linux/kernel/delayacct.c
Yang Yang a3b2aeac9d delayacct: track delays from IRQ/SOFTIRQ
Delay accounting does not track the delay of IRQ/SOFTIRQ.  While
IRQ/SOFTIRQ could have obvious impact on some workloads productivity, such
as when workloads are running on system which is busy handling network
IRQ/SOFTIRQ.

Get the delay of IRQ/SOFTIRQ could help users to reduce such delay.  Such
as setting interrupt affinity or task affinity, using kernel thread for
NAPI etc.  This is inspired by "sched/psi: Add PSI_IRQ to track
IRQ/SOFTIRQ pressure"[1].  Also fix some code indent problems of older
code.

And update tools/accounting/getdelays.c:
    / # ./getdelays -p 156 -di
    print delayacct stats ON
    printing IO accounting
    PID     156

    CPU             count     real total  virtual total    delay total  delay average
                       15       15836008       16218149      275700790         18.380ms
    IO              count    delay total  delay average
                        0              0          0.000ms
    SWAP            count    delay total  delay average
                        0              0          0.000ms
    RECLAIM         count    delay total  delay average
                        0              0          0.000ms
    THRASHING       count    delay total  delay average
                        0              0          0.000ms
    COMPACT         count    delay total  delay average
                        0              0          0.000ms
    WPCOPY          count    delay total  delay average
                       36        7586118          0.211ms
    IRQ             count    delay total  delay average
                       42         929161          0.022ms

[1] commit 52b1364ba0b1("sched/psi: Add PSI_IRQ to track IRQ/SOFTIRQ pressure")

Link: https://lkml.kernel.org/r/202304081728353557233@zte.com.cn
Signed-off-by: Yang Yang <yang.yang29@zte.com.cn>
Cc: Jiang Xuexin <jiang.xuexin@zte.com.cn>
Cc: wangyong <wang.yong12@zte.com.cn>
Cc: junhua huang <huang.junhua@zte.com.cn>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
2023-04-18 16:39:34 -07:00

291 lines
7.4 KiB
C

// SPDX-License-Identifier: GPL-2.0-or-later
/* delayacct.c - per-task delay accounting
*
* Copyright (C) Shailabh Nagar, IBM Corp. 2006
*/
#include <linux/sched.h>
#include <linux/sched/task.h>
#include <linux/sched/cputime.h>
#include <linux/sched/clock.h>
#include <linux/slab.h>
#include <linux/taskstats.h>
#include <linux/sysctl.h>
#include <linux/delayacct.h>
#include <linux/module.h>
DEFINE_STATIC_KEY_FALSE(delayacct_key);
int delayacct_on __read_mostly; /* Delay accounting turned on/off */
struct kmem_cache *delayacct_cache;
static void set_delayacct(bool enabled)
{
if (enabled) {
static_branch_enable(&delayacct_key);
delayacct_on = 1;
} else {
delayacct_on = 0;
static_branch_disable(&delayacct_key);
}
}
static int __init delayacct_setup_enable(char *str)
{
delayacct_on = 1;
return 1;
}
__setup("delayacct", delayacct_setup_enable);
void delayacct_init(void)
{
delayacct_cache = KMEM_CACHE(task_delay_info, SLAB_PANIC|SLAB_ACCOUNT);
delayacct_tsk_init(&init_task);
set_delayacct(delayacct_on);
}
#ifdef CONFIG_PROC_SYSCTL
static int sysctl_delayacct(struct ctl_table *table, int write, void *buffer,
size_t *lenp, loff_t *ppos)
{
int state = delayacct_on;
struct ctl_table t;
int err;
if (write && !capable(CAP_SYS_ADMIN))
return -EPERM;
t = *table;
t.data = &state;
err = proc_dointvec_minmax(&t, write, buffer, lenp, ppos);
if (err < 0)
return err;
if (write)
set_delayacct(state);
return err;
}
static struct ctl_table kern_delayacct_table[] = {
{
.procname = "task_delayacct",
.data = NULL,
.maxlen = sizeof(unsigned int),
.mode = 0644,
.proc_handler = sysctl_delayacct,
.extra1 = SYSCTL_ZERO,
.extra2 = SYSCTL_ONE,
},
{ }
};
static __init int kernel_delayacct_sysctls_init(void)
{
register_sysctl_init("kernel", kern_delayacct_table);
return 0;
}
late_initcall(kernel_delayacct_sysctls_init);
#endif
void __delayacct_tsk_init(struct task_struct *tsk)
{
tsk->delays = kmem_cache_zalloc(delayacct_cache, GFP_KERNEL);
if (tsk->delays)
raw_spin_lock_init(&tsk->delays->lock);
}
/*
* Finish delay accounting for a statistic using its timestamps (@start),
* accumalator (@total) and @count
*/
static void delayacct_end(raw_spinlock_t *lock, u64 *start, u64 *total, u32 *count)
{
s64 ns = local_clock() - *start;
unsigned long flags;
if (ns > 0) {
raw_spin_lock_irqsave(lock, flags);
*total += ns;
(*count)++;
raw_spin_unlock_irqrestore(lock, flags);
}
}
void __delayacct_blkio_start(void)
{
current->delays->blkio_start = local_clock();
}
/*
* We cannot rely on the `current` macro, as we haven't yet switched back to
* the process being woken.
*/
void __delayacct_blkio_end(struct task_struct *p)
{
delayacct_end(&p->delays->lock,
&p->delays->blkio_start,
&p->delays->blkio_delay,
&p->delays->blkio_count);
}
int delayacct_add_tsk(struct taskstats *d, struct task_struct *tsk)
{
u64 utime, stime, stimescaled, utimescaled;
unsigned long long t2, t3;
unsigned long flags, t1;
s64 tmp;
task_cputime(tsk, &utime, &stime);
tmp = (s64)d->cpu_run_real_total;
tmp += utime + stime;
d->cpu_run_real_total = (tmp < (s64)d->cpu_run_real_total) ? 0 : tmp;
task_cputime_scaled(tsk, &utimescaled, &stimescaled);
tmp = (s64)d->cpu_scaled_run_real_total;
tmp += utimescaled + stimescaled;
d->cpu_scaled_run_real_total =
(tmp < (s64)d->cpu_scaled_run_real_total) ? 0 : tmp;
/*
* No locking available for sched_info (and too expensive to add one)
* Mitigate by taking snapshot of values
*/
t1 = tsk->sched_info.pcount;
t2 = tsk->sched_info.run_delay;
t3 = tsk->se.sum_exec_runtime;
d->cpu_count += t1;
tmp = (s64)d->cpu_delay_total + t2;
d->cpu_delay_total = (tmp < (s64)d->cpu_delay_total) ? 0 : tmp;
tmp = (s64)d->cpu_run_virtual_total + t3;
d->cpu_run_virtual_total =
(tmp < (s64)d->cpu_run_virtual_total) ? 0 : tmp;
if (!tsk->delays)
return 0;
/* zero XXX_total, non-zero XXX_count implies XXX stat overflowed */
raw_spin_lock_irqsave(&tsk->delays->lock, flags);
tmp = d->blkio_delay_total + tsk->delays->blkio_delay;
d->blkio_delay_total = (tmp < d->blkio_delay_total) ? 0 : tmp;
tmp = d->swapin_delay_total + tsk->delays->swapin_delay;
d->swapin_delay_total = (tmp < d->swapin_delay_total) ? 0 : tmp;
tmp = d->freepages_delay_total + tsk->delays->freepages_delay;
d->freepages_delay_total = (tmp < d->freepages_delay_total) ? 0 : tmp;
tmp = d->thrashing_delay_total + tsk->delays->thrashing_delay;
d->thrashing_delay_total = (tmp < d->thrashing_delay_total) ? 0 : tmp;
tmp = d->compact_delay_total + tsk->delays->compact_delay;
d->compact_delay_total = (tmp < d->compact_delay_total) ? 0 : tmp;
tmp = d->wpcopy_delay_total + tsk->delays->wpcopy_delay;
d->wpcopy_delay_total = (tmp < d->wpcopy_delay_total) ? 0 : tmp;
tmp = d->irq_delay_total + tsk->delays->irq_delay;
d->irq_delay_total = (tmp < d->irq_delay_total) ? 0 : tmp;
d->blkio_count += tsk->delays->blkio_count;
d->swapin_count += tsk->delays->swapin_count;
d->freepages_count += tsk->delays->freepages_count;
d->thrashing_count += tsk->delays->thrashing_count;
d->compact_count += tsk->delays->compact_count;
d->wpcopy_count += tsk->delays->wpcopy_count;
d->irq_count += tsk->delays->irq_count;
raw_spin_unlock_irqrestore(&tsk->delays->lock, flags);
return 0;
}
__u64 __delayacct_blkio_ticks(struct task_struct *tsk)
{
__u64 ret;
unsigned long flags;
raw_spin_lock_irqsave(&tsk->delays->lock, flags);
ret = nsec_to_clock_t(tsk->delays->blkio_delay);
raw_spin_unlock_irqrestore(&tsk->delays->lock, flags);
return ret;
}
void __delayacct_freepages_start(void)
{
current->delays->freepages_start = local_clock();
}
void __delayacct_freepages_end(void)
{
delayacct_end(&current->delays->lock,
&current->delays->freepages_start,
&current->delays->freepages_delay,
&current->delays->freepages_count);
}
void __delayacct_thrashing_start(bool *in_thrashing)
{
*in_thrashing = !!current->in_thrashing;
if (*in_thrashing)
return;
current->in_thrashing = 1;
current->delays->thrashing_start = local_clock();
}
void __delayacct_thrashing_end(bool *in_thrashing)
{
if (*in_thrashing)
return;
current->in_thrashing = 0;
delayacct_end(&current->delays->lock,
&current->delays->thrashing_start,
&current->delays->thrashing_delay,
&current->delays->thrashing_count);
}
void __delayacct_swapin_start(void)
{
current->delays->swapin_start = local_clock();
}
void __delayacct_swapin_end(void)
{
delayacct_end(&current->delays->lock,
&current->delays->swapin_start,
&current->delays->swapin_delay,
&current->delays->swapin_count);
}
void __delayacct_compact_start(void)
{
current->delays->compact_start = local_clock();
}
void __delayacct_compact_end(void)
{
delayacct_end(&current->delays->lock,
&current->delays->compact_start,
&current->delays->compact_delay,
&current->delays->compact_count);
}
void __delayacct_wpcopy_start(void)
{
current->delays->wpcopy_start = local_clock();
}
void __delayacct_wpcopy_end(void)
{
delayacct_end(&current->delays->lock,
&current->delays->wpcopy_start,
&current->delays->wpcopy_delay,
&current->delays->wpcopy_count);
}
void __delayacct_irq(struct task_struct *task, u32 delta)
{
unsigned long flags;
raw_spin_lock_irqsave(&task->delays->lock, flags);
task->delays->irq_delay += delta;
task->delays->irq_count++;
raw_spin_unlock_irqrestore(&task->delays->lock, flags);
}