cgroup/cpuset: Change cpuset_rwsem and hotplug lock order

cpuset_rwsem is going to be acquired from sched_setscheduler() with a
following patch. There are however paths (e.g., spawn_ksoftirqd) in
which sched_scheduler() is eventually called while holding hotplug lock;
this creates a dependecy between hotplug lock (to be always acquired
first) and cpuset_rwsem (to be always acquired after hotplug lock).

Fix paths which currently take the two locks in the wrong order (after
a following patch is applied).
Tested-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: bristot@redhat.com
Cc: claudio@evidence.eu.com
Cc: lizefan@huawei.com
Cc: longman@redhat.com
Cc: luca.abeni@santannapisa.it
Cc: mathieu.poirier@linaro.org
Cc: rostedt@goodmis.org
Cc: tj@kernel.org
Cc: tommaso.cucinotta@santannapisa.it
Link: https://lkml.kernel.org/r/20190719140000.31694-7-juri.lelli@redhat.comSigned-off-by: Ingo Molnar <mingo@kernel.org>

include/linux/cpuset.h

@@ -40,14 +40,14 @@ static inline bool cpusets_enabled(void)
 
 static inline void cpuset_inc(void)
 {
-	static_branch_inc(&cpusets_pre_enable_key);
-	static_branch_inc(&cpusets_enabled_key);
+	static_branch_inc_cpuslocked(&cpusets_pre_enable_key);
+	static_branch_inc_cpuslocked(&cpusets_enabled_key);
 }
 
 static inline void cpuset_dec(void)
 {
-	static_branch_dec(&cpusets_enabled_key);
-	static_branch_dec(&cpusets_pre_enable_key);
+	static_branch_dec_cpuslocked(&cpusets_enabled_key);
+	static_branch_dec_cpuslocked(&cpusets_pre_enable_key);
 }
 
 extern int cpuset_init(void);

kernel/cgroup/cpuset.c

@@ -973,8 +973,8 @@ static void rebuild_sched_domains_locked(void)
 	cpumask_var_t *doms;
 	int ndoms;
 
+	lockdep_assert_cpus_held();
 	percpu_rwsem_assert_held(&cpuset_rwsem);
-	get_online_cpus();
 
 	/*
 	 * We have raced with CPU hotplug. Don't do anything to avoid
@@ -983,19 +983,17 @@ static void rebuild_sched_domains_locked(void)
 	 */
 	if (!top_cpuset.nr_subparts_cpus &&
 	    !cpumask_equal(top_cpuset.effective_cpus, cpu_active_mask))
-		goto out;
+		return;
 
 	if (top_cpuset.nr_subparts_cpus &&
 	   !cpumask_subset(top_cpuset.effective_cpus, cpu_active_mask))
-		goto out;
+		return;
 
 	/* Generate domain masks and attrs */
 	ndoms = generate_sched_domains(&doms, &attr);
 
 	/* Have scheduler rebuild the domains */
 	partition_and_rebuild_sched_domains(ndoms, doms, attr);
-out:
-	put_online_cpus();
 }
 #else /* !CONFIG_SMP */
 static void rebuild_sched_domains_locked(void)
@@ -1005,9 +1003,11 @@ static void rebuild_sched_domains_locked(void)
 
 void rebuild_sched_domains(void)
 {
+	get_online_cpus();
 	percpu_down_write(&cpuset_rwsem);
 	rebuild_sched_domains_locked();
 	percpu_up_write(&cpuset_rwsem);
+	put_online_cpus();
 }
 
 /**
@@ -2245,6 +2245,7 @@ static int cpuset_write_u64(struct cgroup_subsys_state *css, struct cftype *cft,
 	cpuset_filetype_t type = cft->private;
 	int retval = 0;
 
+	get_online_cpus();
 	percpu_down_write(&cpuset_rwsem);
 	if (!is_cpuset_online(cs)) {
 		retval = -ENODEV;
@@ -2282,6 +2283,7 @@ static int cpuset_write_u64(struct cgroup_subsys_state *css, struct cftype *cft,
 	}
 out_unlock:
 	percpu_up_write(&cpuset_rwsem);
+	put_online_cpus();
 	return retval;
 }
 
@@ -2292,6 +2294,7 @@ static int cpuset_write_s64(struct cgroup_subsys_state *css, struct cftype *cft,
 	cpuset_filetype_t type = cft->private;
 	int retval = -ENODEV;
 
+	get_online_cpus();
 	percpu_down_write(&cpuset_rwsem);
 	if (!is_cpuset_online(cs))
 		goto out_unlock;
@@ -2306,6 +2309,7 @@ static int cpuset_write_s64(struct cgroup_subsys_state *css, struct cftype *cft,
 	}
 out_unlock:
 	percpu_up_write(&cpuset_rwsem);
+	put_online_cpus();
 	return retval;
 }
 
@@ -2344,6 +2348,7 @@ static ssize_t cpuset_write_resmask(struct kernfs_open_file *of,
 	kernfs_break_active_protection(of->kn);
 	flush_work(&cpuset_hotplug_work);
 
+	get_online_cpus();
 	percpu_down_write(&cpuset_rwsem);
 	if (!is_cpuset_online(cs))
 		goto out_unlock;
@@ -2369,6 +2374,7 @@ static ssize_t cpuset_write_resmask(struct kernfs_open_file *of,
 	free_cpuset(trialcs);
 out_unlock:
 	percpu_up_write(&cpuset_rwsem);
+	put_online_cpus();
 	kernfs_unbreak_active_protection(of->kn);
 	css_put(&cs->css);
 	flush_workqueue(cpuset_migrate_mm_wq);
@@ -2499,6 +2505,7 @@ static ssize_t sched_partition_write(struct kernfs_open_file *of, char *buf,
 		return -EINVAL;
 
 	css_get(&cs->css);
+	get_online_cpus();
 	percpu_down_write(&cpuset_rwsem);
 	if (!is_cpuset_online(cs))
 		goto out_unlock;
@@ -2506,6 +2513,7 @@ static ssize_t sched_partition_write(struct kernfs_open_file *of, char *buf,
 	retval = update_prstate(cs, val);
 out_unlock:
 	percpu_up_write(&cpuset_rwsem);
+	put_online_cpus();
 	css_put(&cs->css);
 	return retval ?: nbytes;
 }
@@ -2711,6 +2719,7 @@ static int cpuset_css_online(struct cgroup_subsys_state *css)
 	if (!parent)
 		return 0;
 
+	get_online_cpus();
 	percpu_down_write(&cpuset_rwsem);
 
 	set_bit(CS_ONLINE, &cs->flags);
@@ -2763,6 +2772,7 @@ static int cpuset_css_online(struct cgroup_subsys_state *css)
 	spin_unlock_irq(&callback_lock);
 out_unlock:
 	percpu_up_write(&cpuset_rwsem);
+	put_online_cpus();
 	return 0;
 }
 
@@ -2781,6 +2791,7 @@ static void cpuset_css_offline(struct cgroup_subsys_state *css)
 {
 	struct cpuset *cs = css_cs(css);
 
+	get_online_cpus();
 	percpu_down_write(&cpuset_rwsem);
 
 	if (is_partition_root(cs))
@@ -2801,6 +2812,7 @@ static void cpuset_css_offline(struct cgroup_subsys_state *css)
 	clear_bit(CS_ONLINE, &cs->flags);
 
 	percpu_up_write(&cpuset_rwsem);
+	put_online_cpus();
 }
 
 static void cpuset_css_free(struct cgroup_subsys_state *css)