[PATCH] kernel/rtmutex: Handle non enqueued waiters gracefully in remove_waiter()

From: Sebastian Andrzej Siewior
Date: Mon Mar 12 2018 - 10:28:52 EST


In -RT task_blocks_on_rt_mutex() may return with -EAGAIN due to
(->pi_blocked_on == PI_WAKEUP_INPROGRESS) before it added itself as a
waiter. In such a case we must not call remove_waiter() because without
a waiter it will trigger the BUG_ON() statement.

This was initially reported by Yimin Deng. Thomas Gleixner fixed it then
with an explicit check for waiters before calling remove_waiter() with
the following note:

| Guard it with rt_mutex_has_waiters(). This is a quick fix which is
| easy to backport. The proper fix is to have a central check in
| remove_waiter() so we can call it unconditionally.

This is the suggested change.
Now that it is possible to call remove_waiter() unconditionally I also
remove that check from rt_mutex_slowlock().

Link: https://lkml.kernel.org/r/CAAh1qt=DCL9aUXNxanP5BKtiPp3m+qj4yB+gDohhXPVFCxWwzg@xxxxxxxxxxxxxx
Reported-and-debugged-by: Yimin Deng <yimin11.deng@xxxxxxxxx>
Suggested-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
---
kernel/locking/rtmutex.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/kernel/locking/rtmutex.c b/kernel/locking/rtmutex.c
index 65cc0cb984e6..57d28d8f5280 100644
--- a/kernel/locking/rtmutex.c
+++ b/kernel/locking/rtmutex.c
@@ -1068,12 +1068,15 @@ static void mark_wakeup_next_waiter(struct wake_q_head *wake_q,
static void remove_waiter(struct rt_mutex *lock,
struct rt_mutex_waiter *waiter)
{
- bool is_top_waiter = (waiter == rt_mutex_top_waiter(lock));
+ bool is_top_waiter = false;
struct task_struct *owner = rt_mutex_owner(lock);
struct rt_mutex *next_lock;

lockdep_assert_held(&lock->wait_lock);

+ if (rt_mutex_has_waiters(lock))
+ is_top_waiter = waiter == rt_mutex_top_waiter(lock);
+
raw_spin_lock(&current->pi_lock);
rt_mutex_dequeue(lock, waiter);
current->pi_blocked_on = NULL;
@@ -1268,8 +1271,7 @@ rt_mutex_slowlock(struct rt_mutex *lock, int state,

if (unlikely(ret)) {
__set_current_state(TASK_RUNNING);
- if (rt_mutex_has_waiters(lock))
- remove_waiter(lock, &waiter);
+ remove_waiter(lock, &waiter);
rt_mutex_handle_deadlock(ret, chwalk, &waiter);
}

--
2.16.2