linux/kernel/time
John Stultz 7bd3601446 timekeeping: Fix HRTICK related deadlock from ntp lock changes
Gerlando Falauto reported that when HRTICK is enabled, it is
possible to trigger system deadlocks. These were hard to
reproduce, as HRTICK has been broken in the past, but seemed
to be connected to the timekeeping_seq lock.

Since seqlock/seqcount's aren't supported w/ lockdep, I added
some extra spinlock based locking and triggered the following
lockdep output:

[   15.849182] ntpd/4062 is trying to acquire lock:
[   15.849765]  (&(&pool->lock)->rlock){..-...}, at: [<ffffffff810aa9b5>] __queue_work+0x145/0x480
[   15.850051]
[   15.850051] but task is already holding lock:
[   15.850051]  (timekeeper_lock){-.-.-.}, at: [<ffffffff810df6df>] do_adjtimex+0x7f/0x100

<snip>

[   15.850051] Chain exists of: &(&pool->lock)->rlock --> &p->pi_lock --> timekeeper_lock
[   15.850051]  Possible unsafe locking scenario:
[   15.850051]
[   15.850051]        CPU0                    CPU1
[   15.850051]        ----                    ----
[   15.850051]   lock(timekeeper_lock);
[   15.850051]                                lock(&p->pi_lock);
[   15.850051] lock(timekeeper_lock);
[   15.850051] lock(&(&pool->lock)->rlock);
[   15.850051]
[   15.850051]  *** DEADLOCK ***

The deadlock was introduced by 06c017fdd4 ("timekeeping:
Hold timekeepering locks in do_adjtimex and hardpps") in 3.10

This patch avoids this deadlock, by moving the call to
schedule_delayed_work() outside of the timekeeper lock
critical section.

Reported-by: Gerlando Falauto <gerlando.falauto@keymile.com>
Tested-by: Lin Ming <minggr@gmail.com>
Signed-off-by: John Stultz <john.stultz@linaro.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: stable <stable@vger.kernel.org> #3.11, 3.10
Link: http://lkml.kernel.org/r/1378943457-27314-1-git-send-email-john.stultz@linaro.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2013-09-12 07:49:51 +02:00
..
alarmtimer.c alarmtimer: Export symbols of functions declared in linux/alarmtimer.h 2013-06-12 14:02:12 -07:00
clockevents.c clockevents: Implement unbind functionality 2013-05-16 11:09:18 +02:00
clocksource.c clocksource: Reselect clocksource when watchdog validated high-res capability 2013-07-05 11:09:28 +02:00
jiffies.c time: Kill xtime_lock, replacing it with jiffies_lock 2012-11-13 14:08:23 -05:00
Kconfig context_tracking: Remove full dynticks' hacky dependency on wide context tracking 2013-08-13 00:54:34 +02:00
Makefile sched_clock: Make ARM's sched_clock generic for all architectures 2013-06-12 14:02:13 -07:00
ntp_internal.h ntp: Rework do_adjtimex to take timespec and tai arguments 2013-04-04 13:18:15 -07:00
ntp.c timekeeping: Fix HRTICK related deadlock from ntp lock changes 2013-09-12 07:49:51 +02:00
posix-clock.c kernel: Fix files explicitly needing EXPORT_SYMBOL infrastructure 2011-10-31 19:30:05 -04:00
sched_clock.c sched_clock: Fix integer overflow 2013-07-22 16:24:22 -07:00
tick-broadcast.c tick: broadcast: Check broadcast mode on CPU hotplug 2013-07-12 12:35:40 +02:00
tick-common.c tick: Sanitize broadcast control logic 2013-07-02 14:26:45 +02:00
tick-internal.h clockevents: Define CS_NAME_LEN unconditionally 2013-05-28 09:28:02 +02:00
tick-oneshot.c clockevents: Make minimum delay adjustments configurable 2011-09-08 11:10:56 +02:00
tick-sched.c nohz: Include local CPU in full dynticks global kick 2013-08-16 17:55:33 +02:00
timeconv.c time: add function to convert between calendar time and broken-down time for universal use 2009-09-24 07:20:56 -07:00
timekeeping_debug.c power: Add option to log time spent in suspend 2013-05-29 12:57:34 -07:00
timekeeping_internal.h power: Add option to log time spent in suspend 2013-05-29 12:57:34 -07:00
timekeeping.c timekeeping: Fix HRTICK related deadlock from ntp lock changes 2013-09-12 07:49:51 +02:00
timer_list.c timer_list: Convert timer list to be a proper seq_file 2013-04-17 20:51:02 +02:00
timer_stats.c locking, timer_stats: Annotate table_lock as raw 2011-09-13 11:12:00 +02:00