)]}' { "log": [ { "commit": "dce48a84adf1806676319f6f480e30a6daa012f9", "tree": "79151f5d31d9c3dcdc723ab8877cb943b944890e", "parents": [ "2ff799d3cff1ecb274049378b28120ee5c1c5e5f" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Apr 11 10:43:41 2009 +0200" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri May 15 15:32:45 2009 +0200" }, "message": "sched, timers: move calc_load() to scheduler\n\nDimitri Sivanich noticed that xtime_lock is held write locked across\ncalc_load() which iterates over all online CPUs. That can cause long\nlatencies for xtime_lock readers on large SMP systems. \n\nThe load average calculation is an rough estimate anyway so there is\nno real need to protect the readers vs. the update. It\u0027s not a problem\nwhen the avenrun array is updated while a reader copies the values.\n\nInstead of iterating over all online CPUs let the scheduler_tick code\nupdate the number of active tasks shortly before the avenrun update\nhappens. The avenrun update itself is handled by the CPU which calls\ndo_timer().\n\n[ Impact: reduce xtime_lock write locked section ]\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nAcked-by: Peter Zijlstra \u003cpeterz@infradead.org\u003e\n" }, { "commit": "4ce72a2c063a7fa8e42a9435440ae3364115a58d", "tree": "427b25960945524a205c7d8ed20ed5f02325175e", "parents": [ "0c4b83da58ec2e96ce9c44c211d6eac5f9dae478" ], "author": { "name": "Li Zefan", "email": "lizf@cn.fujitsu.com", "time": "Wed Oct 22 15:25:26 2008 +0800" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Wed Oct 22 10:01:52 2008 +0200" }, "message": "sched: add CONFIG_SMP consistency\n\na patch from Henrik Austad did this:\n\n\u003e\u003e Do not declare select_task_rq as part of sched_class when CONFIG_SMP is\n\u003e\u003e not set.\n\nPeter observed:\n\n\u003e While a proper cleanup, could you do it by re-arranging the methods so\n\u003e as to not create an additional ifdef?\n\nDo not declare select_task_rq and some other methods as part of sched_class\nwhen CONFIG_SMP is not set.\n\nAlso gather those methods to avoid CONFIG_SMP mess.\n\nIdea-by: Henrik Austad \u003chenrik.austad@gmail.com\u003e\nSigned-off-by: Li Zefan \u003clizf@cn.fujitsu.com\u003e\nAcked-by: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nAcked-by: Henrik Austad \u003chenrik@austad.us\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "15afe09bf496ae10c989e1a375a6b5da7bd3e16e", "tree": "4565659d1084e357eea42e6321a4d304ac950faa", "parents": [ "09b22a2f678ae733801b888c44756d0abd686b8a" ], "author": { "name": "Peter Zijlstra", "email": "a.p.zijlstra@chello.nl", "time": "Sat Sep 20 23:38:02 2008 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon Sep 22 16:28:32 2008 +0200" }, "message": "sched: wakeup preempt when small overlap\n\nLin Ming reported a 10% OLTP regression against 2.6.27-rc4.\n\nThe difference seems to come from different preemption agressiveness,\nwhich affects the cache footprint of the workload and its effective\ncache trashing.\n\nAggresively preempt a task if its avg overlap is very small, this should\navoid the task going to sleep and find it still running when we schedule\nback to it - saving a wakeup.\n\nReported-by: Lin Ming \u003cming.m.lin@intel.com\u003e\nSigned-off-by: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "2abdad0a4cd8f9413f778cc998e0ee7d60b28417", "tree": "c7bd3839e765c2991d3dfbfb1f5f83d58f1211e3", "parents": [ "e05510d01ad1565e5e086a939261084d67ba2b10" ], "author": { "name": "Harvey Harrison", "email": "harvey.harrison@gmail.com", "time": "Fri Apr 25 10:53:13 2008 -0700" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon May 05 23:56:17 2008 +0200" }, "message": "sched: make rt_sched_class, idle_sched_class static\n\nThe C files are included directly in sched.c, so they are\neffectively static.\n\nSigned-off-by: Harvey Harrison \u003charvey.harrison@gmail.com\u003e\nAcked-by: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "8f4d37ec073c17e2d4aa8851df5837d798606d6f", "tree": "a9ac9063eca53e4d0110e8086f55241ea70ba993", "parents": [ "02b67cc3ba36bdba351d6c3a00593f4ec550d9d3" ], "author": { "name": "Peter Zijlstra", "email": "a.p.zijlstra@chello.nl", "time": "Fri Jan 25 21:08:29 2008 +0100" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Fri Jan 25 21:08:29 2008 +0100" }, "message": "sched: high-res preemption tick\n\nUse HR-timers (when available) to deliver an accurate preemption tick.\n\nThe regular scheduler tick that runs at 1/HZ can be too coarse when nice\nlevel are used. The fairness system will still keep the cpu utilisation \u0027fair\u0027\nby then delaying the task that got an excessive amount of CPU time but try to\nminimize this by delivering preemption points spot-on.\n\nThe average frequency of this extra interrupt is sched_latency / nr_latency.\nWhich need not be higher than 1/HZ, its just that the distribution within the\nsched_latency period is important.\n\nSigned-off-by: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "cb46984504048db946cd551c261df4e70d59a8ea", "tree": "e07343cc5967f74370c6b0290b67a225d868a99d", "parents": [ "9a897c5a6701bcb6f099f7ca20194999102729fd" ], "author": { "name": "Steven Rostedt", "email": "rostedt@goodmis.org", "time": "Fri Jan 25 21:08:22 2008 +0100" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Fri Jan 25 21:08:22 2008 +0100" }, "message": "sched: RT-balance, add new methods to sched_class\n\nDmitry Adamushko found that the current implementation of the RT\nbalancing code left out changes to the sched_setscheduler and\nrt_mutex_setprio.\n\nThis patch addresses this issue by adding methods to the schedule classes\nto handle being switched out of (switched_from) and being switched into\n(switched_to) a sched_class. Also a method for changing of priorities\nis also added (prio_changed).\n\nThis patch also removes some duplicate logic between rt_mutex_setprio and\nsched_setscheduler.\n\nSigned-off-by: Steven Rostedt \u003csrostedt@redhat.com\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "e7693a362ec84bb5b6fd441d8a8b4b9d568a7a0c", "tree": "078940540641a59aaf199695bfc6de3f062a987b", "parents": [ "697f0a487f294e634a342764472b79375bb3158a" ], "author": { "name": "Gregory Haskins", "email": "ghaskins@novell.com", "time": "Fri Jan 25 21:08:09 2008 +0100" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Fri Jan 25 21:08:09 2008 +0100" }, "message": "sched: de-SCHED_OTHER-ize the RT path\n\nThe current wake-up code path tries to determine if it can optimize the\nwake-up to \"this_cpu\" by computing load calculations. The problem is that\nthese calculations are only relevant to SCHED_OTHER tasks where load is king.\nFor RT tasks, priority is king. So the load calculation is completely wasted\nbandwidth.\n\nTherefore, we create a new sched_class interface to help with\npre-wakeup routing decisions and move the load calculation as a function\nof CFS task\u0027s class.\n\nSigned-off-by: Gregory Haskins \u003cghaskins@novell.com\u003e\nSigned-off-by: Steven Rostedt \u003csrostedt@redhat.com\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "681f3e68541d6f03e3e05d21fe15093578b8b539", "tree": "82e13a7b15a2bb15233ba964dd1eb352b007cae4", "parents": [ "e1d1484f72127a5580d37c379f6a5b2c2786434c" ], "author": { "name": "Peter Williams", "email": "pwil3058@bigpond.net.au", "time": "Wed Oct 24 18:23:51 2007 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Wed Oct 24 18:23:51 2007 +0200" }, "message": "sched: isolate SMP balancing code a bit more\n\nAt the moment, a lot of load balancing code that is irrelevant to non\nSMP systems gets included during non SMP builds.\n\nThis patch addresses this issue and reduces the binary size on non\nSMP systems:\n\n text data bss dec hex filename\n 10983 28 1192 12203 2fab sched.o.before\n 10739 28 1192 11959 2eb7 sched.o.after\n\nSigned-off-by: Peter Williams \u003cpwil3058@bigpond.net.au\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "e1d1484f72127a5580d37c379f6a5b2c2786434c", "tree": "e3e6529c5b9079f35b2c60bbd346a3c51c2b2c13", "parents": [ "a0f846aa76c3e03d54c1700a87cab3a46ccd71e2" ], "author": { "name": "Peter Williams", "email": "pwil3058@bigpond.net.au", "time": "Wed Oct 24 18:23:51 2007 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Wed Oct 24 18:23:51 2007 +0200" }, "message": "sched: reduce balance-tasks overhead\n\nAt the moment, balance_tasks() provides low level functionality for both\n move_tasks() and move_one_task() (indirectly) via the load_balance()\nfunction (in the sched_class interface) which also provides dual\nfunctionality. This dual functionality complicates the interfaces and\ninternal mechanisms and makes the run time overhead of operations that\nare called with two run queue locks held.\n\nThis patch addresses this issue and reduces the overhead of these\noperations.\n\nSigned-off-by: Peter Williams \u003cpwil3058@bigpond.net.au\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "5522d5d5f70005faeffff3ffc0cfa8eec0155de4", "tree": "d05fc41d19b8a70102a0ad3ad7ed8f6fd2d1cf7d", "parents": [ "b9fa3df33f9166daf81bfa8253d339f5a7726122" ], "author": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon Oct 15 17:00:12 2007 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon Oct 15 17:00:12 2007 +0200" }, "message": "sched: mark scheduling classes as const\n\nmark scheduling classes as const. The speeds up the code\na bit and shrinks it:\n\n text data bss dec hex filename\n 40027 4018 292 44337 ad31 sched.o.before\n 40190 3842 292 44324 ad24 sched.o.after\n\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\nReviewed-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "83b699ed20f5218580a1b7042064082e2e05f8c5", "tree": "544c42283f41065f22e96e02af0badb7ed70a371", "parents": [ "edcb60a309769a5f6e7c9e76d7c98b34d1757448" ], "author": { "name": "Srivatsa Vaddagiri", "email": "vatsa@linux.vnet.ibm.com", "time": "Mon Oct 15 17:00:08 2007 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon Oct 15 17:00:08 2007 +0200" }, "message": "sched: revert recent removal of set_curr_task()\n\nRevert removal of set_curr_task.\nUse put_prev_task/set_curr_task when changing groups/policies\n\nSigned-off-by: Srivatsa Vaddagiri \u003c vatsa@linux.vnet.ibm.com\u003e\nSigned-off-by: Dhaval Giani \u003cdhaval@linux.vnet.ibm.com\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\nSigned-off-by: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\n" }, { "commit": "f6b53205e17c8ca481c69ed579a35a650a4b481a", "tree": "d7820281757e3a2c7cfff7e38275261dcd077eab", "parents": [ "4530d7ab0fb8d5056b68c376949e2d5c4db7817e" ], "author": { "name": "Dmitry Adamushko", "email": "dmitry.adamushko@gmail.com", "time": "Mon Oct 15 17:00:08 2007 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon Oct 15 17:00:08 2007 +0200" }, "message": "sched: rework enqueue/dequeue_entity() to get rid of set_curr_task()\n\nrework enqueue/dequeue_entity() to get rid of \nsched_class::set_curr_task(). This simplifies sched_setscheduler(), \nrt_mutex_setprio() and sched_move_tasks().\n\n text data bss dec hex filename\n 24330 2734 20 27084 69cc sched.o.before\n 24233 2730 20 26983 6967 sched.o.after\n\nSigned-off-by: Dmitry Adamushko \u003cdmitry.adamushko@gmail.com\u003e\nSigned-off-by: Srivatsa Vaddagiri \u003cvatsa@linux.vnet.ibm.com\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\nSigned-off-by: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nReviewed-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\n" }, { "commit": "29f59db3a74b0bdf78a1f5b53ef773caa82692dc", "tree": "5ac877639bac41d3749d08b7f624bd13052e5ec6", "parents": [ "119fe5e06800afc197781ebc8c2d8ca7d03497c8" ], "author": { "name": "Srivatsa Vaddagiri", "email": "vatsa@linux.vnet.ibm.com", "time": "Mon Oct 15 17:00:07 2007 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon Oct 15 17:00:07 2007 +0200" }, "message": "sched: group-scheduler core\n\nAdd interface to control cpu bandwidth allocation to task-groups.\n\n(not yet configurable, due to missing CONFIG_CONTAINERS)\n\nSigned-off-by: Srivatsa Vaddagiri \u003cvatsa@linux.vnet.ibm.com\u003e\nSigned-off-by: Dhaval Giani \u003cdhaval@linux.vnet.ibm.com\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\nSigned-off-by: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\n" }, { "commit": "31ee529cc2254e8b62880535ec8f21a4c5e1c091", "tree": "df9a19be47c3111d0e34ce9f13498b236733a850", "parents": [ "ff95f3df54609d9d4b9572f8a67d09922a645043" ], "author": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Thu Aug 09 11:16:49 2007 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Thu Aug 09 11:16:49 2007 +0200" }, "message": "sched: remove the \u0027u64 now\u0027 parameter from -\u003eput_prev_task()\n\nremove the \u0027u64 now\u0027 parameter from -\u003eput_prev_task().\n\n( identity transformation that causes no change in functionality. )\n\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "fb8d47240246e20f864f0724a23a7220cd1c59ac", "tree": "0c94ecb977355a2152655738bfcee6072c10f41e", "parents": [ "f02231e51a280f1a0fee4d03ad8f50048e06cced" ], "author": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Thu Aug 09 11:16:48 2007 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Thu Aug 09 11:16:48 2007 +0200" }, "message": "sched: remove the \u0027u64 now\u0027 parameter from -\u003epick_next_task()\n\nremove the \u0027u64 now\u0027 parameter from -\u003epick_next_task().\n\n( identity transformation that causes no change in functionality. )\n\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "f02231e51a280f1a0fee4d03ad8f50048e06cced", "tree": "becd8bf748828e3dfe34f73df11d260fd6300026", "parents": [ "fd390f6a04f22fb457d6fd1855964f79536525de" ], "author": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Thu Aug 09 11:16:48 2007 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Thu Aug 09 11:16:48 2007 +0200" }, "message": "sched: remove the \u0027u64 now\u0027 parameter from -\u003edequeue_task()\n\nremove the \u0027u64 now\u0027 parameter from -\u003edequeue_task().\n\n( identity transformation that causes no change in functionality. )\n\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "a4ac01c36e286dd1b9a1d5cd7422c5af51dc55f8", "tree": "0c275d58a4835a3d604d9cac4e1dd7c25714e150", "parents": [ "aea25401c3347d9f3a64ebdc81043be246a9f631" ], "author": { "name": "Peter Williams", "email": "pwil3058@bigpond.net.au", "time": "Thu Aug 09 11:16:46 2007 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Thu Aug 09 11:16:46 2007 +0200" }, "message": "sched: fix bug in balance_tasks()\n\nThere are two problems with balance_tasks() and how it used:\n\n1. The variables best_prio and best_prio_seen (inherited from the old\nmove_tasks()) were only required to handle problems caused by the\nactive/expired arrays, the order in which they were processed and the\npossibility that the task with the highest priority could be on either.\n These issues are no longer present and the extra overhead associated\nwith their use is unnecessary (and possibly wrong).\n\n2. In the absence of CONFIG_FAIR_GROUP_SCHED being set, the same\nthis_best_prio variable needs to be used by all scheduling classes or\nthere is a risk of moving too much load. E.g. if the highest priority\ntask on this at the beginning is a fairly low priority task and the rt\nclass migrates a task (during its turn) then that moved task becomes the\nnew highest priority task on this_rq but when the sched_fair class\ninitializes its copy of this_best_prio it will get the priority of the\noriginal highest priority task as, due to the run queue locks being\nheld, the reschedule triggered by pull_task() will not have taken place.\n This could result in inappropriate overriding of skip_for_load and\nexcessive load being moved.\n\nThe attached patch addresses these problems by deleting all reference to\nbest_prio and best_prio_seen and making this_best_prio a reference\nparameter to the various functions involved.\n\nload_balance_fair() has also been modified so that this_best_prio is\nonly reset (in the loop) if CONFIG_FAIR_GROUP_SCHED is set. This should\npreserve the effect of helping spread groups\u0027 higher priority tasks\naround the available CPUs while improving system performance when\nCONFIG_FAIR_GROUP_SCHED isn\u0027t set.\n\nSigned-off-by: Peter Williams \u003cpwil3058@bigpond.net.au\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "4301065920b0cbde3986519582347e883b166f3e", "tree": "415b8e3a2796709673015e100a3b5b28e556104a", "parents": [ "f1a438d813d416fa9f4be4e6dbd10b54c5938d89" ], "author": { "name": "Peter Williams", "email": "pwil3058@bigpond.net.au", "time": "Thu Aug 09 11:16:46 2007 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Thu Aug 09 11:16:46 2007 +0200" }, "message": "sched: simplify move_tasks()\n\nThe move_tasks() function is currently multiplexed with two distinct\ncapabilities:\n\n1. attempt to move a specified amount of weighted load from one run\nqueue to another; and\n2. attempt to move a specified number of tasks from one run queue to\nanother.\n\nThe first of these capabilities is used in two places, load_balance()\nand load_balance_idle(), and in both of these cases the return value of\nmove_tasks() is used purely to decide if tasks/load were moved and no\nnotice of the actual number of tasks moved is taken.\n\nThe second capability is used in exactly one place,\nactive_load_balance(), to attempt to move exactly one task and, as\nbefore, the return value is only used as an indicator of success or failure.\n\nThis multiplexing of sched_task() was introduced, by me, as part of the\nsmpnice patches and was motivated by the fact that the alternative, one\nfunction to move specified load and one to move a single task, would\nhave led to two functions of roughly the same complexity as the old\nmove_tasks() (or the new balance_tasks()). However, the new modular\ndesign of the new CFS scheduler allows a simpler solution to be adopted\nand this patch addresses that solution by:\n\n1. adding a new function, move_one_task(), to be used by\nactive_load_balance(); and\n2. making move_tasks() a single purpose function that tries to move a\nspecified weighted load and returns 1 for success and 0 for failure.\n\nOne of the consequences of these changes is that neither move_one_task()\nor the new move_tasks() care how many tasks sched_class.load_balance()\nmoves and this enables its interface to be simplified by returning the\namount of load moved as its result and removing the load_moved pointer\nfrom the argument list. This helps simplify the new move_tasks() and\nslightly reduces the amount of work done in each of\nsched_class.load_balance()\u0027s implementations.\n\nFurther simplification, e.g. changes to balance_tasks(), are possible\nbut (slightly) complicated by the special needs of load_balance_fair()\nso I\u0027ve left them to a later patch (if this one gets accepted).\n\nNB Since move_tasks() gets called with two run queue locks held even\nsmall reductions in overhead are worthwhile.\n\n[ mingo@elte.hu ]\n\nthis change also reduces code size nicely:\n\n text data bss dec hex filename\n 39216 3618 24 42858 a76a sched.o.before\n 39173 3618 24 42815 a73f sched.o.after\n\nSigned-off-by: Peter Williams \u003cpwil3058@bigpond.net.au\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" }, { "commit": "fa72e9e484c16f0c9aee23981917d8c8c03f0482", "tree": "2bcf45d83b9738af2b3455df0e3ada975a100306", "parents": [ "bb44e5d1c6b3b748e0facf8f516b3162009feb27" ], "author": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon Jul 09 18:51:58 2007 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@elte.hu", "time": "Mon Jul 09 18:51:58 2007 +0200" }, "message": "sched: cfs core, kernel/sched_idletask.c\n\nadd kernel/sched_idletask.c - which implements the idle thread\nscheduling class. This further simplifies sched.c (under CFS),\nfor example a number of \u0027if (p \u003d\u003d rq-\u003eidle)\u0027 type of special-cases\ncan be removed from sched.c, and schedule() gets simpler too.\n\nSigned-off-by: Ingo Molnar \u003cmingo@elte.hu\u003e\n" } ] }