All of lore.kernel.org
 help / color / mirror / Atom feed
From: Yimin Deng <yimin11.deng@gmail.com>
To: linux-rt-users@vger.kernel.org
Subject: kernel BUG at kernel/rtmutex_common.h:75
Date: Wed, 4 Nov 2015 22:35:57 +0800	[thread overview]
Message-ID: <CAAh1qt=DCL9aUXNxanP5BKtiPp3m+qj4yB+gDohhXPVFCxWwzg@mail.gmail.com> (raw)

I encountered “kernel BUG” which was reported in the
rt_mutex_top_waiter() at kernel/rtmutex_common.h:75.

Linux version: 3.12.37-rt51; CONFIG_PREEMPT_RT_FULL is disabled.
Architecture: PowerPC

We ported an application from pSOS RTOS to Linux using the
Xenomai-Mercury (=library to map pSOS task to POSIX threads). And We
have several threads running in the real-time priority domain.
ThreadA: running at prio -59. pthread_mutex_lock() +
pthread_cond_timedwait() + pthread_mutex_unlock()
ThreadB: running at prio -84. pthread_mutex_lock() +
pthread_cond_signal() + pthread_mutex_unlock()

ThreadA:
------ ------
futex_wait_requeue_pi()
    futex_wait_queue_me()
    <timed out>

    raw_spin_lock_irq(&current->pi_lock);
    if (current->pi_blocked_on) {
        raw_spin_unlock_irq(&current->pi_lock);
    } else {
        current->pi_blocked_on = PI_WAKEUP_INPROGRESS;
        raw_spin_unlock_irq(&current->pi_lock);
            <-- ThreadA was interrupted and preempted!
        spin_lock(&hb->lock);


ThreadB:
------ ------
rt_mutex_start_proxy_lock();
    task_blocks_on_rt_mutex();  <-- return "-EAGAIN" due to
"task->pi_blocked_on == PI_WAKEUP_INPROGRESS"
    ...
    if (unlikely(ret))
        remove_waiter(lock, waiter);
            int first = (waiter == rt_mutex_top_waiter(lock));  <--
BUG_ON(w->lock != lock);

It seems that the purpose to call the remove_waiter() is to remove the
waiter added by “plist_add(&waiter->list_entry, &lock->wait_list);” in
the task_blocks_on_rt_mutex(). But in the scenario above there's no
waiter on the lock yet and
the waiter has not been added into the wait list of the lock in the
task_blocks_on_rt_mutex() due to the failure “-EAGAIN”. So it reported
kernel BUG in the rt_mutex_top_waiter().

I modified it as below and the issue seems disappear.
- if (unlikely(ret))
+ if (unlikely(ret && (-EAGAIN != ret)))
       remove_waiter(lock, waiter);

Could the scenario above be possible? If so, how to resolve this issue?
Thanks!

B.R.
Yimin Deng
--
To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

             reply	other threads:[~2015-11-04 14:35 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-11-04 14:35 Yimin Deng [this message]
2015-11-06 14:41 ` kernel BUG at kernel/rtmutex_common.h:75 Thomas Gleixner
2015-11-07 18:09   ` Thomas Gleixner
2015-11-08  3:31     ` Yimin Deng
2018-03-12 14:28 [PATCH] kernel/rtmutex: Handle non enqueued waiters gracefully in remove_waiter() Sebastian Andrzej Siewior
2018-03-16 12:20 ` Peter Zijlstra
2018-03-19 15:11   ` Thomas Gleixner
2018-03-27 12:14     ` [PATCH v2] locking/rtmutex: " Sebastian Andrzej Siewior
2018-03-28 21:07       ` [tip:locking/core] " tip-bot for Peter Zijlstra

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='CAAh1qt=DCL9aUXNxanP5BKtiPp3m+qj4yB+gDohhXPVFCxWwzg@mail.gmail.com' \
    --to=yimin11.deng@gmail.com \
    --cc=linux-rt-users@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.