From: Rick Lindsley <ricklind@us.ibm.com>
It adds lots of CPU scheduler stats in /proc/pid/stat. They are described in
the new Documentation//sched-stats.txt
We were carrying this patch offline for some time, but as there's still
considerable ongoing work in this area, and as the new stats are a
configuration option, I think it's best that this capability be in the base
kernel.
Nick removed a fair amount of statistics that he wasn't using. The full patch
gathers more information. In particular, his patch doesn't include the code
to measure the latency between the time a process is made runnable and the
time it hits a processor which will be key to measuring interactivity changes.
He passed his changes back to me and I got finished merging his changes with
the current statistics patches just before OLS. I believe this is largely a
superset of the patch you grabbed and should port relatively easily too.
Versions also exist for
2.6.8-rc2
2.6.8-rc2-mm1
2.6.8-rc2-mm2
at
http://eaglet.rain.com/rick/linux/schedstat/patches/
and within 24 hours at
http://oss.software.ibm.com/linux/patches/?patch_id=730&show=all
The version below is for 2.6.8-rc2-mm2 without the staircase code and has
been compiled cleanly but not yet run.
From: Ingo Molnar <mingo@elte.hu>
this code needs a couple of cleanups before it can go into mainline:
fs/proc/array.c, fs/proc/base.c, fs/proc/proc_misc.c:
- moved the new /proc/<PID>/stat fields to /proc/<PID>/schedstat,
because the new fields break older procps. It's cleaner this way
anyway. This moving of fields necessiated a bump to version 10.
Documentation/sched-stats.txt:
- updated sched-stats.txt for version 10
- wake_up_forked_thread() => wake_up_new_task()
- updated the per-process field description
Kconfig:
- removed the default y and made the option dependent on DEBUG_KERNEL.
This is really for scheduler analysis, normal users dont need the
overhead.
include/linux/sched.h:
- moved the definitions into kernel/sched.c - this fixes UP compilation
and is cleaner.
- also moved the sched-domain definitions to sched.c - now that the
sched-domains internals are not exposed to architectures this is
doable. It's also necessary due to the previous change.
kernel/fork.c:
- moved the ->sched_info init to sched_fork() where it belongs.
kernel/sched.c:
- wake_up_forked_thread() -> wake_up_new_task(), wuft_cnt -> wunt_cnt,
wuft_moved -> wunt_moved.
- wunt_cnt and wunt_moved were defined by never updated - added the
missing code to wake_up_new_task().
- whitespace/style police
- removed whitespace changes done to code not related to schedstats -
i'll send a separate patch for these (and more).
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
---
25-akpm/Documentation/sched-stats.txt | 149 ++++++++++
25-akpm/arch/i386/Kconfig.debug | 12
25-akpm/arch/ppc/Kconfig.debug | 12
25-akpm/arch/ppc64/Kconfig | 12
25-akpm/arch/x86_64/Kconfig.debug | 12
25-akpm/fs/proc/base.c | 32 ++
25-akpm/fs/proc/proc_misc.c | 3
25-akpm/include/linux/sched.h | 131 +--------
25-akpm/kernel/sched.c | 491 ++++++++++++++++++++++++++++++++--
arch/i386/Kconfig | 0
arch/ppc/Kconfig | 0
arch/x86_64/Kconfig | 0
12 files changed, 720 insertions(+), 134 deletions(-)
diff -puN arch/i386/Kconfig~schedstat-v10 arch/i386/Kconfig
diff -puN arch/ppc64/Kconfig~schedstat-v10 arch/ppc64/Kconfig
--- 25/arch/ppc64/Kconfig~schedstat-v10 2004-08-19 23:25:31.729223568 -0700
+++ 25-akpm/arch/ppc64/Kconfig 2004-08-19 23:25:31.746220984 -0700
@@ -334,6 +334,18 @@ config VIOTAPE
If you are running Linux on an iSeries system and you want Linux
to read and/or write a tape drive owned by OS/400, say Y here.
+config SCHEDSTATS
+ bool "Collect scheduler statistics"
+ depends on DEBUG_KERNEL && PROC_FS
+ help
+ If you say Y here, additional code will be inserted into the
+ scheduler and related routines to collect statistics about
+ scheduler behavior and provide them in /proc/schedstat. These
+ stats may be useful for both tuning and debugging the scheduler
+ If you aren't debugging the scheduler or trying to tune a specific
+ application, you can say N to avoid the very slight overhead
+ this adds.
+
endmenu
config VIOPATH
diff -puN arch/ppc/Kconfig~schedstat-v10 arch/ppc/Kconfig
diff -puN arch/x86_64/Kconfig~schedstat-v10 arch/x86_64/Kconfig
diff -puN /dev/null Documentation/sched-stats.txt
--- /dev/null 2003-09-15 06:40:47.000000000 -0700
+++ 25-akpm/Documentation/sched-stats.txt 2004-08-19 23:25:31.747220832 -0700
@@ -0,0 +1,149 @@
+Version 10 of schedstats includes support for sched_domains, which
+hit the mainline kernel in 2.6.7. Some counters make more sense to be
+per-runqueue; other to be per-domain.
+
+In version 10 of schedstat, there is at least one level of domain
+statistics for each cpu listed, and there may well be more than one
+domain. Domains have no particular names in this implementation, but
+the highest numbered one typically arbitrates balancing across all the
+cpus on the machine, while domain0 is the most tightly focused domain,
+sometimes balancing only between pairs of cpus. At this time, there
+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
+by that domain.
+
+These fields are counters, and only increment. Programs which make use
+of these will need to start with a baseline observation and then calculate
+the change in the counters at each subsequent observation. A perl script
+which does this for many of the fields is available at
+
+ http://eaglet.rain.com/rick/linux/schedstat/
+
+Note that any such script will necessarily be version-specific, as the main
+reason to change versions is changes in the output format. For those wishing
+to write their own scripts, the fields are described here.
+
+CPU statistics
+--------------
+cpu<N> 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
+
+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 four are schedule() statistics:
+ 5) # of times the active queue had at least one other process on it
+ 6) # of times we switched to the expired queue and reused it
+ 7) # of times schedule() was called
+ 8) # of times schedule() left the processor idle
+
+Next four are active_load_balance() statistics:
+ 9) # of times active_load_balance() was called
+ 10) # of times active_load_balance() caused this cpu to gain a task
+ 11) # of times active_load_balance() caused this cpu to lose a task
+ 12) # of times active_load_balance() tried to move a task and failed
+
+Next three are try_to_wake_up() statistics:
+ 13) # of times try_to_wake_up() was called
+ 14) # of times try_to_wake_up() successfully moved the awakening task
+ 15) # of times try_to_wake_up() attempted to move the awakening task
+
+Next two are wake_up_new_task() statistics:
+ 16) # of times wake_up_new_task() was called
+ 17) # of times wake_up_new_task() successfully moved the new task
+
+Next one is a sched_migrate_task() statistic:
+ 18) # of times sched_migrate_task() was called
+
+Next one is a sched_balance_exec() statistic:
+ 19) # of times sched_balance_exec() was called
+
+Next three are statistics describing scheduling latency:
+ 20) sum of all time spent running by tasks on this processor (in ms)
+ 21) sum of all time spent waiting to run by tasks on this processor (in ms)
+ 22) # of tasks (not necessarily unique) given to the processor
+
+The last six are statistics dealing with pull_task():
+ 23) # of times pull_task() moved a task to this cpu when newly idle
+ 24) # of times pull_task() stole a task from this cpu when another cpu
+ was newly idle
+ 25) # of times pull_task() moved a task to this cpu when idle
+ 26) # of times pull_task() stole a task from this cpu when another cpu
+ was idle
+ 27) # of times pull_task() moved a task to this cpu when busy
+ 28) # of times pull_task() stole a task from this cpu when another cpu
+ was busy
+
+
+Domain statistics
+-----------------
+One of these is produced per domain for each cpu described.
+
+domain<N> 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
+
+The first field is a bit mask indicating what cpus this domain operates over.
+
+The next fifteen are a variety of load_balance() statistics:
+
+ 1) # of times in this domain load_balance() was called when the cpu
+ was idle
+ 2) # of times in this domain load_balance() was called when the cpu
+ was busy
+ 3) # of times in this domain load_balance() was called when the cpu
+ was just becoming idle
+ 4) # of times in this domain load_balance() tried to move one or more
+ tasks and failed, when the cpu was idle
+ 5) # of times in this domain load_balance() tried to move one or more
+ tasks and failed, when the cpu was busy
+ 6) # of times in this domain load_balance() tried to move one or more
+ tasks and failed, when the cpu was just becoming idle
+ 7) sum of imbalances discovered (if any) with each call to
+ load_balance() in this domain when the cpu was idle
+ 8) sum of imbalances discovered (if any) with each call to
+ load_balance() in this domain when the cpu was busy
+ 9) sum of imbalances discovered (if any) with each call to
+ load_balance() in this domain when the cpu was just becoming idle
+ 10) # of times in this domain load_balance() was called but did not find
+ a busier queue while the cpu was idle
+ 11) # of times in this domain load_balance() was called but did not find
+ a busier queue while the cpu was busy
+ 12) # of times in this domain load_balance() was called but did not find
+ a busier queue while the cpu was just becoming idle
+ 13) # of times in this domain a busier queue was found while the cpu was
+ idle but no busier group was found
+ 14) # of times in this domain a busier queue was found while the cpu was
+ busy but no busier group was found
+ 15) # of times in this domain a busier queue was found while the cpu was
+ just becoming idle but no busier group was found
+
+Next two are sched_balance_exec() statistics:
+ 17) # of times in this domain sched_balance_exec() successfully pushed
+ a task to a new cpu
+ 18) # of times in this domain sched_balance_exec() tried but failed to
+ push a task to a new cpu
+
+Next two are try_to_wake_up() statistics:
+ 19) # of times in this domain try_to_wake_up() tried to move a task based
+ on affinity and cache warmth
+ 20) # of times in this domain try_to_wake_up() tried to move a task based
+ on load balancing
+
+
+/proc/<pid>/schedstat
+----------------
+schedstats also adds a new /proc/<pid/schedstat file to include some of
+the same information on a per-process level. There are three fields in
+this file correlating to fields 20, 21, and 22 in the CPU fields, but
+they only apply for that process.
+
+A program could be easily written to make use of these extra fields to
+report on how well a particular process or set of processes is faring
+under the scheduler's policies. A simple version of such a program is
+available at
+ http://eaglet.rain.com/rick/linux/schedstat/v10/latency.c
diff -puN fs/proc/base.c~schedstat-v10 fs/proc/base.c
--- 25/fs/proc/base.c~schedstat-v10 2004-08-19 23:25:31.734222808 -0700
+++ 25-akpm/fs/proc/base.c 2004-08-19 23:25:31.749220528 -0700
@@ -60,6 +60,9 @@ enum pid_directory_inos {
PROC_TGID_MAPS,
PROC_TGID_MOUNTS,
PROC_TGID_WCHAN,
+#ifdef CONFIG_SCHEDSTATS
+ PROC_TGID_SCHEDSTAT,
+#endif
#ifdef CONFIG_SECURITY
PROC_TGID_ATTR,
PROC_TGID_ATTR_CURRENT,
@@ -83,6 +86,9 @@ enum pid_directory_inos {
PROC_TID_MAPS,
PROC_TID_MOUNTS,
PROC_TID_WCHAN,
+#ifdef CONFIG_SCHEDSTATS
+ PROC_TID_SCHEDSTAT,
+#endif
#ifdef CONFIG_SECURITY
PROC_TID_ATTR,
PROC_TID_ATTR_CURRENT,
@@ -123,6 +129,9 @@ static struct pid_entry tgid_base_stuff[
#ifdef CONFIG_KALLSYMS
E(PROC_TGID_WCHAN, "wchan", S_IFREG|S_IRUGO),
#endif
+#ifdef CONFIG_SCHEDSTATS
+ E(PROC_TGID_SCHEDSTAT, "schedstat", S_IFREG|S_IRUGO),
+#endif
{0,0,NULL,0}
};
static struct pid_entry tid_base_stuff[] = {
@@ -145,6 +154,9 @@ static struct pid_entry tid_base_stuff[]
#ifdef CONFIG_KALLSYMS
E(PROC_TID_WCHAN, "wchan", S_IFREG|S_IRUGO),
#endif
+#ifdef CONFIG_SCHEDSTATS
+ E(PROC_TID_SCHEDSTAT, "schedstat",S_IFREG|S_IRUGO),
+#endif
{0,0,NULL,0}
};
@@ -397,6 +409,19 @@ static int proc_pid_wchan(struct task_st
}
#endif /* CONFIG_KALLSYMS */
+#ifdef CONFIG_SCHEDSTATS
+/*
+ * Provides /proc/PID/schedstat
+ */
+static int proc_pid_schedstat(struct task_struct *task, char *buffer)
+{
+ return sprintf(buffer, "%lu %lu %lu\n",
+ task->sched_info.cpu_time,
+ task->sched_info.run_delay,
+ task->sched_info.pcnt);
+}
+#endif
+
/************************************************************************/
/* Here the fs part begins */
/************************************************************************/
@@ -1375,6 +1400,13 @@ static struct dentry *proc_pident_lookup
ei->op.proc_read = proc_pid_wchan;
break;
#endif
+#ifdef CONFIG_SCHEDSTATS
+ case PROC_TID_SCHEDSTAT:
+ case PROC_TGID_SCHEDSTAT:
+ inode->i_fop = &proc_info_file_operations;
+ ei->op.proc_read = proc_pid_schedstat;
+ break;
+#endif
default:
printk("procfs: impossible type (%d)",p->type);
iput(inode);
diff -puN fs/proc/proc_misc.c~schedstat-v10 fs/proc/proc_misc.c
--- 25/fs/proc/proc_misc.c~schedstat-v10 2004-08-19 23:25:31.735222656 -0700
+++ 25-akpm/fs/proc/proc_misc.c 2004-08-19 23:25:31.750220376 -0700
@@ -711,6 +711,9 @@ void __init proc_misc_init(void)
#ifdef CONFIG_MODULES
create_seq_entry("modules", 0, &proc_modules_operations);
#endif
+#ifdef CONFIG_SCHEDSTATS
+ create_seq_entry("schedstat", 0, &proc_schedstat_operations);
+#endif
#ifdef CONFIG_PROC_KCORE
proc_root_kcore = create_proc_entry("kcore", S_IRUSR, NULL);
if (proc_root_kcore) {
diff -puN include/linux/sched.h~schedstat-v10 include/linux/sched.h
--- 25/include/linux/sched.h~schedstat-v10 2004-08-19 23:25:31.736222504 -0700
+++ 25-akpm/include/linux/sched.h 2004-08-19 23:25:31.751220224 -0700
@@ -347,6 +347,20 @@ struct k_itimer {
struct timespec wall_to_prev; /* wall_to_monotonic used when set */
};
+#ifdef CONFIG_SCHEDSTATS
+struct sched_info {
+ /* cumulative counters */
+ unsigned long cpu_time, /* time spent on the cpu */
+ run_delay, /* time spent waiting on a runqueue */
+ pcnt; /* # of timeslices run on this cpu */
+
+ /* timestamps */
+ unsigned long last_arrival, /* when we last ran on a cpu */
+ last_queued; /* when we were last queued to run */
+};
+
+extern struct file_operations proc_schedstat_operations;
+#endif
struct io_context; /* See blkdev.h */
void exit_io_context(void);
@@ -409,6 +423,10 @@ struct task_struct {
cpumask_t cpus_allowed;
unsigned int time_slice, first_time_slice;
+#ifdef CONFIG_SCHEDSTATS
+ struct sched_info sched_info;
+#endif
+
struct list_head tasks;
/*
* ptrace_list/ptrace_children forms the list of my children
@@ -565,119 +583,6 @@ do { if (atomic_dec_and_test(&(tsk)->usa
#define PF_SYNCWRITE 0x00200000 /* I am doing a sync write */
#ifdef CONFIG_SMP
-#define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
-
-#define SD_BALANCE_NEWIDLE 1 /* Balance when about to become idle */
-#define SD_BALANCE_EXEC 2 /* Balance on exec */
-#define SD_WAKE_IDLE 4 /* Wake to idle CPU on task wakeup */
-#define SD_WAKE_AFFINE 8 /* Wake task to waking CPU */
-#define SD_WAKE_BALANCE 16 /* Perform balancing at task wakeup */
-#define SD_SHARE_CPUPOWER 32 /* Domain members share cpu power */
-
-struct sched_group {
- struct sched_group *next; /* Must be a circular list */
- cpumask_t cpumask;
-
- /*
- * CPU power of this group, SCHED_LOAD_SCALE being max power for a
- * single CPU. This should be read only (except for setup). Although
- * it will need to be written to at cpu hot(un)plug time, perhaps the
- * cpucontrol semaphore will provide enough exclusion?
- */
- unsigned long cpu_power;
-};
-
-struct sched_domain {
- /* These fields must be setup */
- struct sched_domain *parent; /* top domain must be null terminated */
- struct sched_group *groups; /* the balancing groups of the domain */
- cpumask_t span; /* span of all CPUs in this domain */
- unsigned long min_interval; /* Minimum balance interval ms */
- unsigned long max_interval; /* Maximum balance interval ms */
- unsigned int busy_factor; /* less balancing by factor if busy */
- unsigned int imbalance_pct; /* No balance until over watermark */
- unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
- unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
- unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
- int flags; /* See SD_* */
-
- /* Runtime fields. */
- unsigned long last_balance; /* init to jiffies. units in jiffies */
- unsigned int balance_interval; /* initialise to 1. units in ms. */
- unsigned int nr_balance_failed; /* initialise to 0 */
-};
-
-#ifndef ARCH_HAS_SCHED_TUNE
-#ifdef CONFIG_SCHED_SMT
-#define ARCH_HAS_SCHED_WAKE_IDLE
-/* Common values for SMT siblings */
-#define SD_SIBLING_INIT (struct sched_domain) { \
- .span = CPU_MASK_NONE, \
- .parent = NULL, \
- .groups = NULL, \
- .min_interval = 1, \
- .max_interval = 2, \
- .busy_factor = 8, \
- .imbalance_pct = 110, \
- .cache_hot_time = 0, \
- .cache_nice_tries = 0, \
- .per_cpu_gain = 25, \
- .flags = SD_BALANCE_NEWIDLE \
- | SD_BALANCE_EXEC \
- | SD_WAKE_AFFINE \
- | SD_WAKE_IDLE \
- | SD_SHARE_CPUPOWER, \
- .last_balance = jiffies, \
- .balance_interval = 1, \
- .nr_balance_failed = 0, \
-}
-#endif
-
-/* Common values for CPUs */
-#define SD_CPU_INIT (struct sched_domain) { \
- .span = CPU_MASK_NONE, \
- .parent = NULL, \
- .groups = NULL, \
- .min_interval = 1, \
- .max_interval = 4, \
- .busy_factor = 64, \
- .imbalance_pct = 125, \
- .cache_hot_time = (5*1000000/2), \
- .cache_nice_tries = 1, \
- .per_cpu_gain = 100, \
- .flags = SD_BALANCE_NEWIDLE \
- | SD_BALANCE_EXEC \
- | SD_WAKE_AFFINE \
- | SD_WAKE_BALANCE, \
- .last_balance = jiffies, \
- .balance_interval = 1, \
- .nr_balance_failed = 0, \
-}
-
-#ifdef CONFIG_NUMA
-/* Common values for NUMA nodes */
-#define SD_NODE_INIT (struct sched_domain) { \
- .span = CPU_MASK_NONE, \
- .parent = NULL, \
- .groups = NULL, \
- .min_interval = 8, \
- .max_interval = 32, \
- .busy_factor = 32, \
- .imbalance_pct = 125, \
- .cache_hot_time = (10*1000000), \
- .cache_nice_tries = 1, \
- .per_cpu_gain = 100, \
- .flags = SD_BALANCE_EXEC \
- | SD_WAKE_BALANCE, \
- .last_balance = jiffies, \
- .balance_interval = 1, \
- .nr_balance_failed = 0, \
-}
-#endif
-#endif /* ARCH_HAS_SCHED_TUNE */
-
-extern void cpu_attach_domain(struct sched_domain *sd, int cpu);
-
extern int set_cpus_allowed(task_t *p, cpumask_t new_mask);
#else
static inline int set_cpus_allowed(task_t *p, cpumask_t new_mask)
diff -puN kernel/sched.c~schedstat-v10 kernel/sched.c
--- 25/kernel/sched.c~schedstat-v10 2004-08-19 23:25:31.738222200 -0700
+++ 25-akpm/kernel/sched.c 2004-08-19 23:25:55.400624968 -0700
@@ -41,6 +41,8 @@
#include <linux/percpu.h>
#include <linux/perfctr.h>
#include <linux/kthread.h>
+#include <linux/seq_file.h>
+#include <linux/times.h>
#include <asm/tlb.h>
#include <asm/unistd.h>
@@ -183,6 +185,16 @@ static unsigned int task_timeslice(task_
#define task_hot(p, now, sd) ((now) - (p)->timestamp < (sd)->cache_hot_time)
+enum idle_type
+{
+ IDLE,
+ NOT_IDLE,
+ NEWLY_IDLE,
+ MAX_IDLE_TYPES
+};
+
+struct sched_domain;
+
/*
* These are the runqueue data structures:
*/
@@ -234,10 +246,186 @@ struct runqueue {
task_t *migration_thread;
struct list_head migration_queue;
#endif
+
+#ifdef CONFIG_SCHEDSTATS
+ /* latency stats */
+ struct sched_info rq_sched_info;
+
+ /* sys_sched_yield() stats */
+ unsigned long yld_exp_empty;
+ unsigned long yld_act_empty;
+ unsigned long yld_both_empty;
+ unsigned long yld_cnt;
+
+ /* schedule() stats */
+ unsigned long sched_noswitch;
+ unsigned long sched_switch;
+ unsigned long sched_cnt;
+ unsigned long sched_goidle;
+
+ /* pull_task() stats */
+ unsigned long pt_gained[MAX_IDLE_TYPES];
+ unsigned long pt_lost[MAX_IDLE_TYPES];
+
+ /* active_load_balance() stats */
+ unsigned long alb_cnt;
+ unsigned long alb_lost;
+ unsigned long alb_gained;
+ unsigned long alb_failed;
+
+ /* try_to_wake_up() stats */
+ unsigned long ttwu_cnt;
+ unsigned long ttwu_attempts;
+ unsigned long ttwu_moved;
+
+ /* wake_up_new_task() stats */
+ unsigned long wunt_cnt;
+ unsigned long wunt_moved;
+
+ /* sched_migrate_task() stats */
+ unsigned long smt_cnt;
+
+ /* sched_balance_exec() stats */
+ unsigned long sbe_cnt;
+#endif
};
static DEFINE_PER_CPU(struct runqueue, runqueues);
+/*
+ * sched-domains (multiprocessor balancing) declarations:
+ */
+#ifdef CONFIG_SMP
+#define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
+
+#define SD_BALANCE_NEWIDLE 1 /* Balance when about to become idle */
+#define SD_BALANCE_EXEC 2 /* Balance on exec */
+#define SD_WAKE_IDLE 4 /* Wake to idle CPU on task wakeup */
+#define SD_WAKE_AFFINE 8 /* Wake task to waking CPU */
+#define SD_WAKE_BALANCE 16 /* Perform balancing at task wakeup */
+#define SD_SHARE_CPUPOWER 32 /* Domain members share cpu power */
+
+struct sched_group {
+ struct sched_group *next; /* Must be a circular list */
+ cpumask_t cpumask;
+
+ /*
+ * CPU power of this group, SCHED_LOAD_SCALE being max power for a
+ * single CPU. This should be read only (except for setup). Although
+ * it will need to be written to at cpu hot(un)plug time, perhaps the
+ * cpucontrol semaphore will provide enough exclusion?
+ */
+ unsigned long cpu_power;
+};
+
+struct sched_domain {
+ /* These fields must be setup */
+ struct sched_domain *parent; /* top domain must be null terminated */
+ struct sched_group *groups; /* the balancing groups of the domain */
+ cpumask_t span; /* span of all CPUs in this domain */
+ unsigned long min_interval; /* Minimum balance interval ms */
+ unsigned long max_interval; /* Maximum balance interval ms */
+ unsigned int busy_factor; /* less balancing by factor if busy */
+ unsigned int imbalance_pct; /* No balance until over watermark */
+ unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
+ unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
+ unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
+ int flags; /* See SD_* */
+
+ /* Runtime fields. */
+ unsigned long last_balance; /* init to jiffies. units in jiffies */
+ unsigned int balance_interval; /* initialise to 1. units in ms. */
+ unsigned int nr_balance_failed; /* initialise to 0 */
+
+#ifdef CONFIG_SCHEDSTATS
+ /* load_balance() stats */
+ unsigned long lb_cnt[MAX_IDLE_TYPES];
+ unsigned long lb_failed[MAX_IDLE_TYPES];
+ unsigned long lb_imbalance[MAX_IDLE_TYPES];
+ unsigned long lb_nobusyg[MAX_IDLE_TYPES];
+ unsigned long lb_nobusyq[MAX_IDLE_TYPES];
+
+ /* sched_balance_exec() stats */
+ unsigned long sbe_attempts;
+ unsigned long sbe_pushed;
+
+ /* try_to_wake_up() stats */
+ unsigned long ttwu_wake_affine;
+ unsigned long ttwu_wake_balance;
+#endif
+};
+
+#ifndef ARCH_HAS_SCHED_TUNE
+#ifdef CONFIG_SCHED_SMT
+#define ARCH_HAS_SCHED_WAKE_IDLE
+/* Common values for SMT siblings */
+#define SD_SIBLING_INIT (struct sched_domain) { \
+ .span = CPU_MASK_NONE, \
+ .parent = NULL, \
+ .groups = NULL, \
+ .min_interval = 1, \
+ .max_interval = 2, \
+ .busy_factor = 8, \
+ .imbalance_pct = 110, \
+ .cache_hot_time = 0, \
+ .cache_nice_tries = 0, \
+ .per_cpu_gain = 25, \
+ .flags = SD_BALANCE_NEWIDLE \
+ | SD_BALANCE_EXEC \
+ | SD_WAKE_AFFINE \
+ | SD_WAKE_IDLE \
+ | SD_SHARE_CPUPOWER, \
+ .last_balance = jiffies, \
+ .balance_interval = 1, \
+ .nr_balance_failed = 0, \
+}
+#endif
+
+/* Common values for CPUs */
+#define SD_CPU_INIT (struct sched_domain) { \
+ .span = CPU_MASK_NONE, \
+ .parent = NULL, \
+ .groups = NULL, \
+ .min_interval = 1, \
+ .max_interval = 4, \
+ .busy_factor = 64, \
+ .imbalance_pct = 125, \
+ .cache_hot_time = (5*1000000/2), \
+ .cache_nice_tries = 1, \
+ .per_cpu_gain = 100, \
+ .flags = SD_BALANCE_NEWIDLE \
+ | SD_BALANCE_EXEC \
+ | SD_WAKE_AFFINE \
+ | SD_WAKE_BALANCE, \
+ .last_balance = jiffies, \
+ .balance_interval = 1, \
+ .nr_balance_failed = 0, \
+}
+
+/* Arch can override this macro in processor.h */
+#if defined(CONFIG_NUMA) && !defined(SD_NODE_INIT)
+#define SD_NODE_INIT (struct sched_domain) { \
+ .span = CPU_MASK_NONE, \
+ .parent = NULL, \
+ .groups = NULL, \
+ .min_interval = 8, \
+ .max_interval = 32, \
+ .busy_factor = 32, \
+ .imbalance_pct = 125, \
+ .cache_hot_time = (10*1000000), \
+ .cache_nice_tries = 1, \
+ .per_cpu_gain = 100, \
+ .flags = SD_BALANCE_EXEC \
+ | SD_WAKE_BALANCE, \
+ .last_balance = jiffies, \
+ .balance_interval = 1, \
+ .nr_balance_failed = 0, \
+}
+#endif
+#endif /* ARCH_HAS_SCHED_TUNE */
+#endif
+
+
#define for_each_domain(cpu, domain) \
for (domain = cpu_rq(cpu)->sd; domain; domain = domain->parent)
@@ -280,6 +468,100 @@ static inline void task_rq_unlock(runque
spin_unlock_irqrestore(&rq->lock, *flags);
}
+#ifdef CONFIG_SCHEDSTATS
+/*
+ * bump this up when changing the output format or the meaning of an existing
+ * format, so that tools can adapt (or abort)
+ */
+#define SCHEDSTAT_VERSION 10
+
+static int show_schedstat(struct seq_file *seq, void *v)
+{
+ int cpu;
+ enum idle_type itype;
+
+ seq_printf(seq, "version %d\n", SCHEDSTAT_VERSION);
+ seq_printf(seq, "timestamp %lu\n", jiffies);
+ for_each_online_cpu(cpu) {
+ runqueue_t *rq = cpu_rq(cpu);
+ struct sched_domain *sd;
+ int dcnt = 0;
+
+ /* runqueue-specific stats */
+ seq_printf(seq,
+ "cpu%d %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu "
+ "%lu %lu %lu %lu %lu %lu %lu %lu %lu %lu",
+ cpu, rq->yld_both_empty,
+ rq->yld_act_empty, rq->yld_exp_empty,
+ rq->yld_cnt, rq->sched_noswitch,
+ rq->sched_switch, rq->sched_cnt, rq->sched_goidle,
+ rq->alb_cnt, rq->alb_gained, rq->alb_lost,
+ rq->alb_failed,
+ rq->ttwu_cnt, rq->ttwu_moved, rq->ttwu_attempts,
+ rq->wunt_cnt, rq->wunt_moved,
+ rq->smt_cnt, rq->sbe_cnt, rq->rq_sched_info.cpu_time,
+ rq->rq_sched_info.run_delay, rq->rq_sched_info.pcnt);
+
+ for (itype = IDLE; itype < MAX_IDLE_TYPES; itype++)
+ seq_printf(seq, " %lu %lu", rq->pt_gained[itype],
+ rq->pt_lost[itype]);
+ seq_printf(seq, "\n");
+
+ /* domain-specific stats */
+ for_each_domain(cpu, sd) {
+ char mask_str[NR_CPUS];
+
+ cpumask_scnprintf(mask_str, NR_CPUS, sd->span);
+ seq_printf(seq, "domain%d %s", dcnt++, mask_str);
+ for (itype = IDLE; itype < MAX_IDLE_TYPES; itype++) {
+ seq_printf(seq, " %lu %lu %lu %lu %lu",
+ sd->lb_cnt[itype],
+ sd->lb_failed[itype],
+ sd->lb_imbalance[itype],
+ sd->lb_nobusyq[itype],
+ sd->lb_nobusyg[itype]);
+ }
+ seq_printf(seq, " %lu %lu %lu %lu\n",
+ sd->sbe_pushed, sd->sbe_attempts,
+ sd->ttwu_wake_affine, sd->ttwu_wake_balance);
+ }
+ }
+ return 0;
+}
+
+static int schedstat_open(struct inode *inode, struct file *file)
+{
+ unsigned int size = PAGE_SIZE * (1 + num_online_cpus() / 32);
+ char *buf = kmalloc(size, GFP_KERNEL);
+ struct seq_file *m;
+ int res;
+
+ if (!buf)
+ return -ENOMEM;
+ res = single_open(file, show_schedstat, NULL);
+ if (!res) {
+ m = file->private_data;
+ m->buf = buf;
+ m->size = size;
+ } else
+ kfree(buf);
+ return res;
+}
+
+struct file_operations proc_schedstat_operations = {
+ .open = schedstat_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = single_release,
+};
+
+# define schedstat_inc(rq, field) rq->field++;
+# define schedstat_add(rq, field, amt) rq->field += amt;
+#else /* !CONFIG_SCHEDSTATS */
+# define schedstat_inc(rq, field) do { } while (0);
+# define schedstat_add(rq, field, amt) do { } while (0);
+#endif
+
/*
* rq_lock - lock a given runqueue and disable interrupts.
*/
@@ -299,6 +581,112 @@ static inline void rq_unlock(runqueue_t
spin_unlock_irq(&rq->lock);
}
+#ifdef CONFIG_SCHEDSTATS
+/*
+ * Called when a process is dequeued from the active array and given
+ * the cpu. We should note that with the exception of interactive
+ * tasks, the expired queue will become the active queue after the active
+ * queue is empty, without explicitly dequeuing and requeuing tasks in the
+ * expired queue. (Interactive tasks may be requeued directly to the
+ * active queue, thus delaying tasks in the expired queue from running;
+ * see scheduler_tick()).
+ *
+ * This function is only called from sched_info_arrive(), rather than
+ * dequeue_task(). Even though a task may be queued and dequeued multiple
+ * times as it is shuffled about, we're really interested in knowing how
+ * long it was from the *first* time it was queued to the time that it
+ * finally hit a cpu.
+ */
+static inline void sched_info_dequeued(task_t *t)
+{
+ t->sched_info.last_queued = 0;
+}
+
+/*
+ * Called when a task finally hits the cpu. We can now calculate how
+ * long it was waiting to run. We also note when it began so that we
+ * can keep stats on how long its timeslice is.
+ */
+static inline void sched_info_arrive(task_t *t)
+{
+ unsigned long now = jiffies, diff = 0;
+ struct runqueue *rq = task_rq(t);
+
+ if (t->sched_info.last_queued)
+ diff = now - t->sched_info.last_queued;
+ sched_info_dequeued(t);
+ t->sched_info.run_delay += diff;
+ t->sched_info.last_arrival = now;
+ t->sched_info.pcnt++;
+
+ if (!rq)
+ return;
+
+ rq->rq_sched_info.run_delay += diff;
+ rq->rq_sched_info.pcnt++;
+}
+
+/*
+ * Called when a process is queued into either the active or expired
+ * array. The time is noted and later used to determine how long we
+ * had to wait for us to reach the cpu. Since the expired queue will
+ * become the active queue after active queue is empty, without dequeuing
+ * and requeuing any tasks, we are interested in queuing to either. It
+ * is unusual but not impossible for tasks to be dequeued and immediately
+ * requeued in the same or another array: this can happen in sched_yield(),
+ * set_user_nice(), and even load_balance() as it moves tasks from runqueue
+ * to runqueue.
+ *
+ * This function is only called from enqueue_task(), but also only updates
+ * the timestamp if it is already not set. It's assumed that
+ * sched_info_dequeued() will clear that stamp when appropriate.
+ */
+static inline void sched_info_queued(task_t *t)
+{
+ if (!t->sched_info.last_queued)
+ t->sched_info.last_queued = jiffies;
+}
+
+/*
+ * Called when a process ceases being the active-running process, either
+ * voluntarily or involuntarily. Now we can calculate how long we ran.
+ */
+static inline void sched_info_depart(task_t *t)
+{
+ struct runqueue *rq = task_rq(t);
+ unsigned long diff = jiffies - t->sched_info.last_arrival;
+
+ t->sched_info.cpu_time += diff;
+
+ if (rq)
+ rq->rq_sched_info.cpu_time += diff;
+}
+
+/*
+ * Called when tasks are switched involuntarily due, typically, to expiring
+ * their time slice. (This may also be called when switching to or from
+ * the idle task.) We are only called when prev != next.
+ */
+static inline void sched_info_switch(task_t *prev, task_t *next)
+{
+ struct runqueue *rq = task_rq(prev);
+
+ /*
+ * prev now departs the cpu. It's not interesting to record
+ * stats about how efficient we were at scheduling the idle
+ * process, however.
+ */
+ if (prev != rq->idle)
+ sched_info_depart(prev);
+
+ if (next != rq->idle)
+ sched_info_arrive(next);
+}
+#else
+#define sched_info_queued(t) do { } while (0)
+#define sched_info_switch(t, next) do { } while (0)
+#endif /* CONFIG_SCHEDSTATS */
+
/*
* Adding/removing a task to/from a priority array:
*/
@@ -312,6 +700,7 @@ static void dequeue_task(struct task_str
static void enqueue_task(struct task_struct *p, prio_array_t *array)
{
+ sched_info_queued(p);
list_add_tail(&p->run_list, array->queue + p->prio);
__set_bit(p->prio, array->bitmap);
array->nr_active++;
@@ -741,6 +1130,7 @@ static int try_to_wake_up(task_t * p, un
#endif
rq = task_rq_lock(p, &flags);
+ schedstat_inc(rq, ttwu_cnt);
old_state = p->state;
if (!(old_state & state))
goto out;
@@ -788,23 +1178,35 @@ static int try_to_wake_up(task_t * p, un
*/
imbalance = sd->imbalance_pct + (sd->imbalance_pct - 100) / 2;
- if ( ((sd->flags & SD_WAKE_AFFINE) &&
- !task_hot(p, rq->timestamp_last_tick, sd))
- || ((sd->flags & SD_WAKE_BALANCE) &&
- imbalance*this_load <= 100*load) ) {
+ if ((sd->flags & SD_WAKE_AFFINE) &&
+ !task_hot(p, rq->timestamp_last_tick, sd)) {
+ /*
+ * This domain has SD_WAKE_AFFINE and p is cache cold
+ * in this domain.
+ */
+ if (cpu_isset(cpu, sd->span)) {
+ schedstat_inc(sd, ttwu_wake_affine);
+ goto out_set_cpu;
+ }
+ } else if ((sd->flags & SD_WAKE_BALANCE) &&
+ imbalance*this_load <= 100*load) {
/*
- * Now sd has SD_WAKE_AFFINE and p is cache cold in sd
- * or sd has SD_WAKE_BALANCE and there is an imbalance
+ * This domain has SD_WAKE_BALANCE and there is
+ * an imbalance.
*/
- if (cpu_isset(cpu, sd->span))
+ if (cpu_isset(cpu, sd->span)) {
+ schedstat_inc(sd, ttwu_wake_balance);
goto out_set_cpu;
+ }
}
}
new_cpu = cpu; /* Could not wake to this_cpu. Wake to cpu instead */
out_set_cpu:
+ schedstat_inc(rq, ttwu_attempts);
new_cpu = wake_idle(new_cpu, p);
if (new_cpu != cpu && cpu_isset(new_cpu, p->cpus_allowed)) {
+ schedstat_inc(rq, ttwu_moved);
set_task_cpu(p, new_cpu);
task_rq_unlock(rq, &flags);
/* might preempt at this point */
@@ -887,6 +1289,9 @@ void fastcall sched_fork(task_t *p)
INIT_LIST_HEAD(&p->run_list);
p->array = NULL;
spin_lock_init(&p->switch_lock);
+#ifdef CONFIG_SCHEDSTATS
+ memset(&p->sched_info, 0, sizeof(p->sched_info));
+#endif
#ifdef CONFIG_PREEMPT
/*
* During context-switch we hold precisely one spinlock, which
@@ -944,6 +1349,7 @@ void fastcall wake_up_new_task(task_t *
BUG_ON(p->state != TASK_RUNNING);
+ schedstat_inc(rq, wunt_cnt);
/*
* We decrease the sleep average of forking parents
* and children as well, to keep max-interactive tasks
@@ -992,6 +1398,7 @@ void fastcall wake_up_new_task(task_t *
current->sleep_avg = JIFFIES_TO_NS(CURRENT_BONUS(current) *
PARENT_PENALTY / 100 * MAX_SLEEP_AVG / MAX_BONUS);
+ schedstat_inc(rq, wunt_moved);
}
if (unlikely(cpu != this_cpu)) {
@@ -1162,13 +1569,6 @@ unsigned long nr_iowait(void)
return sum;
}
-enum idle_type
-{
- IDLE,
- NOT_IDLE,
- NEWLY_IDLE,
-};
-
#ifdef CONFIG_SMP
/*
@@ -1283,6 +1683,7 @@ static void sched_migrate_task(task_t *p
|| unlikely(cpu_is_offline(dest_cpu)))
goto out;
+ schedstat_inc(rq, smt_cnt);
/* force the process onto the specified CPU */
if (migrate_task(p, dest_cpu, &req)) {
/* Need to wait for migration thread (might exit: take ref). */
@@ -1310,6 +1711,7 @@ void sched_exec(void)
struct sched_domain *tmp, *sd = NULL;
int new_cpu, this_cpu = get_cpu();
+ schedstat_inc(this_rq(), sbe_cnt);
/* Prefer the current CPU if there's only this task running */
if (this_rq()->nr_running <= 1)
goto out;
@@ -1318,9 +1720,11 @@ void sched_exec(void)
if (tmp->flags & SD_BALANCE_EXEC)
sd = tmp;
+ schedstat_inc(sd, sbe_attempts);
if (sd) {
new_cpu = find_idlest_cpu(current, this_cpu, sd);
if (new_cpu != this_cpu) {
+ schedstat_inc(sd, sbe_pushed);
put_cpu();
sched_migrate_task(current, new_cpu);
return;
@@ -1444,6 +1848,15 @@ skip_queue:
idx++;
goto skip_bitmap;
}
+
+ /*
+ * Right now, this is the only place pull_task() is called,
+ * so we can safely collect pull_task() stats here rather than
+ * inside pull_task().
+ */
+ schedstat_inc(this_rq, pt_gained[idle]);
+ schedstat_inc(busiest, pt_lost[idle]);
+
pull_task(busiest, array, tmp, this_rq, dst_array, this_cpu);
pulled++;
@@ -1638,14 +2051,20 @@ static int load_balance(int this_cpu, ru
int nr_moved;
spin_lock(&this_rq->lock);
+ schedstat_inc(sd, lb_cnt[idle]);
group = find_busiest_group(sd, this_cpu, &imbalance, idle);
- if (!group)
+ if (!group) {
+ schedstat_inc(sd, lb_nobusyg[idle]);
goto out_balanced;
+ }
busiest = find_busiest_queue(group);
- if (!busiest)
+ if (!busiest) {
+ schedstat_inc(sd, lb_nobusyq[idle]);
goto out_balanced;
+ }
+
/*
* This should be "impossible", but since load
* balancing is inherently racy and statistical,
@@ -1656,6 +2075,8 @@ static int load_balance(int this_cpu, ru
goto out_balanced;
}
+ schedstat_add(sd, lb_imbalance[idle], imbalance);
+
nr_moved = 0;
if (busiest->nr_running > 1) {
/*
@@ -1672,6 +2093,7 @@ static int load_balance(int this_cpu, ru
spin_unlock(&this_rq->lock);
if (!nr_moved) {
+ schedstat_inc(sd, lb_failed[idle]);
sd->nr_balance_failed++;
if (unlikely(sd->nr_balance_failed > sd->cache_nice_tries+2)) {
@@ -1726,19 +2148,27 @@ static int load_balance_newidle(int this
unsigned long imbalance;
int nr_moved = 0;
+ schedstat_inc(sd, lb_cnt[NEWLY_IDLE]);
group = find_busiest_group(sd, this_cpu, &imbalance, NEWLY_IDLE);
- if (!group)
+ if (!group) {
+ schedstat_inc(sd, lb_nobusyg[NEWLY_IDLE]);
goto out;
+ }
busiest = find_busiest_queue(group);
- if (!busiest || busiest == this_rq)
+ if (!busiest || busiest == this_rq) {
+ schedstat_inc(sd, lb_nobusyq[NEWLY_IDLE]);
goto out;
+ }
/* Attempt to move tasks */
double_lock_balance(this_rq, busiest);
+ schedstat_add(sd, lb_imbalance[NEWLY_IDLE], imbalance);
nr_moved = move_tasks(this_rq, this_cpu, busiest,
imbalance, sd, NEWLY_IDLE);
+ if (!nr_moved)
+ schedstat_inc(sd, lb_failed[NEWLY_IDLE]);
spin_unlock(&busiest->lock);
@@ -1778,6 +2208,7 @@ static void active_load_balance(runqueue
struct sched_group *group, *busy_group;
int i;
+ schedstat_inc(busiest, alb_cnt);
if (busiest->nr_running <= 1)
return;
@@ -1822,7 +2253,12 @@ static void active_load_balance(runqueue
if (unlikely(busiest == rq))
goto next_group;
double_lock_balance(busiest, rq);
- move_tasks(rq, push_cpu, busiest, 1, sd, IDLE);
+ if (move_tasks(rq, push_cpu, busiest, 1, sd, IDLE)) {
+ schedstat_inc(busiest, alb_lost);
+ schedstat_inc(rq, alb_gained);
+ } else {
+ schedstat_inc(busiest, alb_failed);
+ }
spin_unlock(&rq->lock);
next_group:
group = group->next;
@@ -2175,6 +2611,7 @@ need_resched:
}
release_kernel_lock(prev);
+ schedstat_inc(rq, sched_cnt);
now = sched_clock();
if (likely(now - prev->timestamp < NS_MAX_SLEEP_AVG))
run_time = now - prev->timestamp;
@@ -2221,18 +2658,21 @@ need_resched:
/*
* Switch the active and expired arrays.
*/
+ schedstat_inc(rq, sched_switch);
rq->active = rq->expired;
rq->expired = array;
array = rq->active;
rq->expired_timestamp = 0;
rq->best_expired_prio = MAX_PRIO;
- }
+ } else
+ schedstat_inc(rq, sched_noswitch);
idx = sched_find_first_bit(array->bitmap);
queue = array->queue + idx;
next = list_entry(queue->next, task_t, run_list);
if (dependent_sleeper(cpu, rq, next)) {
+ schedstat_inc(rq, sched_goidle);
next = rq->idle;
goto switch_tasks;
}
@@ -2262,6 +2702,7 @@ switch_tasks:
}
prev->timestamp = now;
+ sched_info_switch(prev, next);
if (likely(prev != next)) {
next->timestamp = now;
rq->nr_switches++;
@@ -2980,6 +3421,7 @@ asmlinkage long sys_sched_yield(void)
prio_array_t *array = current->array;
prio_array_t *target = rq->expired;
+ schedstat_inc(rq, yld_cnt);
/*
* We implement yielding by moving the task into the expired
* queue.
@@ -2990,6 +3432,13 @@ asmlinkage long sys_sched_yield(void)
if (rt_task(current))
target = rq->active;
+ if (current->array->nr_active == 1) {
+ schedstat_inc(rq, yld_act_empty);
+ if (!rq->expired->nr_active)
+ schedstat_inc(rq, yld_both_empty);
+ } else if (!rq->expired->nr_active)
+ schedstat_inc(rq, yld_exp_empty);
+
dequeue_task(current, array);
enqueue_task(current, target);
@@ -3633,7 +4082,7 @@ EXPORT_SYMBOL(kernel_flag);
#ifdef CONFIG_SMP
/* Attach the domain 'sd' to 'cpu' as its base domain */
-void cpu_attach_domain(struct sched_domain *sd, int cpu)
+static void cpu_attach_domain(struct sched_domain *sd, int cpu)
{
migration_req_t req;
unsigned long flags;
diff -puN arch/i386/Kconfig.debug~schedstat-v10 arch/i386/Kconfig.debug
--- 25/arch/i386/Kconfig.debug~schedstat-v10 2004-08-19 23:25:31.740221896 -0700
+++ 25-akpm/arch/i386/Kconfig.debug 2004-08-19 23:25:31.758219160 -0700
@@ -62,6 +62,18 @@ config LOCKMETER
Say Y to enable kernel lock metering, which adds overhead to SMP locks,
but allows you to see various statistics using the lockstat command.
+config SCHEDSTATS
+ bool "Collect scheduler statistics"
+ depends on DEBUG_KERNEL && PROC_FS
+ help
+ If you say Y here, additional code will be inserted into the
+ scheduler and related routines to collect statistics about
+ scheduler behavior and provide them in /proc/schedstat. These
+ stats may be useful for both tuning and debugging the scheduler
+ If you aren't debugging the scheduler or trying to tune a specific
+ application, you can say N to avoid the very slight overhead
+ this adds.
+
config X86_FIND_SMP_CONFIG
bool
depends on X86_LOCAL_APIC || X86_VOYAGER
diff -puN arch/ppc/Kconfig.debug~schedstat-v10 arch/ppc/Kconfig.debug
--- 25/arch/ppc/Kconfig.debug~schedstat-v10 2004-08-19 23:25:31.741221744 -0700
+++ 25-akpm/arch/ppc/Kconfig.debug 2004-08-19 23:25:31.758219160 -0700
@@ -53,6 +53,18 @@ config BDI_SWITCH
Unless you are intending to debug the kernel with one of these
machines, say N here.
+config SCHEDSTATS
+ bool "Collect scheduler statistics"
+ depends on DEBUG_KERNEL && PROC_FS
+ help
+ If you say Y here, additional code will be inserted into the
+ scheduler and related routines to collect statistics about
+ scheduler behavior and provide them in /proc/schedstat. These
+ stats may be useful for both tuning and debugging the scheduler
+ If you aren't debugging the scheduler or trying to tune a specific
+ application, you can say N to avoid the very slight overhead
+ this adds.
+
config BOOTX_TEXT
bool "Support for early boot text console (BootX or OpenFirmware only)"
depends PPC_OF
diff -puN arch/x86_64/Kconfig.debug~schedstat-v10 arch/x86_64/Kconfig.debug
--- 25/arch/x86_64/Kconfig.debug~schedstat-v10 2004-08-19 23:25:31.742221592 -0700
+++ 25-akpm/arch/x86_64/Kconfig.debug 2004-08-19 23:25:31.759219008 -0700
@@ -18,6 +18,18 @@ config INIT_DEBUG
Fill __init and __initdata at the end of boot. This helps debugging
illegal uses of __init and __initdata after initialization.
+config SCHEDSTATS
+ bool "Collect scheduler statistics"
+ depends on DEBUG_KERNEL && PROC_FS
+ help
+ If you say Y here, additional code will be inserted into the
+ scheduler and related routines to collect statistics about
+ scheduler behavior and provide them in /proc/schedstat. These
+ stats may be useful for both tuning and debugging the scheduler
+ If you aren't debugging the scheduler or trying to tune a specific
+ application, you can say N to avoid the very slight overhead
+ this adds.
+
config FRAME_POINTER
bool "Compile the kernel with frame pointers"
help
_