fs/proc/uptime.c: fix idle time reporting in /proc/uptime
Subsystems (2)
PATCH v2 1/1 fs/proc/uptime.c: fix idle time reporting in /proc/uptime
/proc/uptime reports idle time by reading the CPUTIME_IDLE field from the per-cpu kcpustats. However, on NO_HZ systems, idle time is not continually updated on idle cpus, leading this value to appear incorrectly small.
/proc/stat performs an accounting update when reading idle time; we can use the same approach for uptime.
With this patch, /proc/stat and /proc/uptime now agree on idle time. Additionally, the following shows idle time tick up consistently on an idle machine:
(while true; do cat /proc/uptime; sleep 1; done) | awk '{print $2-prev; prev=$2}'
---
v2:
- Move get_idle_time() from cputime.c back into stat.c
- Use kcpustat_cpu_fetch
fs/proc/stat.c | 4 ++-- fs/proc/uptime.c | 14 +++++++++----- include/linux/kernel_stat.h | 1 + 3 files changed, 12 insertions(+), 7 deletions(-)
fs/proc/stat.c +2 -2
@@ -24,7 +24,7 @@#ifdef arch_idle_time -static u64 get_idle_time (struct kernel_cpustat *kcs,int cpu) +u64 get_idle_time (struct kernel_cpustat *kcs,int cpu) {u64 idle;
@@ -46,7 +46,7 @@ static u64 get_iowait_time(struct kernel_cpustat *kcs, int cpu)#else -static u64 get_idle_time (struct kernel_cpustat *kcs,int cpu) +u64 get_idle_time (struct kernel_cpustat *kcs,int cpu) {u64 idle, idle_usecs =-1ULL ;
fs/proc/uptime.c +9 -5
@@ -12,18 +12,22 @@ static int uptime_proc_show(struct seq_file *m, void *v) {struct timespec64 uptime;struct timespec64 idle; -u64 nsec; +u64 idle_nsec;u32 rem;int i; - nsec =0 ; -for_each_possible_cpu (i) - nsec += (__force u64 )kcpustat_cpu (i).cpustat[CPUTIME_IDLE ]; + idle_nsec =0 ; +for_each_possible_cpu (i) { +struct kernel_cpustat kcs; + +kcpustat_cpu_fetch (&kcs, i); + idle_nsec +=get_idle_time (&kcs, i); + }ktime_get_boottime_ts64 (&uptime);timens_add_boottime (&uptime); - idle.tv_sec =div_u64_rem (nsec,NSEC_PER_SEC , &rem); + idle.tv_sec =div_u64_rem (idle_nsec,NSEC_PER_SEC , &rem); idle.tv_nsec = rem; seq_printf(m, "%lu.%02lu %lu.%02lu\n", (unsigned long) uptime.tv_sec,
include/linux/kernel_stat.h +1 -0
@@ -102,6 +102,7 @@ extern void account_system_index_time(struct task_struct *, u64, enum cpu_usage_stat);extern void account_steal_time (u64 );extern void account_idle_time (u64 ); +extern u64 get_idle_time (struct kernel_cpustat *kcs,int cpu);#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE static inline void account_process_tick (struct task_struct *tsk,int user)
Signature · 1 line
2.33.0.259.gc128427fd7-goog
Quoted 9 lines · Josh Don
/proc/uptimereports idle time by reading theCPUTIME_IDLEfield from the per-cpu kcpustats. However, onNO_HZsystems, idle time is not continually updated on idle cpus, leading this value to appear incorrectly small.
/proc/statperforms an accounting update when reading idle time; we can use the same approach for uptime.
Quoted 9 lines
With this patch,
/proc/statand/proc/uptimenow agree on idle time. Additionally, the following shows idle time tick up consistently on an idle machine:
(while true; do cat/proc/uptime; sleep 1; done) | awk '{print $2-prev; prev=$2}'Reported-by Luigi Rizzo <[email protected]>Signed-off-by Josh Don <[email protected]>---
SGTM thanks !
[tip: sched/core] fs/proc/uptime.c: Fix idle time reporting in /proc/uptime
In reply to Josh DonThe following commit has been merged into the sched/core branch of tip:
The following commit has been merged into the sched/core branch of tip:
Commit-ID: 1f74f9ea5a0cc9147f5bda4b5715ccb5cc34605d Gitweb: https://git.kernel.org/tip/1f74f9ea5a0cc9147f5bda4b5715ccb5cc34605d Author: Josh Don <[email protected]> AuthorDate: Fri, 27 Aug 2021 09:54:38 -07:00 Committer: Peter Zijlstra <[email protected]> CommitterDate: Thu, 09 Sep 2021 11:27:31 +02:00
fs/proc/uptime.c: Fix idle time reporting in /proc/uptime
/proc/uptime reports idle time by reading the CPUTIME_IDLE field from the per-cpu kcpustats. However, on NO_HZ systems, idle time is not continually updated on idle cpus, leading this value to appear incorrectly small.
/proc/stat performs an accounting update when reading idle time; we can use the same approach for uptime.
With this patch, /proc/stat and /proc/uptime now agree on idle time. Additionally, the following shows idle time tick up consistently on an idle machine:
(while true; do cat /proc/uptime; sleep 1; done) | awk '{print $2-prev; prev=$2}'
--- fs/proc/stat.c | 4 ++-- fs/proc/uptime.c | 14 +++++++++----- include/linux/kernel_stat.h | 1 + 3 files changed, 12 insertions(+), 7 deletions(-)
fs/proc/stat.c +2 -2
@@ -24,7 +24,7 @@#ifdef arch_idle_time -static u64 get_idle_time (struct kernel_cpustat *kcs,int cpu) +u64 get_idle_time (struct kernel_cpustat *kcs,int cpu) {u64 idle;
@@ -46,7 +46,7 @@ static u64 get_iowait_time(struct kernel_cpustat *kcs, int cpu)#else -static u64 get_idle_time (struct kernel_cpustat *kcs,int cpu) +u64 get_idle_time (struct kernel_cpustat *kcs,int cpu) {u64 idle, idle_usecs =-1ULL ;
fs/proc/uptime.c +9 -5
@@ -12,18 +12,22 @@ static int uptime_proc_show(struct seq_file *m, void *v) {struct timespec64 uptime;struct timespec64 idle; -u64 nsec; +u64 idle_nsec;u32 rem;int i; - nsec =0 ; -for_each_possible_cpu (i) - nsec += (__force u64 )kcpustat_cpu (i).cpustat[CPUTIME_IDLE ]; + idle_nsec =0 ; +for_each_possible_cpu (i) { +struct kernel_cpustat kcs; + +kcpustat_cpu_fetch (&kcs, i); + idle_nsec +=get_idle_time (&kcs, i); + }ktime_get_boottime_ts64 (&uptime);timens_add_boottime (&uptime); - idle.tv_sec =div_u64_rem (nsec,NSEC_PER_SEC , &rem); + idle.tv_sec =div_u64_rem (idle_nsec,NSEC_PER_SEC , &rem); idle.tv_nsec = rem; seq_printf(m, "%lu.%02lu %lu.%02lu\n", (unsigned long) uptime.tv_sec,
include/linux/kernel_stat.h +1 -0
@@ -102,6 +102,7 @@ extern void account_system_index_time(struct task_struct *, u64, enum cpu_usage_stat);extern void account_steal_time (u64 );extern void account_idle_time (u64 ); +extern u64 get_idle_time (struct kernel_cpustat *kcs,int cpu);#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE static inline void account_process_tick (struct task_struct *tsk,int user)
[tip: sched/core] fs/proc/uptime.c: Fix idle time reporting in /proc/uptime
In reply to Josh DonThe following commit has been merged into the sched/core branch of tip:
The following commit has been merged into the sched/core branch of tip:
Commit-ID: a130e8fbc7de796eb6e680724d87f4737a26d0ac Gitweb: https://git.kernel.org/tip/a130e8fbc7de796eb6e680724d87f4737a26d0ac Author: Josh Don <[email protected]> AuthorDate: Fri, 27 Aug 2021 09:54:38 -07:00 Committer: Peter Zijlstra <[email protected]> CommitterDate: Tue, 05 Oct 2021 15:51:35 +02:00
fs/proc/uptime.c: Fix idle time reporting in /proc/uptime
/proc/uptime reports idle time by reading the CPUTIME_IDLE field from the per-cpu kcpustats. However, on NO_HZ systems, idle time is not continually updated on idle cpus, leading this value to appear incorrectly small.
/proc/stat performs an accounting update when reading idle time; we can use the same approach for uptime.
With this patch, /proc/stat and /proc/uptime now agree on idle time. Additionally, the following shows idle time tick up consistently on an idle machine:
(while true; do cat /proc/uptime; sleep 1; done) | awk '{print $2-prev; prev=$2}'
--- fs/proc/stat.c | 4 ++-- fs/proc/uptime.c | 14 +++++++++----- include/linux/kernel_stat.h | 1 + 3 files changed, 12 insertions(+), 7 deletions(-)
fs/proc/stat.c +2 -2
@@ -24,7 +24,7 @@#ifdef arch_idle_time -static u64 get_idle_time (struct kernel_cpustat *kcs,int cpu) +u64 get_idle_time (struct kernel_cpustat *kcs,int cpu) {u64 idle;
@@ -46,7 +46,7 @@ static u64 get_iowait_time(struct kernel_cpustat *kcs, int cpu)#else -static u64 get_idle_time (struct kernel_cpustat *kcs,int cpu) +u64 get_idle_time (struct kernel_cpustat *kcs,int cpu) {u64 idle, idle_usecs =-1ULL ;
fs/proc/uptime.c +9 -5
@@ -12,18 +12,22 @@ static int uptime_proc_show(struct seq_file *m, void *v) {struct timespec64 uptime;struct timespec64 idle; -u64 nsec; +u64 idle_nsec;u32 rem;int i; - nsec =0 ; -for_each_possible_cpu (i) - nsec += (__force u64 )kcpustat_cpu (i).cpustat[CPUTIME_IDLE ]; + idle_nsec =0 ; +for_each_possible_cpu (i) { +struct kernel_cpustat kcs; + +kcpustat_cpu_fetch (&kcs, i); + idle_nsec +=get_idle_time (&kcs, i); + }ktime_get_boottime_ts64 (&uptime);timens_add_boottime (&uptime); - idle.tv_sec =div_u64_rem (nsec,NSEC_PER_SEC , &rem); + idle.tv_sec =div_u64_rem (idle_nsec,NSEC_PER_SEC , &rem); idle.tv_nsec = rem; seq_printf(m, "%lu.%02lu %lu.%02lu\n", (unsigned long) uptime.tv_sec,
include/linux/kernel_stat.h +1 -0
@@ -102,6 +102,7 @@ extern void account_system_index_time(struct task_struct *, u64, enum cpu_usage_stat);extern void account_steal_time (u64 );extern void account_idle_time (u64 ); +extern u64 get_idle_time (struct kernel_cpustat *kcs,int cpu);#ifdef CONFIG_VIRT_CPU_ACCOUNTING_NATIVE static inline void account_process_tick (struct task_struct *tsk,int user)