---
Subject: sched: Fix balance vs hotplug race
From: Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
Date: Wed Nov 25 13:31:39 CET 2009
Since (e761b77: cpu hotplug, sched: Introduce cpu_active_map and redo
sched domain managment) we have cpu_active_mask which is suppose to
rule scheduler migration and load-balancing, except it never did.
The particular problem being solved here is a crash in
try_to_wake_up() where select_task_rq() ends up selecting an offline
cpu because select_task_rq_fair() trusts the sched_domain tree to reflect
the current state of affairs, similarly select_task_rq_rt() trusts the
root_domain.
However, the sched_domains are updated from CPU_DEAD, which is after
the cpu is taken offline and after stop_machine is done. Therefore it
can race perfectly well with code assuming the domains are right.
Cure this by building the domains from cpu_active_mask on
CPU_DOWN_PREPARE.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
---
include/linux/cpumask.h | 2 ++
kernel/cpu.c | 18 +++++++++++++-----
kernel/cpuset.c | 16 +++++++++-------
kernel/sched.c | 44 ++++++++++++++++++++++++++------------------
4 files changed, 50 insertions(+), 30 deletions(-)
Index: linux-2.6/include/linux/cpumask.h
===================================================================
--- linux-2.6.orig/include/linux/cpumask.h
+++ linux-2.6/include/linux/cpumask.h
@@ -84,6 +84,7 @@ extern const struct cpumask *const cpu_a
#define num_online_cpus() cpumask_weight(cpu_online_mask)
#define num_possible_cpus() cpumask_weight(cpu_possible_mask)
#define num_present_cpus() cpumask_weight(cpu_present_mask)
+#define num_active_cpus() cpumask_weight(cpu_active_mask)
#define cpu_online(cpu) cpumask_test_cpu((cpu), cpu_online_mask)
#define cpu_possible(cpu) cpumask_test_cpu((cpu), cpu_possible_mask)
#define cpu_present(cpu) cpumask_test_cpu((cpu), cpu_present_mask)
@@ -92,6 +93,7 @@ extern const struct cpumask *const cpu_a
#define num_online_cpus() 1
#define num_possible_cpus() 1
#define num_present_cpus() 1
+#define num_active_cpus() 1
#define cpu_online(cpu) ((cpu) == 0)
#define cpu_possible(cpu) ((cpu) == 0)
#define cpu_present(cpu) ((cpu) == 0)
Index: linux-2.6/kernel/cpuset.c
===================================================================
--- linux-2.6.orig/kernel/cpuset.c
+++ linux-2.6/kernel/cpuset.c
@@ -872,7 +872,7 @@ static int update_cpumask(struct cpuset
if (retval < 0)
return retval;
- if (!cpumask_subset(trialcs->cpus_allowed, cpu_online_mask))
+ if (!cpumask_subset(trialcs->cpus_allowed, cpu_active_mask))
return -EINVAL;
}
retval = validate_change(cs, trialcs);
@@ -2010,7 +2010,7 @@ static void scan_for_empty_cpusets(struc
}
/* Continue past cpusets with all cpus, mems online */
- if (cpumask_subset(cp->cpus_allowed, cpu_online_mask) &&
+ if (cpumask_subset(cp->cpus_allowed, cpu_active_mask) &&
nodes_subset(cp->mems_allowed, node_states[N_HIGH_MEMORY]))
continue;
@@ -2019,7 +2019,7 @@ static void scan_for_empty_cpusets(struc
/* Remove offline cpus and mems from this cpuset. */
mutex_lock(&callback_mutex);
cpumask_and(cp->cpus_allowed, cp->cpus_allowed,
- cpu_online_mask);
+ cpu_active_mask);
nodes_and(cp->mems_allowed, cp->mems_allowed,
node_states[N_HIGH_MEMORY]);
mutex_unlock(&callback_mutex);
@@ -2057,8 +2057,10 @@ static int cpuset_track_online_cpus(stru
switch (phase) {
case CPU_ONLINE:
case CPU_ONLINE_FROZEN:
- case CPU_DEAD:
- case CPU_DEAD_FROZEN:
+ case CPU_DOWN_PREPARE:
+ case CPU_DOWN_PREPARE_FROZEN:
+ case CPU_DOWN_FAILED:
+ case CPU_DOWN_FAILED_FROZEN:
break;
default:
@@ -2067,7 +2069,7 @@ static int cpuset_track_online_cpus(stru
cgroup_lock();
mutex_lock(&callback_mutex);
- cpumask_copy(top_cpuset.cpus_allowed, cpu_online_mask);
+ cpumask_copy(top_cpuset.cpus_allowed, cpu_active_mask);
mutex_unlock(&callback_mutex);
scan_for_empty_cpusets(&top_cpuset);
ndoms = generate_sched_domains(&doms, &attr);
@@ -2114,7 +2116,7 @@ static int cpuset_track_online_nodes(str
void __init cpuset_init_smp(void)
{
- cpumask_copy(top_cpuset.cpus_allowed, cpu_online_mask);
+ cpumask_copy(top_cpuset.cpus_allowed, cpu_active_mask);
top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
hotcpu_notifier(cpuset_track_online_cpus, 0);
Index: linux-2.6/kernel/sched.c
===================================================================
--- linux-2.6.orig/kernel/sched.c
+++ linux-2.6/kernel/sched.c
@@ -2323,6 +2323,12 @@ void task_oncpu_function_call(struct tas
preempt_enable();
}
+static inline
+int select_task_rq(struct task_struct *p, int sd_flags, int wake_flags)
+{
+ return p->sched_class->select_task_rq(p, sd_flags, wake_flags);
+}
+
/***
* try_to_wake_up - wake up a thread
* @p: the to-be-woken-up thread
@@ -2376,7 +2382,7 @@ static int try_to_wake_up(struct task_st
p->state = TASK_WAKING;
task_rq_unlock(rq, &flags);
- cpu = p->sched_class->select_task_rq(p, SD_BALANCE_WAKE, wake_flags);
+ cpu = select_task_rq(p, SD_BALANCE_WAKE, wake_flags);
if (cpu != orig_cpu) {
local_irq_save(flags);
rq = cpu_rq(cpu);
@@ -2593,7 +2599,7 @@ void sched_fork(struct task_struct *p, i
p->sched_class = &fair_sched_class;
#ifdef CONFIG_SMP
- cpu = p->sched_class->select_task_rq(p, SD_BALANCE_FORK, 0);
+ cpu = select_task_rq(p, SD_BALANCE_FORK, 0);
#endif
local_irq_save(flags);
update_rq_clock(cpu_rq(cpu));
@@ -3156,7 +3162,7 @@ out:
void sched_exec(void)
{
int new_cpu, this_cpu = get_cpu();
- new_cpu = current->sched_class->select_task_rq(current, SD_BALANCE_EXEC, 0);
+ new_cpu = select_task_rq(current, SD_BALANCE_EXEC, 0);
put_cpu();
if (new_cpu != this_cpu)
sched_migrate_task(current, new_cpu);
@@ -4134,7 +4140,7 @@ static int load_balance(int this_cpu, st
unsigned long flags;
struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask);
- cpumask_copy(cpus, cpu_online_mask);
+ cpumask_copy(cpus, cpu_active_mask);
/*
* When power savings policy is enabled for the parent domain, idle
@@ -4297,7 +4303,7 @@ load_balance_newidle(int this_cpu, struc
int all_pinned = 0;
struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask);
- cpumask_copy(cpus, cpu_online_mask);
+ cpumask_copy(cpus, cpu_active_mask);
/*
* When power savings policy is enabled for the parent domain, idle
@@ -4694,7 +4700,7 @@ int select_nohz_load_balancer(int stop_t
cpumask_set_cpu(cpu, nohz.cpu_mask);
/* time for ilb owner also to sleep */
- if (cpumask_weight(nohz.cpu_mask) == num_online_cpus()) {
+ if (cpumask_weight(nohz.cpu_mask) == num_active_cpus()) {
if (atomic_read(&nohz.load_balancer) == cpu)
atomic_set(&nohz.load_balancer, -1);
return 0;
@@ -7071,7 +7077,7 @@ int set_cpus_allowed_ptr(struct task_str
int ret = 0;
rq = task_rq_lock(p, &flags);
- if (!cpumask_intersects(new_mask, cpu_online_mask)) {
+ if (!cpumask_intersects(new_mask, cpu_active_mask)) {
ret = -EINVAL;
goto out;
}
@@ -7093,7 +7099,7 @@ int set_cpus_allowed_ptr(struct task_str
if (cpumask_test_cpu(task_cpu(p), new_mask))
goto out;
- if (migrate_task(p, cpumask_any_and(cpu_online_mask, new_mask), &req)) {
+ if (migrate_task(p, cpumask_any_and(cpu_active_mask, new_mask), &req)) {
/* Need help from migration thread: drop lock and wait. */
struct task_struct *mt = rq->migration_thread;
@@ -7247,19 +7253,19 @@ static void move_task_off_dead_cpu(int d
again:
/* Look for allowed, online CPU in same node. */
- for_each_cpu_and(dest_cpu, nodemask, cpu_online_mask)
+ for_each_cpu_and(dest_cpu, nodemask, cpu_active_mask)
if (cpumask_test_cpu(dest_cpu, &p->cpus_allowed))
goto move;
/* Any allowed, online CPU? */
- dest_cpu = cpumask_any_and(&p->cpus_allowed, cpu_online_mask);
+ dest_cpu = cpumask_any_and(&p->cpus_allowed, cpu_active_mask);
if (dest_cpu < nr_cpu_ids)
goto move;
/* No more Mr. Nice Guy. */
if (dest_cpu >= nr_cpu_ids) {
cpuset_cpus_allowed_locked(p, &p->cpus_allowed);
- dest_cpu = cpumask_any_and(cpu_online_mask, &p->cpus_allowed);
+ dest_cpu = cpumask_any_and(cpu_active_mask, &p->cpus_allowed);
/*
* Don't tell them about moving exiting tasks or
@@ -7288,7 +7294,7 @@ move:
*/
static void migrate_nr_uninterruptible(struct rq *rq_src)
{
- struct rq *rq_dest = cpu_rq(cpumask_any(cpu_online_mask));
+ struct rq *rq_dest = cpu_rq(cpumask_any(cpu_active_mask));
unsigned long flags;
local_irq_save(flags);
@@ -7542,7 +7548,7 @@ static ctl_table *sd_alloc_ctl_cpu_table
static struct ctl_table_header *sd_sysctl_header;
static void register_sched_domain_sysctl(void)
{
- int i, cpu_num = num_online_cpus();
+ int i, cpu_num = num_possible_cpus();
struct ctl_table *entry = sd_alloc_ctl_entry(cpu_num + 1);
char buf[32];
@@ -7552,7 +7558,7 @@ static void register_sched_domain_sysctl
if (entry == NULL)
return;
- for_each_online_cpu(i) {
+ for_each_possible_cpu(i) {
snprintf(buf, 32, "cpu%d", i);
entry->procname = kstrdup(buf, GFP_KERNEL);
entry->mode = 0555;
@@ -9064,7 +9070,7 @@ match1:
if (doms_new == NULL) {
ndoms_cur = 0;
doms_new = &fallback_doms;
- cpumask_andnot(doms_new[0], cpu_online_mask, cpu_isolated_map);
+ cpumask_andnot(doms_new[0], cpu_active_mask, cpu_isolated_map);
WARN_ON_ONCE(dattr_new);
}
@@ -9195,8 +9201,10 @@ static int update_sched_domains(struct n
switch (action) {
case CPU_ONLINE:
case CPU_ONLINE_FROZEN:
- case CPU_DEAD:
- case CPU_DEAD_FROZEN:
+ case CPU_DOWN_PREPARE:
+ case CPU_DOWN_PREPARE_FROZEN:
+ case CPU_DOWN_FAILED:
+ case CPU_DOWN_FAILED_FROZEN:
partition_sched_domains(1, NULL, NULL);
return NOTIFY_OK;
@@ -9243,7 +9251,7 @@ void __init sched_init_smp(void)
#endif
get_online_cpus();
mutex_lock(&sched_domains_mutex);
- arch_init_sched_domains(cpu_online_mask);
+ arch_init_sched_domains(cpu_active_mask);
cpumask_andnot(non_isolated_cpus, cpu_possible_mask, cpu_isolated_map);
if (cpumask_empty(non_isolated_cpus))
cpumask_set_cpu(smp_processor_id(), non_isolated_cpus);
Index: linux-2.6/kernel/cpu.c
===================================================================
--- linux-2.6.orig/kernel/cpu.c
+++ linux-2.6/kernel/cpu.c
@@ -212,6 +212,8 @@ static int __ref _cpu_down(unsigned int
err = __raw_notifier_call_chain(&cpu_chain, CPU_DOWN_PREPARE | mod,
hcpu, -1, &nr_calls);
if (err == NOTIFY_BAD) {
+ set_cpu_active(cpu, true);
+
nr_calls--;
__raw_notifier_call_chain(&cpu_chain, CPU_DOWN_FAILED | mod,
hcpu, nr_calls, NULL);
@@ -223,11 +225,11 @@ static int __ref _cpu_down(unsigned int
/* Ensure that we are not runnable on dying cpu */
cpumask_copy(old_allowed, ¤t->cpus_allowed);
- set_cpus_allowed_ptr(current,
- cpumask_of(cpumask_any_but(cpu_online_mask, cpu)));
+ set_cpus_allowed_ptr(current, cpu_active_mask);
err = __stop_machine(take_cpu_down, &tcd_param, cpumask_of(cpu));
if (err) {
+ set_cpu_active(cpu, true);
/* CPU didn't die: tell everyone. Can't complain. */
if (raw_notifier_call_chain(&cpu_chain, CPU_DOWN_FAILED | mod,
hcpu) == NOTIFY_BAD)
@@ -292,9 +294,6 @@ int __ref cpu_down(unsigned int cpu)
err = _cpu_down(cpu, 0);
- if (cpu_online(cpu))
- set_cpu_active(cpu, true);
-
out:
cpu_maps_update_done();
stop_machine_destroy();
@@ -387,6 +386,15 @@ int disable_nonboot_cpus(void)
* with the userspace trying to use the CPU hotplug at the same time
*/
cpumask_clear(frozen_cpus);
+
+ for_each_online_cpu(cpu) {
+ if (cpu == first_cpu)
+ continue;
+ set_cpu_active(cpu, false);
+ }
+
+ synchronize_sched();
+
printk("Disabling non-boot CPUs ...\n");
for_each_online_cpu(cpu) {
if (cpu == first_cpu)
