freezer,sched: Rewrite core freezer logic

Rewrite the core freezer to behave better wrt thawing and be simpler
in general.

By replacing PF_FROZEN with TASK_FROZEN, a special block state, it is
ensured frozen tasks stay frozen until thawed and don't randomly wake
up early, as is currently possible.

As such, it does away with PF_FROZEN and PF_FREEZER_SKIP, freeing up
two PF_flags (yay!).

Specifically; the current scheme works a little like:

	freezer_do_not_count();
	schedule();
	freezer_count();

And either the task is blocked, or it lands in try_to_freezer()
through freezer_count(). Now, when it is blocked, the freezer
considers it frozen and continues.

However, on thawing, once pm_freezing is cleared, freezer_count()
stops working, and any random/spurious wakeup will let a task run
before its time.

That is, thawing tries to thaw things in explicit order; kernel
threads and workqueues before doing bringing SMP back before userspace
etc.. However due to the above mentioned races it is entirely possible
for userspace tasks to thaw (by accident) before SMP is back.

This can be a fatal problem in asymmetric ISA architectures (eg ARMv9)
where the userspace task requires a special CPU to run.

As said; replace this with a special task state TASK_FROZEN and add
the following state transitions:

	TASK_FREEZABLE	-> TASK_FROZEN
	__TASK_STOPPED	-> TASK_FROZEN
	__TASK_TRACED	-> TASK_FROZEN

The new TASK_FREEZABLE can be set on any state part of TASK_NORMAL
(IOW. TASK_INTERRUPTIBLE and TASK_UNINTERRUPTIBLE) -- any such state
is already required to deal with spurious wakeups and the freezer
causes one such when thawing the task (since the original state is
lost).

The special __TASK_{STOPPED,TRACED} states *can* be restored since
their canonical state is in ->jobctl.

With this, frozen tasks need an explicit TASK_FROZEN wakeup and are
free of undue (early / spurious) wakeups.

Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20220822114649.055452969@infradead.org

1 files changed, 96 insertions, 37 deletions

diff --git a/kernel/freezer.c b/kernel/freezer.c
index 45ab36ffd0e7..4fad0e6fca64 100644
--- a/kernel/freezer.c
+++ b/kernel/freezer.c
@@ -13,10 +13,11 @@
 #include <linux/kthread.h>
 
 /* total number of freezing conditions in effect */
-atomic_t system_freezing_cnt = ATOMIC_INIT(0);
-EXPORT_SYMBOL(system_freezing_cnt);
+DEFINE_STATIC_KEY_FALSE(freezer_active);
+EXPORT_SYMBOL(freezer_active);
 
-/* indicate whether PM freezing is in effect, protected by
+/*
+ * indicate whether PM freezing is in effect, protected by
  * system_transition_mutex
  */
 bool pm_freezing;
@@ -29,7 +30,7 @@ static DEFINE_SPINLOCK(freezer_lock);
  * freezing_slow_path - slow path for testing whether a task needs to be frozen
  * @p: task to be tested
  *
- * This function is called by freezing() if system_freezing_cnt isn't zero
+ * This function is called by freezing() if freezer_active isn't zero
  * and tests whether @p needs to enter and stay in frozen state.  Can be
  * called under any context.  The freezers are responsible for ensuring the
  * target tasks see the updated state.
@@ -52,41 +53,40 @@ bool freezing_slow_path(struct task_struct *p)
 }
 EXPORT_SYMBOL(freezing_slow_path);
 
+bool frozen(struct task_struct *p)
+{
+	return READ_ONCE(p->__state) & TASK_FROZEN;
+}
+
 /* Refrigerator is place where frozen processes are stored :-). */
 bool __refrigerator(bool check_kthr_stop)
 {
-	/* Hmm, should we be allowed to suspend when there are realtime
-	   processes around? */
+	unsigned int state = get_current_state();
 	bool was_frozen = false;
-	unsigned int save = get_current_state();
 
 	pr_debug("%s entered refrigerator\n", current->comm);
 
+	WARN_ON_ONCE(state && !(state & TASK_NORMAL));
+
 	for (;;) {
-		set_current_state(TASK_UNINTERRUPTIBLE);
+		bool freeze;
+
+		set_current_state(TASK_FROZEN);
 
 		spin_lock_irq(&freezer_lock);
-		current->flags |= PF_FROZEN;
-		if (!freezing(current) ||
-		    (check_kthr_stop && kthread_should_stop()))
-			current->flags &= ~PF_FROZEN;
+		freeze = freezing(current) && !(check_kthr_stop && kthread_should_stop());
 		spin_unlock_irq(&freezer_lock);
 
-		if (!(current->flags & PF_FROZEN))
+		if (!freeze)
 			break;
+
 		was_frozen = true;
 		schedule();
 	}
+	__set_current_state(TASK_RUNNING);
 
 	pr_debug("%s left refrigerator\n", current->comm);
 
-	/*
-	 * Restore saved task state before returning.  The mb'd version
-	 * needs to be used; otherwise, it might silently break
-	 * synchronization which depends on ordered task state change.
-	 */
-	set_current_state(save);
-
 	return was_frozen;
 }
 EXPORT_SYMBOL(__refrigerator);
@@ -101,6 +101,44 @@ static void fake_signal_wake_up(struct task_struct *p)
 	}
 }
 
+static int __set_task_frozen(struct task_struct *p, void *arg)
+{
+	unsigned int state = READ_ONCE(p->__state);
+
+	if (p->on_rq)
+		return 0;
+
+	if (p != current && task_curr(p))
+		return 0;
+
+	if (!(state & (TASK_FREEZABLE | __TASK_STOPPED | __TASK_TRACED)))
+		return 0;
+
+	/*
+	 * Only TASK_NORMAL can be augmented with TASK_FREEZABLE, since they
+	 * can suffer spurious wakeups.
+	 */
+	if (state & TASK_FREEZABLE)
+		WARN_ON_ONCE(!(state & TASK_NORMAL));
+
+#ifdef CONFIG_LOCKDEP
+	/*
+	 * It's dangerous to freeze with locks held; there be dragons there.
+	 */
+	if (!(state & __TASK_FREEZABLE_UNSAFE))
+		WARN_ON_ONCE(debug_locks && p->lockdep_depth);
+#endif
+
+	WRITE_ONCE(p->__state, TASK_FROZEN);
+	return TASK_FROZEN;
+}
+
+static bool __freeze_task(struct task_struct *p)
+{
+	/* TASK_FREEZABLE|TASK_STOPPED|TASK_TRACED -> TASK_FROZEN */
+	return task_call_func(p, __set_task_frozen, NULL);
+}
+
 /**
  * freeze_task - send a freeze request to given task
  * @p: task to send the request to
@@ -116,20 +154,8 @@ bool freeze_task(struct task_struct *p)
 {
 	unsigned long flags;
 
-	/*
-	 * This check can race with freezer_do_not_count, but worst case that
-	 * will result in an extra wakeup being sent to the task.  It does not
-	 * race with freezer_count(), the barriers in freezer_count() and
-	 * freezer_should_skip() ensure that either freezer_count() sees
-	 * freezing == true in try_to_freeze() and freezes, or
-	 * freezer_should_skip() sees !PF_FREEZE_SKIP and freezes the task
-	 * normally.
-	 */
-	if (freezer_should_skip(p))
-		return false;
-
 	spin_lock_irqsave(&freezer_lock, flags);
-	if (!freezing(p) || frozen(p)) {
+	if (!freezing(p) || frozen(p) || __freeze_task(p)) {
 		spin_unlock_irqrestore(&freezer_lock, flags);
 		return false;
 	}
@@ -137,19 +163,52 @@ bool freeze_task(struct task_struct *p)
 	if (!(p->flags & PF_KTHREAD))
 		fake_signal_wake_up(p);
 	else
-		wake_up_state(p, TASK_INTERRUPTIBLE);
+		wake_up_state(p, TASK_NORMAL);
 
 	spin_unlock_irqrestore(&freezer_lock, flags);
 	return true;
 }
 
+/*
+ * The special task states (TASK_STOPPED, TASK_TRACED) keep their canonical
+ * state in p->jobctl. If either of them got a wakeup that was missed because
+ * TASK_FROZEN, then their canonical state reflects that and the below will
+ * refuse to restore the special state and instead issue the wakeup.
+ */
+static int __set_task_special(struct task_struct *p, void *arg)
+{
+	unsigned int state = 0;
+
+	if (p->jobctl & JOBCTL_TRACED)
+		state = TASK_TRACED;
+
+	else if (p->jobctl & JOBCTL_STOPPED)
+		state = TASK_STOPPED;
+
+	if (state)
+		WRITE_ONCE(p->__state, state);
+
+	return state;
+}
+
 void __thaw_task(struct task_struct *p)
 {
-	unsigned long flags;
+	unsigned long flags, flags2;
 
 	spin_lock_irqsave(&freezer_lock, flags);
-	if (frozen(p))
-		wake_up_process(p);
+	if (WARN_ON_ONCE(freezing(p)))
+		goto unlock;
+
+	if (lock_task_sighand(p, &flags2)) {
+		/* TASK_FROZEN -> TASK_{STOPPED,TRACED} */
+		bool ret = task_call_func(p, __set_task_special, NULL);
+		unlock_task_sighand(p, &flags2);
+		if (ret)
+			goto unlock;
+	}
+
+	wake_up_state(p, TASK_FROZEN);
+unlock:
 	spin_unlock_irqrestore(&freezer_lock, flags);
 }