blob: 36068d01ac5efe0442f1d997c7d2946c79ec7549 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
|
From: Thomas Gleixner <tglx@linutronix.de>
Date: Fri, 1 Mar 2013 11:17:42 +0100
Subject: futex: Ensure lock/unlock symetry versus pi_lock and hash bucket lock
Origin: https://www.kernel.org/pub/linux/kernel/projects/rt/4.19/older/patches-4.19.1-rt3.tar.xz
In exit_pi_state_list() we have the following locking construct:
spin_lock(&hb->lock);
raw_spin_lock_irq(&curr->pi_lock);
...
spin_unlock(&hb->lock);
In !RT this works, but on RT the migrate_enable() function which is
called from spin_unlock() sees atomic context due to the held pi_lock
and just decrements the migrate_disable_atomic counter of the
task. Now the next call to migrate_disable() sees the counter being
negative and issues a warning. That check should be in
migrate_enable() already.
Fix this by dropping pi_lock before unlocking hb->lock and reaquire
pi_lock after that again. This is safe as the loop code reevaluates
head again under the pi_lock.
Reported-by: Yong Zhang <yong.zhang@windriver.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
kernel/futex.c | 2 ++
1 file changed, 2 insertions(+)
--- a/kernel/futex.c
+++ b/kernel/futex.c
@@ -918,7 +918,9 @@ void exit_pi_state_list(struct task_stru
if (head->next != next) {
/* retain curr->pi_lock for the loop invariant */
raw_spin_unlock(&pi_state->pi_mutex.wait_lock);
+ raw_spin_unlock_irq(&curr->pi_lock);
spin_unlock(&hb->lock);
+ raw_spin_lock_irq(&curr->pi_lock);
put_pi_state(pi_state);
continue;
}
|