Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/avi/kvm
[pandora-kernel.git] / kernel / timer.c
index 1258371..fb4e67d 100644 (file)
@@ -26,6 +26,7 @@
 #include <linux/init.h>
 #include <linux/mm.h>
 #include <linux/swap.h>
+#include <linux/pid_namespace.h>
 #include <linux/notifier.h>
 #include <linux/thread_info.h>
 #include <linux/time.h>
@@ -103,14 +104,14 @@ static inline tvec_base_t *tbase_get_base(tvec_base_t *base)
 static inline void timer_set_deferrable(struct timer_list *timer)
 {
        timer->base = ((tvec_base_t *)((unsigned long)(timer->base) |
-                                      TBASE_DEFERRABLE_FLAG));
+                                      TBASE_DEFERRABLE_FLAG));
 }
 
 static inline void
 timer_set_base(struct timer_list *timer, tvec_base_t *new_base)
 {
        timer->base = (tvec_base_t *)((unsigned long)(new_base) |
-                                     tbase_get_deferrable(timer->base));
+                                     tbase_get_deferrable(timer->base));
 }
 
 /**
@@ -445,10 +446,10 @@ EXPORT_SYMBOL(__mod_timer);
 void add_timer_on(struct timer_list *timer, int cpu)
 {
        tvec_base_t *base = per_cpu(tvec_bases, cpu);
-       unsigned long flags;
+       unsigned long flags;
 
        timer_stats_timer_set_start_info(timer);
-       BUG_ON(timer_pending(timer) || !timer->function);
+       BUG_ON(timer_pending(timer) || !timer->function);
        spin_lock_irqsave(&base->lock, flags);
        timer_set_base(timer, base);
        internal_add_timer(base, timer);
@@ -627,7 +628,7 @@ static inline void __run_timers(tvec_base_t *base)
        while (time_after_eq(jiffies, base->timer_jiffies)) {
                struct list_head work_list;
                struct list_head *head = &work_list;
-               int index = base->timer_jiffies & TVR_MASK;
+               int index = base->timer_jiffies & TVR_MASK;
 
                /*
                 * Cascade timers:
@@ -644,8 +645,8 @@ static inline void __run_timers(tvec_base_t *base)
                        unsigned long data;
 
                        timer = list_first_entry(head, struct timer_list,entry);
-                       fn = timer->function;
-                       data = timer->data;
+                       fn = timer->function;
+                       data = timer->data;
 
                        timer_stats_account_timer(timer);
 
@@ -689,8 +690,8 @@ static unsigned long __next_timer_interrupt(tvec_base_t *base)
        index = slot = timer_jiffies & TVR_MASK;
        do {
                list_for_each_entry(nte, base->tv1.vec + slot, entry) {
-                       if (tbase_get_deferrable(nte->base))
-                               continue;
+                       if (tbase_get_deferrable(nte->base))
+                               continue;
 
                        found = 1;
                        expires = nte->expires;
@@ -817,7 +818,7 @@ unsigned long next_timer_interrupt(void)
 #endif
 
 /*
- * Called from the timer interrupt handler to charge one tick to the current 
+ * Called from the timer interrupt handler to charge one tick to the current
  * process.  user_tick is 1 if the tick is user time, 0 for system.
  */
 void update_process_times(int user_tick)
@@ -826,15 +827,18 @@ void update_process_times(int user_tick)
        int cpu = smp_processor_id();
 
        /* Note: this timer irq context must be accounted for as well. */
-       if (user_tick)
+       if (user_tick) {
                account_user_time(p, jiffies_to_cputime(1));
-       else
+               account_user_time_scaled(p, jiffies_to_cputime(1));
+       } else {
                account_system_time(p, HARDIRQ_OFFSET, jiffies_to_cputime(1));
+               account_system_time_scaled(p, jiffies_to_cputime(1));
+       }
        run_local_timers();
        if (rcu_pending(cpu))
                rcu_check_callbacks(cpu, user_tick);
        scheduler_tick();
-       run_posix_cpu_timers(p);
+       run_posix_cpu_timers(p);
 }
 
 /*
@@ -909,7 +913,7 @@ static inline void update_times(unsigned long ticks)
        update_wall_time();
        calc_load(ticks);
 }
-  
+
 /*
  * The 64-bit jiffies value is not atomic - you MUST NOT read it
  * without sampling the sequence number in xtime_lock.
@@ -953,7 +957,7 @@ asmlinkage unsigned long sys_alarm(unsigned int seconds)
  */
 asmlinkage long sys_getpid(void)
 {
-       return current->tgid;
+       return task_tgid_vnr(current);
 }
 
 /*
@@ -967,7 +971,7 @@ asmlinkage long sys_getppid(void)
        int pid;
 
        rcu_read_lock();
-       pid = rcu_dereference(current->real_parent)->tgid;
+       pid = task_ppid_nr_ns(current, current->nsproxy->pid_ns);
        rcu_read_unlock();
 
        return pid;
@@ -1099,13 +1103,13 @@ EXPORT_SYMBOL(schedule_timeout_uninterruptible);
 /* Thread ID - the internal kernel "pid" */
 asmlinkage long sys_gettid(void)
 {
-       return current->pid;
+       return task_pid_vnr(current);
 }
 
 /**
  * do_sysinfo - fill in sysinfo struct
  * @info: pointer to buffer to fill
- */ 
+ */
 int do_sysinfo(struct sysinfo *info)
 {
        unsigned long mem_total, sav_total;
@@ -1221,7 +1225,8 @@ static int __devinit init_timers_cpu(int cpu)
                        /*
                         * The APs use this path later in boot
                         */
-                       base = kmalloc_node(sizeof(*base), GFP_KERNEL,
+                       base = kmalloc_node(sizeof(*base),
+                                               GFP_KERNEL | __GFP_ZERO,
                                                cpu_to_node(cpu));
                        if (!base)
                                return -ENOMEM;
@@ -1232,7 +1237,6 @@ static int __devinit init_timers_cpu(int cpu)
                                kfree(base);
                                return -ENOMEM;
                        }
-                       memset(base, 0, sizeof(*base));
                        per_cpu(tvec_bases, cpu) = base;
                } else {
                        /*
@@ -1349,194 +1353,6 @@ void __init init_timers(void)
        open_softirq(TIMER_SOFTIRQ, run_timer_softirq, NULL);
 }
 
-#ifdef CONFIG_TIME_INTERPOLATION
-
-struct time_interpolator *time_interpolator __read_mostly;
-static struct time_interpolator *time_interpolator_list __read_mostly;
-static DEFINE_SPINLOCK(time_interpolator_lock);
-
-static inline cycles_t time_interpolator_get_cycles(unsigned int src)
-{
-       unsigned long (*x)(void);
-
-       switch (src)
-       {
-               case TIME_SOURCE_FUNCTION:
-                       x = time_interpolator->addr;
-                       return x();
-
-               case TIME_SOURCE_MMIO64 :
-                       return readq_relaxed((void __iomem *)time_interpolator->addr);
-
-               case TIME_SOURCE_MMIO32 :
-                       return readl_relaxed((void __iomem *)time_interpolator->addr);
-
-               default: return get_cycles();
-       }
-}
-
-static inline u64 time_interpolator_get_counter(int writelock)
-{
-       unsigned int src = time_interpolator->source;
-
-       if (time_interpolator->jitter)
-       {
-               cycles_t lcycle;
-               cycles_t now;
-
-               do {
-                       lcycle = time_interpolator->last_cycle;
-                       now = time_interpolator_get_cycles(src);
-                       if (lcycle && time_after(lcycle, now))
-                               return lcycle;
-
-                       /* When holding the xtime write lock, there's no need
-                        * to add the overhead of the cmpxchg.  Readers are
-                        * force to retry until the write lock is released.
-                        */
-                       if (writelock) {
-                               time_interpolator->last_cycle = now;
-                               return now;
-                       }
-                       /* Keep track of the last timer value returned. The use of cmpxchg here
-                        * will cause contention in an SMP environment.
-                        */
-               } while (unlikely(cmpxchg(&time_interpolator->last_cycle, lcycle, now) != lcycle));
-               return now;
-       }
-       else
-               return time_interpolator_get_cycles(src);
-}
-
-void time_interpolator_reset(void)
-{
-       time_interpolator->offset = 0;
-       time_interpolator->last_counter = time_interpolator_get_counter(1);
-}
-
-#define GET_TI_NSECS(count,i) (((((count) - i->last_counter) & (i)->mask) * (i)->nsec_per_cyc) >> (i)->shift)
-
-unsigned long time_interpolator_get_offset(void)
-{
-       /* If we do not have a time interpolator set up then just return zero */
-       if (!time_interpolator)
-               return 0;
-
-       return time_interpolator->offset +
-               GET_TI_NSECS(time_interpolator_get_counter(0), time_interpolator);
-}
-
-#define INTERPOLATOR_ADJUST 65536
-#define INTERPOLATOR_MAX_SKIP 10*INTERPOLATOR_ADJUST
-
-void time_interpolator_update(long delta_nsec)
-{
-       u64 counter;
-       unsigned long offset;
-
-       /* If there is no time interpolator set up then do nothing */
-       if (!time_interpolator)
-               return;
-
-       /*
-        * The interpolator compensates for late ticks by accumulating the late
-        * time in time_interpolator->offset. A tick earlier than expected will
-        * lead to a reset of the offset and a corresponding jump of the clock
-        * forward. Again this only works if the interpolator clock is running
-        * slightly slower than the regular clock and the tuning logic insures
-        * that.
-        */
-
-       counter = time_interpolator_get_counter(1);
-       offset = time_interpolator->offset +
-                       GET_TI_NSECS(counter, time_interpolator);
-
-       if (delta_nsec < 0 || (unsigned long) delta_nsec < offset)
-               time_interpolator->offset = offset - delta_nsec;
-       else {
-               time_interpolator->skips++;
-               time_interpolator->ns_skipped += delta_nsec - offset;
-               time_interpolator->offset = 0;
-       }
-       time_interpolator->last_counter = counter;
-
-       /* Tuning logic for time interpolator invoked every minute or so.
-        * Decrease interpolator clock speed if no skips occurred and an offset is carried.
-        * Increase interpolator clock speed if we skip too much time.
-        */
-       if (jiffies % INTERPOLATOR_ADJUST == 0)
-       {
-               if (time_interpolator->skips == 0 && time_interpolator->offset > tick_nsec)
-                       time_interpolator->nsec_per_cyc--;
-               if (time_interpolator->ns_skipped > INTERPOLATOR_MAX_SKIP && time_interpolator->offset == 0)
-                       time_interpolator->nsec_per_cyc++;
-               time_interpolator->skips = 0;
-               time_interpolator->ns_skipped = 0;
-       }
-}
-
-static inline int
-is_better_time_interpolator(struct time_interpolator *new)
-{
-       if (!time_interpolator)
-               return 1;
-       return new->frequency > 2*time_interpolator->frequency ||
-           (unsigned long)new->drift < (unsigned long)time_interpolator->drift;
-}
-
-void
-register_time_interpolator(struct time_interpolator *ti)
-{
-       unsigned long flags;
-
-       /* Sanity check */
-       BUG_ON(ti->frequency == 0 || ti->mask == 0);
-
-       ti->nsec_per_cyc = ((u64)NSEC_PER_SEC << ti->shift) / ti->frequency;
-       spin_lock(&time_interpolator_lock);
-       write_seqlock_irqsave(&xtime_lock, flags);
-       if (is_better_time_interpolator(ti)) {
-               time_interpolator = ti;
-               time_interpolator_reset();
-       }
-       write_sequnlock_irqrestore(&xtime_lock, flags);
-
-       ti->next = time_interpolator_list;
-       time_interpolator_list = ti;
-       spin_unlock(&time_interpolator_lock);
-}
-
-void
-unregister_time_interpolator(struct time_interpolator *ti)
-{
-       struct time_interpolator *curr, **prev;
-       unsigned long flags;
-
-       spin_lock(&time_interpolator_lock);
-       prev = &time_interpolator_list;
-       for (curr = *prev; curr; curr = curr->next) {
-               if (curr == ti) {
-                       *prev = curr->next;
-                       break;
-               }
-               prev = &curr->next;
-       }
-
-       write_seqlock_irqsave(&xtime_lock, flags);
-       if (ti == time_interpolator) {
-               /* we lost the best time-interpolator: */
-               time_interpolator = NULL;
-               /* find the next-best interpolator */
-               for (curr = time_interpolator_list; curr; curr = curr->next)
-                       if (is_better_time_interpolator(curr))
-                               time_interpolator = curr;
-               time_interpolator_reset();
-       }
-       write_sequnlock_irqrestore(&xtime_lock, flags);
-       spin_unlock(&time_interpolator_lock);
-}
-#endif /* CONFIG_TIME_INTERPOLATION */
-
 /**
  * msleep - sleep safely even with waitqueue interruptions
  * @msecs: Time in milliseconds to sleep for