Upstream commit c725dafc95f1b37027840aaeaa8b7e4e9cd20516 PREEMPT_RT does not spin and wait until a running timer completes its callback but instead it blocks on a sleeping lock to prevent a livelock in the case that the task waiting for the callback completion preempted the callback. This cannot be done for timers flagged with TIMER_IRQSAFE. These timers can be canceled from an interrupt disabled context even on RT kernels. The expiry callback of such timers is invoked with interrupts disabled so there is no need to use the expiry lock mechanism because obviously the callback cannot be preempted even on RT kernels. Do not use the timer_base::expiry_lock mechanism when waiting for a running callback to complete if the timer is flagged with TIMER_IRQSAFE. Also add a lockdep assertion for RT kernels to validate that the expiry lock mechanism is always invoked in preemptible context. [bigeasy: The logic in v4.9 is slightly different but the outcome is the same as we must not sleep while waiting for the irqsafe timer to complete. The IRQSAFE timer can not be preempted. The "lockdep annotation" is not available and has been replaced with might_sleep()] Reported-by: Mike Galbraith <efault@xxxxxx> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx> Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx> Link: https://lore.kernel.org/r/20201103190937.hga67rqhvknki3tp@xxxxxxxxxxxxx Signed-off-by: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx> --- kernel/time/timer.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/kernel/time/timer.c b/kernel/time/timer.c index 613139c7538eb..401917af2abcf 100644 --- a/kernel/time/timer.c +++ b/kernel/time/timer.c @@ -1179,9 +1179,9 @@ EXPORT_SYMBOL_GPL(add_timer_on); static void wait_for_running_timer(struct timer_list *timer) { struct timer_base *base; - u32 tf = timer->flags; + u32 tf = READ_ONCE(timer->flags); - if (tf & TIMER_MIGRATING) + if (tf & (TIMER_MIGRATING | TIMER_IRQSAFE)) return; base = get_timer_base(tf); @@ -1312,6 +1312,13 @@ int del_timer_sync(struct timer_list *timer) * could lead to deadlock. */ WARN_ON(in_irq() && !(timer->flags & TIMER_IRQSAFE)); + /* + * Must be able to sleep on PREEMPT_RT because of the slowpath in + * del_timer_wait_running(). + */ + if (IS_ENABLED(CONFIG_PREEMPT_RT) && !(timer->flags & TIMER_IRQSAFE)) + might_sleep(); + for (;;) { int ret = try_to_del_timer_sync(timer); if (ret >= 0) -- 2.37.2
