patch fixes it up.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "320cc51d90832231cece478f0db6550ef367f8f3", "tree": "b0aa81b277cc7f3c666f820cfc44e82299ca7ecb", "parents": [ "32e688b8c1afafa389223a4813b97e8c128a1636" ], "author": { "name": "Greg Thelen", "email": "gthelen@google.com", "time": "Mon Mar 15 15:27:28 2010 +0100" }, "committer": { "name": "Jiri Kosina", "email": "jkosina@suse.cz", "time": "Mon Mar 15 15:27:28 2010 +0100" }, "message": "mm: fix typo in refill_stock() comment\n\nChange refill_stock() comment: s/consumt_stock()/consume_stock()/\n\nSigned-off-by: Greg Thelen \u003cgthelen@google.com\u003e\nAcked-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Jiri Kosina \u003cjkosina@suse.cz\u003e\n" }, { "commit": "867578cbccb0893cc14fc29c670f7185809c90d6", "tree": "e4d0cefac265fc64399223bc82ed714a88ebe20c", "parents": [ "0263c12c12ccc90edc9d856fa839f8936183e6d1" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed Mar 10 15:22:39 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Mar 12 15:52:38 2010 -0800" }, "message": "memcg: fix oom kill behavior\n\nIn current page-fault code,\n\n\thandle_mm_fault()\n\t\t-\u003e ...\n\t\t-\u003e mem_cgroup_charge()\n\t\t-\u003e map page or handle error.\n\t-\u003e check return code.\n\nIf page fault\u0027s return code is VM_FAULT_OOM, page_fault_out_of_memory() is\ncalled. But if it\u0027s caused by memcg, OOM should have been already\ninvoked.\n\nThen, I added a patch: a636b327f731143ccc544b966cfd8de6cb6d72c6. That\npatch records last_oom_jiffies for memcg\u0027s sub-hierarchy and prevents\npage_fault_out_of_memory from being invoked in near future.\n\nBut Nishimura-san reported that check by jiffies is not enough when the\nsystem is terribly heavy.\n\nThis patch changes memcg\u0027s oom logic as.\n * If memcg causes OOM-kill, continue to retry.\n * remove jiffies check which is used now.\n * add memcg-oom-lock which works like perzone oom lock.\n * If current is killed(as a process), bypass charge.\n\nSomething more sophisticated can be added but this pactch does\nfundamental things.\nTODO:\n - add oom notifier\n - add permemcg disable-oom-kill flag and freezer at oom.\n - more chances for wake up oom waiter (when changing memory limit etc..)\n\nReviewed-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nTested-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: David Rientjes \u003crientjes@google.com\u003e\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "a0a4db548edcce067c1201ef25cf2bc29f32dca4", "tree": "6f8139a582179666cd248d978332ed0e32ad9f0f", "parents": [ "4ab78683c17d739c2a2077141dcf81a02b7fb57e" ], "author": { "name": "Kirill A. Shutemov", "email": "kirill@shutemov.name", "time": "Wed Mar 10 15:22:34 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Mar 12 15:52:37 2010 -0800" }, "message": "cgroups: remove events before destroying subsystem state objects\n\nEvents should be removed after rmdir of cgroup directory, but before\ndestroying subsystem state objects. Let\u0027s take reference to cgroup\ndirectory dentry to do that.\n\nSigned-off-by: Kirill A. Shutemov \u003ckirill@shutemov.name\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hioryu@jp.fujitsu.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nAcked-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nCc: Dan Malek \u003cdan@embeddedalley.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "d2265e6fa3f220ea5fd37522d13390e9675adcf7", "tree": "30e0264bcd379e21507b22243891880b7418f68f", "parents": [ "430e48631e72aeab74d844c57b441f98a2e36eee" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed Mar 10 15:22:31 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Mar 12 15:52:37 2010 -0800" }, "message": "memcg : share event counter rather than duplicate\n\nMemcg has 2 eventcountes which counts \"the same\" event. Just usages are\ndifferent from each other. This patch tries to reduce event counter.\n\nNow logic uses \"only increment, no reset\" counter and masks for each\nchecks. Softlimit chesk was done per 1000 evetns. So, the similar check\ncan be done by !(new_counter \u0026 0x3ff). Threshold check was done per 100\nevents. So, the similar check can be done by (!new_counter \u0026 0x7f)\n\nALL event checks are done right after EVENT percpu counter is updated.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Kirill A. Shutemov \u003ckirill@shutemov.name\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "430e48631e72aeab74d844c57b441f98a2e36eee", "tree": "e1b3eadc1b5c1871db0bf1247af8684b53cae12a", "parents": [ "c62b1a3b31b5e27a6c5c2e91cc5ce05fdb6344d0" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed Mar 10 15:22:30 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Mar 12 15:52:37 2010 -0800" }, "message": "memcg: update threshold and softlimit at commit\n\nPresently, move_task does \"batched\" precharge. Because res_counter or\ncss\u0027s refcnt are not-scalable jobs for memcg, try_charge_().. tend to be\ndone in batched manner if allowed.\n\nNow, softlimit and threshold check their event counter in try_charge, but\nthe charge is not a per-page event. And event counter is not updated at\ncharge(). Moreover, precharge doesn\u0027t pass \"page\" to try_charge() and\nsoftlimit tree will be never updated until uncharge() causes an event.\"\n\nSo the best place to check the event counter is commit_charge(). This is\nper-page event by its nature. This patch move checks to there.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Kirill A. Shutemov \u003ckirill@shutemov.name\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "c62b1a3b31b5e27a6c5c2e91cc5ce05fdb6344d0", "tree": "6887141ac2afc2a8f3a1a7dfa39a56f6899a9864", "parents": [ "6a6135b64fda39d931a79090f4da37f1c6da4a8c" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed Mar 10 15:22:29 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Mar 12 15:52:37 2010 -0800" }, "message": "memcg: use generic percpu instead of private implementation\n\nWhen per-cpu counter for memcg was implemneted, dynamic percpu allocator\nwas not very good. But now, we have good one and useful macros. This\npatch replaces memcg\u0027s private percpu counter implementation with generic\ndynamic percpu allocator.\n\nThe benefits are\n\t- We can remove private implementation.\n\t- The counters will be NUMA-aware. (Current one is not...)\n\t- This patch makes sizeof struct mem_cgroup smaller. Then,\n\t struct mem_cgroup may be fit in page size on small config.\n - About basic performance aspects, see below.\n\n [Before]\n # size mm/memcontrol.o\n text data bss dec hex filename\n 24373 2528 4132 31033 7939 mm/memcontrol.o\n\n [page-fault-throuput test on 8cpu/SMP in root cgroup]\n # /root/bin/perf stat -a -e page-faults,cache-misses --repeat 5 ./multi-fault-fork 8\n\n Performance counter stats for \u0027./multi-fault-fork 8\u0027 (5 runs):\n\n 45878618 page-faults ( +- 0.110% )\n 602635826 cache-misses ( +- 0.105% )\n\n 61.005373262 seconds time elapsed ( +- 0.004% )\n\n Then cache-miss/page fault \u003d 13.14\n\n [After]\n #size mm/memcontrol.o\n text data bss dec hex filename\n 23913 2528 4132 30573 776d mm/memcontrol.o\n # /root/bin/perf stat -a -e page-faults,cache-misses --repeat 5 ./multi-fault-fork 8\n\n Performance counter stats for \u0027./multi-fault-fork 8\u0027 (5 runs):\n\n 48179400 page-faults ( +- 0.271% )\n 588628407 cache-misses ( +- 0.136% )\n\n 61.004615021 seconds time elapsed ( +- 0.004% )\n\n Then cache-miss/page fault \u003d 12.22\n\n Text size is reduced.\n This performance improvement is not big and will be invisible in real world\n applications. But this result shows this patch has some good effect even\n on (small) SMP.\n\nHere is a test program I used.\n\n 1. fork() processes on each cpus.\n 2. do page fault repeatedly on each process.\n 3. after 60secs, kill all childredn and exit.\n\n(3 is necessary for getting stable data, this is improvement from previous one.)\n\n#define _GNU_SOURCE\n#include \u003cstdio.h\u003e\n#include \u003csched.h\u003e\n#include \u003csys/mman.h\u003e\n#include \u003csys/types.h\u003e\n#include \u003csys/stat.h\u003e\n#include \u003cfcntl.h\u003e\n#include \u003csignal.h\u003e\n#include \u003cstdlib.h\u003e\n\n/*\n * For avoiding contention in page table lock, FAULT area is\n * sparse. If FAULT_LENGTH is too large for your cpus, decrease it.\n */\n#define FAULT_LENGTH\t(2 * 1024 * 1024)\n#define PAGE_SIZE\t4096\n#define MAXNUM\t\t(128)\n\nvoid alarm_handler(int sig)\n{\n}\n\nvoid *worker(int cpu, int ppid)\n{\n\tvoid *start, *end;\n\tchar *c;\n\tcpu_set_t set;\n\tint i;\n\n\tCPU_ZERO(\u0026set);\n\tCPU_SET(cpu, \u0026set);\n\tsched_setaffinity(0, sizeof(set), \u0026set);\n\n\tstart \u003d mmap(NULL, FAULT_LENGTH, PROT_READ|PROT_WRITE,\n\t\t\tMAP_PRIVATE | MAP_ANONYMOUS, 0, 0);\n\tif (start \u003d\u003d MAP_FAILED) {\n\t\tperror(\"mmap\");\n\t\texit(1);\n\t}\n\tend \u003d start + FAULT_LENGTH;\n\n\tpause();\n\t//fprintf(stderr, \"run%d\", cpu);\n\twhile (1) {\n\t\tfor (c \u003d (char*)start; (void *)c \u003c end; c +\u003d PAGE_SIZE)\n\t\t\t*c \u003d 0;\n\t\tmadvise(start, FAULT_LENGTH, MADV_DONTNEED);\n\t}\n\treturn NULL;\n}\n\nint main(int argc, char *argv[])\n{\n\tint num, i, ret, pid, status;\n\tint pids[MAXNUM];\n\n\tif (argc \u003c 2)\n\t\treturn 0;\n\n\tsetpgid(0, 0);\n\tsignal(SIGALRM, alarm_handler);\n\tnum \u003d atoi(argv[1]);\n\tpid \u003d getpid();\n\n\tfor (i \u003d 0; i \u003c num; ++i) {\n\t\tret \u003d fork();\n\t\tif (!ret) {\n\t\t\tworker(i, pid);\n\t\t\texit(0);\n\t\t}\n\t\tpids[i] \u003d ret;\n\t}\n\tsleep(1);\n\tkill(-pid, SIGALRM);\n\tsleep(60);\n\tfor (i \u003d 0; i \u003c num; i++)\n\t\tkill(pids[i], SIGKILL);\n\tfor (i \u003d 0; i \u003c num; i++)\n\t\twaitpid(pids[i], \u0026status, 0);\n\treturn 0;\n}\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "6a6135b64fda39d931a79090f4da37f1c6da4a8c", "tree": "b5ade9fdb3a46a86e0cae9ed4e645ea1e153a7ff", "parents": [ "2e72b6347c9459e6cff5634ddc815485bae6985f" ], "author": { "name": "Kirill A. Shutemov", "email": "kirill@shutemov.name", "time": "Wed Mar 10 15:22:25 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Mar 12 15:52:37 2010 -0800" }, "message": "memcg: typo in comment to mem_cgroup_print_oom_info()\n\ns/mem_cgroup_print_mem_info/mem_cgroup_print_oom_info/\n\nSigned-off-by: Kirill A. Shutemov \u003ckirill@shutemov.name\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "2e72b6347c9459e6cff5634ddc815485bae6985f", "tree": "48ceef47f2cd0d33f2a1dd816c1a05ca2cb3e0e2", "parents": [ "378ce724bc2a0ef1243e11c09d58a70bb6be007a" ], "author": { "name": "Kirill A. Shutemov", "email": "kirill@shutemov.name", "time": "Wed Mar 10 15:22:24 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Mar 12 15:52:37 2010 -0800" }, "message": "memcg: implement memory thresholds\n\nIt allows to register multiple memory and memsw thresholds and gets\nnotifications when it crosses.\n\nTo register a threshold application need:\n- create an eventfd;\n- open memory.usage_in_bytes or memory.memsw.usage_in_bytes;\n- write string like \"\u003cevent_fd\u003e \u003cmemory.usage_in_bytes\u003e \u003cthreshold\u003e\" to\n cgroup.event_control.\n\nApplication will be notified through eventfd when memory usage crosses\nthreshold in any direction.\n\nIt\u0027s applicable for root and non-root cgroup.\n\nIt uses stats to track memory usage, simmilar to soft limits. It checks\nif we need to send event to userspace on every 100 page in/out. I guess\nit\u0027s good compromise between performance and accuracy of thresholds.\n\n[akpm@linux-foundation.org: coding-style fixes]\n[nishimura@mxp.nes.nec.co.jp: fix documentation merge issue]\nSigned-off-by: Kirill A. Shutemov \u003ckirill@shutemov.name\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nCc: Dan Malek \u003cdan@embeddedalley.com\u003e\nCc: Vladislav Buzov \u003cvbuzov@embeddedalley.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Alexander Shishkin \u003cvirtuoso@slind.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "378ce724bc2a0ef1243e11c09d58a70bb6be007a", "tree": "6a0a239547179842a55f9ffc9501cc9d2b0f5499", "parents": [ "104f39284e830f425085886ef72c49aee6631575" ], "author": { "name": "Kirill A. Shutemov", "email": "kirill@shutemov.name", "time": "Wed Mar 10 15:22:23 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Mar 12 15:52:37 2010 -0800" }, "message": "memcg: rework usage of stats by soft limit\n\nInstead of incrementing counter on each page in/out and comparing it with\nconstant, we set counter to constant, decrement counter on each page\nin/out and compare it with zero. We want to make comparing as fast as\npossible. On many RISC systems (probably not only RISC) comparing with\nzero is more effective than comparing with a constant, since not every\nconstant can be immediate operand for compare instruction.\n\nAlso, I\u0027ve renamed MEM_CGROUP_STAT_EVENTS to MEM_CGROUP_STAT_SOFTLIMIT,\nsince really it\u0027s not a generic counter.\n\nSigned-off-by: Kirill A. Shutemov \u003ckirill@shutemov.name\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nCc: Dan Malek \u003cdan@embeddedalley.com\u003e\nCc: Vladislav Buzov \u003cvbuzov@embeddedalley.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Alexander Shishkin \u003cvirtuoso@slind.org\u003e\nCc: Davide Libenzi \u003cdavidel@xmailserver.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "104f39284e830f425085886ef72c49aee6631575", "tree": "6b5fff422dea5c5513a3af45f055fd939b32b316", "parents": [ "0dea116876eefc9c7ca9c5d74fe665481e499fa3" ], "author": { "name": "Kirill A. Shutemov", "email": "kirill@shutemov.name", "time": "Wed Mar 10 15:22:21 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Mar 12 15:52:37 2010 -0800" }, "message": "memcg: extract mem_group_usage() from mem_cgroup_read()\n\nHelper to get memory or mem+swap usage of the cgroup.\n\nSigned-off-by: Kirill A. Shutemov \u003ckirill@shutemov.name\u003e\nAcked-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nCc: Dan Malek \u003cdan@embeddedalley.com\u003e\nCc: Vladislav Buzov \u003cvbuzov@embeddedalley.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Alexander Shishkin \u003cvirtuoso@slind.org\u003e\nCc: Davide Libenzi \u003cdavidel@xmailserver.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "483c30b514bd3037fa3f19fa42327c94c10f51c8", "tree": "aaf96db52bf4bdb0c83f209bc9c6a1237867718f", "parents": [ "024914477e15ef8b17f271ec47f1bb8a589f0806" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Wed Mar 10 15:22:18 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Mar 12 15:52:36 2010 -0800" }, "message": "memcg: improve performance in moving swap charge\n\nTry to reduce overheads in moving swap charge by:\n\n- Adds a new function(__mem_cgroup_put), which takes \"count\" as a arg and\n decrement mem-\u003erefcnt by \"count\".\n- Removed res_counter_uncharge, css_put, and mem_cgroup_put from the path\n of moving swap account, and consolidate all of them into mem_cgroup_clear_mc.\n We cannot do that about mc.to-\u003erefcnt.\n\nThese changes reduces the overhead from 1.35sec to 0.9sec to move charges\nof 1G anonymous memory(including 500MB swap) in my test environment.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "024914477e15ef8b17f271ec47f1bb8a589f0806", "tree": "9a6a8b4224c94fcdd1b8c3127b301ee3537f8cc2", "parents": [ "8033b97c9b5ef063e3f4bf2efe1cd0a22093aaff" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Wed Mar 10 15:22:17 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Mar 12 15:52:36 2010 -0800" }, "message": "memcg: move charges of anonymous swap\n\nThis patch is another core part of this move-charge-at-task-migration\nfeature. It enables moving charges of anonymous swaps.\n\nTo move the charge of swap, we need to exchange swap_cgroup\u0027s record.\n\nIn current implementation, swap_cgroup\u0027s record is protected by:\n\n - page lock: if the entry is on swap cache.\n - swap_lock: if the entry is not on swap cache.\n\nThis works well in usual swap-in/out activity.\n\nBut this behavior make the feature of moving swap charge check many\nconditions to exchange swap_cgroup\u0027s record safely.\n\nSo I changed modification of swap_cgroup\u0027s recored(swap_cgroup_record())\nto use xchg, and define a new function to cmpxchg swap_cgroup\u0027s record.\n\nThis patch also enables moving charge of non pte_present but not uncharged\nswap caches, which can be exist on swap-out path, by getting the target\npages via find_get_page() as do_mincore() does.\n\n[kosaki.motohiro@jp.fujitsu.com: fix ia64 build]\n[akpm@linux-foundation.org: fix typos]\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "8033b97c9b5ef063e3f4bf2efe1cd0a22093aaff", "tree": "0d3dbf74d6c307d425d5d13a5c7c1efbf6c079f2", "parents": [ "854ffa8d104e44111fec96764c0e0cb29223d54c" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Wed Mar 10 15:22:16 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Mar 12 15:52:36 2010 -0800" }, "message": "memcg: avoid oom during moving charge\n\nThis move-charge-at-task-migration feature has extra charges on\n\"to\"(pre-charges) and \"from\"(left-over charges) during moving charge.\nThis means unnecessary oom can happen.\n\nThis patch tries to avoid such oom.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "854ffa8d104e44111fec96764c0e0cb29223d54c", "tree": "ca4d19407e35e1888e210fed150d5fb98be36cc8", "parents": [ "4ffef5feff4e4240e767d2f1144b1634a41762e3" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Wed Mar 10 15:22:15 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Mar 12 15:52:36 2010 -0800" }, "message": "memcg: improve performance in moving charge\n\nTry to reduce overheads in moving charge by:\n\n- Instead of calling res_counter_uncharge() against the old cgroup in\n __mem_cgroup_move_account() everytime, call res_counter_uncharge() at the end\n of task migration once.\n- removed css_get(\u0026to-\u003ecss) from __mem_cgroup_move_account() because callers\n should have already called css_get(). And removed css_put(\u0026to-\u003ecss) too,\n which was called by callers of move_account on success of move_account.\n- Instead of calling __mem_cgroup_try_charge(), i.e. res_counter_charge(),\n repeatedly, call res_counter_charge(PAGE_SIZE * count) in can_attach() if\n possible.\n- Instead of calling css_get()/css_put() repeatedly, make use of coalesce\n __css_get()/__css_put() if possible.\n\nThese changes reduces the overhead from 1.7sec to 0.6sec to move charges\nof 1G anonymous memory in my test environment.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "4ffef5feff4e4240e767d2f1144b1634a41762e3", "tree": "14793120e5809008c2587d89162d8d57130d6fc8", "parents": [ "7dc74be032bfcaa2f9d9e4296ff5bbddfa9e2f19" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Wed Mar 10 15:22:14 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Mar 12 15:52:36 2010 -0800" }, "message": "memcg: move charges of anonymous page\n\nThis patch is the core part of this move-charge-at-task-migration feature.\n It implements functions to move charges of anonymous pages mapped only by\nthe target task.\n\nImplementation:\n- define struct move_charge_struct and a valuable of it(mc) to remember the\n count of pre-charges and other information.\n- At can_attach(), get anon_rss of the target mm, call __mem_cgroup_try_charge()\n repeatedly and count up mc.precharge.\n- At attach(), parse the page table, find a target page to be move, and call\n mem_cgroup_move_account() about the page.\n- Cancel all precharges if mc.precharge \u003e 0 on failure or at the end of\n task move.\n\n[akpm@linux-foundation.org: a little simplification]\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "7dc74be032bfcaa2f9d9e4296ff5bbddfa9e2f19", "tree": "8bd10dcca3cfb3f4f6a0e6222c7f2048995a7d5f", "parents": [ "b70cc5fdb445a6929a01e9c406593265b136c99d" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Wed Mar 10 15:22:13 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Mar 12 15:52:36 2010 -0800" }, "message": "memcg: add interface to move charge at task migration\n\nIn current memcg, charges associated with a task aren\u0027t moved to the new\ncgroup at task migration. Some users feel this behavior to be strange.\nThese patches are for this feature, that is, for charging to the new\ncgroup and, of course, uncharging from the old cgroup at task migration.\n\nThis patch adds \"memory.move_charge_at_immigrate\" file, which is a flag\nfile to determine whether charges should be moved to the new cgroup at\ntask migration or not and what type of charges should be moved. This\npatch also adds read and write handlers of the file.\n\nThis patch also adds no-op handlers for this feature. These handlers will\nbe implemented in later patches. And you cannot write any values other\nthan 0 to move_charge_at_immigrate yet.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "648bcc771145172a14bc35eeb849ed08f6aa4f1e", "tree": "3cc7da55d1ca18b46cb0518b3b35c276adc0aad1", "parents": [ "19adf9c5d5793657118f2002237c0ee49c3b6185" ], "author": { "name": "Thiago Farina", "email": "tfransosi@gmail.com", "time": "Fri Mar 05 13:42:04 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Mar 06 11:26:26 2010 -0800" }, "message": "mm/memcontrol.c: fix \"integer as NULL pointer\" sparse warning\n\nmm/memcontrol.c:2548:32: warning: Using plain integer as NULL pointer\n\nSigned-off-by: Thiago Farina \u003ctfransosi@gmail.com\u003e\nAcked-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "fce66477578d081f19aef5ea218664ff7758c33a", "tree": "ef8914767f7834fcd65c2f72eeee59dd2248076b", "parents": [ "d817cd525589765aa5f6798734e422c867685a58" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Fri Jan 15 17:01:30 2010 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jan 16 12:15:39 2010 -0800" }, "message": "memcg: ensure list is empty at rmdir\n\nCurrent mem_cgroup_force_empty() only ensures mem-\u003eres.usage \u003d\u003d 0 on\nsuccess. But this doesn\u0027t guarantee memcg\u0027s LRU is really empty, because\nthere are some cases in which !PageCgrupUsed pages exist on memcg\u0027s LRU.\n\nFor example:\n- Pages can be uncharged by its owner process while they are on LRU.\n- race between mem_cgroup_add_lru_list() and __mem_cgroup_uncharge_common().\n\nSo there can be a case in which the usage is zero but some of the LRUs are not empty.\n\nOTOH, mem_cgroup_del_lru_list(), which can be called asynchronously with\nrmdir, accesses the mem_cgroup, so this access can cause a problem if it\nraces with rmdir because the mem_cgroup might have been freed by rmdir.\n\nActually, I saw a bug which seems to be caused by this race.\n\n\t[1530745.949906] BUG: unable to handle kernel NULL pointer dereference at 0000000000000230\n\t[1530745.950651] IP: [\u003cffffffff810fbc11\u003e] mem_cgroup_del_lru_list+0x30/0x80\n\t[1530745.950651] PGD 3863de067 PUD 3862c7067 PMD 0\n\t[1530745.950651] Oops: 0002 [#1] SMP\n\t[1530745.950651] last sysfs file: /sys/devices/system/cpu/cpu7/cache/index1/shared_cpu_map\n\t[1530745.950651] CPU 3\n\t[1530745.950651] Modules linked in: configs ipt_REJECT xt_tcpudp iptable_filter ip_tables x_tables bridge stp nfsd nfs_acl auth_rpcgss exportfs autofs4 hidp rfcomm l2cap crc16 bluetooth lockd sunrpc ib_iser rdma_cm ib_cm iw_cm ib_sa ib_mad ib_core ib_addr iscsi_tcp bnx2i cnic uio ipv6 cxgb3i cxgb3 mdio libiscsi_tcp libiscsi scsi_transport_iscsi dm_mirror dm_multipath scsi_dh video output sbs sbshc battery ac lp kvm_intel kvm sg ide_cd_mod cdrom serio_raw tpm_tis tpm tpm_bios acpi_memhotplug button parport_pc parport rtc_cmos rtc_core rtc_lib e1000 i2c_i801 i2c_core pcspkr dm_region_hash dm_log dm_mod ata_piix libata shpchp megaraid_mbox sd_mod scsi_mod megaraid_mm ext3 jbd uhci_hcd ohci_hcd ehci_hcd [last unloaded: freq_table]\n\t[1530745.950651] Pid: 19653, comm: shmem_test_02 Tainted: G M 2.6.32-mm1-00701-g2b04386 #3 Express5800/140Rd-4 [N8100-1065]\n\t[1530745.950651] RIP: 0010:[\u003cffffffff810fbc11\u003e] [\u003cffffffff810fbc11\u003e] mem_cgroup_del_lru_list+0x30/0x80\n\t[1530745.950651] RSP: 0018:ffff8803863ddcb8 EFLAGS: 00010002\n\t[1530745.950651] RAX: 00000000000001e0 RBX: ffff8803abc02238 RCX: 00000000000001e0\n\t[1530745.950651] RDX: 0000000000000000 RSI: ffff88038611a000 RDI: ffff8803abc02238\n\t[1530745.950651] RBP: ffff8803863ddcc8 R08: 0000000000000002 R09: ffff8803a04c8643\n\t[1530745.950651] R10: 0000000000000000 R11: ffffffff810c7333 R12: 0000000000000000\n\t[1530745.950651] R13: ffff880000017f00 R14: 0000000000000092 R15: ffff8800179d0310\n\t[1530745.950651] FS: 0000000000000000(0000) GS:ffff880017800000(0000) knlGS:0000000000000000\n\t[1530745.950651] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b\n\t[1530745.950651] CR2: 0000000000000230 CR3: 0000000379d87000 CR4: 00000000000006e0\n\t[1530745.950651] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000\n\t[1530745.950651] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400\n\t[1530745.950651] Process shmem_test_02 (pid: 19653, threadinfo ffff8803863dc000, task ffff88038612a8a0)\n\t[1530745.950651] Stack:\n\t[1530745.950651] ffffea00040c2fe8 0000000000000000 ffff8803863ddd98 ffffffff810c739a\n\t[1530745.950651] \u003c0\u003e 00000000863ddd18 000000000000000c 0000000000000000 0000000000000000\n\t[1530745.950651] \u003c0\u003e 0000000000000002 0000000000000000 ffff8803863ddd68 0000000000000046\n\t[1530745.950651] Call Trace:\n\t[1530745.950651] [\u003cffffffff810c739a\u003e] release_pages+0x142/0x1e7\n\t[1530745.950651] [\u003cffffffff810c778f\u003e] ? pagevec_move_tail+0x6e/0x112\n\t[1530745.950651] [\u003cffffffff810c781e\u003e] pagevec_move_tail+0xfd/0x112\n\t[1530745.950651] [\u003cffffffff810c78a9\u003e] lru_add_drain+0x76/0x94\n\t[1530745.950651] [\u003cffffffff810dba0c\u003e] exit_mmap+0x6e/0x145\n\t[1530745.950651] [\u003cffffffff8103f52d\u003e] mmput+0x5e/0xcf\n\t[1530745.950651] [\u003cffffffff81043ea8\u003e] exit_mm+0x11c/0x129\n\t[1530745.950651] [\u003cffffffff8108fb29\u003e] ? audit_free+0x196/0x1c9\n\t[1530745.950651] [\u003cffffffff81045353\u003e] do_exit+0x1f5/0x6b7\n\t[1530745.950651] [\u003cffffffff8106133f\u003e] ? up_read+0x2b/0x2f\n\t[1530745.950651] [\u003cffffffff8137d187\u003e] ? lockdep_sys_exit_thunk+0x35/0x67\n\t[1530745.950651] [\u003cffffffff81045898\u003e] do_group_exit+0x83/0xb0\n\t[1530745.950651] [\u003cffffffff810458dc\u003e] sys_exit_group+0x17/0x1b\n\t[1530745.950651] [\u003cffffffff81002c1b\u003e] system_call_fastpath+0x16/0x1b\n\t[1530745.950651] Code: 54 53 0f 1f 44 00 00 83 3d cc 29 7c 00 00 41 89 f4 75 63 eb 4e 48 83 7b 08 00 75 04 0f 0b eb fe 48 89 df e8 18 f3 ff ff 44 89 e2 \u003c48\u003e ff 4c d0 50 48 8b 05 2b 2d 7c 00 48 39 43 08 74 39 48 8b 4b\n\t[1530745.950651] RIP [\u003cffffffff810fbc11\u003e] mem_cgroup_del_lru_list+0x30/0x80\n\t[1530745.950651] RSP \u003cffff8803863ddcb8\u003e\n\t[1530745.950651] CR2: 0000000000000230\n\t[1530745.950651] ---[ end trace c3419c1bb8acc34f ]---\n\t[1530745.950651] Fixing recursive fault but reboot is needed!\n\nThe problem here is pages on LRU may contain pointer to stale memcg. To\nmake res-\u003eusage to be 0, all pages on memcg must be uncharged or moved to\nanother(parent) memcg. Moved page_cgroup have already removed from\noriginal LRU, but uncharged page_cgroup contains pointer to memcg withou\nPCG_USED bit. (This asynchronous LRU work is for improving performance.)\nIf PCG_USED bit is not set, page_cgroup will never be added to memcg\u0027s\nLRU. So, about pages not on LRU, they never access stale pointer. Then,\nwhat we have to take care of is page_cgroup _on_ LRU list. This patch\nfixes this problem by making mem_cgroup_force_empty() visit all LRUs\nbefore exiting its loop and guarantee there are no pages on its LRU.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: \u003cstable@kernel.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "d4220f987cf473c65a342ca69e3eb13dea919a49", "tree": "dbb004a9c805d6de3f6e3955398fee1084a29f16", "parents": [ "61cf693159d6a968a7014e24905143f71ed8ddcf", "f2c03debdfb387fa2e35cac6382779072b8b9209" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Dec 16 12:36:49 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Dec 16 12:36:49 2009 -0800" }, "message": "Merge branch \u0027hwpoison\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ak/linux-mce-2.6\n\n* \u0027hwpoison\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ak/linux-mce-2.6: (34 commits)\n HWPOISON: Remove stray phrase in a comment\n HWPOISON: Try to allocate migration page on the same node\n HWPOISON: Don\u0027t do early filtering if filter is disabled\n HWPOISON: Add a madvise() injector for soft page offlining\n HWPOISON: Add soft page offline support\n HWPOISON: Undefine short-hand macros after use to avoid namespace conflict\n HWPOISON: Use new shake_page in memory_failure\n HWPOISON: Use correct name for MADV_HWPOISON in documentation\n HWPOISON: mention HWPoison in Kconfig entry\n HWPOISON: Use get_user_page_fast in hwpoison madvise\n HWPOISON: add an interface to switch off/on all the page filters\n HWPOISON: add memory cgroup filter\n memcg: add accessor to mem_cgroup.css\n memcg: rename and export try_get_mem_cgroup_from_page()\n HWPOISON: add page flags filter\n mm: export stable page flags\n HWPOISON: limit hwpoison injector to known page types\n HWPOISON: add fs/device filters\n HWPOISON: return 0 to indicate success reliably\n HWPOISON: make semantics of IGNORED/DELAYED clear\n ...\n" }, { "commit": "aa20d489ceb024f91aae084ee00c47fc6a12255c", "tree": "1fc8e826114b1020d0e938d3b915e8a2016ded57", "parents": [ "9ab322caa347c4b580bcaf08f2253ea4cbd9e9ad" ], "author": { "name": "Bob Liu", "email": "lliubbo@gmail.com", "time": "Tue Dec 15 16:47:14 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Dec 16 07:20:08 2009 -0800" }, "message": "memcg: code clean, remove unused variable in mem_cgroup_resize_limit()\n\nVariable `progress\u0027 isn\u0027t used in mem_cgroup_resize_limit() any more.\nRemove it.\n\n[akpm@linux-foundation.org: cleanup]\nSigned-off-by: Bob Liu \u003clliubbo@gmail.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nReviewed-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "9ab322caa347c4b580bcaf08f2253ea4cbd9e9ad", "tree": "3d3344e4ab05fc1047465b439b5cc7f000c44ca2", "parents": [ "d31f56dbf8bafaacb0c617f9a6f137498d5c7aed" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Tue Dec 15 16:47:13 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Dec 16 07:20:07 2009 -0800" }, "message": "memcg: remove memcg_tasklist\n\nmemcg_tasklist was introduced at commit 7f4d454d(memcg: avoid deadlock\ncaused by race between oom and cpuset_attach) instead of cgroup_mutex to\nfix a deadlock problem. The cgroup_mutex, which was removed by the\ncommit, in mem_cgroup_out_of_memory() was originally introduced at commit\nc7ba5c9e (Memory controller: OOM handling).\n\nIIUC, the intention of this cgroup_mutex was to prevent task move during\nselect_bad_process() so that situations like below can be avoided.\n\n Assume cgroup \"foo\" has exceeded its limit and is about to trigger oom.\n 1. Process A, which has been in cgroup \"baa\" and uses large memory, is just\n moved to cgroup \"foo\". Process A can be the candidates for being killed.\n 2. Process B, which has been in cgroup \"foo\" and uses large memory, is just\n moved from cgroup \"foo\". Process B can be excluded from the candidates for\n being killed.\n\nBut these race window exists anyway even if we hold a lock, because\n__mem_cgroup_try_charge() decides wether it should trigger oom or not\noutside of the lock. So the original cgroup_mutex in\nmem_cgroup_out_of_memory and thus current memcg_tasklist has no use. And\nIMHO, those races are not so critical for users.\n\nThis patch removes it and make codes simpler.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "d31f56dbf8bafaacb0c617f9a6f137498d5c7aed", "tree": "88d095c2208d27362e58ff7431407040ead9d848", "parents": [ "57f9fd7d25ac9a0d7e3a4ced580e780ab4524e3b" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Tue Dec 15 16:47:12 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Dec 16 07:20:07 2009 -0800" }, "message": "memcg: avoid oom-killing innocent task in case of use_hierarchy\n\ntask_in_mem_cgroup(), which is called by select_bad_process() to check\nwhether a task can be a candidate for being oom-killed from memcg\u0027s limit,\nchecks \"curr-\u003euse_hierarchy\"(\"curr\" is the mem_cgroup the task belongs\nto).\n\nBut this check return true(it\u0027s false positive) when:\n\n\t\u003csome path\u003e/aa\t\tuse_hierarchy \u003d\u003d 0\t\u003c- hitting limit\n\t \u003csome path\u003e/aa/00\tuse_hierarchy \u003d\u003d 1\t\u003c- the task belongs to\n\nThis leads to killing an innocent task in aa/00. This patch is a fix for\nthis bug. And this patch also fixes the arg for\nmem_cgroup_print_oom_info(). We should print information of mem_cgroup\nwhich the task being killed, not current, belongs to.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nAcked-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "57f9fd7d25ac9a0d7e3a4ced580e780ab4524e3b", "tree": "ff95e62e7326ba7e77903f7f767e9650c0d9b3dd", "parents": [ "a3032a2c15c6967f9f0c0c28375b1a5c833a3112" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Tue Dec 15 16:47:11 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Dec 16 07:20:07 2009 -0800" }, "message": "memcg: cleanup mem_cgroup_move_parent()\n\nmem_cgroup_move_parent() calls try_charge first and cancel_charge on\nfailure. IMHO, charge/uncharge(especially charge) is high cost operation,\nso we should avoid it as far as possible.\n\nThis patch tries to delay try_charge in mem_cgroup_move_parent() by\nre-ordering checks it does.\n\nAnd this patch renames mem_cgroup_move_account() to\n__mem_cgroup_move_account(), changes the return value of\n__mem_cgroup_move_account() from int to void, and adds a new\nwrapper(mem_cgroup_move_account()), which checks whether a @pc is valid\nfor moving account and calls __mem_cgroup_move_account().\n\nThis patch removes the last caller of trylock_page_cgroup(), so removes\nits definition too.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "a3032a2c15c6967f9f0c0c28375b1a5c833a3112", "tree": "32d279856032fd9d84ebd5d6746b2bd382baaac9", "parents": [ "d8046582d5ee24448800e71c6933fdb6813aa062" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Tue Dec 15 16:47:10 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Dec 16 07:20:07 2009 -0800" }, "message": "memcg: add mem_cgroup_cancel_charge()\n\nThere are some places calling both res_counter_uncharge() and css_put() to\ncancel the charge and the refcnt we have got by mem_cgroup_tyr_charge().\n\nThis patch introduces mem_cgroup_cancel_charge() and call it in those\nplaces.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nReviewed-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "d8046582d5ee24448800e71c6933fdb6813aa062", "tree": "857d014a850deee5ddb6da6550d81ce0122f03b1", "parents": [ "cdec2e4265dfa09490601b00aeabd8a8d4af30f0" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Tue Dec 15 16:47:09 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Dec 16 07:20:07 2009 -0800" }, "message": "memcg: make memcg\u0027s file mapped consistent with global VM\n\nIn global VM, FILE_MAPPED is used but memcg uses MAPPED_FILE. This makes\ngrep difficult. Replace memcg\u0027s MAPPED_FILE with FILE_MAPPED\n\nAnd in global VM, mapped shared memory is accounted into FILE_MAPPED.\nBut memcg doesn\u0027t. fix it.\nNote:\n page_is_file_cache() just checks SwapBacked or not.\n So, we need to check PageAnon.\n\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nReviewed-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "cdec2e4265dfa09490601b00aeabd8a8d4af30f0", "tree": "d120759b91e3e24d33f3d76010c9da990f115706", "parents": [ "569b846df54ffb2827b83ce3244c5f032394cba4" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Tue Dec 15 16:47:08 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Dec 16 07:20:07 2009 -0800" }, "message": "memcg: coalesce charging via percpu storage\n\nThis is a patch for coalescing access to res_counter at charging by percpu\ncaching. At charge, memcg charges 64pages and remember it in percpu\ncache. Because it\u0027s cache, drain/flush if necessary.\n\nThis version uses public percpu area.\n 2 benefits for using public percpu area.\n 1. Sum of stocked charge in the system is limited to # of cpus\n not to the number of memcg. This shows better synchonization.\n 2. drain code for flush/cpuhotplug is very easy (and quick)\n\nThe most important point of this patch is that we never touch res_counter\nin fast path. The res_counter is system-wide shared counter which is modified\nvery frequently. We shouldn\u0027t touch it as far as we can for avoiding\nfalse sharing.\n\nOn x86-64 8cpu server, I tested overheads of memcg at page fault by\nrunning a program which does map/fault/unmap in a loop. Running\na task per a cpu by taskset and see sum of the number of page faults\nin 60secs.\n\n[without memcg config]\n 40156968 page-faults # 0.085 M/sec ( +- 0.046% )\n 27.67 cache-miss/faults\n\n[root cgroup]\n 36659599 page-faults # 0.077 M/sec ( +- 0.247% )\n 31.58 cache miss/faults\n\n[in a child cgroup]\n 18444157 page-faults # 0.039 M/sec ( +- 0.133% )\n 69.96 cache miss/faults\n\n[ + coalescing uncharge patch]\n 27133719 page-faults # 0.057 M/sec ( +- 0.155% )\n 47.16 cache miss/faults\n\n[ + coalescing uncharge patch + this patch ]\n 34224709 page-faults # 0.072 M/sec ( +- 0.173% )\n 34.69 cache miss/faults\n\nChangelog (since Oct/2):\n - updated comments\n - replaced get_cpu_var() with __get_cpu_var() if possible.\n - removed mutex for system-wide drain. adds a counter instead of it.\n - removed CONFIG_HOTPLUG_CPU\n\nChangelog (old):\n - rebased onto the latest mmotm\n - moved charge size check before __GFP_WAIT check for avoiding unnecesary\n - added asynchronous flush routine.\n - fixed bugs pointed out by Nishimura-san.\n\n[akpm@linux-foundation.org: tweak comments]\n[nishimura@mxp.nes.nec.co.jp: don\u0027t do INIT_WORK() repeatedly against the same work_struct]\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "569b846df54ffb2827b83ce3244c5f032394cba4", "tree": "77c5d373a5edf97710fab8777912971b99e84828", "parents": [ "cd9b45b78a61e8df250e69385c74e729e5b66abf" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Tue Dec 15 16:47:03 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Dec 16 07:20:07 2009 -0800" }, "message": "memcg: coalesce uncharge during unmap/truncate\n\nIn massive parallel enviroment, res_counter can be a performance\nbottleneck. One strong techinque to reduce lock contention is reducing\ncalls by coalescing some amount of calls into one.\n\nConsidering charge/uncharge chatacteristic,\n\t- charge is done one by one via demand-paging.\n\t- uncharge is done by\n\t\t- in chunk at munmap, truncate, exit, execve...\n\t\t- one by one via vmscan/paging.\n\nIt seems we have a chance to coalesce uncharges for improving scalability\nat unmap/truncation.\n\nThis patch is a for coalescing uncharge. For avoiding scattering memcg\u0027s\nstructure to functions under /mm, this patch adds memcg batch uncharge\ninformation to the task. A reason for per-task batching is for making use\nof caller\u0027s context information. We do batched uncharge (deleyed\nuncharge) when truncation/unmap occurs but do direct uncharge when\nuncharge is called by memory reclaim (vmscan.c).\n\nThe degree of coalescing depends on callers\n - at invalidate/trucate... pagevec size\n - at unmap ....ZAP_BLOCK_SIZE\n(memory itself will be freed in this degree.)\nThen, we\u0027ll not coalescing too much.\n\nOn x86-64 8cpu server, I tested overheads of memcg at page fault by\nrunning a program which does map/fault/unmap in a loop. Running\na task per a cpu by taskset and see sum of the number of page faults\nin 60secs.\n\n[without memcg config]\n 40156968 page-faults # 0.085 M/sec ( +- 0.046% )\n 27.67 cache-miss/faults\n[root cgroup]\n 36659599 page-faults # 0.077 M/sec ( +- 0.247% )\n 31.58 miss/faults\n[in a child cgroup]\n 18444157 page-faults # 0.039 M/sec ( +- 0.133% )\n 69.96 miss/faults\n[child with this patch]\n 27133719 page-faults # 0.057 M/sec ( +- 0.155% )\n 47.16 miss/faults\n\nWe can see some amounts of improvement.\n(root cgroup doesn\u0027t affected by this patch)\nAnother patch for \"charge\" will follow this and above will be improved more.\n\nChangelog(since 2009/10/02):\n - renamed filed of memcg_batch (as pages to bytes, memsw to memsw_bytes)\n - some clean up and commentary/description updates.\n - added initialize code to copy_process(). (possible bug fix)\n\nChangelog(old):\n - fixed !CONFIG_MEM_CGROUP case.\n - rebased onto the latest mmotm + softlimit fix patches.\n - unified patch for callers\n - added commetns.\n - make -\u003edo_batch as bool.\n - removed css_get() at el. We don\u0027t need it.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "cd9b45b78a61e8df250e69385c74e729e5b66abf", "tree": "8cb92363dd4b936814de9a0823016889b3441fde", "parents": [ "6be4b78993498c253e99b12c4d0f7684a36955e2" ], "author": { "name": "Kirill A. Shutemov", "email": "kirill@shutemov.name", "time": "Tue Dec 15 16:47:01 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Dec 16 07:20:07 2009 -0800" }, "message": "memcg: fix memory.memsw.usage_in_bytes for root cgroup\n\nA memory cgroup has a memory.memsw.usage_in_bytes file. It shows the sum\nof the usage of pages and swapents in the cgroup. Presently the root\ncgroup\u0027s memsw.usage_in_bytes shows the wrong value - the number of\nswapents are not added.\n\nSo take MEM_CGROUP_STAT_SWAPOUT into account.\n\nSigned-off-by: Kirill A. Shutemov \u003ckirill@shutemov.name\u003e\nReviewed-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: \u003cstable@kernel.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "d324236b3333e87c8825b35f2104184734020d35", "tree": "e8ad6b50e50f975b692cc6a15d5812a51d0047ad", "parents": [ "e42d9d5d47961fb5db0be65b56dd52fe7b2421f1" ], "author": { "name": "Wu Fengguang", "email": "fengguang.wu@intel.com", "time": "Wed Dec 16 12:19:59 2009 +0100" }, "committer": { "name": "Andi Kleen", "email": "ak@linux.intel.com", "time": "Wed Dec 16 12:19:59 2009 +0100" }, "message": "memcg: add accessor to mem_cgroup.css\n\nSo that an outside user can free the reference count grabbed by\ntry_get_mem_cgroup_from_page().\n\nCC: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCC: Hugh Dickins \u003chugh.dickins@tiscali.co.uk\u003e\nCC: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCC: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Wu Fengguang \u003cfengguang.wu@intel.com\u003e\nSigned-off-by: Andi Kleen \u003cak@linux.intel.com\u003e\n" }, { "commit": "e42d9d5d47961fb5db0be65b56dd52fe7b2421f1", "tree": "6a898c8d43ee0533581076342b9a7a97cca0509e", "parents": [ "478c5ffc0b50527bd2390f2daa46cc16276b8413" ], "author": { "name": "Wu Fengguang", "email": "fengguang.wu@intel.com", "time": "Wed Dec 16 12:19:59 2009 +0100" }, "committer": { "name": "Andi Kleen", "email": "ak@linux.intel.com", "time": "Wed Dec 16 12:19:59 2009 +0100" }, "message": "memcg: rename and export try_get_mem_cgroup_from_page()\n\nSo that the hwpoison injector can get mem_cgroup for arbitrary page\nand thus know whether it is owned by some mem_cgroup task(s).\n\n[AK: Merged with latest git tree]\n\nCC: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCC: Hugh Dickins \u003chugh.dickins@tiscali.co.uk\u003e\nCC: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCC: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Wu Fengguang \u003cfengguang.wu@intel.com\u003e\nSigned-off-by: Andi Kleen \u003cak@linux.intel.com\u003e\n" }, { "commit": "407f9c8b0889ced1dbe2f9157e4e60c61329d5c9", "tree": "30c0c9ad224a73de621bc23909cd89cd81eda1aa", "parents": [ "80e148226028257ec0a1909d99b2c40d0ffe17f2" ], "author": { "name": "Hugh Dickins", "email": "hugh.dickins@tiscali.co.uk", "time": "Mon Dec 14 17:59:30 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Dec 15 08:53:19 2009 -0800" }, "message": "ksm: mem cgroup charge swapin copy\n\nBut ksm swapping does require one small change in mem cgroup handling.\nWhen do_swap_page()\u0027s call to ksm_might_need_to_copy() does indeed\nsubstitute a duplicate page to accommodate a different anon_vma (or a the\n!PageSwapCache check in mem_cgroup_try_charge_swapin().\n\nThat was returning success without charging, on the assumption that\npte_same() would fail after, which is not the case here. Originally I\nproposed that success, so that an unshrinkable mem cgroup at its limit\nwould not fail unnecessarily; but that\u0027s a minor point, and there are\nplenty of other places where we may fail an overallocation which might\nlater prove unnecessary. So just go ahead and do what all the other\nexceptions do: proceed to charge current mm.\n\nSigned-off-by: Hugh Dickins \u003chugh.dickins@tiscali.co.uk\u003e\nCc: Izik Eidus \u003cieidus@redhat.com\u003e\nCc: Andrea Arcangeli \u003caarcange@redhat.com\u003e\nCc: Chris Wright \u003cchrisw@redhat.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "af901ca181d92aac3a7dc265144a9081a86d8f39", "tree": "380054af22521144fbe1364c3bcd55ad24c9bde4", "parents": [ "972b94ffb90ea6d20c589d9a47215df103388ddd" ], "author": { "name": "André Goddard Rosa", "email": "andre.goddard@gmail.com", "time": "Sat Nov 14 13:09:05 2009 -0200" }, "committer": { "name": "Jiri Kosina", "email": "jkosina@suse.cz", "time": "Fri Dec 04 15:39:55 2009 +0100" }, "message": "tree-wide: fix assorted typos all over the place\n\nThat is \"success\", \"unknown\", \"through\", \"performance\", \"[re|un]mapping\"\n, \"access\", \"default\", \"reasonable\", \"[con]currently\", \"temperature\"\n, \"channel\", \"[un]used\", \"application\", \"example\",\"hierarchy\", \"therefore\"\n, \"[over|under]flow\", \"contiguous\", \"threshold\", \"enough\" and others.\n\nSigned-off-by: André Goddard Rosa \u003candre.goddard@gmail.com\u003e\nSigned-off-by: Jiri Kosina \u003cjkosina@suse.cz\u003e\n" }, { "commit": "21ae2956ce289f61f11863cc67080f9a28101ae0", "tree": "c6f66176b5f151a1326879de334a94201eeb0f26", "parents": [ "0e6779bbcb78b434efdc84fff1809e342684d6b3" ], "author": { "name": "Uwe Kleine-König", "email": "u.kleine-koenig@pengutronix.de", "time": "Wed Oct 07 15:21:09 2009 +0200" }, "committer": { "name": "Jiri Kosina", "email": "jkosina@suse.cz", "time": "Mon Nov 09 09:40:57 2009 +0100" }, "message": "tree-wide: fix typos \"aquire\" -\u003e \"acquire\", \"cumsumed\" -\u003e \"consumed\"\n\nThis patch was generated by\n\n\tgit grep -E -i -l \u0027[Aa]quire\u0027 | xargs -r perl -p -i -e \u0027s/([Aa])quire/$1cquire/\u0027\n\nand the cumsumed was found by checking the diff for aquire.\n\nSigned-off-by: Uwe Kleine-König \u003cu.kleine-koenig@pengutronix.de\u003e\nSigned-off-by: Jiri Kosina \u003cjkosina@suse.cz\u003e\n" }, { "commit": "ef8745c1e7fc5413d760b3b958f3fd3a0beaad72", "tree": "a1f1998dbcf06e84fe3539192e440e9d1bb876f2", "parents": [ "4e649152cbaa1aedd01821d200ab9d597fe469e4" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Thu Oct 01 15:44:12 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Oct 01 16:11:13 2009 -0700" }, "message": "memcg: reduce check for softlimit excess\n\nIn charge/uncharge/reclaim path, usage_in_excess is calculated repeatedly\nand it takes res_counter\u0027s spin_lock every time.\n\nThis patch removes unnecessary calls for res_count_soft_limit_excess.\n\nReviewed-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "4e649152cbaa1aedd01821d200ab9d597fe469e4", "tree": "635fa7d75acda929e81b8b0db7e641b7d4e07b35", "parents": [ "3dece8347df6a16239fab10dadb370854f1c969c" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Thu Oct 01 15:44:11 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Oct 01 16:11:13 2009 -0700" }, "message": "memcg: some modification to softlimit under hierarchical memory reclaim.\n\nThis patch clean up/fixes for memcg\u0027s uncharge soft limit path.\n\nProblems:\n Now, res_counter_charge()/uncharge() handles softlimit information at\n charge/uncharge and softlimit-check is done when event counter per memcg\n goes over limit. Now, event counter per memcg is updated only when\n memory usage is over soft limit. Here, considering hierarchical memcg\n management, ancesotors should be taken care of.\n\n Now, ancerstors(hierarchy) are handled in charge() but not in uncharge().\n This is not good.\n\n Prolems:\n 1. memcg\u0027s event counter incremented only when softlimit hits. That\u0027s bad.\n It makes event counter hard to be reused for other purpose.\n\n 2. At uncharge, only the lowest level rescounter is handled. This is bug.\n Because ancesotor\u0027s event counter is not incremented, children should\n take care of them.\n\n 3. res_counter_uncharge()\u0027s 3rd argument is NULL in most case.\n ops under res_counter-\u003elock should be small. No \"if\" sentense is better.\n\nFixes:\n * Removed soft_limit_xx poitner and checks in charge and uncharge.\n Do-check-only-when-necessary scheme works enough well without them.\n\n * make event-counter of memcg incremented at every charge/uncharge.\n (per-cpu area will be accessed soon anyway)\n\n * All ancestors are checked at soft-limit-check. This is necessary because\n ancesotor\u0027s event counter may never be modified. Then, they should be\n checked at the same time.\n\nReviewed-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "26251eaf98e26dc2ce2dc26d63bc502700760704", "tree": "f783cab10c5244c170e7d1a92b6a80c574cdf114", "parents": [ "447e4460a4ef44a275f81d992d227f34673be2a8" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Thu Oct 01 15:44:08 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Oct 01 16:11:12 2009 -0700" }, "message": "memcg: fix refcnt going negative\n\n__mem_cgroup_largest_soft_limit_node() returns a mem_cgroup_per_zone \"mz\"\nwith incremnted mz-\u003emem-\u003ecss\u0027s refcnt. Then, the caller of this function\nhas to call css_put(mz-\u003emem-\u003ecss).\n\nBut, mz can be !NULL even if \"not found\" i.e. without css_get(). By\nthis, css-\u003erefcnt will go down to minus.\n\nThis may cause various things...one of results will be\ninitite-loop in css_tryget() as this.\n\nINFO: RCU detected CPU 0 stall (t\u003d10000 jiffies)\nsending NMI to all CPUs:\nNMI backtrace for cpu 0\nCPU 0:\n\u003csnip\u003e\n\n \u003c\u003cEOE\u003e\u003e \u003cIRQ\u003e [\u003cffffffff810884bd\u003e] trace_hardirqs_off+0xd/0x10\n [\u003cffffffff8102a940\u003e] flat_send_IPI_mask+0x90/0xb0\n [\u003cffffffff8102a9c9\u003e] flat_send_IPI_all+0x69/0x70\n [\u003cffffffff81027372\u003e] arch_trigger_all_cpu_backtrace+0x62/0xa0\n [\u003cffffffff810bff8e\u003e] __rcu_pending+0x7e/0x370\n [\u003cffffffff810c02c7\u003e] rcu_check_callbacks+0x47/0x130\n [\u003cffffffff81063a26\u003e] update_process_times+0x46/0x70\n [\u003cffffffff81085930\u003e] tick_sched_timer+0x60/0x160\n [\u003cffffffff810858d0\u003e] ? tick_sched_timer+0x0/0x160\n [\u003cffffffff8107a03a\u003e] __run_hrtimer+0xba/0x150\n [\u003cffffffff8107a325\u003e] hrtimer_interrupt+0xd5/0x1b0\n [\u003cffffffff81426dfe\u003e] ? trace_hardirqs_off_thunk+0x3a/0x3c\n [\u003cffffffff8142cacd\u003e] smp_apic_timer_interrupt+0x6d/0x9b\n [\u003cffffffff8100cb33\u003e] apic_timer_interrupt+0x13/0x20\n \u003cEOI\u003e [\u003cffffffff811317b6\u003e] ? mem_cgroup_walk_tree+0x156/0x180\n [\u003cffffffff811316d3\u003e] ? mem_cgroup_walk_tree+0x73/0x180\n [\u003cffffffff81131692\u003e] ? mem_cgroup_walk_tree+0x32/0x180\n [\u003cffffffff81131a00\u003e] ? mem_cgroup_get_local_stat+0x0/0x110\n [\u003cffffffff81131d5b\u003e] ? mem_control_stat_show+0x14b/0x330\n [\u003cffffffff810a57fd\u003e] ? cgroup_seqfile_show+0x3d/0x60\n\nAbove shows CPU0 caught in css_tryget()\u0027s inifinite loop because\nof bad refcnt.\n\nThis is a fix to set mz\u003dNULL at the top of retry path.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "1dd3a27326d307952f8ad2499478c84dc7311517", "tree": "ae59d52935c790dd3fd0c8e12af5855cfe1abfcb", "parents": [ "0c3e73e84fe3f64cf1c2e8bb4e91e8901cbcdc38" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Wed Sep 23 15:56:43 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Sep 24 07:20:59 2009 -0700" }, "message": "memcg: show swap usage in stat file\n\nWe now count MEM_CGROUP_STAT_SWAPOUT, so we can show swap usage. It would\nbe useful for users to show swap usage in memory.stat file, because they\ndon\u0027t need calculate memsw.usage - res.usage to know swap usage.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nReviewed-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nReviewed-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "0c3e73e84fe3f64cf1c2e8bb4e91e8901cbcdc38", "tree": "1c3e27fce3babd3703b34c9a353f253fcefb6325", "parents": [ "4e41695356fb4e0b153be1440ad027e46e0a7ea2" ], "author": { "name": "Balbir Singh", "email": "balbir@linux.vnet.ibm.com", "time": "Wed Sep 23 15:56:42 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Sep 24 07:20:59 2009 -0700" }, "message": "memcg: improve resource counter scalability\n\nReduce the resource counter overhead (mostly spinlock) associated with the\nroot cgroup. This is a part of the several patches to reduce mem cgroup\noverhead. I had posted other approaches earlier (including using percpu\ncounters). Those patches will be a natural addition and will be added\niteratively on top of these.\n\nThe patch stops resource counter accounting for the root cgroup. The data\nfor display is derived from the statisitcs we maintain via\nmem_cgroup_charge_statistics (which is more scalable). What happens today\nis that, we do double accounting, once using res_counter_charge() and once\nusing memory_cgroup_charge_statistics(). For the root, since we don\u0027t\nimplement limits any more, we don\u0027t need to track every charge via\nres_counter_charge() and check for limit being exceeded and reclaim.\n\nThe main mem-\u003eres usage_in_bytes can be derived by summing the cache and\nrss usage data from memory statistics (MEM_CGROUP_STAT_RSS and\nMEM_CGROUP_STAT_CACHE). However, for memsw-\u003eres usage_in_bytes, we need\nadditional data about swapped out memory. This patch adds a\nMEM_CGROUP_STAT_SWAPOUT and uses that along with MEM_CGROUP_STAT_RSS and\nMEM_CGROUP_STAT_CACHE to derive the memsw data. This data is computed\nrecursively when hierarchy is enabled.\n\nThe tests results I see on a 24 way show that\n\n1. The lock contention disappears from /proc/lock_stats\n2. The results of the test are comparable to running with\n cgroup_disable\u003dmemory.\n\nHere is a sample of my program runs\n\nWithout Patch\n\n Performance counter stats for \u0027/home/balbir/parallel_pagefault\u0027:\n\n 7192804.124144 task-clock-msecs # 23.937 CPUs\n 424691 context-switches # 0.000 M/sec\n 267 CPU-migrations # 0.000 M/sec\n 28498113 page-faults # 0.004 M/sec\n 5826093739340 cycles # 809.989 M/sec\n 408883496292 instructions # 0.070 IPC\n 7057079452 cache-references # 0.981 M/sec\n 3036086243 cache-misses # 0.422 M/sec\n\n 300.485365680 seconds time elapsed\n\nWith cgroup_disable\u003dmemory\n\n Performance counter stats for \u0027/home/balbir/parallel_pagefault\u0027:\n\n 7182183.546587 task-clock-msecs # 23.915 CPUs\n 425458 context-switches # 0.000 M/sec\n 203 CPU-migrations # 0.000 M/sec\n 92545093 page-faults # 0.013 M/sec\n 6034363609986 cycles # 840.185 M/sec\n 437204346785 instructions # 0.072 IPC\n 6636073192 cache-references # 0.924 M/sec\n 2358117732 cache-misses # 0.328 M/sec\n\n 300.320905827 seconds time elapsed\n\nWith this patch applied\n\n Performance counter stats for \u0027/home/balbir/parallel_pagefault\u0027:\n\n 7191619.223977 task-clock-msecs # 23.955 CPUs\n 422579 context-switches # 0.000 M/sec\n 88 CPU-migrations # 0.000 M/sec\n 91946060 page-faults # 0.013 M/sec\n 5957054385619 cycles # 828.333 M/sec\n 1058117350365 instructions # 0.178 IPC\n 9161776218 cache-references # 1.274 M/sec\n 1920494280 cache-misses # 0.267 M/sec\n\n 300.218764862 seconds time elapsed\n\nData from Prarit (kernel compile with make -j64 on a 64\nCPU/32G machine)\n\nFor a single run\n\nWithout patch\n\nreal 27m8.988s\nuser 87m24.916s\nsys 382m6.037s\n\nWith patch\n\nreal 4m18.607s\nuser 84m58.943s\nsys 50m52.682s\n\nWith config turned off\n\nreal 4m54.972s\nuser 90m13.456s\nsys 50m19.711s\n\nNOTE: The data looks counterintuitive due to the increased performance\nwith the patch, even over the config being turned off. We probably need\nmore runs, but so far all testing has shown that the patches definitely\nhelp.\n\n[akpm@linux-foundation.org: coding-style fixes]\nSigned-off-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Prarit Bhargava \u003cprarit@redhat.com\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nReviewed-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nReviewed-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "4e41695356fb4e0b153be1440ad027e46e0a7ea2", "tree": "547dae77d1655a1acb260ea8b266c7b8a48f2d2c", "parents": [ "75822b4495b62e8721e9b88e3cf9e653a0c85b73" ], "author": { "name": "Balbir Singh", "email": "balbir@linux.vnet.ibm.com", "time": "Wed Sep 23 15:56:39 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Sep 24 07:20:59 2009 -0700" }, "message": "memory controller: soft limit reclaim on contention\n\nImplement reclaim from groups over their soft limit\n\nPermit reclaim from memory cgroups on contention (via the direct reclaim\npath).\n\nmemory cgroup soft limit reclaim finds the group that exceeds its soft\nlimit by the largest number of pages and reclaims pages from it and then\nreinserts the cgroup into its correct place in the rbtree.\n\nAdd additional checks to mem_cgroup_hierarchical_reclaim() to detect long\nloops in case all swap is turned off. The code has been refactored and\nthe loop check (loop \u003c 2) has been enhanced for soft limits. For soft\nlimits, we try to do more targetted reclaim. Instead of bailing out after\ntwo loops, the routine now reclaims memory proportional to the size by\nwhich the soft limit is exceeded. The proportion has been empirically\ndetermined.\n\n[akpm@linux-foundation.org: build fix]\n[kamezawa.hiroyu@jp.fujitsu.com: fix softlimit css refcnt handling]\n[nishimura@mxp.nes.nec.co.jp: refcount of the \"victim\" should be decremented before exiting the loop]\nSigned-off-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nAcked-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "75822b4495b62e8721e9b88e3cf9e653a0c85b73", "tree": "b8d4c10f03a4e289f0a66b982243fd8980f9df07", "parents": [ "f64c3f54940d6929a2b6dcffaab942bd62be2e66" ], "author": { "name": "Balbir Singh", "email": "balbir@linux.vnet.ibm.com", "time": "Wed Sep 23 15:56:38 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Sep 24 07:20:59 2009 -0700" }, "message": "memory controller: soft limit refactor reclaim flags\n\nRefactor mem_cgroup_hierarchical_reclaim()\n\nRefactor the arguments passed to mem_cgroup_hierarchical_reclaim() into\nflags, so that new parameters don\u0027t have to be passed as we make the\nreclaim routine more flexible\n\nSigned-off-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "f64c3f54940d6929a2b6dcffaab942bd62be2e66", "tree": "7b3587700b08639970580be6c87f36df80ca8c74", "parents": [ "296c81d89f4f14269f7346f81442910158c0a83a" ], "author": { "name": "Balbir Singh", "email": "balbir@linux.vnet.ibm.com", "time": "Wed Sep 23 15:56:37 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Sep 24 07:20:59 2009 -0700" }, "message": "memory controller: soft limit organize cgroups\n\nOrganize cgroups over soft limit in a RB-Tree\n\nIntroduce an RB-Tree for storing memory cgroups that are over their soft\nlimit. The overall goal is to\n\n1. Add a memory cgroup to the RB-Tree when the soft limit is exceeded.\n We are careful about updates, updates take place only after a particular\n time interval has passed\n2. We remove the node from the RB-Tree when the usage goes below the soft\n limit\n\nThe next set of patches will exploit the RB-Tree to get the group that is\nover its soft limit by the largest amount and reclaim from it, when we\nface memory contention.\n\n[hugh.dickins@tiscali.co.uk: CONFIG_CGROUP_MEM_RES_CTLR\u003dy CONFIG_PREEMPT\u003dy fails to boot]\nSigned-off-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nSigned-off-by: Hugh Dickins \u003chugh.dickins@tiscali.co.uk\u003e\nCc: Jiri Slaby \u003cjirislaby@gmail.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "296c81d89f4f14269f7346f81442910158c0a83a", "tree": "cf0e1facd1fed8282c1885bc4126e7bca1928712", "parents": [ "a6df63615b943dbef22df04c19f4506330fe835e" ], "author": { "name": "Balbir Singh", "email": "balbir@linux.vnet.ibm.com", "time": "Wed Sep 23 15:56:36 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Sep 24 07:20:59 2009 -0700" }, "message": "memory controller: soft limit interface\n\nAdd an interface to allow get/set of soft limits. Soft limits for memory\nplus swap controller (memsw) is currently not supported. Resource\ncounters have been enhanced to support soft limits and new type\nRES_SOFT_LIMIT has been added. Unlike hard limits, soft limits can be\ndirectly set and do not need any reclaim or checks before setting them to\na newer value.\n\nKamezawa-San raised a question as to whether soft limit should belong to\nres_counter. Since all resources understand the basic concepts of hard\nand soft limits, it is justified to add soft limits here. Soft limits are\na generic resource usage feature, even file system quotas support soft\nlimits.\n\nSigned-off-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "261fb61a8bf6d3bd964ae6f1e6af49585d30db51", "tree": "417cebd2e18bc2457b5b09d8359ad71ba6afd56d", "parents": [ "4b3bde4c983de36c59e6c1a24701f6fe816f9f55" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed Sep 23 15:56:33 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Sep 24 07:20:58 2009 -0700" }, "message": "memcg: add comments explaining memory barriers\n\nAdd comments for the reason of smp_wmb() in mem_cgroup_commit_charge().\n\n[akpm@linux-foundation.org: coding-style fixes]\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "4b3bde4c983de36c59e6c1a24701f6fe816f9f55", "tree": "e759c5ceccf57331b868b7feac61cae5e932c6d4", "parents": [ "be367d09927023d081f9199665c8500f69f14d22" ], "author": { "name": "Balbir Singh", "email": "balbir@linux.vnet.ibm.com", "time": "Wed Sep 23 15:56:32 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Sep 24 07:20:58 2009 -0700" }, "message": "memcg: remove the overhead associated with the root cgroup\n\nChange the memory cgroup to remove the overhead associated with accounting\nall pages in the root cgroup. As a side-effect, we can no longer set a\nmemory hard limit in the root cgroup.\n\nA new flag to track whether the page has been accounted or not has been\nadded as well. Flags are now set atomically for page_cgroup,\npcg_default_flags is now obsolete and removed.\n\n[akpm@linux-foundation.org: fix a few documentation glitches]\nSigned-off-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nReviewed-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" } ], "next": "be367d09927023d081f9199665c8500f69f14d22" }