Signed-off-by: Ciju Rajan K --- diff -Naurp a/Documentation/scheduler/sched-stats.txt b/Documentation/scheduler/sched-stats.txt --- a/Documentation/scheduler/sched-stats.txt 2011-01-17 01:07:47.000000000 +0530 +++ b/Documentation/scheduler/sched-stats.txt 2011-01-17 15:32:05.000000000 +0530 @@ -26,119 +26,81 @@ Note that any such script will necessari reason to change versions is changes in the output format. For those wishing to write their own scripts, the fields are described here. +The first two fields of /proc/schedstat indicates the version (current +version is 16) and jiffies values. The following values are from +cpu & domain statistics. + CPU statistics -------------- -cpu 1 2 3 4 5 6 7 8 9 10 11 12 +The format is like this: + +cpu 1 2 3 4 5 6 7 8 -NOTE: In the sched_yield() statistics, the active queue is considered empty - if it has only one process in it, since obviously the process calling - sched_yield() is that process. - -First four fields are sched_yield() statistics: - 1) # of times both the active and the expired queue were empty - 2) # of times just the active queue was empty - 3) # of times just the expired queue was empty - 4) # of times sched_yield() was called - -Next three are schedule() statistics: - 5) # of times we switched to the expired queue and reused it - 6) # of times schedule() was called - 7) # of times schedule() left the processor idle - -Next two are try_to_wake_up() statistics: - 8) # of times try_to_wake_up() was called - 9) # of times try_to_wake_up() was called to wake up the local cpu - -Next three are statistics describing scheduling latency: - 10) sum of all time spent running by tasks on this processor (in jiffies) - 11) sum of all time spent waiting to run by tasks on this processor (in - jiffies) - 12) # of timeslices run on this cpu +NOTE: In the sched_yield() statistics, the active queue is considered + empty if it has only one process in it, since obviously the + process calling sched_yield() is that process. + + 1) # of times sched_yield() was called on this CPU + 2) # of times scheduler runs on this CPU + 3) # of times scheduler picks idle task as next task on this CPU + 4) # of times try_to_wake_up() is run on this CPU + (Number of times task wakeup is attempted from this CPU) + 5) # of times try_to_wake_up() wakes up a task on the same CPU + (local wakeup) + 6) Time(ns) for which tasks have run on this CPU + 7) Time(ns) for which tasks on this CPU's runqueue have waited + before getting to run on the CPU + 8) # of tasks that have run on this CPU Domain statistics ----------------- -One of these is produced per domain for each cpu described. (Note that if -CONFIG_SMP is not defined, *no* domains are utilized and these lines -will not appear in the output.) +One of these is produced per domain for each cpu described. +(Note that if CONFIG_SMP is not defined, *no* domains are utilized + and these lines will not appear in the output.) -domain 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 +domain 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 The first field is a bit mask indicating what cpus this domain operates over. -The next 24 are a variety of load_balance() statistics in grouped into types -of idleness (idle, busy, and newly idle): +The next 24 are a variety of load_balance() statistics grouped into +types of idleness (idle, busy, and newly idle). The three idle +states are: + +CPU_NEWLY_IDLE: Load balancer is being run on a CPU which is + about to enter IDLE state +CPU_IDLE: This state is entered after CPU_NEWLY_IDLE + state fails to find a new task for this CPU +CPU_NOT_IDLE: Load balancer is being run on a CPU when it is + not in IDLE state (busy times) + +There are eight stats available for each of the three idle states: - 1) # of times in this domain load_balance() was called when the - cpu was idle + 1) # of times in this domain load_balance() was called 2) # of times in this domain load_balance() checked but found - the load did not require balancing when the cpu was idle + the load did not require balancing 3) # of times in this domain load_balance() tried to move one or - more tasks and failed, when the cpu was idle + more tasks and failed 4) sum of imbalances discovered (if any) with each call to - load_balance() in this domain when the cpu was idle - 5) # of times in this domain pull_task() was called when the cpu - was idle + load_balance() in this domain + 5) # of times in this domain pull_task() was called 6) # of times in this domain pull_task() was called even though - the target task was cache-hot when idle + the target task was cache-hot 7) # of times in this domain load_balance() was called but did - not find a busier queue while the cpu was idle - 8) # of times in this domain a busier queue was found while the - cpu was idle but no busier group was found - - 9) # of times in this domain load_balance() was called when the - cpu was busy - 10) # of times in this domain load_balance() checked but found the - load did not require balancing when busy - 11) # of times in this domain load_balance() tried to move one or - more tasks and failed, when the cpu was busy - 12) sum of imbalances discovered (if any) with each call to - load_balance() in this domain when the cpu was busy - 13) # of times in this domain pull_task() was called when busy - 14) # of times in this domain pull_task() was called even though the - target task was cache-hot when busy - 15) # of times in this domain load_balance() was called but did not - find a busier queue while the cpu was busy - 16) # of times in this domain a busier queue was found while the cpu - was busy but no busier group was found - - 17) # of times in this domain load_balance() was called when the - cpu was just becoming idle - 18) # of times in this domain load_balance() checked but found the - load did not require balancing when the cpu was just becoming idle - 19) # of times in this domain load_balance() tried to move one or more - tasks and failed, when the cpu was just becoming idle - 20) sum of imbalances discovered (if any) with each call to - load_balance() in this domain when the cpu was just becoming idle - 21) # of times in this domain pull_task() was called when newly idle - 22) # of times in this domain pull_task() was called even though the - target task was cache-hot when just becoming idle - 23) # of times in this domain load_balance() was called but did not - find a busier queue while the cpu was just becoming idle - 24) # of times in this domain a busier queue was found while the cpu - was just becoming idle but no busier group was found - + not find a busier queue + 8) # of times in this domain a busier queue was found but no + busier group was found + Next three are active_load_balance() statistics: 25) # of times active_load_balance() was called 26) # of times active_load_balance() tried to move a task and failed 27) # of times active_load_balance() successfully moved a task - Next three are sched_balance_exec() statistics: - 28) sbe_cnt is not used - 29) sbe_balanced is not used - 30) sbe_pushed is not used - - Next three are sched_balance_fork() statistics: - 31) sbf_cnt is not used - 32) sbf_balanced is not used - 33) sbf_pushed is not used - - Next three are try_to_wake_up() statistics: - 34) # of times in this domain try_to_wake_up() awoke a task that + Next two are try_to_wake_up() statistics: + 28) # of times in this domain try_to_wake_up() awoke a task that last ran on a different cpu in this domain - 35) # of times in this domain try_to_wake_up() moved a task to the + 29) # of times in this domain try_to_wake_up() moved a task to the waking cpu because it was cache-cold on its own cpu anyway - 36) # of times in this domain try_to_wake_up() started passive balancing /proc//schedstat ----------------