1 files changed, 229 insertions, 206 deletions

diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index 486f403a778b..9c882f20803e 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -8,8 +8,6 @@
  */
 #include "sched.h"
 
-static DEFINE_SPINLOCK(sched_debug_lock);
-
 /*
  * This allows printing both to /proc/sched_debug and
  * to the console
@@ -169,245 +167,258 @@ static const struct file_operations sched_feat_fops = {
 	.release	= single_release,
 };
 
-__read_mostly bool sched_debug_enabled;
+#ifdef CONFIG_SMP
 
-static __init int sched_init_debug(void)
+static ssize_t sched_scaling_write(struct file *filp, const char __user *ubuf,
+				   size_t cnt, loff_t *ppos)
 {
-	debugfs_create_file("sched_features", 0644, NULL, NULL,
-			&sched_feat_fops);
+	char buf[16];
 
-	debugfs_create_bool("sched_debug", 0644, NULL,
-			&sched_debug_enabled);
+	if (cnt > 15)
+		cnt = 15;
 
-	return 0;
+	if (copy_from_user(&buf, ubuf, cnt))
+		return -EFAULT;
+
+	if (kstrtouint(buf, 10, &sysctl_sched_tunable_scaling))
+		return -EINVAL;
+
+	if (sched_update_scaling())
+		return -EINVAL;
+
+	*ppos += cnt;
+	return cnt;
 }
-late_initcall(sched_init_debug);
 
-#ifdef CONFIG_SMP
+static int sched_scaling_show(struct seq_file *m, void *v)
+{
+	seq_printf(m, "%d\n", sysctl_sched_tunable_scaling);
+	return 0;
+}
 
-#ifdef CONFIG_SYSCTL
+static int sched_scaling_open(struct inode *inode, struct file *filp)
+{
+	return single_open(filp, sched_scaling_show, NULL);
+}
 
-static struct ctl_table sd_ctl_dir[] = {
-	{
-		.procname	= "sched_domain",
-		.mode		= 0555,
-	},
-	{}
+static const struct file_operations sched_scaling_fops = {
+	.open		= sched_scaling_open,
+	.write		= sched_scaling_write,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
 };
 
-static struct ctl_table sd_ctl_root[] = {
-	{
-		.procname	= "kernel",
-		.mode		= 0555,
-		.child		= sd_ctl_dir,
-	},
-	{}
-};
+#endif /* SMP */
 
-static struct ctl_table *sd_alloc_ctl_entry(int n)
+#ifdef CONFIG_PREEMPT_DYNAMIC
+
+static ssize_t sched_dynamic_write(struct file *filp, const char __user *ubuf,
+				   size_t cnt, loff_t *ppos)
 {
-	struct ctl_table *entry =
-		kcalloc(n, sizeof(struct ctl_table), GFP_KERNEL);
+	char buf[16];
+	int mode;
+
+	if (cnt > 15)
+		cnt = 15;
+
+	if (copy_from_user(&buf, ubuf, cnt))
+		return -EFAULT;
 
-	return entry;
+	buf[cnt] = 0;
+	mode = sched_dynamic_mode(strstrip(buf));
+	if (mode < 0)
+		return mode;
+
+	sched_dynamic_update(mode);
+
+	*ppos += cnt;
+
+	return cnt;
 }
 
-static void sd_free_ctl_entry(struct ctl_table **tablep)
+static int sched_dynamic_show(struct seq_file *m, void *v)
 {
-	struct ctl_table *entry;
-
-	/*
-	 * In the intermediate directories, both the child directory and
-	 * procname are dynamically allocated and could fail but the mode
-	 * will always be set. In the lowest directory the names are
-	 * static strings and all have proc handlers.
-	 */
-	for (entry = *tablep; entry->mode; entry++) {
-		if (entry->child)
-			sd_free_ctl_entry(&entry->child);
-		if (entry->proc_handler == NULL)
-			kfree(entry->procname);
+	static const char * preempt_modes[] = {
+		"none", "voluntary", "full"
+	};
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(preempt_modes); i++) {
+		if (preempt_dynamic_mode == i)
+			seq_puts(m, "(");
+		seq_puts(m, preempt_modes[i]);
+		if (preempt_dynamic_mode == i)
+			seq_puts(m, ")");
+
+		seq_puts(m, " ");
 	}
 
-	kfree(*tablep);
-	*tablep = NULL;
+	seq_puts(m, "\n");
+	return 0;
 }
 
-static void
-set_table_entry(struct ctl_table *entry,
-		const char *procname, void *data, int maxlen,
-		umode_t mode, proc_handler *proc_handler)
+static int sched_dynamic_open(struct inode *inode, struct file *filp)
 {
-	entry->procname = procname;
-	entry->data = data;
-	entry->maxlen = maxlen;
-	entry->mode = mode;
-	entry->proc_handler = proc_handler;
+	return single_open(filp, sched_dynamic_show, NULL);
 }
 
-static int sd_ctl_doflags(struct ctl_table *table, int write,
-			  void *buffer, size_t *lenp, loff_t *ppos)
+static const struct file_operations sched_dynamic_fops = {
+	.open		= sched_dynamic_open,
+	.write		= sched_dynamic_write,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
+#endif /* CONFIG_PREEMPT_DYNAMIC */
+
+__read_mostly bool sched_debug_verbose;
+
+static const struct seq_operations sched_debug_sops;
+
+static int sched_debug_open(struct inode *inode, struct file *filp)
 {
-	unsigned long flags = *(unsigned long *)table->data;
-	size_t data_size = 0;
-	size_t len = 0;
-	char *tmp, *buf;
-	int idx;
+	return seq_open(filp, &sched_debug_sops);
+}
 
-	if (write)
-		return 0;
+static const struct file_operations sched_debug_fops = {
+	.open		= sched_debug_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= seq_release,
+};
 
-	for_each_set_bit(idx, &flags, __SD_FLAG_CNT) {
-		char *name = sd_flag_debug[idx].name;
+static struct dentry *debugfs_sched;
 
-		/* Name plus whitespace */
-		data_size += strlen(name) + 1;
-	}
+static __init int sched_init_debug(void)
+{
+	struct dentry __maybe_unused *numa;
 
-	if (*ppos > data_size) {
-		*lenp = 0;
-		return 0;
-	}
+	debugfs_sched = debugfs_create_dir("sched", NULL);
 
-	buf = kcalloc(data_size + 1, sizeof(*buf), GFP_KERNEL);
-	if (!buf)
-		return -ENOMEM;
+	debugfs_create_file("features", 0644, debugfs_sched, NULL, &sched_feat_fops);
+	debugfs_create_bool("verbose", 0644, debugfs_sched, &sched_debug_verbose);
+#ifdef CONFIG_PREEMPT_DYNAMIC
+	debugfs_create_file("preempt", 0644, debugfs_sched, NULL, &sched_dynamic_fops);
+#endif
 
-	for_each_set_bit(idx, &flags, __SD_FLAG_CNT) {
-		char *name = sd_flag_debug[idx].name;
+	debugfs_create_u32("latency_ns", 0644, debugfs_sched, &sysctl_sched_latency);
+	debugfs_create_u32("min_granularity_ns", 0644, debugfs_sched, &sysctl_sched_min_granularity);
+	debugfs_create_u32("wakeup_granularity_ns", 0644, debugfs_sched, &sysctl_sched_wakeup_granularity);
 
-		len += snprintf(buf + len, strlen(name) + 2, "%s ", name);
-	}
+	debugfs_create_u32("latency_warn_ms", 0644, debugfs_sched, &sysctl_resched_latency_warn_ms);
+	debugfs_create_u32("latency_warn_once", 0644, debugfs_sched, &sysctl_resched_latency_warn_once);
 
-	tmp = buf + *ppos;
-	len -= *ppos;
+#ifdef CONFIG_SMP
+	debugfs_create_file("tunable_scaling", 0644, debugfs_sched, NULL, &sched_scaling_fops);
+	debugfs_create_u32("migration_cost_ns", 0644, debugfs_sched, &sysctl_sched_migration_cost);
+	debugfs_create_u32("nr_migrate", 0644, debugfs_sched, &sysctl_sched_nr_migrate);
 
-	if (len > *lenp)
-		len = *lenp;
-	if (len)
-		memcpy(buffer, tmp, len);
-	if (len < *lenp) {
-		((char *)buffer)[len] = '\n';
-		len++;
-	}
+	mutex_lock(&sched_domains_mutex);
+	update_sched_domain_debugfs();
+	mutex_unlock(&sched_domains_mutex);
+#endif
 
-	*lenp = len;
-	*ppos += len;
+#ifdef CONFIG_NUMA_BALANCING
+	numa = debugfs_create_dir("numa_balancing", debugfs_sched);
 
-	kfree(buf);
+	debugfs_create_u32("scan_delay_ms", 0644, numa, &sysctl_numa_balancing_scan_delay);
+	debugfs_create_u32("scan_period_min_ms", 0644, numa, &sysctl_numa_balancing_scan_period_min);
+	debugfs_create_u32("scan_period_max_ms", 0644, numa, &sysctl_numa_balancing_scan_period_max);
+	debugfs_create_u32("scan_size_mb", 0644, numa, &sysctl_numa_balancing_scan_size);
+#endif
+
+	debugfs_create_file("debug", 0444, debugfs_sched, NULL, &sched_debug_fops);
 
 	return 0;
 }
+late_initcall(sched_init_debug);
+
+#ifdef CONFIG_SMP
 
-static struct ctl_table *
-sd_alloc_ctl_domain_table(struct sched_domain *sd)
+static cpumask_var_t		sd_sysctl_cpus;
+static struct dentry		*sd_dentry;
+
+static int sd_flags_show(struct seq_file *m, void *v)
 {
-	struct ctl_table *table = sd_alloc_ctl_entry(9);
-
-	if (table == NULL)
-		return NULL;
-
-	set_table_entry(&table[0], "min_interval",	  &sd->min_interval,	    sizeof(long), 0644, proc_doulongvec_minmax);
-	set_table_entry(&table[1], "max_interval",	  &sd->max_interval,	    sizeof(long), 0644, proc_doulongvec_minmax);
-	set_table_entry(&table[2], "busy_factor",	  &sd->busy_factor,	    sizeof(int),  0644, proc_dointvec_minmax);
-	set_table_entry(&table[3], "imbalance_pct",	  &sd->imbalance_pct,	    sizeof(int),  0644, proc_dointvec_minmax);
-	set_table_entry(&table[4], "cache_nice_tries",	  &sd->cache_nice_tries,    sizeof(int),  0644, proc_dointvec_minmax);
-	set_table_entry(&table[5], "flags",		  &sd->flags,		    sizeof(int),  0444, sd_ctl_doflags);
-	set_table_entry(&table[6], "max_newidle_lb_cost", &sd->max_newidle_lb_cost, sizeof(long), 0644, proc_doulongvec_minmax);
-	set_table_entry(&table[7], "name",		  sd->name,	       CORENAME_MAX_SIZE, 0444, proc_dostring);
-	/* &table[8] is terminator */
-
-	return table;
+	unsigned long flags = *(unsigned int *)m->private;
+	int idx;
+
+	for_each_set_bit(idx, &flags, __SD_FLAG_CNT) {
+		seq_puts(m, sd_flag_debug[idx].name);
+		seq_puts(m, " ");
+	}
+	seq_puts(m, "\n");
+
+	return 0;
 }
 
-static struct ctl_table *sd_alloc_ctl_cpu_table(int cpu)
+static int sd_flags_open(struct inode *inode, struct file *file)
 {
-	struct ctl_table *entry, *table;
-	struct sched_domain *sd;
-	int domain_num = 0, i;
-	char buf[32];
-
-	for_each_domain(cpu, sd)
-		domain_num++;
-	entry = table = sd_alloc_ctl_entry(domain_num + 1);
-	if (table == NULL)
-		return NULL;
-
-	i = 0;
-	for_each_domain(cpu, sd) {
-		snprintf(buf, 32, "domain%d", i);
-		entry->procname = kstrdup(buf, GFP_KERNEL);
-		entry->mode = 0555;
-		entry->child = sd_alloc_ctl_domain_table(sd);
-		entry++;
-		i++;
-	}
-	return table;
+	return single_open(file, sd_flags_show, inode->i_private);
 }
 
-static cpumask_var_t		sd_sysctl_cpus;
-static struct ctl_table_header	*sd_sysctl_header;
+static const struct file_operations sd_flags_fops = {
+	.open		= sd_flags_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
 
-void register_sched_domain_sysctl(void)
+static void register_sd(struct sched_domain *sd, struct dentry *parent)
 {
-	static struct ctl_table *cpu_entries;
-	static struct ctl_table **cpu_idx;
-	static bool init_done = false;
-	char buf[32];
-	int i;
+#define SDM(type, mode, member)	\
+	debugfs_create_##type(#member, mode, parent, &sd->member)
 
-	if (!cpu_entries) {
-		cpu_entries = sd_alloc_ctl_entry(num_possible_cpus() + 1);
-		if (!cpu_entries)
-			return;
+	SDM(ulong, 0644, min_interval);
+	SDM(ulong, 0644, max_interval);
+	SDM(u64,   0644, max_newidle_lb_cost);
+	SDM(u32,   0644, busy_factor);
+	SDM(u32,   0644, imbalance_pct);
+	SDM(u32,   0644, cache_nice_tries);
+	SDM(str,   0444, name);
 
-		WARN_ON(sd_ctl_dir[0].child);
-		sd_ctl_dir[0].child = cpu_entries;
-	}
+#undef SDM
 
-	if (!cpu_idx) {
-		struct ctl_table *e = cpu_entries;
-
-		cpu_idx = kcalloc(nr_cpu_ids, sizeof(struct ctl_table*), GFP_KERNEL);
-		if (!cpu_idx)
-			return;
+	debugfs_create_file("flags", 0444, parent, &sd->flags, &sd_flags_fops);
+}
 
-		/* deal with sparse possible map */
-		for_each_possible_cpu(i) {
-			cpu_idx[i] = e;
-			e++;
-		}
-	}
+void update_sched_domain_debugfs(void)
+{
+	int cpu, i;
 
 	if (!cpumask_available(sd_sysctl_cpus)) {
 		if (!alloc_cpumask_var(&sd_sysctl_cpus, GFP_KERNEL))
 			return;
-	}
-
-	if (!init_done) {
-		init_done = true;
-		/* init to possible to not have holes in @cpu_entries */
 		cpumask_copy(sd_sysctl_cpus, cpu_possible_mask);
 	}
 
-	for_each_cpu(i, sd_sysctl_cpus) {
-		struct ctl_table *e = cpu_idx[i];
+	if (!sd_dentry)
+		sd_dentry = debugfs_create_dir("domains", debugfs_sched);
 
-		if (e->child)
-			sd_free_ctl_entry(&e->child);
+	for_each_cpu(cpu, sd_sysctl_cpus) {
+		struct sched_domain *sd;
+		struct dentry *d_cpu;
+		char buf[32];
 
-		if (!e->procname) {
-			snprintf(buf, 32, "cpu%d", i);
-			e->procname = kstrdup(buf, GFP_KERNEL);
+		snprintf(buf, sizeof(buf), "cpu%d", cpu);
+		debugfs_remove(debugfs_lookup(buf, sd_dentry));
+		d_cpu = debugfs_create_dir(buf, sd_dentry);
+
+		i = 0;
+		for_each_domain(cpu, sd) {
+			struct dentry *d_sd;
+
+			snprintf(buf, sizeof(buf), "domain%d", i);
+			d_sd = debugfs_create_dir(buf, d_cpu);
+
+			register_sd(sd, d_sd);
+			i++;
 		}
-		e->mode = 0555;
-		e->child = sd_alloc_ctl_cpu_table(i);
 
-		__cpumask_clear_cpu(i, sd_sysctl_cpus);
+		__cpumask_clear_cpu(cpu, sd_sysctl_cpus);
 	}
-
-	WARN_ON(sd_sysctl_header);
-	sd_sysctl_header = register_sysctl_table(sd_ctl_root);
 }
 
 void dirty_sched_domain_sysctl(int cpu)
@@ -416,13 +427,6 @@ void dirty_sched_domain_sysctl(int cpu)
 		__cpumask_set_cpu(cpu, sd_sysctl_cpus);
 }
 
-/* may be called multiple times per register */
-void unregister_sched_domain_sysctl(void)
-{
-	unregister_sysctl_table(sd_sysctl_header);
-	sd_sysctl_header = NULL;
-}
-#endif /* CONFIG_SYSCTL */
 #endif /* CONFIG_SMP */
 
 #ifdef CONFIG_FAIR_GROUP_SCHED
@@ -470,16 +474,37 @@ static void print_cfs_group_stats(struct seq_file *m, int cpu, struct task_group
 #endif
 
 #ifdef CONFIG_CGROUP_SCHED
+static DEFINE_SPINLOCK(sched_debug_lock);
 static char group_path[PATH_MAX];
 
-static char *task_group_path(struct task_group *tg)
+static void task_group_path(struct task_group *tg, char *path, int plen)
 {
-	if (autogroup_path(tg, group_path, PATH_MAX))
-		return group_path;
+	if (autogroup_path(tg, path, plen))
+		return;
 
-	cgroup_path(tg->css.cgroup, group_path, PATH_MAX);
+	cgroup_path(tg->css.cgroup, path, plen);
+}
 
-	return group_path;
+/*
+ * Only 1 SEQ_printf_task_group_path() caller can use the full length
+ * group_path[] for cgroup path. Other simultaneous callers will have
+ * to use a shorter stack buffer. A "..." suffix is appended at the end
+ * of the stack buffer so that it will show up in case the output length
+ * matches the given buffer size to indicate possible path name truncation.
+ */
+#define SEQ_printf_task_group_path(m, tg, fmt...)			\
+{									\
+	if (spin_trylock(&sched_debug_lock)) {				\
+		task_group_path(tg, group_path, sizeof(group_path));	\
+		SEQ_printf(m, fmt, group_path);				\
+		spin_unlock(&sched_debug_lock);				\
+	} else {							\
+		char buf[128];						\
+		char *bufend = buf + sizeof(buf) - 3;			\
+		task_group_path(tg, buf, bufend - buf);			\
+		strcpy(bufend - 1, "...");				\
+		SEQ_printf(m, fmt, buf);				\
+	}								\
 }
 #endif
 
@@ -506,7 +531,7 @@ print_task(struct seq_file *m, struct rq *rq, struct task_struct *p)
 	SEQ_printf(m, " %d %d", task_node(p), task_numa_group_id(p));
 #endif
 #ifdef CONFIG_CGROUP_SCHED
-	SEQ_printf(m, " %s", task_group_path(task_group(p)));
+	SEQ_printf_task_group_path(m, task_group(p), " %s")
 #endif
 
 	SEQ_printf(m, "\n");
@@ -543,7 +568,7 @@ void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
 
 #ifdef CONFIG_FAIR_GROUP_SCHED
 	SEQ_printf(m, "\n");
-	SEQ_printf(m, "cfs_rq[%d]:%s\n", cpu, task_group_path(cfs_rq->tg));
+	SEQ_printf_task_group_path(m, cfs_rq->tg, "cfs_rq[%d]:%s\n", cpu);
 #else
 	SEQ_printf(m, "\n");
 	SEQ_printf(m, "cfs_rq[%d]:\n", cpu);
@@ -614,7 +639,7 @@ void print_rt_rq(struct seq_file *m, int cpu, struct rt_rq *rt_rq)
 {
 #ifdef CONFIG_RT_GROUP_SCHED
 	SEQ_printf(m, "\n");
-	SEQ_printf(m, "rt_rq[%d]:%s\n", cpu, task_group_path(rt_rq->tg));
+	SEQ_printf_task_group_path(m, rt_rq->tg, "rt_rq[%d]:%s\n", cpu);
 #else
 	SEQ_printf(m, "\n");
 	SEQ_printf(m, "rt_rq[%d]:\n", cpu);
@@ -666,7 +691,6 @@ void print_dl_rq(struct seq_file *m, int cpu, struct dl_rq *dl_rq)
 static void print_cpu(struct seq_file *m, int cpu)
 {
 	struct rq *rq = cpu_rq(cpu);
-	unsigned long flags;
 
 #ifdef CONFIG_X86
 	{
@@ -717,13 +741,11 @@ do {									\
 	}
 #undef P
 
-	spin_lock_irqsave(&sched_debug_lock, flags);
 	print_cfs_stats(m, cpu);
 	print_rt_stats(m, cpu);
 	print_dl_stats(m, cpu);
 
 	print_rq(m, rq, cpu);
-	spin_unlock_irqrestore(&sched_debug_lock, flags);
 	SEQ_printf(m, "\n");
 }
 
@@ -815,7 +837,7 @@ void sysrq_sched_debug_show(void)
 }
 
 /*
- * This itererator needs some explanation.
+ * This iterator needs some explanation.
  * It returns 1 for the header position.
  * This means 2 is CPU 0.
  * In a hotplugged system some CPUs, including CPU 0, may be missing so we have
@@ -860,15 +882,6 @@ static const struct seq_operations sched_debug_sops = {
 	.show		= sched_debug_show,
 };
 
-static int __init init_sched_debug_procfs(void)
-{
-	if (!proc_create_seq("sched_debug", 0444, NULL, &sched_debug_sops))
-		return -ENOMEM;
-	return 0;
-}
-
-__initcall(init_sched_debug_procfs);
-
 #define __PS(S, F) SEQ_printf(m, "%-45s:%21Ld\n", S, (long long)(F))
 #define __P(F) __PS(#F, F)
 #define   P(F) __PS(#F, p->F)
@@ -1033,3 +1046,13 @@ void proc_sched_set_task(struct task_struct *p)
 	memset(&p->se.statistics, 0, sizeof(p->se.statistics));
 #endif
 }
+
+void resched_latency_warn(int cpu, u64 latency)
+{
+	static DEFINE_RATELIMIT_STATE(latency_check_ratelimit, 60 * 60 * HZ, 1);
+
+	WARN(__ratelimit(&latency_check_ratelimit),
+	     "sched: CPU %d need_resched set for > %llu ns (%d ticks) "
+	     "without schedule\n",
+	     cpu, latency, cpu_rq(cpu)->ticks_without_resched);
+}