[PATCH] too much timer simplification...

Noted by David Mosberger:

 "If someone happens to arm a periodic timer at exactly 256 jiffies (as
  ohci happens to do on platforms with HZ=1024), then you end up getting
  an endless loop of timer activations, causing a machine hang.

  The problem is that __run_timers updates base->timer_jiffies _before_
  running the callback routines.  If a callback re-arms the timer at
  exactly 256 jiffies, add_timers() will reinsert the timer into the list
  that we're currently processing, which of course will cause the timer to
  expire immediately again, etc., etc., ad naseum... "

The answer here is to move the whole expired list to a local header and
to not look back.

kernel/timer.c

@@ -397,7 +397,8 @@ static inline void __run_timers(tvec_base_t *base)
 
 	spin_lock_irq(&base->lock);
 	while (time_after_eq(jiffies, base->timer_jiffies)) {
-		struct list_head *head;
+		struct list_head work_list = LIST_HEAD_INIT(work_list);
+		struct list_head *head = &work_list;
  		int index = base->timer_jiffies & TVR_MASK;
  
 		/*
@@ -409,8 +410,8 @@ static inline void __run_timers(tvec_base_t *base)
 					!cascade(base, &base->tv4, INDEX(2)))
 			cascade(base, &base->tv5, INDEX(3));
 		++base->timer_jiffies; 
+		list_splice_init(base->tv1.vec + index, &work_list);
 repeat:
-		head = base->tv1.vec + index;
 		if (!list_empty(head)) {
 			void (*fn)(unsigned long);
 			unsigned long data;