[PATCH 3.2 057/106] alarmtimer: Prevent overflow of relative timers

[Ben Hutchings]

3.2.93-rc1 review patch.  If anyone has any objections, please let me know.

------------------

From: Thomas Gleixner <tglx@xxxxxxxxxxxxx>

commit f4781e76f90df7aec400635d73ea4c35ee1d4765 upstream.

Andrey reported a alartimer related RCU stall while fuzzing the kernel with
syzkaller.

The reason for this is an overflow in ktime_add() which brings the
resulting time into negative space and causes immediate expiry of the
timer. The following rearm with a small interval does not bring the timer
back into positive space due to the same issue.

This results in a permanent firing alarmtimer which hogs the CPU.

Use ktime_add_safe() instead which detects the overflow and clamps the
result to KTIME_SEC_MAX.

Reported-by: Andrey Konovalov <andreyknvl@xxxxxxxxxx>
Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Cc: Kostya Serebryany <kcc@xxxxxxxxxx>
Cc: syzkaller <syzkaller@xxxxxxxxxxxxxxxx>
Cc: John Stultz <john.stultz@xxxxxxxxxx>
Cc: Dmitry Vyukov <dvyukov@xxxxxxxxxx>
Link: http://lkml.kernel.org/r/20170530211655.802921648@xxxxxxxxxxxxx
[bwh: Backported to 3.2: drop change in alarm_start_relative()]
Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
---
--- a/kernel/time/alarmtimer.c
+++ b/kernel/time/alarmtimer.c
@@ -419,7 +419,7 @@ u64 alarm_forward(struct alarm *alarm, k
 		overrun++;
 	}
 
-	alarm->node.expires = ktime_add(alarm->node.expires, interval);
+	alarm->node.expires = ktime_add_safe(alarm->node.expires, interval);
 	return overrun;
 }
 
@@ -604,7 +604,7 @@ static int alarm_timer_set(struct k_itim
 		ktime_t now;
 
 		now = alarm_bases[timr->it.alarm.alarmtimer.type].gettime();
-		exp = ktime_add(now, exp);
+		exp = ktime_add_safe(now, exp);
 	}
 
 	alarm_start(&timr->it.alarm.alarmtimer, exp);