)]}' { "log": [ { "commit": "811158b147a503fbdf9773224004ffd32002d1fe", "tree": "0a11dcfefe721bfc38ea9f1f4a238822dbae0dda", "parents": [ "4e76c5ccd5ac9bd003467d3bb0f49b18572dd4cd", "b26e0ed4936b743b693a4cc1413561fa3e4eaf65" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Apr 03 15:24:35 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Apr 03 15:24:35 2009 -0700" }, "message": "Merge branch \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial\n\n* \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial: (28 commits)\n trivial: Update my email address\n trivial: NULL noise: drivers/mtd/tests/mtd_*test.c\n trivial: NULL noise: drivers/media/dvb/frontends/drx397xD_fw.h\n trivial: Fix misspelling of \"Celsius\".\n trivial: remove unused variable \u0027path\u0027 in alloc_file()\n trivial: fix a pdlfush -\u003e pdflush typo in comment\n trivial: jbd header comment typo fix for JBD_PARANOID_IOFAIL\n trivial: wusb: Storage class should be before const qualifier\n trivial: drivers/char/bsr.c: Storage class should be before const qualifier\n trivial: h8300: Storage class should be before const qualifier\n trivial: fix where cgroup documentation is not correctly referred to\n trivial: Give the right path in Documentation example\n trivial: MTD: remove EOL from MODULE_DESCRIPTION\n trivial: Fix typo in bio_split()\u0027s documentation\n trivial: PWM: fix of #endif comment\n trivial: fix typos/grammar errors in Kconfig texts\n trivial: Fix misspelling of firmware\n trivial: cgroups: documentation typo and spelling corrections\n trivial: Update contact info for Jochen Hein\n trivial: fix typo \"resgister\" -\u003e \"register\"\n ...\n" }, { "commit": "bd1a8ab73edd449fecda633449cc277b856ad4f5", "tree": "66095c2a59f8706138e89190ec6c1d9cfaab6916", "parents": [ "0b4217b3fdddc4a58939720d3ed809537577d48b" ], "author": { "name": "Li Zefan", "email": "lizf@cn.fujitsu.com", "time": "Thu Apr 02 16:57:50 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Apr 02 19:04:56 2009 -0700" }, "message": "cgroups: add \u0027data\u0027 field to struct cgroup_scanner\n\nWe need to pass some data to test_task() or process_task() in some cases.\nWill be used later.\n\nSigned-off-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "0b7f569e45bb6be142d87017030669a6a7d327a1", "tree": "8df7877b95c093ebf4cb4e1006cea16f75fc79b7", "parents": [ "81d39c20f5ee2437d71709beb82597e2a38efbbc" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Thu Apr 02 16:57:38 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Apr 02 19:04:55 2009 -0700" }, "message": "memcg: fix OOM killer under memcg\n\nThis patch tries to fix OOM Killer problems caused by hierarchy.\nNow, memcg itself has OOM KILL function (in oom_kill.c) and tries to\nkill a task in memcg.\n\nBut, when hierarchy is used, it\u0027s broken and correct task cannot\nbe killed. For example, in following cgroup\n\n\t/groupA/\thierarchy\u003d1, limit\u003d1G,\n\t\t01\tnolimit\n\t\t02\tnolimit\nAll tasks\u0027 memory usage under /groupA, /groupA/01, groupA/02 is limited to\ngroupA\u0027s 1Gbytes but OOM Killer just kills tasks in groupA.\n\nThis patch provides makes the bad process be selected from all tasks\nunder hierarchy. BTW, currently, oom_jiffies is updated against groupA\nin above case. oom_jiffies of tree should be updated.\n\nTo see how oom_jiffies is used, please check mem_cgroup_oom_called()\ncallers.\n\n[akpm@linux-foundation.org: build fix]\n[akpm@linux-foundation.org: const fix]\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nCc: David Rientjes \u003crientjes@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "099fca3225b39f7a3ed853036038054172b55581", "tree": "c7a8863f9ca8a5c745297e9ee43b63494b6022d1", "parents": [ "b6719ec1ad54e47e40633b19703f2c1254708842" ], "author": { "name": "Li Zefan", "email": "lizf@cn.fujitsu.com", "time": "Thu Apr 02 16:57:29 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Apr 02 19:04:54 2009 -0700" }, "message": "cgroups: show correct file mode\n\nWe have some read-only files and write-only files, but currently they are\nall set to 0644, which is counter-intuitive and cause trouble for some\ncgroup tools like libcgroup.\n\nThis patch adds \u0027mode\u0027 to struct cftype to allow cgroup subsys to set it\u0027s\nown files\u0027 file mode, and for the most cases cft-\u003emode can be default to 0\nand cgroup will figure out proper mode.\n\nAcked-by: Paul Menage \u003cmenage@google.com\u003e\nReviewed-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "ec64f51545fffbc4cb968f0cea56341a4b07e85a", "tree": "575d890a6759d81f3324fa2a22ca6ab14a41eefc", "parents": [ "38460b48d06440de46b34cb778bd6c4855030754" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Thu Apr 02 16:57:26 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Apr 02 19:04:54 2009 -0700" }, "message": "cgroup: fix frequent -EBUSY at rmdir\n\nIn following situation, with memory subsystem,\n\n\t/groupA use_hierarchy\u003d\u003d1\n\t\t/01 some tasks\n\t\t/02 some tasks\n\t\t/03 some tasks\n\t\t/04 empty\n\nWhen tasks under 01/02/03 hit limit on /groupA, hierarchical reclaim\nis triggered and the kernel walks tree under groupA. In this case,\nrmdir /groupA/04 fails with -EBUSY frequently because of temporal\nrefcnt from the kernel.\n\nIn general. cgroup can be rmdir\u0027d if there are no children groups and\nno tasks. Frequent fails of rmdir() is not useful to users.\n(And the reason for -EBUSY is unknown to users.....in most cases)\n\nThis patch tries to modify above behavior, by\n\t- retries if css_refcnt is got by someone.\n\t- add \"return value\" to pre_destroy() and allows subsystem to\n\t say \"we\u0027re really busy!\"\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\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": "38460b48d06440de46b34cb778bd6c4855030754", "tree": "8f3362a446b5b03879f715c3f7279e70842bcca9", "parents": [ "313e924c0852943e67335fad9d2608701f0dfe8e" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Thu Apr 02 16:57:25 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Apr 02 19:04:53 2009 -0700" }, "message": "cgroup: CSS ID support\n\nPatch for Per-CSS(Cgroup Subsys State) ID and private hierarchy code.\n\nThis patch attaches unique ID to each css and provides following.\n\n - css_lookup(subsys, id)\n returns pointer to struct cgroup_subysys_state of id.\n - css_get_next(subsys, id, rootid, depth, foundid)\n returns the next css under \"root\" by scanning\n\nWhen cgroup_subsys-\u003euse_id is set, an id for css is maintained.\n\nThe cgroup framework only parepares\n\t- css_id of root css for subsys\n\t- id is automatically attached at creation of css.\n\t- id is *not* freed automatically. Because the cgroup framework\n\t don\u0027t know lifetime of cgroup_subsys_state.\n\t free_css_id() function is provided. This must be called by subsys.\n\nThere are several reasons to develop this.\n\t- Saving space .... For example, memcg\u0027s swap_cgroup is array of\n\t pointers to cgroup. But it is not necessary to be very fast.\n\t By replacing pointers(8bytes per ent) to ID (2byes per ent), we can\n\t reduce much amount of memory usage.\n\n\t- Scanning without lock.\n\t CSS_ID provides \"scan id under this ROOT\" function. By this, scanning\n\t css under root can be written without locks.\n\t ex)\n\t do {\n\t\trcu_read_lock();\n\t\tnext \u003d cgroup_get_next(subsys, id, root, \u0026found);\n\t\t/* check sanity of next here */\n\t\tcss_tryget();\n\t\trcu_read_unlock();\n\t\tid \u003d found + 1\n\t } while(...)\n\nCharacteristics:\n\t- Each css has unique ID under subsys.\n\t- Lifetime of ID is controlled by subsys.\n\t- css ID contains \"ID\" and \"Depth in hierarchy\" and stack of hierarchy\n\t- Allowed ID is 1-65535, ID 0 is UNUSED ID.\n\nDesign Choices:\n\t- scan-by-ID v.s. scan-by-tree-walk.\n\t As /proc\u0027s pid scan does, scan-by-ID is robust when scanning is done\n\t by following kind of routine.\n\t scan -\u003e rest a while(release a lock) -\u003e conitunue from interrupted\n\t memcg\u0027s hierarchical reclaim does this.\n\n\t- When subsys-\u003euse_id is set, # of css in the system is limited to\n\t 65535.\n\n[bharata@linux.vnet.ibm.com: remove rcu_read_lock() from css_get_next()]\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Daisuke Nishimura \u003cnishimura@mxp.nes.nec.co.jp\u003e\nSigned-off-by: Bharata B Rao \u003cbharata@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": "313e924c0852943e67335fad9d2608701f0dfe8e", "tree": "fa4c3f65a7ed6edea52ae78b012138ebab1420c3", "parents": [ "d20a390a0ee2bf2f692c539c6ce1c829e1080bb5" ], "author": { "name": "Grzegorz Nosek", "email": "root@localdomain.pl", "time": "Thu Apr 02 16:57:23 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Apr 02 19:04:53 2009 -0700" }, "message": "cgroups: relax ns_can_attach checks to allow attaching to grandchild cgroups\n\nThe ns_proxy cgroup allows moving processes to child cgroups only one\nlevel deep at a time. This commit relaxes this restriction and makes it\npossible to attach tasks directly to grandchild cgroups, e.g.:\n\n($pid is in the root cgroup)\necho $pid \u003e /cgroup/CG1/CG2/tasks\n\nPreviously this operation would fail with -EPERM and would have to be\nperformed as two steps:\necho $pid \u003e /cgroup/CG1/tasks\necho $pid \u003e /cgroup/CG1/CG2/tasks\n\nAlso, the target cgroup no longer needs to be empty to move a task there.\n\nSigned-off-by: Grzegorz Nosek \u003croot@localdomain.pl\u003e\nAcked-by: Serge Hallyn \u003cserue@us.ibm.com\u003e\nReviewed-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "d20a390a0ee2bf2f692c539c6ce1c829e1080bb5", "tree": "50fc49556959daf7da2f7194e8f5d02b2e52e3d8", "parents": [ "45dad7bd9d9b65a30d6e790b111f6f2d8f746d22" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Thu Apr 02 16:57:22 2009 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Apr 02 19:04:53 2009 -0700" }, "message": "cgroups: fix cgroup.h comments\n\nFix the style of some multi-line comments in cgroup.h to match\nDocumentation/CodingStyle\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "21acb9caa2e30b100e9a1943d995bb99d40f4035", "tree": "0c09bde664cb75b0b5e27745628ab458e2472f60", "parents": [ "6d5e147dd034d9ceedc89fe39f4284700944f0c8" ], "author": { "name": "Thadeu Lima de Souza Cascardo", "email": "cascardo@holoscopio.com", "time": "Wed Feb 04 10:12:08 2009 +0100" }, "committer": { "name": "Jiri Kosina", "email": "jkosina@suse.cz", "time": "Mon Mar 30 15:22:02 2009 +0200" }, "message": "trivial: fix where cgroup documentation is not correctly referred to\n\ncgroup documentation was moved to Documentation/cgroups/. There are some\nplaces that still refer to Documentation/controllers/,\nDocumentation/cgroups.txt and Documentation/cpusets.txt. Fix those.\n\nSigned-off-by: Thadeu Lima de Souza Cascardo \u003ccascardo@holoscopio.com\u003e\nReviewed-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nAcked-by: Paul Menage \u003cmenage@google.com\u003e\nSigned-off-by: Jiri Kosina \u003cjkosina@suse.cz\u003e\n" }, { "commit": "cfebe563bd0a3ff97e1bc167123120d59c7a84db", "tree": "6178bf45bcccaf3d43d87cfe0eef059d849c7140", "parents": [ "01c4a4283137d24c9cc3785f1f312e895a18f273" ], "author": { "name": "Li Zefan", "email": "lizf@cn.fujitsu.com", "time": "Wed Feb 11 13:04:36 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Feb 11 14:25:36 2009 -0800" }, "message": "cgroups: fix lockdep subclasses overflow\n\nI enabled all cgroup subsystems when compiling kernel, and then:\n # mount -t cgroup -o net_cls xxx /mnt\n # mkdir /mnt/0\n\nThis showed up immediately:\n BUG: MAX_LOCKDEP_SUBCLASSES too low!\n turning off the locking correctness validator.\n\nIt\u0027s caused by the cgroup hierarchy lock:\n\tfor (i \u003d 0; i \u003c CGROUP_SUBSYS_COUNT; i++) {\n\t\tstruct cgroup_subsys *ss \u003d subsys[i];\n\t\tif (ss-\u003eroot \u003d\u003d root)\n\t\t\tmutex_lock_nested(\u0026ss-\u003ehierarchy_mutex, i);\n\t}\n\nNow we have 9 cgroup subsystems, and the above \u0027i\u0027 for net_cls is 8, but\nMAX_LOCKDEP_SUBCLASSES is 8.\n\nThis patch uses different lockdep keys for different subsystems.\n\nSigned-off-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nAcked-by: Paul Menage \u003cmenage@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "804b3c28a4e4fa1c224571bf76edb534b9c4b1ed", "tree": "59e0dabb227a5067cfdc7cce13f149f141ef8cb4", "parents": [ "1404f06565ee89e0ce04d4a5859c00b0e3a0dc8d" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Thu Jan 29 14:25:21 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 29 18:04:45 2009 -0800" }, "message": "cgroups: add cpu_relax() calls in css_tryget() and cgroup_clear_css_refs()\n\ncss_tryget() and cgroup_clear_css_refs() contain polling loops; these\nloops should have cpu_relax calls in them to reduce cross-cache traffic.\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "e7c5ec9193d32b9559a3bb8893ceedbda85201ff", "tree": "fd7505c03e2c7525a110a702907c0604c233e2d9", "parents": [ "2cb378c862777d050c20db903b119a029845fdcb" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Wed Jan 07 18:08:38 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:10 2009 -0800" }, "message": "cgroups: add css_tryget()\n\nAdd css_tryget(), that obtains a counted reference on a CSS. It is used\nin situations where the caller has a \"weak\" reference to the CSS, i.e.\none that does not protect the cgroup from removal via a reference count,\nbut would instead be cleaned up by a destroy() callback.\n\ncss_tryget() will return true on success, or false if the cgroup is being\nremoved.\n\nThis is similar to Kamezawa Hiroyuki\u0027s patch from a week or two ago, but\nwith the difference that in the event of css_tryget() racing with a\ncgroup_rmdir(), css_tryget() will only return false if the cgroup really\ndoes get removed.\n\nThis implementation is done by biasing css-\u003erefcnt, so that a refcnt of 1\nmeans \"releasable\" and 0 means \"released or releasing\". In the event of a\nrace, css_tryget() distinguishes between \"released\" and \"releasing\" by\nchecking for the CSS_REMOVED flag in css-\u003eflags.\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nTested-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\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "999cd8a450f8f93701669a61cac4d3b19eca07e8", "tree": "990e5b08e6db971d2e9943f89abf39e7c8f4cb1e", "parents": [ "b5a84319a4343a0db753436fd8147e61eaafa7ea" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Wed Jan 07 18:08:36 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:10 2009 -0800" }, "message": "cgroups: add a per-subsystem hierarchy_mutex\n\nThese patches introduce new locking/refcount support for cgroups to\nreduce the need for subsystems to call cgroup_lock(). This will\nultimately allow the atomicity of cgroup_rmdir() (which was removed\nrecently) to be restored.\n\nThese three patches give:\n\n1/3 - introduce a per-subsystem hierarchy_mutex which a subsystem can\n use to prevent changes to its own cgroup tree\n\n2/3 - use hierarchy_mutex in place of calling cgroup_lock() in the\n memory controller\n\n3/3 - introduce a css_tryget() function similar to the one recently\n proposed by Kamezawa, but avoiding spurious refcount failures in\n the event of a race between a css_tryget() and an unsuccessful\n cgroup_rmdir()\n\nFuture patches will likely involve:\n\n- using hierarchy mutex in place of cgroup_lock() in more subsystems\n where appropriate\n\n- restoring the atomicity of cgroup_rmdir() with respect to cgroup_create()\n\nThis patch:\n\nAdd a hierarchy_mutex to the cgroup_subsys object that protects changes to\nthe hierarchy observed by that subsystem. It is taken by the cgroup\nsubsystem (in addition to cgroup_mutex) for the following operations:\n\n- linking a cgroup into that subsystem\u0027s cgroup tree\n- unlinking a cgroup from that subsystem\u0027s cgroup tree\n- moving the subsystem to/from a hierarchy (including across the\n bind() callback)\n\nThus if the subsystem holds its own hierarchy_mutex, it can safely\ntraverse its own hierarchy.\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nTested-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\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "a47295e6bc42ad35f9c15ac66f598aa24debd4e2", "tree": "cb765e996ef35ae88e29d60796655d0d35e8cf5e", "parents": [ "e7b80bb695a5b64c92e314838e083b2f3bdf29b2" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Wed Jan 07 18:07:44 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:03 2009 -0800" }, "message": "cgroups: make cgroup_path() RCU-safe\n\nFix races between /proc/sched_debug by freeing cgroup objects via an RCU\ncallback. Thus any cgroup reference obtained from an RCU-safe source will\nremain valid during the RCU section. Since dentries are also RCU-safe,\nthis allows us to traverse up the tree safely.\n\nAdditionally, make cgroup_path() check for a NULL cgrp-\u003edentry to avoid\ntrying to report a path for a partially-created cgroup.\n\n[lizf@cn.fujitsu.com: call deactive_super() in cgroup_diput()]\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nReviewed-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nTested-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nSigned-off-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "b2aa30f7bb381e04c93eed106089ba55553955f1", "tree": "eda632102a0bc6e325a24e77a9c2819eab5aeef3", "parents": [ "2019f634ce5904c19eba4e86f51b1a119a53a9f1" ], "author": { "name": "Lai Jiangshan", "email": "laijs@cn.fujitsu.com", "time": "Wed Jan 07 18:07:37 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jan 08 08:31:02 2009 -0800" }, "message": "cgroups: don\u0027t put struct cgroupfs_root protected by RCU\n\nWe don\u0027t access struct cgroupfs_root in fast path, so we should not put\nstruct cgroupfs_root protected by RCU\n\nBut the comment in struct cgroup_subsys.root confuse us.\n\nstruct cgroup_subsys.root is used in these places:\n\n1 find_css_set(): if (ss-\u003eroot-\u003esubsys_list.next \u003d\u003d \u0026ss-\u003esibling)\n2 rebind_subsystems(): if (ss-\u003eroot !\u003d \u0026rootnode)\n rcu_assign_pointer(ss-\u003eroot, root);\n rcu_assign_pointer(subsys[i]-\u003eroot, \u0026rootnode);\n3 cgroup_has_css_refs(): if (ss-\u003eroot !\u003d cgrp-\u003eroot)\n4 cgroup_init_subsys(): ss-\u003eroot \u003d \u0026rootnode;\n5 proc_cgroupstats_show(): ss-\u003ename, ss-\u003eroot-\u003esubsys_bits,\n ss-\u003eroot-\u003enumber_of_cgroups, !ss-\u003edisabled);\n6 cgroup_clone(): root \u003d subsys-\u003eroot;\n if ((root !\u003d subsys-\u003eroot) ||\n\nAll these place we have held cgroup_lock() or we don\u0027t dereference to\nstruct cgroupfs_root. It\u0027s means wo don\u0027t need RCU when use struct\ncgroup_subsys.root, and we should not put struct cgroupfs_root protected\nby RCU.\n\nSigned-off-by: Lai Jiangshan \u003claijs@cn.fujitsu.com\u003e\nReviewed-by: Paul Menage \u003cmenage@google.com\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\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": "e5991371ee0d1c0ce19e133c6f9075b49c5b4ae8", "tree": "3e5c9c3293c025f537ff4416660a96d33c694b5f", "parents": [ "39f0dee2d8abe902617622b71f8f6f73985ec71c" ], "author": { "name": "Hugh Dickins", "email": "hugh@veritas.com", "time": "Tue Jan 06 14:39:22 2009 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jan 06 15:59:01 2009 -0800" }, "message": "mm: remove cgroup_mm_owner_callbacks\n\ncgroup_mm_owner_callbacks() was brought in to support the memrlimit\ncontroller, but sneaked into mainline ahead of it. That controller has\nnow been shelved, and the mm_owner_changed() args were inadequate for it\nanyway (they needed an mm pointer instead of a task pointer).\n\nRemove the dead code, and restore mm_update_next_owner() locking to how it\nwas before: taking mmap_sem there does nothing for memcontrol.c, now the\nonly user of mm-\u003eowner.\n\nSigned-off-by: Hugh Dickins \u003chugh@veritas.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "9b913735e53ab0da4a792bac0de8e178cc13dcfb", "tree": "8ecc6e20f99379674b1977631a31e6d33de43174", "parents": [ "51308ee59dee1136ed599d875ea8968d7be55c91" ], "author": { "name": "Li Zefan", "email": "lizf@cn.fujitsu.com", "time": "Wed Oct 29 14:00:54 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Oct 30 11:38:45 2008 -0700" }, "message": "cgroups: tiny cleanups\n\n- remove \u0027private\u0027 field from struct subsys\n- remove cgroup_init_smp()\n\nSigned-off-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nAcked-by: Paul Menage \u003cmenage@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "886465f407e57d6c3c81013c919ea670ce1ae0d0", "tree": "c8981531dee58b1fc19a383068d6770c030bd534", "parents": [ "cc31edceee04a7b87f2be48f9489ebb72d264844" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Sat Oct 18 20:28:05 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Oct 20 08:52:38 2008 -0700" }, "message": "cgroups: fix declaration of cgroup_mm_owner_callbacks\n\nThe choice of real/dummy declaration for cgroup_mm_owner_callbacks()\nshouldn\u0027t be based on CONFIG_MM_OWNER, but on CONFIG_CGROUPS. Otherwise\nkernel/exit.c fails to compile when something other than a cgroups\ncontroller selects CONFIG_MM_OWNER\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nAcked-by: Pekka Enberg \u003cpenberg@cs.helsinki.fi\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "cc31edceee04a7b87f2be48f9489ebb72d264844", "tree": "5d37791218c420281e509899645d89aee7902f2b", "parents": [ "146aa1bd0511f88ddb4e92fafa2b8aad4f2f65f3" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Sat Oct 18 20:28:04 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Oct 20 08:52:38 2008 -0700" }, "message": "cgroups: convert tasks file to use a seq_file with shared pid array\n\nRather than pre-generating the entire text for the \"tasks\" file each\ntime the file is opened, we instead just generate/update the array of\nprocess ids and use a seq_file to report these to userspace. All open\nfile handles on the same \"tasks\" file can share a pid array, which may\nbe updated any time that no thread is actively reading the array. By\nsharing the array, the potential for userspace to DoS the system by\nopening many handles on the same \"tasks\" file is removed.\n\n[Based on a patch by Lai Jiangshan, extended to use seq_file]\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nReviewed-by: Lai Jiangshan \u003claijs@cn.fujitsu.com\u003e\nCc: Serge Hallyn \u003cserue@us.ibm.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": "146aa1bd0511f88ddb4e92fafa2b8aad4f2f65f3", "tree": "d7deb46b9a38f82f109b2126317899efbbce41c2", "parents": [ "248736c2a57206388c86f8cdd3392ee986e84f9f" ], "author": { "name": "Lai Jiangshan", "email": "laijs@cn.fujitsu.com", "time": "Sat Oct 18 20:28:03 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Oct 20 08:52:38 2008 -0700" }, "message": "cgroups: fix probable race with put_css_set[_taskexit] and find_css_set\n\nput_css_set_taskexit may be called when find_css_set is called on other\ncpu. And the race will occur:\n\nput_css_set_taskexit side find_css_set side\n\n |\natomic_dec_and_test(\u0026kref-\u003erefcount) |\n /* kref-\u003erefcount \u003d 0 */ |\n....................................................................\n | read_lock(\u0026css_set_lock)\n | find_existing_css_set\n | get_css_set\n | read_unlock(\u0026css_set_lock);\n....................................................................\n__release_css_set |\n....................................................................\n | /* use a released css_set */\n |\n\n[put_css_set is the same. But in the current code, all put_css_set are\nput into cgroup mutex critical region as the same as find_css_set.]\n\n[akpm@linux-foundation.org: repair comments]\n[menage@google.com: eliminate race in css_set refcounting]\nSigned-off-by: Lai Jiangshan \u003claijs@cn.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "9363b9f23c9cc36cc8ef6c05fdf879ee4a96ae92", "tree": "482746b06d6fdd8be606de4dff584a3a40054c4c", "parents": [ "1648993fb05c487947c1cec6307aca29d8002abe" ], "author": { "name": "Balbir Singh", "email": "balbir@linux.vnet.ibm.com", "time": "Wed Oct 15 22:01:05 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Oct 16 11:21:28 2008 -0700" }, "message": "memrlimit: cgroup mm owner callback changes to add task info\n\nThis patch adds an additional field to the mm_owner callbacks. This field\nis required to get to the mm that changed. Hold mmap_sem in write mode\nbefore calling the mm_owner_changed callback\n\n[hugh@veritas.com: fix mmap_sem deadlock]\nSigned-off-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Sudhir Kumar \u003cskumar@linux.vnet.ibm.com\u003e\nCc: YAMAMOTO Takashi \u003cyamamoto@valinux.co.jp\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nCc: Pavel Emelianov \u003cxemul@openvz.org\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: David Rientjes \u003crientjes@google.com\u003e\nCc: Vivek Goyal \u003cvgoyal@redhat.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": "e885dcde75685e09f23cffae1f6d5169c105b8a0", "tree": "711a91e83fad632c194700839d3e47631aee677a", "parents": [ "856c13aa1ff6136c1968414fdea5938ea9d5ebf2" ], "author": { "name": "Serge E. Hallyn", "email": "serue@us.ibm.com", "time": "Fri Jul 25 01:47:06 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jul 25 10:53:37 2008 -0700" }, "message": "cgroup_clone: use pid of newly created task for new cgroup\n\ncgroup_clone creates a new cgroup with the pid of the task. This works\ncorrectly for unshare, but for clone cgroup_clone is called from\ncopy_namespaces inside copy_process, which happens before the new pid is\ncreated. As a result, the new cgroup was created with current\u0027s pid.\nThis patch:\n\n\t1. Moves the call inside copy_process to after the new pid\n\t is created\n\t2. Passes the struct pid into ns_cgroup_clone (as it is not\n\t yet attached to the task)\n\t3. Passes a name from ns_cgroup_clone() into cgroup_clone()\n\t so as to keep cgroup_clone() itself simpler\n\t4. Uses pid_vnr() to get the process id value, so that the\n\t pid used to name the new cgroup is always the pid as it\n\t would be known to the task which did the cloning or\n\t unsharing. I think that is the most intuitive thing to\n\t do. This way, task t1 does clone(CLONE_NEWPID) to get\n\t t2, which does clone(CLONE_NEWPID) to get t3, then the\n\t cgroup for t3 will be named for the pid by which t2 knows\n\t t3.\n\n(Thanks to Dan Smith for finding the main bug)\n\nChangelog:\n\tJune 11: Incorporate Paul Menage\u0027s feedback: don\u0027t pass\n\t NULL to ns_cgroup_clone from unshare, and reduce\n\t\t patch size by using \u0027nodename\u0027 in cgroup_clone.\n\tJune 10: Original version\n\n[akpm@linux-foundation.org: build fix]\n[akpm@linux-foundation.org: coding-style fixes]\nSigned-off-by: Serge Hallyn \u003cserge@us.ibm.com\u003e\nAcked-by: Paul Menage \u003cmenage@google.com\u003e\nTested-by: Dan Smith \u003cdanms@us.ibm.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": "84eea842886ac35020be6043e04748ed22014359", "tree": "acd2bb7c3bc267a032b5e637c598708f38c68cb3", "parents": [ "e788e066c651b1bbf4a927dc95395c1aa13be436" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Fri Jul 25 01:47:00 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jul 25 10:53:35 2008 -0700" }, "message": "cgroups: misc cleanups to write_string patchset\n\nThis patch contains cleanups suggested by reviewers for the recent\nwrite_string() patchset:\n\n- pair cgroup_lock_live_group() with cgroup_unlock() in cgroup.c for\n clarity, rather than directly unlocking cgroup_mutex.\n\n- make the return type of cgroup_lock_live_group() a bool\n\n- use a #define\u0027d constant for the local buffer size in read/write functions\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nAcked-by: Serge Hallyn \u003cserue@us.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": "e788e066c651b1bbf4a927dc95395c1aa13be436", "tree": "8ffb80dc5917ba5b77e17ab2b71afcea3aadfef8", "parents": [ "db3b14978abc02041046ed8353f0899cb58ffffc" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Fri Jul 25 01:46:59 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jul 25 10:53:35 2008 -0700" }, "message": "cgroup files: move the release_agent file to use typed handlers\n\nAdds cgroup_release_agent_write() and cgroup_release_agent_show()\nmethods to handle writing/reading the path to a cgroup hierarchy\u0027s\nrelease agent. As a result, cgroup_common_file_read() is now unnecessary.\n\nAs part of the change, a previously-tolerated race in\ncgroup_release_agent() is avoided by copying the current\nrelease_agent_path prior to calling call_usermode_helper().\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nAcked-by: Serge Hallyn \u003cserue@us.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": "db3b14978abc02041046ed8353f0899cb58ffffc", "tree": "355ba027c6354bd13bcbb5318e8478b3de4599e9", "parents": [ "ce16b49d37e748574f7fabc2726268d542d0aa1a" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Fri Jul 25 01:46:58 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jul 25 10:53:35 2008 -0700" }, "message": "cgroup files: add write_string cgroup control file method\n\nThis patch adds a write_string() method for cgroups control files. The\nsemantics are that a buffer is copied from userspace to kernelspace\nand the handler function invoked on that buffer. The buffer is\nguaranteed to be nul-terminated, and no longer than max_write_len\n(defaulting to 64 bytes if unspecified). Later patches will convert\nexisting raw file write handlers in control group subsystems to use\nthis method.\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nAcked-by: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nAcked-by: Serge Hallyn \u003cserue@us.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": "ce16b49d37e748574f7fabc2726268d542d0aa1a", "tree": "dc079efb12661484efa7dd2659a38cfd7b3f3fd5", "parents": [ "8947f9d5b361ce927be6d5c11fed57905b7a4100" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Fri Jul 25 01:46:57 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jul 25 10:53:35 2008 -0700" }, "message": "cgroup files: clean up whitespace in struct cftype\n\nThis patch removes some extraneous spaces from method declarations in\nstruct cftype, to fit in with conventional kernel style.\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Serge Hallyn \u003cserue@us.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": "cf475ad28ac35cc9ba612d67158f29b73b38b05d", "tree": "2c7cd568d00357bd42643ea602884e731cc24f26", "parents": [ "29486df325e1fe6e1764afcb19e3370804c2b002" ], "author": { "name": "Balbir Singh", "email": "balbir@linux.vnet.ibm.com", "time": "Tue Apr 29 01:00:16 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Apr 29 08:06:10 2008 -0700" }, "message": "cgroups: add an owner to the mm_struct\n\nRemove the mem_cgroup member from mm_struct and instead adds an owner.\n\nThis approach was suggested by Paul Menage. The advantage of this approach\nis that, once the mm-\u003eowner is known, using the subsystem id, the cgroup\ncan be determined. It also allows several control groups that are\nvirtually grouped by mm_struct, to exist independent of the memory\ncontroller i.e., without adding mem_cgroup\u0027s for each controller, to\nmm_struct.\n\nA new config option CONFIG_MM_OWNER is added and the memory resource\ncontroller selects this config option.\n\nThis patch also adds cgroup callbacks to notify subsystems when mm-\u003eowner\nchanges. The mm_cgroup_changed callback is called with the task_lock() of\nthe new task held and is called just prior to changing the mm-\u003eowner.\n\nI am indebted to Paul Menage for the several reviews of this patchset and\nhelping me make it lighter and simpler.\n\nThis patch was tested on a powerpc box, it was compiled with both the\nMM_OWNER config turned on and off.\n\nAfter the thread group leader exits, it\u0027s moved to init_css_state by\ncgroup_exit(), thus all future charges from runnings threads would be\nredirected to the init_css_set\u0027s subsystem.\n\nSigned-off-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Pavel Emelianov \u003cxemul@openvz.org\u003e\nCc: Hugh Dickins \u003chugh@veritas.com\u003e\nCc: Sudhir Kumar \u003cskumar@linux.vnet.ibm.com\u003e\nCc: YAMAMOTO Takashi \u003cyamamoto@valinux.co.jp\u003e\nCc: Hirokazu Takahashi \u003ctaka@valinux.co.jp\u003e\nCc: David Rientjes \u003crientjes@google.com\u003e,\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nAcked-by: Pekka Enberg \u003cpenberg@cs.helsinki.fi\u003e\nReviewed-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Oleg Nesterov \u003coleg@tv-sign.ru\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "29486df325e1fe6e1764afcb19e3370804c2b002", "tree": "d69a96bb829940f3ae5171fde481edb20a9e468a", "parents": [ "28fd5dfc12bde391981dfdcf20755952b6e916af" ], "author": { "name": "Serge E. Hallyn", "email": "serue@us.ibm.com", "time": "Tue Apr 29 01:00:14 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Apr 29 08:06:10 2008 -0700" }, "message": "cgroups: introduce cft-\u003eread_seq()\n\nIntroduce a read_seq() helper in cftype, which uses seq_file to print out\nlists. Use it in the devices cgroup. Also split devices.allow into two\nfiles, so now devices.deny and devices.allow are the ones to use to manipulate\nthe whitelist, while devices.list outputs the cgroup\u0027s current whitelist.\n\nSigned-off-by: Serge E. Hallyn \u003cserue@us.ibm.com\u003e\nAcked-by: Paul Menage \u003cmenage@google.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": "28fd5dfc12bde391981dfdcf20755952b6e916af", "tree": "ce0637c3f8c04d802b467cd518a9d0c076f1eeea", "parents": [ "e8d55fdeb882cfcb5e8db5a5ce16edfba78aafc5" ], "author": { "name": "Li Zefan", "email": "lizf@cn.fujitsu.com", "time": "Tue Apr 29 01:00:13 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Apr 29 08:06:10 2008 -0700" }, "message": "cgroups: remove the css_set linked-list\n\nNow we can run through the hash table instead of running through the\nlinked-list.\n\nSigned-off-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nReviewed-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "472b1053f3c319cc60bfb2a0bb062fed77a93eb6", "tree": "2ef88bfdb7e397d3718a1bed38f13194f894097e", "parents": [ "08ce5f16ee466ffc5bf243800deeecd77d9eaf50" ], "author": { "name": "Li Zefan", "email": "lizf@cn.fujitsu.com", "time": "Tue Apr 29 01:00:11 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Apr 29 08:06:09 2008 -0700" }, "message": "cgroups: use a hash table for css_set finding\n\nWhen we attach a process to a different cgroup, the css_set linked-list will\nbe run through to find a suitable existing css_set to use. This patch\nimplements a hash table for better performance.\n\nThe following benchmarks have been tested:\n\nFor N in 1, 5, 10, 50, 100, 500, 1000, create N cgroups with one sleeping\ntask in each, and then move an additional task through each cgroup in\nturn.\n\nHere is a test result:\n\nN\tLoop\torig - Time(s)\thash - Time(s)\n----------------------------------------------\n1\t10000\t1.201231728\t1.196311177\n5\t2000\t1.065743872\t1.040566424\n10\t1000\t0.991054735\t0.986876440\n50\t200\t0.976554203\t0.969608733\n100\t100\t0.998504680\t0.969218270\n500\t20\t1.157347764\t0.962602963\n1000\t10\t1.619521852\t1.085140172\n\nSigned-off-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nReviewed-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "d447ea2f30ec60370ddb99a668e5ac12995f043d", "tree": "f5d53c03d32323f3e916237f8ec0e0b2572f93b0", "parents": [ "46ae220bea40bd1cf4abec2d5cdfb4f9396c7115" ], "author": { "name": "Pavel Emelyanov", "email": "xemul@openvz.org", "time": "Tue Apr 29 01:00:08 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Apr 29 08:06:09 2008 -0700" }, "message": "cgroups: add the trigger callback to struct cftype\n\nTrigger callback can be used to receive a kick-up from the user space. The\nstring written is ignored.\n\nThe cftype-\u003eprivate is used for multiplexing events.\n\nSigned-off-by: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nAcked-by: Paul Menage \u003cmenage@google.com\u003e\nAcked-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "e73d2c61d1fcbd3621688ae457b49509c8d4c601", "tree": "e54b1543e786afd036b7884776a6753d8a8973f6", "parents": [ "418d7d875ce7f33ef0d48d7cc3a95f31302dcf56" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Tue Apr 29 01:00:06 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Apr 29 08:06:09 2008 -0700" }, "message": "CGroups _s64 files: add cgroups read_s64/write_s64 file methods\n\nThese patches add cgroups read_s64 and write_s64 control file methods (the\nsigned equivalent of read_u64/write_u64) and use them to implement the\ncpu.rt_runtime_us control file in the CFS cgroup subsystem.\n\nThis patch:\n\nThese are the signed equivalents of the read_u64/write_u64 methods\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nAcked-by: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nCc: Ingo Molnar \u003cmingo@elte.hu\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "3116f0e3df0a67ad56f15dd4c5f6cefb04bb4a98", "tree": "27dc8e1000eb7c3e36157963415a0bc8b8555e44", "parents": [ "c27e8818a09bbdfe7c07c629cb2c27e1a742e156" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Tue Apr 29 01:00:04 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Apr 29 08:06:09 2008 -0700" }, "message": "CGroup API files: move \"releasable\" to cgroup_debug subsystem\n\nThe \"releasable\" control file provided by the cgroup framework exports the\nstate of a per-cgroup flag that\u0027s related to the notify-on-release feature.\nThis isn\u0027t really generally useful, unless you\u0027re trying to debug this\nparticular feature of cgroups.\n\nThis patch moves the \"releasable\" file to the cgroup_debug subsystem.\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nCc: \"Li Zefan\" \u003clizf@cn.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.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": "9179656961adcea3c25403365597e486d851ac5e", "tree": "13acbdd7ae045ed46fa679e613783717eb3b9138", "parents": [ "700fe1ab99240c1a9c4d155e2a0612a1b044bb69" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Tue Apr 29 01:00:01 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Apr 29 08:06:08 2008 -0700" }, "message": "CGroup API files: add cgroup map data type\n\nAdds a new type of supported control file representation, a map from strings\nto u64 values.\n\nEach map entry is printed as a line in a similar format to /proc/vmstat, i.e.\n\"$key $value\\n\"\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nCc: \"Li Zefan\" \u003clizf@cn.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.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": "f4c753b7eacc277e506066abdda351cbc1cf8e6a", "tree": "2510264d4fac836dc64acd700846a5fe8d960b1e", "parents": [ "3ff31d0cca38b3c20e88a022bf38c4f7c98492f0" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Tue Apr 29 00:59:56 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Apr 29 08:06:07 2008 -0700" }, "message": "CGroup API files: rename read/write_uint methods to read_write_u64\n\nSeveral people have justifiably complained that the \"_uint\" suffix is\ninappropriate for functions that handle u64 values, so this patch just renames\nall these functions and their users to have the suffic _u64.\n\n[peterz@infradead.org: build fix]\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nCc: \"Li Zefan\" \u003clizf@cn.fujitsu.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: \"YAMAMOTO Takashi\" \u003cyamamoto@valinux.co.jp\u003e\nSigned-off-by: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "8bab8dded67d026c39367bbd5e27d2f6c556c38e", "tree": "d80f8f85f1da496c56bfa8575f0b59eba7c2ef55", "parents": [ "3a143125ddc4e2e0ca1e67fb4bedd45c36e59cc7" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Fri Apr 04 14:29:57 2008 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Apr 04 14:46:26 2008 -0700" }, "message": "cgroups: add cgroup support for enabling controllers at boot time\n\nThe effects of cgroup_disable\u003dfoo are:\n\n- foo isn\u0027t auto-mounted if you mount all cgroups in a single hierarchy\n- foo isn\u0027t visible as an individually mountable subsystem\n\nAs a result there will only ever be one call to foo-\u003ecreate(), at init time;\nall processes will stay in this group, and the group will never be mounted on\na visible hierarchy. Any additional effects (e.g. not allocating metadata)\nare up to the foo subsystem.\n\nThis doesn\u0027t handle early_init subsystems (their \"disabled\" bit isn\u0027t set be,\nbut it could easily be extended to do so if any of the early_init systems\nwanted it - I think it would just involve some nastier parameter processing\nsince it would occur before the command-line argument parser had been run.\n\nHugh said:\n\n Ballpark figures, I\u0027m trying to get this question out rather than\n processing the exact numbers: CONFIG_CGROUP_MEM_RES_CTLR adds 15% overhead\n to the affected paths, booting with cgroup_disable\u003dmemory cuts that back to\n 1% overhead (due to slightly bigger struct page).\n\n I\u0027m no expert on distros, they may have no interest whatever in\n CONFIG_CGROUP_MEM_RES_CTLR\u003dy; and the rest of us can easily build with or\n without it, or apply the cgroup_disable\u003dmemory patches.\n\nUnix bench\u0027s execl test result on x86_64 was\n\n\u003d\u003d just after boot without mounting any cgroup fs.\u003d\u003d\nmem_cgorup\u003doff : Execl Throughput 43.0 3150.1 732.6\nmem_cgroup\u003don : Execl Throughput 43.0 2932.6 682.0\n\u003d\u003d\n\n[lizf@cn.fujitsu.com: fix boot option parsing]\nSigned-off-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Pavel Emelyanov \u003cxemul@openvz.org\u003e\nCc: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: Hugh Dickins \u003chugh@veritas.com\u003e\nCc: Sudhir Kumar \u003cskumar@linux.vnet.ibm.com\u003e\nCc: YAMAMOTO Takashi \u003cyamamoto@valinux.co.jp\u003e\nCc: David Rientjes \u003crientjes@google.com\u003e\nSigned-off-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "ffd2d883399cbbb641e55730676ce1ec4845d99d", "tree": "9c8d7aa567c33214bc20b0a78be0abfb0e782bfa", "parents": [ "a043e3b2c63445512c5592cbe3c8694f3c655e81" ], "author": { "name": "Li Zefan", "email": "lizf@cn.fujitsu.com", "time": "Sat Feb 23 15:24:09 2008 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Sat Feb 23 17:13:24 2008 -0800" }, "message": "cgroup: clean up cgroup.h\n\n- replace old name \u0027cont\u0027 with \u0027cgrp\u0027 (Paul Menage did this cleanup for\n cgroup.c in commit bd89aabc6761de1c35b154fe6f914a445d301510)\n- remove a duplicate declaration of cgroup_path()\n\nSigned-off-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nAcked-by: Paul Menage \u003cmenage@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "a043e3b2c63445512c5592cbe3c8694f3c655e81", "tree": "abfc3d3f475c32c7df14ada9fc4461b731628bd2", "parents": [ "d19e0583300da82a6e27cd2116f558048502edaa" ], "author": { "name": "Li Zefan", "email": "lizf@cn.fujitsu.com", "time": "Sat Feb 23 15:24:09 2008 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Sat Feb 23 17:13:24 2008 -0800" }, "message": "cgroup: fix comments\n\nfix:\n- comments about need_forkexit_callback\n- comments about release agent\n- typo and comment style, etc.\n\nSigned-off-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nAcked-by: Paul Menage \u003cmenage@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "956db3ca0606e78456786ef19fd4dc7a5151a6e1", "tree": "0bef3d107df1115ecf76e342f30ecee67a7f3705", "parents": [ "31a7df01fd0cd786f60873a921aecafac148c290" ], "author": { "name": "Cliff Wickman", "email": "cpw@sgi.com", "time": "Thu Feb 07 00:14:43 2008 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Thu Feb 07 08:42:22 2008 -0800" }, "message": "hotplug cpu: move tasks in empty cpusets to parent\n\nThis patch corrects a situation that occurs when one disables all the cpus in\na cpuset.\n\nCurrently, the disabled (cpu-less) cpuset inherits the cpus of its parent,\nwhich is incorrect because it may then overlap its cpu-exclusive sibling.\n\nTasks of an empty cpuset should be moved to the cpuset which is the parent of\ntheir current cpuset. Or if the parent cpuset has no cpus, to its parent,\netc.\n\nAnd the empty cpuset should be released (if it is flagged notify_on_release).\n\nDepends on the cgroup_scan_tasks() function (proposed by David Rientjes) to\niterate through all tasks in the cpu-less cpuset. We are deliberately\navoiding a walk of the tasklist.\n\n[akpm@linux-foundation.org: coding-style fixes]\nSigned-off-by: Cliff Wickman \u003ccpw@sgi.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nCc: David Rientjes \u003crientjes@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "31a7df01fd0cd786f60873a921aecafac148c290", "tree": "221f00c864c50e7dc4719cb4de09292040567c55", "parents": [ "dfc05c259e424e4160c66eab728f55cc4b53fd75" ], "author": { "name": "Cliff Wickman", "email": "cpw@sgi.com", "time": "Thu Feb 07 00:14:42 2008 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Thu Feb 07 08:42:22 2008 -0800" }, "message": "cgroups: mechanism to process each task in a cgroup\n\nProvide cgroup_scan_tasks(), which iterates through every task in a cgroup,\ncalling a test function and a process function for each. And call the process\nfunction without holding the css_set_lock lock.\n\nThe idea is David Rientjes\u0027, predicting that such a function will make it much\neasier in the future to extend things that require access to each task in a\ncgroup without holding the lock,\n\n[akpm@linux-foundation.org: cleanup]\n[akpm@linux-foundation.org: coding-style fixes]\nSigned-off-by: Cliff Wickman \u003ccpw@sgi.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nAcked-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": "4fca88c87b7969c698912e2de9b1b31088c777cb", "tree": "6a086aaffb4ec17eadd2f2427c5816a02be6606f", "parents": [ "d2ceb9b7ddedbb2e8e590bc6ce33c854043016f9" ], "author": { "name": "KAMEZAWA Hiroyuki", "email": "kamezawa.hiroyu@jp.fujitsu.com", "time": "Thu Feb 07 00:14:27 2008 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Thu Feb 07 08:42:20 2008 -0800" }, "message": "memory cgroup enhancements: add- pre_destroy() handler\n\nAdd a handler \"pre_destroy\" to cgroup_subsys. It is called before\ncgroup_rmdir() checks all subsys\u0027s refcnt.\n\nI think this is useful for subsys which have some extra refs even if there\nare no tasks in cgroup. By adding pre_destroy(), the kernel keeps the rule\n\"destroy() against subsystem is called only when refcnt\u003d0.\" and allows css\nref to be used by other objects than tasks.\n\nSigned-off-by: KAMEZAWA Hiroyuki \u003ckamezawa.hiroyu@jp.fujitsu.com\u003e\nCc: \"Eric W. Biederman\" \u003cebiederm@xmission.com\u003e\nCc: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: David Rientjes \u003crientjes@google.com\u003e\nCc: Herbert Poetzl \u003cherbert@13thfloor.at\u003e\nCc: Kirill Korotaev \u003cdev@sw.ru\u003e\nCc: Nick Piggin \u003cnickpiggin@yahoo.com.au\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Pavel Emelianov \u003cxemul@openvz.org\u003e\nCc: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nCc: Vaidyanathan Srinivasan \u003csvaidy@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": "846c7bb055747989891f5cd2bb6e8d56243ba1e7", "tree": "e044041366efa8298157c4ae86615d68d30dd6d2", "parents": [ "c2e2c7fa1cb2cf2b114a6c9bc132b6601db5a7c8" ], "author": { "name": "Balbir Singh", "email": "balbir@linux.vnet.ibm.com", "time": "Thu Oct 18 23:39:44 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Fri Oct 19 11:53:36 2007 -0700" }, "message": "Add cgroupstats\n\nThis patch is inspired by the discussion at\nhttp://lkml.org/lkml/2007/4/11/187 and implements per cgroup statistics\nas suggested by Andrew Morton in http://lkml.org/lkml/2007/4/11/263. The\npatch is on top of 2.6.21-mm1 with Paul\u0027s cgroups v9 patches (forward\nported)\n\nThis patch implements per cgroup statistics infrastructure and re-uses\ncode from the taskstats interface. A new set of cgroup operations are\nregistered with commands and attributes. It should be very easy to\n*extend* per cgroup statistics, by adding members to the cgroupstats\nstructure.\n\nThe current model for cgroupstats is a pull, a push model (to post\nstatistics on interesting events), should be very easy to add. Currently\nuser space requests for statistics by passing the cgroup file\ndescriptor. Statistics about the state of all the tasks in the cgroup\nis returned to user space.\n\nTODO\u0027s/NOTE:\n\nThis patch provides an infrastructure for implementing cgroup statistics.\nBased on the needs of each controller, we can incrementally add more statistics,\nevent based support for notification of statistics, accumulation of taskstats\ninto cgroup statistics in the future.\n\nSample output\n\n# ./cgroupstats -C /cgroup/a\nsleeping 2, blocked 0, running 1, stopped 0, uninterruptible 0\n\n# ./cgroupstats -C /cgroup/\nsleeping 154, blocked 0, running 0, stopped 0, uninterruptible 0\n\nIf the approach looks good, I\u0027ll enhance and post the user space utility for\nthe same\n\nFeedback, comments, test results are always welcome!\n\n[akpm@linux-foundation.org: build fix]\nSigned-off-by: Balbir Singh \u003cbalbir@linux.vnet.ibm.com\u003e\nCc: Paul Menage \u003cmenage@google.com\u003e\nCc: Jay Lan \u003cjlan@engr.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": "81a6a5cdd2c5cd70874b88afe524ab09e9e869af", "tree": "ba46c47a0692b687a96e52e61bfda4f14457017f", "parents": [ "817929ec274bcfe771586d338bb31d1659615686" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Thu Oct 18 23:39:38 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Fri Oct 19 11:53:36 2007 -0700" }, "message": "Task Control Groups: automatic userspace notification of idle cgroups\n\nAdd the following files to the cgroup filesystem:\n\nnotify_on_release - configures/reports whether the cgroup subsystem should\nattempt to run a release script when this cgroup becomes unused\n\nrelease_agent - configures/reports the release agent to be used for this\nhierarchy (top level in each hierarchy only)\n\nreleasable - reports whether this cgroup would have been auto-released if\nnotify_on_release was true and a release agent was configured (mainly useful\nfor debugging)\n\nTo avoid locking issues, invoking the userspace release agent is done via a\nworkqueue task; cgroups that need to have their release agents invoked by\nthe workqueue task are linked on to a list.\n\n[pj@sgi.com: Need to include kmod.h]\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Serge E. Hallyn \u003cserue@us.ibm.com\u003e\nCc: \"Eric W. Biederman\" \u003cebiederm@xmission.com\u003e\nCc: Dave Hansen \u003chaveblue@us.ibm.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nCc: Kirill Korotaev \u003cdev@openvz.org\u003e\nCc: Herbert Poetzl \u003cherbert@13thfloor.at\u003e\nCc: Srivatsa Vaddagiri \u003cvatsa@in.ibm.com\u003e\nCc: Cedric Le Goater \u003cclg@fr.ibm.com\u003e\nSigned-off-by: Paul Jackson \u003cpj@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": "817929ec274bcfe771586d338bb31d1659615686", "tree": "5a96ed1afd308016e8720437a00bf2f114e907cb", "parents": [ "a424316ca154317367c7ddf89997d1c80e4a8051" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Thu Oct 18 23:39:36 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Fri Oct 19 11:53:36 2007 -0700" }, "message": "Task Control Groups: shared cgroup subsystem group arrays\n\nReplace the struct css_set embedded in task_struct with a pointer; all tasks\nthat have the same set of memberships across all hierarchies will share a\ncss_set object, and will be linked via their css_sets field to the \"tasks\"\nlist_head in the css_set.\n\nAssuming that many tasks share the same cgroup assignments, this reduces\noverall space usage and keeps the size of the task_struct down (three pointers\nadded to task_struct compared to a non-cgroups kernel, no matter how many\nsubsystems are registered).\n\n[akpm@linux-foundation.org: fix a printk]\n[akpm@linux-foundation.org: build fix]\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Serge E. Hallyn \u003cserue@us.ibm.com\u003e\nCc: \"Eric W. Biederman\" \u003cebiederm@xmission.com\u003e\nCc: Dave Hansen \u003chaveblue@us.ibm.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nCc: Kirill Korotaev \u003cdev@openvz.org\u003e\nCc: Herbert Poetzl \u003cherbert@13thfloor.at\u003e\nCc: Srivatsa Vaddagiri \u003cvatsa@in.ibm.com\u003e\nCc: Cedric Le Goater \u003cclg@fr.ibm.com\u003e\nCc: Serge E. Hallyn \u003cserue@us.ibm.com\u003e\nCc: \"Eric W. Biederman\" \u003cebiederm@xmission.com\u003e\nCc: Dave Hansen \u003chaveblue@us.ibm.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nCc: Kirill Korotaev \u003cdev@openvz.org\u003e\nCc: Herbert Poetzl \u003cherbert@13thfloor.at\u003e\nCc: Srivatsa Vaddagiri \u003cvatsa@in.ibm.com\u003e\nCc: Cedric Le Goater \u003cclg@fr.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": "a424316ca154317367c7ddf89997d1c80e4a8051", "tree": "ed349926c41aad5be6d62c9074ff72a0d9ac32c2", "parents": [ "697f41610863c9264a7ae26dac9a387c9dda8c84" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Thu Oct 18 23:39:35 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Fri Oct 19 11:53:36 2007 -0700" }, "message": "Task Control Groups: add procfs interface\n\nAdd:\n\n/proc/cgroups - general system info\n\n/proc/*/cgroup - per-task cgroup membership info\n\n[a.p.zijlstra@chello.nl: cgroups: bdi init hooks]\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Serge E. Hallyn \u003cserue@us.ibm.com\u003e\nCc: \"Eric W. Biederman\" \u003cebiederm@xmission.com\u003e\nCc: Dave Hansen \u003chaveblue@us.ibm.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nCc: Kirill Korotaev \u003cdev@openvz.org\u003e\nCc: Herbert Poetzl \u003cherbert@13thfloor.at\u003e\nCc: Srivatsa Vaddagiri \u003cvatsa@in.ibm.com\u003e\nCc: Cedric Le Goater \u003cclg@fr.ibm.com\u003e\nSigned-off-by: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "697f41610863c9264a7ae26dac9a387c9dda8c84", "tree": "9d4321f93abe36920a8e9d5239d120803f6cabbc", "parents": [ "b4f48b6363c81ca743ef46943ef23fd72e60f679" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Thu Oct 18 23:39:34 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Fri Oct 19 11:53:36 2007 -0700" }, "message": "Task Control Groups: add cgroup_clone() interface\n\nAdd support for cgroup_clone(), a way to create new cgroups intended to\nbe used for systems such as namespace unsharing. A new subsystem callback,\npost_clone(), is added to allow subsystems to automatically configure cloned\ncgroups.\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Serge E. Hallyn \u003cserue@us.ibm.com\u003e\nCc: \"Eric W. Biederman\" \u003cebiederm@xmission.com\u003e\nCc: Dave Hansen \u003chaveblue@us.ibm.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nCc: Kirill Korotaev \u003cdev@openvz.org\u003e\nCc: Herbert Poetzl \u003cherbert@13thfloor.at\u003e\nCc: Srivatsa Vaddagiri \u003cvatsa@in.ibm.com\u003e\nCc: Cedric Le Goater \u003cclg@fr.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": "b4f48b6363c81ca743ef46943ef23fd72e60f679", "tree": "40437b78e2d7a7d9d71e7bd63bc96e1ad02daa94", "parents": [ "355e0c48b757b7fcc79ccb98fda8105ed37a1598" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Thu Oct 18 23:39:33 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Fri Oct 19 11:53:36 2007 -0700" }, "message": "Task Control Groups: add fork()/exit() hooks\n\nThis adds the necessary hooks to the fork() and exit() paths to ensure\nthat new children inherit their parent\u0027s cgroup assignments, and that\nexiting processes release reference counts on their cgroups.\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Serge E. Hallyn \u003cserue@us.ibm.com\u003e\nCc: \"Eric W. Biederman\" \u003cebiederm@xmission.com\u003e\nCc: Dave Hansen \u003chaveblue@us.ibm.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nCc: Kirill Korotaev \u003cdev@openvz.org\u003e\nCc: Herbert Poetzl \u003cherbert@13thfloor.at\u003e\nCc: Srivatsa Vaddagiri \u003cvatsa@in.ibm.com\u003e\nCc: Cedric Le Goater \u003cclg@fr.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": "355e0c48b757b7fcc79ccb98fda8105ed37a1598", "tree": "f9687961979a808e46620f4bdf9af05de2fd68e2", "parents": [ "bbcb81d09104f0d440974b994c1fc508ccbe9503" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Thu Oct 18 23:39:33 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Fri Oct 19 11:53:36 2007 -0700" }, "message": "Add cgroup write_uint() helper method\n\nAdd write_uint() helper method for cgroup subsystems\n\nThis helper is analagous to the read_uint() helper method for\nreporting u64 values to userspace. It\u0027s designed to reduce the amount\nof boilerplate requierd for creating new cgroup subsystems.\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "bbcb81d09104f0d440974b994c1fc508ccbe9503", "tree": "6d9ef3e2c611bb0a8f63519196f7bd7725b7ea1a", "parents": [ "ddbcc7e8e50aefe467c01cac3dec71f118cd8ac2" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Thu Oct 18 23:39:32 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Fri Oct 19 11:53:36 2007 -0700" }, "message": "Task Control Groups: add tasks file interface\n\nAdd the per-directory \"tasks\" file for cgroupfs mounts; this allows the\nuser to determine which tasks are members of a cgroup by reading a\ncgroup\u0027s \"tasks\", and to move a task into a cgroup by writing its pid to\nits \"tasks\".\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Serge E. Hallyn \u003cserue@us.ibm.com\u003e\nCc: \"Eric W. Biederman\" \u003cebiederm@xmission.com\u003e\nCc: Dave Hansen \u003chaveblue@us.ibm.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nCc: Kirill Korotaev \u003cdev@openvz.org\u003e\nCc: Herbert Poetzl \u003cherbert@13thfloor.at\u003e\nCc: Srivatsa Vaddagiri \u003cvatsa@in.ibm.com\u003e\nCc: Cedric Le Goater \u003cclg@fr.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": "ddbcc7e8e50aefe467c01cac3dec71f118cd8ac2", "tree": "0881a031e669582f819d572339e955b04abfc3d2", "parents": [ "55a230aae650157720becc09cadb7d10efbf5013" ], "author": { "name": "Paul Menage", "email": "menage@google.com", "time": "Thu Oct 18 23:39:30 2007 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Fri Oct 19 11:53:36 2007 -0700" }, "message": "Task Control Groups: basic task cgroup framework\n\nGeneric Process Control Groups\n--------------------------\n\nThere have recently been various proposals floating around for\nresource management/accounting and other task grouping subsystems in\nthe kernel, including ResGroups, User BeanCounters, NSProxy\ncgroups, and others. These all need the basic abstraction of being\nable to group together multiple processes in an aggregate, in order to\ntrack/limit the resources permitted to those processes, or control\nother behaviour of the processes, and all implement this grouping in\ndifferent ways.\n\nThis patchset provides a framework for tracking and grouping processes\ninto arbitrary \"cgroups\" and assigning arbitrary state to those\ngroupings, in order to control the behaviour of the cgroup as an\naggregate.\n\nThe intention is that the various resource management and\nvirtualization/cgroup efforts can also become task cgroup\nclients, with the result that:\n\n- the userspace APIs are (somewhat) normalised\n\n- it\u0027s easier to test e.g. the ResGroups CPU controller in\n conjunction with the BeanCounters memory controller, or use either of\nthem as the resource-control portion of a virtual server system.\n\n- the additional kernel footprint of any of the competing resource\n management systems is substantially reduced, since it doesn\u0027t need\n to provide process grouping/containment, hence improving their\n chances of getting into the kernel\n\nThis patch:\n\nAdd the main task cgroups framework - the cgroup filesystem, and the\nbasic structures for tracking membership and associating subsystem state\nobjects to tasks.\n\nSigned-off-by: Paul Menage \u003cmenage@google.com\u003e\nCc: Serge E. Hallyn \u003cserue@us.ibm.com\u003e\nCc: \"Eric W. Biederman\" \u003cebiederm@xmission.com\u003e\nCc: Dave Hansen \u003chaveblue@us.ibm.com\u003e\nCc: Balbir Singh \u003cbalbir@in.ibm.com\u003e\nCc: Paul Jackson \u003cpj@sgi.com\u003e\nCc: Kirill Korotaev \u003cdev@openvz.org\u003e\nCc: Herbert Poetzl \u003cherbert@13thfloor.at\u003e\nCc: Srivatsa Vaddagiri \u003cvatsa@in.ibm.com\u003e\nCc: Cedric Le Goater \u003cclg@fr.ibm.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" } ] }