)]}' { "log": [ { "commit": "6a7b95481d49f73991d3dbf8c1e696a24684ac05", "tree": "32f80d8839daf939bcae8d7711e64b8e96d89ce1", "parents": [ "b05ca7385a2848abdc72051f832722641daed8b0" ], "author": { "name": "Johannes Weiner", "email": "hannes@cmpxchg.org", "time": "Mon Oct 26 16:50:00 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Oct 29 07:39:30 2009 -0700" }, "message": "vmscan: order evictable rescue in LRU putback\n\nIsolators putting a page back to the LRU do not hold the page lock, and if\nthe page is mlocked, another thread might munlock it concurrently.\n\nExpecting this, the putback code re-checks the evictability of a page when\nit just moved it to the unevictable list in order to correct its decision.\n\nThe problem, however, is that ordering is not garuanteed between setting\nPG_lru when moving the page to the list and checking PG_mlocked\nafterwards:\n\n\t#0:\t\t\t\t#1\n\n\tspin_lock()\n\t\t\t\t\tif (TestClearPageMlocked())\n\t\t\t\t\t if (PageLRU())\n\t\t\t\t\t move to evictable list\n\tSetPageLRU()\n\tspin_unlock()\n\tif (!PageMlocked())\n\t move to evictable list\n\nThe PageMlocked() check may get reordered before SetPageLRU() in #0,\nresulting in #0 not moving the still mlocked page, and in #1 failing to\nisolate and move the page as well. The page is now stranded on the\nunevictable list.\n\nThe race condition is very unlikely. The consequence currently is one\npage falling off the reclaim grid and eventually getting freed with\nPG_unevictable set, which triggers a warning in the page allocator.\n\nTestClearPageMlocked() in #1 already provides full memory barrier\nsemantics.\n\nThis patch adds an explicit full barrier to force ordering between\nSetPageLRU() and PageMlocked() so that either one of the competitors\nrescues the page.\n\nSigned-off-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nReviewed-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Hugh Dickins \u003chugh.dickins@tiscali.co.uk\u003e\nCc: Mel Gorman \u003cmel@csn.ul.ie\u003e\nCc: Lee Schermerhorn \u003cLee.Schermerhorn@hp.com\u003e\nCc: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nReviewed-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": "41e20983fe553b39bc2b00e07c7a379f0c86a4bc", "tree": "685433b37a7ebe5a61ac7315dc8017706227345b", "parents": [ "b76146ed1ae7d7acae1d51f9342e31d00c8d5a12" ], "author": { "name": "Wu Fengguang", "email": "fengguang.wu@intel.com", "time": "Mon Oct 26 16:49:53 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Oct 29 07:39:27 2009 -0700" }, "message": "vmscan: limit VM_EXEC protection to file pages\n\nIt is possible to have !Anon but SwapBacked pages, and some apps could\ncreate huge number of such pages with MAP_SHARED|MAP_ANONYMOUS. These\npages go into the ANON lru list, and hence shall not be protected: we only\ncare mapped executable files. Failing to do so may trigger OOM.\n\nTested-by: Christian Borntraeger \u003cborntraeger@de.ibm.com\u003e\nReviewed-by: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: Wu Fengguang \u003cfengguang.wu@intel.com\u003e\nCc: Hugh Dickins \u003chugh.dickins@tiscali.co.uk\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "58355c7876a0754377c37c8af948b4cd423410e2", "tree": "9e49dfe20538307d2c4c8fc92c88432c4298f63b", "parents": [ "0d0df599f1f11f12d589318bacb59a50fb5c0310" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Mon Oct 26 16:49:35 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Oct 29 07:39:25 2009 -0700" }, "message": "congestion_wait(): don\u0027t use WRITE\n\ncommit 8aa7e847d (Fix congestion_wait() sync/async vs read/write\nconfusion) replace WRITE with BLK_RW_ASYNC. Unfortunately, concurrent mm\ndevelopment made the unchanged place accidentally.\n\nThis patch fixes it too.\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nAcked-by: Jens Axboe \u003cjens.axboe@oracle.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": "6d7f18f6ea3a13af95bdf507fc54d42b165e1712", "tree": "8f6f3a6d46835aa767823fa7049609408a87afc2", "parents": [ "53cddfcc0e760d2b364878b6dadbd0c6d087cfae", "56a131dcf7ed36c3c6e36bea448b674ea85ed5bb" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Sep 25 09:27:30 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Sep 25 09:27:30 2009 -0700" }, "message": "Merge branch \u0027writeback\u0027 of git://git.kernel.dk/linux-2.6-block\n\n* \u0027writeback\u0027 of git://git.kernel.dk/linux-2.6-block:\n writeback: writeback_inodes_sb() should use bdi_start_writeback()\n writeback: don\u0027t delay inodes redirtied by a fast dirtier\n writeback: make the super_block pinning more efficient\n writeback: don\u0027t resort for a single super_block in move_expired_inodes()\n writeback: move inodes from one super_block together\n writeback: get rid to incorrect references to pdflush in comments\n writeback: improve readability of the wb_writeback() continue/break logic\n writeback: cleanup writeback_single_inode()\n writeback: kupdate writeback shall not stop when more io is possible\n writeback: stop background writeback when below background threshold\n writeback: balance_dirty_pages() shall write more than dirtied pages\n fs: Fix busyloop in wb_writeback()\n" }, { "commit": "5b0830cb9085f4b69f9d57d7f3aaff322ffbec26", "tree": "10040eb359269d4cd05487790b758144a69e8e39", "parents": [ "71fd05a887e0f3f6bfff76ff81b33776177d0606" ], "author": { "name": "Jens Axboe", "email": "jens.axboe@oracle.com", "time": "Wed Sep 23 19:37:09 2009 +0200" }, "committer": { "name": "Jens Axboe", "email": "jens.axboe@oracle.com", "time": "Fri Sep 25 18:08:25 2009 +0200" }, "message": "writeback: get rid to incorrect references to pdflush in comments\n\nSigned-off-by: Jens Axboe \u003cjens.axboe@oracle.com\u003e\n" }, { "commit": "db16826367fefcb0ddb93d76b66adc52eb4e6339", "tree": "626224c1eb1eb79c522714591f208b4fdbdcd9d4", "parents": [ "cd6045138ed1bb5d8773e940d51c34318eef3ef2", "465fdd97cbe16ef8727221857e96ef62dd352017" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Sep 24 07:53:22 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Sep 24 07:53:22 2009 -0700" }, "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: (21 commits)\n HWPOISON: Enable error_remove_page on btrfs\n HWPOISON: Add simple debugfs interface to inject hwpoison on arbitary PFNs\n HWPOISON: Add madvise() based injector for hardware poisoned pages v4\n HWPOISON: Enable error_remove_page for NFS\n HWPOISON: Enable .remove_error_page for migration aware file systems\n HWPOISON: The high level memory error handler in the VM v7\n HWPOISON: Add PR_MCE_KILL prctl to control early kill behaviour per process\n HWPOISON: shmem: call set_page_dirty() with locked page\n HWPOISON: Define a new error_remove_page address space op for async truncation\n HWPOISON: Add invalidate_inode_page\n HWPOISON: Refactor truncate to allow direct truncating of page v2\n HWPOISON: check and isolate corrupted free pages v2\n HWPOISON: Handle hardware poisoned pages in try_to_unmap\n HWPOISON: Use bitmask/action code for try_to_unmap behaviour\n HWPOISON: x86: Add VM_FAULT_HWPOISON handling to x86 page fault handler v2\n HWPOISON: Add poison check to page fault handling\n HWPOISON: Add basic support for poisoned pages in fault handler v3\n HWPOISON: Add new SIGBUS error codes for hardware poison signals\n HWPOISON: Add support for poison swap entries v2\n HWPOISON: Export some rmap vma locking to outside world\n ...\n" }, { "commit": "8d65af789f3e2cf4cfbdbf71a0f7a61ebcd41d38", "tree": "121df3bfffc7853ac6d2c514ad514d4a748a0933", "parents": [ "c0d0787b6d47d9f4d5e8bd321921104e854a9135" ], "author": { "name": "Alexey Dobriyan", "email": "adobriyan@gmail.com", "time": "Wed Sep 23 15:57:19 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Sep 24 07:21:04 2009 -0700" }, "message": "sysctl: remove \"struct file *\" argument of -\u003eproc_handler\n\nIt\u0027s unused.\n\nIt isn\u0027t needed -- read or write flag is already passed and sysctl\nshouldn\u0027t care about the rest.\n\nIt _was_ used in two places at arch/frv for some reason.\n\nSigned-off-by: Alexey Dobriyan \u003cadobriyan@gmail.com\u003e\nCc: David Howells \u003cdhowells@redhat.com\u003e\nCc: \"Eric W. Biederman\" \u003cebiederm@xmission.com\u003e\nCc: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nCc: Ralf Baechle \u003cralf@linux-mips.org\u003e\nCc: Martin Schwidefsky \u003cschwidefsky@de.ibm.com\u003e\nCc: Ingo Molnar \u003cmingo@elte.hu\u003e\nCc: \"David S. Miller\" \u003cdavem@davemloft.net\u003e\nCc: James Morris \u003cjmorris@namei.org\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": "f168e1b6390e2d79cf57e48e6ae6d9b0a9e2851a", "tree": "8b265240661b26cea90aebd279cd420bfbfcd368", "parents": [ "f86296317434b21585e229f6c49a33cb9ebab4d3" ], "author": { "name": "Vincent Li", "email": "macli@brc.ubc.ca", "time": "Mon Sep 21 17:03:12 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:39 2009 -0700" }, "message": "mm/vmscan: remove page_queue_congested() comment\n\nCommit 084f71ae5c(kill page_queue_congested()) removed\npage_queue_congested(). Remove the page_queue_congested() comment in\nvmscan pageout() too.\n\nSigned-off-by: Vincent Li \u003cmacli@brc.ubc.ca\u003e\nReviewed-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": "f86296317434b21585e229f6c49a33cb9ebab4d3", "tree": "d4fb05d4aee1a8e373ec053e7316dc9847b2c417", "parents": [ "1a8670a29b5277cbe601f74ab63d2c5211fb3005" ], "author": { "name": "Wu Fengguang", "email": "fengguang.wu@intel.com", "time": "Mon Sep 21 17:03:11 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:39 2009 -0700" }, "message": "mm: do batched scans for mem_cgroup\n\nFor mem_cgroup, shrink_zone() may call shrink_list() with nr_to_scan\u003d1, in\nwhich case shrink_list() _still_ calls isolate_pages() with the much\nlarger SWAP_CLUSTER_MAX. It effectively scales up the inactive list scan\nrate by up to 32 times.\n\nFor example, with 16k inactive pages and DEF_PRIORITY\u003d12, (16k \u003e\u003e 12)\u003d4.\nSo when shrink_zone() expects to scan 4 pages in the active/inactive list,\nthe active list will be scanned 4 pages, while the inactive list will be\n(over) scanned SWAP_CLUSTER_MAX\u003d32 pages in effect. And that could break\nthe balance between the two lists.\n\nIt can further impact the scan of anon active list, due to the anon\nactive/inactive ratio rebalance logic in balance_pgdat()/shrink_zone():\n\ninactive anon list over scanned \u003d\u003e inactive_anon_is_low() \u003d\u003d TRUE\n \u003d\u003e shrink_active_list()\n \u003d\u003e active anon list over scanned\n\nSo the end result may be\n\n- anon inactive \u003d\u003e over scanned\n- anon active \u003d\u003e over scanned (maybe not as much)\n- file inactive \u003d\u003e over scanned\n- file active \u003d\u003e under scanned (relatively)\n\nThe accesses to nr_saved_scan are not lock protected and so not 100%\naccurate, however we can tolerate small errors and the resulted small\nimbalanced scan rates between zones.\n\nCc: Rik van Riel \u003criel@redhat.com\u003e\nReviewed-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nAcked-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nReviewed-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: 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": "0b21767637c3c99890a248fe47ac414e51cf5eb7", "tree": "ebedeecddd3e8d7d665464c4fadbea40ee43a317", "parents": [ "2c85f51d222ccdd8c401d77a36b723a89156810d" ], "author": { "name": "Vincent Li", "email": "macli@brc.ubc.ca", "time": "Mon Sep 21 17:03:09 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:38 2009 -0700" }, "message": "mm/vmscan: rename zone_nr_pages() to zone_nr_lru_pages()\n\nThe name `zone_nr_pages\u0027 can be mis-read as zone\u0027s (total) number pages,\nbut it actually returns zone\u0027s LRU list number pages.\n\nSigned-off-by: Vincent Li \u003cmacli@brc.ubc.ca\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Rik van Riel \u003criel@redhat.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": "ceddc3a52d783fabbf1ba623601419b9d6337194", "tree": "651d0fd4d5bf2455468f90364b88f01f982d79af", "parents": [ "edcf4748cd56adcdf0856cc99ef108a4ea3ac7fe" ], "author": { "name": "Johannes Weiner", "email": "hannes@cmpxchg.org", "time": "Mon Sep 21 17:03:00 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:38 2009 -0700" }, "message": "mm: document is_page_cache_freeable()\n\nEnlighten the reader of this code about what reference count makes a page\ncache page freeable.\n\nSigned-off-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nCc: Christoph Lameter \u003ccl@linux-foundation.org\u003e\nReviewed-by: Christoph Lameter \u003ccl@linux-foundation.org\u003e\nReviewed-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": "edcf4748cd56adcdf0856cc99ef108a4ea3ac7fe", "tree": "317d477d08dea82f5eef2e9c17294d0f0639ea81", "parents": [ "6c0b13519d1c755d874e82c8fb8a6dcef0ee402c" ], "author": { "name": "Johannes Weiner", "email": "hannes@cmpxchg.org", "time": "Mon Sep 21 17:02:59 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:38 2009 -0700" }, "message": "mm: return boolean from page_has_private()\n\nMake page_has_private() return a true boolean value and remove the double\nnegations from the two callsites using it for arithmetic.\n\nSigned-off-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nCc: Christoph Lameter \u003ccl@linux-foundation.org\u003e\nReviewed-by: Christoph Lameter \u003ccl@linux-foundation.org\u003e\nReviewed-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": "6c0b13519d1c755d874e82c8fb8a6dcef0ee402c", "tree": "0fe6e6902a488ad6c59ecee971fe64c81edbcce3", "parents": [ "401a8e1c1670085b8177330ca47d4f7c4ac88761" ], "author": { "name": "Johannes Weiner", "email": "hannes@cmpxchg.org", "time": "Mon Sep 21 17:02:59 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:37 2009 -0700" }, "message": "mm: return boolean from page_is_file_cache()\n\npage_is_file_cache() has been used for both boolean checks and LRU\narithmetic, which was always a bit weird.\n\nNow that page_lru_base_type() exists for LRU arithmetic, make\npage_is_file_cache() a real predicate function and adjust the\nboolean-using callsites to drop those pesky double negations.\n\nSigned-off-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nReviewed-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": "401a8e1c1670085b8177330ca47d4f7c4ac88761", "tree": "5883daaed42fa2a186e5769bab1c10535dc39a41", "parents": [ "b7c46d151cb82856a429709d1227ba1648028232" ], "author": { "name": "Johannes Weiner", "email": "hannes@cmpxchg.org", "time": "Mon Sep 21 17:02:58 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:35 2009 -0700" }, "message": "mm: introduce page_lru_base_type()\n\nInstead of abusing page_is_file_cache() for LRU list index arithmetic, add\nanother helper with a more appropriate name and convert the non-boolean\nusers of page_is_file_cache() accordingly.\n\nThis new helper gives the LRU base type a page is supposed to live on,\ninactive anon or inactive file.\n\n[hugh.dickins@tiscali.co.uk: convert del_page_from_lru() also]\nSigned-off-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nReviewed-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nReviewed-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": "b7c46d151cb82856a429709d1227ba1648028232", "tree": "b93012b4bb2ffd603c020b38d8f1f6f3b4714ff7", "parents": [ "bba78819548a59a52e60f0b259997bbd011164ae" ], "author": { "name": "Johannes Weiner", "email": "hannes@cmpxchg.org", "time": "Mon Sep 21 17:02:56 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:35 2009 -0700" }, "message": "mm: drop unneeded double negations\n\nRemove double negations where the operand is already boolean.\n\nSigned-off-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nCc: Mel Gorman \u003cmel@csn.ul.ie\u003e\nReviewed-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": "a26f5320c4ee3d46a0da48fa0f3ac6a00b575793", "tree": "76e4502dae5d278c9a3cebbeda53f01aa1cf3e67", "parents": [ "74a1c48fb4e9f10e3c83dcd39af73487968e35bf" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Mon Sep 21 17:01:46 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:30 2009 -0700" }, "message": "vmscan: kill unnecessary prefetch\n\nThe pages in the list passed move_active_pages_to_lru() are already\ntouched by shrink_active_list(). IOW the prefetch in\nmove_active_pages_to_lru() don\u0027t populate any cache. it\u0027s pointless.\n\nThis patch remove it.\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nReviewed-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nCc: Mel Gorman \u003cmel@csn.ul.ie\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": "74a1c48fb4e9f10e3c83dcd39af73487968e35bf", "tree": "8a9facf333d114c9e5b4be37278951b35b21ff2b", "parents": [ "5205e56eeab04ce02f8bb6b47d1569b216bc0b6a" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Mon Sep 21 17:01:45 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:30 2009 -0700" }, "message": "vmscan: kill unnecessary page flag test\n\nThe page_lru() already evaluate PageActive() and PageSwapBacked(). We\ndon\u0027t need to re-evaluate it.\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nReviewed-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nCc: Mel Gorman \u003cmel@csn.ul.ie\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": "5205e56eeab04ce02f8bb6b47d1569b216bc0b6a", "tree": "ef27909cc2f016ddd7dc6c34fb5a37ca23dcadd3", "parents": [ "de2e7567c7ddf24f0ca80010163ed10da66a14e2" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Mon Sep 21 17:01:44 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:30 2009 -0700" }, "message": "vmscan: move ClearPageActive from move_active_pages() to shrink_active_list()\n\nThe move_active_pages_to_lru() function is called under irq disabled and\nClearPageActive() doesn\u0027t need irq disabling.\n\nThen, this patch move it into shrink_active_list().\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nReviewed-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nCc: Mel Gorman \u003cmel@csn.ul.ie\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": "de2e7567c7ddf24f0ca80010163ed10da66a14e2", "tree": "203eeb5deda2ef11e46ee449134ba9afa0c8579b", "parents": [ "adea02a1bea71a508da32c04d715485a1fe62029" ], "author": { "name": "Minchan Kim", "email": "minchan.kim@gmail.com", "time": "Mon Sep 21 17:01:43 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:30 2009 -0700" }, "message": "vmscan: don\u0027t attempt to reclaim anon page in lumpy reclaim when no swap space is available\n\nThe VM already avoids attempting to reclaim anon pages in various places,\nBut it doesn\u0027t avoid it for lumpy reclaim.\n\nIt shuffles lru list unnecessary so that it is pointless.\n\n[akpm@linux-foundation.org: cleanup]\nSigned-off-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nReviewed-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nReviewed-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: Mel Gorman \u003cmel@csn.ul.ie\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": "adea02a1bea71a508da32c04d715485a1fe62029", "tree": "c78742bbab36bf3b8d20f84b4dc6dc6585bb7cb4", "parents": [ "55c37a840d9ec0ebed5c944355156d490b1ad5d1" ], "author": { "name": "Wu Fengguang", "email": "fengguang.wu@intel.com", "time": "Mon Sep 21 17:01:42 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:30 2009 -0700" }, "message": "mm: count only reclaimable lru pages\n\nglobal_lru_pages() / zone_lru_pages() can be used in two ways:\n- to estimate max reclaimable pages in determine_dirtyable_memory()\n- to calculate the slab scan ratio\n\nWhen swap is full or not present, the anon lru lists are not reclaimable\nand also won\u0027t be scanned. So the anon pages shall not be counted in both\nusage scenarios. Also rename to _reclaimable_pages: now they are counting\nthe possibly reclaimable lru pages.\n\nIt can greatly (and correctly) increase the slab scan rate under high\nmemory pressure (when most file pages have been reclaimed and swap is\nfull/absent), thus reduce false OOM kills.\n\nAcked-by: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nReviewed-by: Rik van Riel \u003criel@redhat.com\u003e\nReviewed-by: Christoph Lameter \u003ccl@linux-foundation.org\u003e\nReviewed-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nSigned-off-by: Wu Fengguang \u003cfengguang.wu@intel.com\u003e\nAcked-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nReviewed-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nReviewed-by: Jesse Barnes \u003cjbarnes@virtuousgeek.org\u003e\nCc: David Howells \u003cdhowells@redhat.com\u003e\nCc: \"Li, Ming Chun\" \u003cmacli@brc.ubc.ca\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "35cd78156c499ef83f60605e4643d5a98fef14fd", "tree": "67cdc9019d4b110b9d57634bd347c8dad6bf8346", "parents": [ "a731286de62294b63d8ceb3c5914ac52cc17e690" ], "author": { "name": "Rik van Riel", "email": "riel@redhat.com", "time": "Mon Sep 21 17:01:38 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:29 2009 -0700" }, "message": "vmscan: throttle direct reclaim when too many pages are isolated already\n\nWhen way too many processes go into direct reclaim, it is possible for all\nof the pages to be taken off the LRU. One result of this is that the next\nprocess in the page reclaim code thinks there are no reclaimable pages\nleft and triggers an out of memory kill.\n\nOne solution to this problem is to never let so many processes into the\npage reclaim path that the entire LRU is emptied. Limiting the system to\nonly having half of each inactive list isolated for reclaim should be\nsafe.\n\nSigned-off-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Wu Fengguang \u003cfengguang.wu@intel.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": "a731286de62294b63d8ceb3c5914ac52cc17e690", "tree": "c321e14500ec264e37fd103ffa71c7b133088010", "parents": [ "b35ea17b7bbf5dea35faa0de11030acc620c3197" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Mon Sep 21 17:01:37 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:29 2009 -0700" }, "message": "mm: vmstat: add isolate pages\n\nIf the system is running a heavy load of processes then concurrent reclaim\ncan isolate a large number of pages from the LRU. /proc/vmstat and the\noutput generated for an OOM do not show how many pages were isolated.\n\nThis has been observed during process fork bomb testing (mstctl11 in LTP).\n\nThis patch shows the information about isolated pages.\n\nReproduced via:\n\n-----------------------\n% ./hackbench 140 process 1000\n \u003d\u003e OOM occur\n\nactive_anon:146 inactive_anon:0 isolated_anon:49245\n active_file:79 inactive_file:18 isolated_file:113\n unevictable:0 dirty:0 writeback:0 unstable:0 buffer:39\n free:370 slab_reclaimable:309 slab_unreclaimable:5492\n mapped:53 shmem:15 pagetables:28140 bounce:0\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nAcked-by: Wu Fengguang \u003cfengguang.wu@intel.com\u003e\nReviewed-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nCc: Hugh Dickins \u003chugh.dickins@tiscali.co.uk\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "b35ea17b7bbf5dea35faa0de11030acc620c3197", "tree": "fba2dabec71193f769c16c3737e152900682bbfa", "parents": [ "44c241f166b31999482c3c40448f4bbb2157a804" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Mon Sep 21 17:01:36 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:28 2009 -0700" }, "message": "mm: shrink_inactive_list() nr_scan accounting fix fix\n\nIf sc-\u003eisolate_pages() return 0, we don\u0027t need to call shrink_page_list().\nIn past days, shrink_inactive_list() handled it properly.\n\nBut commit fb8d14e1 (three years ago commit!) breaked it. current\nshrink_inactive_list() always call shrink_page_list() although\nisolate_pages() return 0.\n\nThis patch restore proper return value check.\n\nRequirements:\n o \"nr_taken \u003d\u003d 0\" condition should stay before calling shrink_page_list().\n o \"nr_taken \u003d\u003d 0\" condition should stay after nr_scan related statistics\n modification.\n\nCc: Wu Fengguang \u003cfengguang.wu@intel.com\u003e\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nReviewed-by: Rik van Riel \u003criel@redhat.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": "44c241f166b31999482c3c40448f4bbb2157a804", "tree": "1bfa7e8940de223e0ac35e0598c942f4e4404e20", "parents": [ "b259fbde0a86085264c89aa2ce9c6e35792a1aad" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Mon Sep 21 17:01:35 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Sep 22 07:17:27 2009 -0700" }, "message": "mm: rename pgmoved variable in shrink_active_list()\n\nCurrently the pgmoved variable has two meanings. It causes harder\nreviewing. This patch separates it.\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nReviewed-by: Rik van Riel \u003criel@redhat.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": "14fa31b89c5ae79e4131da41761378a6df674352", "tree": "c6c79e89e0aa0b2efeaf657d4715250a406ab699", "parents": [ "a6e04aa92965565968573a220a35b4e907385697" ], "author": { "name": "Andi Kleen", "email": "andi@firstfloor.org", "time": "Wed Sep 16 11:50:10 2009 +0200" }, "committer": { "name": "Andi Kleen", "email": "ak@linux.intel.com", "time": "Wed Sep 16 11:50:10 2009 +0200" }, "message": "HWPOISON: Use bitmask/action code for try_to_unmap behaviour\n\ntry_to_unmap currently has multiple modi (migration, munlock, normal unmap)\nwhich are selected by magic flag variables. The logic is not very straight\nforward, because each of these flag change multiple behaviours (e.g.\nmigration turns off aging, not only sets up migration ptes etc.)\nAlso the different flags interact in magic ways.\n\nA later patch in this series adds another mode to try_to_unmap, so\nthis becomes quickly unmanageable.\n\nReplace the different flags with a action code (migration, munlock, munmap)\nand some additional flags as modifiers (ignore mlock, ignore aging).\nThis makes the logic more straight forward and allows easier extension\nto new behaviours. Change all the caller to declare what they want to\ndo.\n\nThis patch is supposed to be a nop in behaviour. If anyone can prove\nit is not that would be a bug.\n\nCc: Lee.Schermerhorn@hp.com\nCc: npiggin@suse.de\n\nSigned-off-by: Andi Kleen \u003cak@linux.intel.com\u003e\n" }, { "commit": "03ba3782e8dcc5b0e1efe440d33084f066e38cae", "tree": "e5a6513b411de16a46199530ec98ef9b7f1efc50", "parents": [ "66f3b8e2e103a0b93b945764d98e9ba46cb926dd" ], "author": { "name": "Jens Axboe", "email": "jens.axboe@oracle.com", "time": "Wed Sep 09 09:08:54 2009 +0200" }, "committer": { "name": "Jens Axboe", "email": "jens.axboe@oracle.com", "time": "Fri Sep 11 09:20:25 2009 +0200" }, "message": "writeback: switch to per-bdi threads for flushing data\n\nThis gets rid of pdflush for bdi writeout and kupdated style cleaning.\npdflush writeout suffers from lack of locality and also requires more\nthreads to handle the same workload, since it has to work in a\nnon-blocking fashion against each queue. This also introduces lumpy\nbehaviour and potential request starvation, since pdflush can be starved\nfor queue access if others are accessing it. A sample ffsb workload that\ndoes random writes to files is about 8% faster here on a simple SATA drive\nduring the benchmark phase. File layout also seems a LOT more smooth in\nvmstat:\n\n r b swpd free buff cache si so bi bo in cs us sy id wa\n 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42\n 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44\n 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37\n 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58\n 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34\n 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37\n 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44\n 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38\n 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41\n 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45\n\nwhere vanilla tends to fluctuate a lot in the creation phase:\n\n r b swpd free buff cache si so bi bo in cs us sy id wa\n 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36\n 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51\n 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40\n 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37\n 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41\n 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49\n 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36\n 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43\n 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39\n 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45\n 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34\n 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54\n\nA 10 disk test with btrfs performs 26% faster with per-bdi flushing. A\nSSD based writeback test on XFS performs over 20% better as well, with\nthe throughput being very stable around 1GB/sec, where pdflush only\nmanages 750MB/sec and fluctuates wildly while doing so. Random buffered\nwrites to many files behave a lot better as well, as does random mmap\u0027ed\nwrites.\n\nA separate thread is added to sync the super blocks. In the long term,\nadding sync_supers_bdi() functionality could get rid of this thread again.\n\nSigned-off-by: Jens Axboe \u003cjens.axboe@oracle.com\u003e\n" }, { "commit": "03ef83af528899aa339e42d8024b37e2f434fba4", "tree": "1a323ead874d690de79cd213d5231b4fc758979b", "parents": [ "b62e408c05228f40e69bb38a48db8961cac6cd23" ], "author": { "name": "Minchan Kim", "email": "minchan.kim@gmail.com", "time": "Wed Aug 26 14:29:23 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Aug 26 20:06:52 2009 -0700" }, "message": "mm: fix for infinite churning of mlocked pages\n\nAn mlocked page might lose the isolatation race. This causes the page to\nclear PG_mlocked while it remains in a VM_LOCKED vma. This means it can\nbe put onto the [in]active list. We can rescue it by using try_to_unmap()\nin shrink_page_list().\n\nBut now, As Wu Fengguang pointed out, vmscan has a bug. If the page has\nPG_referenced, it can\u0027t reach try_to_unmap() in shrink_page_list() but is\nput into the active list. If the page is referenced repeatedly, it can\nremain on the [in]active list without being moving to the unevictable\nlist.\n\nThis patch fixes it.\n\nReported-by: Wu Fengguang \u003cfengguang.wu@intel.com\u003e\nSigned-off-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nReviewed-by: KOSAKI Motohiro \u003c\u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Lee Schermerhorn \u003clee.schermerhorn@hp.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": "8aa7e847d834ed937a9ad37a0f2ad5b8584c1ab0", "tree": "76c8b4f1362a928d426f2201790ab5d128f57724", "parents": [ "c2cc49a2f8a479dde96a599646d30b6cc9dbed78" ], "author": { "name": "Jens Axboe", "email": "jens.axboe@oracle.com", "time": "Thu Jul 09 14:52:32 2009 +0200" }, "committer": { "name": "Jens Axboe", "email": "axboe@carl.(none)", "time": "Fri Jul 10 20:31:53 2009 +0200" }, "message": "Fix congestion_wait() sync/async vs read/write confusion\n\nCommit 1faa16d22877f4839bd433547d770c676d1d964c accidentally broke\nthe bdi congestion wait queue logic, causing us to wait on congestion\nfor WRITE (\u003d\u003d 1) when we really wanted BLK_RW_ASYNC (\u003d\u003d 0) instead.\n\nSigned-off-by: Jens Axboe \u003cjens.axboe@oracle.com\u003e\n" }, { "commit": "cb4cbcf6b3cf79f80c157afdc8dd8221643d8481", "tree": "a328516967ad0daeff9702a2cf5760fb25fd26a7", "parents": [ "d888a4c76c51092993643f8992bf55b3c28da483" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Tue Jun 23 08:57:55 2009 +0900" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 23 10:17:28 2009 -0700" }, "message": "mm: fix incorrect page removal from LRU\n\nThe isolated page is \"cursor_page\" not \"page\".\n\nThis could cause LRU list corruption under memory pressure, caught by\nCONFIG_DEBUG_LIST.\n\nReported-by: Ingo Molnar \u003cmingo@elte.hu\u003e\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nReviewed-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nTested-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "2ffebca6aa7e1687905c842dd8c5c1e811e574e7", "tree": "93ec4e18c54289e82dcc8e602c166d5ffee50ef8", "parents": [ "c5b947b28828e82814605824e5db0bc58d66d8c0" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed Jun 17 16:27:21 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jun 18 13:03:48 2009 -0700" }, "message": "memcg: fix lru rotation in isolate_pages\n\nTry to fix memcg\u0027s lru rotation sanity: make memcg use the same logic as\nthe global LRU does.\n\nNow, at __isolate_lru_page() retruns -EBUSY, the page is rotated to the\ntail of LRU in global LRU\u0027s isolate LRU pages. But in memcg, it\u0027s not\nhandled. This makes memcg do the same behavior as global LRU and rotate\nLRU in the page is busy.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nAcked-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Mel Gorman \u003cmel@csn.ul.ie\u003e\nCc: 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": "ee993b135ec75a93bd5c45e636bb210d2975159b", "tree": "edc7ca495c20b0ecd9c78ce22dad121c01459511", "parents": [ "24cf72518c79cdcda486ed26074ff8151291cf65" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Tue Jun 16 15:33:24 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 19:47:46 2009 -0700" }, "message": "mm: fix lumpy reclaim lru handling at isolate_lru_pages\n\nAt lumpy reclaim, a page failed to be taken by __isolate_lru_page() can be\npushed back to \"src\" list by list_move(). But the page may not be from\n\"src\" list. This pushes the page back to wrong LRU. And list_move()\nitself is unnecessary because the page is not on top of LRU. Then, leave\nit as it is if __isolate_lru_page() fails.\n\nReviewed-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nReviewed-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nAcked-by: Mel Gorman \u003cmel@csn.ul.ie\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": "24cf72518c79cdcda486ed26074ff8151291cf65", "tree": "0db04965fb57135e68edf62c332c3a3825c1e3b5", "parents": [ "fa5e084e43eb14c14942027e1e2e894aeed96097" ], "author": { "name": "Mel Gorman", "email": "mel@csn.ul.ie", "time": "Tue Jun 16 15:33:23 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 19:47:46 2009 -0700" }, "message": "vmscan: count the number of times zone_reclaim() scans and fails\n\nOn NUMA machines, the administrator can configure zone_reclaim_mode that\nis a more targetted form of direct reclaim. On machines with large NUMA\ndistances for example, a zone_reclaim_mode defaults to 1 meaning that\nclean unmapped pages will be reclaimed if the zone watermarks are not\nbeing met.\n\nThere is a heuristic that determines if the scan is worthwhile but it is\npossible that the heuristic will fail and the CPU gets tied up scanning\nuselessly. Detecting the situation requires some guesswork and\nexperimentation so this patch adds a counter \"zreclaim_failed\" to\n/proc/vmstat. If during high CPU utilisation this counter is increasing\nrapidly, then the resolution to the problem may be to set\n/proc/sys/vm/zone_reclaim_mode to 0.\n\n[akpm@linux-foundation.org: name things consistently]\nSigned-off-by: Mel Gorman \u003cmel@csn.ul.ie\u003e\nReviewed-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: Christoph Lameter \u003ccl@linux-foundation.org\u003e\nReviewed-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Wu Fengguang \u003cfengguang.wu@intel.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": "fa5e084e43eb14c14942027e1e2e894aeed96097", "tree": "3e7ebf714858e8dd1de7042fd1ef62294a3ec20f", "parents": [ "90afa5de6f3fa89a733861e843377302479fcf7e" ], "author": { "name": "Mel Gorman", "email": "mel@csn.ul.ie", "time": "Tue Jun 16 15:33:22 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 19:47:45 2009 -0700" }, "message": "vmscan: do not unconditionally treat zones that fail zone_reclaim() as full\n\nOn NUMA machines, the administrator can configure zone_reclaim_mode that\nis a more targetted form of direct reclaim. On machines with large NUMA\ndistances for example, a zone_reclaim_mode defaults to 1 meaning that\nclean unmapped pages will be reclaimed if the zone watermarks are not\nbeing met. The problem is that zone_reclaim() failing at all means the\nzone gets marked full.\n\nThis can cause situations where a zone is usable, but is being skipped\nbecause it has been considered full. Take a situation where a large tmpfs\nmount is occuping a large percentage of memory overall. The pages do not\nget cleaned or reclaimed by zone_reclaim(), but the zone gets marked full\nand the zonelist cache considers them not worth trying in the future.\n\nThis patch makes zone_reclaim() return more fine-grained information about\nwhat occured when zone_reclaim() failued. The zone only gets marked full\nif it really is unreclaimable. If it\u0027s a case that the scan did not occur\nor if enough pages were not reclaimed with the limited reclaim_mode, then\nthe zone is simply skipped.\n\nThere is a side-effect to this patch. Currently, if zone_reclaim()\nsuccessfully reclaimed SWAP_CLUSTER_MAX, an allocation attempt would go\nahead. With this patch applied, zone watermarks are rechecked after\nzone_reclaim() does some work.\n\nThis bug was introduced by commit 9276b1bc96a132f4068fdee00983c532f43d3a26\n(\"memory page_alloc zonelist caching speedup\") way back in 2.6.19 when the\nzonelist_cache was introduced. It was not intended that zone_reclaim()\naggressively consider the zone to be full when it failed as full direct\nreclaim can still be an option. Due to the age of the bug, it should be\nconsidered a -stable candidate.\n\nSigned-off-by: Mel Gorman \u003cmel@csn.ul.ie\u003e\nReviewed-by: Wu Fengguang \u003cfengguang.wu@intel.com\u003e\nReviewed-by: Rik van Riel \u003criel@redhat.com\u003e\nReviewed-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Christoph Lameter \u003ccl@linux-foundation.org\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": "90afa5de6f3fa89a733861e843377302479fcf7e", "tree": "2870878fa3361c27551b5a18c4732073ae1432bd", "parents": [ "84a892456046921a40646114deed65e2df93a1bc" ], "author": { "name": "Mel Gorman", "email": "mel@csn.ul.ie", "time": "Tue Jun 16 15:33:20 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 19:47:45 2009 -0700" }, "message": "vmscan: properly account for the number of page cache pages zone_reclaim() can reclaim\n\nA bug was brought to my attention against a distro kernel but it affects\nmainline and I believe problems like this have been reported in various\nguises on the mailing lists although I don\u0027t have specific examples at the\nmoment.\n\nThe reported problem was that malloc() stalled for a long time (minutes in\nsome cases) if a large tmpfs mount was occupying a large percentage of\nmemory overall. The pages did not get cleaned or reclaimed by\nzone_reclaim() because the zone_reclaim_mode was unsuitable, but the lists\nare uselessly scanned frequencly making the CPU spin at near 100%.\n\nThis patchset intends to address that bug and bring the behaviour of\nzone_reclaim() more in line with expectations which were noticed during\ninvestigation. It is based on top of mmotm and takes advantage of\nKosaki\u0027s work with respect to zone_reclaim().\n\nPatch 1 fixes the heuristics that zone_reclaim() uses to determine if the\n\tscan should go ahead. The broken heuristic is what was causing the\n\tmalloc() stall as it uselessly scanned the LRU constantly. Currently,\n\tzone_reclaim is assuming zone_reclaim_mode is 1 and historically it\n\tcould not deal with tmpfs pages at all. This fixes up the heuristic so\n\tthat an unnecessary scan is more likely to be correctly avoided.\n\nPatch 2 notes that zone_reclaim() returning a failure automatically means\n\tthe zone is marked full. This is not always true. It could have\n\tfailed because the GFP mask or zone_reclaim_mode were unsuitable.\n\nPatch 3 introduces a counter zreclaim_failed that will increment each\n\ttime the zone_reclaim scan-avoidance heuristics fail. If that\n\tcounter is rapidly increasing, then zone_reclaim_mode should be\n\tset to 0 as a temporarily resolution and a bug reported because\n\tthe scan-avoidance heuristic is still broken.\n\nThis patch:\n\nOn NUMA machines, the administrator can configure zone_reclaim_mode that\nis a more targetted form of direct reclaim. On machines with large NUMA\ndistances for example, a zone_reclaim_mode defaults to 1 meaning that\nclean unmapped pages will be reclaimed if the zone watermarks are not\nbeing met.\n\nThere is a heuristic that determines if the scan is worthwhile but the\nproblem is that the heuristic is not being properly applied and is\nbasically assuming zone_reclaim_mode is 1 if it is enabled. The lack of\nproper detection can manfiest as high CPU usage as the LRU list is scanned\nuselessly.\n\nHistorically, once enabled it was depending on NR_FILE_PAGES which may\ninclude swapcache pages that the reclaim_mode cannot deal with. Patch\nvmscan-change-the-number-of-the-unmapped-files-in-zone-reclaim.patch by\nKosaki Motohiro noted that zone_page_state(zone, NR_FILE_PAGES) included\npages that were not file-backed such as swapcache and made a calculation\nbased on the inactive, active and mapped files. This is far superior when\nzone_reclaim\u003d\u003d1 but if RECLAIM_SWAP is set, then NR_FILE_PAGES is a\nreasonable starting figure.\n\nThis patch alters how zone_reclaim() works out how many pages it might be\nable to reclaim given the current reclaim_mode. If RECLAIM_SWAP is set in\nthe reclaim_mode it will either consider NR_FILE_PAGES as potential\ncandidates or else use NR_{IN}ACTIVE}_PAGES-NR_FILE_MAPPED to discount\nswapcache and other non-file-backed pages. If RECLAIM_WRITE is not set,\nthen NR_FILE_DIRTY number of pages are not candidates. If RECLAIM_SWAP is\nnot set, then NR_FILE_MAPPED are not.\n\n[kosaki.motohiro@jp.fujitsu.com: Estimate unmapped pages minus tmpfs pages]\n[fengguang.wu@intel.com: Fix underflow problem in Kosaki\u0027s estimate]\nSigned-off-by: Mel Gorman \u003cmel@csn.ul.ie\u003e\nReviewed-by: Rik van Riel \u003criel@redhat.com\u003e\nAcked-by: Christoph Lameter \u003ccl@linux-foundation.org\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Wu Fengguang \u003cfengguang.wu@intel.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": "9198e96c06744517e3b18fce8be6db61e96a3227", "tree": "3d79e58771cc27204980a0350abafaf6e7524447", "parents": [ "3eb4140f0389bdada022d5e8efd88504ad30df14" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Tue Jun 16 15:33:15 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 19:47:45 2009 -0700" }, "message": "vmscan: handle may_swap more strictly\n\nCommit 2e2e425989080cc534fc0fca154cae515f971cf5 (\"vmscan,memcg:\nreintroduce sc-\u003emay_swap) add may_swap flag and handle it at\nget_scan_ratio().\n\nBut the result of get_scan_ratio() is ignored when priority \u003d\u003d 0, so anon\nlru is scanned even if may_swap \u003d\u003d 0 or nr_swap_pages \u003d\u003d 0. IMHO, this is\nnot an expected behavior.\n\nAs for memcg especially, because of this behavior many and many pages are\nswapped-out just in vain when oom is invoked by mem+swap limit.\n\nThis patch is for handling may_swap flag more strictly.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nReviewed-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nCc: Johannes Weiner \u003channes@cmpxchg.org\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Lee Schermerhorn \u003clee.schermerhorn@hp.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "3eb4140f0389bdada022d5e8efd88504ad30df14", "tree": "6f3f229a2a3bee1bd0a5846c7e7f7987a5cc602b", "parents": [ "8cab4754d24a0f2e05920170c845bd84472814c6" ], "author": { "name": "Wu Fengguang", "email": "fengguang.wu@intel.com", "time": "Tue Jun 16 15:33:13 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 19:47:45 2009 -0700" }, "message": "vmscan: merge duplicate code in shrink_active_list()\n\nThe \"move pages to active list\" and \"move pages to inactive list\" code\nblocks are mostly identical and can be served by a function.\n\nThanks to Andrew Morton for pointing this out.\n\nNote that buffer_heads_over_limit check will also be carried out for\nre-activated pages, which is slightly different from pre-2.6.28 kernels.\nAlso, Rik\u0027s \"vmscan: evict use-once pages first\" patch could totally stop\nscans of active file list when memory pressure is low. So the net effect\ncould be, the number of buffer heads is now more likely to grow large.\n\nHowever that\u0027s fine according to Johannes\u0027 comments:\n\n I don\u0027t think that this could be harmful. We just preserve the buffer\n mappings of what we consider the working set and with low memory\n pressure, as you say, this set is not big.\n\n As to stripping of reactivated pages: the only pages we re-activate\n for now are those VM_EXEC mapped ones. Since we don\u0027t expect IO from\n or to these pages, removing the buffer mappings in case they grow too\n large should be okay, I guess.\n\nCc: Pekka Enberg \u003cpenberg@cs.helsinki.fi\u003e\nAcked-by: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nReviewed-by: Rik van Riel \u003criel@redhat.com\u003e\nReviewed-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nReviewed-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nSigned-off-by: 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": "8cab4754d24a0f2e05920170c845bd84472814c6", "tree": "99b1ec52daf8b81034f634777d2f24dd20d6d7c5", "parents": [ "6fe6b7e35785e3232ffe7f81d3893f1316710a02" ], "author": { "name": "Wu Fengguang", "email": "fengguang.wu@intel.com", "time": "Tue Jun 16 15:33:12 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 19:47:44 2009 -0700" }, "message": "vmscan: make mapped executable pages the first class citizen\n\nProtect referenced PROT_EXEC mapped pages from being deactivated.\n\nPROT_EXEC(or its internal presentation VM_EXEC) pages normally belong to some\ncurrently running executables and their linked libraries, they shall really be\ncached aggressively to provide good user experiences.\n\nThanks to Johannes Weiner for the advice to reuse the VMA walk in\npage_referenced() to get the PROT_EXEC bit.\n\n[more details]\n\n( The consequences of this patch will have to be discussed together with\n Rik van Riel\u0027s recent patch \"vmscan: evict use-once pages first\". )\n\n( Some of the good points and insights are taken into this changelog.\n Thanks to all the involved people for the great LKML discussions. )\n\nthe problem\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nFor a typical desktop, the most precious working set is composed of\n*actively accessed*\n\t(1) memory mapped executables\n\t(2) and their anonymous pages\n\t(3) and other files\n\t(4) and the dcache/icache/.. slabs\nwhile the least important data are\n\t(5) infrequently used or use-once files\n\nFor a typical desktop, one major problem is busty and large amount of (5)\nuse-once files flushing out the working set.\n\nInside the working set, (4) dcache/icache have already been too sticky ;-)\nSo we only have to care (2) anonymous and (1)(3) file pages.\n\nanonymous pages\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nAnonymous pages are effectively immune to the streaming IO attack, because we\nnow have separate file/anon LRU lists. When the use-once files crowd into the\nfile LRU, the list\u0027s \"quality\" is significantly lowered. Therefore the scan\nbalance policy in get_scan_ratio() will choose to scan the (low quality) file\nLRU much more frequently than the anon LRU.\n\nfile pages\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nRik proposed to *not* scan the active file LRU when the inactive list grows\nlarger than active list. This guarantees that when there are use-once streaming\nIO, and the working set is not too large(so that active_size \u003c inactive_size),\nthe active file LRU will *not* be scanned at all. So the not-too-large working\nset can be well protected.\n\nBut there are also situations where the file working set is a bit large so that\n(active_size \u003e\u003d inactive_size), or the streaming IOs are not purely use-once.\nIn these cases, the active list will be scanned slowly. Because the current\nshrink_active_list() policy is to deactivate active pages regardless of their\nreferenced bits. The deactivated pages become susceptible to the streaming IO\nattack: the inactive list could be scanned fast (500MB / 50MBps \u003d 10s) so that\nthe deactivated pages don\u0027t have enough time to get re-referenced. Because a\nuser tend to switch between windows in intervals from seconds to minutes.\n\nThis patch holds mapped executable pages in the active list as long as they\nare referenced during each full scan of the active list. Because the active\nlist is normally scanned much slower, they get longer grace time (eg. 100s)\nfor further references, which better matches the pace of user operations.\n\nTherefore this patch greatly prolongs the in-cache time of executable code,\nwhen there are moderate memory pressures.\n\n\tbefore patch: guaranteed to be cached if reference intervals \u003c I\n\tafter patch: guaranteed to be cached if reference intervals \u003c I+A\n\t\t (except when randomly reclaimed by the lumpy reclaim)\nwhere\n\tA \u003d time to fully scan the active file LRU\n\tI \u003d time to fully scan the inactive file LRU\n\nNote that normally A \u003e\u003e I.\n\nside effects\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nThis patch is safe in general, it restores the pre-2.6.28 mmap() behavior\nbut in a much smaller and well targeted scope.\n\nOne may worry about some one to abuse the PROT_EXEC heuristic. But as\nAndrew Morton stated, there are other tricks to getting that sort of boost.\n\nAnother concern is the PROT_EXEC mapped pages growing large in rare cases,\nand therefore hurting reclaim efficiency. But a sane application targeted for\nlarge audience will never use PROT_EXEC for data mappings. If some home made\napplication tries to abuse that bit, it shall be aware of the consequences.\nIf it is abused to scale of 2/3 total memory, it gains nothing but overheads.\n\nbenchmarks\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\n1) memory tight desktop\n\n1.1) brief summary\n\n- clock time and major faults are reduced by 50%;\n- pswpin numbers are reduced to ~1/3.\n\nThat means X desktop responsiveness is doubled under high memory/swap pressure.\n\n1.2) test scenario\n\n- nfsroot gnome desktop with 512M physical memory\n- run some programs, and switch between the existing windows\n after starting each new program.\n\n1.3) progress timing (seconds)\n\n before after programs\n 0.02 0.02 N xeyes\n 0.75 0.76 N firefox\n 2.02 1.88 N nautilus\n 3.36 3.17 N nautilus --browser\n 5.26 4.89 N gthumb\n 7.12 6.47 N gedit\n 9.22 8.16 N xpdf /usr/share/doc/shared-mime-info/shared-mime-info-spec.pdf\n 13.58 12.55 N xterm\n 15.87 14.57 N mlterm\n 18.63 17.06 N gnome-terminal\n 21.16 18.90 N urxvt\n 26.24 23.48 N gnome-system-monitor\n 28.72 26.52 N gnome-help\n 32.15 29.65 N gnome-dictionary\n 39.66 36.12 N /usr/games/sol\n 43.16 39.27 N /usr/games/gnometris\n 48.65 42.56 N /usr/games/gnect\n 53.31 47.03 N /usr/games/gtali\n 58.60 52.05 N /usr/games/iagno\n 65.77 55.42 N /usr/games/gnotravex\n 70.76 61.47 N /usr/games/mahjongg\n 76.15 67.11 N /usr/games/gnome-sudoku\n 86.32 75.15 N /usr/games/glines\n 92.21 79.70 N /usr/games/glchess\n 103.79 88.48 N /usr/games/gnomine\n 113.84 96.51 N /usr/games/gnotski\n 124.40 102.19 N /usr/games/gnibbles\n 137.41 114.93 N /usr/games/gnobots2\n 155.53 125.02 N /usr/games/blackjack\n 179.85 135.11 N /usr/games/same-gnome\n 224.49 154.50 N /usr/bin/gnome-window-properties\n 248.44 162.09 N /usr/bin/gnome-default-applications-properties\n 282.62 173.29 N /usr/bin/gnome-at-properties\n 323.72 188.21 N /usr/bin/gnome-typing-monitor\n 363.99 199.93 N /usr/bin/gnome-at-visual\n 394.21 206.95 N /usr/bin/gnome-sound-properties\n 435.14 224.49 N /usr/bin/gnome-at-mobility\n 463.05 234.11 N /usr/bin/gnome-keybinding-properties\n 503.75 248.59 N /usr/bin/gnome-about-me\n 554.00 276.27 N /usr/bin/gnome-display-properties\n 615.48 304.39 N /usr/bin/gnome-network-preferences\n 693.03 342.01 N /usr/bin/gnome-mouse-properties\n 759.90 388.58 N /usr/bin/gnome-appearance-properties\n 937.90 508.47 N /usr/bin/gnome-control-center\n 1109.75 587.57 N /usr/bin/gnome-keyboard-properties\n 1399.05 758.16 N : oocalc\n 1524.64 830.03 N : oodraw\n 1684.31 900.03 N : ooimpress\n 1874.04 993.91 N : oomath\n 2115.12 1081.89 N : ooweb\n 2369.02 1161.99 N : oowriter\n\nNote that the last \": oo*\" commands are actually commented out.\n\n1.4) vmstat numbers (some relevant ones are marked with *)\n\n before after\n nr_free_pages 1293 3898\n nr_inactive_anon 59956 53460\n nr_active_anon 26815 30026\n nr_inactive_file 2657 3218\n nr_active_file 2019 2806\n nr_unevictable 4 4\n nr_mlock 4 4\n nr_anon_pages 26706 27859\n*nr_mapped 3542 4469\n nr_file_pages 72232 67681\n nr_dirty 1 0\n nr_writeback 123 19\n nr_slab_reclaimable 3375 3534\n nr_slab_unreclaimable 11405 10665\n nr_page_table_pages 8106 7864\n nr_unstable 0 0\n nr_bounce 0 0\n*nr_vmscan_write 394776 230839\n nr_writeback_temp 0 0\n numa_hit 6843353 3318676\n numa_miss 0 0\n numa_foreign 0 0\n numa_interleave 1719 1719\n numa_local 6843353 3318676\n numa_other 0 0\n*pgpgin 5954683 2057175\n*pgpgout 1578276 922744\n*pswpin 1486615 512238\n*pswpout 394568 230685\n pgalloc_dma 277432 56602\n pgalloc_dma32 6769477 3310348\n pgalloc_normal 0 0\n pgalloc_movable 0 0\n pgfree 7048396 3371118\n pgactivate 2036343 1471492\n pgdeactivate 2189691 1612829\n pgfault 3702176 3100702\n*pgmajfault 452116 201343\n pgrefill_dma 12185 7127\n pgrefill_dma32 334384 653703\n pgrefill_normal 0 0\n pgrefill_movable 0 0\n pgsteal_dma 74214 22179\n pgsteal_dma32 3334164 1638029\n pgsteal_normal 0 0\n pgsteal_movable 0 0\n pgscan_kswapd_dma 1081421 1216199\n pgscan_kswapd_dma32 58979118 46002810\n pgscan_kswapd_normal 0 0\n pgscan_kswapd_movable 0 0\n pgscan_direct_dma 2015438 1086109\n pgscan_direct_dma32 55787823 36101597\n pgscan_direct_normal 0 0\n pgscan_direct_movable 0 0\n pginodesteal 3461 7281\n slabs_scanned 564864 527616\n kswapd_steal 2889797 1448082\n kswapd_inodesteal 14827 14835\n pageoutrun 43459 21562\n allocstall 9653 4032\n pgrotated 384216 228631\n\n1.5) free numbers at the end of the tests\n\nbefore patch:\n total used free shared buffers cached\n Mem: 474 467 7 0 0 236\n -/+ buffers/cache: 230 243\n Swap: 1023 418 605\n\nafter patch:\n total used free shared buffers cached\n Mem: 474 457 16 0 0 236\n -/+ buffers/cache: 221 253\n Swap: 1023 404 619\n\n2) memory flushing in a file server\n\n2.1) brief summary\n\nThe number of major faults from 50 to 3 during 10% cache hot reads.\n\nThat means this patch successfully stops major faults when the active file\nlist is slowly scanned when there are partially cache hot streaming IO.\n\n2.2) test scenario\n\nDo 100000 pread(size\u003d110 pages, offset\u003d(i*100) pages), where 10% of the\npages will be activated:\n\n for i in `seq 0 100 10000000`; do echo $i 110; done \u003e pattern-hot-10\n iotrace.rb --load pattern-hot-10 --play /b/sparse\n\tvmmon nr_mapped nr_active_file nr_inactive_file pgmajfault pgdeactivate pgfree\n\nand monitor /proc/vmstat during the time. The test box has 2G memory.\n\nI carried out tests on fresh booted console as well as X desktop, and\nfetched the vmstat numbers on\n\n(1) begin: shortly after the big read IO starts;\n(2) end: just before the big read IO stops;\n(3) restore: the big read IO stops and the zsh working set restored\n(4) restore X: after IO, switch back and forth between the urxvt and firefox\n windows to restore their working set.\n\n2.3) console mode results\n\n nr_mapped nr_active_file nr_inactive_file pgmajfault pgdeactivate pgfree\n\n2.6.29 VM_EXEC protection ON:\nbegin: 2481 2237 8694 630 0 574299\nend: 275 231976 233914 633 776271 20933042\nrestore: 370 232154 234524 691 777183 20958453\n\n2.6.29 VM_EXEC protection ON (second run):\nbegin: 2434 2237 8493 629 0 574195\nend: 284 231970 233536 632 771918 20896129\nrestore: 399 232218 234789 690 774526 20957909\n\n2.6.30-rc4-mm VM_EXEC protection OFF:\nbegin: 2479 2344 9659 210 0 579643\nend: 284 232010 234142 260 772776 20917184\nrestore: 379 232159 234371 301 774888 20967849\n\nThe above console numbers show that\n\n- The startup pgmajfault of 2.6.30-rc4-mm is merely 1/3 that of 2.6.29.\n I\u0027d attribute that improvement to the mmap readahead improvements :-)\n\n- The pgmajfault increment during the file copy is 633-630\u003d3 vs 260-210\u003d50.\n That\u0027s a huge improvement - which means with the VM_EXEC protection logic,\n active mmap pages is pretty safe even under partially cache hot streaming IO.\n\n- when active:inactive file lru size reaches 1:1, their scan rates is 1:20.8\n under 10% cache hot IO. (computed with formula Dpgdeactivate:Dpgfree)\n That roughly means the active mmap pages get 20.8 more chances to get\n re-referenced to stay in memory.\n\n- The absolute nr_mapped drops considerably to 1/9 during the big IO, and the\n dropped pages are mostly inactive ones. The patch has almost no impact in\n this aspect, that means it won\u0027t unnecessarily increase memory pressure.\n (In contrast, your 20% mmap protection ratio will keep them all, and\n therefore eliminate the extra 41 major faults to restore working set\n of zsh etc.)\n\nThe iotrace.rb read throughput is\n\t151.194384MB/s 284.198252s 100001x 450560b --load pattern-hot-10 --play /b/sparse\nwhich means the inactive list is rotated at the speed of 250MB/s,\nso a full scan of which takes about 3.5 seconds, while a full scan\nof active file list takes about 77 seconds.\n\n2.4) X mode results\n\nWe can reach roughly the same conclusions for X desktop:\n\n nr_mapped nr_active_file nr_inactive_file pgmajfault pgdeactivate pgfree\n\n2.6.30-rc4-mm VM_EXEC protection ON:\nbegin: 9740 8920 64075 561 0 678360\nend: 768 218254 220029 565 798953 21057006\nrestore: 857 218543 220987 606 799462 21075710\nrestore X: 2414 218560 225344 797 799462 21080795\n\n2.6.30-rc4-mm VM_EXEC protection OFF:\nbegin: 9368 5035 26389 554 0 633391\nend: 770 218449 221230 661 646472 17832500\nrestore: 1113 218466 220978 710 649881 17905235\nrestore X: 2687 218650 225484 947 802700 21083584\n\n- the absolute nr_mapped drops considerably (to 1/13 of the original size)\n during the streaming IO.\n- the delta of pgmajfault is 3 vs 107 during IO, or 236 vs 393\n during the whole process.\n\nCc: Elladan \u003celladan@eskimo.com\u003e\nCc: Nick Piggin \u003cnpiggin@suse.de\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nCc: Christoph Lameter \u003ccl@linux-foundation.org\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nAcked-by: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nAcked-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nReviewed-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nReviewed-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nSigned-off-by: 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": "6fe6b7e35785e3232ffe7f81d3893f1316710a02", "tree": "6f47c03735504d8aab8f7b048465b87cc5b15861", "parents": [ "608e8e66a154cbc3d591a59dcebfd9cbc9e3431a" ], "author": { "name": "Wu Fengguang", "email": "fengguang.wu@intel.com", "time": "Tue Jun 16 15:33:05 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 19:47:44 2009 -0700" }, "message": "vmscan: report vm_flags in page_referenced()\n\nCollect vma-\u003evm_flags of the VMAs that actually referenced the page.\n\nThis is preparing for more informed reclaim heuristics, eg. to protect\nexecutable file pages more aggressively. For now only the VM_EXEC bit\nwill be used by the caller.\n\nThanks to Johannes, Peter and Minchan for all the good tips.\n\nAcked-by: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nReviewed-by: Rik van Riel \u003criel@redhat.com\u003e\nReviewed-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nReviewed-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nSigned-off-by: 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": "cb4b86ba47bb0937b71fb825b3ed88adf7a190f0", "tree": "4b8528ba914a315e5857e7fe2a6e7d415f2e6650", "parents": [ "6837765963f1723e80ca97b1fae660f3a60d77df" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Tue Jun 16 15:32:52 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 19:47:42 2009 -0700" }, "message": "mm: add swap cache interface for swap reference\n\nIn a following patch, the usage of swap cache is recorded into swap_map.\nThis patch is for necessary interface changes to do that.\n\n2 interfaces:\n\n - swapcache_prepare()\n - swapcache_free()\n\nare added for allocating/freeing refcnt from swap-cache to existing swap\nentries. But implementation itself is not changed under this patch. At\nadding swapcache_free(), memcg\u0027s hook code is moved under\nswapcache_free(). This is better than using scattered hooks.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nReviewed-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nAcked-by: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Hugh Dickins \u003chugh.dickins@tiscali.co.uk\u003e\nCc: Johannes Weiner \u003channes@cmpxchg.org\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Dhaval Giani \u003cdhaval@linux.vnet.ibm.com\u003e\nCc: YAMAMOTO Takashi \u003cyamamoto@valinux.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": "6837765963f1723e80ca97b1fae660f3a60d77df", "tree": "a9a6ed4b7e3bf188966da78b04bf39298f24375a", "parents": [ "bce7394a3ef82b8477952fbab838e4a6e8cb47d2" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Tue Jun 16 15:32:51 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 19:47:42 2009 -0700" }, "message": "mm: remove CONFIG_UNEVICTABLE_LRU config option\n\nCurrently, nobody wants to turn UNEVICTABLE_LRU off. Thus this\nconfigurability is unnecessary.\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Johannes Weiner \u003channes@cmpxchg.org\u003e\nCc: Andi Kleen \u003candi@firstfloor.org\u003e\nAcked-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nCc: David Woodhouse \u003cdwmw2@infradead.org\u003e\nCc: Matt Mackall \u003cmpm@selenic.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Lee Schermerhorn \u003clee.schermerhorn@hp.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "69c854817566db82c362797b4a6521d0b00fe1d8", "tree": "b4873e257212659ca0eb224151971e029ce3875d", "parents": [ "35282a2de4e5e4e173ab61aa9d7015886021a821" ], "author": { "name": "MinChan Kim", "email": "minchan.kim@gmail.com", "time": "Tue Jun 16 15:32:44 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 19:47:41 2009 -0700" }, "message": "vmscan: prevent shrinking of active anon lru list in case of no swap space V3\n\nshrink_zone() can deactivate active anon pages even if we don\u0027t have a\nswap device. Many embedded products don\u0027t have a swap device. So the\ndeactivation of anon pages is unnecessary.\n\nThis patch prevents unnecessary deactivation of anon lru pages. But, it\ndon\u0027t prevent aging of anon pages to swap out.\n\nSigned-off-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nAcked-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Johannes Weiner \u003channes@cmpxchg.org\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": "af166777cf451f0373b952ce6766dc1c25385686", "tree": "2051064ffccb223e1f29b8af52ef28651eff9062", "parents": [ "08d9ae7cbbd0c5c07573d072ec771e997a9a39e0" ], "author": { "name": "Wu Fengguang", "email": "fengguang.wu@intel.com", "time": "Tue Jun 16 15:32:31 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 19:47:39 2009 -0700" }, "message": "vmscan: ZVC updates in shrink_active_list() can be done once\n\nThis effectively lifts the unit of updates to nr_inactive_* and\npgdeactivate from PAGEVEC_SIZE\u003d14 to SWAP_CLUSTER_MAX\u003d32, or\nMAX_ORDER_NR_PAGES\u003d1024 for reclaim_zone().\n\nCc: Johannes Weiner \u003channes@cmpxchg.org\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nReviewed-by: Minchan Kim \u003cminchan.kim@gmail.com\u003e\nSigned-off-by: Wu Fengguang \u003cfengguang.wu@intel.com\u003e\nCc: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nCc: Christoph Lameter \u003ccl@linux-foundation.org\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Lee Schermerhorn \u003clee.schermerhorn@hp.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "6e08a369ee10b361ac1cdcdf4fabd420fd08beb3", "tree": "9dbf870cad025b64781d9051b6680a8a23927e5a", "parents": [ "56e49d218890f49b0057710a4b6fef31f5ffbfec" ], "author": { "name": "Wu Fengguang", "email": "fengguang.wu@intel.com", "time": "Tue Jun 16 15:32:29 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 19:47:39 2009 -0700" }, "message": "vmscan: cleanup the scan batching code\n\nThe vmscan batching logic is twisting. Move it into a standalone function\nnr_scan_try_batch() and document it. No behavior change.\n\nSigned-off-by: Wu Fengguang \u003cfengguang.wu@intel.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: Nick Piggin \u003cnpiggin@suse.de\u003e\nCc: Christoph Lameter \u003ccl@linux-foundation.org\u003e\nAcked-by: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nAcked-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": "56e49d218890f49b0057710a4b6fef31f5ffbfec", "tree": "a0525dd9a140352276bdfb76a3d65230c3e5121d", "parents": [ "35efa5e993a7a00a50b87d2b7725c3eafc80b083" ], "author": { "name": "Rik van Riel", "email": "riel@redhat.com", "time": "Tue Jun 16 15:32:28 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 19:47:38 2009 -0700" }, "message": "vmscan: evict use-once pages first\n\nWhen the file LRU lists are dominated by streaming IO pages, evict those\npages first, before considering evicting other pages.\n\nThis should be safe from deadlocks or performance problems\nbecause only three things can happen to an inactive file page:\n\n1) referenced twice and promoted to the active list\n2) evicted by the pageout code\n3) under IO, after which it will get evicted or promoted\n\nThe pages freed in this way can either be reused for streaming IO, or\nallocated for something else. If the pages are used for streaming IO,\nthis pageout pattern continues. Otherwise, we will fall back to the\nnormal pageout pattern.\n\nSigned-off-by: Rik van Riel \u003criel@redhat.com\u003e\nReported-by: Elladan \u003celladan@eskimo.com\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nCc: Lee Schermerhorn \u003clee.schermerhorn@hp.com\u003e\nAcked-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "418589663d6011de9006425b6c5721e1544fb47a", "tree": "ef37fb026d3e38191d6b5c99bc95c190fa98d0fb", "parents": [ "a3af9c389a7f3e675313f442fdd8c247c1cdb66b" ], "author": { "name": "Mel Gorman", "email": "mel@csn.ul.ie", "time": "Tue Jun 16 15:32:12 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 19:47:35 2009 -0700" }, "message": "page allocator: use allocation flags as an index to the zone watermark\n\nALLOC_WMARK_MIN, ALLOC_WMARK_LOW and ALLOC_WMARK_HIGH determin whether\npages_min, pages_low or pages_high is used as the zone watermark when\nallocating the pages. Two branches in the allocator hotpath determine\nwhich watermark to use.\n\nThis patch uses the flags as an array index into a watermark array that is\nindexed with WMARK_* defines accessed via helpers. All call sites that\nuse zone-\u003epages_* are updated to use the helpers for accessing the values\nand the array offsets for setting.\n\nSigned-off-by: Mel Gorman \u003cmel@csn.ul.ie\u003e\nReviewed-by: Christoph Lameter \u003ccl@linux-foundation.org\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Pekka Enberg \u003cpenberg@cs.helsinki.fi\u003e\nCc: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nCc: Nick Piggin \u003cnickpiggin@yahoo.com.au\u003e\nCc: Dave Hansen \u003cdave@linux.vnet.ibm.com\u003e\nCc: Lee Schermerhorn \u003cLee.Schermerhorn@hp.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "78dc583d3ab43115579cb5f3f7bd12e3548dd5a5", "tree": "ef8886bd9fce4bd8e4faa30bafcacd90aee54e25", "parents": [ "d2bf6be8ab63aa84e6149aac934649aadf3828b1" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Tue Jun 16 15:31:40 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 16 19:47:31 2009 -0700" }, "message": "vmscan: low order lumpy reclaim also should use PAGEOUT_IO_SYNC\n\nCommit 33c120ed2843090e2bd316de1588b8bf8b96cbde (\"more aggressively use\nlumpy reclaim\") increased how aggressive lumpy reclaim was by isolating\nboth active and inactive pages for asynchronous lumpy reclaim on\ncostly-high-order pages and for cheap-high-order when memory pressure is\nhigh. However, if the system is under heavy pressure and there are dirty\npages, asynchronous IO may not be sufficient to reclaim a suitable page in\ntime.\n\nThis patch causes the caller to enter synchronous lumpy reclaim for\ncostly-high-order pages and for cheap-high-order pages when under memory\npressure.\n\nMinchan.kim@gmail.com said:\n\nAndy added synchronous lumpy reclaim with\nc661b078fd62abe06fd11fab4ac5e4eeafe26b6d. At that time, lumpy reclaim is\nnot agressive. His intension is just for high-order users.(above\nPAGE_ALLOC_COSTLY_ORDER).\n\nAfter some time, Rik added aggressive lumpy reclaim with\n33c120ed2843090e2bd316de1588b8bf8b96cbde. His intention was to do lumpy\nreclaim when high-order users and trouble getting a small set of\ncontiguous pages.\n\nSo we also have to add synchronous pageout for small set of contiguous\npages.\n\nCc: Lee Schermerhorn \u003cLee.Schermerhorn@hp.com\u003e\nCc: Andy Whitcroft \u003capw@shadowen.org\u003e\nAcked-by: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nReviewed-by: Minchan Kim \u003cMinchan.kim@gmail.com\u003e\nReviewed-by: Mel Gorman \u003cmel@csn.ul.ie\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "c6f37f12197ac3bd2e5a35f2f0e195ae63d437de", "tree": "e9609ca63d211911582afcc4b8314887b85a6c46", "parents": [ "e240b58c79144708530138e05f17c6d0d8d744a8" ], "author": { "name": "Rafael J. Wysocki", "email": "rjw@sisk.pl", "time": "Sun May 24 22:16:31 2009 +0200" }, "committer": { "name": "Rafael J. Wysocki", "email": "rjw@sisk.pl", "time": "Fri Jun 12 21:32:32 2009 +0200" }, "message": "PM/Suspend: Do not shrink memory before suspend\n\nRemove the shrinking of memory from the suspend-to-RAM code, where\nit is not really necessary.\n\nSigned-off-by: Rafael J. Wysocki \u003crjw@sisk.pl\u003e\nAcked-by: Nigel Cunningham \u003cnigel@tuxonice.net\u003e\nAcked-by: Wu Fengguang \u003cfengguang.wu@intel.com\u003e\n" }, { "commit": "e767e0561d7fd2333df1921f1ab4176211f9036b", "tree": "3b936733f80ceb1ee61ce99f927d002d2296250e", "parents": [ "bd6daba909d8484bd2ccf6017db4028d7a420927" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Thu May 28 14:34:28 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri May 29 08:40:02 2009 -0700" }, "message": "memcg: fix deadlock between lock_page_cgroup and mapping tree_lock\n\nmapping-\u003etree_lock can be acquired from interrupt context. Then,\nfollowing dead lock can occur.\n\nAssume \"A\" as a page.\n\n CPU0:\n lock_page_cgroup(A)\n\t\tinterrupted\n\t\t\t-\u003e take mapping-\u003etree_lock.\n CPU1:\n take mapping-\u003etree_lock\n\t\t-\u003e lock_page_cgroup(A)\n\nThis patch tries to fix above deadlock by moving memcg\u0027s hook to out of\nmapping-\u003etree_lock. charge/uncharge of pagecache/swapcache is protected\nby page lock, not tree_lock.\n\nAfter this patch, lock_page_cgroup() is not called under mapping-\u003etree_lock.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-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\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "8713e01295140f674a41f2199b0f7ca99dfb69d5", "tree": "0e222bd3469534f040c087c334b68e869e2dbb5f", "parents": [ "0ae05fb254a5f2617fc8fcfff7be959b54a5e963" ], "author": { "name": "Andrew Morton", "email": "akpm@linux-foundation.org", "time": "Thu Apr 30 15:08:55 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat May 02 15:36:10 2009 -0700" }, "message": "vmscan: avoid multiplication overflow in shrink_zone()\n\nLocal variable `scan\u0027 can overflow on zones which are larger than\n\n\t(2G * 4k) / 100 \u003d 80GB.\n\nMaking it 64-bit on 64-bit will fix that up.\n\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Wu Fengguang \u003cfengguang.wu@intel.com\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Lee Schermerhorn \u003clee.schermerhorn@hp.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "2e2e425989080cc534fc0fca154cae515f971cf5", "tree": "c4acbd403164d1d210e3bc274b7d93487071153e", "parents": [ "55e5750b3e979bac853c0809ad0ef75b7cebd18c" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Tue Apr 21 12:24:57 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Apr 21 13:41:51 2009 -0700" }, "message": "vmscan,memcg: reintroduce sc-\u003emay_swap\n\nCommit a6dc60f8975ad96d162915e07703a4439c80dcf0 (\"vmscan: rename\nsc.may_swap to may_unmap\") removed the may_swap flag, but memcg had used\nit as a flag for \"we need to use swap?\", as the name indicate.\n\nAnd in the current implementation, memcg cannot reclaim mapped file\ncaches when mem+swap hits the limit.\n\nre-introduce may_swap flag and handle it at get_scan_ratio(). This\npatch doesn\u0027t influence any scan_control users other than memcg.\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nAcked-by: Johannes Weiner \u003channes@cmpxchg.org\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": "a21e25536169432cf9174d631972bc1cd4c75062", "tree": "5414335eae4b34be3e0cc0374b81d8303c6a95af", "parents": [ "aefe6475720bd5eb8aacbc881488f3aa65618562" ], "author": { "name": "Rafael J. Wysocki", "email": "rjw@sisk.pl", "time": "Sat Apr 18 17:23:41 2009 +0200" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Apr 18 11:36:58 2009 -0700" }, "message": "PM/Hibernate: Fix memory shrinking\n\nCommit d979677c4c0 (\"mm: shrink_all_memory(): use sc.nr_reclaimed\")\nbroke the memory shrinking used by hibernation, becuse it did not update\nshrink_all_zones() in accordance with the other changes it made.\n\nFix this by making shrink_all_zones() update sc-\u003enr_reclaimed instead of\noverwriting its value.\n\nThis fixes http://bugzilla.kernel.org/show_bug.cgi?id\u003d13058\n\nReported-and-tested-by: Alan Jenkins \u003calan-jenkins@tuffmail.co.uk\u003e\nSigned-off-by: Rafael J. Wysocki \u003crjw@sisk.pl\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "90975ef71246c5c688ead04e8ff6f36dc92d28b3", "tree": "eda44b2efe91509719b0e62219c2efec13a9e762", "parents": [ "cab4e4c43f92582a2bfc026137b3d8a175bd0360", "558f6ab9106e6be701acb0257e7171df1bbccf04" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Apr 05 10:33:07 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Apr 05 10:33:07 2009 -0700" }, "message": "Merge git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux-2.6-cpumask\n\n* git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux-2.6-cpumask: (36 commits)\n cpumask: remove cpumask allocation from idle_balance, fix\n numa, cpumask: move numa_node_id default implementation to topology.h, fix\n cpumask: remove cpumask allocation from idle_balance\n x86: cpumask: x86 mmio-mod.c use cpumask_var_t for downed_cpus\n x86: cpumask: update 32-bit APM not to mug current-\u003ecpus_allowed\n x86: microcode: cleanup\n x86: cpumask: use work_on_cpu in arch/x86/kernel/microcode_core.c\n cpumask: fix CONFIG_CPUMASK_OFFSTACK\u003dy cpu hotunplug crash\n numa, cpumask: move numa_node_id default implementation to topology.h\n cpumask: convert node_to_cpumask_map[] to cpumask_var_t\n cpumask: remove x86 cpumask_t uses.\n cpumask: use cpumask_var_t in uv_flush_tlb_others.\n cpumask: remove cpumask_t assignment from vector_allocation_domain()\n cpumask: make Xen use the new operators.\n cpumask: clean up summit\u0027s send_IPI functions\n cpumask: use new cpumask functions throughout x86\n x86: unify cpu_callin_mask/cpu_callout_mask/cpu_initialized_mask/cpu_sibling_setup_mask\n cpumask: convert struct cpuinfo_x86\u0027s llc_shared_map to cpumask_var_t\n cpumask: convert node_to_cpumask_map[] to cpumask_var_t\n x86: unify 32 and 64-bit node_to_cpumask_map\n ...\n" }, { "commit": "266cf658efcf6ac33541a46740f74f50c79d2b6b", "tree": "5c83b0879892d509e598dfd54be3ba3679ecd348", "parents": [ "03fb3d2af96c2783c3a5bc03f3d984cf422f0e69" ], "author": { "name": "David Howells", "email": "dhowells@redhat.com", "time": "Fri Apr 03 16:42:36 2009 +0100" }, "committer": { "name": "David Howells", "email": "dhowells@redhat.com", "time": "Fri Apr 03 16:42:36 2009 +0100" }, "message": "FS-Cache: Recruit a page flags for cache management\n\nRecruit a page flag to aid in cache management. The following extra flag is\ndefined:\n\n (1) PG_fscache (PG_private_2)\n\n The marked page is backed by a local cache and is pinning resources in the\n cache driver.\n\nIf PG_fscache is set, then things that checked for PG_private will now also\ncheck for that. This includes things like truncation and page invalidation.\nThe function page_has_private() had been added to make the checks for both\nPG_private and PG_private_2 at the same time.\n\nSigned-off-by: David Howells \u003cdhowells@redhat.com\u003e\nAcked-by: Steve Dickson \u003csteved@redhat.com\u003e\nAcked-by: Trond Myklebust \u003cTrond.Myklebust@netapp.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nAcked-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nTested-by: Daire Byrne \u003cDaire.Byrne@framestore.com\u003e\n" }, { "commit": "327c0e968645f2601a43f5ea7c19c7b3a5fa0a34", "tree": "acc6789c120a6d5000ceebf51e690d14c6cfcacb", "parents": [ "2678958e1225f350806d90f211a3b475f64aee80" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Tue Mar 31 15:23:31 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Apr 01 08:59:15 2009 -0700" }, "message": "vmscan: fix it to take care of nodemask\n\ntry_to_free_pages() is used for the direct reclaim of up to\nSWAP_CLUSTER_MAX pages when watermarks are low. The caller to\nalloc_pages_nodemask() can specify a nodemask of nodes that are allowed to\nbe used but this is not passed to try_to_free_pages(). This can lead to\nunnecessary reclaim of pages that are unusable by the caller and int the\nworst case lead to allocation failure as progress was not been make where\nit is needed.\n\nThis patch passes the nodemask used for alloc_pages_nodemask() to\ntry_to_free_pages().\n\nReviewed-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nAcked-by: Mel Gorman \u003cmel@csn.ul.ie\u003e\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: 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": "88c3bd707c2552bcef93cc3724647903aece159d", "tree": "68c104f559854b1f5fea01ed8568584e535ac597", "parents": [ "71aa653c6bfa6743d838342105ebc067145394e4" ], "author": { "name": "David Rientjes", "email": "rientjes@google.com", "time": "Tue Mar 31 15:23:29 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Apr 01 08:59:15 2009 -0700" }, "message": "vmscan: print shrink_slab symbol name on negative shrinker objects\n\nWhen a shrinker has a negative number of objects to delete, the symbol\nname of the shrinker should be printed, not shrink_slab. This also makes\nthe error message slightly more informative.\n\nCc: Ingo Molnar \u003cmingo@elte.hu\u003e\nSigned-off-by: 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": "ad1c3544d0a85da7738ce8cff6f8a148da57935c", "tree": "f1ac66e22e172945f200f1b2fd9a2ece23a0880b", "parents": [ "2443462b0a04ef0f82ad48f4fd0ef4ac5b24c4b7" ], "author": { "name": "Johannes Weiner", "email": "hannes@cmpxchg.org", "time": "Tue Mar 31 15:23:13 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Apr 01 08:59:13 2009 -0700" }, "message": "mm: don\u0027t free swap slots on page deactivation\n\nThe pagevec_swap_free() at the end of shrink_active_list() was introduced\nin 68a22394 \"vmscan: free swap space on swap-in/activation\" when\nshrink_active_list() was still rotating referenced active pages.\n\nIn 7e9cd48 \"vmscan: fix pagecache reclaim referenced bit check\" this was\nchanged, the rotating removed but the pagevec_swap_free() after the\nrotation loop was forgotten, applying now to the pagevec of the\ndeactivation loop instead.\n\nNow swap space is freed for deactivated pages. And only for those that\nhappen to be on the pagevec after the deactivation loop.\n\nComplete 7e9cd48 and remove the rest of the swap freeing.\n\nSigned-off-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: Hugh Dickins \u003chugh@veritas.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "2443462b0a04ef0f82ad48f4fd0ef4ac5b24c4b7", "tree": "3be4cdc338bf004f7b8c6c70baf7fb78ad252844", "parents": [ "2584e517320bd48dc8d20e38a2621a2dbe58fade" ], "author": { "name": "Johannes Weiner", "email": "hannes@cmpxchg.org", "time": "Tue Mar 31 15:23:12 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Apr 01 08:59:13 2009 -0700" }, "message": "mm: move pagevec stripping to save unlock-relock\n\nIn shrink_active_list() after the deactivation loop, we strip buffer heads\nfrom the potentially remaining pages in the pagevec.\n\nCurrently, this drops the zone\u0027s lru lock for stripping, only to reacquire\nit again afterwards to update statistics.\n\nIt is not necessary to strip the pages before updating the stats, so move\nthe whole thing out of the protected region and save the extra locking.\n\nSigned-off-by: Johannes Weiner \u003channes@cmpxchg.org\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": "bd2f6199cf9af472aeefa1b642c9f504f19e6008", "tree": "ba8cd5e7d940bcac161e28872998f86a2910362b", "parents": [ "bd775c42ea5f7c766d03a287083837cf05e7e738" ], "author": { "name": "Johannes Weiner", "email": "hannes@cmpxchg.org", "time": "Tue Mar 31 15:19:38 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Apr 01 08:59:12 2009 -0700" }, "message": "vmscan: respect higher order in zone_reclaim()\n\nDuring page allocation, there are two stages of direct reclaim that are\napplied to each zone in the preferred list. The first stage using\nzone_reclaim() reclaims unmapped file backed pages and slab pages if over\ndefined limits as these are cheaper to reclaim. The caller specifies the\norder of the target allocation but the scan control is not being correctly\ninitialised.\n\nThe impact is that the correct number of pages are being reclaimed but\nthat lumpy reclaim is not being applied. This increases the chances of a\nfull direct reclaim via try_to_free_pages() is required.\n\nThis patch initialises the order field of the scan control as requested by\nthe caller.\n\n[mel@csn.ul.ie: rewrote changelog]\nSigned-off-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nAcked-by: Mel Gorman \u003cmel@csn.ul.ie\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Andy Whitcroft \u003capw@shadowen.org\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.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": "9786bf841da57fac3457a1dac41acb4c1f2eced6", "tree": "763cb1dd3f6a3c4faf1c31b44ad2e7bef0d206d2", "parents": [ "d979677c4c02f0a72db5a03ecd8184bd9d6695c8" ], "author": { "name": "Johannes Weiner", "email": "hannes@cmpxchg.org", "time": "Tue Mar 31 15:19:35 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Apr 01 08:59:12 2009 -0700" }, "message": "vmscan: clip swap_cluster_max in shrink_all_memory()\n\nshrink_inactive_list() scans in sc-\u003eswap_cluster_max chunks until it hits\nthe scan limit it was passed.\n\nshrink_inactive_list()\n{\n\tdo {\n\t\tisolate_pages(swap_cluster_max)\n\t\tshrink_page_list()\n\t} while (nr_scanned \u003c max_scan);\n}\n\nThis assumes that swap_cluster_max is not bigger than the scan limit\nbecause the latter is checked only after at least one iteration.\n\nIn shrink_all_memory() sc-\u003eswap_cluster_max is initialized to the overall\nreclaim goal in the beginning but not decreased while reclaim is making\nprogress which leads to subsequent calls to shrink_inactive_list()\nreclaiming way too much in the one iteration that is done unconditionally.\n\nSet sc-\u003eswap_cluster_max always to the proper goal before doing\n shrink_all_zones()\n shrink_list()\n shrink_inactive_list().\n\nWhile the current shrink_all_memory() happily reclaims more than actually\nrequested, this patch fixes it to never exceed the goal:\n\nunpatched\n wanted\u003d10000 reclaimed\u003d13356\n wanted\u003d10000 reclaimed\u003d19711\n wanted\u003d10000 reclaimed\u003d10289\n wanted\u003d10000 reclaimed\u003d17306\n wanted\u003d10000 reclaimed\u003d10700\n wanted\u003d10000 reclaimed\u003d10004\n wanted\u003d10000 reclaimed\u003d13301\n wanted\u003d10000 reclaimed\u003d10976\n wanted\u003d10000 reclaimed\u003d10605\n wanted\u003d10000 reclaimed\u003d10088\n wanted\u003d10000 reclaimed\u003d15000\n\npatched\n wanted\u003d10000 reclaimed\u003d10000\n wanted\u003d10000 reclaimed\u003d9599\n wanted\u003d10000 reclaimed\u003d8476\n wanted\u003d10000 reclaimed\u003d8326\n wanted\u003d10000 reclaimed\u003d10000\n wanted\u003d10000 reclaimed\u003d10000\n wanted\u003d10000 reclaimed\u003d9919\n wanted\u003d10000 reclaimed\u003d10000\n wanted\u003d10000 reclaimed\u003d10000\n wanted\u003d10000 reclaimed\u003d10000\n wanted\u003d10000 reclaimed\u003d10000\n wanted\u003d10000 reclaimed\u003d9624\n wanted\u003d10000 reclaimed\u003d10000\n wanted\u003d10000 reclaimed\u003d10000\n wanted\u003d8500 reclaimed\u003d8092\n wanted\u003d316 reclaimed\u003d316\n\nSigned-off-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nReviewed-by: MinChan Kim \u003cminchan.kim@gmail.com\u003e\nAcked-by: Nigel Cunningham \u003cncunningham@crca.org.au\u003e\nAcked-by: \"Rafael J. Wysocki\" \u003crjw@sisk.pl\u003e\nReviewed-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: 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": "d979677c4c02f0a72db5a03ecd8184bd9d6695c8", "tree": "370c936bb17210db0db80df0e9b488c6ee5c11a1", "parents": [ "0a0dd05dd7e1a800241888cbf515bf8d3dc2e59c" ], "author": { "name": "MinChan Kim", "email": "minchan.kim@gmail.com", "time": "Tue Mar 31 15:19:34 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Apr 01 08:59:12 2009 -0700" }, "message": "mm: shrink_all_memory(): use sc.nr_reclaimed\n\nCommit a79311c14eae4bb946a97af25f3e1b17d625985d \"vmscan: bail out of\ndirect reclaim after swap_cluster_max pages\" moved the nr_reclaimed\ncounter into the scan control to accumulate the number of all reclaimed\npages in a reclaim invocation.\n\nshrink_all_memory() can use the same mechanism. it increase code\nconsistency and redability.\n\n[akpm@linux-foundation.org: coding-style fixes]\nSigned-off-by: MinChan Kim \u003cminchan.kim@gmail.com\u003e\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nSigned-off-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nCc: \"Rafael J. Wysocki\" \u003crjw@sisk.pl\u003e\nCc: 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": "ee99c71c59f897436ec65debb99372b3146f9985", "tree": "051f1c43b7c7658689d4b2c23b3d8585d6464a89", "parents": [ "a6dc60f8975ad96d162915e07703a4439c80dcf0" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Tue Mar 31 15:19:31 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Apr 01 08:59:11 2009 -0700" }, "message": "mm: introduce for_each_populated_zone() macro\n\nImpact: cleanup\n\nIn almost cases, for_each_zone() is used with populated_zone(). It\u0027s\nbecause almost function doesn\u0027t need memoryless node information.\nTherefore, for_each_populated_zone() can help to make code simplify.\n\nThis patch has no functional change.\n\n[akpm@linux-foundation.org: small cleanup]\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Mel Gorman \u003cmel@csn.ul.ie\u003e\nReviewed-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "a6dc60f8975ad96d162915e07703a4439c80dcf0", "tree": "b28cae1e69e85e87d6991dbbb199bfe73fb433af", "parents": [ "9de1581e75ba9d7979766d4ab6d56f0f2d87f7c6" ], "author": { "name": "Johannes Weiner", "email": "hannes@cmpxchg.org", "time": "Tue Mar 31 15:19:30 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Apr 01 08:59:11 2009 -0700" }, "message": "vmscan: rename sc.may_swap to may_unmap\n\nsc.may_swap does not only influence reclaiming of anon pages but pages\nmapped into pagetables in general, which also includes mapped file pages.\n\nIn shrink_page_list():\n\n\t\tif (!sc-\u003emay_swap \u0026\u0026 page_mapped(page))\n\t\t\tgoto keep_locked;\n\nFor anon pages, this makes sense as they are always mapped and reclaiming\nthem always requires swapping.\n\nBut mapped file pages are skipped here as well and it has nothing to do\nwith swapping.\n\nThe real effect of the knob is whether mapped pages are unmapped and\nreclaimed or not. Rename it to `may_unmap\u0027 to have its name match its\nactual meaning more precisely.\n\nSigned-off-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nReviewed-by: MinChan Kim \u003cminchan.kim@gmail.com\u003e\nReviewed-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Lee Schermerhorn \u003clee.schermerhorn@hp.com\u003e\nCc: 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": "558f6ab9106e6be701acb0257e7171df1bbccf04", "tree": "6e811633baeb676693c493f6c82bf785cab2771d", "parents": [ "15f7176eb1cccec0a332541285ee752b935c1c85", "65fb0d23fcddd8697c871047b700c78817bdaa43" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Tue Mar 31 13:33:50 2009 +1030" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Tue Mar 31 13:33:50 2009 +1030" }, "message": "Merge branch \u0027cpumask-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip\n\nConflicts:\n\n\tarch/x86/include/asm/topology.h\n\tdrivers/oprofile/buffer_sync.c\n(Both cases: changed in Linus\u0027 tree, removed in Ingo\u0027s).\n" }, { "commit": "c4e1aa67ed9e4e542a064bc271ddbf152b677e91", "tree": "2a2ca00bed0fc22b4eb83db092c9178868d8f76b", "parents": [ "cf2f7d7c90279cdbc12429de278f3d27ac2050ae", "2f8501815256af8498904e68bd0984b1afffd6f8" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Mar 30 17:17:35 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Mar 30 17:17:35 2009 -0700" }, "message": "Merge branch \u0027locking-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip\n\n* \u0027locking-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (33 commits)\n lockdep: fix deadlock in lockdep_trace_alloc\n lockdep: annotate reclaim context (__GFP_NOFS), fix SLOB\n lockdep: annotate reclaim context (__GFP_NOFS), fix\n lockdep: build fix for !PROVE_LOCKING\n lockstat: warn about disabled lock debugging\n lockdep: use stringify.h\n lockdep: simplify check_prev_add_irq()\n lockdep: get_user_chars() redo\n lockdep: simplify get_user_chars()\n lockdep: add comments to mark_lock_irq()\n lockdep: remove macro usage from mark_held_locks()\n lockdep: fully reduce mark_lock_irq()\n lockdep: merge the !_READ mark_lock_irq() helpers\n lockdep: merge the _READ mark_lock_irq() helpers\n lockdep: simplify mark_lock_irq() helpers #3\n lockdep: further simplify mark_lock_irq() helpers\n lockdep: simplify the mark_lock_irq() helpers\n lockdep: split up mark_lock_irq()\n lockdep: generate usage strings\n lockdep: generate the state bit definitions\n ...\n" }, { "commit": "65fb0d23fcddd8697c871047b700c78817bdaa43", "tree": "119e6e5f276622c4c862f6c9b6d795264ba1603a", "parents": [ "8c083f081d0014057901c68a0a3e0f8ca7ac8d23", "dfbbe89e197a77f2c8046a51c74e33e35f878080" ], "author": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon Mar 30 23:53:32 2009 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon Mar 30 23:53:32 2009 +0200" }, "message": "Merge branch \u0027linus\u0027 into cpumask-for-linus\n\nConflicts:\n\tarch/x86/kernel/cpu/common.c\n" }, { "commit": "1d885526f2f3fffacee2ecb541270bd00168adff", "tree": "0f87c5d264361fd75c628b4485ef5bf91eb0c95d", "parents": [ "84814d642a4f1f294bd675ab11aae1ca54c6cedb" ], "author": { "name": "Daisuke Nishimura", "email": "nishimura@mxp.nes.nec.co.jp", "time": "Fri Mar 13 13:52:00 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Mar 14 11:57:22 2009 -0700" }, "message": "vmscan: pgmoved should be cleared after updating recent_rotated\n\npgmoved should be cleared after updating recent_rotated.\n\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Lee Schermerhorn \u003cLee.Schermerhorn@hp.com\u003e\nAcked-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": "238a5b4bff2e0929a9ceb41f340137f36a6a75d9", "tree": "825d444ffd70579ab3e9a84f829d0cc6091295a8", "parents": [ "17d85bc7564571a1cce23ffdb2d2a33301876925", "73e907de7d5cecef43d9949ab8f4fdca508168c7" ], "author": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Fri Mar 13 05:54:55 2009 +0100" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Fri Mar 13 05:54:55 2009 +0100" }, "message": "Merge branch \u0027cpus4096\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux-2.6-x86 into cpus4096\n" }, { "commit": "a70f730282019f487aa33a84e5ac9a5e89c5abd0", "tree": "e6891ec5db5383c6f39617d0cc9671e1a0d1a988", "parents": [ "c69fc56de1df5769f2ec69c915c7ad5afe63804c" ], "author": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Mar 13 14:49:46 2009 +1030" }, "committer": { "name": "Rusty Russell", "email": "rusty@rustcorp.com.au", "time": "Fri Mar 13 14:49:46 2009 +1030" }, "message": "cpumask: replace node_to_cpumask with cpumask_of_node.\n\nImpact: cleanup\n\nnode_to_cpumask (and the blecherous node_to_cpumask_ptr which\ncontained a declaration) are replaced now everyone implements\ncpumask_of_node.\n\nSigned-off-by: Rusty Russell \u003crusty@rustcorp.com.au\u003e\n" }, { "commit": "f272b7bc447553410dde691aa31fc531adf9c175", "tree": "95bd28ba8cea8471c0f461cfe6a9f2f0c2f00dff", "parents": [ "02d46e07e538c285accb5c000a7db3a97eff1fbf" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Thu Mar 12 14:31:36 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Mar 12 16:20:24 2009 -0700" }, "message": "memcg: use correct scan number at reclaim\n\nEven when page reclaim is under mem_cgroup, # of scan page is determined by\nstatus of global LRU. Fix that.\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\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": "0cb57258fe01e9b21076b6a15b6aec7a24168228", "tree": "9744786d92356a21d3d2bcd2d3cc2935d81a489f", "parents": [ "3049103ddfc9aac111916bd2f39ac6976c431517" ], "author": { "name": "Johannes Weiner", "email": "hannes@cmpxchg.org", "time": "Sat Feb 14 02:04:10 2009 +0100" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 21 14:17:17 2009 -0800" }, "message": "swsusp: clean up shrink_all_zones()\n\nMove local variables to innermost possible scopes and use local\nvariables to cache calculations/reads done more than once.\n\nNo change in functionality (intended).\n\nSigned-off-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nReviewed-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nSigned-off-by: Rafael J. Wysocki \u003crjw@sisk.pl\u003e\nCc: Len Brown \u003clenb@kernel.org\u003e\nCc: Greg KH \u003cgregkh@suse.de\u003e\nAcked-by: Pavel Machek \u003cpavel@ucw.cz\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "3049103ddfc9aac111916bd2f39ac6976c431517", "tree": "709bfc2b40389cac7643e27be5307c81a35c6c4b", "parents": [ "09664fda48c5dd63277f1f42888ca9d5dca6037a" ], "author": { "name": "Johannes Weiner", "email": "hannes@cmpxchg.org", "time": "Sat Feb 14 02:03:08 2009 +0100" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Feb 21 14:17:17 2009 -0800" }, "message": "swsusp: dont fiddle with swappiness\n\nsc.swappiness is not used in the swsusp memory shrinking path, do not\nset it.\n\nSigned-off-by: Johannes Weiner \u003channes@cmpxchg.org\u003e\nReviewed-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nSigned-off-by: Rafael J. Wysocki \u003crjw@sisk.pl\u003e\nCc: Len Brown \u003clenb@kernel.org\u003e\nCc: Greg KH \u003cgregkh@suse.de\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "6700ec65c207068a81a535e9dca616fefac21671", "tree": "e7acfee960655670d8ce0e336f656f9325e823f0", "parents": [ "868a23a8043f2a3042dae60105c89bd4680187ba" ], "author": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Sun Feb 15 21:18:17 2009 +0100" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Sun Feb 15 21:22:48 2009 +0100" }, "message": "lockdep: annotate reclaim context (__GFP_NOFS), fix\n\nImpact: fix build warning\n\nFix:\n\n mm/vmscan.c: In function ‘kswapd’:\n mm/vmscan.c:1969: warning: ISO C90 forbids mixed declarations and code\n\nnode_to_cpumask_ptr(cpumask, pgdat-\u003enode_id), has a side-effect: it\ndefines the \u0027cpumask\u0027 local variable as well, so it has to go into\nthe variable definition section.\n\nSidenote: it might make sense to make this purpose of these macros\nmore apparent, by naming them the standard way, such as:\n\n DEFINE_node_to_cpumask_ptr(cpumask, pgdat-\u003enode_id);\n\n(But that is outside the scope of this patch.)\n\nCc: Rusty Russell \u003crusty@rustcorp.com.au\u003e\nCc: Mike Travis \u003ctravis@sgi.com\u003e\nCc: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nCc: Nick Piggin \u003cnpiggin@suse.de\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "cf40bd16fdad42c053040bcd3988f5fdedbb6c57", "tree": "d97ab25726981712be806d77650b7f65167cfc88", "parents": [ "6f2b9b9a9d750a9175dc79c74bfed5add840983c" ], "author": { "name": "Nick Piggin", "email": "npiggin@suse.de", "time": "Wed Jan 21 08:12:39 2009 +0100" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Sat Feb 14 23:27:49 2009 +0100" }, "message": "lockdep: annotate reclaim context (__GFP_NOFS)\n\nHere is another version, with the incremental patch rolled up, and\nadded reclaim context annotation to kswapd, and allocation tracing\nto slab allocators (which may only ever reach the page allocator\nin rare cases, so it is good to put annotations here too).\n\nHaven\u0027t tested this version as such, but it should be getting closer\nto merge worthy ;)\n\n--\nAfter noticing some code in mm/filemap.c accidentally perform a __GFP_FS\nallocation when it should not have been, I thought it might be a good idea to\ntry to catch this kind of thing with lockdep.\n\nI coded up a little idea that seems to work. Unfortunately the system has to\nactually be in __GFP_FS page reclaim, then take the lock, before it will mark\nit. But at least that might still be some orders of magnitude more common\n(and more debuggable) than an actual deadlock condition, so we have some\nimprovement I hope (the concept is no less complete than discovery of a lock\u0027s\ninterrupt contexts).\n\nI guess we could even do the same thing with __GFP_IO (normal reclaim), and\neven GFP_NOIO locks too... but filesystems will have the most locks and fiddly\ncode paths, so let\u0027s start there and see how it goes.\n\nIt *seems* to work. I did a quick test.\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n[ INFO: inconsistent lock state ]\n2.6.28-rc6-00007-ged31348-dirty #26\n---------------------------------\ninconsistent {in-reclaim-W} -\u003e {ov-reclaim-W} usage.\nmodprobe/8526 [HC0[0]:SC0[0]:HE1:SE1] takes:\n (testlock){--..}, at: [\u003cffffffffa0020055\u003e] brd_init+0x55/0x216 [brd]\n{in-reclaim-W} state was registered at:\n [\u003cffffffff80267bdb\u003e] __lock_acquire+0x75b/0x1a60\n [\u003cffffffff80268f71\u003e] lock_acquire+0x91/0xc0\n [\u003cffffffff8070f0e1\u003e] mutex_lock_nested+0xb1/0x310\n [\u003cffffffffa002002b\u003e] brd_init+0x2b/0x216 [brd]\n [\u003cffffffff8020903b\u003e] _stext+0x3b/0x170\n [\u003cffffffff80272ebf\u003e] sys_init_module+0xaf/0x1e0\n [\u003cffffffff8020c3fb\u003e] system_call_fastpath+0x16/0x1b\n [\u003cffffffffffffffff\u003e] 0xffffffffffffffff\nirq event stamp: 3929\nhardirqs last enabled at (3929): [\u003cffffffff8070f2b5\u003e] mutex_lock_nested+0x285/0x310\nhardirqs last disabled at (3928): [\u003cffffffff8070f089\u003e] mutex_lock_nested+0x59/0x310\nsoftirqs last enabled at (3732): [\u003cffffffff8061f623\u003e] sk_filter+0x83/0xe0\nsoftirqs last disabled at (3730): [\u003cffffffff8061f5b6\u003e] sk_filter+0x16/0xe0\n\nother info that might help us debug this:\n1 lock held by modprobe/8526:\n #0: (testlock){--..}, at: [\u003cffffffffa0020055\u003e] brd_init+0x55/0x216 [brd]\n\nstack backtrace:\nPid: 8526, comm: modprobe Not tainted 2.6.28-rc6-00007-ged31348-dirty #26\nCall Trace:\n [\u003cffffffff80265483\u003e] print_usage_bug+0x193/0x1d0\n [\u003cffffffff80266530\u003e] mark_lock+0xaf0/0xca0\n [\u003cffffffff80266735\u003e] mark_held_locks+0x55/0xc0\n [\u003cffffffffa0020000\u003e] ? brd_init+0x0/0x216 [brd]\n [\u003cffffffff802667ca\u003e] trace_reclaim_fs+0x2a/0x60\n [\u003cffffffff80285005\u003e] __alloc_pages_internal+0x475/0x580\n [\u003cffffffff8070f29e\u003e] ? mutex_lock_nested+0x26e/0x310\n [\u003cffffffffa0020000\u003e] ? brd_init+0x0/0x216 [brd]\n [\u003cffffffffa002006a\u003e] brd_init+0x6a/0x216 [brd]\n [\u003cffffffffa0020000\u003e] ? brd_init+0x0/0x216 [brd]\n [\u003cffffffff8020903b\u003e] _stext+0x3b/0x170\n [\u003cffffffff8070f8b9\u003e] ? mutex_unlock+0x9/0x10\n [\u003cffffffff8070f83d\u003e] ? __mutex_unlock_slowpath+0x10d/0x180\n [\u003cffffffff802669ec\u003e] ? trace_hardirqs_on_caller+0x12c/0x190\n [\u003cffffffff80272ebf\u003e] sys_init_module+0xaf/0x1e0\n [\u003cffffffff8020c3fb\u003e] system_call_fastpath+0x16/0x1b\n\nSigned-off-by: Nick Piggin \u003cnpiggin@suse.de\u003e\nSigned-off-by: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "c772be939e078afd2505ede7d596a30f8f61de95", "tree": "2cf3157fdd2b3570e105f2d91576585750526fb3", "parents": [ "a7885eb8ad465ec9db99ac5b5e6680f0ca8e11c8" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Wed Jan 07 18:08:25 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:09 2009 -0800" }, "message": "memcg: fix calculation of active_ratio\n\nCurrently, inactive_ratio of memcg is calculated at setting limit.\nbecause page_alloc.c does so and current implementation is straightforward\nporting.\n\nHowever, memcg introduced hierarchy feature recently. In hierarchy\nrestriction, memory limit is not only decided memory.limit_in_bytes of\ncurrent cgroup, but also parent limit and sibling memory usage.\n\nThen, The optimal inactive_ratio is changed frequently. So, everytime\ncalculation is better.\n\nTested-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@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": "a7885eb8ad465ec9db99ac5b5e6680f0ca8e11c8", "tree": "4f3ffaa399fbc16003cc1787228f10543dc9c3ef", "parents": [ "2733c06ac864ed40b9dfbbd5270f3f16949bd4a1" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Wed Jan 07 18:08:24 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:08 2009 -0800" }, "message": "memcg: swappiness\n\nCurrently, /proc/sys/vm/swappiness can change swappiness ratio for global\nreclaim. However, memcg reclaim doesn\u0027t have tuning parameter for itself.\n\nIn general, the optimal swappiness depend on workload. (e.g. hpc\nworkload need to low swappiness than the others.)\n\nThen, per cgroup swappiness improve administrator tunability.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Hugh Dickins \u003chugh@veritas.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": "e72e2bd6747c7a5c432197b6614cf3a387e61a0e", "tree": "2ab1a184ff5844c01a7c8b5a9392d1c906b37abe", "parents": [ "7f016ee8b6a9a43f768e6252021f169abec4fa1f" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed Jan 07 18:08:23 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:08 2009 -0800" }, "message": "memcg: rename scan global lru\n\nRename scan_global_lru() to scanning_global_lru().\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: Hugh Dickins \u003chugh@veritas.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": "9439c1c95b5c25b8031b2a7eb7e1590eb84be7f5", "tree": "79eaae56278407e0a223e562a2d0079834ca5529", "parents": [ "3e2f41f1f64744f7942980d93cc93dd3e5924560" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Wed Jan 07 18:08:21 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:08 2009 -0800" }, "message": "memcg: remove mem_cgroup_cal_reclaim()\n\nNow, get_scan_ratio() return correct value although memcg reclaim. Then,\nmem_cgroup_calc_reclaim() can be removed.\n\nSo, memcg reclaim get the same capability of anon/file reclaim balancing\nas global reclaim now.\n\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@redhat.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Hugh Dickins \u003chugh@veritas.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": "3e2f41f1f64744f7942980d93cc93dd3e5924560", "tree": "7b605c407b7470877fd9c5c853407f75edcbeb49", "parents": [ "a3d8e0549d913e30968fa02e505dfe02c0a23e0d" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Wed Jan 07 18:08:20 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:08 2009 -0800" }, "message": "memcg: add zone_reclaim_stat\n\nIntroduce mem_cgroup_per_zone::reclaim_stat member and its statics\ncollecting function.\n\nNow, get_scan_ratio() can calculate correct value on memcg reclaim.\n\n[hugh@veritas.com: avoid reclaim_stat oops when disabled]\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Hugh Dickins \u003chugh@veritas.com\u003e\nCc: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "a3d8e0549d913e30968fa02e505dfe02c0a23e0d", "tree": "e4d3df33666a18614289ae7bc76c27ad7950eb73", "parents": [ "14797e2363c2b2f1ce139fd1c5a215e4e05aa1d9" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Wed Jan 07 18:08:19 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:08 2009 -0800" }, "message": "memcg: add mem_cgroup_zone_nr_pages()\n\nIntroduce mem_cgroup_zone_nr_pages(). It is called by zone_nr_pages()\nhelper function.\n\nThis patch doesn\u0027t have any behavior change.\n\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nAcked-by: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Hugh Dickins \u003chugh@veritas.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": "14797e2363c2b2f1ce139fd1c5a215e4e05aa1d9", "tree": "a56edaa680c7c338a5a3043aa24897d7f668b6c9", "parents": [ "549927620b04a8f8073ce2ee2a8977f209af2ee5" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Wed Jan 07 18:08:18 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:08 2009 -0800" }, "message": "memcg: add inactive_anon_is_low()\n\nThe inactive_anon_is_low() is key component of active/inactive anon\nbalancing on reclaim. However current inactive_anon_is_low() function\nonly consider global reclaim.\n\nTherefore, we need following ugly scan_global_lru() condition.\n\n\tif (lru \u003d\u003d LRU_ACTIVE_ANON \u0026\u0026\n\t (!scan_global_lru(sc) || inactive_anon_is_low(zone))) {\n\t\tshrink_active_list(nr_to_scan, zone, sc, priority, file);\n\t\treturn 0;\n\nit cause that memcg reclaim always deactivate pages when shrink_list() is\ncalled. To make mem_cgroup_inactive_anon_is_low() improve active/inactive\nanon balancing of memcgroup.\n\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Cyrill Gorcunov \u003cgorcunov@gmail.com\u003e\nCc: \"Pekka Enberg\" \u003cpenberg@cs.helsinki.fi\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Hugh Dickins \u003chugh@veritas.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": "eeee9a8cd1e93c8b94e7788790fa9e2f8910c779", "tree": "2ef0a61a4ce12410ecfa48014a0181c03e73a3cb", "parents": [ "c9f299d9862deadf9fbee3ca28d915fdb006975a" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Wed Jan 07 18:08:17 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:07 2009 -0800" }, "message": "mm: make get_scan_ratio() safe for memcg\n\nCurrently, get_scan_ratio() always calculate the balancing value for\nglobal reclaim and memcg reclaim doesn\u0027t use it. Therefore it doesn\u0027t\nhave scan_global_lru() condition.\n\nHowever, we plan to expand get_scan_ratio() to be usable for memcg too,\nlatter. Then, The dependency code of global reclaim in the\nget_scan_ratio() insert into scan_global_lru() condision explictly.\n\nThis patch doesn\u0027t have any functional change.\n\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Hugh Dickins \u003chugh@veritas.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": "c9f299d9862deadf9fbee3ca28d915fdb006975a", "tree": "2979e1eb8fe7e15369f01bcb405a029c18ca3e66", "parents": [ "6e9015716ae9b59e9635d692fddfcfb9582c146c" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Wed Jan 07 18:08:16 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:07 2009 -0800" }, "message": "mm: add zone nr_pages helper function\n\nAdd zone_nr_pages() helper function.\n\nIt is used by a later patch. This patch doesn\u0027t have any functional\nchange.\n\nReviewed-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Hugh Dickins \u003chugh@veritas.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": "6e9015716ae9b59e9635d692fddfcfb9582c146c", "tree": "e1876d3822c46a20e1c35b41580f5ef6b2f6e053", "parents": [ "f89eb90e33fd4e4e0cc1a6d20afd63c5a561885a" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Wed Jan 07 18:08:15 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:07 2009 -0800" }, "message": "mm: introduce zone_reclaim struct\n\nAdd zone_reclam_stat struct for later enhancement.\n\nA later patch uses this. This patch doesn\u0027t any behavior change (yet).\n\nReviewed-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Hugh Dickins \u003chugh@veritas.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": "f89eb90e33fd4e4e0cc1a6d20afd63c5a561885a", "tree": "2e39e02196dda539c900383ab03f08a311124cd9", "parents": [ "670ec2f170301425fc4fdfa63d40652071fe85f6" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Wed Jan 07 18:08:14 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:07 2009 -0800" }, "message": "inactive_anon_is_low: move to vmscan\n\nThe inactive_anon_is_low() is called only vmscan. Then it can move to\nvmscan.c\n\nThis patch doesn\u0027t have any functional change.\n\nReviewd-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: Hugh Dickins \u003chugh@veritas.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": "08e552c69c6930d64722de3ec18c51844d06ee28", "tree": "a744d57ed4b23401115f1033dcaac9e85d550e09", "parents": [ "8c7c6e34a1256a5082d38c8e9bd1474476912715" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed Jan 07 18:08:01 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:05 2009 -0800" }, "message": "memcg: synchronized LRU\n\nA big patch for changing memcg\u0027s LRU semantics.\n\nNow,\n - page_cgroup is linked to mem_cgroup\u0027s its own LRU (per zone).\n\n - LRU of page_cgroup is not synchronous with global LRU.\n\n - page and page_cgroup is one-to-one and statically allocated.\n\n - To find page_cgroup is on what LRU, you have to check pc-\u003emem_cgroup as\n - lru \u003d page_cgroup_zoneinfo(pc, nid_of_pc, zid_of_pc);\n\n - SwapCache is handled.\n\nAnd, when we handle LRU list of page_cgroup, we do following.\n\n\tpc \u003d lookup_page_cgroup(page);\n\tlock_page_cgroup(pc); .....................(1)\n\tmz \u003d page_cgroup_zoneinfo(pc);\n\tspin_lock(\u0026mz-\u003elru_lock);\n\t.....add to LRU\n\tspin_unlock(\u0026mz-\u003elru_lock);\n\tunlock_page_cgroup(pc);\n\nBut (1) is spin_lock and we have to be afraid of dead-lock with zone-\u003elru_lock.\nSo, trylock() is used at (1), now. Without (1), we can\u0027t trust \"mz\" is correct.\n\nThis is a trial to remove this dirty nesting of locks.\nThis patch changes mz-\u003elru_lock to be zone-\u003elru_lock.\nThen, above sequence will be written as\n\n spin_lock(\u0026zone-\u003elru_lock); # in vmscan.c or swap.c via global LRU\n\tmem_cgroup_add/remove/etc_lru() {\n\t\tpc \u003d lookup_page_cgroup(page);\n\t\tmz \u003d page_cgroup_zoneinfo(pc);\n\t\tif (PageCgroupUsed(pc)) {\n\t\t\t....add to LRU\n\t\t}\n spin_lock(\u0026zone-\u003elru_lock); # in vmscan.c or swap.c via global LRU\n\nThis is much simpler.\n(*) We\u0027re safe even if we don\u0027t take lock_page_cgroup(pc). Because..\n 1. When pc-\u003emem_cgroup can be modified.\n - at charge.\n - at account_move().\n 2. at charge\n the PCG_USED bit is not set before pc-\u003emem_cgroup is fixed.\n 3. at account_move()\n the page is isolated and not on LRU.\n\nPros.\n - easy for maintenance.\n - memcg can make use of laziness of pagevec.\n - we don\u0027t have to duplicated LRU/Active/Unevictable bit in page_cgroup.\n - LRU status of memcg will be synchronized with global LRU\u0027s one.\n - # of locks are reduced.\n - account_move() is simplified very much.\nCons.\n - may increase cost of LRU rotation.\n (no impact if memcg is not configured.)\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\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": "8c7c6e34a1256a5082d38c8e9bd1474476912715", "tree": "09f53c7c4bac5532a9ecbdadb4450702c744ea6f", "parents": [ "27a7faa0779dd13729196c1a818c294f44bbd1ee" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Wed Jan 07 18:08:00 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:05 2009 -0800" }, "message": "memcg: mem+swap controller core\n\nThis patch implements per cgroup limit for usage of memory+swap. However\nthere are SwapCache, double counting of swap-cache and swap-entry is\navoided.\n\nMem+Swap controller works as following.\n - memory usage is limited by memory.limit_in_bytes.\n - memory + swap usage is limited by memory.memsw_limit_in_bytes.\n\nThis has following benefits.\n - A user can limit total resource usage of mem+swap.\n\n Without this, because memory resource controller doesn\u0027t take care of\n usage of swap, a process can exhaust all the swap (by memory leak.)\n We can avoid this case.\n\n And Swap is shared resource but it cannot be reclaimed (goes back to memory)\n until it\u0027s used. This characteristic can be trouble when the memory\n is divided into some parts by cpuset or memcg.\n Assume group A and group B.\n After some application executes, the system can be..\n\n Group A -- very large free memory space but occupy 99% of swap.\n Group B -- under memory shortage but cannot use swap...it\u0027s nearly full.\n\n Ability to set appropriate swap limit for each group is required.\n\nMaybe someone wonder \"why not swap but mem+swap ?\"\n\n - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means\n to move account from memory to swap...there is no change in usage of\n mem+swap.\n\n In other words, when we want to limit the usage of swap without affecting\n global LRU, mem+swap limit is better than just limiting swap.\n\nAccounting target information is stored in swap_cgroup which is\nper swap entry record.\n\nCharge is done as following.\n map\n - charge page and memsw.\n\n unmap\n - uncharge page/memsw if not SwapCache.\n\n swap-out (__delete_from_swap_cache)\n - uncharge page\n - record mem_cgroup information to swap_cgroup.\n\n swap-in (do_swap_page)\n - charged as page and memsw.\n record in swap_cgroup is cleared.\n memsw accounting is decremented.\n\n swap-free (swap_free())\n - if swap entry is freed, memsw is uncharged by PAGE_SIZE.\n\nThere are people work under never-swap environments and consider swap as\nsomething bad. For such people, this mem+swap controller extension is just an\noverhead. This overhead is avoided by config or boot option.\n(see Kconfig. detail is not in this patch.)\n\nTODO:\n - maybe more optimization can be don in swap-in path. (but not very safe.)\n But we just do simple accounting at this stage.\n\n[nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex]\n[hugh@veritas.com: memswap controller core swapcache fixes]\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nSigned-off-by: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "73ce02e96fe34a983199a9855b2ae738f960a6ee", "tree": "06dbbc2c38d55f8eacab09c55a68c736156b9540", "parents": [ "594fe1a044325bb0a1a49ca7d086e3df4f1df59a" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Tue Jan 06 14:40:33 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jan 06 15:59:10 2009 -0800" }, "message": "mm: stop kswapd\u0027s infinite loop at high order allocation\n\nWassim Dagash reported following kswapd infinite loop problem.\n\n kswapd runs in some infinite loop trying to swap until order 10 of zone\n highmem is OK.... kswapd will continue to try to balance order 10 of zone\n highmem forever (or until someone release a very large chunk of highmem).\n\nFor non order-0 allocations, the system may never be balanced due to\nfragmentation but kswapd should not infinitely loop as a result.\n\nInstead, recheck all watermarks at order-0 as they are the most important.\nIf watermarks are ok, kswapd will go back to sleep.\n\n[akpm@linux-foundation.org: fix comment]\nReported-by: wassim dagash \u003cwassim.dagash@gmail.com\u003e\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nReviewed-by: Nick Piggin \u003cnpiggin@suse.de\u003e\nSigned-off-by: Mel Gorman \u003cmel@csn.ul.ie\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "b555749aac87d7c2637f153e44bd77c7fdf4c65b", "tree": "b65a437fd70795eac749f5dcf327151d458ec742", "parents": [ "1e9e63650d6cb88e6d6d2ca6cc3ee276c26de4a3" ], "author": { "name": "Andrew Morton", "email": "akpm@linux-foundation.org", "time": "Tue Jan 06 14:40:13 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jan 06 15:59:08 2009 -0800" }, "message": "vmscan: shrink_active_list(): reduce lru_lock hold time\n\nThese three statements manipulate local variables and do not need the lock\ncoverage.\n\nCc: Johannes Weiner \u003channes@cmpxchg.org\u003e\nCc: Lee Schermerhorn \u003cLee.Schermerhorn@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "09f445e7f5107c91be12ed386350de6cd055e0a4", "tree": "4a1114bb896f51e18efbdc05d57ec358abc5aced", "parents": [ "01dbe5c9b1004dab045cb7f38428258ca9cddc02" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Tue Jan 06 14:40:03 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jan 06 15:59:06 2009 -0800" }, "message": "mm: kill zone_is_near_oom()\n\nzone_is_near_oom() is unused.\n\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": "01dbe5c9b1004dab045cb7f38428258ca9cddc02", "tree": "526bba2afd30a7a58e47597f0ab9255a37a21d55", "parents": [ "a79311c14eae4bb946a97af25f3e1b17d625985d" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Tue Jan 06 14:40:02 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jan 06 15:59:06 2009 -0800" }, "message": "vmscan: improve reclaim throughput to bail out patch\n\nThe vmscan bail out patch move nr_reclaimed variable to struct\nscan_control. Unfortunately, indirect access can easily happen cache\nmiss.\n\nif heavy memory pressure happend, that\u0027s ok.\ncache miss already plenty. it is not observable.\n\nbut, if memory pressure is lite, performance degression is obserbable.\n\nI compared following three pattern (it was mesured 10 times each)\n\nhackbench 125 process 3000\nhackbench 130 process 3000\nhackbench 135 process 3000\n\n 2.6.28-rc6 bail-out\n\n\t125\t130\t135\t\t125\t130\t135\n \u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\t71.866\t75.86\t81.274\t\t93.414\t73.254\t193.382\n\t74.145\t78.295\t77.27\t\t74.897\t75.021\t80.17\n\t70.305\t77.643\t75.855\t\t70.134\t77.571\t79.896\n\t74.288\t73.986\t75.955\t\t77.222\t78.48\t80.619\n\t72.029\t79.947\t78.312\t\t75.128\t82.172\t79.708\n\t71.499\t77.615\t77.042\t\t74.177\t76.532\t77.306\n\t76.188\t74.471\t83.562\t\t73.839\t72.43\t79.833\n\t73.236\t75.606\t78.743\t\t76.001\t76.557\t82.726\n\t69.427\t77.271\t76.691\t\t76.236\t79.371\t103.189\n\t72.473\t76.978\t80.643\t\t69.128\t78.932\t75.736\n\navg\t72.545\t76.767\t78.534\t\t76.017\t77.03\t93.256\nstd\t1.89\t1.71\t2.41\t\t6.29\t2.79\t34.16\nmin\t69.427\t73.986\t75.855\t\t69.128\t72.43\t75.736\nmax\t76.188\t79.947\t83.562\t\t93.414\t82.172\t193.382\n\nabout 4-5% degression.\n\nThen, this patch introduces a temporary local variable.\n\nresult:\n\n 2.6.28-rc6 this patch\n\nnum\t125\t130\t135\t\t125\t130\t135\n \u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\t71.866\t75.86\t81.274\t\t67.302\t68.269\t77.161\n\t74.145\t78.295\t77.27 \t72.616\t72.712\t79.06\n\t70.305\t77.643\t75.855 \t72.475\t75.712\t77.735\n\t74.288\t73.986\t75.955 \t69.229\t73.062\t78.814\n\t72.029\t79.947\t78.312 \t71.551\t74.392\t78.564\n\t71.499\t77.615\t77.042 \t69.227\t74.31\t78.837\n\t76.188\t74.471\t83.562 \t70.759\t75.256\t76.6\n\t73.236\t75.606\t78.743 \t69.966\t76.001\t78.464\n\t69.427\t77.271\t76.691 \t69.068\t75.218\t80.321\n\t72.473\t76.978\t80.643 \t72.057\t77.151\t79.068\n\navg\t72.545\t76.767\t78.534 \t\t70.425\t74.2083\t78.462\nstd \t1.89\t1.71\t2.41 \t1.66\t2.34\t1.00\nmin \t69.427\t73.986\t75.855 \t67.302\t68.269\t76.6\nmax \t76.188\t79.947\t83.562 \t72.616\t77.151\t80.321\n\nOK. the degression is disappeared.\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: Mel Gorman \u003cmel@csn.ul.ie\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "a79311c14eae4bb946a97af25f3e1b17d625985d", "tree": "f12391481dfd7a1c3ffcff75bd04d3a88957d16c", "parents": [ "ebdd4aea8d736e3b5ce27ab0a26860c9fded341b" ], "author": { "name": "Rik van Riel", "email": "riel@redhat.com", "time": "Tue Jan 06 14:40:01 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jan 06 15:59:06 2009 -0800" }, "message": "vmscan: bail out of direct reclaim after swap_cluster_max pages\n\nWhen the VM is under pressure, it can happen that several direct reclaim\nprocesses are in the pageout code simultaneously. It also happens that\nthe reclaiming processes run into mostly referenced, mapped and dirty\npages in the first round.\n\nThis results in multiple direct reclaim processes having a lower\npageout priority, which corresponds to a higher target of pages to\nscan.\n\nThis in turn can result in each direct reclaim process freeing\nmany pages. Together, they can end up freeing way too many pages.\n\nThis kicks useful data out of memory (in some cases more than half\nof all memory is swapped out). It also impacts performance by\nkeeping tasks stuck in the pageout code for too long.\n\nA 30% improvement in hackbench has been observed with this patch.\n\nThe fix is relatively simple: in shrink_zone() we can check how many\npages we have already freed, direct reclaim tasks break out of the\nscanning loop if they have already freed enough pages and have reached\na lower priority level.\n\nWe do not break out of shrink_zone() when priority \u003d\u003d DEF_PRIORITY,\nto ensure that equal pressure is applied to every zone in the common\ncase.\n\nHowever, in order to do this we do need to know how many pages we already\nfreed, so move nr_reclaimed into scan_control.\n\nakpm: a historical interlude...\n\nWe tried this in 2004:\n\n:commit e468e46a9bea3297011d5918663ce6d19094cf87\n:Author: akpm \u003cakpm\u003e\n:Date: Thu Jun 24 15:53:52 2004 +0000\n:\n:[PATCH] vmscan.c: dont reclaim too many pages\n:\n: The shrink_zone() logic can, under some circumstances, cause far too many\n: pages to be reclaimed. Say, we\u0027re scanning at high priority and suddenly hit\n: a large number of reclaimable pages on the LRU.\n: Change things so we bale out when SWAP_CLUSTER_MAX pages have been reclaimed.\n\nAnd we reverted it in 2006:\n\n:commit 210fe530305ee50cd889fe9250168228b2994f32\n:Author: Andrew Morton \u003cakpm@osdl.org\u003e\n:Date: Fri Jan 6 00:11:14 2006 -0800\n:\n: [PATCH] vmscan: balancing fix\n:\n: Revert a patch which went into 2.6.8-rc1. The changelog for that patch was:\n:\n: The shrink_zone() logic can, under some circumstances, cause far too many\n: pages to be reclaimed. Say, we\u0027re scanning at high priority and suddenly\n: hit a large number of reclaimable pages on the LRU.\n:\n: Change things so we bale out when SWAP_CLUSTER_MAX pages have been\n: reclaimed.\n:\n: Problem is, this change caused significant imbalance in inter-zone scan\n: balancing by truncating scans of larger zones.\n:\n: Suppose, for example, ZONE_HIGHMEM is 10x the size of ZONE_NORMAL. The zone\n: balancing algorithm would require that if we\u0027re scanning 100 pages of\n: ZONE_HIGHMEM, we should scan 10 pages of ZONE_NORMAL. But this logic will\n: cause the scanning of ZONE_HIGHMEM to bale out after only 32 pages are\n: reclaimed. Thus effectively causing smaller zones to be scanned relatively\n: harder than large ones.\n:\n: Now I need to remember what the workload was which caused me to write this\n: patch originally, then fix it up in a different way...\n\nAnd we haven\u0027t demonstrated that whatever problem caused that reversion is\nnot being reintroduced by this change in 2008.\n\nSigned-off-by: Rik van Riel \u003criel@redhat.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": "14b90b22ec0f359ef4791033ab386b2b627bae07", "tree": "aa8bf58d54743b914b1d539b05541e6f2a8a68a9", "parents": [ "ff30153bf9647c8646538810d4c01015a5e44787" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Tue Jan 06 14:39:45 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jan 06 15:59:04 2009 -0800" }, "message": "mm: make scan_zone_unevictable_pages() static\n\nsparse output following warning\n\n\tmm/vmscan.c:2507:6: warning: symbol \u0027scan_zone_unevictable_pages\u0027 was not declared. Should it be static?\n\ncleanup here.\n\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": "ff30153bf9647c8646538810d4c01015a5e44787", "tree": "8f423cd925332660c3d46b6b987a7fac6794073f", "parents": [ "d38d2a7582012ecf53aac33683ca5c689093cf65" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Tue Jan 06 14:39:44 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jan 06 15:59:04 2009 -0800" }, "message": "mm: make scan_all_zones_unevictable_pages() static\n\nsparse output following warning.\n\n\tmm/vmscan.c:2549:6: warning: symbol \u0027scan_all_zones_unevictable_pages\u0027 was not declared. Should it be static?\n\ncleanup here.\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\n\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "077cbc5864cd9188fa4c4e181e48ff58317e6400", "tree": "bcfcdecce7427801451fb8d6fcf5e7648bd2bfb3", "parents": [ "b962716b459505a8d83aea313fea0abe76749f42" ], "author": { "name": "KOSAKI Motohiro", "email": "kosaki.motohiro@jp.fujitsu.com", "time": "Tue Jan 06 14:39:42 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jan 06 15:59:04 2009 -0800" }, "message": "memcg: reclaim shouldn\u0027t change zone-\u003erecent_rotated statistics\n\nmemcg reclaim shouldn\u0027t change zone-\u003erecent_rotated statistics. If\nmemcgroup reclaim changes zone statistics, global reclaim can get a bit\nconfused.\n\nSigned-off-by: KOSAKI Motohiro \u003ckosaki.motohiro@jp.fujitsu.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: Balbir Singh \u003cbalbir@in.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": "b962716b459505a8d83aea313fea0abe76749f42", "tree": "b401e882f2ee8ff5a11209057496e603c211e8fe", "parents": [ "60371d971a3d01afd102f0bbf2681f32ecc31d78" ], "author": { "name": "Hugh Dickins", "email": "hugh@veritas.com", "time": "Tue Jan 06 14:39:41 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jan 06 15:59:04 2009 -0800" }, "message": "mm: optimize get_scan_ratio for no swap\n\nRik suggests a simplified get_scan_ratio() for !CONFIG_SWAP. Yes, the gcc\noptimizer gives us that, when nr_swap_pages is #defined as 0L. Move usual\ndeclaration to swapfile.c: it never belonged in page_alloc.c.\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nCc: Lee Schermerhorn \u003clee.schermerhorn@hp.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: Nick Piggin \u003cnickpiggin@yahoo.com.au\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Robin Holt \u003cholt@sgi.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "60371d971a3d01afd102f0bbf2681f32ecc31d78", "tree": "999cd75fbf80877f99b518e632c445ae849eeb4d", "parents": [ "ac47b003d03c2a4f28aef1d505b66d24ad191c4f" ], "author": { "name": "Hugh Dickins", "email": "hugh@veritas.com", "time": "Tue Jan 06 14:39:40 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jan 06 15:59:04 2009 -0800" }, "message": "mm: add add_to_swap stub\n\nIf we add a failing stub for add_to_swap(), then we can remove the #ifdef\nCONFIG_SWAP from mm/vmscan.c.\n\nThis was intended as a source cleanup, but looking more closely, it turns\nout that the !CONFIG_SWAP case was going to keep_locked for an anonymous\npage, whereas now it goes to the more suitable activate_locked, like the\nCONFIG_SWAP nr_swap_pages 0 case.\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nCc: Lee Schermerhorn \u003clee.schermerhorn@hp.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: Nick Piggin \u003cnickpiggin@yahoo.com.au\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Robin Holt \u003cholt@sgi.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "ac47b003d03c2a4f28aef1d505b66d24ad191c4f", "tree": "12c31491b756f4dc7a55a9e11b69b193743ff647", "parents": [ "63d6c5ad7fc27455ce5cb4706884671fb7e0df08" ], "author": { "name": "Hugh Dickins", "email": "hugh@veritas.com", "time": "Tue Jan 06 14:39:39 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jan 06 15:59:04 2009 -0800" }, "message": "mm: remove gfp_mask from add_to_swap\n\nRemove gfp_mask argument from add_to_swap(): it\u0027s misleading because its\nonly caller, shrink_page_list(), is not atomic at that point; and in due\ncourse (implementing discard) we\u0027ll sometimes want to allocate some memory\nwith GFP_NOIO (as is used in swap_writepage) when allocating swap.\n\nNo change to the gfp_mask passed down to add_to_swap_cache(): still use\n__GFP_HIGH without __GFP_WAIT (with nomemalloc and nowarn as before):\nthough it\u0027s not obvious if that\u0027s the best combination to ask for here.\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nCc: Lee Schermerhorn \u003clee.schermerhorn@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Nick Piggin \u003cnickpiggin@yahoo.com.au\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Robin Holt \u003cholt@sgi.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "63d6c5ad7fc27455ce5cb4706884671fb7e0df08", "tree": "82cf04209a4825ab4daf2a94977fc44550d2c9b7", "parents": [ "68bdc8d64742ccc5e340c5d122ebbab3f0cf2a74" ], "author": { "name": "Hugh Dickins", "email": "hugh@veritas.com", "time": "Tue Jan 06 14:39:38 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jan 06 15:59:03 2009 -0800" }, "message": "mm: remove try_to_munlock from vmscan\n\nAn unfortunate feature of the Unevictable LRU work was that reclaiming an\nanonymous page involved an extra scan through the anon_vma: to check that\nthe page is evictable before allocating swap, because the swap could not\nbe freed reliably soon afterwards.\n\nNow try_to_free_swap() has replaced remove_exclusive_swap_page(), that\u0027s\nnot an issue any more: remove try_to_munlock() call from\nshrink_page_list(), leaving it to try_to_munmap() to discover if the page\nis one to be culled to the unevictable list - in which case then\ntry_to_free_swap().\n\nUpdate unevictable-lru.txt to remove comments on the try_to_munlock() in\nshrink_page_list(), and shorten some lines over 80 columns.\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nCc: Lee Schermerhorn \u003clee.schermerhorn@hp.com\u003e\nAcked-by: Rik van Riel \u003criel@redhat.com\u003e\nCc: Nick Piggin \u003cnickpiggin@yahoo.com.au\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Robin Holt \u003cholt@sgi.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "a2c43eed8334e878702fca713b212ae2a11d84b9", "tree": "4eb3f9b9153df5e7a638b15a6f7c9aa924066a94", "parents": [ "7b1fe59793e61f826bef053107b57b23954833bb" ], "author": { "name": "Hugh Dickins", "email": "hugh@veritas.com", "time": "Tue Jan 06 14:39:36 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jan 06 15:59:03 2009 -0800" }, "message": "mm: try_to_free_swap replaces remove_exclusive_swap_page\n\nremove_exclusive_swap_page(): its problem is in living up to its name.\n\nIt doesn\u0027t matter if someone else has a reference to the page (raised\npage_count); it doesn\u0027t matter if the page is mapped into userspace\n(raised page_mapcount - though that hints it may be worth keeping the\nswap): all that matters is that there be no more references to the swap\n(and no writeback in progress).\n\nswapoff (try_to_unuse) has been removing pages from swapcache for years,\nwith no concern for page count or page mapcount, and we used to have a\ncomment in lookup_swap_cache() recognizing that: if you go for a page of\nswapcache, you\u0027ll get the right page, but it could have been removed from\nswapcache by the time you get page lock.\n\nSo, give up asking for exclusivity: get rid of\nremove_exclusive_swap_page(), and remove_exclusive_swap_page_ref() and\nremove_exclusive_swap_page_count() which were spawned for the recent LRU\nwork: replace them by the simpler try_to_free_swap() which just checks\npage_swapcount().\n\nSimilarly, remove the page_count limitation from free_swap_and_count(),\nbut assume that it\u0027s worth holding on to the swap if page is mapped and\nswap nowhere near full. Add a vm_swap_full() test in free_swap_cache()?\nIt would be consistent, but I think we probably have enough for now.\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nCc: Lee Schermerhorn \u003clee.schermerhorn@hp.com\u003e\nCc: Rik van Riel \u003criel@redhat.com\u003e\nCc: Nick Piggin \u003cnickpiggin@yahoo.com.au\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Robin Holt \u003cholt@sgi.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" } ], "next": "174596a0b9f21e8844d70566a6bb29bf48a87750" }