timers: split process wide cpu clocks/timers

Change the process wide cpu timers/clocks so that we:

 1) don't mess up the kernel with too many threads,
 2) don't have a per-cpu allocation for each process,
 3) have no impact when not used.

In order to accomplish this we're going to split it into two parts:

 - clocks; which can take all the time they want since they run
           from user context -- ie. sys_clock_gettime(CLOCK_PROCESS_CPUTIME_ID)

 - timers; which need constant time sampling but since they're
           explicity used, the user can pay the overhead.

The clock readout will go back to a full sum of the thread group, while the
timers will run of a global 'clock' that only runs when needed, so only
programs that make use of the facility pay the price.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Reviewed-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

5 files changed, 155 insertions, 54 deletions

diff --git a/include/linux/init_task.h b/include/linux/init_task.h
index ea0ea1a4c36f..e752d973fa21 100644
--- a/include/linux/init_task.h
+++ b/include/linux/init_task.h
@@ -48,12 +48,11 @@ extern struct fs_struct init_fs;
         .posix_timers    = LIST_HEAD_INIT(sig.posix_timers),            \
         .cpu_timers     = INIT_CPU_TIMERS(sig.cpu_timers),              \
         .rlim           = INIT_RLIMITS,                                 \
-        .cputime        = { .totals = {                                 \
-                .utime = cputime_zero,                                  \
-                .stime = cputime_zero,                                  \
-                .sum_exec_runtime = 0,                                  \
-                .lock = __SPIN_LOCK_UNLOCKED(sig.cputime.totals.lock),  \
-        }, },                                                           \
+        .cputimer       = {                                             \
+                .cputime = INIT_CPUTIME,                                \
+                .running = 0,                                           \
+                .lock = __SPIN_LOCK_UNLOCKED(sig.cputimer.lock),        \
+        },                                                              \
 }
 
 extern struct nsproxy init_nsproxy;
diff --git a/include/linux/sched.h b/include/linux/sched.h
index 2e0646a30314..082d7619b3a1 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -443,7 +443,6 @@ struct pacct_struct {
  * @utime:              time spent in user mode, in &cputime_t units
  * @stime:              time spent in kernel mode, in &cputime_t units
  * @sum_exec_runtime:   total time spent on the CPU, in nanoseconds
- * @lock:               lock for fields in this struct
  *
  * This structure groups together three kinds of CPU time that are
  * tracked for threads and thread groups.  Most things considering
@@ -454,23 +453,33 @@ struct task_cputime {
         cputime_t utime;
         cputime_t stime;
         unsigned long long sum_exec_runtime;
-        spinlock_t lock;
 };
 /* Alternate field names when used to cache expirations. */
 #define prof_exp        stime
 #define virt_exp        utime
 #define sched_exp       sum_exec_runtime
 
+#define INIT_CPUTIME    \
+        (struct task_cputime) {                                 \
+                .utime = cputime_zero,                          \
+                .stime = cputime_zero,                          \
+                .sum_exec_runtime = 0,                          \
+        }
+
 /**
- * struct thread_group_cputime - thread group interval timer counts
- * @totals:             thread group interval timers; substructure for
- *                      uniprocessor kernel, per-cpu for SMP kernel.
+ * struct thread_group_cputimer - thread group interval timer counts
+ * @cputime:            thread group interval timers.
+ * @running:            non-zero when there are timers running and
+ *                      @cputime receives updates.
+ * @lock:               lock for fields in this struct.
  *
  * This structure contains the version of task_cputime, above, that is
- * used for thread group CPU clock calculations.
+ * used for thread group CPU timer calculations.
  */
-struct thread_group_cputime {
-        struct task_cputime totals;
+struct thread_group_cputimer {
+        struct task_cputime cputime;
+        int running;
+        spinlock_t lock;
 };
 
 /*
@@ -519,10 +528,10 @@ struct signal_struct {
         cputime_t it_prof_incr, it_virt_incr;
 
         /*
-         * Thread group totals for process CPU clocks.
-         * See thread_group_cputime(), et al, for details.
+         * Thread group totals for process CPU timers.
+         * See thread_group_cputimer(), et al, for details.
          */
-        struct thread_group_cputime cputime;
+        struct thread_group_cputimer cputimer;
 
         /* Earliest-expiration cache. */
         struct task_cputime cputime_expires;
@@ -2191,27 +2200,26 @@ static inline int spin_needbreak(spinlock_t *lock)
 /*
  * Thread group CPU time accounting.
  */
+void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times);
 
 static inline
-void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times)
+void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times)
 {
-        struct task_cputime *totals = &tsk->signal->cputime.totals;
+        struct thread_group_cputimer *cputimer = &tsk->signal->cputimer;
         unsigned long flags;
 
-        spin_lock_irqsave(&totals->lock, flags);
-        *times = *totals;
-        spin_unlock_irqrestore(&totals->lock, flags);
+        WARN_ON(!cputimer->running);
+
+        spin_lock_irqsave(&cputimer->lock, flags);
+        *times = cputimer->cputime;
+        spin_unlock_irqrestore(&cputimer->lock, flags);
 }
 
 static inline void thread_group_cputime_init(struct signal_struct *sig)
 {
-        sig->cputime.totals = (struct task_cputime){
-                .utime = cputime_zero,
-                .stime = cputime_zero,
-                .sum_exec_runtime = 0,
-        };
-
-        spin_lock_init(&sig->cputime.totals.lock);
+        sig->cputimer.cputime = INIT_CPUTIME;
+        spin_lock_init(&sig->cputimer.lock);
+        sig->cputimer.running = 0;
 }
 
 static inline void thread_group_cputime_free(struct signal_struct *sig)
diff --git a/kernel/itimer.c b/kernel/itimer.c
index 6a5fe93dd8bd..58762f7077ec 100644
--- a/kernel/itimer.c
+++ b/kernel/itimer.c
@@ -62,7 +62,7 @@ int do_getitimer(int which, struct itimerval *value)
                         struct task_cputime cputime;
                         cputime_t utime;
 
-                        thread_group_cputime(tsk, &cputime);
+                        thread_group_cputimer(tsk, &cputime);
                         utime = cputime.utime;
                         if (cputime_le(cval, utime)) { /* about to fire */
                                 cval = jiffies_to_cputime(1);
@@ -82,7 +82,7 @@ int do_getitimer(int which, struct itimerval *value)
                         struct task_cputime times;
                         cputime_t ptime;
 
-                        thread_group_cputime(tsk, &times);
+                        thread_group_cputimer(tsk, &times);
                         ptime = cputime_add(times.utime, times.stime);
                         if (cputime_le(cval, ptime)) { /* about to fire */
                                 cval = jiffies_to_cputime(1);
diff --git a/kernel/posix-cpu-timers.c b/kernel/posix-cpu-timers.c
index fa07da94d7be..db107c9bbc05 100644
--- a/kernel/posix-cpu-timers.c
+++ b/kernel/posix-cpu-timers.c
@@ -230,6 +230,37 @@ static int cpu_clock_sample(const clockid_t which_clock, struct task_struct *p,
         return 0;
 }
 
+void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times)
+{
+        struct sighand_struct *sighand;
+        struct signal_struct *sig;
+        struct task_struct *t;
+
+        *times = INIT_CPUTIME;
+
+        rcu_read_lock();
+        sighand = rcu_dereference(tsk->sighand);
+        if (!sighand)
+                goto out;
+
+        sig = tsk->signal;
+
+        t = tsk;
+        do {
+                times->utime = cputime_add(times->utime, t->utime);
+                times->stime = cputime_add(times->stime, t->stime);
+                times->sum_exec_runtime += t->se.sum_exec_runtime;
+
+                t = next_thread(t);
+        } while (t != tsk);
+
+        times->utime = cputime_add(times->utime, sig->utime);
+        times->stime = cputime_add(times->stime, sig->stime);
+        times->sum_exec_runtime += sig->sum_sched_runtime;
+out:
+        rcu_read_unlock();
+}
+
 /*
  * Sample a process (thread group) clock for the given group_leader task.
  * Must be called with tasklist_lock held for reading.
@@ -476,6 +507,29 @@ static void clear_dead_task(struct k_itimer *timer, union cpu_time_count now)
 }
 
 /*
+ * Enable the process wide cpu timer accounting.
+ *
+ * serialized using ->sighand->siglock
+ */
+static void start_process_timers(struct task_struct *tsk)
+{
+        tsk->signal->cputimer.running = 1;
+        barrier();
+}
+
+/*
+ * Release the process wide timer accounting -- timer stops ticking when
+ * nobody cares about it.
+ *
+ * serialized using ->sighand->siglock
+ */
+static void stop_process_timers(struct task_struct *tsk)
+{
+        tsk->signal->cputimer.running = 0;
+        barrier();
+}
+
+/*
  * Insert the timer on the appropriate list before any timers that
  * expire later.  This must be called with the tasklist_lock held
  * for reading, and interrupts disabled.
@@ -495,6 +549,9 @@ static void arm_timer(struct k_itimer *timer, union cpu_time_count now)
         BUG_ON(!irqs_disabled());
         spin_lock(&p->sighand->siglock);
 
+        if (!CPUCLOCK_PERTHREAD(timer->it_clock))
+                start_process_timers(p);
+
         listpos = head;
         if (CPUCLOCK_WHICH(timer->it_clock) == CPUCLOCK_SCHED) {
                 list_for_each_entry(next, head, entry) {
@@ -987,13 +1044,15 @@ static void check_process_timers(struct task_struct *tsk,
             sig->rlim[RLIMIT_CPU].rlim_cur == RLIM_INFINITY &&
             list_empty(&timers[CPUCLOCK_VIRT]) &&
             cputime_eq(sig->it_virt_expires, cputime_zero) &&
-            list_empty(&timers[CPUCLOCK_SCHED]))
+            list_empty(&timers[CPUCLOCK_SCHED])) {
+                stop_process_timers(tsk);
                 return;
+        }
 
         /*
          * Collect the current process totals.
          */
-        thread_group_cputime(tsk, &cputime);
+        thread_group_cputimer(tsk, &cputime);
         utime = cputime.utime;
         ptime = cputime_add(utime, cputime.stime);
         sum_sched_runtime = cputime.sum_exec_runtime;
@@ -1259,7 +1318,7 @@ static inline int fastpath_timer_check(struct task_struct *tsk)
         if (!task_cputime_zero(&sig->cputime_expires)) {
                 struct task_cputime group_sample;
 
-                thread_group_cputime(tsk, &group_sample);
+                thread_group_cputimer(tsk, &group_sample);
                 if (task_cputime_expired(&group_sample, &sig->cputime_expires))
                         return 1;
         }
@@ -1329,6 +1388,33 @@ void run_posix_cpu_timers(struct task_struct *tsk)
 }
 
 /*
+ * Sample a process (thread group) timer for the given group_leader task.
+ * Must be called with tasklist_lock held for reading.
+ */
+static int cpu_timer_sample_group(const clockid_t which_clock,
+                                  struct task_struct *p,
+                                  union cpu_time_count *cpu)
+{
+        struct task_cputime cputime;
+
+        thread_group_cputimer(p, &cputime);
+        switch (CPUCLOCK_WHICH(which_clock)) {
+        default:
+                return -EINVAL;
+        case CPUCLOCK_PROF:
+                cpu->cpu = cputime_add(cputime.utime, cputime.stime);
+                break;
+        case CPUCLOCK_VIRT:
+                cpu->cpu = cputime.utime;
+                break;
+        case CPUCLOCK_SCHED:
+                cpu->sched = cputime.sum_exec_runtime + task_delta_exec(p);
+                break;
+        }
+        return 0;
+}
+
+/*
  * Set one of the process-wide special case CPU timers.
  * The tsk->sighand->siglock must be held by the caller.
  * The *newval argument is relative and we update it to be absolute, *oldval
@@ -1341,7 +1427,8 @@ void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx,
         struct list_head *head;
 
         BUG_ON(clock_idx == CPUCLOCK_SCHED);
-        cpu_clock_sample_group(clock_idx, tsk, &now);
+        start_process_timers(tsk);
+        cpu_timer_sample_group(clock_idx, tsk, &now);
 
         if (oldval) {
                 if (!cputime_eq(*oldval, cputime_zero)) {
diff --git a/kernel/sched_stats.h b/kernel/sched_stats.h
index 8ab0cef8ecab..a8f93dd374e1 100644
--- a/kernel/sched_stats.h
+++ b/kernel/sched_stats.h
@@ -296,19 +296,21 @@ sched_info_switch(struct task_struct *prev, struct task_struct *next)
 static inline void account_group_user_time(struct task_struct *tsk,
                                            cputime_t cputime)
 {
-        struct task_cputime *times;
-        struct signal_struct *sig;
+        struct thread_group_cputimer *cputimer;
 
         /* tsk == current, ensure it is safe to use ->signal */
         if (unlikely(tsk->exit_state))
                 return;
 
-        sig = tsk->signal;
-        times = &sig->cputime.totals;
+        cputimer = &tsk->signal->cputimer;
 
-        spin_lock(&times->lock);
-        times->utime = cputime_add(times->utime, cputime);
-        spin_unlock(&times->lock);
+        if (!cputimer->running)
+                return;
+
+        spin_lock(&cputimer->lock);
+        cputimer->cputime.utime =
+                cputime_add(cputimer->cputime.utime, cputime);
+        spin_unlock(&cputimer->lock);
 }
 
 /**
@@ -324,19 +326,21 @@ static inline void account_group_user_time(struct task_struct *tsk,
 static inline void account_group_system_time(struct task_struct *tsk,
                                              cputime_t cputime)
 {
-        struct task_cputime *times;
-        struct signal_struct *sig;
+        struct thread_group_cputimer *cputimer;
 
         /* tsk == current, ensure it is safe to use ->signal */
         if (unlikely(tsk->exit_state))
                 return;
 
-        sig = tsk->signal;
-        times = &sig->cputime.totals;
+        cputimer = &tsk->signal->cputimer;
+
+        if (!cputimer->running)
+                return;
 
-        spin_lock(&times->lock);
-        times->stime = cputime_add(times->stime, cputime);
-        spin_unlock(&times->lock);
+        spin_lock(&cputimer->lock);
+        cputimer->cputime.stime =
+                cputime_add(cputimer->cputime.stime, cputime);
+        spin_unlock(&cputimer->lock);
 }
 
 /**
@@ -352,7 +356,7 @@ static inline void account_group_system_time(struct task_struct *tsk,
 static inline void account_group_exec_runtime(struct task_struct *tsk,
                                               unsigned long long ns)
 {
-        struct task_cputime *times;
+        struct thread_group_cputimer *cputimer;
         struct signal_struct *sig;
 
         sig = tsk->signal;
@@ -361,9 +365,12 @@ static inline void account_group_exec_runtime(struct task_struct *tsk,
         if (unlikely(!sig))
                 return;
 
-        times = &sig->cputime.totals;
+        cputimer = &sig->cputimer;
+
+        if (!cputimer->running)
+                return;
 
-        spin_lock(&times->lock);
-        times->sum_exec_runtime += ns;
-        spin_unlock(&times->lock);
+        spin_lock(&cputimer->lock);
+        cputimer->cputime.sum_exec_runtime += ns;
+        spin_unlock(&cputimer->lock);
 }