Skip to contentSkip to conversation

sched: make nr_running() return "unsigned int"

PATCH In review Alexey Dobriyan 2 participants 4 messages Mar 11, 2020 – Mar 12, 2020

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/2 sched: make nr_running() return "unsigned int"

I don't anyone have been crazy enough to spawn 2^32 threads. It'd require absurd amounts of physical memory.

Meanwhile save few bits on REX prefixes.

And remove "extern" from prototypes while I'm at it.

---

 fs/proc/loadavg.c          |    2 +-
 fs/proc/stat.c             |    2 +-
 include/linux/sched/stat.h |    2 +-
 kernel/sched/core.c        |    4 ++--
 4 files changed, 5 insertions(+), 5 deletions(-)
fs/proc/loadavg.c +1 -1
@@ -16,7 +16,7 @@ static int loadavg_proc_show(struct seq_file *m, void *v)
 
 	get_avenrun(avnrun, FIXED_1/200, 0);
 
-	seq_printf(m, "%lu.%02lu %lu.%02lu %lu.%02lu %ld/%d %d\n",
+	seq_printf(m, "%lu.%02lu %lu.%02lu %lu.%02lu %u/%d %d\n",
 		LOAD_INT(avnrun[0]), LOAD_FRAC(avnrun[0]),
 		LOAD_INT(avnrun[1]), LOAD_FRAC(avnrun[1]),
 		LOAD_INT(avnrun[2]), LOAD_FRAC(avnrun[2]),
fs/proc/stat.c +1 -1
@@ -197,7 +197,7 @@ static int show_stat(struct seq_file *p, void *v)
 		"\nctxt %llu\n"
 		"btime %llu\n"
 		"processes %lu\n"
-		"procs_running %lu\n"
+		"procs_running %u\n"
 		"procs_blocked %lu\n",
 		nr_context_switches(),
 		(unsigned long long)boottime.tv_sec,
include/linux/sched/stat.h +1 -1
@@ -16,7 +16,7 @@ extern unsigned long total_forks;
 extern int nr_threads;
 DECLARE_PER_CPU(unsigned long, process_counts);
 extern int nr_processes(void);
-extern unsigned long nr_running(void);
+unsigned int nr_running(void);
 extern bool single_task_running(void);
 extern unsigned long nr_iowait(void);
 extern unsigned long nr_iowait_cpu(int cpu);
kernel/sched/core.c +2 -2
@@ -3389,9 +3389,9 @@ context_switch(struct rq *rq, struct task_struct *prev,
  * externally visible scheduler statistics: current number of runnable
  * threads, total number of context switches performed since bootup.
  */
-unsigned long nr_running(void)
+unsigned int nr_running(void)
 {
-	unsigned long i, sum = 0;
+	unsigned int i, sum = 0;
 
 	for_each_online_cpu(i)
 		sum += cpu_rq(i)->nr_running;
#1 lore raw

PATCH 2/2 sched: make nr_iowait_cpu() return "unsigned int"

Same logic: 2^32 threads stuck waiting in runqueue implies 2^32+ processes total which is absurd.

Per-runqueue ->nr_iowait member being 32-bit hints that it is correct change!

---

 fs/proc/stat.c             |    2 +-
 include/linux/sched/stat.h |    4 ++--
 kernel/sched/core.c        |    6 +++---
 3 files changed, 6 insertions(+), 6 deletions(-)
fs/proc/stat.c +1 -1
@@ -198,7 +198,7 @@ static int show_stat(struct seq_file *p, void *v)
 		"btime %llu\n"
 		"processes %lu\n"
 		"procs_running %u\n"
-		"procs_blocked %lu\n",
+		"procs_blocked %u\n",
 		nr_context_switches(),
 		(unsigned long long)boottime.tv_sec,
 		total_forks,
include/linux/sched/stat.h +2 -2
@@ -18,8 +18,8 @@ DECLARE_PER_CPU(unsigned long, process_counts);
 extern int nr_processes(void);
 unsigned int nr_running(void);
 extern bool single_task_running(void);
-extern unsigned long nr_iowait(void);
-extern unsigned long nr_iowait_cpu(int cpu);
+unsigned int nr_iowait(void);
+unsigned int nr_iowait_cpu(int cpu);
 
 static inline int sched_info_on(void)
 {
kernel/sched/core.c +3 -3
@@ -3436,7 +3436,7 @@ unsigned long long nr_context_switches(void)
  * it does become runnable.
  */
 
-unsigned long nr_iowait_cpu(int cpu)
+unsigned int nr_iowait_cpu(int cpu)
 {
 	return atomic_read(&cpu_rq(cpu)->nr_iowait);
 }
@@ -3471,9 +3471,9 @@ unsigned long nr_iowait_cpu(int cpu)
  * Task CPU affinities can make all that even more 'interesting'.
  */
 
-unsigned long nr_iowait(void)
+unsigned int nr_iowait(void)
 {
-	unsigned long i, sum = 0;
+	unsigned int i, sum = 0;
 
 	for_each_possible_cpu(i)
 		sum += nr_iowait_cpu(i);
maintainer
#2 lore raw
In reply to Alexey Dobriyan
Quoted 2 lines · Alexey Dobriyan

I don't anyone have been crazy enough to spawn 2^32 threads. It'd require absurd amounts of physical memory.

And we're going to 5 level page-tables because 48 bits physical isn't enough. 57 bits of physical is plenty space to spawn that many tasks and still have some left over.

Now 32 bit tasks is indeed insane, but memory isn't the problem. The actual limit is the pid-space, which is 30 bits.

#3 lore raw
Quoted 10 lines · Peter Zijlstra
Quoted 2 lines · Alexey Dobriyan

I don't anyone have been crazy enough to spawn 2^32 threads. It'd require absurd amounts of physical memory.

And we're going to 5 level page-tables because 48 bits physical isn't enough. 57 bits of physical is plenty space to spawn that many tasks and still have some left over.

Now 32 bit tasks is indeed insane, but memory isn't the problem. The actual limit is the pid-space, which is 30 bits.

Indeed, pid space limit is even better!