Merge master.kernel.org:/home/rmk/linux-2.6-arm
[linux-2.6.git] / kernel / softlockup.c
index 884e6cd..0d4c789 100644 (file)
 #include <linux/lockdep.h>
 #include <linux/notifier.h>
 #include <linux/module.h>
+#include <linux/sysctl.h>
 
 #include <asm/irq_regs.h>
 
 static DEFINE_SPINLOCK(print_lock);
 
-static DEFINE_PER_CPU(unsigned long, touch_timestamp);
-static DEFINE_PER_CPU(unsigned long, print_timestamp);
-static DEFINE_PER_CPU(struct task_struct *, watchdog_task);
+static DEFINE_PER_CPU(unsigned long, softlockup_touch_ts); /* touch timestamp */
+static DEFINE_PER_CPU(unsigned long, softlockup_print_ts); /* print timestamp */
+static DEFINE_PER_CPU(struct task_struct *, softlockup_watchdog);
+static DEFINE_PER_CPU(bool, softlock_touch_sync);
 
 static int __read_mostly did_panic;
 int __read_mostly softlockup_thresh = 60;
@@ -69,25 +71,39 @@ static void __touch_softlockup_watchdog(void)
 {
        int this_cpu = raw_smp_processor_id();
 
-       __raw_get_cpu_var(touch_timestamp) = get_timestamp(this_cpu);
+       __raw_get_cpu_var(softlockup_touch_ts) = get_timestamp(this_cpu);
 }
 
 void touch_softlockup_watchdog(void)
 {
-       __raw_get_cpu_var(touch_timestamp) = 0;
+       __raw_get_cpu_var(softlockup_touch_ts) = 0;
 }
 EXPORT_SYMBOL(touch_softlockup_watchdog);
 
+void touch_softlockup_watchdog_sync(void)
+{
+       __raw_get_cpu_var(softlock_touch_sync) = true;
+       __raw_get_cpu_var(softlockup_touch_ts) = 0;
+}
+
 void touch_all_softlockup_watchdogs(void)
 {
        int cpu;
 
        /* Cause each CPU to re-update its timestamp rather than complain */
        for_each_online_cpu(cpu)
-               per_cpu(touch_timestamp, cpu) = 0;
+               per_cpu(softlockup_touch_ts, cpu) = 0;
 }
 EXPORT_SYMBOL(touch_all_softlockup_watchdogs);
 
+int proc_dosoftlockup_thresh(struct ctl_table *table, int write,
+                            void __user *buffer,
+                            size_t *lenp, loff_t *ppos)
+{
+       touch_all_softlockup_watchdogs();
+       return proc_dointvec_minmax(table, write, buffer, lenp, ppos);
+}
+
 /*
  * This callback runs from the timer interrupt, and checks
  * whether the watchdog thread has hung or not:
@@ -95,28 +111,36 @@ EXPORT_SYMBOL(touch_all_softlockup_watchdogs);
 void softlockup_tick(void)
 {
        int this_cpu = smp_processor_id();
-       unsigned long touch_timestamp = per_cpu(touch_timestamp, this_cpu);
-       unsigned long print_timestamp;
+       unsigned long touch_ts = per_cpu(softlockup_touch_ts, this_cpu);
+       unsigned long print_ts;
        struct pt_regs *regs = get_irq_regs();
        unsigned long now;
 
        /* Is detection switched off? */
-       if (!per_cpu(watchdog_task, this_cpu) || softlockup_thresh <= 0) {
+       if (!per_cpu(softlockup_watchdog, this_cpu) || softlockup_thresh <= 0) {
                /* Be sure we don't false trigger if switched back on */
-               if (touch_timestamp)
-                       per_cpu(touch_timestamp, this_cpu) = 0;
+               if (touch_ts)
+                       per_cpu(softlockup_touch_ts, this_cpu) = 0;
                return;
        }
 
-       if (touch_timestamp == 0) {
+       if (touch_ts == 0) {
+               if (unlikely(per_cpu(softlock_touch_sync, this_cpu))) {
+                       /*
+                        * If the time stamp was touched atomically
+                        * make sure the scheduler tick is up to date.
+                        */
+                       per_cpu(softlock_touch_sync, this_cpu) = false;
+                       sched_clock_tick();
+               }
                __touch_softlockup_watchdog();
                return;
        }
 
-       print_timestamp = per_cpu(print_timestamp, this_cpu);
+       print_ts = per_cpu(softlockup_print_ts, this_cpu);
 
        /* report at most once a second */
-       if (print_timestamp == touch_timestamp || did_panic)
+       if (print_ts == touch_ts || did_panic)
                return;
 
        /* do not print during early bootup: */
@@ -131,18 +155,18 @@ void softlockup_tick(void)
         * Wake up the high-prio watchdog task twice per
         * threshold timespan.
         */
-       if (now > touch_timestamp + softlockup_thresh/2)
-               wake_up_process(per_cpu(watchdog_task, this_cpu));
+       if (now > touch_ts + softlockup_thresh/2)
+               wake_up_process(per_cpu(softlockup_watchdog, this_cpu));
 
        /* Warn about unreasonable delays: */
-       if (now <= (touch_timestamp + softlockup_thresh))
+       if (now <= (touch_ts + softlockup_thresh))
                return;
 
-       per_cpu(print_timestamp, this_cpu) = touch_timestamp;
+       per_cpu(softlockup_print_ts, this_cpu) = touch_ts;
 
        spin_lock(&print_lock);
        printk(KERN_ERR "BUG: soft lockup - CPU#%d stuck for %lus! [%s:%d]\n",
-                       this_cpu, now - touch_timestamp,
+                       this_cpu, now - touch_ts,
                        current->comm, task_pid_nr(current));
        print_modules();
        print_irqtrace_events(current);
@@ -157,97 +181,11 @@ void softlockup_tick(void)
 }
 
 /*
- * Have a reasonable limit on the number of tasks checked:
- */
-unsigned long __read_mostly sysctl_hung_task_check_count = 1024;
-
-/*
- * Zero means infinite timeout - no checking done:
- */
-unsigned long __read_mostly sysctl_hung_task_timeout_secs = 480;
-
-unsigned long __read_mostly sysctl_hung_task_warnings = 10;
-
-/*
- * Only do the hung-tasks check on one CPU:
- */
-static int check_cpu __read_mostly = -1;
-
-static void check_hung_task(struct task_struct *t, unsigned long now)
-{
-       unsigned long switch_count = t->nvcsw + t->nivcsw;
-
-       if (t->flags & PF_FROZEN)
-               return;
-
-       if (switch_count != t->last_switch_count || !t->last_switch_timestamp) {
-               t->last_switch_count = switch_count;
-               t->last_switch_timestamp = now;
-               return;
-       }
-       if ((long)(now - t->last_switch_timestamp) <
-                                       sysctl_hung_task_timeout_secs)
-               return;
-       if (sysctl_hung_task_warnings < 0)
-               return;
-       sysctl_hung_task_warnings--;
-
-       /*
-        * Ok, the task did not get scheduled for more than 2 minutes,
-        * complain:
-        */
-       printk(KERN_ERR "INFO: task %s:%d blocked for more than "
-                       "%ld seconds.\n", t->comm, t->pid,
-                       sysctl_hung_task_timeout_secs);
-       printk(KERN_ERR "\"echo 0 > /proc/sys/kernel/hung_task_timeout_secs\""
-                       " disables this message.\n");
-       sched_show_task(t);
-       __debug_show_held_locks(t);
-
-       t->last_switch_timestamp = now;
-       touch_nmi_watchdog();
-
-       if (softlockup_panic)
-               panic("softlockup: blocked tasks");
-}
-
-/*
- * Check whether a TASK_UNINTERRUPTIBLE does not get woken up for
- * a really long time (120 seconds). If that happens, print out
- * a warning.
- */
-static void check_hung_uninterruptible_tasks(int this_cpu)
-{
-       int max_count = sysctl_hung_task_check_count;
-       unsigned long now = get_timestamp(this_cpu);
-       struct task_struct *g, *t;
-
-       /*
-        * If the system crashed already then all bets are off,
-        * do not report extra hung tasks:
-        */
-       if (test_taint(TAINT_DIE) || did_panic)
-               return;
-
-       read_lock(&tasklist_lock);
-       do_each_thread(g, t) {
-               if (!--max_count)
-                       goto unlock;
-               /* use "==" to skip the TASK_KILLABLE tasks waiting on NFS */
-               if (t->state == TASK_UNINTERRUPTIBLE)
-                       check_hung_task(t, now);
-       } while_each_thread(g, t);
- unlock:
-       read_unlock(&tasklist_lock);
-}
-
-/*
  * The watchdog thread - runs every second and touches the timestamp.
  */
 static int watchdog(void *__bind_cpu)
 {
        struct sched_param param = { .sched_priority = MAX_RT_PRIO-1 };
-       int this_cpu = (long)__bind_cpu;
 
        sched_setscheduler(current, SCHED_FIFO, &param);
 
@@ -267,11 +205,6 @@ static int watchdog(void *__bind_cpu)
                if (kthread_should_stop())
                        break;
 
-               if (this_cpu == check_cpu) {
-                       if (sysctl_hung_task_timeout_secs)
-                               check_hung_uninterruptible_tasks(this_cpu);
-               }
-
                set_current_state(TASK_INTERRUPTIBLE);
        }
        __set_current_state(TASK_RUNNING);
@@ -291,43 +224,32 @@ cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu)
        switch (action) {
        case CPU_UP_PREPARE:
        case CPU_UP_PREPARE_FROZEN:
-               BUG_ON(per_cpu(watchdog_task, hotcpu));
+               BUG_ON(per_cpu(softlockup_watchdog, hotcpu));
                p = kthread_create(watchdog, hcpu, "watchdog/%d", hotcpu);
                if (IS_ERR(p)) {
                        printk(KERN_ERR "watchdog for %i failed\n", hotcpu);
                        return NOTIFY_BAD;
                }
-               per_cpu(touch_timestamp, hotcpu) = 0;
-               per_cpu(watchdog_task, hotcpu) = p;
+               per_cpu(softlockup_touch_ts, hotcpu) = 0;
+               per_cpu(softlockup_watchdog, hotcpu) = p;
                kthread_bind(p, hotcpu);
                break;
        case CPU_ONLINE:
        case CPU_ONLINE_FROZEN:
-               check_cpu = any_online_cpu(cpu_online_map);
-               wake_up_process(per_cpu(watchdog_task, hotcpu));
+               wake_up_process(per_cpu(softlockup_watchdog, hotcpu));
                break;
 #ifdef CONFIG_HOTPLUG_CPU
-       case CPU_DOWN_PREPARE:
-       case CPU_DOWN_PREPARE_FROZEN:
-               if (hotcpu == check_cpu) {
-                       cpumask_t temp_cpu_online_map = cpu_online_map;
-
-                       cpu_clear(hotcpu, temp_cpu_online_map);
-                       check_cpu = any_online_cpu(temp_cpu_online_map);
-               }
-               break;
-
        case CPU_UP_CANCELED:
        case CPU_UP_CANCELED_FROZEN:
-               if (!per_cpu(watchdog_task, hotcpu))
+               if (!per_cpu(softlockup_watchdog, hotcpu))
                        break;
                /* Unbind so it can run.  Fall thru. */
-               kthread_bind(per_cpu(watchdog_task, hotcpu),
-                            any_online_cpu(cpu_online_map));
+               kthread_bind(per_cpu(softlockup_watchdog, hotcpu),
+                            cpumask_any(cpu_online_mask));
        case CPU_DEAD:
        case CPU_DEAD_FROZEN:
-               p = per_cpu(watchdog_task, hotcpu);
-               per_cpu(watchdog_task, hotcpu) = NULL;
+               p = per_cpu(softlockup_watchdog, hotcpu);
+               per_cpu(softlockup_watchdog, hotcpu) = NULL;
                kthread_stop(p);
                break;
 #endif /* CONFIG_HOTPLUG_CPU */