1 files changed, 412 insertions, 125 deletions

diff --git a/kernel/stop_machine.c b/kernel/stop_machine.c
index 9bb9fb1bd79c..b4e7431e7c78 100644
--- a/kernel/stop_machine.c
+++ b/kernel/stop_machine.c
@@ -1,17 +1,384 @@
-/* Copyright 2008, 2005 Rusty Russell rusty@rustcorp.com.au IBM Corporation.
- * GPL v2 and any later version.
+/*
+ * kernel/stop_machine.c
+ *
+ * Copyright (C) 2008, 2005     IBM Corporation.
+ * Copyright (C) 2008, 2005     Rusty Russell rusty@rustcorp.com.au
+ * Copyright (C) 2010           SUSE Linux Products GmbH
+ * Copyright (C) 2010           Tejun Heo <tj@kernel.org>
+ *
+ * This file is released under the GPLv2 and any later version.
  */
+#include <linux/completion.h>
 #include <linux/cpu.h>
-#include <linux/err.h>
+#include <linux/init.h>
 #include <linux/kthread.h>
 #include <linux/module.h>
+#include <linux/percpu.h>
 #include <linux/sched.h>
 #include <linux/stop_machine.h>
-#include <linux/syscalls.h>
 #include <linux/interrupt.h>
+#include <linux/kallsyms.h>
 
 #include <asm/atomic.h>
-#include <asm/uaccess.h>
+
+/*
+ * Structure to determine completion condition and record errors.  May
+ * be shared by works on different cpus.
+ */
+struct cpu_stop_done {
+        atomic_t                nr_todo;        /* nr left to execute */
+        bool                    executed;       /* actually executed? */
+        int                     ret;            /* collected return value */
+        struct completion       completion;     /* fired if nr_todo reaches 0 */
+};
+
+/* the actual stopper, one per every possible cpu, enabled on online cpus */
+struct cpu_stopper {
+        spinlock_t              lock;
+        struct list_head        works;          /* list of pending works */
+        struct task_struct      *thread;        /* stopper thread */
+        bool                    enabled;        /* is this stopper enabled? */
+};
+
+static DEFINE_PER_CPU(struct cpu_stopper, cpu_stopper);
+
+static void cpu_stop_init_done(struct cpu_stop_done *done, unsigned int nr_todo)
+{
+        memset(done, 0, sizeof(*done));
+        atomic_set(&done->nr_todo, nr_todo);
+        init_completion(&done->completion);
+}
+
+/* signal completion unless @done is NULL */
+static void cpu_stop_signal_done(struct cpu_stop_done *done, bool executed)
+{
+        if (done) {
+                if (executed)
+                        done->executed = true;
+                if (atomic_dec_and_test(&done->nr_todo))
+                        complete(&done->completion);
+        }
+}
+
+/* queue @work to @stopper.  if offline, @work is completed immediately */
+static void cpu_stop_queue_work(struct cpu_stopper *stopper,
+                                struct cpu_stop_work *work)
+{
+        unsigned long flags;
+
+        spin_lock_irqsave(&stopper->lock, flags);
+
+        if (stopper->enabled) {
+                list_add_tail(&work->list, &stopper->works);
+                wake_up_process(stopper->thread);
+        } else
+                cpu_stop_signal_done(work->done, false);
+
+        spin_unlock_irqrestore(&stopper->lock, flags);
+}
+
+/**
+ * stop_one_cpu - stop a cpu
+ * @cpu: cpu to stop
+ * @fn: function to execute
+ * @arg: argument to @fn
+ *
+ * Execute @fn(@arg) on @cpu.  @fn is run in a process context with
+ * the highest priority preempting any task on the cpu and
+ * monopolizing it.  This function returns after the execution is
+ * complete.
+ *
+ * This function doesn't guarantee @cpu stays online till @fn
+ * completes.  If @cpu goes down in the middle, execution may happen
+ * partially or fully on different cpus.  @fn should either be ready
+ * for that or the caller should ensure that @cpu stays online until
+ * this function completes.
+ *
+ * CONTEXT:
+ * Might sleep.
+ *
+ * RETURNS:
+ * -ENOENT if @fn(@arg) was not executed because @cpu was offline;
+ * otherwise, the return value of @fn.
+ */
+int stop_one_cpu(unsigned int cpu, cpu_stop_fn_t fn, void *arg)
+{
+        struct cpu_stop_done done;
+        struct cpu_stop_work work = { .fn = fn, .arg = arg, .done = &done };
+
+        cpu_stop_init_done(&done, 1);
+        cpu_stop_queue_work(&per_cpu(cpu_stopper, cpu), &work);
+        wait_for_completion(&done.completion);
+        return done.executed ? done.ret : -ENOENT;
+}
+
+/**
+ * stop_one_cpu_nowait - stop a cpu but don't wait for completion
+ * @cpu: cpu to stop
+ * @fn: function to execute
+ * @arg: argument to @fn
+ *
+ * Similar to stop_one_cpu() but doesn't wait for completion.  The
+ * caller is responsible for ensuring @work_buf is currently unused
+ * and will remain untouched until stopper starts executing @fn.
+ *
+ * CONTEXT:
+ * Don't care.
+ */
+void stop_one_cpu_nowait(unsigned int cpu, cpu_stop_fn_t fn, void *arg,
+                        struct cpu_stop_work *work_buf)
+{
+        *work_buf = (struct cpu_stop_work){ .fn = fn, .arg = arg, };
+        cpu_stop_queue_work(&per_cpu(cpu_stopper, cpu), work_buf);
+}
+
+/* static data for stop_cpus */
+static DEFINE_MUTEX(stop_cpus_mutex);
+static DEFINE_PER_CPU(struct cpu_stop_work, stop_cpus_work);
+
+int __stop_cpus(const struct cpumask *cpumask, cpu_stop_fn_t fn, void *arg)
+{
+        struct cpu_stop_work *work;
+        struct cpu_stop_done done;
+        unsigned int cpu;
+
+        /* initialize works and done */
+        for_each_cpu(cpu, cpumask) {
+                work = &per_cpu(stop_cpus_work, cpu);
+                work->fn = fn;
+                work->arg = arg;
+                work->done = &done;
+        }
+        cpu_stop_init_done(&done, cpumask_weight(cpumask));
+
+        /*
+         * Disable preemption while queueing to avoid getting
+         * preempted by a stopper which might wait for other stoppers
+         * to enter @fn which can lead to deadlock.
+         */
+        preempt_disable();
+        for_each_cpu(cpu, cpumask)
+                cpu_stop_queue_work(&per_cpu(cpu_stopper, cpu),
+                                    &per_cpu(stop_cpus_work, cpu));
+        preempt_enable();
+
+        wait_for_completion(&done.completion);
+        return done.executed ? done.ret : -ENOENT;
+}
+
+/**
+ * stop_cpus - stop multiple cpus
+ * @cpumask: cpus to stop
+ * @fn: function to execute
+ * @arg: argument to @fn
+ *
+ * Execute @fn(@arg) on online cpus in @cpumask.  On each target cpu,
+ * @fn is run in a process context with the highest priority
+ * preempting any task on the cpu and monopolizing it.  This function
+ * returns after all executions are complete.
+ *
+ * This function doesn't guarantee the cpus in @cpumask stay online
+ * till @fn completes.  If some cpus go down in the middle, execution
+ * on the cpu may happen partially or fully on different cpus.  @fn
+ * should either be ready for that or the caller should ensure that
+ * the cpus stay online until this function completes.
+ *
+ * All stop_cpus() calls are serialized making it safe for @fn to wait
+ * for all cpus to start executing it.
+ *
+ * CONTEXT:
+ * Might sleep.
+ *
+ * RETURNS:
+ * -ENOENT if @fn(@arg) was not executed at all because all cpus in
+ * @cpumask were offline; otherwise, 0 if all executions of @fn
+ * returned 0, any non zero return value if any returned non zero.
+ */
+int stop_cpus(const struct cpumask *cpumask, cpu_stop_fn_t fn, void *arg)
+{
+        int ret;
+
+        /* static works are used, process one request at a time */
+        mutex_lock(&stop_cpus_mutex);
+        ret = __stop_cpus(cpumask, fn, arg);
+        mutex_unlock(&stop_cpus_mutex);
+        return ret;
+}
+
+/**
+ * try_stop_cpus - try to stop multiple cpus
+ * @cpumask: cpus to stop
+ * @fn: function to execute
+ * @arg: argument to @fn
+ *
+ * Identical to stop_cpus() except that it fails with -EAGAIN if
+ * someone else is already using the facility.
+ *
+ * CONTEXT:
+ * Might sleep.
+ *
+ * RETURNS:
+ * -EAGAIN if someone else is already stopping cpus, -ENOENT if
+ * @fn(@arg) was not executed at all because all cpus in @cpumask were
+ * offline; otherwise, 0 if all executions of @fn returned 0, any non
+ * zero return value if any returned non zero.
+ */
+int try_stop_cpus(const struct cpumask *cpumask, cpu_stop_fn_t fn, void *arg)
+{
+        int ret;
+
+        /* static works are used, process one request at a time */
+        if (!mutex_trylock(&stop_cpus_mutex))
+                return -EAGAIN;
+        ret = __stop_cpus(cpumask, fn, arg);
+        mutex_unlock(&stop_cpus_mutex);
+        return ret;
+}
+
+static int cpu_stopper_thread(void *data)
+{
+        struct cpu_stopper *stopper = data;
+        struct cpu_stop_work *work;
+        int ret;
+
+repeat:
+        set_current_state(TASK_INTERRUPTIBLE);  /* mb paired w/ kthread_stop */
+
+        if (kthread_should_stop()) {
+                __set_current_state(TASK_RUNNING);
+                return 0;
+        }
+
+        work = NULL;
+        spin_lock_irq(&stopper->lock);
+        if (!list_empty(&stopper->works)) {
+                work = list_first_entry(&stopper->works,
+                                        struct cpu_stop_work, list);
+                list_del_init(&work->list);
+        }
+        spin_unlock_irq(&stopper->lock);
+
+        if (work) {
+                cpu_stop_fn_t fn = work->fn;
+                void *arg = work->arg;
+                struct cpu_stop_done *done = work->done;
+                char ksym_buf[KSYM_NAME_LEN];
+
+                __set_current_state(TASK_RUNNING);
+
+                /* cpu stop callbacks are not allowed to sleep */
+                preempt_disable();
+
+                ret = fn(arg);
+                if (ret)
+                        done->ret = ret;
+
+                /* restore preemption and check it's still balanced */
+                preempt_enable();
+                WARN_ONCE(preempt_count(),
+                          "cpu_stop: %s(%p) leaked preempt count\n",
+                          kallsyms_lookup((unsigned long)fn, NULL, NULL, NULL,
+                                          ksym_buf), arg);
+
+                cpu_stop_signal_done(done, true);
+        } else
+                schedule();
+
+        goto repeat;
+}
+
+/* manage stopper for a cpu, mostly lifted from sched migration thread mgmt */
+static int __cpuinit cpu_stop_cpu_callback(struct notifier_block *nfb,
+                                           unsigned long action, void *hcpu)
+{
+        struct sched_param param = { .sched_priority = MAX_RT_PRIO - 1 };
+        unsigned int cpu = (unsigned long)hcpu;
+        struct cpu_stopper *stopper = &per_cpu(cpu_stopper, cpu);
+        struct task_struct *p;
+
+        switch (action & ~CPU_TASKS_FROZEN) {
+        case CPU_UP_PREPARE:
+                BUG_ON(stopper->thread || stopper->enabled ||
+                       !list_empty(&stopper->works));
+                p = kthread_create(cpu_stopper_thread, stopper, "migration/%d",
+                                   cpu);
+                if (IS_ERR(p))
+                        return NOTIFY_BAD;
+                sched_setscheduler_nocheck(p, SCHED_FIFO, &param);
+                get_task_struct(p);
+                stopper->thread = p;
+                break;
+
+        case CPU_ONLINE:
+                kthread_bind(stopper->thread, cpu);
+                /* strictly unnecessary, as first user will wake it */
+                wake_up_process(stopper->thread);
+                /* mark enabled */
+                spin_lock_irq(&stopper->lock);
+                stopper->enabled = true;
+                spin_unlock_irq(&stopper->lock);
+                break;
+
+#ifdef CONFIG_HOTPLUG_CPU
+        case CPU_UP_CANCELED:
+        case CPU_DEAD:
+        {
+                struct cpu_stop_work *work;
+
+                /* kill the stopper */
+                kthread_stop(stopper->thread);
+                /* drain remaining works */
+                spin_lock_irq(&stopper->lock);
+                list_for_each_entry(work, &stopper->works, list)
+                        cpu_stop_signal_done(work->done, false);
+                stopper->enabled = false;
+                spin_unlock_irq(&stopper->lock);
+                /* release the stopper */
+                put_task_struct(stopper->thread);
+                stopper->thread = NULL;
+                break;
+        }
+#endif
+        }
+
+        return NOTIFY_OK;
+}
+
+/*
+ * Give it a higher priority so that cpu stopper is available to other
+ * cpu notifiers.  It currently shares the same priority as sched
+ * migration_notifier.
+ */
+static struct notifier_block __cpuinitdata cpu_stop_cpu_notifier = {
+        .notifier_call  = cpu_stop_cpu_callback,
+        .priority       = 10,
+};
+
+static int __init cpu_stop_init(void)
+{
+        void *bcpu = (void *)(long)smp_processor_id();
+        unsigned int cpu;
+        int err;
+
+        for_each_possible_cpu(cpu) {
+                struct cpu_stopper *stopper = &per_cpu(cpu_stopper, cpu);
+
+                spin_lock_init(&stopper->lock);
+                INIT_LIST_HEAD(&stopper->works);
+        }
+
+        /* start one for the boot cpu */
+        err = cpu_stop_cpu_callback(&cpu_stop_cpu_notifier, CPU_UP_PREPARE,
+                                    bcpu);
+        BUG_ON(err == NOTIFY_BAD);
+        cpu_stop_cpu_callback(&cpu_stop_cpu_notifier, CPU_ONLINE, bcpu);
+        register_cpu_notifier(&cpu_stop_cpu_notifier);
+
+        return 0;
+}
+early_initcall(cpu_stop_init);
+
+#ifdef CONFIG_STOP_MACHINE
 
 /* This controls the threads on each CPU. */
 enum stopmachine_state {
@@ -26,174 +393,94 @@ enum stopmachine_state {
         /* Exit */
         STOPMACHINE_EXIT,
 };
-static enum stopmachine_state state;
 
 struct stop_machine_data {
-        int (*fn)(void *);
-        void *data;
-        int fnret;
+        int                     (*fn)(void *);
+        void                    *data;
+        /* Like num_online_cpus(), but hotplug cpu uses us, so we need this. */
+        unsigned int            num_threads;
+        const struct cpumask    *active_cpus;
+
+        enum stopmachine_state  state;
+        atomic_t                thread_ack;
 };
 
-/* Like num_online_cpus(), but hotplug cpu uses us, so we need this. */
-static unsigned int num_threads;
-static atomic_t thread_ack;
-static DEFINE_MUTEX(lock);
-/* setup_lock protects refcount, stop_machine_wq and stop_machine_work. */
-static DEFINE_MUTEX(setup_lock);
-/* Users of stop_machine. */
-static int refcount;
-static struct workqueue_struct *stop_machine_wq;
-static struct stop_machine_data active, idle;
-static const struct cpumask *active_cpus;
-static void __percpu *stop_machine_work;
-
-static void set_state(enum stopmachine_state newstate)
+static void set_state(struct stop_machine_data *smdata,
+                      enum stopmachine_state newstate)
 {
         /* Reset ack counter. */
-        atomic_set(&thread_ack, num_threads);
+        atomic_set(&smdata->thread_ack, smdata->num_threads);
         smp_wmb();
-        state = newstate;
+        smdata->state = newstate;
 }
 
 /* Last one to ack a state moves to the next state. */
-static void ack_state(void)
+static void ack_state(struct stop_machine_data *smdata)
 {
-        if (atomic_dec_and_test(&thread_ack))
-                set_state(state + 1);
+        if (atomic_dec_and_test(&smdata->thread_ack))
+                set_state(smdata, smdata->state + 1);
 }
 
-/* This is the actual function which stops the CPU. It runs
- * in the context of a dedicated stopmachine workqueue. */
-static void stop_cpu(struct work_struct *unused)
+/* This is the cpu_stop function which stops the CPU. */
+static int stop_machine_cpu_stop(void *data)
 {
+        struct stop_machine_data *smdata = data;
         enum stopmachine_state curstate = STOPMACHINE_NONE;
-        struct stop_machine_data *smdata = &idle;
-        int cpu = smp_processor_id();
-        int err;
+        int cpu = smp_processor_id(), err = 0;
+        bool is_active;
+
+        if (!smdata->active_cpus)
+                is_active = cpu == cpumask_first(cpu_online_mask);
+        else
+                is_active = cpumask_test_cpu(cpu, smdata->active_cpus);
 
-        if (!active_cpus) {
-                if (cpu == cpumask_first(cpu_online_mask))
-                        smdata = &active;
-        } else {
-                if (cpumask_test_cpu(cpu, active_cpus))
-                        smdata = &active;
-        }
         /* Simple state machine */
         do {
                 /* Chill out and ensure we re-read stopmachine_state. */
                 cpu_relax();
-                if (state != curstate) {
-                        curstate = state;
+                if (smdata->state != curstate) {
+                        curstate = smdata->state;
                         switch (curstate) {
                         case STOPMACHINE_DISABLE_IRQ:
                                 local_irq_disable();
                                 hard_irq_disable();
                                 break;
                         case STOPMACHINE_RUN:
-                                /* On multiple CPUs only a single error code
-                                 * is needed to tell that something failed. */
-                                err = smdata->fn(smdata->data);
-                                if (err)
-                                        smdata->fnret = err;
+                                if (is_active)
+                                        err = smdata->fn(smdata->data);
                                 break;
                         default:
                                 break;
                         }
-                        ack_state();
+                        ack_state(smdata);
                 }
         } while (curstate != STOPMACHINE_EXIT);
 
         local_irq_enable();
+        return err;
 }
 
-/* Callback for CPUs which aren't supposed to do anything. */
-static int chill(void *unused)
-{
-        return 0;
-}
-
-int stop_machine_create(void)
-{
-        mutex_lock(&setup_lock);
-        if (refcount)
-                goto done;
-        stop_machine_wq = create_rt_workqueue("kstop");
-        if (!stop_machine_wq)
-                goto err_out;
-        stop_machine_work = alloc_percpu(struct work_struct);
-        if (!stop_machine_work)
-                goto err_out;
-done:
-        refcount++;
-        mutex_unlock(&setup_lock);
-        return 0;
-
-err_out:
-        if (stop_machine_wq)
-                destroy_workqueue(stop_machine_wq);
-        mutex_unlock(&setup_lock);
-        return -ENOMEM;
-}
-EXPORT_SYMBOL_GPL(stop_machine_create);
-
-void stop_machine_destroy(void)
-{
-        mutex_lock(&setup_lock);
-        refcount--;
-        if (refcount)
-                goto done;
-        destroy_workqueue(stop_machine_wq);
-        free_percpu(stop_machine_work);
-done:
-        mutex_unlock(&setup_lock);
-}
-EXPORT_SYMBOL_GPL(stop_machine_destroy);
-
 int __stop_machine(int (*fn)(void *), void *data, const struct cpumask *cpus)
 {
-        struct work_struct *sm_work;
-        int i, ret;
-
-        /* Set up initial state. */
-        mutex_lock(&lock);
-        num_threads = num_online_cpus();
-        active_cpus = cpus;
-        active.fn = fn;
-        active.data = data;
-        active.fnret = 0;
-        idle.fn = chill;
-        idle.data = NULL;
-
-        set_state(STOPMACHINE_PREPARE);
-
-        /* Schedule the stop_cpu work on all cpus: hold this CPU so one
-         * doesn't hit this CPU until we're ready. */
-        get_cpu();
-        for_each_online_cpu(i) {
-                sm_work = per_cpu_ptr(stop_machine_work, i);
-                INIT_WORK(sm_work, stop_cpu);
-                queue_work_on(i, stop_machine_wq, sm_work);
-        }
-        /* This will release the thread on our CPU. */
-        put_cpu();
-        flush_workqueue(stop_machine_wq);
-        ret = active.fnret;
-        mutex_unlock(&lock);
-        return ret;
+        struct stop_machine_data smdata = { .fn = fn, .data = data,
+                                            .num_threads = num_online_cpus(),
+                                            .active_cpus = cpus };
+
+        /* Set the initial state and stop all online cpus. */
+        set_state(&smdata, STOPMACHINE_PREPARE);
+        return stop_cpus(cpu_online_mask, stop_machine_cpu_stop, &smdata);
 }
 
 int stop_machine(int (*fn)(void *), void *data, const struct cpumask *cpus)
 {
         int ret;
 
-        ret = stop_machine_create();
-        if (ret)
-                return ret;
         /* No CPUs can come up or down during this. */
         get_online_cpus();
         ret = __stop_machine(fn, data, cpus);
         put_online_cpus();
-        stop_machine_destroy();
         return ret;
 }
 EXPORT_SYMBOL_GPL(stop_machine);
+
+#endif  /* CONFIG_STOP_MACHINE */