tick: Fix oneshot broadcast setup really
authorThomas Gleixner <tglx@linutronix.de>
Wed, 18 Apr 2012 10:08:23 +0000 (12:08 +0200)
committerThomas Gleixner <tglx@linutronix.de>
Wed, 18 Apr 2012 12:00:56 +0000 (14:00 +0200)
commitb435092f70ec5ebbfb6d075d5bf3c631b49a51de
treec9cfd15c405dd218003fd84585196b5bffcc9bbd
parent592fe8980688e7cba46897685d014c7fb3018a67
tick: Fix oneshot broadcast setup really

Sven Joachim reported, that suspend/resume on rc3 trips over a NULL
pointer dereference. Linus spotted the clockevent handler being NULL.

commit fa4da365b(clockevents: tTack broadcast device mode change in
tick_broadcast_switch_to_oneshot()) tried to fix a problem with the
broadcast device setup, which was introduced in commit 77b0d60c5(
clockevents: Leave the broadcast device in shutdown mode when not
needed).

The initial commit avoided to set up the broadcast device when no
broadcast request bits were set, but that left the broadcast device
disfunctional. In consequence deep idle states which need the
broadcast device were not woken up.

commit fa4da365b tried to fix that by initializing the state of the
broadcast facility, but that missed the fact, that nothing initializes
the event handler and some other state of the underlying clock event
device.

The fix is to revert both commits and make only the mode setting of
the clock event device conditional on the state of active broadcast
users.

That initializes everything except the low level device mode, but this
happens when the broadcast functionality is invoked by deep idle.

Reported-and-tested-by: Sven Joachim <svenjoac@gmx.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Rafael J. Wysocki <rjw@sisk.pl>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/alpine.LFD.2.02.1204181205540.2542@ionos
kernel/time/tick-broadcast.c