alarmtimer: Rate limit periodic intervals

[ Upstream commit ff86bf0 ]

The alarmtimer code has another source of potentially rearming itself too
fast. Interval timers with a very samll interval have a similar CPU hog
effect as the previously fixed overflow issue.

The reason is that alarmtimers do not implement the normal protection
against this kind of problem which the other posix timer use:

  timer expires -> queue signal -> deliver signal -> rearm timer

This scheme brings the rearming under scheduler control and prevents
permanently firing timers which hog the CPU.

Bringing this scheme to the alarm timer code is a major overhaul because it
lacks all the necessary mechanisms completely.

So for a quick fix limit the interval to one jiffie. This is not
problematic in practice as alarmtimers are usually backed by an RTC for
suspend which have 1 second resolution. It could be therefor argued that
the resolution of this clock should be set to 1 second in general, but
that's outside the scope of this fix.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Kostya Serebryany <kcc@google.com>
Cc: syzkaller <syzkaller@googlegroups.com>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: stable@vger.kernel.org
Link: http://lkml.kernel.org/r/20170530211655.896767100@linutronix.de
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>

kernel/time/alarmtimer.c

@@ -614,6 +614,14 @@ static int alarm_timer_set(struct k_itimer *timr, int flags,
 
 	/* start the timer */
 	timr->it.alarm.interval = timespec_to_ktime(new_setting->it_interval);
+
+	/*
+	 * Rate limit to the tick as a hot fix to prevent DOS. Will be
+	 * mopped up later.
+	 */
+	if (ktime_to_ns(timr->it.alarm.interval) < TICK_NSEC)
+		timr->it.alarm.interval = ktime_set(0, TICK_NSEC);
+
 	exp = timespec_to_ktime(new_setting->it_value);
 	/* Convert (if necessary) to absolute time */
 	if (flags != TIMER_ABSTIME) {