kthread: Prevent unpark race which puts threads on the wrong cpu

The smpboot threads rely on the park/unpark mechanism which binds per
cpu threads on a particular core. Though the functionality is racy:

CPU0	       	 	CPU1  	     	    CPU2
unpark(T)				    wake_up_process(T)
  clear(SHOULD_PARK)	T runs
			leave parkme() due to !SHOULD_PARK  
  bind_to(CPU2)		BUG_ON(wrong CPU)						    

We cannot let the tasks move themself to the target CPU as one of
those tasks is actually the migration thread itself, which requires
that it starts running on the target cpu right away.

The solution to this problem is to prevent wakeups in park mode which
are not from unpark(). That way we can guarantee that the association
of the task to the target cpu is working correctly.

Add a new task state (TASK_PARKED) which prevents other wakeups and
use this state explicitly for the unpark wakeup.

Peter noticed: Also, since the task state is visible to userspace and
all the parked tasks are still in the PID space, its a good hint in ps
and friends that these tasks aren't really there for the moment.

The migration thread has another related issue.

CPU0	      	     	 CPU1
Bring up CPU2
create_thread(T)
park(T)
 wait_for_completion()
			 parkme()
			 complete()
sched_set_stop_task()
			 schedule(TASK_PARKED)

The sched_set_stop_task() call is issued while the task is on the
runqueue of CPU1 and that confuses the hell out of the stop_task class
on that cpu. So we need the same synchronizaion before
sched_set_stop_task().

Reported-by: Dave Jones <davej@redhat.com>
Reported-and-tested-by: Dave Hansen <dave@sr71.net>
Reported-and-tested-by: Borislav Petkov <bp@alien8.de>
Acked-by: Peter Ziljstra <peterz@infradead.org>
Cc: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Cc: dhillf@gmail.com
Cc: Ingo Molnar <mingo@kernel.org>
Cc: stable@vger.kernel.org
Link: http://lkml.kernel.org/r/alpine.LFD.2.02.1304091635430.21884@ionos
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

5 files changed, 45 insertions, 29 deletions

diff --git a/fs/proc/array.c b/fs/proc/array.c
index f7ed9ee46eb..cbd0f1b324b 100644
--- a/fs/proc/array.c
+++ b/fs/proc/array.c
@@ -143,6 +143,7 @@ static const char * const task_state_array[] = {
 	"x (dead)",		/*  64 */
 	"K (wakekill)",		/* 128 */
 	"W (waking)",		/* 256 */
+	"P (parked)",		/* 512 */
 };
 
 static inline const char *get_task_state(struct task_struct *tsk)
diff --git a/include/linux/sched.h b/include/linux/sched.h
index d35d2b6ddbf..e692a022527 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -163,9 +163,10 @@ print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
 #define TASK_DEAD		64
 #define TASK_WAKEKILL		128
 #define TASK_WAKING		256
-#define TASK_STATE_MAX		512
+#define TASK_PARKED		512
+#define TASK_STATE_MAX		1024
 
-#define TASK_STATE_TO_CHAR_STR "RSDTtZXxKW"
+#define TASK_STATE_TO_CHAR_STR "RSDTtZXxKWP"
 
 extern char ___assert_task_state[1 - 2*!!(
 		sizeof(TASK_STATE_TO_CHAR_STR)-1 != ilog2(TASK_STATE_MAX)+1)];
diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h
index 5a8671e8a67..e5586caff67 100644
--- a/include/trace/events/sched.h
+++ b/include/trace/events/sched.h
@@ -147,7 +147,7 @@ TRACE_EVENT(sched_switch,
 		  __print_flags(__entry->prev_state & (TASK_STATE_MAX-1), "|",
 				{ 1, "S"} , { 2, "D" }, { 4, "T" }, { 8, "t" },
 				{ 16, "Z" }, { 32, "X" }, { 64, "x" },
-				{ 128, "W" }) : "R",
+				{ 128, "K" }, { 256, "W" }, { 512, "P" }) : "R",
 		__entry->prev_state & TASK_STATE_MAX ? "+" : "",
 		__entry->next_comm, __entry->next_pid, __entry->next_prio)
 );
diff --git a/kernel/kthread.c b/kernel/kthread.c
index 691dc2ef9ba..9eb7fed0bba 100644
--- a/kernel/kthread.c
+++ b/kernel/kthread.c
@@ -124,12 +124,12 @@ void *kthread_data(struct task_struct *task)
 
 static void __kthread_parkme(struct kthread *self)
 {
-	__set_current_state(TASK_INTERRUPTIBLE);
+	__set_current_state(TASK_PARKED);
 	while (test_bit(KTHREAD_SHOULD_PARK, &self->flags)) {
 		if (!test_and_set_bit(KTHREAD_IS_PARKED, &self->flags))
 			complete(&self->parked);
 		schedule();
-		__set_current_state(TASK_INTERRUPTIBLE);
+		__set_current_state(TASK_PARKED);
 	}
 	clear_bit(KTHREAD_IS_PARKED, &self->flags);
 	__set_current_state(TASK_RUNNING);
@@ -256,8 +256,13 @@ struct task_struct *kthread_create_on_node(int (*threadfn)(void *data),
 }
 EXPORT_SYMBOL(kthread_create_on_node);
 
-static void __kthread_bind(struct task_struct *p, unsigned int cpu)
+static void __kthread_bind(struct task_struct *p, unsigned int cpu, long state)
 {
+	/* Must have done schedule() in kthread() before we set_task_cpu */
+	if (!wait_task_inactive(p, state)) {
+		WARN_ON(1);
+		return;
+	}
 	/* It's safe because the task is inactive. */
 	do_set_cpus_allowed(p, cpumask_of(cpu));
 	p->flags |= PF_THREAD_BOUND;
@@ -274,12 +279,7 @@ static void __kthread_bind(struct task_struct *p, unsigned int cpu)
  */
 void kthread_bind(struct task_struct *p, unsigned int cpu)
 {
-	/* Must have done schedule() in kthread() before we set_task_cpu */
-	if (!wait_task_inactive(p, TASK_UNINTERRUPTIBLE)) {
-		WARN_ON(1);
-		return;
-	}
-	__kthread_bind(p, cpu);
+	__kthread_bind(p, cpu, TASK_UNINTERRUPTIBLE);
 }
 EXPORT_SYMBOL(kthread_bind);
 
@@ -324,6 +324,22 @@ static struct kthread *task_get_live_kthread(struct task_struct *k)
 	return NULL;
 }
 
+static void __kthread_unpark(struct task_struct *k, struct kthread *kthread)
+{
+	clear_bit(KTHREAD_SHOULD_PARK, &kthread->flags);
+	/*
+	 * We clear the IS_PARKED bit here as we don't wait
+	 * until the task has left the park code. So if we'd
+	 * park before that happens we'd see the IS_PARKED bit
+	 * which might be about to be cleared.
+	 */
+	if (test_and_clear_bit(KTHREAD_IS_PARKED, &kthread->flags)) {
+		if (test_bit(KTHREAD_IS_PER_CPU, &kthread->flags))
+			__kthread_bind(k, kthread->cpu, TASK_PARKED);
+		wake_up_state(k, TASK_PARKED);
+	}
+}
+
 /**
  * kthread_unpark - unpark a thread created by kthread_create().
  * @k:		thread created by kthread_create().
@@ -336,20 +352,8 @@ void kthread_unpark(struct task_struct *k)
 {
 	struct kthread *kthread = task_get_live_kthread(k);
 
-	if (kthread) {
-		clear_bit(KTHREAD_SHOULD_PARK, &kthread->flags);
-		/*
-		 * We clear the IS_PARKED bit here as we don't wait
-		 * until the task has left the park code. So if we'd
-		 * park before that happens we'd see the IS_PARKED bit
-		 * which might be about to be cleared.
-		 */
-		if (test_and_clear_bit(KTHREAD_IS_PARKED, &kthread->flags)) {
-			if (test_bit(KTHREAD_IS_PER_CPU, &kthread->flags))
-				__kthread_bind(k, kthread->cpu);
-			wake_up_process(k);
-		}
-	}
+	if (kthread)
+		__kthread_unpark(k, kthread);
 	put_task_struct(k);
 }
 
@@ -407,7 +411,7 @@ int kthread_stop(struct task_struct *k)
 	trace_sched_kthread_stop(k);
 	if (kthread) {
 		set_bit(KTHREAD_SHOULD_STOP, &kthread->flags);
-		clear_bit(KTHREAD_SHOULD_PARK, &kthread->flags);
+		__kthread_unpark(k, kthread);
 		wake_up_process(k);
 		wait_for_completion(&kthread->exited);
 	}
diff --git a/kernel/smpboot.c b/kernel/smpboot.c
index 8eaed9aa9cf..02fc5c93367 100644
--- a/kernel/smpboot.c
+++ b/kernel/smpboot.c
@@ -185,8 +185,18 @@ __smpboot_create_thread(struct smp_hotplug_thread *ht, unsigned int cpu)
 	}
 	get_task_struct(tsk);
 	*per_cpu_ptr(ht->store, cpu) = tsk;
-	if (ht->create)
-		ht->create(cpu);
+	if (ht->create) {
+		/*
+		 * Make sure that the task has actually scheduled out
+		 * into park position, before calling the create
+		 * callback. At least the migration thread callback
+		 * requires that the task is off the runqueue.
+		 */
+		if (!wait_task_inactive(tsk, TASK_PARKED))
+			WARN_ON(1);
+		else
+			ht->create(cpu);
+	}
 	return 0;
 }