Kernel/linux-next
1
0
Fork 0
mirror of https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git synced 2025-01-15 02:05:33 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

docs: Update Schedstat version to 17

Browse Source 
Update the Schedstat version to 17 as more fields are added to report
different kinds of imbalances in the sched domain. Also domain field
started printing corresponding domain name.

Signed-off-by: Swapnil Sapkal <swapnil.sapkal@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20241220063224.17767-7-swapnil.sapkal@amd.com
This commit is contained in:
Swapnil Sapkal 2024-12-20 06:32:24 +00:00 committed by Peter Zijlstra
parent 011b3a14dc
commit 7c8cd569ff
2 changed files with 77 additions and 53 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

128
Documentation/scheduler/sched-stats.rst
View File

@ -2,6 +2,12 @@
Scheduler Statistics Scheduler Statistics
==================== ====================
Version 17 of schedstats removed 'lb_imbalance' field as it has no
significance anymore and instead added more relevant fields namely
'lb_imbalance_load', 'lb_imbalance_util', 'lb_imbalance_task' and
'lb_imbalance_misfit'. The domain field prints the name of the
corresponding sched domain from this version onwards.
Version 16 of schedstats changed the order of definitions within Version 16 of schedstats changed the order of definitions within
'enum cpu_idle_type', which changed the order of [CPU_MAX_IDLE_TYPES] 'enum cpu_idle_type', which changed the order of [CPU_MAX_IDLE_TYPES]
columns in show_schedstat(). In particular the position of CPU_IDLE columns in show_schedstat(). In particular the position of CPU_IDLE
@ -9,7 +15,9 @@ and __CPU_NOT_IDLE changed places. The size of the array is unchanged.
Version 15 of schedstats dropped counters for some sched_yield: Version 15 of schedstats dropped counters for some sched_yield:
yld_exp_empty, yld_act_empty and yld_both_empty. Otherwise, it is yld_exp_empty, yld_act_empty and yld_both_empty. Otherwise, it is
identical to version 14. identical to version 14. Details are available at
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/scheduler/sched-stats.txt?id=1e1dbb259c79b
Version 14 of schedstats includes support for sched_domains, which hit the Version 14 of schedstats includes support for sched_domains, which hit the
mainline kernel in 2.6.20 although it is identical to the stats from version mainline kernel in 2.6.20 although it is identical to the stats from version
@ -26,7 +34,14 @@ cpus on the machine, while domain0 is the most tightly focused domain,
sometimes balancing only between pairs of cpus. At this time, there sometimes balancing only between pairs of cpus. At this time, there
are no architectures which need more than three domain levels. The first are no architectures which need more than three domain levels. The first
field in the domain stats is a bit map indicating which cpus are affected field in the domain stats is a bit map indicating which cpus are affected
by that domain. by that domain. Details are available at
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/sched-stats.txt?id=b762f3ffb797c
The schedstat documentation is maintained version 10 onwards and is not
updated for version 11 and 12. The details for version 10 are available at
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/sched-stats.txt?id=1da177e4c3f4
These fields are counters, and only increment. Programs which make use These fields are counters, and only increment. Programs which make use
of these will need to start with a baseline observation and then calculate of these will need to start with a baseline observation and then calculate
@ -71,88 +86,97 @@ Domain statistics
----------------- -----------------
One of these is produced per domain for each cpu described. (Note that if 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 CONFIG_SMP is not defined, *no* domains are utilized and these lines
will not appear in the output.) will not appear in the output. <name> is an extension to the domain field
that prints the name of the corresponding sched domain. It can appear in
schedstat version 17 and above, and requires CONFIG_SCHED_DEBUG.)
domain<N> <cpumask> 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<N> <name> <cpumask> 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 37 38 39 40 41 42 43 44 45
The first field is a bit mask indicating what cpus this domain operates over. The first field is a bit mask indicating what cpus this domain operates over.
The next 24 are a variety of sched_balance_rq() statistics in grouped into types The next 33 are a variety of sched_balance_rq() statistics in grouped into types
of idleness (idle, busy, and newly idle): of idleness (busy, idle and newly idle):
1) # of times in this domain sched_balance_rq() was called when the 1) # of times in this domain sched_balance_rq() was called when the
cpu was idle
2) # of times in this domain sched_balance_rq() checked but found
the load did not require balancing when the cpu was idle
3) # of times in this domain sched_balance_rq() tried to move one or
more tasks and failed, when the cpu was idle
4) sum of imbalances discovered (if any) with each call to
sched_balance_rq() in this domain when the cpu was idle
5) # of times in this domain pull_task() was called when the cpu
was idle
6) # of times in this domain pull_task() was called even though
the target task was cache-hot when idle
7) # of times in this domain sched_balance_rq() 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 sched_balance_rq() was called when the
cpu was busy cpu was busy
10) # of times in this domain sched_balance_rq() checked but found the 2) # of times in this domain sched_balance_rq() checked but found the
load did not require balancing when busy load did not require balancing when busy
11) # of times in this domain sched_balance_rq() tried to move one or 3) # of times in this domain sched_balance_rq() tried to move one or
more tasks and failed, when the cpu was busy more tasks and failed, when the cpu was busy
12) sum of imbalances discovered (if any) with each call to 4) Total imbalance in load when the cpu was busy
sched_balance_rq() in this domain when the cpu was busy 5) Total imbalance in utilization when the cpu was busy
13) # of times in this domain pull_task() was called when busy 6) Total imbalance in number of tasks when the cpu was busy
14) # of times in this domain pull_task() was called even though the 7) Total imbalance due to misfit tasks when the cpu was busy
8) # of times in this domain pull_task() was called when busy
9) # of times in this domain pull_task() was called even though the
target task was cache-hot when busy target task was cache-hot when busy
15) # of times in this domain sched_balance_rq() was called but did not 10) # of times in this domain sched_balance_rq() was called but did not
find a busier queue while the cpu was busy find a busier queue while the cpu was busy
16) # of times in this domain a busier queue was found while the cpu 11) # of times in this domain a busier queue was found while the cpu
was busy but no busier group was found was busy but no busier group was found
17) # of times in this domain sched_balance_rq() was called when the 12) # of times in this domain sched_balance_rq() was called when the
cpu was just becoming idle cpu was idle
18) # of times in this domain sched_balance_rq() checked but found the 13) # of times in this domain sched_balance_rq() checked but found
the load did not require balancing when the cpu was idle
14) # of times in this domain sched_balance_rq() tried to move one or
more tasks and failed, when the cpu was idle
15) Total imbalance in load when the cpu was idle
16) Total imbalance in utilization when the cpu was idle
17) Total imbalance in number of tasks when the cpu was idle
18) Total imbalance due to misfit tasks when the cpu was idle
19) # of times in this domain pull_task() was called when the cpu
was idle
20) # of times in this domain pull_task() was called even though
the target task was cache-hot when idle
21) # of times in this domain sched_balance_rq() was called but did
not find a busier queue while the cpu was idle
22) # of times in this domain a busier queue was found while the
cpu was idle but no busier group was found
23) # of times in this domain sched_balance_rq() was called when the
was just becoming idle
24) # of times in this domain sched_balance_rq() checked but found the
load did not require balancing when the cpu was just becoming idle load did not require balancing when the cpu was just becoming idle
19) # of times in this domain sched_balance_rq() tried to move one or more 25) # of times in this domain sched_balance_rq() tried to move one or more
tasks and failed, when the cpu was just becoming idle tasks and failed, when the cpu was just becoming idle
20) sum of imbalances discovered (if any) with each call to 26) Total imbalance in load when the cpu was just becoming idle
sched_balance_rq() in this domain when the cpu was just becoming idle 27) Total imbalance in utilization when the cpu was just becoming idle
21) # of times in this domain pull_task() was called when newly idle 28) Total imbalance in number of tasks when the cpu was just becoming idle
22) # of times in this domain pull_task() was called even though the 29) Total imbalance due to misfit tasks when the cpu was just becoming idle
30) # of times in this domain pull_task() was called when newly idle
31) # of times in this domain pull_task() was called even though the
target task was cache-hot when just becoming idle target task was cache-hot when just becoming idle
23) # of times in this domain sched_balance_rq() was called but did not 32) # of times in this domain sched_balance_rq() was called but did not
find a busier queue while the cpu was just becoming idle 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 33) # of times in this domain a busier queue was found while the cpu
was just becoming idle but no busier group was found was just becoming idle but no busier group was found
Next three are active_load_balance() statistics: Next three are active_load_balance() statistics:
25) # of times active_load_balance() was called 34) # of times active_load_balance() was called
26) # of times active_load_balance() tried to move a task and failed 35) # of times active_load_balance() tried to move a task and failed
27) # of times active_load_balance() successfully moved a task 36) # of times active_load_balance() successfully moved a task
Next three are sched_balance_exec() statistics: Next three are sched_balance_exec() statistics:
28) sbe_cnt is not used 37) sbe_cnt is not used
29) sbe_balanced is not used 38) sbe_balanced is not used
30) sbe_pushed is not used 39) sbe_pushed is not used
Next three are sched_balance_fork() statistics: Next three are sched_balance_fork() statistics:
31) sbf_cnt is not used 40) sbf_cnt is not used
32) sbf_balanced is not used 41) sbf_balanced is not used
33) sbf_pushed is not used 42) sbf_pushed is not used
Next three are try_to_wake_up() statistics: Next three are try_to_wake_up() statistics:
34) # of times in this domain try_to_wake_up() awoke a task that 43) # of times in this domain try_to_wake_up() awoke a task that
last ran on a different cpu in this domain last ran on a different cpu in this domain
35) # of times in this domain try_to_wake_up() moved a task to the 44) # 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 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 45) # of times in this domain try_to_wake_up() started passive balancing
/proc/<pid>/schedstat /proc/<pid>/schedstat
--------------------- ---------------------

2
kernel/sched/stats.c
View File

@ -103,7 +103,7 @@ void __update_stats_enqueue_sleeper(struct rq *rq, struct task_struct *p,
* Bump this up when changing the output format or the meaning of an existing * Bump this up when changing the output format or the meaning of an existing
* format, so that tools can adapt (or abort) * format, so that tools can adapt (or abort)
*/ */
#define SCHEDSTAT_VERSION 16 #define SCHEDSTAT_VERSION 17
static int show_schedstat(struct seq_file *seq, void *v) static int show_schedstat(struct seq_file *seq, void *v)
{ {