diff options
Diffstat (limited to 'kernel')
47 files changed, 4179 insertions, 3570 deletions
diff --git a/kernel/Makefile b/kernel/Makefile index 864ff75d65f2..6aebdeb2aa34 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -100,6 +100,7 @@ obj-$(CONFIG_SLOW_WORK_DEBUG) += slow-work-debugfs.o obj-$(CONFIG_PERF_EVENTS) += perf_event.o obj-$(CONFIG_HAVE_HW_BREAKPOINT) += hw_breakpoint.o obj-$(CONFIG_USER_RETURN_NOTIFIER) += user-return-notifier.o +obj-$(CONFIG_PADATA) += padata.o ifneq ($(CONFIG_SCHED_OMIT_FRAME_POINTER),y) # According to Alan Modra <alan@linuxcare.com.au>, the -fno-omit-frame-pointer is diff --git a/kernel/cgroup.c b/kernel/cgroup.c index aa3bee566446..4fd90e129772 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -23,6 +23,7 @@ */ #include <linux/cgroup.h> +#include <linux/module.h> #include <linux/ctype.h> #include <linux/errno.h> #include <linux/fs.h> @@ -166,6 +167,20 @@ static DEFINE_SPINLOCK(hierarchy_id_lock); */ static int need_forkexit_callback __read_mostly; +#ifdef CONFIG_PROVE_LOCKING +int cgroup_lock_is_held(void) +{ + return lockdep_is_held(&cgroup_mutex); +} +#else /* #ifdef CONFIG_PROVE_LOCKING */ +int cgroup_lock_is_held(void) +{ + return mutex_is_locked(&cgroup_mutex); +} +#endif /* #else #ifdef CONFIG_PROVE_LOCKING */ + +EXPORT_SYMBOL_GPL(cgroup_lock_is_held); + /* convenient tests for these bits */ inline int cgroup_is_removed(const struct cgroup *cgrp) { diff --git a/kernel/exit.c b/kernel/exit.c index 546774a31a66..45ed043b8bf5 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -85,7 +85,9 @@ static void __exit_signal(struct task_struct *tsk) BUG_ON(!sig); BUG_ON(!atomic_read(&sig->count)); - sighand = rcu_dereference(tsk->sighand); + sighand = rcu_dereference_check(tsk->sighand, + rcu_read_lock_held() || + lockdep_is_held(&tasklist_lock)); spin_lock(&sighand->siglock); posix_cpu_timers_exit(tsk); @@ -170,8 +172,10 @@ void release_task(struct task_struct * p) repeat: tracehook_prepare_release_task(p); /* don't need to get the RCU readlock here - the process is dead and - * can't be modifying its own credentials */ + * can't be modifying its own credentials. But shut RCU-lockdep up */ + rcu_read_lock(); atomic_dec(&__task_cred(p)->user->processes); + rcu_read_unlock(); proc_flush_task(p); @@ -473,9 +477,11 @@ static void close_files(struct files_struct * files) /* * It is safe to dereference the fd table without RCU or * ->file_lock because this is the last reference to the - * files structure. + * files structure. But use RCU to shut RCU-lockdep up. */ + rcu_read_lock(); fdt = files_fdtable(files); + rcu_read_unlock(); for (;;) { unsigned long set; i = j * __NFDBITS; @@ -521,10 +527,12 @@ void put_files_struct(struct files_struct *files) * at the end of the RCU grace period. Otherwise, * you can free files immediately. */ + rcu_read_lock(); fdt = files_fdtable(files); if (fdt != &files->fdtab) kmem_cache_free(files_cachep, files); free_fdtable(fdt); + rcu_read_unlock(); } } diff --git a/kernel/fork.c b/kernel/fork.c index f88bd984df35..17bbf093356d 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -86,6 +86,7 @@ int max_threads; /* tunable limit on nr_threads */ DEFINE_PER_CPU(unsigned long, process_counts) = 0; __cacheline_aligned DEFINE_RWLOCK(tasklist_lock); /* outer */ +EXPORT_SYMBOL_GPL(tasklist_lock); int nr_processes(void) { diff --git a/kernel/kprobes.c b/kernel/kprobes.c index b7df302a0204..ccec774c716d 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -44,6 +44,7 @@ #include <linux/debugfs.h> #include <linux/kdebug.h> #include <linux/memory.h> +#include <linux/ftrace.h> #include <asm-generic/sections.h> #include <asm/cacheflush.h> @@ -93,6 +94,7 @@ static struct kprobe_blackpoint kprobe_blacklist[] = { {"native_get_debugreg",}, {"irq_entries_start",}, {"common_interrupt",}, + {"mcount",}, /* mcount can be called from everywhere */ {NULL} /* Terminator */ }; @@ -124,30 +126,6 @@ static LIST_HEAD(kprobe_insn_pages); static int kprobe_garbage_slots; static int collect_garbage_slots(void); -static int __kprobes check_safety(void) -{ - int ret = 0; -#if defined(CONFIG_PREEMPT) && defined(CONFIG_FREEZER) - ret = freeze_processes(); - if (ret == 0) { - struct task_struct *p, *q; - do_each_thread(p, q) { - if (p != current && p->state == TASK_RUNNING && - p->pid != 0) { - printk("Check failed: %s is running\n",p->comm); - ret = -1; - goto loop_end; - } - } while_each_thread(p, q); - } -loop_end: - thaw_processes(); -#else - synchronize_sched(); -#endif - return ret; -} - /** * __get_insn_slot() - Find a slot on an executable page for an instruction. * We allocate an executable page if there's no room on existing ones. @@ -235,9 +213,8 @@ static int __kprobes collect_garbage_slots(void) { struct kprobe_insn_page *kip, *next; - /* Ensure no-one is preepmted on the garbages */ - if (check_safety()) - return -EAGAIN; + /* Ensure no-one is interrupted on the garbages */ + synchronize_sched(); list_for_each_entry_safe(kip, next, &kprobe_insn_pages, list) { int i; @@ -728,7 +705,8 @@ int __kprobes register_kprobe(struct kprobe *p) preempt_disable(); if (!kernel_text_address((unsigned long) p->addr) || - in_kprobes_functions((unsigned long) p->addr)) { + in_kprobes_functions((unsigned long) p->addr) || + ftrace_text_reserved(p->addr, p->addr)) { preempt_enable(); return -EINVAL; } diff --git a/kernel/ksysfs.c b/kernel/ksysfs.c index 3feaf5a74514..6b1ccc3f0205 100644 --- a/kernel/ksysfs.c +++ b/kernel/ksysfs.c @@ -197,16 +197,8 @@ static int __init ksysfs_init(void) goto group_exit; } - /* create the /sys/kernel/uids/ directory */ - error = uids_sysfs_init(); - if (error) - goto notes_exit; - return 0; -notes_exit: - if (notes_size > 0) - sysfs_remove_bin_file(kernel_kobj, ¬es_attr); group_exit: sysfs_remove_group(kernel_kobj, &kernel_attr_group); kset_exit: diff --git a/kernel/kthread.c b/kernel/kthread.c index fbb6222fe7e0..82ed0ea15194 100644 --- a/kernel/kthread.c +++ b/kernel/kthread.c @@ -101,7 +101,7 @@ static void create_kthread(struct kthread_create_info *create) * * Description: This helper function creates and names a kernel * thread. The thread will be stopped: use wake_up_process() to start - * it. See also kthread_run(), kthread_create_on_cpu(). + * it. See also kthread_run(). * * When woken, the thread will run @threadfn() with @data as its * argument. @threadfn() can either call do_exit() directly if it is a diff --git a/kernel/lockdep.c b/kernel/lockdep.c index c62ec14609b9..0c30d0455de1 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -3809,3 +3809,21 @@ void lockdep_sys_exit(void) lockdep_print_held_locks(curr); } } + +void lockdep_rcu_dereference(const char *file, const int line) +{ + struct task_struct *curr = current; + + if (!debug_locks_off()) + return; + printk("\n===================================================\n"); + printk( "[ INFO: suspicious rcu_dereference_check() usage. ]\n"); + printk( "---------------------------------------------------\n"); + printk("%s:%d invoked rcu_dereference_check() without protection!\n", + file, line); + printk("\nother info that might help us debug this:\n\n"); + lockdep_print_held_locks(curr); + printk("\nstack backtrace:\n"); + dump_stack(); +} +EXPORT_SYMBOL_GPL(lockdep_rcu_dereference); diff --git a/kernel/notifier.c b/kernel/notifier.c index acd24e7643eb..2488ba7eb568 100644 --- a/kernel/notifier.c +++ b/kernel/notifier.c @@ -78,10 +78,10 @@ static int __kprobes notifier_call_chain(struct notifier_block **nl, int ret = NOTIFY_DONE; struct notifier_block *nb, *next_nb; - nb = rcu_dereference(*nl); + nb = rcu_dereference_raw(*nl); while (nb && nr_to_call) { - next_nb = rcu_dereference(nb->next); + next_nb = rcu_dereference_raw(nb->next); #ifdef CONFIG_DEBUG_NOTIFIERS if (unlikely(!func_ptr_is_kernel_text(nb->notifier_call))) { @@ -309,7 +309,7 @@ int __blocking_notifier_call_chain(struct blocking_notifier_head *nh, * racy then it does not matter what the result of the test * is, we re-check the list after having taken the lock anyway: */ - if (rcu_dereference(nh->head)) { + if (rcu_dereference_raw(nh->head)) { down_read(&nh->rwsem); ret = notifier_call_chain(&nh->head, val, v, nr_to_call, nr_calls); diff --git a/kernel/padata.c b/kernel/padata.c new file mode 100644 index 000000000000..6f9bcb8313d6 --- /dev/null +++ b/kernel/padata.c @@ -0,0 +1,690 @@ +/* + * padata.c - generic interface to process data streams in parallel + * + * Copyright (C) 2008, 2009 secunet Security Networks AG + * Copyright (C) 2008, 2009 Steffen Klassert <steffen.klassert@secunet.com> + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include <linux/module.h> +#include <linux/cpumask.h> +#include <linux/err.h> +#include <linux/cpu.h> +#include <linux/padata.h> +#include <linux/mutex.h> +#include <linux/sched.h> +#include <linux/rcupdate.h> + +#define MAX_SEQ_NR INT_MAX - NR_CPUS +#define MAX_OBJ_NUM 10000 * NR_CPUS + +static int padata_index_to_cpu(struct parallel_data *pd, int cpu_index) +{ + int cpu, target_cpu; + + target_cpu = cpumask_first(pd->cpumask); + for (cpu = 0; cpu < cpu_index; cpu++) + target_cpu = cpumask_next(target_cpu, pd->cpumask); + + return target_cpu; +} + +static int padata_cpu_hash(struct padata_priv *padata) +{ + int cpu_index; + struct parallel_data *pd; + + pd = padata->pd; + + /* + * Hash the sequence numbers to the cpus by taking + * seq_nr mod. number of cpus in use. + */ + cpu_index = padata->seq_nr % cpumask_weight(pd->cpumask); + + return padata_index_to_cpu(pd, cpu_index); +} + +static void padata_parallel_worker(struct work_struct *work) +{ + struct padata_queue *queue; + struct parallel_data *pd; + struct padata_instance *pinst; + LIST_HEAD(local_list); + + local_bh_disable(); + queue = container_of(work, struct padata_queue, pwork); + pd = queue->pd; + pinst = pd->pinst; + + spin_lock(&queue->parallel.lock); + list_replace_init(&queue->parallel.list, &local_list); + spin_unlock(&queue->parallel.lock); + + while (!list_empty(&local_list)) { + struct padata_priv *padata; + + padata = list_entry(local_list.next, + struct padata_priv, list); + + list_del_init(&padata->list); + + padata->parallel(padata); + } + + local_bh_enable(); +} + +/* + * padata_do_parallel - padata parallelization function + * + * @pinst: padata instance + * @padata: object to be parallelized + * @cb_cpu: cpu the serialization callback function will run on, + * must be in the cpumask of padata. + * + * The parallelization callback function will run with BHs off. + * Note: Every object which is parallelized by padata_do_parallel + * must be seen by padata_do_serial. + */ +int padata_do_parallel(struct padata_instance *pinst, + struct padata_priv *padata, int cb_cpu) +{ + int target_cpu, err; + struct padata_queue *queue; + struct parallel_data *pd; + + rcu_read_lock_bh(); + + pd = rcu_dereference(pinst->pd); + + err = 0; + if (!(pinst->flags & PADATA_INIT)) + goto out; + + err = -EBUSY; + if ((pinst->flags & PADATA_RESET)) + goto out; + + if (atomic_read(&pd->refcnt) >= MAX_OBJ_NUM) + goto out; + + err = -EINVAL; + if (!cpumask_test_cpu(cb_cpu, pd->cpumask)) + goto out; + + err = -EINPROGRESS; + atomic_inc(&pd->refcnt); + padata->pd = pd; + padata->cb_cpu = cb_cpu; + + if (unlikely(atomic_read(&pd->seq_nr) == pd->max_seq_nr)) + atomic_set(&pd->seq_nr, -1); + + padata->seq_nr = atomic_inc_return(&pd->seq_nr); + + target_cpu = padata_cpu_hash(padata); + queue = per_cpu_ptr(pd->queue, target_cpu); + + spin_lock(&queue->parallel.lock); + list_add_tail(&padata->list, &queue->parallel.list); + spin_unlock(&queue->parallel.lock); + + queue_work_on(target_cpu, pinst->wq, &queue->pwork); + +out: + rcu_read_unlock_bh(); + + return err; +} +EXPORT_SYMBOL(padata_do_parallel); + +static struct padata_priv *padata_get_next(struct parallel_data *pd) +{ + int cpu, num_cpus, empty, calc_seq_nr; + int seq_nr, next_nr, overrun, next_overrun; + struct padata_queue *queue, *next_queue; + struct padata_priv *padata; + struct padata_list *reorder; + + empty = 0; + next_nr = -1; + next_overrun = 0; + next_queue = NULL; + + num_cpus = cpumask_weight(pd->cpumask); + + for_each_cpu(cpu, pd->cpumask) { + queue = per_cpu_ptr(pd->queue, cpu); + reorder = &queue->reorder; + + /* + * Calculate the seq_nr of the object that should be + * next in this queue. + */ + overrun = 0; + calc_seq_nr = (atomic_read(&queue->num_obj) * num_cpus) + + queue->cpu_index; + + if (unlikely(calc_seq_nr > pd->max_seq_nr)) { + calc_seq_nr = calc_seq_nr - pd->max_seq_nr - 1; + overrun = 1; + } + + if (!list_empty(&reorder->list)) { + padata = list_entry(reorder->list.next, + struct padata_priv, list); + + seq_nr = padata->seq_nr; + BUG_ON(calc_seq_nr != seq_nr); + } else { + seq_nr = calc_seq_nr; + empty++; + } + + if (next_nr < 0 || seq_nr < next_nr + || (next_overrun && !overrun)) { + next_nr = seq_nr; + next_overrun = overrun; + next_queue = queue; + } + } + + padata = NULL; + + if (empty == num_cpus) + goto out; + + reorder = &next_queue->reorder; + + if (!list_empty(&reorder->list)) { + padata = list_entry(reorder->list.next, + struct padata_priv, list); + + if (unlikely(next_overrun)) { + for_each_cpu(cpu, pd->cpumask) { + queue = per_cpu_ptr(pd->queue, cpu); + atomic_set(&queue->num_obj, 0); + } + } + + spin_lock(&reorder->lock); + list_del_init(&padata->list); + atomic_dec(&pd->reorder_objects); + spin_unlock(&reorder->lock); + + atomic_inc(&next_queue->num_obj); + + goto out; + } + + if (next_nr % num_cpus == next_queue->cpu_index) { + padata = ERR_PTR(-ENODATA); + goto out; + } + + padata = ERR_PTR(-EINPROGRESS); +out: + return padata; +} + +static void padata_reorder(struct parallel_data *pd) +{ + struct padata_priv *padata; + struct padata_queue *queue; + struct padata_instance *pinst = pd->pinst; + +try_again: + if (!spin_trylock_bh(&pd->lock)) + goto out; + + while (1) { + padata = padata_get_next(pd); + + if (!padata || PTR_ERR(padata) == -EINPROGRESS) + break; + + if (PTR_ERR(padata) == -ENODATA) { + spin_unlock_bh(&pd->lock); + goto out; + } + + queue = per_cpu_ptr(pd->queue, padata->cb_cpu); + + spin_lock(&queue->serial.lock); + list_add_tail(&padata->list, &queue->serial.list); + spin_unlock(&queue->serial.lock); + + queue_work_on(padata->cb_cpu, pinst->wq, &queue->swork); + } + + spin_unlock_bh(&pd->lock); + + if (atomic_read(&pd->reorder_objects)) + goto try_again; + +out: + return; +} + +static void padata_serial_worker(struct work_struct *work) +{ + struct padata_queue *queue; + struct parallel_data *pd; + LIST_HEAD(local_list); + + local_bh_disable(); + queue = container_of(work, struct padata_queue, swork); + pd = queue->pd; + + spin_lock(&queue->serial.lock); + list_replace_init(&queue->serial.list, &local_list); + spin_unlock(&queue->serial.lock); + + while (!list_empty(&local_list)) { + struct padata_priv *padata; + + padata = list_entry(local_list.next, + struct padata_priv, list); + + list_del_init(&padata->list); + + padata->serial(padata); + atomic_dec(&pd->refcnt); + } + local_bh_enable(); +} + +/* + * padata_do_serial - padata serialization function + * + * @padata: object to be serialized. + * + * padata_do_serial must be called for every parallelized object. + * The serialization callback function will run with BHs off. + */ +void padata_do_serial(struct padata_priv *padata) +{ + int cpu; + struct padata_queue *queue; + struct parallel_data *pd; + + pd = padata->pd; + + cpu = get_cpu(); + queue = per_cpu_ptr(pd->queue, cpu); + + spin_lock(&queue->reorder.lock); + atomic_inc(&pd->reorder_objects); + list_add_tail(&padata->list, &queue->reorder.list); + spin_unlock(&queue->reorder.lock); + + put_cpu(); + + padata_reorder(pd); +} +EXPORT_SYMBOL(padata_do_serial); + +static struct parallel_data *padata_alloc_pd(struct padata_instance *pinst, + const struct cpumask *cpumask) +{ + int cpu, cpu_index, num_cpus; + struct padata_queue *queue; + struct parallel_data *pd; + + cpu_index = 0; + + pd = kzalloc(sizeof(struct parallel_data), GFP_KERNEL); + if (!pd) + goto err; + + pd->queue = alloc_percpu(struct padata_queue); + if (!pd->queue) + goto err_free_pd; + + if (!alloc_cpumask_var(&pd->cpumask, GFP_KERNEL)) + goto err_free_queue; + + for_each_possible_cpu(cpu) { + queue = per_cpu_ptr(pd->queue, cpu); + + queue->pd = pd; + + if (cpumask_test_cpu(cpu, cpumask) + && cpumask_test_cpu(cpu, cpu_active_mask)) { + queue->cpu_index = cpu_index; + cpu_index++; + } else + queue->cpu_index = -1; + + INIT_LIST_HEAD(&queue->reorder.list); + INIT_LIST_HEAD(&queue->parallel.list); + INIT_LIST_HEAD(&queue->serial.list); + spin_lock_init(&queue->reorder.lock); + spin_lock_init(&queue->parallel.lock); + spin_lock_init(&queue->serial.lock); + + INIT_WORK(&queue->pwork, padata_parallel_worker); + INIT_WORK(&queue->swork, padata_serial_worker); + atomic_set(&queue->num_obj, 0); + } + + cpumask_and(pd->cpumask, cpumask, cpu_active_mask); + + num_cpus = cpumask_weight(pd->cpumask); + pd->max_seq_nr = (MAX_SEQ_NR / num_cpus) * num_cpus - 1; + + atomic_set(&pd->seq_nr, -1); + atomic_set(&pd->reorder_objects, 0); + atomic_set(&pd->refcnt, 0); + pd->pinst = pinst; + spin_lock_init(&pd->lock); + + return pd; + +err_free_queue: + free_percpu(pd->queue); +err_free_pd: + kfree(pd); +err: + return NULL; +} + +static void padata_free_pd(struct parallel_data *pd) +{ + free_cpumask_var(pd->cpumask); + free_percpu(pd->queue); + kfree(pd); +} + +static void padata_replace(struct padata_instance *pinst, + struct parallel_data *pd_new) +{ + struct parallel_data *pd_old = pinst->pd; + + pinst->flags |= PADATA_RESET; + + rcu_assign_pointer(pinst->pd, pd_new); + + synchronize_rcu(); + + while (atomic_read(&pd_old->refcnt) != 0) + yield(); + + flush_workqueue(pinst->wq); + + padata_free_pd(pd_old); + + pinst->flags &= ~PADATA_RESET; +} + +/* + * padata_set_cpumask - set the cpumask that padata should use + * + * @pinst: padata instance + * @cpumask: the cpumask to use + */ +int padata_set_cpumask(struct padata_instance *pinst, + cpumask_var_t cpumask) +{ + struct parallel_data *pd; + int err = 0; + + might_sleep(); + + mutex_lock(&pinst->lock); + + pd = padata_alloc_pd(pinst, cpumask); + if (!pd) { + err = -ENOMEM; + goto out; + } + + cpumask_copy(pinst->cpumask, cpumask); + + padata_replace(pinst, pd); + +out: + mutex_unlock(&pinst->lock); + + return err; +} +EXPORT_SYMBOL(padata_set_cpumask); + +static int __padata_add_cpu(struct padata_instance *pinst, int cpu) +{ + struct parallel_data *pd; + + if (cpumask_test_cpu(cpu, cpu_active_mask)) { + pd = padata_alloc_pd(pinst, pinst->cpumask); + if (!pd) + return -ENOMEM; + + padata_replace(pinst, pd); + } + + return 0; +} + +/* + * padata_add_cpu - add a cpu to the padata cpumask + * + * @pinst: padata instance + * @cpu: cpu to add + */ +int padata_add_cpu(struct padata_instance *pinst, int cpu) +{ + int err; + + might_sleep(); + + mutex_lock(&pinst->lock); + + cpumask_set_cpu(cpu, pinst->cpumask); + err = __padata_add_cpu(pinst, cpu); + + mutex_unlock(&pinst->lock); + + return err; +} +EXPORT_SYMBOL(padata_add_cpu); + +static int __padata_remove_cpu(struct padata_instance *pinst, int cpu) +{ + struct parallel_data *pd; + + if (cpumask_test_cpu(cpu, cpu_online_mask)) { + pd = padata_alloc_pd(pinst, pinst->cpumask); + if (!pd) + return -ENOMEM; + + padata_replace(pinst, pd); + } + + return 0; +} + +/* + * padata_remove_cpu - remove a cpu from the padata cpumask + * + * @pinst: padata instance + * @cpu: cpu to remove + */ +int padata_remove_cpu(struct padata_instance *pinst, int cpu) +{ + int err; + + might_sleep(); + + mutex_lock(&pinst->lock); + + cpumask_clear_cpu(cpu, pinst->cpumask); + err = __padata_remove_cpu(pinst, cpu); + + mutex_unlock(&pinst->lock); + + return err; +} +EXPORT_SYMBOL(padata_remove_cpu); + +/* + * padata_start - start the parallel processing + * + * @pinst: padata instance to start + */ +void padata_start(struct padata_instance *pinst) +{ + might_sleep(); + + mutex_lock(&pinst->lock); + pinst->flags |= PADATA_INIT; + mutex_unlock(&pinst->lock); +} +EXPORT_SYMBOL(padata_start); + +/* + * padata_stop - stop the parallel processing + * + * @pinst: padata instance to stop + */ +void padata_stop(struct padata_instance *pinst) +{ + might_sleep(); + + mutex_lock(&pinst->lock); + pinst->flags &= ~PADATA_INIT; + mutex_unlock(&pinst->lock); +} +EXPORT_SYMBOL(padata_stop); + +static int __cpuinit padata_cpu_callback(struct notifier_block *nfb, + unsigned long action, void *hcpu) +{ + int err; + struct padata_instance *pinst; + int cpu = (unsigned long)hcpu; + + pinst = container_of(nfb, struct padata_instance, cpu_notifier); + + switch (action) { + case CPU_ONLINE: + case CPU_ONLINE_FROZEN: + if (!cpumask_test_cpu(cpu, pinst->cpumask)) + break; + mutex_lock(&pinst->lock); + err = __padata_add_cpu(pinst, cpu); + mutex_unlock(&pinst->lock); + if (err) + return NOTIFY_BAD; + break; + + case CPU_DOWN_PREPARE: + case CPU_DOWN_PREPARE_FROZEN: + if (!cpumask_test_cpu(cpu, pinst->cpumask)) + break; + mutex_lock(&pinst->lock); + err = __padata_remove_cpu(pinst, cpu); + mutex_unlock(&pinst->lock); + if (err) + return NOTIFY_BAD; + break; + + case CPU_UP_CANCELED: + case CPU_UP_CANCELED_FROZEN: + if (!cpumask_test_cpu(cpu, pinst->cpumask)) + break; + mutex_lock(&pinst->lock); + __padata_remove_cpu(pinst, cpu); + mutex_unlock(&pinst->lock); + + case CPU_DOWN_FAILED: + case CPU_DOWN_FAILED_FROZEN: + if (!cpumask_test_cpu(cpu, pinst->cpumask)) + break; + mutex_lock(&pinst->lock); + __padata_add_cpu(pinst, cpu); + mutex_unlock(&pinst->lock); + } + + return NOTIFY_OK; +} + +/* + * padata_alloc - allocate and initialize a padata instance + * + * @cpumask: cpumask that padata uses for parallelization + * @wq: workqueue to use for the allocated padata instance + */ +struct padata_instance *padata_alloc(const struct cpumask *cpumask, + struct workqueue_struct *wq) +{ + int err; + struct padata_instance *pinst; + struct parallel_data *pd; + + pinst = kzalloc(sizeof(struct padata_instance), GFP_KERNEL); + if (!pinst) + goto err; + + pd = padata_alloc_pd(pinst, cpumask); + if (!pd) + goto err_free_inst; + + rcu_assign_pointer(pinst->pd, pd); + + pinst->wq = wq; + + cpumask_copy(pinst->cpumask, cpumask); + + pinst->flags = 0; + + pinst->cpu_notifier.notifier_call = padata_cpu_callback; + pinst->cpu_notifier.priority = 0; + err = register_hotcpu_notifier(&pinst->cpu_notifier); + if (err) + goto err_free_pd; + + mutex_init(&pinst->lock); + + return pinst; + +err_free_pd: + padata_free_pd(pd); +err_free_inst: + kfree(pinst); +err: + return NULL; +} +EXPORT_SYMBOL(padata_alloc); + +/* + * padata_free - free a padata instance + * + * @ padata_inst: padata instance to free + */ +void padata_free(struct padata_instance *pinst) +{ + padata_stop(pinst); + + synchronize_rcu(); + + while (atomic_read(&pinst->pd->refcnt) != 0) + yield(); + + unregister_hotcpu_notifier(&pinst->cpu_notifier); + padata_free_pd(pinst->pd); + kfree(pinst); +} +EXPORT_SYMBOL(padata_free); diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 2b19297742cb..a661e7991865 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -98,11 +98,12 @@ void __weak hw_perf_enable(void) { barrier(); } void __weak hw_perf_event_setup(int cpu) { barrier(); } void __weak hw_perf_event_setup_online(int cpu) { barrier(); } +void __weak hw_perf_event_setup_offline(int cpu) { barrier(); } int __weak hw_perf_group_sched_in(struct perf_event *group_leader, struct perf_cpu_context *cpuctx, - struct perf_event_context *ctx, int cpu) + struct perf_event_context *ctx) { return 0; } @@ -248,7 +249,7 @@ static void perf_unpin_context(struct perf_event_context *ctx) static inline u64 perf_clock(void) { - return cpu_clock(smp_processor_id()); + return cpu_clock(raw_smp_processor_id()); } /* @@ -289,6 +290,15 @@ static void update_event_times(struct perf_event *event) event->total_time_running = run_end - event->tstamp_running; } +static struct list_head * +ctx_group_list(struct perf_event *event, struct perf_event_context *ctx) +{ + if (event->attr.pinned) + return &ctx->pinned_groups; + else + return &ctx->flexible_groups; +} + /* * Add a event from the lists for its context. * Must be called with ctx->mutex and ctx->lock held. @@ -303,9 +313,19 @@ list_add_event(struct perf_event *event, struct perf_event_context *ctx) * add it straight to the context's event list, or to the group * leader's sibling list: */ - if (group_leader == event) - list_add_tail(&event->group_entry, &ctx->group_list); - else { + if (group_leader == event) { + struct list_head *list; + + if (is_software_event(event)) + event->group_flags |= PERF_GROUP_SOFTWARE; + + list = ctx_group_list(event, ctx); + list_add_tail(&event->group_entry, list); + } else { + if (group_leader->group_flags & PERF_GROUP_SOFTWARE && + !is_software_event(event)) + group_leader->group_flags &= ~PERF_GROUP_SOFTWARE; + list_add_tail(&event->group_entry, &group_leader->sibling_list); group_leader->nr_siblings++; } @@ -355,9 +375,14 @@ list_del_event(struct perf_event *event, struct perf_event_context *ctx) * to the context list directly: */ list_for_each_entry_safe(sibling, tmp, &event->sibling_list, group_entry) { + struct list_head *list; - list_move_tail(&sibling->group_entry, &ctx->group_list); + list = ctx_group_list(event, ctx); + list_move_tail(&sibling->group_entry, list); sibling->group_leader = sibling; + + /* Inherit group flags from the previous leader */ + sibling->group_flags = event->group_flags; } } @@ -608,14 +633,13 @@ void perf_event_disable(struct perf_event *event) static int event_sched_in(struct perf_event *event, struct perf_cpu_context *cpuctx, - struct perf_event_context *ctx, - int cpu) + struct perf_event_context *ctx) { if (event->state <= PERF_EVENT_STATE_OFF) return 0; event->state = PERF_EVENT_STATE_ACTIVE; - event->oncpu = cpu; /* TODO: put 'cpu' into cpuctx->cpu */ + event->oncpu = smp_processor_id(); /* * The new state must be visible before we turn it on in the hardware: */ @@ -642,8 +666,7 @@ event_sched_in(struct perf_event *event, static int group_sched_in(struct perf_event *group_event, struct perf_cpu_context *cpuctx, - struct perf_event_context *ctx, - int cpu) + struct perf_event_context *ctx) { struct perf_event *event, *partial_group; int ret; @@ -651,18 +674,18 @@ group_sched_in(struct perf_event *group_event, if (group_event->state == PERF_EVENT_STATE_OFF) return 0; - ret = hw_perf_group_sched_in(group_event, cpuctx, ctx, cpu); + ret = hw_perf_group_sched_in(group_event, cpuctx, ctx); if (ret) return ret < 0 ? ret : 0; - if (event_sched_in(group_event, cpuctx, ctx, cpu)) + if (event_sched_in(group_event, cpuctx, ctx)) return -EAGAIN; /* * Schedule in siblings as one group (if any): */ list_for_each_entry(event, &group_event->sibling_list, group_entry) { - if (event_sched_in(event, cpuctx, ctx, cpu)) { + if (event_sched_in(event, cpuctx, ctx)) { partial_group = event; goto group_error; } @@ -686,24 +709,6 @@ group_error: } /* - * Return 1 for a group consisting entirely of software events, - * 0 if the group contains any hardware events. - */ -static int is_software_only_group(struct perf_event *leader) -{ - struct perf_event *event; - - if (!is_software_event(leader)) - return 0; - - list_for_each_entry(event, &leader->sibling_list, group_entry) - if (!is_software_event(event)) - return 0; - - return 1; -} - -/* * Work out whether we can put this event group on the CPU now. */ static int group_can_go_on(struct perf_event *event, @@ -713,7 +718,7 @@ static int group_can_go_on(struct perf_event *event, /* * Groups consisting entirely of software events can always go on. */ - if (is_software_only_group(event)) + if (event->group_flags & PERF_GROUP_SOFTWARE) return 1; /* * If an exclusive group is already on, no other hardware @@ -754,7 +759,6 @@ static void __perf_install_in_context(void *info) struct perf_event *event = info; struct perf_event_context *ctx = event->ctx; struct perf_event *leader = event->group_leader; - int cpu = smp_processor_id(); int err; /* @@ -801,7 +805,7 @@ static void __perf_install_in_context(void *info) if (!group_can_go_on(event, cpuctx, 1)) err = -EEXIST; else - err = event_sched_in(event, cpuctx, ctx, cpu); + err = event_sched_in(event, cpuctx, ctx); if (err) { /* @@ -943,11 +947,9 @@ static void __perf_event_enable(void *info) } else { perf_disable(); if (event == leader) - err = group_sched_in(event, cpuctx, ctx, - smp_processor_id()); + err = group_sched_in(event, cpuctx, ctx); else - err = event_sched_in(event, cpuctx, ctx, - smp_processor_id()); + err = event_sched_in(event, cpuctx, ctx); perf_enable(); } @@ -1043,8 +1045,15 @@ static int perf_event_refresh(struct perf_event *event, int refresh) return 0; } -void __perf_event_sched_out(struct perf_event_context *ctx, - struct perf_cpu_context *cpuctx) +enum event_type_t { + EVENT_FLEXIBLE = 0x1, + EVENT_PINNED = 0x2, + EVENT_ALL = EVENT_FLEXIBLE | EVENT_PINNED, +}; + +static void ctx_sched_out(struct perf_event_context *ctx, + struct perf_cpu_context *cpuctx, + enum event_type_t event_type) { struct perf_event *event; @@ -1055,10 +1064,18 @@ void __perf_event_sched_out(struct perf_event_context *ctx, update_context_time(ctx); perf_disable(); - if (ctx->nr_active) { - list_for_each_entry(event, &ctx->group_list, group_entry) + if (!ctx->nr_active) + goto out_enable; + + if (event_type & EVENT_PINNED) + list_for_each_entry(event, &ctx->pinned_groups, group_entry) group_sched_out(event, cpuctx, ctx); - } + + if (event_type & EVENT_FLEXIBLE) + list_for_each_entry(event, &ctx->flexible_groups, group_entry) + group_sched_out(event, cpuctx, ctx); + + out_enable: perf_enable(); out: raw_spin_unlock(&ctx->lock); @@ -1170,9 +1187,9 @@ static void perf_event_sync_stat(struct perf_event_context *ctx, * not restart the event. */ void perf_event_task_sched_out(struct task_struct *task, - struct task_struct *next, int cpu) + struct task_struct *next) { - struct perf_cpu_context *cpuctx = &per_cpu(perf_cpu_context, cpu); + struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); struct perf_event_context *ctx = task->perf_event_ctxp; struct perf_event_context *next_ctx; struct perf_event_context *parent; @@ -1220,15 +1237,13 @@ void perf_event_task_sched_out(struct task_struct *task, rcu_read_unlock(); if (do_switch) { - __perf_event_sched_out(ctx, cpuctx); + ctx_sched_out(ctx, cpuctx, EVENT_ALL); cpuctx->task_ctx = NULL; } } -/* - * Called with IRQs disabled - */ -static void __perf_event_task_sched_out(struct perf_event_context *ctx) +static void task_ctx_sched_out(struct perf_event_context *ctx, + enum event_type_t event_type) { struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); @@ -1238,47 +1253,41 @@ static void __perf_event_task_sched_out(struct perf_event_context *ctx) if (WARN_ON_ONCE(ctx != cpuctx->task_ctx)) return; - __perf_event_sched_out(ctx, cpuctx); + ctx_sched_out(ctx, cpuctx, event_type); cpuctx->task_ctx = NULL; } /* * Called with IRQs disabled */ -static void perf_event_cpu_sched_out(struct perf_cpu_context *cpuctx) +static void __perf_event_task_sched_out(struct perf_event_context *ctx) +{ + task_ctx_sched_out(ctx, EVENT_ALL); +} + +/* + * Called with IRQs disabled + */ +static void cpu_ctx_sched_out(struct perf_cpu_context *cpuctx, + enum event_type_t event_type) { - __perf_event_sched_out(&cpuctx->ctx, cpuctx); + ctx_sched_out(&cpuctx->ctx, cpuctx, event_type); } static void -__perf_event_sched_in(struct perf_event_context *ctx, - struct perf_cpu_context *cpuctx, int cpu) +ctx_pinned_sched_in(struct perf_event_context *ctx, + struct perf_cpu_context *cpuctx) { struct perf_event *event; - int can_add_hw = 1; - - raw_spin_lock(&ctx->lock); - ctx->is_active = 1; - if (likely(!ctx->nr_events)) - goto out; - - ctx->timestamp = perf_clock(); - - perf_disable(); - /* - * First go through the list and put on any pinned groups - * in order to give them the best chance of going on. - */ - list_for_each_entry(event, &ctx->group_list, group_entry) { - if (event->state <= PERF_EVENT_STATE_OFF || - !event->attr.pinned) + list_for_each_entry(event, &ctx->pinned_groups, group_entry) { + if (event->state <= PERF_EVENT_STATE_OFF) continue; - if (event->cpu != -1 && event->cpu != cpu) + if (event->cpu != -1 && event->cpu != smp_processor_id()) continue; if (group_can_go_on(event, cpuctx, 1)) - group_sched_in(event, cpuctx, ctx, cpu); + group_sched_in(event, cpuctx, ctx); /* * If this pinned group hasn't been scheduled, @@ -1289,32 +1298,83 @@ __perf_event_sched_in(struct perf_event_context *ctx, event->state = PERF_EVENT_STATE_ERROR; } } +} - list_for_each_entry(event, &ctx->group_list, group_entry) { - /* - * Ignore events in OFF or ERROR state, and - * ignore pinned events since we did them already. - */ - if (event->state <= PERF_EVENT_STATE_OFF || - event->attr.pinned) - continue; +static void +ctx_flexible_sched_in(struct perf_event_context *ctx, + struct perf_cpu_context *cpuctx) +{ + struct perf_event *event; + int can_add_hw = 1; + list_for_each_entry(event, &ctx->flexible_groups, group_entry) { + /* Ignore events in OFF or ERROR state */ + if (event->state <= PERF_EVENT_STATE_OFF) + continue; /* * Listen to the 'cpu' scheduling filter constraint * of events: */ - if (event->cpu != -1 && event->cpu != cpu) + if (event->cpu != -1 && event->cpu != smp_processor_id()) continue; if (group_can_go_on(event, cpuctx, can_add_hw)) - if (group_sched_in(event, cpuctx, ctx, cpu)) + if (group_sched_in(event, cpuctx, ctx)) can_add_hw = 0; } +} + +static void +ctx_sched_in(struct perf_event_context *ctx, + struct perf_cpu_context *cpuctx, + enum event_type_t event_type) +{ + raw_spin_lock(&ctx->lock); + ctx->is_active = 1; + if (likely(!ctx->nr_events)) + goto out; + + ctx->timestamp = perf_clock(); + + perf_disable(); + + /* + * First go through the list and put on any pinned groups + * in order to give them the best chance of going on. + */ + if (event_type & EVENT_PINNED) + ctx_pinned_sched_in(ctx, cpuctx); + + /* Then walk through the lower prio flexible groups */ + if (event_type & EVENT_FLEXIBLE) + ctx_flexible_sched_in(ctx, cpuctx); + perf_enable(); out: raw_spin_unlock(&ctx->lock); } +static void cpu_ctx_sched_in(struct perf_cpu_context *cpuctx, + enum event_type_t event_type) +{ + struct perf_event_context *ctx = &cpuctx->ctx; + + ctx_sched_in(ctx, cpuctx, event_type); +} + +static void task_ctx_sched_in(struct task_struct *task, + enum event_type_t event_type) +{ + struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); + struct perf_event_context *ctx = task->perf_event_ctxp; + + if (likely(!ctx)) + return; + if (cpuctx->task_ctx == ctx) + return; + ctx_sched_in(ctx, cpuctx, event_type); + cpuctx->task_ctx = ctx; +} /* * Called from scheduler to add the events of the current task * with interrupts disabled. @@ -1326,38 +1386,128 @@ __perf_event_sched_in(struct perf_event_context *ctx, * accessing the event control register. If a NMI hits, then it will * keep the event running. */ -void perf_event_task_sched_in(struct task_struct *task, int cpu) +void perf_event_task_sched_in(struct task_struct *task) { - struct perf_cpu_context *cpuctx = &per_cpu(perf_cpu_context, cpu); + struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); struct perf_event_context *ctx = task->perf_event_ctxp; if (likely(!ctx)) return; + if (cpuctx->task_ctx == ctx) return; - __perf_event_sched_in(ctx, cpuctx, cpu); + + /* + * We want to keep the following priority order: + * cpu pinned (that don't need to move), task pinned, + * cpu flexible, task flexible. + */ + cpu_ctx_sched_out(cpuctx, EVENT_FLEXIBLE); + + ctx_sched_in(ctx, cpuctx, EVENT_PINNED); + cpu_ctx_sched_in(cpuctx, EVENT_FLEXIBLE); + ctx_sched_in(ctx, cpuctx, EVENT_FLEXIBLE); + cpuctx->task_ctx = ctx; } -static void perf_event_cpu_sched_in(struct perf_cpu_context *cpuctx, int cpu) +#define MAX_INTERRUPTS (~0ULL) + +static void perf_log_throttle(struct perf_event *event, int enable); + +static u64 perf_calculate_period(struct perf_event *event, u64 nsec, u64 count) { - struct perf_event_context *ctx = &cpuctx->ctx; + u64 frequency = event->attr.sample_freq; + u64 sec = NSEC_PER_SEC; + u64 divisor, dividend; + + int count_fls, nsec_fls, frequency_fls, sec_fls; + + count_fls = fls64(count); + nsec_fls = fls64(nsec); + frequency_fls = fls64(frequency); + sec_fls = 30; + + /* + * We got @count in @nsec, with a target of sample_freq HZ + * the target period becomes: + * + * @count * 10^9 + * period = ------------------- + * @nsec * sample_freq + * + */ + + /* + * Reduce accuracy by one bit such that @a and @b converge + * to a similar magnitude. + */ +#define REDUCE_FLS(a, b) \ +do { \ + if (a##_fls > b##_fls) { \ + a >>= 1; \ + a##_fls--; \ + } else { \ + b >>= 1; \ + b##_fls--; \ + } \ +} while (0) + + /* + * Reduce accuracy until either term fits in a u64, then proceed with + * the other, so that finally we can do a u64/u64 division. + */ + while (count_fls + sec_fls > 64 && nsec_fls + frequency_fls > 64) { + REDUCE_FLS(nsec, frequency); + REDUCE_FLS(sec, count); + } + + if (count_fls + sec_fls > 64) { + divisor = nsec * frequency; - __perf_event_sched_in(ctx, cpuctx, cpu); + while (count_fls + sec_fls > 64) { + REDUCE_FLS(count, sec); + divisor >>= 1; + } + + dividend = count * sec; + } else { + dividend = count * sec; + + while (nsec_fls + frequency_fls > 64) { + REDUCE_FLS(nsec, frequency); + dividend >>= 1; + } + + divisor = nsec * frequency; + } + + return div64_u64(dividend, divisor); } -#define MAX_INTERRUPTS (~0ULL) +static void perf_event_stop(struct perf_event *event) +{ + if (!event->pmu->stop) + return event->pmu->disable(event); -static void perf_log_throttle(struct perf_event *event, int enable); + return event->pmu->stop(event); +} + +static int perf_event_start(struct perf_event *event) +{ + if (!event->pmu->start) + return event->pmu->enable(event); + + return event->pmu->start(event); +} -static void perf_adjust_period(struct perf_event *event, u64 events) +static void perf_adjust_period(struct perf_event *event, u64 nsec, u64 count) { struct hw_perf_event *hwc = &event->hw; u64 period, sample_period; s64 delta; - events *= hwc->sample_period; - period = div64_u64(events, event->attr.sample_freq); + period = perf_calculate_period(event, nsec, count); delta = (s64)(period - hwc->sample_period); delta = (delta + 7) / 8; /* low pass filter */ @@ -1368,13 +1518,22 @@ static void perf_adjust_period(struct perf_event *event, u64 events) sample_period = 1; hwc->sample_period = sample_period; + + if (atomic64_read(&hwc->period_left) > 8*sample_period) { + perf_disable(); + perf_event_stop(event); + atomic64_set(&hwc->period_left, 0); + perf_event_start(event); + perf_enable(); + } } static void perf_ctx_adjust_freq(struct perf_event_context *ctx) { struct perf_event *event; struct hw_perf_event *hwc; - u64 interrupts, freq; + u64 interrupts, now; + s64 delta; raw_spin_lock(&ctx->lock); list_for_each_entry_rcu(event, &ctx->event_list, event_entry) { @@ -1395,44 +1554,18 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx) if (interrupts == MAX_INTERRUPTS) { perf_log_throttle(event, 1); event->pmu->unthrottle(event); - interrupts = 2*sysctl_perf_event_sample_rate/HZ; } if (!event->attr.freq || !event->attr.sample_freq) continue; - /* - * if the specified freq < HZ then we need to skip ticks - */ - if (event->attr.sample_freq < HZ) { - freq = event->attr.sample_freq; - - hwc->freq_count += freq; - hwc->freq_interrupts += interrupts; - - if (hwc->freq_count < HZ) - continue; - - interrupts = hwc->freq_interrupts; - hwc->freq_interrupts = 0; - hwc->freq_count -= HZ; - } else - freq = HZ; - - perf_adjust_period(event, freq * interrupts); + event->pmu->read(event); + now = atomic64_read(&event->count); + delta = now - hwc->freq_count_stamp; + hwc->freq_count_stamp = now; - /* - * In order to avoid being stalled by an (accidental) huge - * sample period, force reset the sample period if we didn't - * get any events in this freq period. - */ - if (!interrupts) { - perf_disable(); - event->pmu->disable(event); - atomic64_set(&hwc->period_left, 0); - event->pmu->enable(event); - perf_enable(); - } + if (delta > 0) + perf_adjust_period(event, TICK_NSEC, delta); } raw_spin_unlock(&ctx->lock); } @@ -1442,26 +1575,18 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx) */ static void rotate_ctx(struct perf_event_context *ctx) { - struct perf_event *event; - if (!ctx->nr_events) return; raw_spin_lock(&ctx->lock); - /* - * Rotate the first entry last (works just fine for group events too): - */ - perf_disable(); - list_for_each_entry(event, &ctx->group_list, group_entry) { - list_move_tail(&event->group_entry, &ctx->group_list); - break; - } - perf_enable(); + + /* Rotate the first entry last of non-pinned groups */ + list_rotate_left(&ctx->flexible_groups); raw_spin_unlock(&ctx->lock); } -void perf_event_task_tick(struct task_struct *curr, int cpu) +void perf_event_task_tick(struct task_struct *curr) { struct perf_cpu_context *cpuctx; struct perf_event_context *ctx; @@ -1469,24 +1594,43 @@ void perf_event_task_tick(struct task_struct *curr, int cpu) if (!atomic_read(&nr_events)) return; - cpuctx = &per_cpu(perf_cpu_context, cpu); + cpuctx = &__get_cpu_var(perf_cpu_context); ctx = curr->perf_event_ctxp; + perf_disable(); + perf_ctx_adjust_freq(&cpuctx->ctx); if (ctx) perf_ctx_adjust_freq(ctx); - perf_event_cpu_sched_out(cpuctx); + cpu_ctx_sched_out(cpuctx, EVENT_FLEXIBLE); if (ctx) - __perf_event_task_sched_out(ctx); + task_ctx_sched_out(ctx, EVENT_FLEXIBLE); rotate_ctx(&cpuctx->ctx); if (ctx) rotate_ctx(ctx); - perf_event_cpu_sched_in(cpuctx, cpu); + cpu_ctx_sched_in(cpuctx, EVENT_FLEXIBLE); if (ctx) - perf_event_task_sched_in(curr, cpu); + task_ctx_sched_in(curr, EVENT_FLEXIBLE); + + perf_enable(); +} + +static int event_enable_on_exec(struct perf_event *event, + struct perf_event_context *ctx) +{ + if (!event->attr.enable_on_exec) + return 0; + + event->attr.enable_on_exec = 0; + if (event->state >= PERF_EVENT_STATE_INACTIVE) + return 0; + + __perf_event_mark_enabled(event, ctx); + + return 1; } /* @@ -1499,6 +1643,7 @@ static void perf_event_enable_on_exec(struct task_struct *task) struct perf_event *event; unsigned long flags; int enabled = 0; + int ret; local_irq_save(flags); ctx = task->perf_event_ctxp; @@ -1509,14 +1654,16 @@ static void perf_event_enable_on_exec(struct task_struct *task) raw_spin_lock(&ctx->lock); - list_for_each_entry(event, &ctx->group_list, group_entry) { - if (!event->attr.enable_on_exec) - continue; - event->attr.enable_on_exec = 0; - if (event->state >= PERF_EVENT_STATE_INACTIVE) - continue; - __perf_event_mark_enabled(event, ctx); - enabled = 1; + list_for_each_entry(event, &ctx->pinned_groups, group_entry) { + ret = event_enable_on_exec(event, ctx); + if (ret) + enabled = 1; + } + + list_for_each_entry(event, &ctx->flexible_groups, group_entry) { + ret = event_enable_on_exec(event, ctx); + if (ret) + enabled = 1; } /* @@ -1527,7 +1674,7 @@ static void perf_event_enable_on_exec(struct task_struct *task) raw_spin_unlock(&ctx->lock); - perf_event_task_sched_in(task, smp_processor_id()); + perf_event_task_sched_in(task); out: local_irq_restore(flags); } @@ -1590,7 +1737,8 @@ __perf_event_init_context(struct perf_event_context *ctx, { raw_spin_lock_init(&ctx->lock); mutex_init(&ctx->mutex); - INIT_LIST_HEAD(&ctx->group_list); + INIT_LIST_HEAD(&ctx->pinned_groups); + INIT_LIST_HEAD(&ctx->flexible_groups); INIT_LIST_HEAD(&ctx->event_list); atomic_set(&ctx->refcount, 1); ctx->task = task; @@ -3259,8 +3407,6 @@ static void perf_event_task_output(struct perf_event *event, task_event->event_id.tid = perf_event_tid(event, task); task_event->event_id.ptid = perf_event_tid(event, current); - task_event->event_id.time = perf_clock(); - perf_output_put(&handle, task_event->event_id); perf_output_end(&handle); @@ -3268,7 +3414,7 @@ static void perf_event_task_output(struct perf_event *event, static int perf_event_task_match(struct perf_event *event) { - if (event->state != PERF_EVENT_STATE_ACTIVE) + if (event->state < PERF_EVENT_STATE_INACTIVE) return 0; if (event->cpu != -1 && event->cpu != smp_processor_id()) @@ -3300,7 +3446,7 @@ static void perf_event_task_event(struct perf_task_event *task_event) cpuctx = &get_cpu_var(perf_cpu_context); perf_event_task_ctx(&cpuctx->ctx, task_event); if (!ctx) - ctx = rcu_dereference(task_event->task->perf_event_ctxp); + ctx = rcu_dereference(current->perf_event_ctxp); if (ctx) perf_event_task_ctx(ctx, task_event); put_cpu_var(perf_cpu_context); @@ -3331,6 +3477,7 @@ static void perf_event_task(struct task_struct *task, /* .ppid */ /* .tid */ /* .ptid */ + .time = perf_clock(), }, }; @@ -3380,7 +3527,7 @@ static void perf_event_comm_output(struct perf_event *event, static int perf_event_comm_match(struct perf_event *event) { - if (event->state != PERF_EVENT_STATE_ACTIVE) + if (event->state < PERF_EVENT_STATE_INACTIVE) return 0; if (event->cpu != -1 && event->cpu != smp_processor_id()) @@ -3500,7 +3647,7 @@ static void perf_event_mmap_output(struct perf_event *event, static int perf_event_mmap_match(struct perf_event *event, struct perf_mmap_event *mmap_event) { - if (event->state != PERF_EVENT_STATE_ACTIVE) + if (event->state < PERF_EVENT_STATE_INACTIVE) return 0; if (event->cpu != -1 && event->cpu != smp_processor_id()) @@ -3609,7 +3756,7 @@ void __perf_event_mmap(struct vm_area_struct *vma) /* .tid */ .start = vma->vm_start, .len = vma->vm_end - vma->vm_start, - .pgoff = vma->vm_pgoff, + .pgoff = (u64)vma->vm_pgoff << PAGE_SHIFT, }, }; @@ -3689,12 +3836,12 @@ static int __perf_event_overflow(struct perf_event *event, int nmi, if (event->attr.freq) { u64 now = perf_clock(); - s64 delta = now - hwc->freq_stamp; + s64 delta = now - hwc->freq_time_stamp; - hwc->freq_stamp = now; + hwc->freq_time_stamp = now; - if (delta > 0 && delta < TICK_NSEC) - perf_adjust_period(event, NSEC_PER_SEC / (int)delta); + if (delta > 0 && delta < 2*TICK_NSEC) + perf_adjust_period(event, delta, hwc->last_period); } /* @@ -4185,7 +4332,7 @@ static const struct pmu perf_ops_task_clock = { .read = task_clock_perf_event_read, }; -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_EVENT_TRACING void perf_tp_event(int event_id, u64 addr, u64 count, void *record, int entry_size) @@ -4290,7 +4437,7 @@ static void perf_event_free_filter(struct perf_event *event) { } -#endif /* CONFIG_EVENT_PROFILE */ +#endif /* CONFIG_EVENT_TRACING */ #ifdef CONFIG_HAVE_HW_BREAKPOINT static void bp_perf_event_destroy(struct perf_event *event) @@ -4871,8 +5018,15 @@ inherit_event(struct perf_event *parent_event, else child_event->state = PERF_EVENT_STATE_OFF; - if (parent_event->attr.freq) - child_event->hw.sample_period = parent_event->hw.sample_period; + if (parent_event->attr.freq) { + u64 sample_period = parent_event->hw.sample_period; + struct hw_perf_event *hwc = &child_event->hw; + + hwc->sample_period = sample_period; + hwc->last_period = sample_period; + + atomic64_set(&hwc->period_left, sample_period); + } child_event->overflow_handler = parent_event->overflow_handler; @@ -5040,7 +5194,11 @@ void perf_event_exit_task(struct task_struct *child) mutex_lock_nested(&child_ctx->mutex, SINGLE_DEPTH_NESTING); again: - list_for_each_entry_safe(child_event, tmp, &child_ctx->group_list, + list_for_each_entry_safe(child_event, tmp, &child_ctx->pinned_groups, + group_entry) + __perf_event_exit_task(child_event, child_ctx, child); + + list_for_each_entry_safe(child_event, tmp, &child_ctx->flexible_groups, group_entry) __perf_event_exit_task(child_event, child_ctx, child); @@ -5049,7 +5207,8 @@ again: * its siblings to the list, but we obtained 'tmp' before that which * will still point to the list head terminating the iteration. */ - if (!list_empty(&child_ctx->group_list)) + if (!list_empty(&child_ctx->pinned_groups) || + !list_empty(&child_ctx->flexible_groups)) goto again; mutex_unlock(&child_ctx->mutex); @@ -5057,6 +5216,24 @@ again: put_ctx(child_ctx); } +static void perf_free_event(struct perf_event *event, + struct perf_event_context *ctx) +{ + struct perf_event *parent = event->parent; + + if (WARN_ON_ONCE(!parent)) + return; + + mutex_lock(&parent->child_mutex); + list_del_init(&event->child_list); + mutex_unlock(&parent->child_mutex); + + fput(parent->filp); + + list_del_event(event, ctx); + free_event(event); +} + /* * free an unexposed, unused context as created by inheritance by * init_task below, used by fork() in case of fail. @@ -5071,36 +5248,70 @@ void perf_event_free_task(struct task_struct *task) mutex_lock(&ctx->mutex); again: - list_for_each_entry_safe(event, tmp, &ctx->group_list, group_entry) { - struct perf_event *parent = event->parent; + list_for_each_entry_safe(event, tmp, &ctx->pinned_groups, group_entry) + perf_free_event(event, ctx); - if (WARN_ON_ONCE(!parent)) - continue; + list_for_each_entry_safe(event, tmp, &ctx->flexible_groups, + group_entry) + perf_free_event(event, ctx); + + if (!list_empty(&ctx->pinned_groups) || + !list_empty(&ctx->flexible_groups)) + goto again; - mutex_lock(&parent->child_mutex); - list_del_init(&event->child_list); - mutex_unlock(&parent->child_mutex); + mutex_unlock(&ctx->mutex); - fput(parent->filp); + put_ctx(ctx); +} - list_del_event(event, ctx); - free_event(event); +static int +inherit_task_group(struct perf_event *event, struct task_struct *parent, + struct perf_event_context *parent_ctx, + struct task_struct *child, + int *inherited_all) +{ + int ret; + struct perf_event_context *child_ctx = child->perf_event_ctxp; + + if (!event->attr.inherit) { + *inherited_all = 0; + return 0; } - if (!list_empty(&ctx->group_list)) - goto again; + if (!child_ctx) { + /* + * This is executed from the parent task context, so + * inherit events that have been marked for cloning. + * First allocate and initialize a context for the + * child. + */ - mutex_unlock(&ctx->mutex); + child_ctx = kzalloc(sizeof(struct perf_event_context), + GFP_KERNEL); + if (!child_ctx) + return -ENOMEM; - put_ctx(ctx); + __perf_event_init_context(child_ctx, child); + child->perf_event_ctxp = child_ctx; + get_task_struct(child); + } + + ret = inherit_group(event, parent, parent_ctx, + child, child_ctx); + + if (ret) + *inherited_all = 0; + + return ret; } + /* * Initialize the perf_event context in task_struct */ int perf_event_init_task(struct task_struct *child) { - struct perf_event_context *child_ctx = NULL, *parent_ctx; + struct perf_event_context *child_ctx, *parent_ctx; struct perf_event_context *cloned_ctx; struct perf_event *event; struct task_struct *parent = current; @@ -5138,41 +5349,22 @@ int perf_event_init_task(struct task_struct *child) * We dont have to disable NMIs - we are only looking at * the list, not manipulating it: */ - list_for_each_entry(event, &parent_ctx->group_list, group_entry) { - - if (!event->attr.inherit) { - inherited_all = 0; - continue; - } - - if (!child->perf_event_ctxp) { - /* - * This is executed from the parent task context, so - * inherit events that have been marked for cloning. - * First allocate and initialize a context for the - * child. - */ - - child_ctx = kzalloc(sizeof(struct perf_event_context), - GFP_KERNEL); - if (!child_ctx) { - ret = -ENOMEM; - break; - } - - __perf_event_init_context(child_ctx, child); - child->perf_event_ctxp = child_ctx; - get_task_struct(child); - } + list_for_each_entry(event, &parent_ctx->pinned_groups, group_entry) { + ret = inherit_task_group(event, parent, parent_ctx, child, + &inherited_all); + if (ret) + break; + } - ret = inherit_group(event, parent, parent_ctx, - child, child_ctx); - if (ret) { - inherited_all = 0; + list_for_each_entry(event, &parent_ctx->flexible_groups, group_entry) { + ret = inherit_task_group(event, parent, parent_ctx, child, + &inherited_all); + if (ret) break; - } } + child_ctx = child->perf_event_ctxp; + if (child_ctx && inherited_all) { /* * Mark the child context as a clone of the parent @@ -5221,7 +5413,9 @@ static void __perf_event_exit_cpu(void *info) struct perf_event_context *ctx = &cpuctx->ctx; struct perf_event *event, *tmp; - list_for_each_entry_safe(event, tmp, &ctx->group_list, group_entry) + list_for_each_entry_safe(event, tmp, &ctx->pinned_groups, group_entry) + __perf_event_remove_from_context(event); + list_for_each_entry_safe(event, tmp, &ctx->flexible_groups, group_entry) __perf_event_remove_from_context(event); } static void perf_event_exit_cpu(int cpu) @@ -5259,6 +5453,10 @@ perf_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) perf_event_exit_cpu(cpu); break; + case CPU_DEAD: + hw_perf_event_setup_offline(cpu); + break; + default: break; } diff --git a/kernel/pid.c b/kernel/pid.c index 2e17c9c92cbe..b08e697cd83f 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -367,7 +367,7 @@ struct task_struct *pid_task(struct pid *pid, enum pid_type type) struct task_struct *result = NULL; if (pid) { struct hlist_node *first; - first = rcu_dereference(pid->tasks[type].first); + first = rcu_dereference_check(pid->tasks[type].first, rcu_read_lock_held() || lockdep_is_held(&tasklist_lock)); if (first) result = hlist_entry(first, struct task_struct, pids[(type)].node); } diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig index 91e09d3b2eb2..5c36ea9d55d2 100644 --- a/kernel/power/Kconfig +++ b/kernel/power/Kconfig @@ -27,6 +27,15 @@ config PM_DEBUG code. This is helpful when debugging and reporting PM bugs, like suspend support. +config PM_ADVANCED_DEBUG + bool "Extra PM attributes in sysfs for low-level debugging/testing" + depends on PM_DEBUG + default n + ---help--- + Add extra sysfs attributes allowing one to access some Power Management + fields of device objects from user space. If you are not a kernel + developer interested in debugging/testing Power Management, say "no". + config PM_VERBOSE bool "Verbose Power Management debugging" depends on PM_DEBUG @@ -85,6 +94,11 @@ config PM_SLEEP depends on SUSPEND || HIBERNATION || XEN_SAVE_RESTORE default y +config PM_SLEEP_ADVANCED_DEBUG + bool + depends on PM_ADVANCED_DEBUG + default n + config SUSPEND bool "Suspend to RAM and standby" depends on PM && ARCH_SUSPEND_POSSIBLE @@ -222,3 +236,8 @@ config PM_RUNTIME and the bus type drivers of the buses the devices are on are responsible for the actual handling of the autosuspend requests and wake-up events. + +config PM_OPS + bool + depends on PM_SLEEP || PM_RUNTIME + default y diff --git a/kernel/power/main.c b/kernel/power/main.c index 0998c7139053..b58800b21fc0 100644 --- a/kernel/power/main.c +++ b/kernel/power/main.c @@ -44,6 +44,32 @@ int pm_notifier_call_chain(unsigned long val) == NOTIFY_BAD) ? -EINVAL : 0; } +/* If set, devices may be suspended and resumed asynchronously. */ +int pm_async_enabled = 1; + +static ssize_t pm_async_show(struct kobject *kobj, struct kobj_attribute *attr, + char *buf) +{ + return sprintf(buf, "%d\n", pm_async_enabled); +} + +static ssize_t pm_async_store(struct kobject *kobj, struct kobj_attribute *attr, + const char *buf, size_t n) +{ + unsigned long val; + + if (strict_strtoul(buf, 10, &val)) + return -EINVAL; + + if (val > 1) + return -EINVAL; + + pm_async_enabled = val; + return n; +} + +power_attr(pm_async); + #ifdef CONFIG_PM_DEBUG int pm_test_level = TEST_NONE; @@ -208,9 +234,12 @@ static struct attribute * g[] = { #ifdef CONFIG_PM_TRACE &pm_trace_attr.attr, #endif -#if defined(CONFIG_PM_SLEEP) && defined(CONFIG_PM_DEBUG) +#ifdef CONFIG_PM_SLEEP + &pm_async_attr.attr, +#ifdef CONFIG_PM_DEBUG &pm_test_attr.attr, #endif +#endif NULL, }; diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c index 36cb168e4330..830cadecbdfc 100644 --- a/kernel/power/snapshot.c +++ b/kernel/power/snapshot.c @@ -1181,7 +1181,7 @@ static void free_unnecessary_pages(void) memory_bm_position_reset(©_bm); - while (to_free_normal > 0 && to_free_highmem > 0) { + while (to_free_normal > 0 || to_free_highmem > 0) { unsigned long pfn = memory_bm_next_pfn(©_bm); struct page *page = pfn_to_page(pfn); @@ -1500,7 +1500,7 @@ asmlinkage int swsusp_save(void) { unsigned int nr_pages, nr_highmem; - printk(KERN_INFO "PM: Creating hibernation image: \n"); + printk(KERN_INFO "PM: Creating hibernation image:\n"); drain_local_pages(NULL); nr_pages = count_data_pages(); diff --git a/kernel/power/swap.c b/kernel/power/swap.c index 09b2b0ae9e9d..1d575733d4e1 100644 --- a/kernel/power/swap.c +++ b/kernel/power/swap.c @@ -657,10 +657,6 @@ int swsusp_read(unsigned int *flags_p) struct swsusp_info *header; *flags_p = swsusp_header->flags; - if (IS_ERR(resume_bdev)) { - pr_debug("PM: Image device not initialised\n"); - return PTR_ERR(resume_bdev); - } memset(&snapshot, 0, sizeof(struct snapshot_handle)); error = snapshot_write_next(&snapshot, PAGE_SIZE); diff --git a/kernel/power/swsusp.c b/kernel/power/swsusp.c deleted file mode 100644 index 5b3601bd1893..000000000000 --- a/kernel/power/swsusp.c +++ /dev/null @@ -1,58 +0,0 @@ -/* - * linux/kernel/power/swsusp.c - * - * This file provides code to write suspend image to swap and read it back. - * - * Copyright (C) 1998-2001 Gabor Kuti <seasons@fornax.hu> - * Copyright (C) 1998,2001-2005 Pavel Machek <pavel@suse.cz> - * - * This file is released under the GPLv2. - * - * I'd like to thank the following people for their work: - * - * Pavel Machek <pavel@ucw.cz>: - * Modifications, defectiveness pointing, being with me at the very beginning, - * suspend to swap space, stop all tasks. Port to 2.4.18-ac and 2.5.17. - * - * Steve Doddi <dirk@loth.demon.co.uk>: - * Support the possibility of hardware state restoring. - * - * Raph <grey.havens@earthling.net>: - * Support for preserving states of network devices and virtual console - * (including X and svgatextmode) - * - * Kurt Garloff <garloff@suse.de>: - * Straightened the critical function in order to prevent compilers from - * playing tricks with local variables. - * - * Andreas Mohr <a.mohr@mailto.de> - * - * Alex Badea <vampire@go.ro>: - * Fixed runaway init - * - * Rafael J. Wysocki <rjw@sisk.pl> - * Reworked the freeing of memory and the handling of swap - * - * More state savers are welcome. Especially for the scsi layer... - * - * For TODOs,FIXMEs also look in Documentation/power/swsusp.txt - */ - -#include <linux/mm.h> -#include <linux/suspend.h> -#include <linux/spinlock.h> -#include <linux/kernel.h> -#include <linux/major.h> -#include <linux/swap.h> -#include <linux/pm.h> -#include <linux/swapops.h> -#include <linux/bootmem.h> -#include <linux/syscalls.h> -#include <linux/highmem.h> -#include <linux/time.h> -#include <linux/rbtree.h> -#include <linux/io.h> - -#include "power.h" - -int in_suspend __nosavedata = 0; diff --git a/kernel/power/user.c b/kernel/power/user.c index bf0014d6a5f0..4d2289626a84 100644 --- a/kernel/power/user.c +++ b/kernel/power/user.c @@ -195,6 +195,15 @@ static ssize_t snapshot_write(struct file *filp, const char __user *buf, return res; } +static void snapshot_deprecated_ioctl(unsigned int cmd) +{ + if (printk_ratelimit()) + printk(KERN_NOTICE "%pf: ioctl '%.8x' is deprecated and will " + "be removed soon, update your suspend-to-disk " + "utilities\n", + __builtin_return_address(0), cmd); +} + static long snapshot_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) { @@ -246,8 +255,9 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, data->frozen = 0; break; - case SNAPSHOT_CREATE_IMAGE: case SNAPSHOT_ATOMIC_SNAPSHOT: + snapshot_deprecated_ioctl(cmd); + case SNAPSHOT_CREATE_IMAGE: if (data->mode != O_RDONLY || !data->frozen || data->ready) { error = -EPERM; break; @@ -275,8 +285,9 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, data->ready = 0; break; - case SNAPSHOT_PREF_IMAGE_SIZE: case SNAPSHOT_SET_IMAGE_SIZE: + snapshot_deprecated_ioctl(cmd); + case SNAPSHOT_PREF_IMAGE_SIZE: image_size = arg; break; @@ -290,15 +301,17 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, error = put_user(size, (loff_t __user *)arg); break; - case SNAPSHOT_AVAIL_SWAP_SIZE: case SNAPSHOT_AVAIL_SWAP: + snapshot_deprecated_ioctl(cmd); + case SNAPSHOT_AVAIL_SWAP_SIZE: size = count_swap_pages(data->swap, 1); size <<= PAGE_SHIFT; error = put_user(size, (loff_t __user *)arg); break; - case SNAPSHOT_ALLOC_SWAP_PAGE: case SNAPSHOT_GET_SWAP_PAGE: + snapshot_deprecated_ioctl(cmd); + case SNAPSHOT_ALLOC_SWAP_PAGE: if (data->swap < 0 || data->swap >= MAX_SWAPFILES) { error = -ENODEV; break; @@ -321,6 +334,7 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, break; case SNAPSHOT_SET_SWAP_FILE: /* This ioctl is deprecated */ + snapshot_deprecated_ioctl(cmd); if (!swsusp_swap_in_use()) { /* * User space encodes device types as two-byte values, @@ -362,6 +376,7 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd, break; case SNAPSHOT_PMOPS: /* This ioctl is deprecated */ + snapshot_deprecated_ioctl(cmd); error = -EINVAL; switch (arg) { diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index 9b7fd4723878..f1125c1a6321 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -44,14 +44,43 @@ #include <linux/cpu.h> #include <linux/mutex.h> #include <linux/module.h> +#include <linux/kernel_stat.h> #ifdef CONFIG_DEBUG_LOCK_ALLOC static struct lock_class_key rcu_lock_key; struct lockdep_map rcu_lock_map = STATIC_LOCKDEP_MAP_INIT("rcu_read_lock", &rcu_lock_key); EXPORT_SYMBOL_GPL(rcu_lock_map); + +static struct lock_class_key rcu_bh_lock_key; +struct lockdep_map rcu_bh_lock_map = + STATIC_LOCKDEP_MAP_INIT("rcu_read_lock_bh", &rcu_bh_lock_key); +EXPORT_SYMBOL_GPL(rcu_bh_lock_map); + +static struct lock_class_key rcu_sched_lock_key; +struct lockdep_map rcu_sched_lock_map = + STATIC_LOCKDEP_MAP_INIT("rcu_read_lock_sched", &rcu_sched_lock_key); +EXPORT_SYMBOL_GPL(rcu_sched_lock_map); #endif +int rcu_scheduler_active __read_mostly; +EXPORT_SYMBOL_GPL(rcu_scheduler_active); + +/* + * This function is invoked towards the end of the scheduler's initialization + * process. Before this is called, the idle task might contain + * RCU read-side critical sections (during which time, this idle + * task is booting the system). After this function is called, the + * idle tasks are prohibited from containing RCU read-side critical + * sections. + */ +void rcu_scheduler_starting(void) +{ + WARN_ON(num_online_cpus() != 1); + WARN_ON(nr_context_switches() > 0); + rcu_scheduler_active = 1; +} + /* * Awaken the corresponding synchronize_rcu() instance now that a * grace period has elapsed. diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index 9bb52177af02..258cdf0a91eb 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -61,6 +61,9 @@ static int test_no_idle_hz; /* Test RCU's support for tickless idle CPUs. */ static int shuffle_interval = 3; /* Interval between shuffles (in sec)*/ static int stutter = 5; /* Start/stop testing interval (in sec) */ static int irqreader = 1; /* RCU readers from irq (timers). */ +static int fqs_duration = 0; /* Duration of bursts (us), 0 to disable. */ +static int fqs_holdoff = 0; /* Hold time within burst (us). */ +static int fqs_stutter = 3; /* Wait time between bursts (s). */ static char *torture_type = "rcu"; /* What RCU implementation to torture. */ module_param(nreaders, int, 0444); @@ -79,6 +82,12 @@ module_param(stutter, int, 0444); MODULE_PARM_DESC(stutter, "Number of seconds to run/halt test"); module_param(irqreader, int, 0444); MODULE_PARM_DESC(irqreader, "Allow RCU readers from irq handlers"); +module_param(fqs_duration, int, 0444); +MODULE_PARM_DESC(fqs_duration, "Duration of fqs bursts (us)"); +module_param(fqs_holdoff, int, 0444); +MODULE_PARM_DESC(fqs_holdoff, "Holdoff time within fqs bursts (us)"); +module_param(fqs_stutter, int, 0444); +MODULE_PARM_DESC(fqs_stutter, "Wait time between fqs bursts (s)"); module_param(torture_type, charp, 0444); MODULE_PARM_DESC(torture_type, "Type of RCU to torture (rcu, rcu_bh, srcu)"); @@ -99,6 +108,7 @@ static struct task_struct **reader_tasks; static struct task_struct *stats_task; static struct task_struct *shuffler_task; static struct task_struct *stutter_task; +static struct task_struct *fqs_task; #define RCU_TORTURE_PIPE_LEN 10 @@ -263,6 +273,7 @@ struct rcu_torture_ops { void (*deferred_free)(struct rcu_torture *p); void (*sync)(void); void (*cb_barrier)(void); + void (*fqs)(void); int (*stats)(char *page); int irq_capable; char *name; @@ -347,6 +358,7 @@ static struct rcu_torture_ops rcu_ops = { .deferred_free = rcu_torture_deferred_free, .sync = synchronize_rcu, .cb_barrier = rcu_barrier, + .fqs = rcu_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "rcu" @@ -388,6 +400,7 @@ static struct rcu_torture_ops rcu_sync_ops = { .deferred_free = rcu_sync_torture_deferred_free, .sync = synchronize_rcu, .cb_barrier = NULL, + .fqs = rcu_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "rcu_sync" @@ -403,6 +416,7 @@ static struct rcu_torture_ops rcu_expedited_ops = { .deferred_free = rcu_sync_torture_deferred_free, .sync = synchronize_rcu_expedited, .cb_barrier = NULL, + .fqs = rcu_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "rcu_expedited" @@ -465,6 +479,7 @@ static struct rcu_torture_ops rcu_bh_ops = { .deferred_free = rcu_bh_torture_deferred_free, .sync = rcu_bh_torture_synchronize, .cb_barrier = rcu_barrier_bh, + .fqs = rcu_bh_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "rcu_bh" @@ -480,6 +495,7 @@ static struct rcu_torture_ops rcu_bh_sync_ops = { .deferred_free = rcu_sync_torture_deferred_free, .sync = rcu_bh_torture_synchronize, .cb_barrier = NULL, + .fqs = rcu_bh_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "rcu_bh_sync" @@ -621,6 +637,7 @@ static struct rcu_torture_ops sched_ops = { .deferred_free = rcu_sched_torture_deferred_free, .sync = sched_torture_synchronize, .cb_barrier = rcu_barrier_sched, + .fqs = rcu_sched_force_quiescent_state, .stats = NULL, .irq_capable = 1, .name = "sched" @@ -636,6 +653,7 @@ static struct rcu_torture_ops sched_sync_ops = { .deferred_free = rcu_sync_torture_deferred_free, .sync = sched_torture_synchronize, .cb_barrier = NULL, + .fqs = rcu_sched_force_quiescent_state, .stats = NULL, .name = "sched_sync" }; @@ -650,12 +668,45 @@ static struct rcu_torture_ops sched_expedited_ops = { .deferred_free = rcu_sync_torture_deferred_free, .sync = synchronize_sched_expedited, .cb_barrier = NULL, + .fqs = rcu_sched_force_quiescent_state, .stats = rcu_expedited_torture_stats, .irq_capable = 1, .name = "sched_expedited" }; /* + * RCU torture force-quiescent-state kthread. Repeatedly induces + * bursts of calls to force_quiescent_state(), increasing the probability + * of occurrence of some important types of race conditions. + */ +static int +rcu_torture_fqs(void *arg) +{ + unsigned long fqs_resume_time; + int fqs_burst_remaining; + + VERBOSE_PRINTK_STRING("rcu_torture_fqs task started"); + do { + fqs_resume_time = jiffies + fqs_stutter * HZ; + while (jiffies - fqs_resume_time > LONG_MAX) { + schedule_timeout_interruptible(1); + } + fqs_burst_remaining = fqs_duration; + while (fqs_burst_remaining > 0) { + cur_ops->fqs(); + udelay(fqs_holdoff); + fqs_burst_remaining -= fqs_holdoff; + } + rcu_stutter_wait("rcu_torture_fqs"); + } while (!kthread_should_stop() && fullstop == FULLSTOP_DONTSTOP); + VERBOSE_PRINTK_STRING("rcu_torture_fqs task stopping"); + rcutorture_shutdown_absorb("rcu_torture_fqs"); + while (!kthread_should_stop()) + schedule_timeout_uninterruptible(1); + return 0; +} + +/* * RCU torture writer kthread. Repeatedly substitutes a new structure * for that pointed to by rcu_torture_current, freeing the old structure * after a series of grace periods (the "pipeline"). @@ -745,7 +796,11 @@ static void rcu_torture_timer(unsigned long unused) idx = cur_ops->readlock(); completed = cur_ops->completed(); - p = rcu_dereference(rcu_torture_current); + p = rcu_dereference_check(rcu_torture_current, + rcu_read_lock_held() || + rcu_read_lock_bh_held() || + rcu_read_lock_sched_held() || + srcu_read_lock_held(&srcu_ctl)); if (p == NULL) { /* Leave because rcu_torture_writer is not yet underway */ cur_ops->readunlock(idx); @@ -798,11 +853,15 @@ rcu_torture_reader(void *arg) do { if (irqreader && cur_ops->irq_capable) { if (!timer_pending(&t)) - mod_timer(&t, 1); + mod_timer(&t, jiffies + 1); } idx = cur_ops->readlock(); completed = cur_ops->completed(); - p = rcu_dereference(rcu_torture_current); + p = rcu_dereference_check(rcu_torture_current, + rcu_read_lock_held() || + rcu_read_lock_bh_held() || + rcu_read_lock_sched_held() || + srcu_read_lock_held(&srcu_ctl)); if (p == NULL) { /* Wait for rcu_torture_writer to get underway */ cur_ops->readunlock(idx); @@ -1030,10 +1089,11 @@ rcu_torture_print_module_parms(char *tag) printk(KERN_ALERT "%s" TORTURE_FLAG "--- %s: nreaders=%d nfakewriters=%d " "stat_interval=%d verbose=%d test_no_idle_hz=%d " - "shuffle_interval=%d stutter=%d irqreader=%d\n", + "shuffle_interval=%d stutter=%d irqreader=%d " + "fqs_duration=%d fqs_holdoff=%d fqs_stutter=%d\n", torture_type, tag, nrealreaders, nfakewriters, stat_interval, verbose, test_no_idle_hz, shuffle_interval, - stutter, irqreader); + stutter, irqreader, fqs_duration, fqs_holdoff, fqs_stutter); } static struct notifier_block rcutorture_nb = { @@ -1109,6 +1169,12 @@ rcu_torture_cleanup(void) } stats_task = NULL; + if (fqs_task) { + VERBOSE_PRINTK_STRING("Stopping rcu_torture_fqs task"); + kthread_stop(fqs_task); + } + fqs_task = NULL; + /* Wait for all RCU callbacks to fire. */ if (cur_ops->cb_barrier != NULL) @@ -1154,6 +1220,11 @@ rcu_torture_init(void) mutex_unlock(&fullstop_mutex); return -EINVAL; } + if (cur_ops->fqs == NULL && fqs_duration != 0) { + printk(KERN_ALERT "rcu-torture: ->fqs NULL and non-zero " + "fqs_duration, fqs disabled.\n"); + fqs_duration = 0; + } if (cur_ops->init) cur_ops->init(); /* no "goto unwind" prior to this point!!! */ @@ -1282,6 +1353,19 @@ rcu_torture_init(void) goto unwind; } } + if (fqs_duration < 0) + fqs_duration = 0; + if (fqs_duration) { + /* Create the stutter thread */ + fqs_task = kthread_run(rcu_torture_fqs, NULL, + "rcu_torture_fqs"); + if (IS_ERR(fqs_task)) { + firsterr = PTR_ERR(fqs_task); + VERBOSE_PRINTK_ERRSTRING("Failed to create fqs"); + fqs_task = NULL; + goto unwind; + } + } register_reboot_notifier(&rcutorture_nb); mutex_unlock(&fullstop_mutex); return 0; diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 53ae9598f798..3ec8160fc75f 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -46,7 +46,6 @@ #include <linux/cpu.h> #include <linux/mutex.h> #include <linux/time.h> -#include <linux/kernel_stat.h> #include "rcutree.h" @@ -66,11 +65,11 @@ static struct lock_class_key rcu_node_class[NUM_RCU_LVLS]; .signaled = RCU_GP_IDLE, \ .gpnum = -300, \ .completed = -300, \ - .onofflock = __SPIN_LOCK_UNLOCKED(&name.onofflock), \ + .onofflock = __RAW_SPIN_LOCK_UNLOCKED(&name.onofflock), \ .orphan_cbs_list = NULL, \ .orphan_cbs_tail = &name.orphan_cbs_list, \ .orphan_qlen = 0, \ - .fqslock = __SPIN_LOCK_UNLOCKED(&name.fqslock), \ + .fqslock = __RAW_SPIN_LOCK_UNLOCKED(&name.fqslock), \ .n_force_qs = 0, \ .n_force_qs_ngp = 0, \ } @@ -81,9 +80,6 @@ DEFINE_PER_CPU(struct rcu_data, rcu_sched_data); struct rcu_state rcu_bh_state = RCU_STATE_INITIALIZER(rcu_bh_state); DEFINE_PER_CPU(struct rcu_data, rcu_bh_data); -static int rcu_scheduler_active __read_mostly; - - /* * Return true if an RCU grace period is in progress. The ACCESS_ONCE()s * permit this function to be invoked without holding the root rcu_node @@ -157,6 +153,24 @@ long rcu_batches_completed_bh(void) EXPORT_SYMBOL_GPL(rcu_batches_completed_bh); /* + * Force a quiescent state for RCU BH. + */ +void rcu_bh_force_quiescent_state(void) +{ + force_quiescent_state(&rcu_bh_state, 0); +} +EXPORT_SYMBOL_GPL(rcu_bh_force_quiescent_state); + +/* + * Force a quiescent state for RCU-sched. + */ +void rcu_sched_force_quiescent_state(void) +{ + force_quiescent_state(&rcu_sched_state, 0); +} +EXPORT_SYMBOL_GPL(rcu_sched_force_quiescent_state); + +/* * Does the CPU have callbacks ready to be invoked? */ static int @@ -439,10 +453,10 @@ static void print_other_cpu_stall(struct rcu_state *rsp) /* Only let one CPU complain about others per time interval. */ - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); delta = jiffies - rsp->jiffies_stall; if (delta < RCU_STALL_RAT_DELAY || !rcu_gp_in_progress(rsp)) { - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); return; } rsp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_RECHECK; @@ -452,13 +466,15 @@ static void print_other_cpu_stall(struct rcu_state *rsp) * due to CPU offlining. */ rcu_print_task_stall(rnp); - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); /* OK, time to rat on our buddy... */ printk(KERN_ERR "INFO: RCU detected CPU stalls:"); rcu_for_each_leaf_node(rsp, rnp) { + raw_spin_lock_irqsave(&rnp->lock, flags); rcu_print_task_stall(rnp); + raw_spin_unlock_irqrestore(&rnp->lock, flags); if (rnp->qsmask == 0) continue; for (cpu = 0; cpu <= rnp->grphi - rnp->grplo; cpu++) @@ -469,6 +485,10 @@ static void print_other_cpu_stall(struct rcu_state *rsp) smp_processor_id(), (long)(jiffies - rsp->gp_start)); trigger_all_cpu_backtrace(); + /* If so configured, complain about tasks blocking the grace period. */ + + rcu_print_detail_task_stall(rsp); + force_quiescent_state(rsp, 0); /* Kick them all. */ } @@ -481,11 +501,11 @@ static void print_cpu_stall(struct rcu_state *rsp) smp_processor_id(), jiffies - rsp->gp_start); trigger_all_cpu_backtrace(); - spin_lock_irqsave(&rnp->lock, flags); - if ((long)(jiffies - rsp->jiffies_stall) >= 0) + raw_spin_lock_irqsave(&rnp->lock, flags); + if (ULONG_CMP_GE(jiffies, rsp->jiffies_stall)) rsp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_RECHECK; - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); set_need_resched(); /* kick ourselves to get things going. */ } @@ -545,12 +565,12 @@ static void note_new_gpnum(struct rcu_state *rsp, struct rcu_data *rdp) local_irq_save(flags); rnp = rdp->mynode; if (rdp->gpnum == ACCESS_ONCE(rnp->gpnum) || /* outside lock. */ - !spin_trylock(&rnp->lock)) { /* irqs already off, retry later. */ + !raw_spin_trylock(&rnp->lock)) { /* irqs already off, so later. */ local_irq_restore(flags); return; } __note_new_gpnum(rsp, rnp, rdp); - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } /* @@ -609,12 +629,12 @@ rcu_process_gp_end(struct rcu_state *rsp, struct rcu_data *rdp) local_irq_save(flags); rnp = rdp->mynode; if (rdp->completed == ACCESS_ONCE(rnp->completed) || /* outside lock. */ - !spin_trylock(&rnp->lock)) { /* irqs already off, retry later. */ + !raw_spin_trylock(&rnp->lock)) { /* irqs already off, so later. */ local_irq_restore(flags); return; } __rcu_process_gp_end(rsp, rnp, rdp); - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } /* @@ -659,12 +679,14 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags) struct rcu_data *rdp = rsp->rda[smp_processor_id()]; struct rcu_node *rnp = rcu_get_root(rsp); - if (!cpu_needs_another_gp(rsp, rdp)) { + if (!cpu_needs_another_gp(rsp, rdp) || rsp->fqs_active) { + if (cpu_needs_another_gp(rsp, rdp)) + rsp->fqs_need_gp = 1; if (rnp->completed == rsp->completed) { - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); return; } - spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ /* * Propagate new ->completed value to rcu_node structures @@ -672,9 +694,9 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags) * of the next grace period to process their callbacks. */ rcu_for_each_node_breadth_first(rsp, rnp) { - spin_lock(&rnp->lock); /* irqs already disabled. */ + raw_spin_lock(&rnp->lock); /* irqs already disabled. */ rnp->completed = rsp->completed; - spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ } local_irq_restore(flags); return; @@ -695,15 +717,15 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags) rnp->completed = rsp->completed; rsp->signaled = RCU_SIGNAL_INIT; /* force_quiescent_state OK. */ rcu_start_gp_per_cpu(rsp, rnp, rdp); - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); return; } - spin_unlock(&rnp->lock); /* leave irqs disabled. */ + raw_spin_unlock(&rnp->lock); /* leave irqs disabled. */ /* Exclude any concurrent CPU-hotplug operations. */ - spin_lock(&rsp->onofflock); /* irqs already disabled. */ + raw_spin_lock(&rsp->onofflock); /* irqs already disabled. */ /* * Set the quiescent-state-needed bits in all the rcu_node @@ -723,21 +745,21 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags) * irqs disabled. */ rcu_for_each_node_breadth_first(rsp, rnp) { - spin_lock(&rnp->lock); /* irqs already disabled. */ + raw_spin_lock(&rnp->lock); /* irqs already disabled. */ rcu_preempt_check_blocked_tasks(rnp); rnp->qsmask = rnp->qsmaskinit; rnp->gpnum = rsp->gpnum; rnp->completed = rsp->completed; if (rnp == rdp->mynode) rcu_start_gp_per_cpu(rsp, rnp, rdp); - spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ } rnp = rcu_get_root(rsp); - spin_lock(&rnp->lock); /* irqs already disabled. */ + raw_spin_lock(&rnp->lock); /* irqs already disabled. */ rsp->signaled = RCU_SIGNAL_INIT; /* force_quiescent_state now OK. */ - spin_unlock(&rnp->lock); /* irqs remain disabled. */ - spin_unlock_irqrestore(&rsp->onofflock, flags); + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock_irqrestore(&rsp->onofflock, flags); } /* @@ -776,14 +798,14 @@ rcu_report_qs_rnp(unsigned long mask, struct rcu_state *rsp, if (!(rnp->qsmask & mask)) { /* Our bit has already been cleared, so done. */ - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); return; } rnp->qsmask &= ~mask; if (rnp->qsmask != 0 || rcu_preempted_readers(rnp)) { /* Other bits still set at this level, so done. */ - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); return; } mask = rnp->grpmask; @@ -793,10 +815,10 @@ rcu_report_qs_rnp(unsigned long mask, struct rcu_state *rsp, break; } - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); rnp_c = rnp; rnp = rnp->parent; - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); WARN_ON_ONCE(rnp_c->qsmask); } @@ -825,7 +847,7 @@ rcu_report_qs_rdp(int cpu, struct rcu_state *rsp, struct rcu_data *rdp, long las struct rcu_node *rnp; rnp = rdp->mynode; - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); if (lastcomp != rnp->completed) { /* @@ -837,12 +859,12 @@ rcu_report_qs_rdp(int cpu, struct rcu_state *rsp, struct rcu_data *rdp, long las * race occurred. */ rdp->passed_quiesc = 0; /* try again later! */ - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); return; } mask = rdp->grpmask; if ((rnp->qsmask & mask) == 0) { - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } else { rdp->qs_pending = 0; @@ -906,7 +928,7 @@ static void rcu_send_cbs_to_orphanage(struct rcu_state *rsp) if (rdp->nxtlist == NULL) return; /* irqs disabled, so comparison is stable. */ - spin_lock(&rsp->onofflock); /* irqs already disabled. */ + raw_spin_lock(&rsp->onofflock); /* irqs already disabled. */ *rsp->orphan_cbs_tail = rdp->nxtlist; rsp->orphan_cbs_tail = rdp->nxttail[RCU_NEXT_TAIL]; rdp->nxtlist = NULL; @@ -914,7 +936,7 @@ static void rcu_send_cbs_to_orphanage(struct rcu_state *rsp) rdp->nxttail[i] = &rdp->nxtlist; rsp->orphan_qlen += rdp->qlen; rdp->qlen = 0; - spin_unlock(&rsp->onofflock); /* irqs remain disabled. */ + raw_spin_unlock(&rsp->onofflock); /* irqs remain disabled. */ } /* @@ -925,10 +947,10 @@ static void rcu_adopt_orphan_cbs(struct rcu_state *rsp) unsigned long flags; struct rcu_data *rdp; - spin_lock_irqsave(&rsp->onofflock, flags); + raw_spin_lock_irqsave(&rsp->onofflock, flags); rdp = rsp->rda[smp_processor_id()]; if (rsp->orphan_cbs_list == NULL) { - spin_unlock_irqrestore(&rsp->onofflock, flags); + raw_spin_unlock_irqrestore(&rsp->onofflock, flags); return; } *rdp->nxttail[RCU_NEXT_TAIL] = rsp->orphan_cbs_list; @@ -937,7 +959,7 @@ static void rcu_adopt_orphan_cbs(struct rcu_state *rsp) rsp->orphan_cbs_list = NULL; rsp->orphan_cbs_tail = &rsp->orphan_cbs_list; rsp->orphan_qlen = 0; - spin_unlock_irqrestore(&rsp->onofflock, flags); + raw_spin_unlock_irqrestore(&rsp->onofflock, flags); } /* @@ -953,23 +975,23 @@ static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp) struct rcu_node *rnp; /* Exclude any attempts to start a new grace period. */ - spin_lock_irqsave(&rsp->onofflock, flags); + raw_spin_lock_irqsave(&rsp->onofflock, flags); /* Remove the outgoing CPU from the masks in the rcu_node hierarchy. */ rnp = rdp->mynode; /* this is the outgoing CPU's rnp. */ mask = rdp->grpmask; /* rnp->grplo is constant. */ do { - spin_lock(&rnp->lock); /* irqs already disabled. */ + raw_spin_lock(&rnp->lock); /* irqs already disabled. */ rnp->qsmaskinit &= ~mask; if (rnp->qsmaskinit != 0) { if (rnp != rdp->mynode) - spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ break; } if (rnp == rdp->mynode) need_report = rcu_preempt_offline_tasks(rsp, rnp, rdp); else - spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ mask = rnp->grpmask; rnp = rnp->parent; } while (rnp != NULL); @@ -980,12 +1002,12 @@ static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp) * because invoking rcu_report_unblock_qs_rnp() with ->onofflock * held leads to deadlock. */ - spin_unlock(&rsp->onofflock); /* irqs remain disabled. */ + raw_spin_unlock(&rsp->onofflock); /* irqs remain disabled. */ rnp = rdp->mynode; if (need_report & RCU_OFL_TASKS_NORM_GP) rcu_report_unblock_qs_rnp(rnp, flags); else - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); if (need_report & RCU_OFL_TASKS_EXP_GP) rcu_report_exp_rnp(rsp, rnp); @@ -1144,11 +1166,9 @@ void rcu_check_callbacks(int cpu, int user) /* * Scan the leaf rcu_node structures, processing dyntick state for any that * have not yet encountered a quiescent state, using the function specified. - * Returns 1 if the current grace period ends while scanning (possibly - * because we made it end). + * The caller must have suppressed start of new grace periods. */ -static int rcu_process_dyntick(struct rcu_state *rsp, long lastcomp, - int (*f)(struct rcu_data *)) +static void force_qs_rnp(struct rcu_state *rsp, int (*f)(struct rcu_data *)) { unsigned long bit; int cpu; @@ -1158,13 +1178,13 @@ static int rcu_process_dyntick(struct rcu_state *rsp, long lastcomp, rcu_for_each_leaf_node(rsp, rnp) { mask = 0; - spin_lock_irqsave(&rnp->lock, flags); - if (rnp->completed != lastcomp) { - spin_unlock_irqrestore(&rnp->lock, flags); - return 1; + raw_spin_lock_irqsave(&rnp->lock, flags); + if (!rcu_gp_in_progress(rsp)) { + raw_spin_unlock_irqrestore(&rnp->lock, flags); + return; } if (rnp->qsmask == 0) { - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); continue; } cpu = rnp->grplo; @@ -1173,15 +1193,14 @@ static int rcu_process_dyntick(struct rcu_state *rsp, long lastcomp, if ((rnp->qsmask & bit) != 0 && f(rsp->rda[cpu])) mask |= bit; } - if (mask != 0 && rnp->completed == lastcomp) { + if (mask != 0) { /* rcu_report_qs_rnp() releases rnp->lock. */ rcu_report_qs_rnp(mask, rsp, rnp, flags); continue; } - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } - return 0; } /* @@ -1191,32 +1210,26 @@ static int rcu_process_dyntick(struct rcu_state *rsp, long lastcomp, static void force_quiescent_state(struct rcu_state *rsp, int relaxed) { unsigned long flags; - long lastcomp; struct rcu_node *rnp = rcu_get_root(rsp); - u8 signaled; - u8 forcenow; if (!rcu_gp_in_progress(rsp)) return; /* No grace period in progress, nothing to force. */ - if (!spin_trylock_irqsave(&rsp->fqslock, flags)) { + if (!raw_spin_trylock_irqsave(&rsp->fqslock, flags)) { rsp->n_force_qs_lh++; /* Inexact, can lose counts. Tough! */ return; /* Someone else is already on the job. */ } - if (relaxed && - (long)(rsp->jiffies_force_qs - jiffies) >= 0) - goto unlock_ret; /* no emergency and done recently. */ + if (relaxed && ULONG_CMP_GE(rsp->jiffies_force_qs, jiffies)) + goto unlock_fqs_ret; /* no emergency and done recently. */ rsp->n_force_qs++; - spin_lock(&rnp->lock); - lastcomp = rsp->gpnum - 1; - signaled = rsp->signaled; + raw_spin_lock(&rnp->lock); /* irqs already disabled */ rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS; if(!rcu_gp_in_progress(rsp)) { rsp->n_force_qs_ngp++; - spin_unlock(&rnp->lock); - goto unlock_ret; /* no GP in progress, time updated. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled */ + goto unlock_fqs_ret; /* no GP in progress, time updated. */ } - spin_unlock(&rnp->lock); - switch (signaled) { + rsp->fqs_active = 1; + switch (rsp->signaled) { case RCU_GP_IDLE: case RCU_GP_INIT: @@ -1224,45 +1237,38 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) case RCU_SAVE_DYNTICK: + raw_spin_unlock(&rnp->lock); /* irqs remain disabled */ if (RCU_SIGNAL_INIT != RCU_SAVE_DYNTICK) break; /* So gcc recognizes the dead code. */ /* Record dyntick-idle state. */ - if (rcu_process_dyntick(rsp, lastcomp, - dyntick_save_progress_counter)) - goto unlock_ret; - /* fall into next case. */ - - case RCU_SAVE_COMPLETED: - - /* Update state, record completion counter. */ - forcenow = 0; - spin_lock(&rnp->lock); - if (lastcomp + 1 == rsp->gpnum && - lastcomp == rsp->completed && - rsp->signaled == signaled) { + force_qs_rnp(rsp, dyntick_save_progress_counter); + raw_spin_lock(&rnp->lock); /* irqs already disabled */ + if (rcu_gp_in_progress(rsp)) rsp->signaled = RCU_FORCE_QS; - rsp->completed_fqs = lastcomp; - forcenow = signaled == RCU_SAVE_COMPLETED; - } - spin_unlock(&rnp->lock); - if (!forcenow) - break; - /* fall into next case. */ + break; case RCU_FORCE_QS: /* Check dyntick-idle state, send IPI to laggarts. */ - if (rcu_process_dyntick(rsp, rsp->completed_fqs, - rcu_implicit_dynticks_qs)) - goto unlock_ret; + raw_spin_unlock(&rnp->lock); /* irqs remain disabled */ + force_qs_rnp(rsp, rcu_implicit_dynticks_qs); /* Leave state in case more forcing is required. */ + raw_spin_lock(&rnp->lock); /* irqs already disabled */ break; } -unlock_ret: - spin_unlock_irqrestore(&rsp->fqslock, flags); + rsp->fqs_active = 0; + if (rsp->fqs_need_gp) { + raw_spin_unlock(&rsp->fqslock); /* irqs remain disabled */ + rsp->fqs_need_gp = 0; + rcu_start_gp(rsp, flags); /* releases rnp->lock */ + return; + } + raw_spin_unlock(&rnp->lock); /* irqs remain disabled */ +unlock_fqs_ret: + raw_spin_unlock_irqrestore(&rsp->fqslock, flags); } #else /* #ifdef CONFIG_SMP */ @@ -1290,7 +1296,7 @@ __rcu_process_callbacks(struct rcu_state *rsp, struct rcu_data *rdp) * If an RCU GP has gone long enough, go check for dyntick * idle CPUs and, if needed, send resched IPIs. */ - if ((long)(ACCESS_ONCE(rsp->jiffies_force_qs) - jiffies) < 0) + if (ULONG_CMP_LT(ACCESS_ONCE(rsp->jiffies_force_qs), jiffies)) force_quiescent_state(rsp, 1); /* @@ -1304,7 +1310,7 @@ __rcu_process_callbacks(struct rcu_state *rsp, struct rcu_data *rdp) /* Does this CPU require a not-yet-started grace period? */ if (cpu_needs_another_gp(rsp, rdp)) { - spin_lock_irqsave(&rcu_get_root(rsp)->lock, flags); + raw_spin_lock_irqsave(&rcu_get_root(rsp)->lock, flags); rcu_start_gp(rsp, flags); /* releases above lock */ } @@ -1335,6 +1341,9 @@ static void rcu_process_callbacks(struct softirq_action *unused) * grace-period manipulations above. */ smp_mb(); /* See above block comment. */ + + /* If we are last CPU on way to dyntick-idle mode, accelerate it. */ + rcu_needs_cpu_flush(); } static void @@ -1369,7 +1378,7 @@ __call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu), unsigned long nestflag; struct rcu_node *rnp_root = rcu_get_root(rsp); - spin_lock_irqsave(&rnp_root->lock, nestflag); + raw_spin_lock_irqsave(&rnp_root->lock, nestflag); rcu_start_gp(rsp, nestflag); /* releases rnp_root->lock. */ } @@ -1387,7 +1396,7 @@ __call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu), force_quiescent_state(rsp, 0); rdp->n_force_qs_snap = rsp->n_force_qs; rdp->qlen_last_fqs_check = rdp->qlen; - } else if ((long)(ACCESS_ONCE(rsp->jiffies_force_qs) - jiffies) < 0) + } else if (ULONG_CMP_LT(ACCESS_ONCE(rsp->jiffies_force_qs), jiffies)) force_quiescent_state(rsp, 1); local_irq_restore(flags); } @@ -1520,7 +1529,7 @@ static int __rcu_pending(struct rcu_state *rsp, struct rcu_data *rdp) /* Has an RCU GP gone long enough to send resched IPIs &c? */ if (rcu_gp_in_progress(rsp) && - ((long)(ACCESS_ONCE(rsp->jiffies_force_qs) - jiffies) < 0)) { + ULONG_CMP_LT(ACCESS_ONCE(rsp->jiffies_force_qs), jiffies)) { rdp->n_rp_need_fqs++; return 1; } @@ -1545,10 +1554,9 @@ static int rcu_pending(int cpu) /* * Check to see if any future RCU-related work will need to be done * by the current CPU, even if none need be done immediately, returning - * 1 if so. This function is part of the RCU implementation; it is -not- - * an exported member of the RCU API. + * 1 if so. */ -int rcu_needs_cpu(int cpu) +static int rcu_needs_cpu_quick_check(int cpu) { /* RCU callbacks either ready or pending? */ return per_cpu(rcu_sched_data, cpu).nxtlist || @@ -1556,21 +1564,6 @@ int rcu_needs_cpu(int cpu) rcu_preempt_needs_cpu(cpu); } -/* - * This function is invoked towards the end of the scheduler's initialization - * process. Before this is called, the idle task might contain - * RCU read-side critical sections (during which time, this idle - * task is booting the system). After this function is called, the - * idle tasks are prohibited from containing RCU read-side critical - * sections. - */ -void rcu_scheduler_starting(void) -{ - WARN_ON(num_online_cpus() != 1); - WARN_ON(nr_context_switches() > 0); - rcu_scheduler_active = 1; -} - static DEFINE_PER_CPU(struct rcu_head, rcu_barrier_head) = {NULL}; static atomic_t rcu_barrier_cpu_count; static DEFINE_MUTEX(rcu_barrier_mutex); @@ -1659,7 +1652,7 @@ rcu_boot_init_percpu_data(int cpu, struct rcu_state *rsp) struct rcu_node *rnp = rcu_get_root(rsp); /* Set up local state, ensuring consistent view of global state. */ - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); rdp->grpmask = 1UL << (cpu - rdp->mynode->grplo); rdp->nxtlist = NULL; for (i = 0; i < RCU_NEXT_SIZE; i++) @@ -1669,7 +1662,7 @@ rcu_boot_init_percpu_data(int cpu, struct rcu_state *rsp) rdp->dynticks = &per_cpu(rcu_dynticks, cpu); #endif /* #ifdef CONFIG_NO_HZ */ rdp->cpu = cpu; - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } /* @@ -1687,7 +1680,7 @@ rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptable) struct rcu_node *rnp = rcu_get_root(rsp); /* Set up local state, ensuring consistent view of global state. */ - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); rdp->passed_quiesc = 0; /* We could be racing with new GP, */ rdp->qs_pending = 1; /* so set up to respond to current GP. */ rdp->beenonline = 1; /* We have now been online. */ @@ -1695,7 +1688,7 @@ rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptable) rdp->qlen_last_fqs_check = 0; rdp->n_force_qs_snap = rsp->n_force_qs; rdp->blimit = blimit; - spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ /* * A new grace period might start here. If so, we won't be part @@ -1703,14 +1696,14 @@ rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptable) */ /* Exclude any attempts to start a new GP on large systems. */ - spin_lock(&rsp->onofflock); /* irqs already disabled. */ + raw_spin_lock(&rsp->onofflock); /* irqs already disabled. */ /* Add CPU to rcu_node bitmasks. */ rnp = rdp->mynode; mask = rdp->grpmask; do { /* Exclude any attempts to start a new GP on small systems. */ - spin_lock(&rnp->lock); /* irqs already disabled. */ + raw_spin_lock(&rnp->lock); /* irqs already disabled. */ rnp->qsmaskinit |= mask; mask = rnp->grpmask; if (rnp == rdp->mynode) { @@ -1718,11 +1711,11 @@ rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptable) rdp->completed = rnp->completed; rdp->passed_quiesc_completed = rnp->completed - 1; } - spin_unlock(&rnp->lock); /* irqs already disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs already disabled. */ rnp = rnp->parent; } while (rnp != NULL && !(rnp->qsmaskinit & mask)); - spin_unlock_irqrestore(&rsp->onofflock, flags); + raw_spin_unlock_irqrestore(&rsp->onofflock, flags); } static void __cpuinit rcu_online_cpu(int cpu) @@ -1806,11 +1799,17 @@ static void __init rcu_init_levelspread(struct rcu_state *rsp) */ static void __init rcu_init_one(struct rcu_state *rsp) { + static char *buf[] = { "rcu_node_level_0", + "rcu_node_level_1", + "rcu_node_level_2", + "rcu_node_level_3" }; /* Match MAX_RCU_LVLS */ int cpustride = 1; int i; int j; struct rcu_node *rnp; + BUILD_BUG_ON(MAX_RCU_LVLS > ARRAY_SIZE(buf)); /* Fix buf[] init! */ + /* Initialize the level-tracking arrays. */ for (i = 1; i < NUM_RCU_LVLS; i++) @@ -1823,8 +1822,9 @@ static void __init rcu_init_one(struct rcu_state *rsp) cpustride *= rsp->levelspread[i]; rnp = rsp->level[i]; for (j = 0; j < rsp->levelcnt[i]; j++, rnp++) { - spin_lock_init(&rnp->lock); - lockdep_set_class(&rnp->lock, &rcu_node_class[i]); + raw_spin_lock_init(&rnp->lock); + lockdep_set_class_and_name(&rnp->lock, + &rcu_node_class[i], buf[i]); rnp->gpnum = 0; rnp->qsmask = 0; rnp->qsmaskinit = 0; @@ -1876,7 +1876,7 @@ do { \ void __init rcu_init(void) { - int i; + int cpu; rcu_bootup_announce(); #ifdef CONFIG_RCU_CPU_STALL_DETECTOR @@ -1896,8 +1896,8 @@ void __init rcu_init(void) * or the scheduler are operational. */ cpu_notifier(rcu_cpu_notify, 0); - for_each_online_cpu(i) - rcu_cpu_notify(NULL, CPU_UP_PREPARE, (void *)(long)i); + for_each_online_cpu(cpu) + rcu_cpu_notify(NULL, CPU_UP_PREPARE, (void *)(long)cpu); } #include "rcutree_plugin.h" diff --git a/kernel/rcutree.h b/kernel/rcutree.h index d2a0046f63b2..1439eb504c22 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -90,12 +90,12 @@ struct rcu_dynticks { * Definition for node within the RCU grace-period-detection hierarchy. */ struct rcu_node { - spinlock_t lock; /* Root rcu_node's lock protects some */ + raw_spinlock_t lock; /* Root rcu_node's lock protects some */ /* rcu_state fields as well as following. */ - long gpnum; /* Current grace period for this node. */ + unsigned long gpnum; /* Current grace period for this node. */ /* This will either be equal to or one */ /* behind the root rcu_node's gpnum. */ - long completed; /* Last grace period completed for this node. */ + unsigned long completed; /* Last GP completed for this node. */ /* This will either be equal to or one */ /* behind the root rcu_node's gpnum. */ unsigned long qsmask; /* CPUs or groups that need to switch in */ @@ -161,11 +161,11 @@ struct rcu_node { /* Per-CPU data for read-copy update. */ struct rcu_data { /* 1) quiescent-state and grace-period handling : */ - long completed; /* Track rsp->completed gp number */ + unsigned long completed; /* Track rsp->completed gp number */ /* in order to detect GP end. */ - long gpnum; /* Highest gp number that this CPU */ + unsigned long gpnum; /* Highest gp number that this CPU */ /* is aware of having started. */ - long passed_quiesc_completed; + unsigned long passed_quiesc_completed; /* Value of completed at time of qs. */ bool passed_quiesc; /* User-mode/idle loop etc. */ bool qs_pending; /* Core waits for quiesc state. */ @@ -221,14 +221,14 @@ struct rcu_data { unsigned long resched_ipi; /* Sent a resched IPI. */ /* 5) __rcu_pending() statistics. */ - long n_rcu_pending; /* rcu_pending() calls since boot. */ - long n_rp_qs_pending; - long n_rp_cb_ready; - long n_rp_cpu_needs_gp; - long n_rp_gp_completed; - long n_rp_gp_started; - long n_rp_need_fqs; - long n_rp_need_nothing; + unsigned long n_rcu_pending; /* rcu_pending() calls since boot. */ + unsigned long n_rp_qs_pending; + unsigned long n_rp_cb_ready; + unsigned long n_rp_cpu_needs_gp; + unsigned long n_rp_gp_completed; + unsigned long n_rp_gp_started; + unsigned long n_rp_need_fqs; + unsigned long n_rp_need_nothing; int cpu; }; @@ -237,12 +237,11 @@ struct rcu_data { #define RCU_GP_IDLE 0 /* No grace period in progress. */ #define RCU_GP_INIT 1 /* Grace period being initialized. */ #define RCU_SAVE_DYNTICK 2 /* Need to scan dyntick state. */ -#define RCU_SAVE_COMPLETED 3 /* Need to save rsp->completed. */ -#define RCU_FORCE_QS 4 /* Need to force quiescent state. */ +#define RCU_FORCE_QS 3 /* Need to force quiescent state. */ #ifdef CONFIG_NO_HZ #define RCU_SIGNAL_INIT RCU_SAVE_DYNTICK #else /* #ifdef CONFIG_NO_HZ */ -#define RCU_SIGNAL_INIT RCU_SAVE_COMPLETED +#define RCU_SIGNAL_INIT RCU_FORCE_QS #endif /* #else #ifdef CONFIG_NO_HZ */ #define RCU_JIFFIES_TILL_FORCE_QS 3 /* for rsp->jiffies_force_qs */ @@ -256,6 +255,9 @@ struct rcu_data { #endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ +#define ULONG_CMP_GE(a, b) (ULONG_MAX / 2 >= (a) - (b)) +#define ULONG_CMP_LT(a, b) (ULONG_MAX / 2 < (a) - (b)) + /* * RCU global state, including node hierarchy. This hierarchy is * represented in "heap" form in a dense array. The root (first level) @@ -277,12 +279,19 @@ struct rcu_state { u8 signaled ____cacheline_internodealigned_in_smp; /* Force QS state. */ - long gpnum; /* Current gp number. */ - long completed; /* # of last completed gp. */ + u8 fqs_active; /* force_quiescent_state() */ + /* is running. */ + u8 fqs_need_gp; /* A CPU was prevented from */ + /* starting a new grace */ + /* period because */ + /* force_quiescent_state() */ + /* was running. */ + unsigned long gpnum; /* Current gp number. */ + unsigned long completed; /* # of last completed gp. */ /* End of fields guarded by root rcu_node's lock. */ - spinlock_t onofflock; /* exclude on/offline and */ + raw_spinlock_t onofflock; /* exclude on/offline and */ /* starting new GP. Also */ /* protects the following */ /* orphan_cbs fields. */ @@ -292,10 +301,8 @@ struct rcu_state { /* going offline. */ struct rcu_head **orphan_cbs_tail; /* And tail pointer. */ long orphan_qlen; /* Number of orphaned cbs. */ - spinlock_t fqslock; /* Only one task forcing */ + raw_spinlock_t fqslock; /* Only one task forcing */ /* quiescent states. */ - long completed_fqs; /* Value of completed @ snap. */ - /* Protected by fqslock. */ unsigned long jiffies_force_qs; /* Time at which to invoke */ /* force_quiescent_state(). */ unsigned long n_force_qs; /* Number of calls to */ @@ -319,8 +326,6 @@ struct rcu_state { #define RCU_OFL_TASKS_EXP_GP 0x2 /* Tasks blocking expedited */ /* GP were moved to root. */ -#ifdef RCU_TREE_NONCORE - /* * RCU implementation internal declarations: */ @@ -335,7 +340,7 @@ extern struct rcu_state rcu_preempt_state; DECLARE_PER_CPU(struct rcu_data, rcu_preempt_data); #endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */ -#else /* #ifdef RCU_TREE_NONCORE */ +#ifndef RCU_TREE_NONCORE /* Forward declarations for rcutree_plugin.h */ static void rcu_bootup_announce(void); @@ -347,6 +352,7 @@ static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags); #endif /* #ifdef CONFIG_HOTPLUG_CPU */ #ifdef CONFIG_RCU_CPU_STALL_DETECTOR +static void rcu_print_detail_task_stall(struct rcu_state *rsp); static void rcu_print_task_stall(struct rcu_node *rnp); #endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ static void rcu_preempt_check_blocked_tasks(struct rcu_node *rnp); @@ -367,5 +373,6 @@ static int rcu_preempt_needs_cpu(int cpu); static void __cpuinit rcu_preempt_init_percpu_data(int cpu); static void rcu_preempt_send_cbs_to_orphanage(void); static void __init __rcu_init_preempt(void); +static void rcu_needs_cpu_flush(void); -#endif /* #else #ifdef RCU_TREE_NONCORE */ +#endif /* #ifndef RCU_TREE_NONCORE */ diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 37fbccdf41d5..464ad2cdee00 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -62,6 +62,15 @@ long rcu_batches_completed(void) EXPORT_SYMBOL_GPL(rcu_batches_completed); /* + * Force a quiescent state for preemptible RCU. + */ +void rcu_force_quiescent_state(void) +{ + force_quiescent_state(&rcu_preempt_state, 0); +} +EXPORT_SYMBOL_GPL(rcu_force_quiescent_state); + +/* * Record a preemptable-RCU quiescent state for the specified CPU. Note * that this just means that the task currently running on the CPU is * not in a quiescent state. There might be any number of tasks blocked @@ -102,7 +111,7 @@ static void rcu_preempt_note_context_switch(int cpu) /* Possibly blocking in an RCU read-side critical section. */ rdp = rcu_preempt_state.rda[cpu]; rnp = rdp->mynode; - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); t->rcu_read_unlock_special |= RCU_READ_UNLOCK_BLOCKED; t->rcu_blocked_node = rnp; @@ -123,7 +132,7 @@ static void rcu_preempt_note_context_switch(int cpu) WARN_ON_ONCE(!list_empty(&t->rcu_node_entry)); phase = (rnp->gpnum + !(rnp->qsmask & rdp->grpmask)) & 0x1; list_add(&t->rcu_node_entry, &rnp->blocked_tasks[phase]); - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } /* @@ -180,7 +189,7 @@ static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags) struct rcu_node *rnp_p; if (rnp->qsmask != 0 || rcu_preempted_readers(rnp)) { - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); return; /* Still need more quiescent states! */ } @@ -197,8 +206,8 @@ static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags) /* Report up the rest of the hierarchy. */ mask = rnp->grpmask; - spin_unlock(&rnp->lock); /* irqs remain disabled. */ - spin_lock(&rnp_p->lock); /* irqs already disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_lock(&rnp_p->lock); /* irqs already disabled. */ rcu_report_qs_rnp(mask, &rcu_preempt_state, rnp_p, flags); } @@ -248,10 +257,10 @@ static void rcu_read_unlock_special(struct task_struct *t) */ for (;;) { rnp = t->rcu_blocked_node; - spin_lock(&rnp->lock); /* irqs already disabled. */ + raw_spin_lock(&rnp->lock); /* irqs already disabled. */ if (rnp == t->rcu_blocked_node) break; - spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ } empty = !rcu_preempted_readers(rnp); empty_exp = !rcu_preempted_readers_exp(rnp); @@ -265,7 +274,7 @@ static void rcu_read_unlock_special(struct task_struct *t) * Note that rcu_report_unblock_qs_rnp() releases rnp->lock. */ if (empty) - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); else rcu_report_unblock_qs_rnp(rnp, flags); @@ -295,29 +304,73 @@ void __rcu_read_unlock(void) if (--ACCESS_ONCE(t->rcu_read_lock_nesting) == 0 && unlikely(ACCESS_ONCE(t->rcu_read_unlock_special))) rcu_read_unlock_special(t); +#ifdef CONFIG_PROVE_LOCKING + WARN_ON_ONCE(ACCESS_ONCE(t->rcu_read_lock_nesting) < 0); +#endif /* #ifdef CONFIG_PROVE_LOCKING */ } EXPORT_SYMBOL_GPL(__rcu_read_unlock); #ifdef CONFIG_RCU_CPU_STALL_DETECTOR +#ifdef CONFIG_RCU_CPU_STALL_VERBOSE + +/* + * Dump detailed information for all tasks blocking the current RCU + * grace period on the specified rcu_node structure. + */ +static void rcu_print_detail_task_stall_rnp(struct rcu_node *rnp) +{ + unsigned long flags; + struct list_head *lp; + int phase; + struct task_struct *t; + + if (rcu_preempted_readers(rnp)) { + raw_spin_lock_irqsave(&rnp->lock, flags); + phase = rnp->gpnum & 0x1; + lp = &rnp->blocked_tasks[phase]; + list_for_each_entry(t, lp, rcu_node_entry) + sched_show_task(t); + raw_spin_unlock_irqrestore(&rnp->lock, flags); + } +} + +/* + * Dump detailed information for all tasks blocking the current RCU + * grace period. + */ +static void rcu_print_detail_task_stall(struct rcu_state *rsp) +{ + struct rcu_node *rnp = rcu_get_root(rsp); + + rcu_print_detail_task_stall_rnp(rnp); + rcu_for_each_leaf_node(rsp, rnp) + rcu_print_detail_task_stall_rnp(rnp); +} + +#else /* #ifdef CONFIG_RCU_CPU_STALL_VERBOSE */ + +static void rcu_print_detail_task_stall(struct rcu_state *rsp) +{ +} + +#endif /* #else #ifdef CONFIG_RCU_CPU_STALL_VERBOSE */ + /* * Scan the current list of tasks blocked within RCU read-side critical * sections, printing out the tid of each. */ static void rcu_print_task_stall(struct rcu_node *rnp) { - unsigned long flags; struct list_head *lp; int phase; struct task_struct *t; if (rcu_preempted_readers(rnp)) { - spin_lock_irqsave(&rnp->lock, flags); phase = rnp->gpnum & 0x1; lp = &rnp->blocked_tasks[phase]; list_for_each_entry(t, lp, rcu_node_entry) printk(" P%d", t->pid); - spin_unlock_irqrestore(&rnp->lock, flags); } } @@ -388,11 +441,11 @@ static int rcu_preempt_offline_tasks(struct rcu_state *rsp, lp_root = &rnp_root->blocked_tasks[i]; while (!list_empty(lp)) { tp = list_entry(lp->next, typeof(*tp), rcu_node_entry); - spin_lock(&rnp_root->lock); /* irqs already disabled */ + raw_spin_lock(&rnp_root->lock); /* irqs already disabled */ list_del(&tp->rcu_node_entry); tp->rcu_blocked_node = rnp_root; list_add(&tp->rcu_node_entry, lp_root); - spin_unlock(&rnp_root->lock); /* irqs remain disabled */ + raw_spin_unlock(&rnp_root->lock); /* irqs remain disabled */ } } return retval; @@ -516,7 +569,7 @@ static void rcu_report_exp_rnp(struct rcu_state *rsp, struct rcu_node *rnp) unsigned long flags; unsigned long mask; - spin_lock_irqsave(&rnp->lock, flags); + raw_spin_lock_irqsave(&rnp->lock, flags); for (;;) { if (!sync_rcu_preempt_exp_done(rnp)) break; @@ -525,12 +578,12 @@ static void rcu_report_exp_rnp(struct rcu_state *rsp, struct rcu_node *rnp) break; } mask = rnp->grpmask; - spin_unlock(&rnp->lock); /* irqs remain disabled */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled */ rnp = rnp->parent; - spin_lock(&rnp->lock); /* irqs already disabled */ + raw_spin_lock(&rnp->lock); /* irqs already disabled */ rnp->expmask &= ~mask; } - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } /* @@ -545,11 +598,11 @@ sync_rcu_preempt_exp_init(struct rcu_state *rsp, struct rcu_node *rnp) { int must_wait; - spin_lock(&rnp->lock); /* irqs already disabled */ + raw_spin_lock(&rnp->lock); /* irqs already disabled */ list_splice_init(&rnp->blocked_tasks[0], &rnp->blocked_tasks[2]); list_splice_init(&rnp->blocked_tasks[1], &rnp->blocked_tasks[3]); must_wait = rcu_preempted_readers_exp(rnp); - spin_unlock(&rnp->lock); /* irqs remain disabled */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled */ if (!must_wait) rcu_report_exp_rnp(rsp, rnp); } @@ -594,13 +647,13 @@ void synchronize_rcu_expedited(void) /* force all RCU readers onto blocked_tasks[]. */ synchronize_sched_expedited(); - spin_lock_irqsave(&rsp->onofflock, flags); + raw_spin_lock_irqsave(&rsp->onofflock, flags); /* Initialize ->expmask for all non-leaf rcu_node structures. */ rcu_for_each_nonleaf_node_breadth_first(rsp, rnp) { - spin_lock(&rnp->lock); /* irqs already disabled. */ + raw_spin_lock(&rnp->lock); /* irqs already disabled. */ rnp->expmask = rnp->qsmaskinit; - spin_unlock(&rnp->lock); /* irqs remain disabled. */ + raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ } /* Snapshot current state of ->blocked_tasks[] lists. */ @@ -609,7 +662,7 @@ void synchronize_rcu_expedited(void) if (NUM_RCU_NODES > 1) sync_rcu_preempt_exp_init(rsp, rcu_get_root(rsp)); - spin_unlock_irqrestore(&rsp->onofflock, flags); + raw_spin_unlock_irqrestore(&rsp->onofflock, flags); /* Wait for snapshotted ->blocked_tasks[] lists to drain. */ rnp = rcu_get_root(rsp); @@ -713,6 +766,16 @@ long rcu_batches_completed(void) EXPORT_SYMBOL_GPL(rcu_batches_completed); /* + * Force a quiescent state for RCU, which, because there is no preemptible + * RCU, becomes the same as rcu-sched. + */ +void rcu_force_quiescent_state(void) +{ + rcu_sched_force_quiescent_state(); +} +EXPORT_SYMBOL_GPL(rcu_force_quiescent_state); + +/* * Because preemptable RCU does not exist, we never have to check for * CPUs being in quiescent states. */ @@ -734,7 +797,7 @@ static int rcu_preempted_readers(struct rcu_node *rnp) /* Because preemptible RCU does not exist, no quieting of tasks. */ static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags) { - spin_unlock_irqrestore(&rnp->lock, flags); + raw_spin_unlock_irqrestore(&rnp->lock, flags); } #endif /* #ifdef CONFIG_HOTPLUG_CPU */ @@ -745,6 +808,14 @@ static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp, unsigned long flags) * Because preemptable RCU does not exist, we never have to check for * tasks blocked within RCU read-side critical sections. */ +static void rcu_print_detail_task_stall(struct rcu_state *rsp) +{ +} + +/* + * Because preemptable RCU does not exist, we never have to check for + * tasks blocked within RCU read-side critical sections. + */ static void rcu_print_task_stall(struct rcu_node *rnp) { } @@ -884,3 +955,113 @@ static void __init __rcu_init_preempt(void) } #endif /* #else #ifdef CONFIG_TREE_PREEMPT_RCU */ + +#if !defined(CONFIG_RCU_FAST_NO_HZ) + +/* + * Check to see if any future RCU-related work will need to be done + * by the current CPU, even if none need be done immediately, returning + * 1 if so. This function is part of the RCU implementation; it is -not- + * an exported member of the RCU API. + * + * Because we have preemptible RCU, just check whether this CPU needs + * any flavor of RCU. Do not chew up lots of CPU cycles with preemption + * disabled in a most-likely vain attempt to cause RCU not to need this CPU. + */ +int rcu_needs_cpu(int cpu) +{ + return rcu_needs_cpu_quick_check(cpu); +} + +/* + * Check to see if we need to continue a callback-flush operations to + * allow the last CPU to enter dyntick-idle mode. But fast dyntick-idle + * entry is not configured, so we never do need to. + */ +static void rcu_needs_cpu_flush(void) +{ +} + +#else /* #if !defined(CONFIG_RCU_FAST_NO_HZ) */ + +#define RCU_NEEDS_CPU_FLUSHES 5 +static DEFINE_PER_CPU(int, rcu_dyntick_drain); +static DEFINE_PER_CPU(unsigned long, rcu_dyntick_holdoff); + +/* + * Check to see if any future RCU-related work will need to be done + * by the current CPU, even if none need be done immediately, returning + * 1 if so. This function is part of the RCU implementation; it is -not- + * an exported member of the RCU API. + * + * Because we are not supporting preemptible RCU, attempt to accelerate + * any current grace periods so that RCU no longer needs this CPU, but + * only if all other CPUs are already in dynticks-idle mode. This will + * allow the CPU cores to be powered down immediately, as opposed to after + * waiting many milliseconds for grace periods to elapse. + * + * Because it is not legal to invoke rcu_process_callbacks() with irqs + * disabled, we do one pass of force_quiescent_state(), then do a + * raise_softirq() to cause rcu_process_callbacks() to be invoked later. + * The per-cpu rcu_dyntick_drain variable controls the sequencing. + */ +int rcu_needs_cpu(int cpu) +{ + int c = 0; + int thatcpu; + + /* Don't bother unless we are the last non-dyntick-idle CPU. */ + for_each_cpu_not(thatcpu, nohz_cpu_mask) + if (thatcpu != cpu) { + per_cpu(rcu_dyntick_drain, cpu) = 0; + per_cpu(rcu_dyntick_holdoff, cpu) = jiffies - 1; + return rcu_needs_cpu_quick_check(cpu); + } + + /* Check and update the rcu_dyntick_drain sequencing. */ + if (per_cpu(rcu_dyntick_drain, cpu) <= 0) { + /* First time through, initialize the counter. */ + per_cpu(rcu_dyntick_drain, cpu) = RCU_NEEDS_CPU_FLUSHES; + } else if (--per_cpu(rcu_dyntick_drain, cpu) <= 0) { + /* We have hit the limit, so time to give up. */ + per_cpu(rcu_dyntick_holdoff, cpu) = jiffies; + return rcu_needs_cpu_quick_check(cpu); + } + + /* Do one step pushing remaining RCU callbacks through. */ + if (per_cpu(rcu_sched_data, cpu).nxtlist) { + rcu_sched_qs(cpu); + force_quiescent_state(&rcu_sched_state, 0); + c = c || per_cpu(rcu_sched_data, cpu).nxtlist; + } + if (per_cpu(rcu_bh_data, cpu).nxtlist) { + rcu_bh_qs(cpu); + force_quiescent_state(&rcu_bh_state, 0); + c = c || per_cpu(rcu_bh_data, cpu).nxtlist; + } + + /* If RCU callbacks are still pending, RCU still needs this CPU. */ + if (c) { + raise_softirq(RCU_SOFTIRQ); + per_cpu(rcu_dyntick_holdoff, cpu) = jiffies; + } + return c; +} + +/* + * Check to see if we need to continue a callback-flush operations to + * allow the last CPU to enter dyntick-idle mode. + */ +static void rcu_needs_cpu_flush(void) +{ + int cpu = smp_processor_id(); + unsigned long flags; + + if (per_cpu(rcu_dyntick_drain, cpu) <= 0) + return; + local_irq_save(flags); + (void)rcu_needs_cpu(cpu); + local_irq_restore(flags); +} + +#endif /* #else #if !defined(CONFIG_RCU_FAST_NO_HZ) */ diff --git a/kernel/rcutree_trace.c b/kernel/rcutree_trace.c index 9d2c88423b31..d45db2e35d27 100644 --- a/kernel/rcutree_trace.c +++ b/kernel/rcutree_trace.c @@ -50,7 +50,7 @@ static void print_one_rcu_data(struct seq_file *m, struct rcu_data *rdp) { if (!rdp->beenonline) return; - seq_printf(m, "%3d%cc=%ld g=%ld pq=%d pqc=%ld qp=%d", + seq_printf(m, "%3d%cc=%lu g=%lu pq=%d pqc=%lu qp=%d", rdp->cpu, cpu_is_offline(rdp->cpu) ? '!' : ' ', rdp->completed, rdp->gpnum, @@ -105,7 +105,7 @@ static void print_one_rcu_data_csv(struct seq_file *m, struct rcu_data *rdp) { if (!rdp->beenonline) return; - seq_printf(m, "%d,%s,%ld,%ld,%d,%ld,%d", + seq_printf(m, "%d,%s,%lu,%lu,%d,%lu,%d", rdp->cpu, cpu_is_offline(rdp->cpu) ? "\"N\"" : "\"Y\"", rdp->completed, rdp->gpnum, @@ -155,13 +155,13 @@ static const struct file_operations rcudata_csv_fops = { static void print_one_rcu_state(struct seq_file *m, struct rcu_state *rsp) { - long gpnum; + unsigned long gpnum; int level = 0; int phase; struct rcu_node *rnp; gpnum = rsp->gpnum; - seq_printf(m, "c=%ld g=%ld s=%d jfq=%ld j=%x " + seq_printf(m, "c=%lu g=%lu s=%d jfq=%ld j=%x " "nfqs=%lu/nfqsng=%lu(%lu) fqlh=%lu oqlen=%ld\n", rsp->completed, gpnum, rsp->signaled, (long)(rsp->jiffies_force_qs - jiffies), @@ -215,12 +215,12 @@ static const struct file_operations rcuhier_fops = { static int show_rcugp(struct seq_file *m, void *unused) { #ifdef CONFIG_TREE_PREEMPT_RCU - seq_printf(m, "rcu_preempt: completed=%ld gpnum=%ld\n", + seq_printf(m, "rcu_preempt: completed=%ld gpnum=%lu\n", rcu_preempt_state.completed, rcu_preempt_state.gpnum); #endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */ - seq_printf(m, "rcu_sched: completed=%ld gpnum=%ld\n", + seq_printf(m, "rcu_sched: completed=%ld gpnum=%lu\n", rcu_sched_state.completed, rcu_sched_state.gpnum); - seq_printf(m, "rcu_bh: completed=%ld gpnum=%ld\n", + seq_printf(m, "rcu_bh: completed=%ld gpnum=%lu\n", rcu_bh_state.completed, rcu_bh_state.gpnum); return 0; } diff --git a/kernel/resource.c b/kernel/resource.c index 03c897f7935e..4e9d87fd7bc5 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -188,6 +188,36 @@ static int __release_resource(struct resource *old) return -EINVAL; } +static void __release_child_resources(struct resource *r) +{ + struct resource *tmp, *p; + resource_size_t size; + + p = r->child; + r->child = NULL; + while (p) { + tmp = p; + p = p->sibling; + + tmp->parent = NULL; + tmp->sibling = NULL; + __release_child_resources(tmp); + + printk(KERN_DEBUG "release child resource %pR\n", tmp); + /* need to restore size, and keep flags */ + size = resource_size(tmp); + tmp->start = 0; + tmp->end = size - 1; + } +} + +void release_child_resources(struct resource *r) +{ + write_lock(&resource_lock); + __release_child_resources(r); + write_unlock(&resource_lock); +} + /** * request_resource - request and reserve an I/O or memory resource * @root: root resource descriptor @@ -316,8 +346,10 @@ int __weak page_is_ram(unsigned long pfn) static int find_resource(struct resource *root, struct resource *new, resource_size_t size, resource_size_t min, resource_size_t max, resource_size_t align, - void (*alignf)(void *, struct resource *, - resource_size_t, resource_size_t), + resource_size_t (*alignf)(void *, + const struct resource *, + resource_size_t, + resource_size_t), void *alignf_data) { struct resource *this = root->child; @@ -343,7 +375,7 @@ static int find_resource(struct resource *root, struct resource *new, tmp.end = max; tmp.start = ALIGN(tmp.start, align); if (alignf) - alignf(alignf_data, &tmp, size, align); + tmp.start = alignf(alignf_data, &tmp, size, align); if (tmp.start < tmp.end && tmp.end - tmp.start >= size - 1) { new->start = tmp.start; new->end = tmp.start + size - 1; @@ -371,8 +403,10 @@ static int find_resource(struct resource *root, struct resource *new, int allocate_resource(struct resource *root, struct resource *new, resource_size_t size, resource_size_t min, resource_size_t max, resource_size_t align, - void (*alignf)(void *, struct resource *, - resource_size_t, resource_size_t), + resource_size_t (*alignf)(void *, + const struct resource *, + resource_size_t, + resource_size_t), void *alignf_data) { int err; diff --git a/kernel/sched.c b/kernel/sched.c index 3a8fb30a91b1..6a212c97f523 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -233,7 +233,7 @@ static void destroy_rt_bandwidth(struct rt_bandwidth *rt_b) */ static DEFINE_MUTEX(sched_domains_mutex); -#ifdef CONFIG_GROUP_SCHED +#ifdef CONFIG_CGROUP_SCHED #include <linux/cgroup.h> @@ -243,13 +243,7 @@ static LIST_HEAD(task_groups); /* task group related information */ struct task_group { -#ifdef CONFIG_CGROUP_SCHED struct cgroup_subsys_state css; -#endif - -#ifdef CONFIG_USER_SCHED - uid_t uid; -#endif #ifdef CONFIG_FAIR_GROUP_SCHED /* schedulable entities of this group on each cpu */ @@ -274,35 +268,7 @@ struct task_group { struct list_head children; }; -#ifdef CONFIG_USER_SCHED - -/* Helper function to pass uid information to create_sched_user() */ -void set_tg_uid(struct user_struct *user) -{ - user->tg->uid = user->uid; -} - -/* - * Root task group. - * Every UID task group (including init_task_group aka UID-0) will - * be a child to this group. - */ -struct task_group root_task_group; - -#ifdef CONFIG_FAIR_GROUP_SCHED -/* Default task group's sched entity on each cpu */ -static DEFINE_PER_CPU(struct sched_entity, init_sched_entity); -/* Default task group's cfs_rq on each cpu */ -static DEFINE_PER_CPU_SHARED_ALIGNED(struct cfs_rq, init_tg_cfs_rq); -#endif /* CONFIG_FAIR_GROUP_SCHED */ - -#ifdef CONFIG_RT_GROUP_SCHED -static DEFINE_PER_CPU(struct sched_rt_entity, init_sched_rt_entity); -static DEFINE_PER_CPU_SHARED_ALIGNED(struct rt_rq, init_rt_rq_var); -#endif /* CONFIG_RT_GROUP_SCHED */ -#else /* !CONFIG_USER_SCHED */ #define root_task_group init_task_group -#endif /* CONFIG_USER_SCHED */ /* task_group_lock serializes add/remove of task groups and also changes to * a task group's cpu shares. @@ -318,11 +284,7 @@ static int root_task_group_empty(void) } #endif -#ifdef CONFIG_USER_SCHED -# define INIT_TASK_GROUP_LOAD (2*NICE_0_LOAD) -#else /* !CONFIG_USER_SCHED */ # define INIT_TASK_GROUP_LOAD NICE_0_LOAD -#endif /* CONFIG_USER_SCHED */ /* * A weight of 0 or 1 can cause arithmetics problems. @@ -348,11 +310,7 @@ static inline struct task_group *task_group(struct task_struct *p) { struct task_group *tg; -#ifdef CONFIG_USER_SCHED - rcu_read_lock(); - tg = __task_cred(p)->user->tg; - rcu_read_unlock(); -#elif defined(CONFIG_CGROUP_SCHED) +#ifdef CONFIG_CGROUP_SCHED tg = container_of(task_subsys_state(p, cpu_cgroup_subsys_id), struct task_group, css); #else @@ -383,7 +341,7 @@ static inline struct task_group *task_group(struct task_struct *p) return NULL; } -#endif /* CONFIG_GROUP_SCHED */ +#endif /* CONFIG_CGROUP_SCHED */ /* CFS-related fields in a runqueue */ struct cfs_rq { @@ -478,7 +436,6 @@ struct rt_rq { struct rq *rq; struct list_head leaf_rt_rq_list; struct task_group *tg; - struct sched_rt_entity *rt_se; #endif }; @@ -645,6 +602,11 @@ static inline int cpu_of(struct rq *rq) #endif } +#define rcu_dereference_check_sched_domain(p) \ + rcu_dereference_check((p), \ + rcu_read_lock_sched_held() || \ + lockdep_is_held(&sched_domains_mutex)) + /* * The domain tree (rq->sd) is protected by RCU's quiescent state transition. * See detach_destroy_domains: synchronize_sched for details. @@ -653,7 +615,7 @@ static inline int cpu_of(struct rq *rq) * preempt-disabled sections. */ #define for_each_domain(cpu, __sd) \ - for (__sd = rcu_dereference(cpu_rq(cpu)->sd); __sd; __sd = __sd->parent) + for (__sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd); __sd; __sd = __sd->parent) #define cpu_rq(cpu) (&per_cpu(runqueues, (cpu))) #define this_rq() (&__get_cpu_var(runqueues)) @@ -941,16 +903,33 @@ static inline void finish_lock_switch(struct rq *rq, struct task_struct *prev) #endif /* __ARCH_WANT_UNLOCKED_CTXSW */ /* + * Check whether the task is waking, we use this to synchronize against + * ttwu() so that task_cpu() reports a stable number. + * + * We need to make an exception for PF_STARTING tasks because the fork + * path might require task_rq_lock() to work, eg. it can call + * set_cpus_allowed_ptr() from the cpuset clone_ns code. + */ +static inline int task_is_waking(struct task_struct *p) +{ + return unlikely((p->state == TASK_WAKING) && !(p->flags & PF_STARTING)); +} + +/* * __task_rq_lock - lock the runqueue a given task resides on. * Must be called interrupts disabled. */ static inline struct rq *__task_rq_lock(struct task_struct *p) __acquires(rq->lock) { + struct rq *rq; + for (;;) { - struct rq *rq = task_rq(p); + while (task_is_waking(p)) + cpu_relax(); + rq = task_rq(p); raw_spin_lock(&rq->lock); - if (likely(rq == task_rq(p))) + if (likely(rq == task_rq(p) && !task_is_waking(p))) return rq; raw_spin_unlock(&rq->lock); } @@ -967,10 +946,12 @@ static struct rq *task_rq_lock(struct task_struct *p, unsigned long *flags) struct rq *rq; for (;;) { + while (task_is_waking(p)) + cpu_relax(); local_irq_save(*flags); rq = task_rq(p); raw_spin_lock(&rq->lock); - if (likely(rq == task_rq(p))) + if (likely(rq == task_rq(p) && !task_is_waking(p))) return rq; raw_spin_unlock_irqrestore(&rq->lock, *flags); } @@ -1390,32 +1371,6 @@ static const u32 prio_to_wmult[40] = { /* 15 */ 119304647, 148102320, 186737708, 238609294, 286331153, }; -static void activate_task(struct rq *rq, struct task_struct *p, int wakeup); - -/* - * runqueue iterator, to support SMP load-balancing between different - * scheduling classes, without having to expose their internal data - * structures to the load-balancing proper: - */ -struct rq_iterator { - void *arg; - struct task_struct *(*start)(void *); - struct task_struct *(*next)(void *); -}; - -#ifdef CONFIG_SMP -static unsigned long -balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, struct sched_domain *sd, - enum cpu_idle_type idle, int *all_pinned, - int *this_best_prio, struct rq_iterator *iterator); - -static int -iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle, - struct rq_iterator *iterator); -#endif - /* Time spent by the tasks of the cpu accounting group executing in ... */ enum cpuacct_stat_index { CPUACCT_STAT_USER, /* ... user mode */ @@ -1531,7 +1486,7 @@ static unsigned long target_load(int cpu, int type) static struct sched_group *group_of(int cpu) { - struct sched_domain *sd = rcu_dereference(cpu_rq(cpu)->sd); + struct sched_domain *sd = rcu_dereference_sched(cpu_rq(cpu)->sd); if (!sd) return NULL; @@ -1701,16 +1656,6 @@ static void update_shares(struct sched_domain *sd) } } -static void update_shares_locked(struct rq *rq, struct sched_domain *sd) -{ - if (root_task_group_empty()) - return; - - raw_spin_unlock(&rq->lock); - update_shares(sd); - raw_spin_lock(&rq->lock); -} - static void update_h_load(long cpu) { if (root_task_group_empty()) @@ -1725,10 +1670,6 @@ static inline void update_shares(struct sched_domain *sd) { } -static inline void update_shares_locked(struct rq *rq, struct sched_domain *sd) -{ -} - #endif #ifdef CONFIG_PREEMPT @@ -1805,6 +1746,51 @@ static inline void double_unlock_balance(struct rq *this_rq, struct rq *busiest) raw_spin_unlock(&busiest->lock); lock_set_subclass(&this_rq->lock.dep_map, 0, _RET_IP_); } + +/* + * double_rq_lock - safely lock two runqueues + * + * Note this does not disable interrupts like task_rq_lock, + * you need to do so manually before calling. + */ +static void double_rq_lock(struct rq *rq1, struct rq *rq2) + __acquires(rq1->lock) + __acquires(rq2->lock) +{ + BUG_ON(!irqs_disabled()); + if (rq1 == rq2) { + raw_spin_lock(&rq1->lock); + __acquire(rq2->lock); /* Fake it out ;) */ + } else { + if (rq1 < rq2) { + raw_spin_lock(&rq1->lock); + raw_spin_lock_nested(&rq2->lock, SINGLE_DEPTH_NESTING); + } else { + raw_spin_lock(&rq2->lock); + raw_spin_lock_nested(&rq1->lock, SINGLE_DEPTH_NESTING); + } + } + update_rq_clock(rq1); + update_rq_clock(rq2); +} + +/* + * double_rq_unlock - safely unlock two runqueues + * + * Note this does not restore interrupts like task_rq_unlock, + * you need to do so manually after calling. + */ +static void double_rq_unlock(struct rq *rq1, struct rq *rq2) + __releases(rq1->lock) + __releases(rq2->lock) +{ + raw_spin_unlock(&rq1->lock); + if (rq1 != rq2) + raw_spin_unlock(&rq2->lock); + else + __release(rq2->lock); +} + #endif #ifdef CONFIG_FAIR_GROUP_SCHED @@ -1834,18 +1820,14 @@ static inline void __set_task_cpu(struct task_struct *p, unsigned int cpu) #endif } -#include "sched_stats.h" -#include "sched_idletask.c" -#include "sched_fair.c" -#include "sched_rt.c" -#ifdef CONFIG_SCHED_DEBUG -# include "sched_debug.c" -#endif +static const struct sched_class rt_sched_class; #define sched_class_highest (&rt_sched_class) #define for_each_class(class) \ for (class = sched_class_highest; class; class = class->next) +#include "sched_stats.h" + static void inc_nr_running(struct rq *rq) { rq->nr_running++; @@ -1883,13 +1865,14 @@ static void update_avg(u64 *avg, u64 sample) *avg += diff >> 3; } -static void enqueue_task(struct rq *rq, struct task_struct *p, int wakeup) +static void +enqueue_task(struct rq *rq, struct task_struct *p, int wakeup, bool head) { if (wakeup) p->se.start_runtime = p->se.sum_exec_runtime; sched_info_queued(p); - p->sched_class->enqueue_task(rq, p, wakeup); + p->sched_class->enqueue_task(rq, p, wakeup, head); p->se.on_rq = 1; } @@ -1912,6 +1895,37 @@ static void dequeue_task(struct rq *rq, struct task_struct *p, int sleep) } /* + * activate_task - move a task to the runqueue. + */ +static void activate_task(struct rq *rq, struct task_struct *p, int wakeup) +{ + if (task_contributes_to_load(p)) + rq->nr_uninterruptible--; + + enqueue_task(rq, p, wakeup, false); + inc_nr_running(rq); +} + +/* + * deactivate_task - remove a task from the runqueue. + */ +static void deactivate_task(struct rq *rq, struct task_struct *p, int sleep) +{ + if (task_contributes_to_load(p)) + rq->nr_uninterruptible++; + + dequeue_task(rq, p, sleep); + dec_nr_running(rq); +} + +#include "sched_idletask.c" +#include "sched_fair.c" +#include "sched_rt.c" +#ifdef CONFIG_SCHED_DEBUG +# include "sched_debug.c" +#endif + +/* * __normal_prio - return the priority that is based on the static prio */ static inline int __normal_prio(struct task_struct *p) @@ -1957,30 +1971,6 @@ static int effective_prio(struct task_struct *p) return p->prio; } -/* - * activate_task - move a task to the runqueue. - */ -static void activate_task(struct rq *rq, struct task_struct *p, int wakeup) -{ - if (task_contributes_to_load(p)) - rq->nr_uninterruptible--; - - enqueue_task(rq, p, wakeup); - inc_nr_running(rq); -} - -/* - * deactivate_task - remove a task from the runqueue. - */ -static void deactivate_task(struct rq *rq, struct task_struct *p, int sleep) -{ - if (task_contributes_to_load(p)) - rq->nr_uninterruptible++; - - dequeue_task(rq, p, sleep); - dec_nr_running(rq); -} - /** * task_curr - is this task currently executing on a CPU? * @p: the task in question. @@ -2408,14 +2398,27 @@ static int try_to_wake_up(struct task_struct *p, unsigned int state, __task_rq_unlock(rq); cpu = select_task_rq(p, SD_BALANCE_WAKE, wake_flags); - if (cpu != orig_cpu) + if (cpu != orig_cpu) { + /* + * Since we migrate the task without holding any rq->lock, + * we need to be careful with task_rq_lock(), since that + * might end up locking an invalid rq. + */ set_task_cpu(p, cpu); + } - rq = __task_rq_lock(p); + rq = cpu_rq(cpu); + raw_spin_lock(&rq->lock); update_rq_clock(rq); + /* + * We migrated the task without holding either rq->lock, however + * since the task is not on the task list itself, nobody else + * will try and migrate the task, hence the rq should match the + * cpu we just moved it to. + */ + WARN_ON(task_cpu(p) != cpu); WARN_ON(p->state != TASK_WAKING); - cpu = task_cpu(p); #ifdef CONFIG_SCHEDSTATS schedstat_inc(rq, ttwu_count); @@ -2663,7 +2666,13 @@ void wake_up_new_task(struct task_struct *p, unsigned long clone_flags) set_task_cpu(p, cpu); #endif - rq = task_rq_lock(p, &flags); + /* + * Since the task is not on the rq and we still have TASK_WAKING set + * nobody else will migrate this task. + */ + rq = cpu_rq(cpu); + raw_spin_lock_irqsave(&rq->lock, flags); + BUG_ON(p->state != TASK_WAKING); p->state = TASK_RUNNING; update_rq_clock(rq); @@ -2794,7 +2803,13 @@ static void finish_task_switch(struct rq *rq, struct task_struct *prev) */ prev_state = prev->state; finish_arch_switch(prev); - perf_event_task_sched_in(current, cpu_of(rq)); +#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW + local_irq_disable(); +#endif /* __ARCH_WANT_INTERRUPTS_ON_CTXSW */ + perf_event_task_sched_in(current); +#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW + local_irq_enable(); +#endif /* __ARCH_WANT_INTERRUPTS_ON_CTXSW */ finish_lock_switch(rq, prev); fire_sched_in_preempt_notifiers(current); @@ -3099,50 +3114,6 @@ static void update_cpu_load(struct rq *this_rq) #ifdef CONFIG_SMP /* - * double_rq_lock - safely lock two runqueues - * - * Note this does not disable interrupts like task_rq_lock, - * you need to do so manually before calling. - */ -static void double_rq_lock(struct rq *rq1, struct rq *rq2) - __acquires(rq1->lock) - __acquires(rq2->lock) -{ - BUG_ON(!irqs_disabled()); - if (rq1 == rq2) { - raw_spin_lock(&rq1->lock); - __acquire(rq2->lock); /* Fake it out ;) */ - } else { - if (rq1 < rq2) { - raw_spin_lock(&rq1->lock); - raw_spin_lock_nested(&rq2->lock, SINGLE_DEPTH_NESTING); - } else { - raw_spin_lock(&rq2->lock); - raw_spin_lock_nested(&rq1->lock, SINGLE_DEPTH_NESTING); - } - } - update_rq_clock(rq1); - update_rq_clock(rq2); -} - -/* - * double_rq_unlock - safely unlock two runqueues - * - * Note this does not restore interrupts like task_rq_unlock, - * you need to do so manually after calling. - */ -static void double_rq_unlock(struct rq *rq1, struct rq *rq2) - __releases(rq1->lock) - __releases(rq2->lock) -{ - raw_spin_unlock(&rq1->lock); - if (rq1 != rq2) - raw_spin_unlock(&rq2->lock); - else - __release(rq2->lock); -} - -/* * sched_exec - execve() is a valuable balancing opportunity, because at * this point the task has the smallest effective memory and cache footprint. */ @@ -3190,1771 +3161,6 @@ again: task_rq_unlock(rq, &flags); } -/* - * pull_task - move a task from a remote runqueue to the local runqueue. - * Both runqueues must be locked. - */ -static void pull_task(struct rq *src_rq, struct task_struct *p, - struct rq *this_rq, int this_cpu) -{ - deactivate_task(src_rq, p, 0); - set_task_cpu(p, this_cpu); - activate_task(this_rq, p, 0); - check_preempt_curr(this_rq, p, 0); -} - -/* - * can_migrate_task - may task p from runqueue rq be migrated to this_cpu? - */ -static -int can_migrate_task(struct task_struct *p, struct rq *rq, int this_cpu, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned) -{ - int tsk_cache_hot = 0; - /* - * We do not migrate tasks that are: - * 1) running (obviously), or - * 2) cannot be migrated to this CPU due to cpus_allowed, or - * 3) are cache-hot on their current CPU. - */ - if (!cpumask_test_cpu(this_cpu, &p->cpus_allowed)) { - schedstat_inc(p, se.nr_failed_migrations_affine); - return 0; - } - *all_pinned = 0; - - if (task_running(rq, p)) { - schedstat_inc(p, se.nr_failed_migrations_running); - return 0; - } - - /* - * Aggressive migration if: - * 1) task is cache cold, or - * 2) too many balance attempts have failed. - */ - - tsk_cache_hot = task_hot(p, rq->clock, sd); - if (!tsk_cache_hot || - sd->nr_balance_failed > sd->cache_nice_tries) { -#ifdef CONFIG_SCHEDSTATS - if (tsk_cache_hot) { - schedstat_inc(sd, lb_hot_gained[idle]); - schedstat_inc(p, se.nr_forced_migrations); - } -#endif - return 1; - } - - if (tsk_cache_hot) { - schedstat_inc(p, se.nr_failed_migrations_hot); - return 0; - } - return 1; -} - -static unsigned long -balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, struct sched_domain *sd, - enum cpu_idle_type idle, int *all_pinned, - int *this_best_prio, struct rq_iterator *iterator) -{ - int loops = 0, pulled = 0, pinned = 0; - struct task_struct *p; - long rem_load_move = max_load_move; - - if (max_load_move == 0) - goto out; - - pinned = 1; - - /* - * Start the load-balancing iterator: - */ - p = iterator->start(iterator->arg); -next: - if (!p || loops++ > sysctl_sched_nr_migrate) - goto out; - - if ((p->se.load.weight >> 1) > rem_load_move || - !can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) { - p = iterator->next(iterator->arg); - goto next; - } - - pull_task(busiest, p, this_rq, this_cpu); - pulled++; - rem_load_move -= p->se.load.weight; - -#ifdef CONFIG_PREEMPT - /* - * NEWIDLE balancing is a source of latency, so preemptible kernels - * will stop after the first task is pulled to minimize the critical - * section. - */ - if (idle == CPU_NEWLY_IDLE) - goto out; -#endif - - /* - * We only want to steal up to the prescribed amount of weighted load. - */ - if (rem_load_move > 0) { - if (p->prio < *this_best_prio) - *this_best_prio = p->prio; - p = iterator->next(iterator->arg); - goto next; - } -out: - /* - * Right now, this is one of only two places pull_task() is called, - * so we can safely collect pull_task() stats here rather than - * inside pull_task(). - */ - schedstat_add(sd, lb_gained[idle], pulled); - - if (all_pinned) - *all_pinned = pinned; - - return max_load_move - rem_load_move; -} - -/* - * move_tasks tries to move up to max_load_move weighted load from busiest to - * this_rq, as part of a balancing operation within domain "sd". - * Returns 1 if successful and 0 otherwise. - * - * Called with both runqueues locked. - */ -static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned) -{ - const struct sched_class *class = sched_class_highest; - unsigned long total_load_moved = 0; - int this_best_prio = this_rq->curr->prio; - - do { - total_load_moved += - class->load_balance(this_rq, this_cpu, busiest, - max_load_move - total_load_moved, - sd, idle, all_pinned, &this_best_prio); - class = class->next; - -#ifdef CONFIG_PREEMPT - /* - * NEWIDLE balancing is a source of latency, so preemptible - * kernels will stop after the first task is pulled to minimize - * the critical section. - */ - if (idle == CPU_NEWLY_IDLE && this_rq->nr_running) - break; -#endif - } while (class && max_load_move > total_load_moved); - - return total_load_moved > 0; -} - -static int -iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle, - struct rq_iterator *iterator) -{ - struct task_struct *p = iterator->start(iterator->arg); - int pinned = 0; - - while (p) { - if (can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) { - pull_task(busiest, p, this_rq, this_cpu); - /* - * Right now, this is only the second place pull_task() - * is called, so we can safely collect pull_task() - * stats here rather than inside pull_task(). - */ - schedstat_inc(sd, lb_gained[idle]); - - return 1; - } - p = iterator->next(iterator->arg); - } - - return 0; -} - -/* - * move_one_task tries to move exactly one task from busiest to this_rq, as - * part of active balancing operations within "domain". - * Returns 1 if successful and 0 otherwise. - * - * Called with both runqueues locked. - */ -static int move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle) -{ - const struct sched_class *class; - - for_each_class(class) { - if (class->move_one_task(this_rq, this_cpu, busiest, sd, idle)) - return 1; - } - - return 0; -} -/********** Helpers for find_busiest_group ************************/ -/* - * sd_lb_stats - Structure to store the statistics of a sched_domain - * during load balancing. - */ -struct sd_lb_stats { - struct sched_group *busiest; /* Busiest group in this sd */ - struct sched_group *this; /* Local group in this sd */ - unsigned long total_load; /* Total load of all groups in sd */ - unsigned long total_pwr; /* Total power of all groups in sd */ - unsigned long avg_load; /* Average load across all groups in sd */ - - /** Statistics of this group */ - unsigned long this_load; - unsigned long this_load_per_task; - unsigned long this_nr_running; - - /* Statistics of the busiest group */ - unsigned long max_load; - unsigned long busiest_load_per_task; - unsigned long busiest_nr_running; - - int group_imb; /* Is there imbalance in this sd */ -#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) - int power_savings_balance; /* Is powersave balance needed for this sd */ - struct sched_group *group_min; /* Least loaded group in sd */ - struct sched_group *group_leader; /* Group which relieves group_min */ - unsigned long min_load_per_task; /* load_per_task in group_min */ - unsigned long leader_nr_running; /* Nr running of group_leader */ - unsigned long min_nr_running; /* Nr running of group_min */ -#endif -}; - -/* - * sg_lb_stats - stats of a sched_group required for load_balancing - */ -struct sg_lb_stats { - unsigned long avg_load; /*Avg load across the CPUs of the group */ - unsigned long group_load; /* Total load over the CPUs of the group */ - unsigned long sum_nr_running; /* Nr tasks running in the group */ - unsigned long sum_weighted_load; /* Weighted load of group's tasks */ - unsigned long group_capacity; - int group_imb; /* Is there an imbalance in the group ? */ -}; - -/** - * group_first_cpu - Returns the first cpu in the cpumask of a sched_group. - * @group: The group whose first cpu is to be returned. - */ -static inline unsigned int group_first_cpu(struct sched_group *group) -{ - return cpumask_first(sched_group_cpus(group)); -} - -/** - * get_sd_load_idx - Obtain the load index for a given sched domain. - * @sd: The sched_domain whose load_idx is to be obtained. - * @idle: The Idle status of the CPU for whose sd load_icx is obtained. - */ -static inline int get_sd_load_idx(struct sched_domain *sd, - enum cpu_idle_type idle) -{ - int load_idx; - - switch (idle) { - case CPU_NOT_IDLE: - load_idx = sd->busy_idx; - break; - - case CPU_NEWLY_IDLE: - load_idx = sd->newidle_idx; - break; - default: - load_idx = sd->idle_idx; - break; - } - - return load_idx; -} - - -#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) -/** - * init_sd_power_savings_stats - Initialize power savings statistics for - * the given sched_domain, during load balancing. - * - * @sd: Sched domain whose power-savings statistics are to be initialized. - * @sds: Variable containing the statistics for sd. - * @idle: Idle status of the CPU at which we're performing load-balancing. - */ -static inline void init_sd_power_savings_stats(struct sched_domain *sd, - struct sd_lb_stats *sds, enum cpu_idle_type idle) -{ - /* - * Busy processors will not participate in power savings - * balance. - */ - if (idle == CPU_NOT_IDLE || !(sd->flags & SD_POWERSAVINGS_BALANCE)) - sds->power_savings_balance = 0; - else { - sds->power_savings_balance = 1; - sds->min_nr_running = ULONG_MAX; - sds->leader_nr_running = 0; - } -} - -/** - * update_sd_power_savings_stats - Update the power saving stats for a - * sched_domain while performing load balancing. - * - * @group: sched_group belonging to the sched_domain under consideration. - * @sds: Variable containing the statistics of the sched_domain - * @local_group: Does group contain the CPU for which we're performing - * load balancing ? - * @sgs: Variable containing the statistics of the group. - */ -static inline void update_sd_power_savings_stats(struct sched_group *group, - struct sd_lb_stats *sds, int local_group, struct sg_lb_stats *sgs) -{ - - if (!sds->power_savings_balance) - return; - - /* - * If the local group is idle or completely loaded - * no need to do power savings balance at this domain - */ - if (local_group && (sds->this_nr_running >= sgs->group_capacity || - !sds->this_nr_running)) - sds->power_savings_balance = 0; - - /* - * If a group is already running at full capacity or idle, - * don't include that group in power savings calculations - */ - if (!sds->power_savings_balance || - sgs->sum_nr_running >= sgs->group_capacity || - !sgs->sum_nr_running) - return; - - /* - * Calculate the group which has the least non-idle load. - * This is the group from where we need to pick up the load - * for saving power - */ - if ((sgs->sum_nr_running < sds->min_nr_running) || - (sgs->sum_nr_running == sds->min_nr_running && - group_first_cpu(group) > group_first_cpu(sds->group_min))) { - sds->group_min = group; - sds->min_nr_running = sgs->sum_nr_running; - sds->min_load_per_task = sgs->sum_weighted_load / - sgs->sum_nr_running; - } - - /* - * Calculate the group which is almost near its - * capacity but still has some space to pick up some load - * from other group and save more power - */ - if (sgs->sum_nr_running + 1 > sgs->group_capacity) - return; - - if (sgs->sum_nr_running > sds->leader_nr_running || - (sgs->sum_nr_running == sds->leader_nr_running && - group_first_cpu(group) < group_first_cpu(sds->group_leader))) { - sds->group_leader = group; - sds->leader_nr_running = sgs->sum_nr_running; - } -} - -/** - * check_power_save_busiest_group - see if there is potential for some power-savings balance - * @sds: Variable containing the statistics of the sched_domain - * under consideration. - * @this_cpu: Cpu at which we're currently performing load-balancing. - * @imbalance: Variable to store the imbalance. - * - * Description: - * Check if we have potential to perform some power-savings balance. - * If yes, set the busiest group to be the least loaded group in the - * sched_domain, so that it's CPUs can be put to idle. - * - * Returns 1 if there is potential to perform power-savings balance. - * Else returns 0. - */ -static inline int check_power_save_busiest_group(struct sd_lb_stats *sds, - int this_cpu, unsigned long *imbalance) -{ - if (!sds->power_savings_balance) - return 0; - - if (sds->this != sds->group_leader || - sds->group_leader == sds->group_min) - return 0; - - *imbalance = sds->min_load_per_task; - sds->busiest = sds->group_min; - - return 1; - -} -#else /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ -static inline void init_sd_power_savings_stats(struct sched_domain *sd, - struct sd_lb_stats *sds, enum cpu_idle_type idle) -{ - return; -} - -static inline void update_sd_power_savings_stats(struct sched_group *group, - struct sd_lb_stats *sds, int local_group, struct sg_lb_stats *sgs) -{ - return; -} - -static inline int check_power_save_busiest_group(struct sd_lb_stats *sds, - int this_cpu, unsigned long *imbalance) -{ - return 0; -} -#endif /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ - - -unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu) -{ - return SCHED_LOAD_SCALE; -} - -unsigned long __weak arch_scale_freq_power(struct sched_domain *sd, int cpu) -{ - return default_scale_freq_power(sd, cpu); -} - -unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu) -{ - unsigned long weight = cpumask_weight(sched_domain_span(sd)); - unsigned long smt_gain = sd->smt_gain; - - smt_gain /= weight; - - return smt_gain; -} - -unsigned long __weak arch_scale_smt_power(struct sched_domain *sd, int cpu) -{ - return default_scale_smt_power(sd, cpu); -} - -unsigned long scale_rt_power(int cpu) -{ - struct rq *rq = cpu_rq(cpu); - u64 total, available; - - sched_avg_update(rq); - - total = sched_avg_period() + (rq->clock - rq->age_stamp); - available = total - rq->rt_avg; - - if (unlikely((s64)total < SCHED_LOAD_SCALE)) - total = SCHED_LOAD_SCALE; - - total >>= SCHED_LOAD_SHIFT; - - return div_u64(available, total); -} - -static void update_cpu_power(struct sched_domain *sd, int cpu) -{ - unsigned long weight = cpumask_weight(sched_domain_span(sd)); - unsigned long power = SCHED_LOAD_SCALE; - struct sched_group *sdg = sd->groups; - - if (sched_feat(ARCH_POWER)) - power *= arch_scale_freq_power(sd, cpu); - else - power *= default_scale_freq_power(sd, cpu); - - power >>= SCHED_LOAD_SHIFT; - - if ((sd->flags & SD_SHARE_CPUPOWER) && weight > 1) { - if (sched_feat(ARCH_POWER)) - power *= arch_scale_smt_power(sd, cpu); - else - power *= default_scale_smt_power(sd, cpu); - - power >>= SCHED_LOAD_SHIFT; - } - - power *= scale_rt_power(cpu); - power >>= SCHED_LOAD_SHIFT; - - if (!power) - power = 1; - - sdg->cpu_power = power; -} - -static void update_group_power(struct sched_domain *sd, int cpu) -{ - struct sched_domain *child = sd->child; - struct sched_group *group, *sdg = sd->groups; - unsigned long power; - - if (!child) { - update_cpu_power(sd, cpu); - return; - } - - power = 0; - - group = child->groups; - do { - power += group->cpu_power; - group = group->next; - } while (group != child->groups); - - sdg->cpu_power = power; -} - -/** - * update_sg_lb_stats - Update sched_group's statistics for load balancing. - * @sd: The sched_domain whose statistics are to be updated. - * @group: sched_group whose statistics are to be updated. - * @this_cpu: Cpu for which load balance is currently performed. - * @idle: Idle status of this_cpu - * @load_idx: Load index of sched_domain of this_cpu for load calc. - * @sd_idle: Idle status of the sched_domain containing group. - * @local_group: Does group contain this_cpu. - * @cpus: Set of cpus considered for load balancing. - * @balance: Should we balance. - * @sgs: variable to hold the statistics for this group. - */ -static inline void update_sg_lb_stats(struct sched_domain *sd, - struct sched_group *group, int this_cpu, - enum cpu_idle_type idle, int load_idx, int *sd_idle, - int local_group, const struct cpumask *cpus, - int *balance, struct sg_lb_stats *sgs) -{ - unsigned long load, max_cpu_load, min_cpu_load; - int i; - unsigned int balance_cpu = -1, first_idle_cpu = 0; - unsigned long sum_avg_load_per_task; - unsigned long avg_load_per_task; - - if (local_group) { - balance_cpu = group_first_cpu(group); - if (balance_cpu == this_cpu) - update_group_power(sd, this_cpu); - } - - /* Tally up the load of all CPUs in the group */ - sum_avg_load_per_task = avg_load_per_task = 0; - max_cpu_load = 0; - min_cpu_load = ~0UL; - - for_each_cpu_and(i, sched_group_cpus(group), cpus) { - struct rq *rq = cpu_rq(i); - - if (*sd_idle && rq->nr_running) - *sd_idle = 0; - - /* Bias balancing toward cpus of our domain */ - if (local_group) { - if (idle_cpu(i) && !first_idle_cpu) { - first_idle_cpu = 1; - balance_cpu = i; - } - - load = target_load(i, load_idx); - } else { - load = source_load(i, load_idx); - if (load > max_cpu_load) - max_cpu_load = load; - if (min_cpu_load > load) - min_cpu_load = load; - } - - sgs->group_load += load; - sgs->sum_nr_running += rq->nr_running; - sgs->sum_weighted_load += weighted_cpuload(i); - - sum_avg_load_per_task += cpu_avg_load_per_task(i); - } - - /* - * First idle cpu or the first cpu(busiest) in this sched group - * is eligible for doing load balancing at this and above - * domains. In the newly idle case, we will allow all the cpu's - * to do the newly idle load balance. - */ - if (idle != CPU_NEWLY_IDLE && local_group && - balance_cpu != this_cpu && balance) { - *balance = 0; - return; - } - - /* Adjust by relative CPU power of the group */ - sgs->avg_load = (sgs->group_load * SCHED_LOAD_SCALE) / group->cpu_power; - - - /* - * Consider the group unbalanced when the imbalance is larger - * than the average weight of two tasks. - * - * APZ: with cgroup the avg task weight can vary wildly and - * might not be a suitable number - should we keep a - * normalized nr_running number somewhere that negates - * the hierarchy? - */ - avg_load_per_task = (sum_avg_load_per_task * SCHED_LOAD_SCALE) / - group->cpu_power; - - if ((max_cpu_load - min_cpu_load) > 2*avg_load_per_task) - sgs->group_imb = 1; - - sgs->group_capacity = - DIV_ROUND_CLOSEST(group->cpu_power, SCHED_LOAD_SCALE); -} - -/** - * update_sd_lb_stats - Update sched_group's statistics for load balancing. - * @sd: sched_domain whose statistics are to be updated. - * @this_cpu: Cpu for which load balance is currently performed. - * @idle: Idle status of this_cpu - * @sd_idle: Idle status of the sched_domain containing group. - * @cpus: Set of cpus considered for load balancing. - * @balance: Should we balance. - * @sds: variable to hold the statistics for this sched_domain. - */ -static inline void update_sd_lb_stats(struct sched_domain *sd, int this_cpu, - enum cpu_idle_type idle, int *sd_idle, - const struct cpumask *cpus, int *balance, - struct sd_lb_stats *sds) -{ - struct sched_domain *child = sd->child; - struct sched_group *group = sd->groups; - struct sg_lb_stats sgs; - int load_idx, prefer_sibling = 0; - - if (child && child->flags & SD_PREFER_SIBLING) - prefer_sibling = 1; - - init_sd_power_savings_stats(sd, sds, idle); - load_idx = get_sd_load_idx(sd, idle); - - do { - int local_group; - - local_group = cpumask_test_cpu(this_cpu, - sched_group_cpus(group)); - memset(&sgs, 0, sizeof(sgs)); - update_sg_lb_stats(sd, group, this_cpu, idle, load_idx, sd_idle, - local_group, cpus, balance, &sgs); - - if (local_group && balance && !(*balance)) - return; - - sds->total_load += sgs.group_load; - sds->total_pwr += group->cpu_power; - - /* - * In case the child domain prefers tasks go to siblings - * first, lower the group capacity to one so that we'll try - * and move all the excess tasks away. - */ - if (prefer_sibling) - sgs.group_capacity = min(sgs.group_capacity, 1UL); - - if (local_group) { - sds->this_load = sgs.avg_load; - sds->this = group; - sds->this_nr_running = sgs.sum_nr_running; - sds->this_load_per_task = sgs.sum_weighted_load; - } else if (sgs.avg_load > sds->max_load && - (sgs.sum_nr_running > sgs.group_capacity || - sgs.group_imb)) { - sds->max_load = sgs.avg_load; - sds->busiest = group; - sds->busiest_nr_running = sgs.sum_nr_running; - sds->busiest_load_per_task = sgs.sum_weighted_load; - sds->group_imb = sgs.group_imb; - } - - update_sd_power_savings_stats(group, sds, local_group, &sgs); - group = group->next; - } while (group != sd->groups); -} - -/** - * fix_small_imbalance - Calculate the minor imbalance that exists - * amongst the groups of a sched_domain, during - * load balancing. - * @sds: Statistics of the sched_domain whose imbalance is to be calculated. - * @this_cpu: The cpu at whose sched_domain we're performing load-balance. - * @imbalance: Variable to store the imbalance. - */ -static inline void fix_small_imbalance(struct sd_lb_stats *sds, - int this_cpu, unsigned long *imbalance) -{ - unsigned long tmp, pwr_now = 0, pwr_move = 0; - unsigned int imbn = 2; - - if (sds->this_nr_running) { - sds->this_load_per_task /= sds->this_nr_running; - if (sds->busiest_load_per_task > - sds->this_load_per_task) - imbn = 1; - } else - sds->this_load_per_task = - cpu_avg_load_per_task(this_cpu); - - if (sds->max_load - sds->this_load + sds->busiest_load_per_task >= - sds->busiest_load_per_task * imbn) { - *imbalance = sds->busiest_load_per_task; - return; - } - - /* - * OK, we don't have enough imbalance to justify moving tasks, - * however we may be able to increase total CPU power used by - * moving them. - */ - - pwr_now += sds->busiest->cpu_power * - min(sds->busiest_load_per_task, sds->max_load); - pwr_now += sds->this->cpu_power * - min(sds->this_load_per_task, sds->this_load); - pwr_now /= SCHED_LOAD_SCALE; - - /* Amount of load we'd subtract */ - tmp = (sds->busiest_load_per_task * SCHED_LOAD_SCALE) / - sds->busiest->cpu_power; - if (sds->max_load > tmp) - pwr_move += sds->busiest->cpu_power * - min(sds->busiest_load_per_task, sds->max_load - tmp); - - /* Amount of load we'd add */ - if (sds->max_load * sds->busiest->cpu_power < - sds->busiest_load_per_task * SCHED_LOAD_SCALE) - tmp = (sds->max_load * sds->busiest->cpu_power) / - sds->this->cpu_power; - else - tmp = (sds->busiest_load_per_task * SCHED_LOAD_SCALE) / - sds->this->cpu_power; - pwr_move += sds->this->cpu_power * - min(sds->this_load_per_task, sds->this_load + tmp); - pwr_move /= SCHED_LOAD_SCALE; - - /* Move if we gain throughput */ - if (pwr_move > pwr_now) - *imbalance = sds->busiest_load_per_task; -} - -/** - * calculate_imbalance - Calculate the amount of imbalance present within the - * groups of a given sched_domain during load balance. - * @sds: statistics of the sched_domain whose imbalance is to be calculated. - * @this_cpu: Cpu for which currently load balance is being performed. - * @imbalance: The variable to store the imbalance. - */ -static inline void calculate_imbalance(struct sd_lb_stats *sds, int this_cpu, - unsigned long *imbalance) -{ - unsigned long max_pull; - /* - * In the presence of smp nice balancing, certain scenarios can have - * max load less than avg load(as we skip the groups at or below - * its cpu_power, while calculating max_load..) - */ - if (sds->max_load < sds->avg_load) { - *imbalance = 0; - return fix_small_imbalance(sds, this_cpu, imbalance); - } - - /* Don't want to pull so many tasks that a group would go idle */ - max_pull = min(sds->max_load - sds->avg_load, - sds->max_load - sds->busiest_load_per_task); - - /* How much load to actually move to equalise the imbalance */ - *imbalance = min(max_pull * sds->busiest->cpu_power, - (sds->avg_load - sds->this_load) * sds->this->cpu_power) - / SCHED_LOAD_SCALE; - - /* - * if *imbalance is less than the average load per runnable task - * there is no gaurantee that any tasks will be moved so we'll have - * a think about bumping its value to force at least one task to be - * moved - */ - if (*imbalance < sds->busiest_load_per_task) - return fix_small_imbalance(sds, this_cpu, imbalance); - -} -/******* find_busiest_group() helpers end here *********************/ - -/** - * find_busiest_group - Returns the busiest group within the sched_domain - * if there is an imbalance. If there isn't an imbalance, and - * the user has opted for power-savings, it returns a group whose - * CPUs can be put to idle by rebalancing those tasks elsewhere, if - * such a group exists. - * - * Also calculates the amount of weighted load which should be moved - * to restore balance. - * - * @sd: The sched_domain whose busiest group is to be returned. - * @this_cpu: The cpu for which load balancing is currently being performed. - * @imbalance: Variable which stores amount of weighted load which should - * be moved to restore balance/put a group to idle. - * @idle: The idle status of this_cpu. - * @sd_idle: The idleness of sd - * @cpus: The set of CPUs under consideration for load-balancing. - * @balance: Pointer to a variable indicating if this_cpu - * is the appropriate cpu to perform load balancing at this_level. - * - * Returns: - the busiest group if imbalance exists. - * - If no imbalance and user has opted for power-savings balance, - * return the least loaded group whose CPUs can be - * put to idle by rebalancing its tasks onto our group. - */ -static struct sched_group * -find_busiest_group(struct sched_domain *sd, int this_cpu, - unsigned long *imbalance, enum cpu_idle_type idle, - int *sd_idle, const struct cpumask *cpus, int *balance) -{ - struct sd_lb_stats sds; - - memset(&sds, 0, sizeof(sds)); - - /* - * Compute the various statistics relavent for load balancing at - * this level. - */ - update_sd_lb_stats(sd, this_cpu, idle, sd_idle, cpus, - balance, &sds); - - /* Cases where imbalance does not exist from POV of this_cpu */ - /* 1) this_cpu is not the appropriate cpu to perform load balancing - * at this level. - * 2) There is no busy sibling group to pull from. - * 3) This group is the busiest group. - * 4) This group is more busy than the avg busieness at this - * sched_domain. - * 5) The imbalance is within the specified limit. - * 6) Any rebalance would lead to ping-pong - */ - if (balance && !(*balance)) - goto ret; - - if (!sds.busiest || sds.busiest_nr_running == 0) - goto out_balanced; - - if (sds.this_load >= sds.max_load) - goto out_balanced; - - sds.avg_load = (SCHED_LOAD_SCALE * sds.total_load) / sds.total_pwr; - - if (sds.this_load >= sds.avg_load) - goto out_balanced; - - if (100 * sds.max_load <= sd->imbalance_pct * sds.this_load) - goto out_balanced; - - sds.busiest_load_per_task /= sds.busiest_nr_running; - if (sds.group_imb) - sds.busiest_load_per_task = - min(sds.busiest_load_per_task, sds.avg_load); - - /* - * We're trying to get all the cpus to the average_load, so we don't - * want to push ourselves above the average load, nor do we wish to - * reduce the max loaded cpu below the average load, as either of these - * actions would just result in more rebalancing later, and ping-pong - * tasks around. Thus we look for the minimum possible imbalance. - * Negative imbalances (*we* are more loaded than anyone else) will - * be counted as no imbalance for these purposes -- we can't fix that - * by pulling tasks to us. Be careful of negative numbers as they'll - * appear as very large values with unsigned longs. - */ - if (sds.max_load <= sds.busiest_load_per_task) - goto out_balanced; - - /* Looks like there is an imbalance. Compute it */ - calculate_imbalance(&sds, this_cpu, imbalance); - return sds.busiest; - -out_balanced: - /* - * There is no obvious imbalance. But check if we can do some balancing - * to save power. - */ - if (check_power_save_busiest_group(&sds, this_cpu, imbalance)) - return sds.busiest; -ret: - *imbalance = 0; - return NULL; -} - -/* - * find_busiest_queue - find the busiest runqueue among the cpus in group. - */ -static struct rq * -find_busiest_queue(struct sched_group *group, enum cpu_idle_type idle, - unsigned long imbalance, const struct cpumask *cpus) -{ - struct rq *busiest = NULL, *rq; - unsigned long max_load = 0; - int i; - - for_each_cpu(i, sched_group_cpus(group)) { - unsigned long power = power_of(i); - unsigned long capacity = DIV_ROUND_CLOSEST(power, SCHED_LOAD_SCALE); - unsigned long wl; - - if (!cpumask_test_cpu(i, cpus)) - continue; - - rq = cpu_rq(i); - wl = weighted_cpuload(i) * SCHED_LOAD_SCALE; - wl /= power; - - if (capacity && rq->nr_running == 1 && wl > imbalance) - continue; - - if (wl > max_load) { - max_load = wl; - busiest = rq; - } - } - - return busiest; -} - -/* - * Max backoff if we encounter pinned tasks. Pretty arbitrary value, but - * so long as it is large enough. - */ -#define MAX_PINNED_INTERVAL 512 - -/* Working cpumask for load_balance and load_balance_newidle. */ -static DEFINE_PER_CPU(cpumask_var_t, load_balance_tmpmask); - -/* - * Check this_cpu to ensure it is balanced within domain. Attempt to move - * tasks if there is an imbalance. - */ -static int load_balance(int this_cpu, struct rq *this_rq, - struct sched_domain *sd, enum cpu_idle_type idle, - int *balance) -{ - int ld_moved, all_pinned = 0, active_balance = 0, sd_idle = 0; - struct sched_group *group; - unsigned long imbalance; - struct rq *busiest; - unsigned long flags; - struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask); - - cpumask_copy(cpus, cpu_active_mask); - - /* - * When power savings policy is enabled for the parent domain, idle - * sibling can pick up load irrespective of busy siblings. In this case, - * let the state of idle sibling percolate up as CPU_IDLE, instead of - * portraying it as CPU_NOT_IDLE. - */ - if (idle != CPU_NOT_IDLE && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - sd_idle = 1; - - schedstat_inc(sd, lb_count[idle]); - -redo: - update_shares(sd); - group = find_busiest_group(sd, this_cpu, &imbalance, idle, &sd_idle, - cpus, balance); - - if (*balance == 0) - goto out_balanced; - - if (!group) { - schedstat_inc(sd, lb_nobusyg[idle]); - goto out_balanced; - } - - busiest = find_busiest_queue(group, idle, imbalance, cpus); - if (!busiest) { - schedstat_inc(sd, lb_nobusyq[idle]); - goto out_balanced; - } - - BUG_ON(busiest == this_rq); - - schedstat_add(sd, lb_imbalance[idle], imbalance); - - ld_moved = 0; - if (busiest->nr_running > 1) { - /* - * Attempt to move tasks. If find_busiest_group has found - * an imbalance but busiest->nr_running <= 1, the group is - * still unbalanced. ld_moved simply stays zero, so it is - * correctly treated as an imbalance. - */ - local_irq_save(flags); - double_rq_lock(this_rq, busiest); - ld_moved = move_tasks(this_rq, this_cpu, busiest, - imbalance, sd, idle, &all_pinned); - double_rq_unlock(this_rq, busiest); - local_irq_restore(flags); - - /* - * some other cpu did the load balance for us. - */ - if (ld_moved && this_cpu != smp_processor_id()) - resched_cpu(this_cpu); - - /* All tasks on this runqueue were pinned by CPU affinity */ - if (unlikely(all_pinned)) { - cpumask_clear_cpu(cpu_of(busiest), cpus); - if (!cpumask_empty(cpus)) - goto redo; - goto out_balanced; - } - } - - if (!ld_moved) { - schedstat_inc(sd, lb_failed[idle]); - sd->nr_balance_failed++; - - if (unlikely(sd->nr_balance_failed > sd->cache_nice_tries+2)) { - - raw_spin_lock_irqsave(&busiest->lock, flags); - - /* don't kick the migration_thread, if the curr - * task on busiest cpu can't be moved to this_cpu - */ - if (!cpumask_test_cpu(this_cpu, - &busiest->curr->cpus_allowed)) { - raw_spin_unlock_irqrestore(&busiest->lock, - flags); - all_pinned = 1; - goto out_one_pinned; - } - - if (!busiest->active_balance) { - busiest->active_balance = 1; - busiest->push_cpu = this_cpu; - active_balance = 1; - } - raw_spin_unlock_irqrestore(&busiest->lock, flags); - if (active_balance) - wake_up_process(busiest->migration_thread); - - /* - * We've kicked active balancing, reset the failure - * counter. - */ - sd->nr_balance_failed = sd->cache_nice_tries+1; - } - } else - sd->nr_balance_failed = 0; - - if (likely(!active_balance)) { - /* We were unbalanced, so reset the balancing interval */ - sd->balance_interval = sd->min_interval; - } else { - /* - * If we've begun active balancing, start to back off. This - * case may not be covered by the all_pinned logic if there - * is only 1 task on the busy runqueue (because we don't call - * move_tasks). - */ - if (sd->balance_interval < sd->max_interval) - sd->balance_interval *= 2; - } - - if (!ld_moved && !sd_idle && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - ld_moved = -1; - - goto out; - -out_balanced: - schedstat_inc(sd, lb_balanced[idle]); - - sd->nr_balance_failed = 0; - -out_one_pinned: - /* tune up the balancing interval */ - if ((all_pinned && sd->balance_interval < MAX_PINNED_INTERVAL) || - (sd->balance_interval < sd->max_interval)) - sd->balance_interval *= 2; - - if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - ld_moved = -1; - else - ld_moved = 0; -out: - if (ld_moved) - update_shares(sd); - return ld_moved; -} - -/* - * Check this_cpu to ensure it is balanced within domain. Attempt to move - * tasks if there is an imbalance. - * - * Called from schedule when this_rq is about to become idle (CPU_NEWLY_IDLE). - * this_rq is locked. - */ -static int -load_balance_newidle(int this_cpu, struct rq *this_rq, struct sched_domain *sd) -{ - struct sched_group *group; - struct rq *busiest = NULL; - unsigned long imbalance; - int ld_moved = 0; - int sd_idle = 0; - int all_pinned = 0; - struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask); - - cpumask_copy(cpus, cpu_active_mask); - - /* - * When power savings policy is enabled for the parent domain, idle - * sibling can pick up load irrespective of busy siblings. In this case, - * let the state of idle sibling percolate up as IDLE, instead of - * portraying it as CPU_NOT_IDLE. - */ - if (sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - sd_idle = 1; - - schedstat_inc(sd, lb_count[CPU_NEWLY_IDLE]); -redo: - update_shares_locked(this_rq, sd); - group = find_busiest_group(sd, this_cpu, &imbalance, CPU_NEWLY_IDLE, - &sd_idle, cpus, NULL); - if (!group) { - schedstat_inc(sd, lb_nobusyg[CPU_NEWLY_IDLE]); - goto out_balanced; - } - - busiest = find_busiest_queue(group, CPU_NEWLY_IDLE, imbalance, cpus); - if (!busiest) { - schedstat_inc(sd, lb_nobusyq[CPU_NEWLY_IDLE]); - goto out_balanced; - } - - BUG_ON(busiest == this_rq); - - schedstat_add(sd, lb_imbalance[CPU_NEWLY_IDLE], imbalance); - - ld_moved = 0; - if (busiest->nr_running > 1) { - /* Attempt to move tasks */ - double_lock_balance(this_rq, busiest); - /* this_rq->clock is already updated */ - update_rq_clock(busiest); - ld_moved = move_tasks(this_rq, this_cpu, busiest, - imbalance, sd, CPU_NEWLY_IDLE, - &all_pinned); - double_unlock_balance(this_rq, busiest); - - if (unlikely(all_pinned)) { - cpumask_clear_cpu(cpu_of(busiest), cpus); - if (!cpumask_empty(cpus)) - goto redo; - } - } - - if (!ld_moved) { - int active_balance = 0; - - schedstat_inc(sd, lb_failed[CPU_NEWLY_IDLE]); - if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - return -1; - - if (sched_mc_power_savings < POWERSAVINGS_BALANCE_WAKEUP) - return -1; - - if (sd->nr_balance_failed++ < 2) - return -1; - - /* - * The only task running in a non-idle cpu can be moved to this - * cpu in an attempt to completely freeup the other CPU - * package. The same method used to move task in load_balance() - * have been extended for load_balance_newidle() to speedup - * consolidation at sched_mc=POWERSAVINGS_BALANCE_WAKEUP (2) - * - * The package power saving logic comes from - * find_busiest_group(). If there are no imbalance, then - * f_b_g() will return NULL. However when sched_mc={1,2} then - * f_b_g() will select a group from which a running task may be - * pulled to this cpu in order to make the other package idle. - * If there is no opportunity to make a package idle and if - * there are no imbalance, then f_b_g() will return NULL and no - * action will be taken in load_balance_newidle(). - * - * Under normal task pull operation due to imbalance, there - * will be more than one task in the source run queue and - * move_tasks() will succeed. ld_moved will be true and this - * active balance code will not be triggered. - */ - - /* Lock busiest in correct order while this_rq is held */ - double_lock_balance(this_rq, busiest); - - /* - * don't kick the migration_thread, if the curr - * task on busiest cpu can't be moved to this_cpu - */ - if (!cpumask_test_cpu(this_cpu, &busiest->curr->cpus_allowed)) { - double_unlock_balance(this_rq, busiest); - all_pinned = 1; - return ld_moved; - } - - if (!busiest->active_balance) { - busiest->active_balance = 1; - busiest->push_cpu = this_cpu; - active_balance = 1; - } - - double_unlock_balance(this_rq, busiest); - /* - * Should not call ttwu while holding a rq->lock - */ - raw_spin_unlock(&this_rq->lock); - if (active_balance) - wake_up_process(busiest->migration_thread); - raw_spin_lock(&this_rq->lock); - - } else - sd->nr_balance_failed = 0; - - update_shares_locked(this_rq, sd); - return ld_moved; - -out_balanced: - schedstat_inc(sd, lb_balanced[CPU_NEWLY_IDLE]); - if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && - !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) - return -1; - sd->nr_balance_failed = 0; - - return 0; -} - -/* - * idle_balance is called by schedule() if this_cpu is about to become - * idle. Attempts to pull tasks from other CPUs. - */ -static void idle_balance(int this_cpu, struct rq *this_rq) -{ - struct sched_domain *sd; - int pulled_task = 0; - unsigned long next_balance = jiffies + HZ; - - this_rq->idle_stamp = this_rq->clock; - - if (this_rq->avg_idle < sysctl_sched_migration_cost) - return; - - for_each_domain(this_cpu, sd) { - unsigned long interval; - - if (!(sd->flags & SD_LOAD_BALANCE)) - continue; - - if (sd->flags & SD_BALANCE_NEWIDLE) - /* If we've pulled tasks over stop searching: */ - pulled_task = load_balance_newidle(this_cpu, this_rq, - sd); - - interval = msecs_to_jiffies(sd->balance_interval); - if (time_after(next_balance, sd->last_balance + interval)) - next_balance = sd->last_balance + interval; - if (pulled_task) { - this_rq->idle_stamp = 0; - break; - } - } - if (pulled_task || time_after(jiffies, this_rq->next_balance)) { - /* - * We are going idle. next_balance may be set based on - * a busy processor. So reset next_balance. - */ - this_rq->next_balance = next_balance; - } -} - -/* - * active_load_balance is run by migration threads. It pushes running tasks - * off the busiest CPU onto idle CPUs. It requires at least 1 task to be - * running on each physical CPU where possible, and avoids physical / - * logical imbalances. - * - * Called with busiest_rq locked. - */ -static void active_load_balance(struct rq *busiest_rq, int busiest_cpu) -{ - int target_cpu = busiest_rq->push_cpu; - struct sched_domain *sd; - struct rq *target_rq; - - /* Is there any task to move? */ - if (busiest_rq->nr_running <= 1) - return; - - target_rq = cpu_rq(target_cpu); - - /* - * This condition is "impossible", if it occurs - * we need to fix it. Originally reported by - * Bjorn Helgaas on a 128-cpu setup. - */ - BUG_ON(busiest_rq == target_rq); - - /* move a task from busiest_rq to target_rq */ - double_lock_balance(busiest_rq, target_rq); - update_rq_clock(busiest_rq); - update_rq_clock(target_rq); - - /* Search for an sd spanning us and the target CPU. */ - for_each_domain(target_cpu, sd) { - if ((sd->flags & SD_LOAD_BALANCE) && - cpumask_test_cpu(busiest_cpu, sched_domain_span(sd))) - break; - } - - if (likely(sd)) { - schedstat_inc(sd, alb_count); - - if (move_one_task(target_rq, target_cpu, busiest_rq, - sd, CPU_IDLE)) - schedstat_inc(sd, alb_pushed); - else - schedstat_inc(sd, alb_failed); - } - double_unlock_balance(busiest_rq, target_rq); -} - -#ifdef CONFIG_NO_HZ -static struct { - atomic_t load_balancer; - cpumask_var_t cpu_mask; - cpumask_var_t ilb_grp_nohz_mask; -} nohz ____cacheline_aligned = { - .load_balancer = ATOMIC_INIT(-1), -}; - -int get_nohz_load_balancer(void) -{ - return atomic_read(&nohz.load_balancer); -} - -#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) -/** - * lowest_flag_domain - Return lowest sched_domain containing flag. - * @cpu: The cpu whose lowest level of sched domain is to - * be returned. - * @flag: The flag to check for the lowest sched_domain - * for the given cpu. - * - * Returns the lowest sched_domain of a cpu which contains the given flag. - */ -static inline struct sched_domain *lowest_flag_domain(int cpu, int flag) -{ - struct sched_domain *sd; - - for_each_domain(cpu, sd) - if (sd && (sd->flags & flag)) - break; - - return sd; -} - -/** - * for_each_flag_domain - Iterates over sched_domains containing the flag. - * @cpu: The cpu whose domains we're iterating over. - * @sd: variable holding the value of the power_savings_sd - * for cpu. - * @flag: The flag to filter the sched_domains to be iterated. - * - * Iterates over all the scheduler domains for a given cpu that has the 'flag' - * set, starting from the lowest sched_domain to the highest. - */ -#define for_each_flag_domain(cpu, sd, flag) \ - for (sd = lowest_flag_domain(cpu, flag); \ - (sd && (sd->flags & flag)); sd = sd->parent) - -/** - * is_semi_idle_group - Checks if the given sched_group is semi-idle. - * @ilb_group: group to be checked for semi-idleness - * - * Returns: 1 if the group is semi-idle. 0 otherwise. - * - * We define a sched_group to be semi idle if it has atleast one idle-CPU - * and atleast one non-idle CPU. This helper function checks if the given - * sched_group is semi-idle or not. - */ -static inline int is_semi_idle_group(struct sched_group *ilb_group) -{ - cpumask_and(nohz.ilb_grp_nohz_mask, nohz.cpu_mask, - sched_group_cpus(ilb_group)); - - /* - * A sched_group is semi-idle when it has atleast one busy cpu - * and atleast one idle cpu. - */ - if (cpumask_empty(nohz.ilb_grp_nohz_mask)) - return 0; - - if (cpumask_equal(nohz.ilb_grp_nohz_mask, sched_group_cpus(ilb_group))) - return 0; - - return 1; -} -/** - * find_new_ilb - Finds the optimum idle load balancer for nomination. - * @cpu: The cpu which is nominating a new idle_load_balancer. - * - * Returns: Returns the id of the idle load balancer if it exists, - * Else, returns >= nr_cpu_ids. - * - * This algorithm picks the idle load balancer such that it belongs to a - * semi-idle powersavings sched_domain. The idea is to try and avoid - * completely idle packages/cores just for the purpose of idle load balancing - * when there are other idle cpu's which are better suited for that job. - */ -static int find_new_ilb(int cpu) -{ - struct sched_domain *sd; - struct sched_group *ilb_group; - - /* - * Have idle load balancer selection from semi-idle packages only - * when power-aware load balancing is enabled - */ - if (!(sched_smt_power_savings || sched_mc_power_savings)) - goto out_done; - - /* - * Optimize for the case when we have no idle CPUs or only one - * idle CPU. Don't walk the sched_domain hierarchy in such cases - */ - if (cpumask_weight(nohz.cpu_mask) < 2) - goto out_done; - - for_each_flag_domain(cpu, sd, SD_POWERSAVINGS_BALANCE) { - ilb_group = sd->groups; - - do { - if (is_semi_idle_group(ilb_group)) - return cpumask_first(nohz.ilb_grp_nohz_mask); - - ilb_group = ilb_group->next; - - } while (ilb_group != sd->groups); - } - -out_done: - return cpumask_first(nohz.cpu_mask); -} -#else /* (CONFIG_SCHED_MC || CONFIG_SCHED_SMT) */ -static inline int find_new_ilb(int call_cpu) -{ - return cpumask_first(nohz.cpu_mask); -} -#endif - -/* - * This routine will try to nominate the ilb (idle load balancing) - * owner among the cpus whose ticks are stopped. ilb owner will do the idle - * load balancing on behalf of all those cpus. If all the cpus in the system - * go into this tickless mode, then there will be no ilb owner (as there is - * no need for one) and all the cpus will sleep till the next wakeup event - * arrives... - * - * For the ilb owner, tick is not stopped. And this tick will be used - * for idle load balancing. ilb owner will still be part of - * nohz.cpu_mask.. - * - * While stopping the tick, this cpu will become the ilb owner if there - * is no other owner. And will be the owner till that cpu becomes busy - * or if all cpus in the system stop their ticks at which point - * there is no need for ilb owner. - * - * When the ilb owner becomes busy, it nominates another owner, during the - * next busy scheduler_tick() - */ -int select_nohz_load_balancer(int stop_tick) -{ - int cpu = smp_processor_id(); - - if (stop_tick) { - cpu_rq(cpu)->in_nohz_recently = 1; - - if (!cpu_active(cpu)) { - if (atomic_read(&nohz.load_balancer) != cpu) - return 0; - - /* - * If we are going offline and still the leader, - * give up! - */ - if (atomic_cmpxchg(&nohz.load_balancer, cpu, -1) != cpu) - BUG(); - - return 0; - } - - cpumask_set_cpu(cpu, nohz.cpu_mask); - - /* time for ilb owner also to sleep */ - if (cpumask_weight(nohz.cpu_mask) == num_active_cpus()) { - if (atomic_read(&nohz.load_balancer) == cpu) - atomic_set(&nohz.load_balancer, -1); - return 0; - } - - if (atomic_read(&nohz.load_balancer) == -1) { - /* make me the ilb owner */ - if (atomic_cmpxchg(&nohz.load_balancer, -1, cpu) == -1) - return 1; - } else if (atomic_read(&nohz.load_balancer) == cpu) { - int new_ilb; - - if (!(sched_smt_power_savings || - sched_mc_power_savings)) - return 1; - /* - * Check to see if there is a more power-efficient - * ilb. - */ - new_ilb = find_new_ilb(cpu); - if (new_ilb < nr_cpu_ids && new_ilb != cpu) { - atomic_set(&nohz.load_balancer, -1); - resched_cpu(new_ilb); - return 0; - } - return 1; - } - } else { - if (!cpumask_test_cpu(cpu, nohz.cpu_mask)) - return 0; - - cpumask_clear_cpu(cpu, nohz.cpu_mask); - - if (atomic_read(&nohz.load_balancer) == cpu) - if (atomic_cmpxchg(&nohz.load_balancer, cpu, -1) != cpu) - BUG(); - } - return 0; -} -#endif - -static DEFINE_SPINLOCK(balancing); - -/* - * It checks each scheduling domain to see if it is due to be balanced, - * and initiates a balancing operation if so. - * - * Balancing parameters are set up in arch_init_sched_domains. - */ -static void rebalance_domains(int cpu, enum cpu_idle_type idle) -{ - int balance = 1; - struct rq *rq = cpu_rq(cpu); - unsigned long interval; - struct sched_domain *sd; - /* Earliest time when we have to do rebalance again */ - unsigned long next_balance = jiffies + 60*HZ; - int update_next_balance = 0; - int need_serialize; - - for_each_domain(cpu, sd) { - if (!(sd->flags & SD_LOAD_BALANCE)) - continue; - - interval = sd->balance_interval; - if (idle != CPU_IDLE) - interval *= sd->busy_factor; - - /* scale ms to jiffies */ - interval = msecs_to_jiffies(interval); - if (unlikely(!interval)) - interval = 1; - if (interval > HZ*NR_CPUS/10) - interval = HZ*NR_CPUS/10; - - need_serialize = sd->flags & SD_SERIALIZE; - - if (need_serialize) { - if (!spin_trylock(&balancing)) - goto out; - } - - if (time_after_eq(jiffies, sd->last_balance + interval)) { - if (load_balance(cpu, rq, sd, idle, &balance)) { - /* - * We've pulled tasks over so either we're no - * longer idle, or one of our SMT siblings is - * not idle. - */ - idle = CPU_NOT_IDLE; - } - sd->last_balance = jiffies; - } - if (need_serialize) - spin_unlock(&balancing); -out: - if (time_after(next_balance, sd->last_balance + interval)) { - next_balance = sd->last_balance + interval; - update_next_balance = 1; - } - - /* - * Stop the load balance at this level. There is another - * CPU in our sched group which is doing load balancing more - * actively. - */ - if (!balance) - break; - } - - /* - * next_balance will be updated only when there is a need. - * When the cpu is attached to null domain for ex, it will not be - * updated. - */ - if (likely(update_next_balance)) - rq->next_balance = next_balance; -} - -/* - * run_rebalance_domains is triggered when needed from the scheduler tick. - * In CONFIG_NO_HZ case, the idle load balance owner will do the - * rebalancing for all the cpus for whom scheduler ticks are stopped. - */ -static void run_rebalance_domains(struct softirq_action *h) -{ - int this_cpu = smp_processor_id(); - struct rq *this_rq = cpu_rq(this_cpu); - enum cpu_idle_type idle = this_rq->idle_at_tick ? - CPU_IDLE : CPU_NOT_IDLE; - - rebalance_domains(this_cpu, idle); - -#ifdef CONFIG_NO_HZ - /* - * If this cpu is the owner for idle load balancing, then do the - * balancing on behalf of the other idle cpus whose ticks are - * stopped. - */ - if (this_rq->idle_at_tick && - atomic_read(&nohz.load_balancer) == this_cpu) { - struct rq *rq; - int balance_cpu; - - for_each_cpu(balance_cpu, nohz.cpu_mask) { - if (balance_cpu == this_cpu) - continue; - - /* - * If this cpu gets work to do, stop the load balancing - * work being done for other cpus. Next load - * balancing owner will pick it up. - */ - if (need_resched()) - break; - - rebalance_domains(balance_cpu, CPU_IDLE); - - rq = cpu_rq(balance_cpu); - if (time_after(this_rq->next_balance, rq->next_balance)) - this_rq->next_balance = rq->next_balance; - } - } -#endif -} - -static inline int on_null_domain(int cpu) -{ - return !rcu_dereference(cpu_rq(cpu)->sd); -} - -/* - * Trigger the SCHED_SOFTIRQ if it is time to do periodic load balancing. - * - * In case of CONFIG_NO_HZ, this is the place where we nominate a new - * idle load balancing owner or decide to stop the periodic load balancing, - * if the whole system is idle. - */ -static inline void trigger_load_balance(struct rq *rq, int cpu) -{ -#ifdef CONFIG_NO_HZ - /* - * If we were in the nohz mode recently and busy at the current - * scheduler tick, then check if we need to nominate new idle - * load balancer. - */ - if (rq->in_nohz_recently && !rq->idle_at_tick) { - rq->in_nohz_recently = 0; - - if (atomic_read(&nohz.load_balancer) == cpu) { - cpumask_clear_cpu(cpu, nohz.cpu_mask); - atomic_set(&nohz.load_balancer, -1); - } - - if (atomic_read(&nohz.load_balancer) == -1) { - int ilb = find_new_ilb(cpu); - - if (ilb < nr_cpu_ids) - resched_cpu(ilb); - } - } - - /* - * If this cpu is idle and doing idle load balancing for all the - * cpus with ticks stopped, is it time for that to stop? - */ - if (rq->idle_at_tick && atomic_read(&nohz.load_balancer) == cpu && - cpumask_weight(nohz.cpu_mask) == num_online_cpus()) { - resched_cpu(cpu); - return; - } - - /* - * If this cpu is idle and the idle load balancing is done by - * someone else, then no need raise the SCHED_SOFTIRQ - */ - if (rq->idle_at_tick && atomic_read(&nohz.load_balancer) != cpu && - cpumask_test_cpu(cpu, nohz.cpu_mask)) - return; -#endif - /* Don't need to rebalance while attached to NULL domain */ - if (time_after_eq(jiffies, rq->next_balance) && - likely(!on_null_domain(cpu))) - raise_softirq(SCHED_SOFTIRQ); -} - -#else /* CONFIG_SMP */ - -/* - * on UP we do not need to balance between CPUs: - */ -static inline void idle_balance(int cpu, struct rq *rq) -{ -} - #endif DEFINE_PER_CPU(struct kernel_stat, kstat); @@ -5309,7 +3515,7 @@ void scheduler_tick(void) curr->sched_class->task_tick(rq, curr, 0); raw_spin_unlock(&rq->lock); - perf_event_task_tick(curr, cpu); + perf_event_task_tick(curr); #ifdef CONFIG_SMP rq->idle_at_tick = idle_cpu(cpu); @@ -5523,7 +3729,7 @@ need_resched_nonpreemptible: if (likely(prev != next)) { sched_info_switch(prev, next); - perf_event_task_sched_out(prev, next, cpu); + perf_event_task_sched_out(prev, next); rq->nr_switches++; rq->curr = next; @@ -6054,7 +4260,7 @@ void rt_mutex_setprio(struct task_struct *p, int prio) unsigned long flags; int oldprio, on_rq, running; struct rq *rq; - const struct sched_class *prev_class = p->sched_class; + const struct sched_class *prev_class; BUG_ON(prio < 0 || prio > MAX_PRIO); @@ -6062,6 +4268,7 @@ void rt_mutex_setprio(struct task_struct *p, int prio) update_rq_clock(rq); oldprio = p->prio; + prev_class = p->sched_class; on_rq = p->se.on_rq; running = task_current(rq, p); if (on_rq) @@ -6079,7 +4286,7 @@ void rt_mutex_setprio(struct task_struct *p, int prio) if (running) p->sched_class->set_curr_task(rq); if (on_rq) { - enqueue_task(rq, p, 0); + enqueue_task(rq, p, 0, oldprio < prio); check_class_changed(rq, p, prev_class, oldprio, running); } @@ -6123,7 +4330,7 @@ void set_user_nice(struct task_struct *p, long nice) delta = p->prio - old_prio; if (on_rq) { - enqueue_task(rq, p, 0); + enqueue_task(rq, p, 0, false); /* * If the task increased its priority or is running and * lowered its priority, then reschedule its CPU: @@ -6281,7 +4488,7 @@ static int __sched_setscheduler(struct task_struct *p, int policy, { int retval, oldprio, oldpolicy = -1, on_rq, running; unsigned long flags; - const struct sched_class *prev_class = p->sched_class; + const struct sched_class *prev_class; struct rq *rq; int reset_on_fork; @@ -6395,6 +4602,7 @@ recheck: p->sched_reset_on_fork = reset_on_fork; oldprio = p->prio; + prev_class = p->sched_class; __setscheduler(rq, p, policy, param->sched_priority); if (running) @@ -7145,27 +5353,8 @@ int set_cpus_allowed_ptr(struct task_struct *p, const struct cpumask *new_mask) struct rq *rq; int ret = 0; - /* - * Since we rely on wake-ups to migrate sleeping tasks, don't change - * the ->cpus_allowed mask from under waking tasks, which would be - * possible when we change rq->lock in ttwu(), so synchronize against - * TASK_WAKING to avoid that. - * - * Make an exception for freshly cloned tasks, since cpuset namespaces - * might move the task about, we have to validate the target in - * wake_up_new_task() anyway since the cpu might have gone away. - */ -again: - while (p->state == TASK_WAKING && !(p->flags & PF_STARTING)) - cpu_relax(); - rq = task_rq_lock(p, &flags); - if (p->state == TASK_WAKING && !(p->flags & PF_STARTING)) { - task_rq_unlock(rq, &flags); - goto again; - } - if (!cpumask_intersects(new_mask, cpu_active_mask)) { ret = -EINVAL; goto out; @@ -9452,7 +7641,6 @@ static void init_tg_rt_entry(struct task_group *tg, struct rt_rq *rt_rq, tg->rt_rq[cpu] = rt_rq; init_rt_rq(rt_rq, rq); rt_rq->tg = tg; - rt_rq->rt_se = rt_se; rt_rq->rt_runtime = tg->rt_bandwidth.rt_runtime; if (add) list_add(&rt_rq->leaf_rt_rq_list, &rq->leaf_rt_rq_list); @@ -9483,9 +7671,6 @@ void __init sched_init(void) #ifdef CONFIG_RT_GROUP_SCHED alloc_size += 2 * nr_cpu_ids * sizeof(void **); #endif -#ifdef CONFIG_USER_SCHED - alloc_size *= 2; -#endif #ifdef CONFIG_CPUMASK_OFFSTACK alloc_size += num_possible_cpus() * cpumask_size(); #endif @@ -9499,13 +7684,6 @@ void __init sched_init(void) init_task_group.cfs_rq = (struct cfs_rq **)ptr; ptr += nr_cpu_ids * sizeof(void **); -#ifdef CONFIG_USER_SCHED - root_task_group.se = (struct sched_entity **)ptr; - ptr += nr_cpu_ids * sizeof(void **); - - root_task_group.cfs_rq = (struct cfs_rq **)ptr; - ptr += nr_cpu_ids * sizeof(void **); -#endif /* CONFIG_USER_SCHED */ #endif /* CONFIG_FAIR_GROUP_SCHED */ #ifdef CONFIG_RT_GROUP_SCHED init_task_group.rt_se = (struct sched_rt_entity **)ptr; @@ -9514,13 +7692,6 @@ void __init sched_init(void) init_task_group.rt_rq = (struct rt_rq **)ptr; ptr += nr_cpu_ids * sizeof(void **); -#ifdef CONFIG_USER_SCHED - root_task_group.rt_se = (struct sched_rt_entity **)ptr; - ptr += nr_cpu_ids * sizeof(void **); - - root_task_group.rt_rq = (struct rt_rq **)ptr; - ptr += nr_cpu_ids * sizeof(void **); -#endif /* CONFIG_USER_SCHED */ #endif /* CONFIG_RT_GROUP_SCHED */ #ifdef CONFIG_CPUMASK_OFFSTACK for_each_possible_cpu(i) { @@ -9540,22 +7711,13 @@ void __init sched_init(void) #ifdef CONFIG_RT_GROUP_SCHED init_rt_bandwidth(&init_task_group.rt_bandwidth, global_rt_period(), global_rt_runtime()); -#ifdef CONFIG_USER_SCHED - init_rt_bandwidth(&root_task_group.rt_bandwidth, - global_rt_period(), RUNTIME_INF); -#endif /* CONFIG_USER_SCHED */ #endif /* CONFIG_RT_GROUP_SCHED */ -#ifdef CONFIG_GROUP_SCHED +#ifdef CONFIG_CGROUP_SCHED list_add(&init_task_group.list, &task_groups); INIT_LIST_HEAD(&init_task_group.children); -#ifdef CONFIG_USER_SCHED - INIT_LIST_HEAD(&root_task_group.children); - init_task_group.parent = &root_task_group; - list_add(&init_task_group.siblings, &root_task_group.children); -#endif /* CONFIG_USER_SCHED */ -#endif /* CONFIG_GROUP_SCHED */ +#endif /* CONFIG_CGROUP_SCHED */ #if defined CONFIG_FAIR_GROUP_SCHED && defined CONFIG_SMP update_shares_data = __alloc_percpu(nr_cpu_ids * sizeof(unsigned long), @@ -9595,25 +7757,6 @@ void __init sched_init(void) * directly in rq->cfs (i.e init_task_group->se[] = NULL). */ init_tg_cfs_entry(&init_task_group, &rq->cfs, NULL, i, 1, NULL); -#elif defined CONFIG_USER_SCHED - root_task_group.shares = NICE_0_LOAD; - init_tg_cfs_entry(&root_task_group, &rq->cfs, NULL, i, 0, NULL); - /* - * In case of task-groups formed thr' the user id of tasks, - * init_task_group represents tasks belonging to root user. - * Hence it forms a sibling of all subsequent groups formed. - * In this case, init_task_group gets only a fraction of overall - * system cpu resource, based on the weight assigned to root - * user's cpu share (INIT_TASK_GROUP_LOAD). This is accomplished - * by letting tasks of init_task_group sit in a separate cfs_rq - * (init_tg_cfs_rq) and having one entity represent this group of - * tasks in rq->cfs (i.e init_task_group->se[] != NULL). - */ - init_tg_cfs_entry(&init_task_group, - &per_cpu(init_tg_cfs_rq, i), - &per_cpu(init_sched_entity, i), i, 1, - root_task_group.se[i]); - #endif #endif /* CONFIG_FAIR_GROUP_SCHED */ @@ -9622,12 +7765,6 @@ void __init sched_init(void) INIT_LIST_HEAD(&rq->leaf_rt_rq_list); #ifdef CONFIG_CGROUP_SCHED init_tg_rt_entry(&init_task_group, &rq->rt, NULL, i, 1, NULL); -#elif defined CONFIG_USER_SCHED - init_tg_rt_entry(&root_task_group, &rq->rt, NULL, i, 0, NULL); - init_tg_rt_entry(&init_task_group, - &per_cpu(init_rt_rq_var, i), - &per_cpu(init_sched_rt_entity, i), i, 1, - root_task_group.rt_se[i]); #endif #endif @@ -9712,7 +7849,7 @@ static inline int preempt_count_equals(int preempt_offset) return (nested == PREEMPT_INATOMIC_BASE + preempt_offset); } -void __might_sleep(char *file, int line, int preempt_offset) +void __might_sleep(const char *file, int line, int preempt_offset) { #ifdef in_atomic static unsigned long prev_jiffy; /* ratelimiting */ @@ -10023,7 +8160,7 @@ static inline void unregister_rt_sched_group(struct task_group *tg, int cpu) } #endif /* CONFIG_RT_GROUP_SCHED */ -#ifdef CONFIG_GROUP_SCHED +#ifdef CONFIG_CGROUP_SCHED static void free_sched_group(struct task_group *tg) { free_fair_sched_group(tg); @@ -10128,11 +8265,11 @@ void sched_move_task(struct task_struct *tsk) if (unlikely(running)) tsk->sched_class->set_curr_task(rq); if (on_rq) - enqueue_task(rq, tsk, 0); + enqueue_task(rq, tsk, 0, false); task_rq_unlock(rq, &flags); } -#endif /* CONFIG_GROUP_SCHED */ +#endif /* CONFIG_CGROUP_SCHED */ #ifdef CONFIG_FAIR_GROUP_SCHED static void __set_se_shares(struct sched_entity *se, unsigned long shares) @@ -10274,13 +8411,6 @@ static int tg_schedulable(struct task_group *tg, void *data) runtime = d->rt_runtime; } -#ifdef CONFIG_USER_SCHED - if (tg == &root_task_group) { - period = global_rt_period(); - runtime = global_rt_runtime(); - } -#endif - /* * Cannot have more runtime than the period. */ @@ -10900,12 +9030,30 @@ static void cpuacct_charge(struct task_struct *tsk, u64 cputime) } /* + * When CONFIG_VIRT_CPU_ACCOUNTING is enabled one jiffy can be very large + * in cputime_t units. As a result, cpuacct_update_stats calls + * percpu_counter_add with values large enough to always overflow the + * per cpu batch limit causing bad SMP scalability. + * + * To fix this we scale percpu_counter_batch by cputime_one_jiffy so we + * batch the same amount of time with CONFIG_VIRT_CPU_ACCOUNTING disabled + * and enabled. We cap it at INT_MAX which is the largest allowed batch value. + */ +#ifdef CONFIG_SMP +#define CPUACCT_BATCH \ + min_t(long, percpu_counter_batch * cputime_one_jiffy, INT_MAX) +#else +#define CPUACCT_BATCH 0 +#endif + +/* * Charge the system/user time to the task's accounting group. */ static void cpuacct_update_stats(struct task_struct *tsk, enum cpuacct_stat_index idx, cputime_t val) { struct cpuacct *ca; + int batch = CPUACCT_BATCH; if (unlikely(!cpuacct_subsys.active)) return; @@ -10914,7 +9062,7 @@ static void cpuacct_update_stats(struct task_struct *tsk, ca = task_ca(tsk); do { - percpu_counter_add(&ca->cpustat[idx], val); + __percpu_counter_add(&ca->cpustat[idx], val, batch); ca = ca->parent; } while (ca); rcu_read_unlock(); diff --git a/kernel/sched_cpupri.c b/kernel/sched_cpupri.c index 597b33099dfa..eeb3506c4834 100644 --- a/kernel/sched_cpupri.c +++ b/kernel/sched_cpupri.c @@ -47,9 +47,7 @@ static int convert_prio(int prio) } #define for_each_cpupri_active(array, idx) \ - for (idx = find_first_bit(array, CPUPRI_NR_PRIORITIES); \ - idx < CPUPRI_NR_PRIORITIES; \ - idx = find_next_bit(array, CPUPRI_NR_PRIORITIES, idx+1)) + for_each_bit(idx, array, CPUPRI_NR_PRIORITIES) /** * cpupri_find - find the best (lowest-pri) CPU in the system diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 8fe7ee81c552..3e1fd96c6cf9 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1053,7 +1053,8 @@ static inline void hrtick_update(struct rq *rq) * increased. Here we update the fair scheduling stats and * then put the task into the rbtree: */ -static void enqueue_task_fair(struct rq *rq, struct task_struct *p, int wakeup) +static void +enqueue_task_fair(struct rq *rq, struct task_struct *p, int wakeup, bool head) { struct cfs_rq *cfs_rq; struct sched_entity *se = &p->se; @@ -1815,57 +1816,164 @@ static void put_prev_task_fair(struct rq *rq, struct task_struct *prev) */ /* - * Load-balancing iterator. Note: while the runqueue stays locked - * during the whole iteration, the current task might be - * dequeued so the iterator has to be dequeue-safe. Here we - * achieve that by always pre-iterating before returning - * the current task: + * pull_task - move a task from a remote runqueue to the local runqueue. + * Both runqueues must be locked. */ -static struct task_struct * -__load_balance_iterator(struct cfs_rq *cfs_rq, struct list_head *next) +static void pull_task(struct rq *src_rq, struct task_struct *p, + struct rq *this_rq, int this_cpu) { - struct task_struct *p = NULL; - struct sched_entity *se; + deactivate_task(src_rq, p, 0); + set_task_cpu(p, this_cpu); + activate_task(this_rq, p, 0); + check_preempt_curr(this_rq, p, 0); +} - if (next == &cfs_rq->tasks) - return NULL; +/* + * can_migrate_task - may task p from runqueue rq be migrated to this_cpu? + */ +static +int can_migrate_task(struct task_struct *p, struct rq *rq, int this_cpu, + struct sched_domain *sd, enum cpu_idle_type idle, + int *all_pinned) +{ + int tsk_cache_hot = 0; + /* + * We do not migrate tasks that are: + * 1) running (obviously), or + * 2) cannot be migrated to this CPU due to cpus_allowed, or + * 3) are cache-hot on their current CPU. + */ + if (!cpumask_test_cpu(this_cpu, &p->cpus_allowed)) { + schedstat_inc(p, se.nr_failed_migrations_affine); + return 0; + } + *all_pinned = 0; - se = list_entry(next, struct sched_entity, group_node); - p = task_of(se); - cfs_rq->balance_iterator = next->next; + if (task_running(rq, p)) { + schedstat_inc(p, se.nr_failed_migrations_running); + return 0; + } - return p; -} + /* + * Aggressive migration if: + * 1) task is cache cold, or + * 2) too many balance attempts have failed. + */ -static struct task_struct *load_balance_start_fair(void *arg) -{ - struct cfs_rq *cfs_rq = arg; + tsk_cache_hot = task_hot(p, rq->clock, sd); + if (!tsk_cache_hot || + sd->nr_balance_failed > sd->cache_nice_tries) { +#ifdef CONFIG_SCHEDSTATS + if (tsk_cache_hot) { + schedstat_inc(sd, lb_hot_gained[idle]); + schedstat_inc(p, se.nr_forced_migrations); + } +#endif + return 1; + } - return __load_balance_iterator(cfs_rq, cfs_rq->tasks.next); + if (tsk_cache_hot) { + schedstat_inc(p, se.nr_failed_migrations_hot); + return 0; + } + return 1; } -static struct task_struct *load_balance_next_fair(void *arg) +/* + * move_one_task tries to move exactly one task from busiest to this_rq, as + * part of active balancing operations within "domain". + * Returns 1 if successful and 0 otherwise. + * + * Called with both runqueues locked. + */ +static int +move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, + struct sched_domain *sd, enum cpu_idle_type idle) { - struct cfs_rq *cfs_rq = arg; + struct task_struct *p, *n; + struct cfs_rq *cfs_rq; + int pinned = 0; + + for_each_leaf_cfs_rq(busiest, cfs_rq) { + list_for_each_entry_safe(p, n, &cfs_rq->tasks, se.group_node) { + + if (!can_migrate_task(p, busiest, this_cpu, + sd, idle, &pinned)) + continue; - return __load_balance_iterator(cfs_rq, cfs_rq->balance_iterator); + pull_task(busiest, p, this_rq, this_cpu); + /* + * Right now, this is only the second place pull_task() + * is called, so we can safely collect pull_task() + * stats here rather than inside pull_task(). + */ + schedstat_inc(sd, lb_gained[idle]); + return 1; + } + } + + return 0; } static unsigned long -__load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, struct sched_domain *sd, - enum cpu_idle_type idle, int *all_pinned, int *this_best_prio, - struct cfs_rq *cfs_rq) +balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, + unsigned long max_load_move, struct sched_domain *sd, + enum cpu_idle_type idle, int *all_pinned, + int *this_best_prio, struct cfs_rq *busiest_cfs_rq) { - struct rq_iterator cfs_rq_iterator; + int loops = 0, pulled = 0, pinned = 0; + long rem_load_move = max_load_move; + struct task_struct *p, *n; - cfs_rq_iterator.start = load_balance_start_fair; - cfs_rq_iterator.next = load_balance_next_fair; - cfs_rq_iterator.arg = cfs_rq; + if (max_load_move == 0) + goto out; - return balance_tasks(this_rq, this_cpu, busiest, - max_load_move, sd, idle, all_pinned, - this_best_prio, &cfs_rq_iterator); + pinned = 1; + + list_for_each_entry_safe(p, n, &busiest_cfs_rq->tasks, se.group_node) { + if (loops++ > sysctl_sched_nr_migrate) + break; + + if ((p->se.load.weight >> 1) > rem_load_move || + !can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) + continue; + + pull_task(busiest, p, this_rq, this_cpu); + pulled++; + rem_load_move -= p->se.load.weight; + +#ifdef CONFIG_PREEMPT + /* + * NEWIDLE balancing is a source of latency, so preemptible + * kernels will stop after the first task is pulled to minimize + * the critical section. + */ + if (idle == CPU_NEWLY_IDLE) + break; +#endif + + /* + * We only want to steal up to the prescribed amount of + * weighted load. + */ + if (rem_load_move <= 0) + break; + + if (p->prio < *this_best_prio) + *this_best_prio = p->prio; + } +out: + /* + * Right now, this is one of only two places pull_task() is called, + * so we can safely collect pull_task() stats here rather than + * inside pull_task(). + */ + schedstat_add(sd, lb_gained[idle], pulled); + + if (all_pinned) + *all_pinned = pinned; + + return max_load_move - rem_load_move; } #ifdef CONFIG_FAIR_GROUP_SCHED @@ -1897,9 +2005,9 @@ load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, rem_load = (u64)rem_load_move * busiest_weight; rem_load = div_u64(rem_load, busiest_h_load + 1); - moved_load = __load_balance_fair(this_rq, this_cpu, busiest, + moved_load = balance_tasks(this_rq, this_cpu, busiest, rem_load, sd, idle, all_pinned, this_best_prio, - tg->cfs_rq[busiest_cpu]); + busiest_cfs_rq); if (!moved_load) continue; @@ -1922,35 +2030,1509 @@ load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, struct sched_domain *sd, enum cpu_idle_type idle, int *all_pinned, int *this_best_prio) { - return __load_balance_fair(this_rq, this_cpu, busiest, + return balance_tasks(this_rq, this_cpu, busiest, max_load_move, sd, idle, all_pinned, this_best_prio, &busiest->cfs); } #endif -static int -move_one_task_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle) +/* + * move_tasks tries to move up to max_load_move weighted load from busiest to + * this_rq, as part of a balancing operation within domain "sd". + * Returns 1 if successful and 0 otherwise. + * + * Called with both runqueues locked. + */ +static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, + unsigned long max_load_move, + struct sched_domain *sd, enum cpu_idle_type idle, + int *all_pinned) { - struct cfs_rq *busy_cfs_rq; - struct rq_iterator cfs_rq_iterator; + unsigned long total_load_moved = 0, load_moved; + int this_best_prio = this_rq->curr->prio; - cfs_rq_iterator.start = load_balance_start_fair; - cfs_rq_iterator.next = load_balance_next_fair; + do { + load_moved = load_balance_fair(this_rq, this_cpu, busiest, + max_load_move - total_load_moved, + sd, idle, all_pinned, &this_best_prio); - for_each_leaf_cfs_rq(busiest, busy_cfs_rq) { + total_load_moved += load_moved; + +#ifdef CONFIG_PREEMPT /* - * pass busy_cfs_rq argument into - * load_balance_[start|next]_fair iterators + * NEWIDLE balancing is a source of latency, so preemptible + * kernels will stop after the first task is pulled to minimize + * the critical section. */ - cfs_rq_iterator.arg = busy_cfs_rq; - if (iter_move_one_task(this_rq, this_cpu, busiest, sd, idle, - &cfs_rq_iterator)) - return 1; + if (idle == CPU_NEWLY_IDLE && this_rq->nr_running) + break; + + if (raw_spin_is_contended(&this_rq->lock) || + raw_spin_is_contended(&busiest->lock)) + break; +#endif + } while (load_moved && max_load_move > total_load_moved); + + return total_load_moved > 0; +} + +/********** Helpers for find_busiest_group ************************/ +/* + * sd_lb_stats - Structure to store the statistics of a sched_domain + * during load balancing. + */ +struct sd_lb_stats { + struct sched_group *busiest; /* Busiest group in this sd */ + struct sched_group *this; /* Local group in this sd */ + unsigned long total_load; /* Total load of all groups in sd */ + unsigned long total_pwr; /* Total power of all groups in sd */ + unsigned long avg_load; /* Average load across all groups in sd */ + + /** Statistics of this group */ + unsigned long this_load; + unsigned long this_load_per_task; + unsigned long this_nr_running; + + /* Statistics of the busiest group */ + unsigned long max_load; + unsigned long busiest_load_per_task; + unsigned long busiest_nr_running; + unsigned long busiest_group_capacity; + + int group_imb; /* Is there imbalance in this sd */ +#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) + int power_savings_balance; /* Is powersave balance needed for this sd */ + struct sched_group *group_min; /* Least loaded group in sd */ + struct sched_group *group_leader; /* Group which relieves group_min */ + unsigned long min_load_per_task; /* load_per_task in group_min */ + unsigned long leader_nr_running; /* Nr running of group_leader */ + unsigned long min_nr_running; /* Nr running of group_min */ +#endif +}; + +/* + * sg_lb_stats - stats of a sched_group required for load_balancing + */ +struct sg_lb_stats { + unsigned long avg_load; /*Avg load across the CPUs of the group */ + unsigned long group_load; /* Total load over the CPUs of the group */ + unsigned long sum_nr_running; /* Nr tasks running in the group */ + unsigned long sum_weighted_load; /* Weighted load of group's tasks */ + unsigned long group_capacity; + int group_imb; /* Is there an imbalance in the group ? */ +}; + +/** + * group_first_cpu - Returns the first cpu in the cpumask of a sched_group. + * @group: The group whose first cpu is to be returned. + */ +static inline unsigned int group_first_cpu(struct sched_group *group) +{ + return cpumask_first(sched_group_cpus(group)); +} + +/** + * get_sd_load_idx - Obtain the load index for a given sched domain. + * @sd: The sched_domain whose load_idx is to be obtained. + * @idle: The Idle status of the CPU for whose sd load_icx is obtained. + */ +static inline int get_sd_load_idx(struct sched_domain *sd, + enum cpu_idle_type idle) +{ + int load_idx; + + switch (idle) { + case CPU_NOT_IDLE: + load_idx = sd->busy_idx; + break; + + case CPU_NEWLY_IDLE: + load_idx = sd->newidle_idx; + break; + default: + load_idx = sd->idle_idx; + break; } + return load_idx; +} + + +#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) +/** + * init_sd_power_savings_stats - Initialize power savings statistics for + * the given sched_domain, during load balancing. + * + * @sd: Sched domain whose power-savings statistics are to be initialized. + * @sds: Variable containing the statistics for sd. + * @idle: Idle status of the CPU at which we're performing load-balancing. + */ +static inline void init_sd_power_savings_stats(struct sched_domain *sd, + struct sd_lb_stats *sds, enum cpu_idle_type idle) +{ + /* + * Busy processors will not participate in power savings + * balance. + */ + if (idle == CPU_NOT_IDLE || !(sd->flags & SD_POWERSAVINGS_BALANCE)) + sds->power_savings_balance = 0; + else { + sds->power_savings_balance = 1; + sds->min_nr_running = ULONG_MAX; + sds->leader_nr_running = 0; + } +} + +/** + * update_sd_power_savings_stats - Update the power saving stats for a + * sched_domain while performing load balancing. + * + * @group: sched_group belonging to the sched_domain under consideration. + * @sds: Variable containing the statistics of the sched_domain + * @local_group: Does group contain the CPU for which we're performing + * load balancing ? + * @sgs: Variable containing the statistics of the group. + */ +static inline void update_sd_power_savings_stats(struct sched_group *group, + struct sd_lb_stats *sds, int local_group, struct sg_lb_stats *sgs) +{ + + if (!sds->power_savings_balance) + return; + + /* + * If the local group is idle or completely loaded + * no need to do power savings balance at this domain + */ + if (local_group && (sds->this_nr_running >= sgs->group_capacity || + !sds->this_nr_running)) + sds->power_savings_balance = 0; + + /* + * If a group is already running at full capacity or idle, + * don't include that group in power savings calculations + */ + if (!sds->power_savings_balance || + sgs->sum_nr_running >= sgs->group_capacity || + !sgs->sum_nr_running) + return; + + /* + * Calculate the group which has the least non-idle load. + * This is the group from where we need to pick up the load + * for saving power + */ + if ((sgs->sum_nr_running < sds->min_nr_running) || + (sgs->sum_nr_running == sds->min_nr_running && + group_first_cpu(group) > group_first_cpu(sds->group_min))) { + sds->group_min = group; + sds->min_nr_running = sgs->sum_nr_running; + sds->min_load_per_task = sgs->sum_weighted_load / + sgs->sum_nr_running; + } + + /* + * Calculate the group which is almost near its + * capacity but still has some space to pick up some load + * from other group and save more power + */ + if (sgs->sum_nr_running + 1 > sgs->group_capacity) + return; + + if (sgs->sum_nr_running > sds->leader_nr_running || + (sgs->sum_nr_running == sds->leader_nr_running && + group_first_cpu(group) < group_first_cpu(sds->group_leader))) { + sds->group_leader = group; + sds->leader_nr_running = sgs->sum_nr_running; + } +} + +/** + * check_power_save_busiest_group - see if there is potential for some power-savings balance + * @sds: Variable containing the statistics of the sched_domain + * under consideration. + * @this_cpu: Cpu at which we're currently performing load-balancing. + * @imbalance: Variable to store the imbalance. + * + * Description: + * Check if we have potential to perform some power-savings balance. + * If yes, set the busiest group to be the least loaded group in the + * sched_domain, so that it's CPUs can be put to idle. + * + * Returns 1 if there is potential to perform power-savings balance. + * Else returns 0. + */ +static inline int check_power_save_busiest_group(struct sd_lb_stats *sds, + int this_cpu, unsigned long *imbalance) +{ + if (!sds->power_savings_balance) + return 0; + + if (sds->this != sds->group_leader || + sds->group_leader == sds->group_min) + return 0; + + *imbalance = sds->min_load_per_task; + sds->busiest = sds->group_min; + + return 1; + +} +#else /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ +static inline void init_sd_power_savings_stats(struct sched_domain *sd, + struct sd_lb_stats *sds, enum cpu_idle_type idle) +{ + return; +} + +static inline void update_sd_power_savings_stats(struct sched_group *group, + struct sd_lb_stats *sds, int local_group, struct sg_lb_stats *sgs) +{ + return; +} + +static inline int check_power_save_busiest_group(struct sd_lb_stats *sds, + int this_cpu, unsigned long *imbalance) +{ return 0; } +#endif /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ + + +unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu) +{ + return SCHED_LOAD_SCALE; +} + +unsigned long __weak arch_scale_freq_power(struct sched_domain *sd, int cpu) +{ + return default_scale_freq_power(sd, cpu); +} + +unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu) +{ + unsigned long weight = cpumask_weight(sched_domain_span(sd)); + unsigned long smt_gain = sd->smt_gain; + + smt_gain /= weight; + + return smt_gain; +} + +unsigned long __weak arch_scale_smt_power(struct sched_domain *sd, int cpu) +{ + return default_scale_smt_power(sd, cpu); +} + +unsigned long scale_rt_power(int cpu) +{ + struct rq *rq = cpu_rq(cpu); + u64 total, available; + + sched_avg_update(rq); + + total = sched_avg_period() + (rq->clock - rq->age_stamp); + available = total - rq->rt_avg; + + if (unlikely((s64)total < SCHED_LOAD_SCALE)) + total = SCHED_LOAD_SCALE; + + total >>= SCHED_LOAD_SHIFT; + + return div_u64(available, total); +} + +static void update_cpu_power(struct sched_domain *sd, int cpu) +{ + unsigned long weight = cpumask_weight(sched_domain_span(sd)); + unsigned long power = SCHED_LOAD_SCALE; + struct sched_group *sdg = sd->groups; + + if (sched_feat(ARCH_POWER)) + power *= arch_scale_freq_power(sd, cpu); + else + power *= default_scale_freq_power(sd, cpu); + + power >>= SCHED_LOAD_SHIFT; + + if ((sd->flags & SD_SHARE_CPUPOWER) && weight > 1) { + if (sched_feat(ARCH_POWER)) + power *= arch_scale_smt_power(sd, cpu); + else + power *= default_scale_smt_power(sd, cpu); + + power >>= SCHED_LOAD_SHIFT; + } + + power *= scale_rt_power(cpu); + power >>= SCHED_LOAD_SHIFT; + + if (!power) + power = 1; + + sdg->cpu_power = power; +} + +static void update_group_power(struct sched_domain *sd, int cpu) +{ + struct sched_domain *child = sd->child; + struct sched_group *group, *sdg = sd->groups; + unsigned long power; + + if (!child) { + update_cpu_power(sd, cpu); + return; + } + + power = 0; + + group = child->groups; + do { + power += group->cpu_power; + group = group->next; + } while (group != child->groups); + + sdg->cpu_power = power; +} + +/** + * update_sg_lb_stats - Update sched_group's statistics for load balancing. + * @sd: The sched_domain whose statistics are to be updated. + * @group: sched_group whose statistics are to be updated. + * @this_cpu: Cpu for which load balance is currently performed. + * @idle: Idle status of this_cpu + * @load_idx: Load index of sched_domain of this_cpu for load calc. + * @sd_idle: Idle status of the sched_domain containing group. + * @local_group: Does group contain this_cpu. + * @cpus: Set of cpus considered for load balancing. + * @balance: Should we balance. + * @sgs: variable to hold the statistics for this group. + */ +static inline void update_sg_lb_stats(struct sched_domain *sd, + struct sched_group *group, int this_cpu, + enum cpu_idle_type idle, int load_idx, int *sd_idle, + int local_group, const struct cpumask *cpus, + int *balance, struct sg_lb_stats *sgs) +{ + unsigned long load, max_cpu_load, min_cpu_load; + int i; + unsigned int balance_cpu = -1, first_idle_cpu = 0; + unsigned long avg_load_per_task = 0; + + if (local_group) + balance_cpu = group_first_cpu(group); + + /* Tally up the load of all CPUs in the group */ + max_cpu_load = 0; + min_cpu_load = ~0UL; + + for_each_cpu_and(i, sched_group_cpus(group), cpus) { + struct rq *rq = cpu_rq(i); + + if (*sd_idle && rq->nr_running) + *sd_idle = 0; + + /* Bias balancing toward cpus of our domain */ + if (local_group) { + if (idle_cpu(i) && !first_idle_cpu) { + first_idle_cpu = 1; + balance_cpu = i; + } + + load = target_load(i, load_idx); + } else { + load = source_load(i, load_idx); + if (load > max_cpu_load) + max_cpu_load = load; + if (min_cpu_load > load) + min_cpu_load = load; + } + + sgs->group_load += load; + sgs->sum_nr_running += rq->nr_running; + sgs->sum_weighted_load += weighted_cpuload(i); + + } + + /* + * First idle cpu or the first cpu(busiest) in this sched group + * is eligible for doing load balancing at this and above + * domains. In the newly idle case, we will allow all the cpu's + * to do the newly idle load balance. + */ + if (idle != CPU_NEWLY_IDLE && local_group && + balance_cpu != this_cpu) { + *balance = 0; + return; + } + + update_group_power(sd, this_cpu); + + /* Adjust by relative CPU power of the group */ + sgs->avg_load = (sgs->group_load * SCHED_LOAD_SCALE) / group->cpu_power; + + /* + * Consider the group unbalanced when the imbalance is larger + * than the average weight of two tasks. + * + * APZ: with cgroup the avg task weight can vary wildly and + * might not be a suitable number - should we keep a + * normalized nr_running number somewhere that negates + * the hierarchy? + */ + if (sgs->sum_nr_running) + avg_load_per_task = sgs->sum_weighted_load / sgs->sum_nr_running; + + if ((max_cpu_load - min_cpu_load) > 2*avg_load_per_task) + sgs->group_imb = 1; + + sgs->group_capacity = + DIV_ROUND_CLOSEST(group->cpu_power, SCHED_LOAD_SCALE); +} + +/** + * update_sd_lb_stats - Update sched_group's statistics for load balancing. + * @sd: sched_domain whose statistics are to be updated. + * @this_cpu: Cpu for which load balance is currently performed. + * @idle: Idle status of this_cpu + * @sd_idle: Idle status of the sched_domain containing group. + * @cpus: Set of cpus considered for load balancing. + * @balance: Should we balance. + * @sds: variable to hold the statistics for this sched_domain. + */ +static inline void update_sd_lb_stats(struct sched_domain *sd, int this_cpu, + enum cpu_idle_type idle, int *sd_idle, + const struct cpumask *cpus, int *balance, + struct sd_lb_stats *sds) +{ + struct sched_domain *child = sd->child; + struct sched_group *group = sd->groups; + struct sg_lb_stats sgs; + int load_idx, prefer_sibling = 0; + + if (child && child->flags & SD_PREFER_SIBLING) + prefer_sibling = 1; + + init_sd_power_savings_stats(sd, sds, idle); + load_idx = get_sd_load_idx(sd, idle); + + do { + int local_group; + + local_group = cpumask_test_cpu(this_cpu, + sched_group_cpus(group)); + memset(&sgs, 0, sizeof(sgs)); + update_sg_lb_stats(sd, group, this_cpu, idle, load_idx, sd_idle, + local_group, cpus, balance, &sgs); + + if (local_group && !(*balance)) + return; + + sds->total_load += sgs.group_load; + sds->total_pwr += group->cpu_power; + + /* + * In case the child domain prefers tasks go to siblings + * first, lower the group capacity to one so that we'll try + * and move all the excess tasks away. + */ + if (prefer_sibling) + sgs.group_capacity = min(sgs.group_capacity, 1UL); + + if (local_group) { + sds->this_load = sgs.avg_load; + sds->this = group; + sds->this_nr_running = sgs.sum_nr_running; + sds->this_load_per_task = sgs.sum_weighted_load; + } else if (sgs.avg_load > sds->max_load && + (sgs.sum_nr_running > sgs.group_capacity || + sgs.group_imb)) { + sds->max_load = sgs.avg_load; + sds->busiest = group; + sds->busiest_nr_running = sgs.sum_nr_running; + sds->busiest_group_capacity = sgs.group_capacity; + sds->busiest_load_per_task = sgs.sum_weighted_load; + sds->group_imb = sgs.group_imb; + } + + update_sd_power_savings_stats(group, sds, local_group, &sgs); + group = group->next; + } while (group != sd->groups); +} + +/** + * fix_small_imbalance - Calculate the minor imbalance that exists + * amongst the groups of a sched_domain, during + * load balancing. + * @sds: Statistics of the sched_domain whose imbalance is to be calculated. + * @this_cpu: The cpu at whose sched_domain we're performing load-balance. + * @imbalance: Variable to store the imbalance. + */ +static inline void fix_small_imbalance(struct sd_lb_stats *sds, + int this_cpu, unsigned long *imbalance) +{ + unsigned long tmp, pwr_now = 0, pwr_move = 0; + unsigned int imbn = 2; + unsigned long scaled_busy_load_per_task; + + if (sds->this_nr_running) { + sds->this_load_per_task /= sds->this_nr_running; + if (sds->busiest_load_per_task > + sds->this_load_per_task) + imbn = 1; + } else + sds->this_load_per_task = + cpu_avg_load_per_task(this_cpu); + + scaled_busy_load_per_task = sds->busiest_load_per_task + * SCHED_LOAD_SCALE; + scaled_busy_load_per_task /= sds->busiest->cpu_power; + + if (sds->max_load - sds->this_load + scaled_busy_load_per_task >= + (scaled_busy_load_per_task * imbn)) { + *imbalance = sds->busiest_load_per_task; + return; + } + + /* + * OK, we don't have enough imbalance to justify moving tasks, + * however we may be able to increase total CPU power used by + * moving them. + */ + + pwr_now += sds->busiest->cpu_power * + min(sds->busiest_load_per_task, sds->max_load); + pwr_now += sds->this->cpu_power * + min(sds->this_load_per_task, sds->this_load); + pwr_now /= SCHED_LOAD_SCALE; + + /* Amount of load we'd subtract */ + tmp = (sds->busiest_load_per_task * SCHED_LOAD_SCALE) / + sds->busiest->cpu_power; + if (sds->max_load > tmp) + pwr_move += sds->busiest->cpu_power * + min(sds->busiest_load_per_task, sds->max_load - tmp); + + /* Amount of load we'd add */ + if (sds->max_load * sds->busiest->cpu_power < + sds->busiest_load_per_task * SCHED_LOAD_SCALE) + tmp = (sds->max_load * sds->busiest->cpu_power) / + sds->this->cpu_power; + else + tmp = (sds->busiest_load_per_task * SCHED_LOAD_SCALE) / + sds->this->cpu_power; + pwr_move += sds->this->cpu_power * + min(sds->this_load_per_task, sds->this_load + tmp); + pwr_move /= SCHED_LOAD_SCALE; + + /* Move if we gain throughput */ + if (pwr_move > pwr_now) + *imbalance = sds->busiest_load_per_task; +} + +/** + * calculate_imbalance - Calculate the amount of imbalance present within the + * groups of a given sched_domain during load balance. + * @sds: statistics of the sched_domain whose imbalance is to be calculated. + * @this_cpu: Cpu for which currently load balance is being performed. + * @imbalance: The variable to store the imbalance. + */ +static inline void calculate_imbalance(struct sd_lb_stats *sds, int this_cpu, + unsigned long *imbalance) +{ + unsigned long max_pull, load_above_capacity = ~0UL; + + sds->busiest_load_per_task /= sds->busiest_nr_running; + if (sds->group_imb) { + sds->busiest_load_per_task = + min(sds->busiest_load_per_task, sds->avg_load); + } + + /* + * In the presence of smp nice balancing, certain scenarios can have + * max load less than avg load(as we skip the groups at or below + * its cpu_power, while calculating max_load..) + */ + if (sds->max_load < sds->avg_load) { + *imbalance = 0; + return fix_small_imbalance(sds, this_cpu, imbalance); + } + + if (!sds->group_imb) { + /* + * Don't want to pull so many tasks that a group would go idle. + */ + load_above_capacity = (sds->busiest_nr_running - + sds->busiest_group_capacity); + + load_above_capacity *= (SCHED_LOAD_SCALE * SCHED_LOAD_SCALE); + + load_above_capacity /= sds->busiest->cpu_power; + } + + /* + * We're trying to get all the cpus to the average_load, so we don't + * want to push ourselves above the average load, nor do we wish to + * reduce the max loaded cpu below the average load. At the same time, + * we also don't want to reduce the group load below the group capacity + * (so that we can implement power-savings policies etc). Thus we look + * for the minimum possible imbalance. + * Be careful of negative numbers as they'll appear as very large values + * with unsigned longs. + */ + max_pull = min(sds->max_load - sds->avg_load, load_above_capacity); + + /* How much load to actually move to equalise the imbalance */ + *imbalance = min(max_pull * sds->busiest->cpu_power, + (sds->avg_load - sds->this_load) * sds->this->cpu_power) + / SCHED_LOAD_SCALE; + + /* + * if *imbalance is less than the average load per runnable task + * there is no gaurantee that any tasks will be moved so we'll have + * a think about bumping its value to force at least one task to be + * moved + */ + if (*imbalance < sds->busiest_load_per_task) + return fix_small_imbalance(sds, this_cpu, imbalance); + +} +/******* find_busiest_group() helpers end here *********************/ + +/** + * find_busiest_group - Returns the busiest group within the sched_domain + * if there is an imbalance. If there isn't an imbalance, and + * the user has opted for power-savings, it returns a group whose + * CPUs can be put to idle by rebalancing those tasks elsewhere, if + * such a group exists. + * + * Also calculates the amount of weighted load which should be moved + * to restore balance. + * + * @sd: The sched_domain whose busiest group is to be returned. + * @this_cpu: The cpu for which load balancing is currently being performed. + * @imbalance: Variable which stores amount of weighted load which should + * be moved to restore balance/put a group to idle. + * @idle: The idle status of this_cpu. + * @sd_idle: The idleness of sd + * @cpus: The set of CPUs under consideration for load-balancing. + * @balance: Pointer to a variable indicating if this_cpu + * is the appropriate cpu to perform load balancing at this_level. + * + * Returns: - the busiest group if imbalance exists. + * - If no imbalance and user has opted for power-savings balance, + * return the least loaded group whose CPUs can be + * put to idle by rebalancing its tasks onto our group. + */ +static struct sched_group * +find_busiest_group(struct sched_domain *sd, int this_cpu, + unsigned long *imbalance, enum cpu_idle_type idle, + int *sd_idle, const struct cpumask *cpus, int *balance) +{ + struct sd_lb_stats sds; + + memset(&sds, 0, sizeof(sds)); + + /* + * Compute the various statistics relavent for load balancing at + * this level. + */ + update_sd_lb_stats(sd, this_cpu, idle, sd_idle, cpus, + balance, &sds); + + /* Cases where imbalance does not exist from POV of this_cpu */ + /* 1) this_cpu is not the appropriate cpu to perform load balancing + * at this level. + * 2) There is no busy sibling group to pull from. + * 3) This group is the busiest group. + * 4) This group is more busy than the avg busieness at this + * sched_domain. + * 5) The imbalance is within the specified limit. + */ + if (!(*balance)) + goto ret; + + if (!sds.busiest || sds.busiest_nr_running == 0) + goto out_balanced; + + if (sds.this_load >= sds.max_load) + goto out_balanced; + + sds.avg_load = (SCHED_LOAD_SCALE * sds.total_load) / sds.total_pwr; + + if (sds.this_load >= sds.avg_load) + goto out_balanced; + + if (100 * sds.max_load <= sd->imbalance_pct * sds.this_load) + goto out_balanced; + + /* Looks like there is an imbalance. Compute it */ + calculate_imbalance(&sds, this_cpu, imbalance); + return sds.busiest; + +out_balanced: + /* + * There is no obvious imbalance. But check if we can do some balancing + * to save power. + */ + if (check_power_save_busiest_group(&sds, this_cpu, imbalance)) + return sds.busiest; +ret: + *imbalance = 0; + return NULL; +} + +/* + * find_busiest_queue - find the busiest runqueue among the cpus in group. + */ +static struct rq * +find_busiest_queue(struct sched_group *group, enum cpu_idle_type idle, + unsigned long imbalance, const struct cpumask *cpus) +{ + struct rq *busiest = NULL, *rq; + unsigned long max_load = 0; + int i; + + for_each_cpu(i, sched_group_cpus(group)) { + unsigned long power = power_of(i); + unsigned long capacity = DIV_ROUND_CLOSEST(power, SCHED_LOAD_SCALE); + unsigned long wl; + + if (!cpumask_test_cpu(i, cpus)) + continue; + + rq = cpu_rq(i); + wl = weighted_cpuload(i); + + /* + * When comparing with imbalance, use weighted_cpuload() + * which is not scaled with the cpu power. + */ + if (capacity && rq->nr_running == 1 && wl > imbalance) + continue; + + /* + * For the load comparisons with the other cpu's, consider + * the weighted_cpuload() scaled with the cpu power, so that + * the load can be moved away from the cpu that is potentially + * running at a lower capacity. + */ + wl = (wl * SCHED_LOAD_SCALE) / power; + + if (wl > max_load) { + max_load = wl; + busiest = rq; + } + } + + return busiest; +} + +/* + * Max backoff if we encounter pinned tasks. Pretty arbitrary value, but + * so long as it is large enough. + */ +#define MAX_PINNED_INTERVAL 512 + +/* Working cpumask for load_balance and load_balance_newidle. */ +static DEFINE_PER_CPU(cpumask_var_t, load_balance_tmpmask); + +static int need_active_balance(struct sched_domain *sd, int sd_idle, int idle) +{ + if (idle == CPU_NEWLY_IDLE) { + /* + * The only task running in a non-idle cpu can be moved to this + * cpu in an attempt to completely freeup the other CPU + * package. + * + * The package power saving logic comes from + * find_busiest_group(). If there are no imbalance, then + * f_b_g() will return NULL. However when sched_mc={1,2} then + * f_b_g() will select a group from which a running task may be + * pulled to this cpu in order to make the other package idle. + * If there is no opportunity to make a package idle and if + * there are no imbalance, then f_b_g() will return NULL and no + * action will be taken in load_balance_newidle(). + * + * Under normal task pull operation due to imbalance, there + * will be more than one task in the source run queue and + * move_tasks() will succeed. ld_moved will be true and this + * active balance code will not be triggered. + */ + if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + return 0; + + if (sched_mc_power_savings < POWERSAVINGS_BALANCE_WAKEUP) + return 0; + } + + return unlikely(sd->nr_balance_failed > sd->cache_nice_tries+2); +} + +/* + * Check this_cpu to ensure it is balanced within domain. Attempt to move + * tasks if there is an imbalance. + */ +static int load_balance(int this_cpu, struct rq *this_rq, + struct sched_domain *sd, enum cpu_idle_type idle, + int *balance) +{ + int ld_moved, all_pinned = 0, active_balance = 0, sd_idle = 0; + struct sched_group *group; + unsigned long imbalance; + struct rq *busiest; + unsigned long flags; + struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask); + + cpumask_copy(cpus, cpu_active_mask); + + /* + * When power savings policy is enabled for the parent domain, idle + * sibling can pick up load irrespective of busy siblings. In this case, + * let the state of idle sibling percolate up as CPU_IDLE, instead of + * portraying it as CPU_NOT_IDLE. + */ + if (idle != CPU_NOT_IDLE && sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + sd_idle = 1; + + schedstat_inc(sd, lb_count[idle]); + +redo: + update_shares(sd); + group = find_busiest_group(sd, this_cpu, &imbalance, idle, &sd_idle, + cpus, balance); + + if (*balance == 0) + goto out_balanced; + + if (!group) { + schedstat_inc(sd, lb_nobusyg[idle]); + goto out_balanced; + } + + busiest = find_busiest_queue(group, idle, imbalance, cpus); + if (!busiest) { + schedstat_inc(sd, lb_nobusyq[idle]); + goto out_balanced; + } + + BUG_ON(busiest == this_rq); + + schedstat_add(sd, lb_imbalance[idle], imbalance); + + ld_moved = 0; + if (busiest->nr_running > 1) { + /* + * Attempt to move tasks. If find_busiest_group has found + * an imbalance but busiest->nr_running <= 1, the group is + * still unbalanced. ld_moved simply stays zero, so it is + * correctly treated as an imbalance. + */ + local_irq_save(flags); + double_rq_lock(this_rq, busiest); + ld_moved = move_tasks(this_rq, this_cpu, busiest, + imbalance, sd, idle, &all_pinned); + double_rq_unlock(this_rq, busiest); + local_irq_restore(flags); + + /* + * some other cpu did the load balance for us. + */ + if (ld_moved && this_cpu != smp_processor_id()) + resched_cpu(this_cpu); + + /* All tasks on this runqueue were pinned by CPU affinity */ + if (unlikely(all_pinned)) { + cpumask_clear_cpu(cpu_of(busiest), cpus); + if (!cpumask_empty(cpus)) + goto redo; + goto out_balanced; + } + } + + if (!ld_moved) { + schedstat_inc(sd, lb_failed[idle]); + sd->nr_balance_failed++; + + if (need_active_balance(sd, sd_idle, idle)) { + raw_spin_lock_irqsave(&busiest->lock, flags); + + /* don't kick the migration_thread, if the curr + * task on busiest cpu can't be moved to this_cpu + */ + if (!cpumask_test_cpu(this_cpu, + &busiest->curr->cpus_allowed)) { + raw_spin_unlock_irqrestore(&busiest->lock, + flags); + all_pinned = 1; + goto out_one_pinned; + } + + if (!busiest->active_balance) { + busiest->active_balance = 1; + busiest->push_cpu = this_cpu; + active_balance = 1; + } + raw_spin_unlock_irqrestore(&busiest->lock, flags); + if (active_balance) + wake_up_process(busiest->migration_thread); + + /* + * We've kicked active balancing, reset the failure + * counter. + */ + sd->nr_balance_failed = sd->cache_nice_tries+1; + } + } else + sd->nr_balance_failed = 0; + + if (likely(!active_balance)) { + /* We were unbalanced, so reset the balancing interval */ + sd->balance_interval = sd->min_interval; + } else { + /* + * If we've begun active balancing, start to back off. This + * case may not be covered by the all_pinned logic if there + * is only 1 task on the busy runqueue (because we don't call + * move_tasks). + */ + if (sd->balance_interval < sd->max_interval) + sd->balance_interval *= 2; + } + + if (!ld_moved && !sd_idle && sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + ld_moved = -1; + + goto out; + +out_balanced: + schedstat_inc(sd, lb_balanced[idle]); + + sd->nr_balance_failed = 0; + +out_one_pinned: + /* tune up the balancing interval */ + if ((all_pinned && sd->balance_interval < MAX_PINNED_INTERVAL) || + (sd->balance_interval < sd->max_interval)) + sd->balance_interval *= 2; + + if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && + !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) + ld_moved = -1; + else + ld_moved = 0; +out: + if (ld_moved) + update_shares(sd); + return ld_moved; +} + +/* + * idle_balance is called by schedule() if this_cpu is about to become + * idle. Attempts to pull tasks from other CPUs. + */ +static void idle_balance(int this_cpu, struct rq *this_rq) +{ + struct sched_domain *sd; + int pulled_task = 0; + unsigned long next_balance = jiffies + HZ; + + this_rq->idle_stamp = this_rq->clock; + + if (this_rq->avg_idle < sysctl_sched_migration_cost) + return; + + /* + * Drop the rq->lock, but keep IRQ/preempt disabled. + */ + raw_spin_unlock(&this_rq->lock); + + for_each_domain(this_cpu, sd) { + unsigned long interval; + int balance = 1; + + if (!(sd->flags & SD_LOAD_BALANCE)) + continue; + + if (sd->flags & SD_BALANCE_NEWIDLE) { + /* If we've pulled tasks over stop searching: */ + pulled_task = load_balance(this_cpu, this_rq, + sd, CPU_NEWLY_IDLE, &balance); + } + + interval = msecs_to_jiffies(sd->balance_interval); + if (time_after(next_balance, sd->last_balance + interval)) + next_balance = sd->last_balance + interval; + if (pulled_task) { + this_rq->idle_stamp = 0; + break; + } + } + + raw_spin_lock(&this_rq->lock); + + if (pulled_task || time_after(jiffies, this_rq->next_balance)) { + /* + * We are going idle. next_balance may be set based on + * a busy processor. So reset next_balance. + */ + this_rq->next_balance = next_balance; + } +} + +/* + * active_load_balance is run by migration threads. It pushes running tasks + * off the busiest CPU onto idle CPUs. It requires at least 1 task to be + * running on each physical CPU where possible, and avoids physical / + * logical imbalances. + * + * Called with busiest_rq locked. + */ +static void active_load_balance(struct rq *busiest_rq, int busiest_cpu) +{ + int target_cpu = busiest_rq->push_cpu; + struct sched_domain *sd; + struct rq *target_rq; + + /* Is there any task to move? */ + if (busiest_rq->nr_running <= 1) + return; + + target_rq = cpu_rq(target_cpu); + + /* + * This condition is "impossible", if it occurs + * we need to fix it. Originally reported by + * Bjorn Helgaas on a 128-cpu setup. + */ + BUG_ON(busiest_rq == target_rq); + + /* move a task from busiest_rq to target_rq */ + double_lock_balance(busiest_rq, target_rq); + update_rq_clock(busiest_rq); + update_rq_clock(target_rq); + + /* Search for an sd spanning us and the target CPU. */ + for_each_domain(target_cpu, sd) { + if ((sd->flags & SD_LOAD_BALANCE) && + cpumask_test_cpu(busiest_cpu, sched_domain_span(sd))) + break; + } + + if (likely(sd)) { + schedstat_inc(sd, alb_count); + + if (move_one_task(target_rq, target_cpu, busiest_rq, + sd, CPU_IDLE)) + schedstat_inc(sd, alb_pushed); + else + schedstat_inc(sd, alb_failed); + } + double_unlock_balance(busiest_rq, target_rq); +} + +#ifdef CONFIG_NO_HZ +static struct { + atomic_t load_balancer; + cpumask_var_t cpu_mask; + cpumask_var_t ilb_grp_nohz_mask; +} nohz ____cacheline_aligned = { + .load_balancer = ATOMIC_INIT(-1), +}; + +int get_nohz_load_balancer(void) +{ + return atomic_read(&nohz.load_balancer); +} + +#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) +/** + * lowest_flag_domain - Return lowest sched_domain containing flag. + * @cpu: The cpu whose lowest level of sched domain is to + * be returned. + * @flag: The flag to check for the lowest sched_domain + * for the given cpu. + * + * Returns the lowest sched_domain of a cpu which contains the given flag. + */ +static inline struct sched_domain *lowest_flag_domain(int cpu, int flag) +{ + struct sched_domain *sd; + + for_each_domain(cpu, sd) + if (sd && (sd->flags & flag)) + break; + + return sd; +} + +/** + * for_each_flag_domain - Iterates over sched_domains containing the flag. + * @cpu: The cpu whose domains we're iterating over. + * @sd: variable holding the value of the power_savings_sd + * for cpu. + * @flag: The flag to filter the sched_domains to be iterated. + * + * Iterates over all the scheduler domains for a given cpu that has the 'flag' + * set, starting from the lowest sched_domain to the highest. + */ +#define for_each_flag_domain(cpu, sd, flag) \ + for (sd = lowest_flag_domain(cpu, flag); \ + (sd && (sd->flags & flag)); sd = sd->parent) + +/** + * is_semi_idle_group - Checks if the given sched_group is semi-idle. + * @ilb_group: group to be checked for semi-idleness + * + * Returns: 1 if the group is semi-idle. 0 otherwise. + * + * We define a sched_group to be semi idle if it has atleast one idle-CPU + * and atleast one non-idle CPU. This helper function checks if the given + * sched_group is semi-idle or not. + */ +static inline int is_semi_idle_group(struct sched_group *ilb_group) +{ + cpumask_and(nohz.ilb_grp_nohz_mask, nohz.cpu_mask, + sched_group_cpus(ilb_group)); + + /* + * A sched_group is semi-idle when it has atleast one busy cpu + * and atleast one idle cpu. + */ + if (cpumask_empty(nohz.ilb_grp_nohz_mask)) + return 0; + + if (cpumask_equal(nohz.ilb_grp_nohz_mask, sched_group_cpus(ilb_group))) + return 0; + + return 1; +} +/** + * find_new_ilb - Finds the optimum idle load balancer for nomination. + * @cpu: The cpu which is nominating a new idle_load_balancer. + * + * Returns: Returns the id of the idle load balancer if it exists, + * Else, returns >= nr_cpu_ids. + * + * This algorithm picks the idle load balancer such that it belongs to a + * semi-idle powersavings sched_domain. The idea is to try and avoid + * completely idle packages/cores just for the purpose of idle load balancing + * when there are other idle cpu's which are better suited for that job. + */ +static int find_new_ilb(int cpu) +{ + struct sched_domain *sd; + struct sched_group *ilb_group; + + /* + * Have idle load balancer selection from semi-idle packages only + * when power-aware load balancing is enabled + */ + if (!(sched_smt_power_savings || sched_mc_power_savings)) + goto out_done; + + /* + * Optimize for the case when we have no idle CPUs or only one + * idle CPU. Don't walk the sched_domain hierarchy in such cases + */ + if (cpumask_weight(nohz.cpu_mask) < 2) + goto out_done; + + for_each_flag_domain(cpu, sd, SD_POWERSAVINGS_BALANCE) { + ilb_group = sd->groups; + + do { + if (is_semi_idle_group(ilb_group)) + return cpumask_first(nohz.ilb_grp_nohz_mask); + + ilb_group = ilb_group->next; + + } while (ilb_group != sd->groups); + } + +out_done: + return cpumask_first(nohz.cpu_mask); +} +#else /* (CONFIG_SCHED_MC || CONFIG_SCHED_SMT) */ +static inline int find_new_ilb(int call_cpu) +{ + return cpumask_first(nohz.cpu_mask); +} +#endif + +/* + * This routine will try to nominate the ilb (idle load balancing) + * owner among the cpus whose ticks are stopped. ilb owner will do the idle + * load balancing on behalf of all those cpus. If all the cpus in the system + * go into this tickless mode, then there will be no ilb owner (as there is + * no need for one) and all the cpus will sleep till the next wakeup event + * arrives... + * + * For the ilb owner, tick is not stopped. And this tick will be used + * for idle load balancing. ilb owner will still be part of + * nohz.cpu_mask.. + * + * While stopping the tick, this cpu will become the ilb owner if there + * is no other owner. And will be the owner till that cpu becomes busy + * or if all cpus in the system stop their ticks at which point + * there is no need for ilb owner. + * + * When the ilb owner becomes busy, it nominates another owner, during the + * next busy scheduler_tick() + */ +int select_nohz_load_balancer(int stop_tick) +{ + int cpu = smp_processor_id(); + + if (stop_tick) { + cpu_rq(cpu)->in_nohz_recently = 1; + + if (!cpu_active(cpu)) { + if (atomic_read(&nohz.load_balancer) != cpu) + return 0; + + /* + * If we are going offline and still the leader, + * give up! + */ + if (atomic_cmpxchg(&nohz.load_balancer, cpu, -1) != cpu) + BUG(); + + return 0; + } + + cpumask_set_cpu(cpu, nohz.cpu_mask); + + /* time for ilb owner also to sleep */ + if (cpumask_weight(nohz.cpu_mask) == num_active_cpus()) { + if (atomic_read(&nohz.load_balancer) == cpu) + atomic_set(&nohz.load_balancer, -1); + return 0; + } + + if (atomic_read(&nohz.load_balancer) == -1) { + /* make me the ilb owner */ + if (atomic_cmpxchg(&nohz.load_balancer, -1, cpu) == -1) + return 1; + } else if (atomic_read(&nohz.load_balancer) == cpu) { + int new_ilb; + + if (!(sched_smt_power_savings || + sched_mc_power_savings)) + return 1; + /* + * Check to see if there is a more power-efficient + * ilb. + */ + new_ilb = find_new_ilb(cpu); + if (new_ilb < nr_cpu_ids && new_ilb != cpu) { + atomic_set(&nohz.load_balancer, -1); + resched_cpu(new_ilb); + return 0; + } + return 1; + } + } else { + if (!cpumask_test_cpu(cpu, nohz.cpu_mask)) + return 0; + + cpumask_clear_cpu(cpu, nohz.cpu_mask); + + if (atomic_read(&nohz.load_balancer) == cpu) + if (atomic_cmpxchg(&nohz.load_balancer, cpu, -1) != cpu) + BUG(); + } + return 0; +} +#endif + +static DEFINE_SPINLOCK(balancing); + +/* + * It checks each scheduling domain to see if it is due to be balanced, + * and initiates a balancing operation if so. + * + * Balancing parameters are set up in arch_init_sched_domains. + */ +static void rebalance_domains(int cpu, enum cpu_idle_type idle) +{ + int balance = 1; + struct rq *rq = cpu_rq(cpu); + unsigned long interval; + struct sched_domain *sd; + /* Earliest time when we have to do rebalance again */ + unsigned long next_balance = jiffies + 60*HZ; + int update_next_balance = 0; + int need_serialize; + + for_each_domain(cpu, sd) { + if (!(sd->flags & SD_LOAD_BALANCE)) + continue; + + interval = sd->balance_interval; + if (idle != CPU_IDLE) + interval *= sd->busy_factor; + + /* scale ms to jiffies */ + interval = msecs_to_jiffies(interval); + if (unlikely(!interval)) + interval = 1; + if (interval > HZ*NR_CPUS/10) + interval = HZ*NR_CPUS/10; + + need_serialize = sd->flags & SD_SERIALIZE; + + if (need_serialize) { + if (!spin_trylock(&balancing)) + goto out; + } + + if (time_after_eq(jiffies, sd->last_balance + interval)) { + if (load_balance(cpu, rq, sd, idle, &balance)) { + /* + * We've pulled tasks over so either we're no + * longer idle, or one of our SMT siblings is + * not idle. + */ + idle = CPU_NOT_IDLE; + } + sd->last_balance = jiffies; + } + if (need_serialize) + spin_unlock(&balancing); +out: + if (time_after(next_balance, sd->last_balance + interval)) { + next_balance = sd->last_balance + interval; + update_next_balance = 1; + } + + /* + * Stop the load balance at this level. There is another + * CPU in our sched group which is doing load balancing more + * actively. + */ + if (!balance) + break; + } + + /* + * next_balance will be updated only when there is a need. + * When the cpu is attached to null domain for ex, it will not be + * updated. + */ + if (likely(update_next_balance)) + rq->next_balance = next_balance; +} + +/* + * run_rebalance_domains is triggered when needed from the scheduler tick. + * In CONFIG_NO_HZ case, the idle load balance owner will do the + * rebalancing for all the cpus for whom scheduler ticks are stopped. + */ +static void run_rebalance_domains(struct softirq_action *h) +{ + int this_cpu = smp_processor_id(); + struct rq *this_rq = cpu_rq(this_cpu); + enum cpu_idle_type idle = this_rq->idle_at_tick ? + CPU_IDLE : CPU_NOT_IDLE; + + rebalance_domains(this_cpu, idle); + +#ifdef CONFIG_NO_HZ + /* + * If this cpu is the owner for idle load balancing, then do the + * balancing on behalf of the other idle cpus whose ticks are + * stopped. + */ + if (this_rq->idle_at_tick && + atomic_read(&nohz.load_balancer) == this_cpu) { + struct rq *rq; + int balance_cpu; + + for_each_cpu(balance_cpu, nohz.cpu_mask) { + if (balance_cpu == this_cpu) + continue; + + /* + * If this cpu gets work to do, stop the load balancing + * work being done for other cpus. Next load + * balancing owner will pick it up. + */ + if (need_resched()) + break; + + rebalance_domains(balance_cpu, CPU_IDLE); + + rq = cpu_rq(balance_cpu); + if (time_after(this_rq->next_balance, rq->next_balance)) + this_rq->next_balance = rq->next_balance; + } + } +#endif +} + +static inline int on_null_domain(int cpu) +{ + return !rcu_dereference(cpu_rq(cpu)->sd); +} + +/* + * Trigger the SCHED_SOFTIRQ if it is time to do periodic load balancing. + * + * In case of CONFIG_NO_HZ, this is the place where we nominate a new + * idle load balancing owner or decide to stop the periodic load balancing, + * if the whole system is idle. + */ +static inline void trigger_load_balance(struct rq *rq, int cpu) +{ +#ifdef CONFIG_NO_HZ + /* + * If we were in the nohz mode recently and busy at the current + * scheduler tick, then check if we need to nominate new idle + * load balancer. + */ + if (rq->in_nohz_recently && !rq->idle_at_tick) { + rq->in_nohz_recently = 0; + + if (atomic_read(&nohz.load_balancer) == cpu) { + cpumask_clear_cpu(cpu, nohz.cpu_mask); + atomic_set(&nohz.load_balancer, -1); + } + + if (atomic_read(&nohz.load_balancer) == -1) { + int ilb = find_new_ilb(cpu); + + if (ilb < nr_cpu_ids) + resched_cpu(ilb); + } + } + + /* + * If this cpu is idle and doing idle load balancing for all the + * cpus with ticks stopped, is it time for that to stop? + */ + if (rq->idle_at_tick && atomic_read(&nohz.load_balancer) == cpu && + cpumask_weight(nohz.cpu_mask) == num_online_cpus()) { + resched_cpu(cpu); + return; + } + + /* + * If this cpu is idle and the idle load balancing is done by + * someone else, then no need raise the SCHED_SOFTIRQ + */ + if (rq->idle_at_tick && atomic_read(&nohz.load_balancer) != cpu && + cpumask_test_cpu(cpu, nohz.cpu_mask)) + return; +#endif + /* Don't need to rebalance while attached to NULL domain */ + if (time_after_eq(jiffies, rq->next_balance) && + likely(!on_null_domain(cpu))) + raise_softirq(SCHED_SOFTIRQ); +} static void rq_online_fair(struct rq *rq) { @@ -1962,6 +3544,15 @@ static void rq_offline_fair(struct rq *rq) update_sysctl(); } +#else /* CONFIG_SMP */ + +/* + * on UP we do not need to balance between CPUs: + */ +static inline void idle_balance(int cpu, struct rq *rq) +{ +} + #endif /* CONFIG_SMP */ /* @@ -2076,7 +3667,7 @@ static void moved_group_fair(struct task_struct *p, int on_rq) } #endif -unsigned int get_rr_interval_fair(struct rq *rq, struct task_struct *task) +static unsigned int get_rr_interval_fair(struct rq *rq, struct task_struct *task) { struct sched_entity *se = &task->se; unsigned int rr_interval = 0; @@ -2108,8 +3699,6 @@ static const struct sched_class fair_sched_class = { #ifdef CONFIG_SMP .select_task_rq = select_task_rq_fair, - .load_balance = load_balance_fair, - .move_one_task = move_one_task_fair, .rq_online = rq_online_fair, .rq_offline = rq_offline_fair, diff --git a/kernel/sched_idletask.c b/kernel/sched_idletask.c index 5f93b570d383..a8a6d8a50947 100644 --- a/kernel/sched_idletask.c +++ b/kernel/sched_idletask.c @@ -44,24 +44,6 @@ static void put_prev_task_idle(struct rq *rq, struct task_struct *prev) { } -#ifdef CONFIG_SMP -static unsigned long -load_balance_idle(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned, int *this_best_prio) -{ - return 0; -} - -static int -move_one_task_idle(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle) -{ - return 0; -} -#endif - static void task_tick_idle(struct rq *rq, struct task_struct *curr, int queued) { } @@ -97,7 +79,7 @@ static void prio_changed_idle(struct rq *rq, struct task_struct *p, check_preempt_curr(rq, p, 0); } -unsigned int get_rr_interval_idle(struct rq *rq, struct task_struct *task) +static unsigned int get_rr_interval_idle(struct rq *rq, struct task_struct *task) { return 0; } @@ -119,9 +101,6 @@ static const struct sched_class idle_sched_class = { #ifdef CONFIG_SMP .select_task_rq = select_task_rq_idle, - - .load_balance = load_balance_idle, - .move_one_task = move_one_task_idle, #endif .set_curr_task = set_curr_task_idle, diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index f48328ac216f..bf3e38fdbe6d 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -194,17 +194,20 @@ static inline struct rt_rq *group_rt_rq(struct sched_rt_entity *rt_se) return rt_se->my_q; } -static void enqueue_rt_entity(struct sched_rt_entity *rt_se); +static void enqueue_rt_entity(struct sched_rt_entity *rt_se, bool head); static void dequeue_rt_entity(struct sched_rt_entity *rt_se); static void sched_rt_rq_enqueue(struct rt_rq *rt_rq) { + int this_cpu = smp_processor_id(); struct task_struct *curr = rq_of_rt_rq(rt_rq)->curr; - struct sched_rt_entity *rt_se = rt_rq->rt_se; + struct sched_rt_entity *rt_se; + + rt_se = rt_rq->tg->rt_se[this_cpu]; if (rt_rq->rt_nr_running) { if (rt_se && !on_rt_rq(rt_se)) - enqueue_rt_entity(rt_se); + enqueue_rt_entity(rt_se, false); if (rt_rq->highest_prio.curr < curr->prio) resched_task(curr); } @@ -212,7 +215,10 @@ static void sched_rt_rq_enqueue(struct rt_rq *rt_rq) static void sched_rt_rq_dequeue(struct rt_rq *rt_rq) { - struct sched_rt_entity *rt_se = rt_rq->rt_se; + int this_cpu = smp_processor_id(); + struct sched_rt_entity *rt_se; + + rt_se = rt_rq->tg->rt_se[this_cpu]; if (rt_se && on_rt_rq(rt_se)) dequeue_rt_entity(rt_se); @@ -803,7 +809,7 @@ void dec_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) dec_rt_group(rt_se, rt_rq); } -static void __enqueue_rt_entity(struct sched_rt_entity *rt_se) +static void __enqueue_rt_entity(struct sched_rt_entity *rt_se, bool head) { struct rt_rq *rt_rq = rt_rq_of_se(rt_se); struct rt_prio_array *array = &rt_rq->active; @@ -819,7 +825,10 @@ static void __enqueue_rt_entity(struct sched_rt_entity *rt_se) if (group_rq && (rt_rq_throttled(group_rq) || !group_rq->rt_nr_running)) return; - list_add_tail(&rt_se->run_list, queue); + if (head) + list_add(&rt_se->run_list, queue); + else + list_add_tail(&rt_se->run_list, queue); __set_bit(rt_se_prio(rt_se), array->bitmap); inc_rt_tasks(rt_se, rt_rq); @@ -856,11 +865,11 @@ static void dequeue_rt_stack(struct sched_rt_entity *rt_se) } } -static void enqueue_rt_entity(struct sched_rt_entity *rt_se) +static void enqueue_rt_entity(struct sched_rt_entity *rt_se, bool head) { dequeue_rt_stack(rt_se); for_each_sched_rt_entity(rt_se) - __enqueue_rt_entity(rt_se); + __enqueue_rt_entity(rt_se, head); } static void dequeue_rt_entity(struct sched_rt_entity *rt_se) @@ -871,21 +880,22 @@ static void dequeue_rt_entity(struct sched_rt_entity *rt_se) struct rt_rq *rt_rq = group_rt_rq(rt_se); if (rt_rq && rt_rq->rt_nr_running) - __enqueue_rt_entity(rt_se); + __enqueue_rt_entity(rt_se, false); } } /* * Adding/removing a task to/from a priority array: */ -static void enqueue_task_rt(struct rq *rq, struct task_struct *p, int wakeup) +static void +enqueue_task_rt(struct rq *rq, struct task_struct *p, int wakeup, bool head) { struct sched_rt_entity *rt_se = &p->rt; if (wakeup) rt_se->timeout = 0; - enqueue_rt_entity(rt_se); + enqueue_rt_entity(rt_se, head); if (!task_current(rq, p) && p->rt.nr_cpus_allowed > 1) enqueue_pushable_task(rq, p); @@ -1481,24 +1491,6 @@ static void task_woken_rt(struct rq *rq, struct task_struct *p) push_rt_tasks(rq); } -static unsigned long -load_balance_rt(struct rq *this_rq, int this_cpu, struct rq *busiest, - unsigned long max_load_move, - struct sched_domain *sd, enum cpu_idle_type idle, - int *all_pinned, int *this_best_prio) -{ - /* don't touch RT tasks */ - return 0; -} - -static int -move_one_task_rt(struct rq *this_rq, int this_cpu, struct rq *busiest, - struct sched_domain *sd, enum cpu_idle_type idle) -{ - /* don't touch RT tasks */ - return 0; -} - static void set_cpus_allowed_rt(struct task_struct *p, const struct cpumask *new_mask) { @@ -1721,7 +1713,7 @@ static void set_curr_task_rt(struct rq *rq) dequeue_pushable_task(rq, p); } -unsigned int get_rr_interval_rt(struct rq *rq, struct task_struct *task) +static unsigned int get_rr_interval_rt(struct rq *rq, struct task_struct *task) { /* * Time slice is 0 for SCHED_FIFO tasks @@ -1746,8 +1738,6 @@ static const struct sched_class rt_sched_class = { #ifdef CONFIG_SMP .select_task_rq = select_task_rq_rt, - .load_balance = load_balance_rt, - .move_one_task = move_one_task_rt, .set_cpus_allowed = set_cpus_allowed_rt, .rq_online = rq_online_rt, .rq_offline = rq_offline_rt, diff --git a/kernel/smp.c b/kernel/smp.c index f10408422444..9867b6bfefce 100644 --- a/kernel/smp.c +++ b/kernel/smp.c @@ -12,8 +12,6 @@ #include <linux/smp.h> #include <linux/cpu.h> -static DEFINE_PER_CPU(struct call_single_queue, call_single_queue); - static struct { struct list_head queue; raw_spinlock_t lock; @@ -33,12 +31,14 @@ struct call_function_data { cpumask_var_t cpumask; }; +static DEFINE_PER_CPU_SHARED_ALIGNED(struct call_function_data, cfd_data); + struct call_single_queue { struct list_head list; raw_spinlock_t lock; }; -static DEFINE_PER_CPU(struct call_function_data, cfd_data); +static DEFINE_PER_CPU_SHARED_ALIGNED(struct call_single_queue, call_single_queue); static int hotplug_cfd(struct notifier_block *nfb, unsigned long action, void *hcpu) @@ -256,7 +256,7 @@ void generic_smp_call_function_single_interrupt(void) } } -static DEFINE_PER_CPU(struct call_single_data, csd_data); +static DEFINE_PER_CPU_SHARED_ALIGNED(struct call_single_data, csd_data); /* * smp_call_function_single - Run a function on a specific CPU diff --git a/kernel/srcu.c b/kernel/srcu.c index 818d7d9aa03c..bde4295774c8 100644 --- a/kernel/srcu.c +++ b/kernel/srcu.c @@ -34,6 +34,30 @@ #include <linux/smp.h> #include <linux/srcu.h> +static int init_srcu_struct_fields(struct srcu_struct *sp) +{ + sp->completed = 0; + mutex_init(&sp->mutex); + sp->per_cpu_ref = alloc_percpu(struct srcu_struct_array); + return sp->per_cpu_ref ? 0 : -ENOMEM; +} + +#ifdef CONFIG_DEBUG_LOCK_ALLOC + +int __init_srcu_struct(struct srcu_struct *sp, const char *name, + struct lock_class_key *key) +{ +#ifdef CONFIG_DEBUG_LOCK_ALLOC + /* Don't re-initialize a lock while it is held. */ + debug_check_no_locks_freed((void *)sp, sizeof(*sp)); + lockdep_init_map(&sp->dep_map, name, key, 0); +#endif /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */ + return init_srcu_struct_fields(sp); +} +EXPORT_SYMBOL_GPL(__init_srcu_struct); + +#else /* #ifdef CONFIG_DEBUG_LOCK_ALLOC */ + /** * init_srcu_struct - initialize a sleep-RCU structure * @sp: structure to initialize. @@ -44,13 +68,12 @@ */ int init_srcu_struct(struct srcu_struct *sp) { - sp->completed = 0; - mutex_init(&sp->mutex); - sp->per_cpu_ref = alloc_percpu(struct srcu_struct_array); - return (sp->per_cpu_ref ? 0 : -ENOMEM); + return init_srcu_struct_fields(sp); } EXPORT_SYMBOL_GPL(init_srcu_struct); +#endif /* #else #ifdef CONFIG_DEBUG_LOCK_ALLOC */ + /* * srcu_readers_active_idx -- returns approximate number of readers * active on the specified rank of per-CPU counters. @@ -100,15 +123,12 @@ void cleanup_srcu_struct(struct srcu_struct *sp) } EXPORT_SYMBOL_GPL(cleanup_srcu_struct); -/** - * srcu_read_lock - register a new reader for an SRCU-protected structure. - * @sp: srcu_struct in which to register the new reader. - * +/* * Counts the new reader in the appropriate per-CPU element of the * srcu_struct. Must be called from process context. * Returns an index that must be passed to the matching srcu_read_unlock(). */ -int srcu_read_lock(struct srcu_struct *sp) +int __srcu_read_lock(struct srcu_struct *sp) { int idx; @@ -120,31 +140,27 @@ int srcu_read_lock(struct srcu_struct *sp) preempt_enable(); return idx; } -EXPORT_SYMBOL_GPL(srcu_read_lock); +EXPORT_SYMBOL_GPL(__srcu_read_lock); -/** - * srcu_read_unlock - unregister a old reader from an SRCU-protected structure. - * @sp: srcu_struct in which to unregister the old reader. - * @idx: return value from corresponding srcu_read_lock(). - * +/* * Removes the count for the old reader from the appropriate per-CPU * element of the srcu_struct. Note that this may well be a different * CPU than that which was incremented by the corresponding srcu_read_lock(). * Must be called from process context. */ -void srcu_read_unlock(struct srcu_struct *sp, int idx) +void __srcu_read_unlock(struct srcu_struct *sp, int idx) { preempt_disable(); srcu_barrier(); /* ensure compiler won't misorder critical section. */ per_cpu_ptr(sp->per_cpu_ref, smp_processor_id())->c[idx]--; preempt_enable(); } -EXPORT_SYMBOL_GPL(srcu_read_unlock); +EXPORT_SYMBOL_GPL(__srcu_read_unlock); /* * Helper function for synchronize_srcu() and synchronize_srcu_expedited(). */ -void __synchronize_srcu(struct srcu_struct *sp, void (*sync_func)(void)) +static void __synchronize_srcu(struct srcu_struct *sp, void (*sync_func)(void)) { int idx; diff --git a/kernel/sys.c b/kernel/sys.c index 26a6b73a6b85..877fe4f8e05e 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -222,6 +222,7 @@ SYSCALL_DEFINE2(getpriority, int, which, int, who) if (which > PRIO_USER || which < PRIO_PROCESS) return -EINVAL; + rcu_read_lock(); read_lock(&tasklist_lock); switch (which) { case PRIO_PROCESS: @@ -267,6 +268,7 @@ SYSCALL_DEFINE2(getpriority, int, which, int, who) } out_unlock: read_unlock(&tasklist_lock); + rcu_read_unlock(); return retval; } @@ -569,11 +571,6 @@ static int set_user(struct cred *new) if (!new_user) return -EAGAIN; - if (!task_can_switch_user(new_user, current)) { - free_uid(new_user); - return -EINVAL; - } - if (atomic_read(&new_user->processes) >= current->signal->rlim[RLIMIT_NPROC].rlim_cur && new_user != INIT_USER) { diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 60e2ce0181ee..13e13d428cd3 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -328,15 +328,6 @@ config BRANCH_TRACER Say N if unsure. -config POWER_TRACER - bool "Trace power consumption behavior" - depends on X86 - select GENERIC_TRACER - help - This tracer helps developers to analyze and optimize the kernel's - power management decisions, specifically the C-state and P-state - behavior. - config KSYM_TRACER bool "Trace read and write access on kernel memory locations" depends on HAVE_HW_BREAKPOINT @@ -449,7 +440,7 @@ config BLK_DEV_IO_TRACE config KPROBE_EVENT depends on KPROBES - depends on X86 + depends on HAVE_REGS_AND_STACK_ACCESS_API bool "Enable kprobes-based dynamic events" select TRACING default y diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index cd9ecd89ec77..d00c6fe23f54 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -51,7 +51,9 @@ endif obj-$(CONFIG_EVENT_TRACING) += trace_events.o obj-$(CONFIG_EVENT_TRACING) += trace_export.o obj-$(CONFIG_FTRACE_SYSCALLS) += trace_syscalls.o -obj-$(CONFIG_EVENT_PROFILE) += trace_event_profile.o +ifeq ($(CONFIG_PERF_EVENTS),y) +obj-$(CONFIG_EVENT_TRACING) += trace_event_profile.o +endif obj-$(CONFIG_EVENT_TRACING) += trace_events_filter.o obj-$(CONFIG_KPROBE_EVENT) += trace_kprobe.o obj-$(CONFIG_KSYM_TRACER) += trace_ksym.o diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 1e6640f80454..83783579378f 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -22,7 +22,6 @@ #include <linux/hardirq.h> #include <linux/kthread.h> #include <linux/uaccess.h> -#include <linux/kprobes.h> #include <linux/ftrace.h> #include <linux/sysctl.h> #include <linux/ctype.h> @@ -898,36 +897,6 @@ static struct dyn_ftrace *ftrace_free_records; } \ } -#ifdef CONFIG_KPROBES - -static int frozen_record_count; - -static inline void freeze_record(struct dyn_ftrace *rec) -{ - if (!(rec->flags & FTRACE_FL_FROZEN)) { - rec->flags |= FTRACE_FL_FROZEN; - frozen_record_count++; - } -} - -static inline void unfreeze_record(struct dyn_ftrace *rec) -{ - if (rec->flags & FTRACE_FL_FROZEN) { - rec->flags &= ~FTRACE_FL_FROZEN; - frozen_record_count--; - } -} - -static inline int record_frozen(struct dyn_ftrace *rec) -{ - return rec->flags & FTRACE_FL_FROZEN; -} -#else -# define freeze_record(rec) ({ 0; }) -# define unfreeze_record(rec) ({ 0; }) -# define record_frozen(rec) ({ 0; }) -#endif /* CONFIG_KPROBES */ - static void ftrace_free_rec(struct dyn_ftrace *rec) { rec->freelist = ftrace_free_records; @@ -1025,6 +994,21 @@ static void ftrace_bug(int failed, unsigned long ip) } +/* Return 1 if the address range is reserved for ftrace */ +int ftrace_text_reserved(void *start, void *end) +{ + struct dyn_ftrace *rec; + struct ftrace_page *pg; + + do_for_each_ftrace_rec(pg, rec) { + if (rec->ip <= (unsigned long)end && + rec->ip + MCOUNT_INSN_SIZE > (unsigned long)start) + return 1; + } while_for_each_ftrace_rec(); + return 0; +} + + static int __ftrace_replace_code(struct dyn_ftrace *rec, int enable) { @@ -1076,14 +1060,6 @@ static void ftrace_replace_code(int enable) !(rec->flags & FTRACE_FL_CONVERTED)) continue; - /* ignore updates to this record's mcount site */ - if (get_kprobe((void *)rec->ip)) { - freeze_record(rec); - continue; - } else { - unfreeze_record(rec); - } - failed = __ftrace_replace_code(rec, enable); if (failed) { rec->flags |= FTRACE_FL_FAILED; @@ -2426,6 +2402,7 @@ static const struct file_operations ftrace_notrace_fops = { static DEFINE_MUTEX(graph_lock); int ftrace_graph_count; +int ftrace_graph_filter_enabled; unsigned long ftrace_graph_funcs[FTRACE_GRAPH_MAX_FUNCS] __read_mostly; static void * @@ -2448,7 +2425,7 @@ static void *g_start(struct seq_file *m, loff_t *pos) mutex_lock(&graph_lock); /* Nothing, tell g_show to print all functions are enabled */ - if (!ftrace_graph_count && !*pos) + if (!ftrace_graph_filter_enabled && !*pos) return (void *)1; return __g_next(m, pos); @@ -2494,6 +2471,7 @@ ftrace_graph_open(struct inode *inode, struct file *file) mutex_lock(&graph_lock); if ((file->f_mode & FMODE_WRITE) && (file->f_flags & O_TRUNC)) { + ftrace_graph_filter_enabled = 0; ftrace_graph_count = 0; memset(ftrace_graph_funcs, 0, sizeof(ftrace_graph_funcs)); } @@ -2519,7 +2497,7 @@ ftrace_set_func(unsigned long *array, int *idx, char *buffer) struct dyn_ftrace *rec; struct ftrace_page *pg; int search_len; - int found = 0; + int fail = 1; int type, not; char *search; bool exists; @@ -2530,37 +2508,51 @@ ftrace_set_func(unsigned long *array, int *idx, char *buffer) /* decode regex */ type = filter_parse_regex(buffer, strlen(buffer), &search, ¬); - if (not) - return -EINVAL; + if (!not && *idx >= FTRACE_GRAPH_MAX_FUNCS) + return -EBUSY; search_len = strlen(search); mutex_lock(&ftrace_lock); do_for_each_ftrace_rec(pg, rec) { - if (*idx >= FTRACE_GRAPH_MAX_FUNCS) - break; - if (rec->flags & (FTRACE_FL_FAILED | FTRACE_FL_FREE)) continue; if (ftrace_match_record(rec, search, search_len, type)) { - /* ensure it is not already in the array */ + /* if it is in the array */ exists = false; - for (i = 0; i < *idx; i++) + for (i = 0; i < *idx; i++) { if (array[i] == rec->ip) { exists = true; break; } - if (!exists) - array[(*idx)++] = rec->ip; - found = 1; + } + + if (!not) { + fail = 0; + if (!exists) { + array[(*idx)++] = rec->ip; + if (*idx >= FTRACE_GRAPH_MAX_FUNCS) + goto out; + } + } else { + if (exists) { + array[i] = array[--(*idx)]; + array[*idx] = 0; + fail = 0; + } + } } } while_for_each_ftrace_rec(); - +out: mutex_unlock(&ftrace_lock); - return found ? 0 : -EINVAL; + if (fail) + return -EINVAL; + + ftrace_graph_filter_enabled = 1; + return 0; } static ssize_t @@ -2570,16 +2562,11 @@ ftrace_graph_write(struct file *file, const char __user *ubuf, struct trace_parser parser; ssize_t read, ret; - if (!cnt || cnt < 0) + if (!cnt) return 0; mutex_lock(&graph_lock); - if (ftrace_graph_count >= FTRACE_GRAPH_MAX_FUNCS) { - ret = -EBUSY; - goto out_unlock; - } - if (trace_parser_get_init(&parser, FTRACE_BUFF_MAX)) { ret = -ENOMEM; goto out_unlock; diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index eac6875cb990..032c57ca6502 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -32,6 +32,7 @@ #include <linux/splice.h> #include <linux/kdebug.h> #include <linux/string.h> +#include <linux/rwsem.h> #include <linux/ctype.h> #include <linux/init.h> #include <linux/poll.h> @@ -102,9 +103,6 @@ static inline void ftrace_enable_cpu(void) static cpumask_var_t __read_mostly tracing_buffer_mask; -/* Define which cpu buffers are currently read in trace_pipe */ -static cpumask_var_t tracing_reader_cpumask; - #define for_each_tracing_cpu(cpu) \ for_each_cpu(cpu, tracing_buffer_mask) @@ -243,12 +241,91 @@ static struct tracer *current_trace __read_mostly; /* * trace_types_lock is used to protect the trace_types list. - * This lock is also used to keep user access serialized. - * Accesses from userspace will grab this lock while userspace - * activities happen inside the kernel. */ static DEFINE_MUTEX(trace_types_lock); +/* + * serialize the access of the ring buffer + * + * ring buffer serializes readers, but it is low level protection. + * The validity of the events (which returns by ring_buffer_peek() ..etc) + * are not protected by ring buffer. + * + * The content of events may become garbage if we allow other process consumes + * these events concurrently: + * A) the page of the consumed events may become a normal page + * (not reader page) in ring buffer, and this page will be rewrited + * by events producer. + * B) The page of the consumed events may become a page for splice_read, + * and this page will be returned to system. + * + * These primitives allow multi process access to different cpu ring buffer + * concurrently. + * + * These primitives don't distinguish read-only and read-consume access. + * Multi read-only access are also serialized. + */ + +#ifdef CONFIG_SMP +static DECLARE_RWSEM(all_cpu_access_lock); +static DEFINE_PER_CPU(struct mutex, cpu_access_lock); + +static inline void trace_access_lock(int cpu) +{ + if (cpu == TRACE_PIPE_ALL_CPU) { + /* gain it for accessing the whole ring buffer. */ + down_write(&all_cpu_access_lock); + } else { + /* gain it for accessing a cpu ring buffer. */ + + /* Firstly block other trace_access_lock(TRACE_PIPE_ALL_CPU). */ + down_read(&all_cpu_access_lock); + + /* Secondly block other access to this @cpu ring buffer. */ + mutex_lock(&per_cpu(cpu_access_lock, cpu)); + } +} + +static inline void trace_access_unlock(int cpu) +{ + if (cpu == TRACE_PIPE_ALL_CPU) { + up_write(&all_cpu_access_lock); + } else { + mutex_unlock(&per_cpu(cpu_access_lock, cpu)); + up_read(&all_cpu_access_lock); + } +} + +static inline void trace_access_lock_init(void) +{ + int cpu; + + for_each_possible_cpu(cpu) + mutex_init(&per_cpu(cpu_access_lock, cpu)); +} + +#else + +static DEFINE_MUTEX(access_lock); + +static inline void trace_access_lock(int cpu) +{ + (void)cpu; + mutex_lock(&access_lock); +} + +static inline void trace_access_unlock(int cpu) +{ + (void)cpu; + mutex_unlock(&access_lock); +} + +static inline void trace_access_lock_init(void) +{ +} + +#endif + /* trace_wait is a waitqueue for tasks blocked on trace_poll */ static DECLARE_WAIT_QUEUE_HEAD(trace_wait); @@ -1320,8 +1397,10 @@ int trace_vbprintk(unsigned long ip, const char *fmt, va_list args) entry->fmt = fmt; memcpy(entry->buf, trace_buf, sizeof(u32) * len); - if (!filter_check_discard(call, entry, buffer, event)) + if (!filter_check_discard(call, entry, buffer, event)) { ring_buffer_unlock_commit(buffer, event); + ftrace_trace_stack(buffer, flags, 6, pc); + } out_unlock: arch_spin_unlock(&trace_buf_lock); @@ -1394,8 +1473,10 @@ int trace_array_vprintk(struct trace_array *tr, memcpy(&entry->buf, trace_buf, len); entry->buf[len] = '\0'; - if (!filter_check_discard(call, entry, buffer, event)) + if (!filter_check_discard(call, entry, buffer, event)) { ring_buffer_unlock_commit(buffer, event); + ftrace_trace_stack(buffer, irq_flags, 6, pc); + } out_unlock: arch_spin_unlock(&trace_buf_lock); @@ -1585,12 +1666,6 @@ static void tracing_iter_reset(struct trace_iterator *iter, int cpu) } /* - * No necessary locking here. The worst thing which can - * happen is loosing events consumed at the same time - * by a trace_pipe reader. - * Other than that, we don't risk to crash the ring buffer - * because it serializes the readers. - * * The current tracer is copied to avoid a global locking * all around. */ @@ -1645,12 +1720,16 @@ static void *s_start(struct seq_file *m, loff_t *pos) } trace_event_read_lock(); + trace_access_lock(cpu_file); return p; } static void s_stop(struct seq_file *m, void *p) { + struct trace_iterator *iter = m->private; + atomic_dec(&trace_record_cmdline_disabled); + trace_access_unlock(iter->cpu_file); trace_event_read_unlock(); } @@ -2841,22 +2920,6 @@ static int tracing_open_pipe(struct inode *inode, struct file *filp) mutex_lock(&trace_types_lock); - /* We only allow one reader per cpu */ - if (cpu_file == TRACE_PIPE_ALL_CPU) { - if (!cpumask_empty(tracing_reader_cpumask)) { - ret = -EBUSY; - goto out; - } - cpumask_setall(tracing_reader_cpumask); - } else { - if (!cpumask_test_cpu(cpu_file, tracing_reader_cpumask)) - cpumask_set_cpu(cpu_file, tracing_reader_cpumask); - else { - ret = -EBUSY; - goto out; - } - } - /* create a buffer to store the information to pass to userspace */ iter = kzalloc(sizeof(*iter), GFP_KERNEL); if (!iter) { @@ -2912,12 +2975,6 @@ static int tracing_release_pipe(struct inode *inode, struct file *file) mutex_lock(&trace_types_lock); - if (iter->cpu_file == TRACE_PIPE_ALL_CPU) - cpumask_clear(tracing_reader_cpumask); - else - cpumask_clear_cpu(iter->cpu_file, tracing_reader_cpumask); - - if (iter->trace->pipe_close) iter->trace->pipe_close(iter); @@ -3079,6 +3136,7 @@ waitagain: iter->pos = -1; trace_event_read_lock(); + trace_access_lock(iter->cpu_file); while (find_next_entry_inc(iter) != NULL) { enum print_line_t ret; int len = iter->seq.len; @@ -3095,6 +3153,7 @@ waitagain: if (iter->seq.len >= cnt) break; } + trace_access_unlock(iter->cpu_file); trace_event_read_unlock(); /* Now copy what we have to the user */ @@ -3220,6 +3279,7 @@ static ssize_t tracing_splice_read_pipe(struct file *filp, } trace_event_read_lock(); + trace_access_lock(iter->cpu_file); /* Fill as many pages as possible. */ for (i = 0, rem = len; i < PIPE_BUFFERS && rem; i++) { @@ -3243,6 +3303,7 @@ static ssize_t tracing_splice_read_pipe(struct file *filp, trace_seq_init(&iter->seq); } + trace_access_unlock(iter->cpu_file); trace_event_read_unlock(); mutex_unlock(&iter->mutex); @@ -3544,10 +3605,12 @@ tracing_buffers_read(struct file *filp, char __user *ubuf, info->read = 0; + trace_access_lock(info->cpu); ret = ring_buffer_read_page(info->tr->buffer, &info->spare, count, info->cpu, 0); + trace_access_unlock(info->cpu); if (ret < 0) return 0; @@ -3675,6 +3738,7 @@ tracing_buffers_splice_read(struct file *file, loff_t *ppos, len &= PAGE_MASK; } + trace_access_lock(info->cpu); entries = ring_buffer_entries_cpu(info->tr->buffer, info->cpu); for (i = 0; i < PIPE_BUFFERS && len && entries; i++, len -= PAGE_SIZE) { @@ -3722,6 +3786,7 @@ tracing_buffers_splice_read(struct file *file, loff_t *ppos, entries = ring_buffer_entries_cpu(info->tr->buffer, info->cpu); } + trace_access_unlock(info->cpu); spd.nr_pages = i; /* did we read anything? */ @@ -4158,6 +4223,8 @@ static __init int tracer_init_debugfs(void) struct dentry *d_tracer; int cpu; + trace_access_lock_init(); + d_tracer = tracing_init_dentry(); trace_create_file("tracing_enabled", 0644, d_tracer, @@ -4392,9 +4459,6 @@ __init static int tracer_alloc_buffers(void) if (!alloc_cpumask_var(&tracing_cpumask, GFP_KERNEL)) goto out_free_buffer_mask; - if (!zalloc_cpumask_var(&tracing_reader_cpumask, GFP_KERNEL)) - goto out_free_tracing_cpumask; - /* To save memory, keep the ring buffer size to its minimum */ if (ring_buffer_expanded) ring_buf_size = trace_buf_size; @@ -4452,8 +4516,6 @@ __init static int tracer_alloc_buffers(void) return 0; out_free_cpumask: - free_cpumask_var(tracing_reader_cpumask); -out_free_tracing_cpumask: free_cpumask_var(tracing_cpumask); out_free_buffer_mask: free_cpumask_var(tracing_buffer_mask); diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 4df6a77eb196..fd05bcaf91b0 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -497,6 +497,7 @@ trace_print_graph_duration(unsigned long long duration, struct trace_seq *s); #ifdef CONFIG_DYNAMIC_FTRACE /* TODO: make this variable */ #define FTRACE_GRAPH_MAX_FUNCS 32 +extern int ftrace_graph_filter_enabled; extern int ftrace_graph_count; extern unsigned long ftrace_graph_funcs[FTRACE_GRAPH_MAX_FUNCS]; @@ -504,7 +505,7 @@ static inline int ftrace_graph_addr(unsigned long addr) { int i; - if (!ftrace_graph_count || test_tsk_trace_graph(current)) + if (!ftrace_graph_filter_enabled) return 1; for (i = 0; i < ftrace_graph_count; i++) { @@ -791,7 +792,8 @@ extern const char *__stop___trace_bprintk_fmt[]; #undef FTRACE_ENTRY #define FTRACE_ENTRY(call, struct_name, id, tstruct, print) \ - extern struct ftrace_event_call event_##call; + extern struct ftrace_event_call \ + __attribute__((__aligned__(4))) event_##call; #undef FTRACE_ENTRY_DUP #define FTRACE_ENTRY_DUP(call, struct_name, id, tstruct, print) \ FTRACE_ENTRY(call, struct_name, id, PARAMS(tstruct), PARAMS(print)) diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index 4a194f08f88c..b9bc4d470177 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -307,8 +307,23 @@ static int annotated_branch_stat_cmp(void *p1, void *p2) return -1; if (percent_a > percent_b) return 1; - else - return 0; + + if (a->incorrect < b->incorrect) + return -1; + if (a->incorrect > b->incorrect) + return 1; + + /* + * Since the above shows worse (incorrect) cases + * first, we continue that by showing best (correct) + * cases last. + */ + if (a->correct > b->correct) + return -1; + if (a->correct < b->correct) + return 1; + + return 0; } static struct tracer_stat annotated_branch_stats = { diff --git a/kernel/trace/trace_event_profile.c b/kernel/trace/trace_event_profile.c index 9e25573242cf..f0d693005075 100644 --- a/kernel/trace/trace_event_profile.c +++ b/kernel/trace/trace_event_profile.c @@ -6,14 +6,12 @@ */ #include <linux/module.h> +#include <linux/kprobes.h> #include "trace.h" -char *perf_trace_buf; -EXPORT_SYMBOL_GPL(perf_trace_buf); - -char *perf_trace_buf_nmi; -EXPORT_SYMBOL_GPL(perf_trace_buf_nmi); +static char *perf_trace_buf; +static char *perf_trace_buf_nmi; typedef typeof(char [FTRACE_MAX_PROFILE_SIZE]) perf_trace_t ; @@ -120,3 +118,47 @@ void ftrace_profile_disable(int event_id) } mutex_unlock(&event_mutex); } + +__kprobes void *ftrace_perf_buf_prepare(int size, unsigned short type, + int *rctxp, unsigned long *irq_flags) +{ + struct trace_entry *entry; + char *trace_buf, *raw_data; + int pc, cpu; + + pc = preempt_count(); + + /* Protect the per cpu buffer, begin the rcu read side */ + local_irq_save(*irq_flags); + + *rctxp = perf_swevent_get_recursion_context(); + if (*rctxp < 0) + goto err_recursion; + + cpu = smp_processor_id(); + + if (in_nmi()) + trace_buf = rcu_dereference(perf_trace_buf_nmi); + else + trace_buf = rcu_dereference(perf_trace_buf); + + if (!trace_buf) + goto err; + + raw_data = per_cpu_ptr(trace_buf, cpu); + + /* zero the dead bytes from align to not leak stack to user */ + *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; + + entry = (struct trace_entry *)raw_data; + tracing_generic_entry_update(entry, *irq_flags, pc); + entry->type = type; + + return raw_data; +err: + perf_swevent_put_recursion_context(*rctxp); +err_recursion: + local_irq_restore(*irq_flags); + return NULL; +} +EXPORT_SYMBOL_GPL(ftrace_perf_buf_prepare); diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 189b09baf4fb..3f972ad98d04 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -60,10 +60,8 @@ int trace_define_field(struct ftrace_event_call *call, const char *type, return 0; err: - if (field) { + if (field) kfree(field->name); - kfree(field->type); - } kfree(field); return -ENOMEM; @@ -520,41 +518,16 @@ out: return ret; } -extern char *__bad_type_size(void); - -#undef FIELD -#define FIELD(type, name) \ - sizeof(type) != sizeof(field.name) ? __bad_type_size() : \ - #type, "common_" #name, offsetof(typeof(field), name), \ - sizeof(field.name), is_signed_type(type) - -static int trace_write_header(struct trace_seq *s) -{ - struct trace_entry field; - - /* struct trace_entry */ - return trace_seq_printf(s, - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\tsigned:%u;\n" - "\n", - FIELD(unsigned short, type), - FIELD(unsigned char, flags), - FIELD(unsigned char, preempt_count), - FIELD(int, pid), - FIELD(int, lock_depth)); -} - static ssize_t event_format_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { struct ftrace_event_call *call = filp->private_data; + struct ftrace_event_field *field; struct trace_seq *s; + int common_field_count = 5; char *buf; - int r; + int r = 0; if (*ppos) return 0; @@ -565,14 +538,48 @@ event_format_read(struct file *filp, char __user *ubuf, size_t cnt, trace_seq_init(s); - /* If any of the first writes fail, so will the show_format. */ - trace_seq_printf(s, "name: %s\n", call->name); trace_seq_printf(s, "ID: %d\n", call->id); trace_seq_printf(s, "format:\n"); - trace_write_header(s); - r = call->show_format(call, s); + list_for_each_entry_reverse(field, &call->fields, link) { + /* + * Smartly shows the array type(except dynamic array). + * Normal: + * field:TYPE VAR + * If TYPE := TYPE[LEN], it is shown: + * field:TYPE VAR[LEN] + */ + const char *array_descriptor = strchr(field->type, '['); + + if (!strncmp(field->type, "__data_loc", 10)) + array_descriptor = NULL; + + if (!array_descriptor) { + r = trace_seq_printf(s, "\tfield:%s %s;\toffset:%u;" + "\tsize:%u;\tsigned:%d;\n", + field->type, field->name, field->offset, + field->size, !!field->is_signed); + } else { + r = trace_seq_printf(s, "\tfield:%.*s %s%s;\toffset:%u;" + "\tsize:%u;\tsigned:%d;\n", + (int)(array_descriptor - field->type), + field->type, field->name, + array_descriptor, field->offset, + field->size, !!field->is_signed); + } + + if (--common_field_count == 0) + r = trace_seq_printf(s, "\n"); + + if (!r) + break; + } + + if (r) + r = trace_seq_printf(s, "\nprint fmt: %s\n", + call->print_fmt); + if (!r) { /* * ug! The format output is bigger than a PAGE!! @@ -948,10 +955,6 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events, filter); } - /* A trace may not want to export its format */ - if (!call->show_format) - return 0; - trace_create_file("format", 0444, call->dir, call, format); diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c index e42af9aad69f..4615f62a04f1 100644 --- a/kernel/trace/trace_events_filter.c +++ b/kernel/trace/trace_events_filter.c @@ -1371,7 +1371,7 @@ out_unlock: return err; } -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS void ftrace_profile_free_filter(struct perf_event *event) { @@ -1439,5 +1439,5 @@ out_unlock: return err; } -#endif /* CONFIG_EVENT_PROFILE */ +#endif /* CONFIG_PERF_EVENTS */ diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c index d4fa5dc1ee4e..e091f64ba6ce 100644 --- a/kernel/trace/trace_export.c +++ b/kernel/trace/trace_export.c @@ -62,78 +62,6 @@ static void __always_unused ____ftrace_check_##name(void) \ #include "trace_entries.h" - -#undef __field -#define __field(type, item) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ - "offset:%zu;\tsize:%zu;\tsigned:%u;\n", \ - offsetof(typeof(field), item), \ - sizeof(field.item), is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __field_desc -#define __field_desc(type, container, item) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ - "offset:%zu;\tsize:%zu;\tsigned:%u;\n", \ - offsetof(typeof(field), container.item), \ - sizeof(field.container.item), \ - is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __array -#define __array(type, item, len) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item "[" #len "];\t" \ - "offset:%zu;\tsize:%zu;\tsigned:%u;\n", \ - offsetof(typeof(field), item), \ - sizeof(field.item), is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __array_desc -#define __array_desc(type, container, item, len) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item "[" #len "];\t" \ - "offset:%zu;\tsize:%zu;\tsigned:%u;\n", \ - offsetof(typeof(field), container.item), \ - sizeof(field.container.item), \ - is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef __dynamic_array -#define __dynamic_array(type, item) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ - "offset:%zu;\tsize:0;\tsigned:%u;\n", \ - offsetof(typeof(field), item), \ - is_signed_type(type)); \ - if (!ret) \ - return 0; - -#undef F_printk -#define F_printk(fmt, args...) "%s, %s\n", #fmt, __stringify(args) - -#undef __entry -#define __entry REC - -#undef FTRACE_ENTRY -#define FTRACE_ENTRY(name, struct_name, id, tstruct, print) \ -static int \ -ftrace_format_##name(struct ftrace_event_call *unused, \ - struct trace_seq *s) \ -{ \ - struct struct_name field __attribute__((unused)); \ - int ret = 0; \ - \ - tstruct; \ - \ - trace_seq_printf(s, "\nprint fmt: " print); \ - \ - return ret; \ -} - -#include "trace_entries.h" - #undef __field #define __field(type, item) \ ret = trace_define_field(event_call, #type, #item, \ @@ -175,7 +103,12 @@ ftrace_format_##name(struct ftrace_event_call *unused, \ return ret; #undef __dynamic_array -#define __dynamic_array(type, item) +#define __dynamic_array(type, item) \ + ret = trace_define_field(event_call, #type, #item, \ + offsetof(typeof(field), item), \ + 0, is_signed_type(type), FILTER_OTHER);\ + if (ret) \ + return ret; #undef FTRACE_ENTRY #define FTRACE_ENTRY(name, struct_name, id, tstruct, print) \ @@ -198,6 +131,9 @@ static int ftrace_raw_init_event(struct ftrace_event_call *call) return 0; } +#undef __entry +#define __entry REC + #undef __field #define __field(type, item) @@ -213,6 +149,9 @@ static int ftrace_raw_init_event(struct ftrace_event_call *call) #undef __dynamic_array #define __dynamic_array(type, item) +#undef F_printk +#define F_printk(fmt, args...) #fmt ", " __stringify(args) + #undef FTRACE_ENTRY #define FTRACE_ENTRY(call, struct_name, type, tstruct, print) \ \ @@ -223,7 +162,7 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .id = type, \ .system = __stringify(TRACE_SYSTEM), \ .raw_init = ftrace_raw_init_event, \ - .show_format = ftrace_format_##call, \ + .print_fmt = print, \ .define_fields = ftrace_define_fields_##call, \ }; \ diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index b1342c5d37cf..e998a824e9db 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -18,6 +18,7 @@ struct fgraph_cpu_data { pid_t last_pid; int depth; int ignore; + unsigned long enter_funcs[FTRACE_RETFUNC_DEPTH]; }; struct fgraph_data { @@ -212,13 +213,11 @@ int trace_graph_entry(struct ftrace_graph_ent *trace) int cpu; int pc; - if (unlikely(!tr)) - return 0; - if (!ftrace_trace_task(current)) return 0; - if (!ftrace_graph_addr(trace->func)) + /* trace it when it is-nested-in or is a function enabled. */ + if (!(trace->depth || ftrace_graph_addr(trace->func))) return 0; local_irq_save(flags); @@ -231,9 +230,6 @@ int trace_graph_entry(struct ftrace_graph_ent *trace) } else { ret = 0; } - /* Only do the atomic if it is not already set */ - if (!test_tsk_trace_graph(current)) - set_tsk_trace_graph(current); atomic_dec(&data->disabled); local_irq_restore(flags); @@ -281,17 +277,24 @@ void trace_graph_return(struct ftrace_graph_ret *trace) pc = preempt_count(); __trace_graph_return(tr, trace, flags, pc); } - if (!trace->depth) - clear_tsk_trace_graph(current); atomic_dec(&data->disabled); local_irq_restore(flags); } +void set_graph_array(struct trace_array *tr) +{ + graph_array = tr; + + /* Make graph_array visible before we start tracing */ + + smp_mb(); +} + static int graph_trace_init(struct trace_array *tr) { int ret; - graph_array = tr; + set_graph_array(tr); ret = register_ftrace_graph(&trace_graph_return, &trace_graph_entry); if (ret) @@ -301,11 +304,6 @@ static int graph_trace_init(struct trace_array *tr) return 0; } -void set_graph_array(struct trace_array *tr) -{ - graph_array = tr; -} - static void graph_trace_reset(struct trace_array *tr) { tracing_stop_cmdline_record(); @@ -673,15 +671,21 @@ print_graph_entry_leaf(struct trace_iterator *iter, duration = graph_ret->rettime - graph_ret->calltime; if (data) { + struct fgraph_cpu_data *cpu_data; int cpu = iter->cpu; - int *depth = &(per_cpu_ptr(data->cpu_data, cpu)->depth); + + cpu_data = per_cpu_ptr(data->cpu_data, cpu); /* * Comments display at + 1 to depth. Since * this is a leaf function, keep the comments * equal to this depth. */ - *depth = call->depth - 1; + cpu_data->depth = call->depth - 1; + + /* No need to keep this function around for this depth */ + if (call->depth < FTRACE_RETFUNC_DEPTH) + cpu_data->enter_funcs[call->depth] = 0; } /* Overhead */ @@ -721,10 +725,15 @@ print_graph_entry_nested(struct trace_iterator *iter, int i; if (data) { + struct fgraph_cpu_data *cpu_data; int cpu = iter->cpu; - int *depth = &(per_cpu_ptr(data->cpu_data, cpu)->depth); - *depth = call->depth; + cpu_data = per_cpu_ptr(data->cpu_data, cpu); + cpu_data->depth = call->depth; + + /* Save this function pointer to see if the exit matches */ + if (call->depth < FTRACE_RETFUNC_DEPTH) + cpu_data->enter_funcs[call->depth] = call->func; } /* No overhead */ @@ -854,19 +863,28 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, struct fgraph_data *data = iter->private; pid_t pid = ent->pid; int cpu = iter->cpu; + int func_match = 1; int ret; int i; if (data) { + struct fgraph_cpu_data *cpu_data; int cpu = iter->cpu; - int *depth = &(per_cpu_ptr(data->cpu_data, cpu)->depth); + + cpu_data = per_cpu_ptr(data->cpu_data, cpu); /* * Comments display at + 1 to depth. This is the * return from a function, we now want the comments * to display at the same level of the bracket. */ - *depth = trace->depth - 1; + cpu_data->depth = trace->depth - 1; + + if (trace->depth < FTRACE_RETFUNC_DEPTH) { + if (cpu_data->enter_funcs[trace->depth] != trace->func) + func_match = 0; + cpu_data->enter_funcs[trace->depth] = 0; + } } if (print_graph_prologue(iter, s, 0, 0)) @@ -891,9 +909,21 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, return TRACE_TYPE_PARTIAL_LINE; } - ret = trace_seq_printf(s, "}\n"); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; + /* + * If the return function does not have a matching entry, + * then the entry was lost. Instead of just printing + * the '}' and letting the user guess what function this + * belongs to, write out the function name. + */ + if (func_match) { + ret = trace_seq_printf(s, "}\n"); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } else { + ret = trace_seq_printf(s, "} (%ps)\n", (void *)trace->func); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } /* Overrun */ if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERRUN) { diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 50b1b8239806..505c92273b1a 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -91,11 +91,6 @@ static __kprobes unsigned long fetch_memory(struct pt_regs *regs, void *addr) return retval; } -static __kprobes unsigned long fetch_argument(struct pt_regs *regs, void *num) -{ - return regs_get_argument_nth(regs, (unsigned int)((unsigned long)num)); -} - static __kprobes unsigned long fetch_retvalue(struct pt_regs *regs, void *dummy) { @@ -231,9 +226,7 @@ static int probe_arg_string(char *buf, size_t n, struct fetch_func *ff) { int ret = -EINVAL; - if (ff->func == fetch_argument) - ret = snprintf(buf, n, "$arg%lu", (unsigned long)ff->data); - else if (ff->func == fetch_register) { + if (ff->func == fetch_register) { const char *name; name = regs_query_register_name((unsigned int)((long)ff->data)); ret = snprintf(buf, n, "%%%s", name); @@ -489,14 +482,6 @@ static int parse_probe_vars(char *arg, struct fetch_func *ff, int is_return) } } else ret = -EINVAL; - } else if (strncmp(arg, "arg", 3) == 0 && isdigit(arg[3])) { - ret = strict_strtoul(arg + 3, 10, ¶m); - if (ret || param > PARAM_MAX_ARGS) - ret = -EINVAL; - else { - ff->func = fetch_argument; - ff->data = (void *)param; - } } else ret = -EINVAL; return ret; @@ -611,7 +596,6 @@ static int create_trace_probe(int argc, char **argv) * - Add kprobe: p[:[GRP/]EVENT] KSYM[+OFFS]|KADDR [FETCHARGS] * - Add kretprobe: r[:[GRP/]EVENT] KSYM[+0] [FETCHARGS] * Fetch args: - * $argN : fetch Nth of function argument. (N:0-) * $retval : fetch return value * $stack : fetch stack address * $stackN : fetch Nth of stack (N:0-) @@ -651,12 +635,12 @@ static int create_trace_probe(int argc, char **argv) event = strchr(group, '/') + 1; event[-1] = '\0'; if (strlen(group) == 0) { - pr_info("Group name is not specifiled\n"); + pr_info("Group name is not specified\n"); return -EINVAL; } } if (strlen(event) == 0) { - pr_info("Event name is not specifiled\n"); + pr_info("Event name is not specified\n"); return -EINVAL; } } @@ -958,7 +942,7 @@ static const struct file_operations kprobe_profile_ops = { }; /* Kprobe handler */ -static __kprobes int kprobe_trace_func(struct kprobe *kp, struct pt_regs *regs) +static __kprobes void kprobe_trace_func(struct kprobe *kp, struct pt_regs *regs) { struct trace_probe *tp = container_of(kp, struct trace_probe, rp.kp); struct kprobe_trace_entry *entry; @@ -978,7 +962,7 @@ static __kprobes int kprobe_trace_func(struct kprobe *kp, struct pt_regs *regs) event = trace_current_buffer_lock_reserve(&buffer, call->id, size, irq_flags, pc); if (!event) - return 0; + return; entry = ring_buffer_event_data(event); entry->nargs = tp->nr_args; @@ -988,11 +972,10 @@ static __kprobes int kprobe_trace_func(struct kprobe *kp, struct pt_regs *regs) if (!filter_current_check_discard(buffer, call, entry, event)) trace_nowake_buffer_unlock_commit(buffer, event, irq_flags, pc); - return 0; } /* Kretprobe handler */ -static __kprobes int kretprobe_trace_func(struct kretprobe_instance *ri, +static __kprobes void kretprobe_trace_func(struct kretprobe_instance *ri, struct pt_regs *regs) { struct trace_probe *tp = container_of(ri->rp, struct trace_probe, rp); @@ -1011,7 +994,7 @@ static __kprobes int kretprobe_trace_func(struct kretprobe_instance *ri, event = trace_current_buffer_lock_reserve(&buffer, call->id, size, irq_flags, pc); if (!event) - return 0; + return; entry = ring_buffer_event_data(event); entry->nargs = tp->nr_args; @@ -1022,8 +1005,6 @@ static __kprobes int kretprobe_trace_func(struct kretprobe_instance *ri, if (!filter_current_check_discard(buffer, call, entry, event)) trace_nowake_buffer_unlock_commit(buffer, event, irq_flags, pc); - - return 0; } /* Event entry printers */ @@ -1174,213 +1155,123 @@ static int kretprobe_event_define_fields(struct ftrace_event_call *event_call) return 0; } -static int __probe_event_show_format(struct trace_seq *s, - struct trace_probe *tp, const char *fmt, - const char *arg) +static int __set_print_fmt(struct trace_probe *tp, char *buf, int len) { int i; + int pos = 0; - /* Show format */ - if (!trace_seq_printf(s, "\nprint fmt: \"%s", fmt)) - return 0; + const char *fmt, *arg; - for (i = 0; i < tp->nr_args; i++) - if (!trace_seq_printf(s, " %s=%%lx", tp->args[i].name)) - return 0; + if (!probe_is_return(tp)) { + fmt = "(%lx)"; + arg = "REC->" FIELD_STRING_IP; + } else { + fmt = "(%lx <- %lx)"; + arg = "REC->" FIELD_STRING_FUNC ", REC->" FIELD_STRING_RETIP; + } - if (!trace_seq_printf(s, "\", %s", arg)) - return 0; + /* When len=0, we just calculate the needed length */ +#define LEN_OR_ZERO (len ? len - pos : 0) - for (i = 0; i < tp->nr_args; i++) - if (!trace_seq_printf(s, ", REC->%s", tp->args[i].name)) - return 0; - - return trace_seq_puts(s, "\n"); -} + pos += snprintf(buf + pos, LEN_OR_ZERO, "\"%s", fmt); -#undef SHOW_FIELD -#define SHOW_FIELD(type, item, name) \ - do { \ - ret = trace_seq_printf(s, "\tfield:" #type " %s;\t" \ - "offset:%u;\tsize:%u;\tsigned:%d;\n", name,\ - (unsigned int)offsetof(typeof(field), item),\ - (unsigned int)sizeof(type), \ - is_signed_type(type)); \ - if (!ret) \ - return 0; \ - } while (0) + for (i = 0; i < tp->nr_args; i++) { + pos += snprintf(buf + pos, LEN_OR_ZERO, " %s=%%lx", + tp->args[i].name); + } -static int kprobe_event_show_format(struct ftrace_event_call *call, - struct trace_seq *s) -{ - struct kprobe_trace_entry field __attribute__((unused)); - int ret, i; - struct trace_probe *tp = (struct trace_probe *)call->data; + pos += snprintf(buf + pos, LEN_OR_ZERO, "\", %s", arg); - SHOW_FIELD(unsigned long, ip, FIELD_STRING_IP); - SHOW_FIELD(int, nargs, FIELD_STRING_NARGS); + for (i = 0; i < tp->nr_args; i++) { + pos += snprintf(buf + pos, LEN_OR_ZERO, ", REC->%s", + tp->args[i].name); + } - /* Show fields */ - for (i = 0; i < tp->nr_args; i++) - SHOW_FIELD(unsigned long, args[i], tp->args[i].name); - trace_seq_puts(s, "\n"); +#undef LEN_OR_ZERO - return __probe_event_show_format(s, tp, "(%lx)", - "REC->" FIELD_STRING_IP); + /* return the length of print_fmt */ + return pos; } -static int kretprobe_event_show_format(struct ftrace_event_call *call, - struct trace_seq *s) +static int set_print_fmt(struct trace_probe *tp) { - struct kretprobe_trace_entry field __attribute__((unused)); - int ret, i; - struct trace_probe *tp = (struct trace_probe *)call->data; + int len; + char *print_fmt; - SHOW_FIELD(unsigned long, func, FIELD_STRING_FUNC); - SHOW_FIELD(unsigned long, ret_ip, FIELD_STRING_RETIP); - SHOW_FIELD(int, nargs, FIELD_STRING_NARGS); + /* First: called with 0 length to calculate the needed length */ + len = __set_print_fmt(tp, NULL, 0); + print_fmt = kmalloc(len + 1, GFP_KERNEL); + if (!print_fmt) + return -ENOMEM; - /* Show fields */ - for (i = 0; i < tp->nr_args; i++) - SHOW_FIELD(unsigned long, args[i], tp->args[i].name); - trace_seq_puts(s, "\n"); + /* Second: actually write the @print_fmt */ + __set_print_fmt(tp, print_fmt, len + 1); + tp->call.print_fmt = print_fmt; - return __probe_event_show_format(s, tp, "(%lx <- %lx)", - "REC->" FIELD_STRING_FUNC - ", REC->" FIELD_STRING_RETIP); + return 0; } -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS /* Kprobe profile handler */ -static __kprobes int kprobe_profile_func(struct kprobe *kp, +static __kprobes void kprobe_profile_func(struct kprobe *kp, struct pt_regs *regs) { struct trace_probe *tp = container_of(kp, struct trace_probe, rp.kp); struct ftrace_event_call *call = &tp->call; struct kprobe_trace_entry *entry; - struct trace_entry *ent; - int size, __size, i, pc, __cpu; + int size, __size, i; unsigned long irq_flags; - char *trace_buf; - char *raw_data; int rctx; - pc = preempt_count(); __size = SIZEOF_KPROBE_TRACE_ENTRY(tp->nr_args); size = ALIGN(__size + sizeof(u32), sizeof(u64)); size -= sizeof(u32); if (WARN_ONCE(size > FTRACE_MAX_PROFILE_SIZE, "profile buffer not large enough")) - return 0; - - /* - * Protect the non nmi buffer - * This also protects the rcu read side - */ - local_irq_save(irq_flags); - - rctx = perf_swevent_get_recursion_context(); - if (rctx < 0) - goto end_recursion; - - __cpu = smp_processor_id(); - - if (in_nmi()) - trace_buf = rcu_dereference(perf_trace_buf_nmi); - else - trace_buf = rcu_dereference(perf_trace_buf); + return; - if (!trace_buf) - goto end; - - raw_data = per_cpu_ptr(trace_buf, __cpu); - - /* Zero dead bytes from alignment to avoid buffer leak to userspace */ - *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; - entry = (struct kprobe_trace_entry *)raw_data; - ent = &entry->ent; + entry = ftrace_perf_buf_prepare(size, call->id, &rctx, &irq_flags); + if (!entry) + return; - tracing_generic_entry_update(ent, irq_flags, pc); - ent->type = call->id; entry->nargs = tp->nr_args; entry->ip = (unsigned long)kp->addr; for (i = 0; i < tp->nr_args; i++) entry->args[i] = call_fetch(&tp->args[i].fetch, regs); - perf_tp_event(call->id, entry->ip, 1, entry, size); - -end: - perf_swevent_put_recursion_context(rctx); -end_recursion: - local_irq_restore(irq_flags); - return 0; + ftrace_perf_buf_submit(entry, size, rctx, entry->ip, 1, irq_flags); } /* Kretprobe profile handler */ -static __kprobes int kretprobe_profile_func(struct kretprobe_instance *ri, +static __kprobes void kretprobe_profile_func(struct kretprobe_instance *ri, struct pt_regs *regs) { struct trace_probe *tp = container_of(ri->rp, struct trace_probe, rp); struct ftrace_event_call *call = &tp->call; struct kretprobe_trace_entry *entry; - struct trace_entry *ent; - int size, __size, i, pc, __cpu; + int size, __size, i; unsigned long irq_flags; - char *trace_buf; - char *raw_data; int rctx; - pc = preempt_count(); __size = SIZEOF_KRETPROBE_TRACE_ENTRY(tp->nr_args); size = ALIGN(__size + sizeof(u32), sizeof(u64)); size -= sizeof(u32); if (WARN_ONCE(size > FTRACE_MAX_PROFILE_SIZE, "profile buffer not large enough")) - return 0; - - /* - * Protect the non nmi buffer - * This also protects the rcu read side - */ - local_irq_save(irq_flags); - - rctx = perf_swevent_get_recursion_context(); - if (rctx < 0) - goto end_recursion; - - __cpu = smp_processor_id(); + return; - if (in_nmi()) - trace_buf = rcu_dereference(perf_trace_buf_nmi); - else - trace_buf = rcu_dereference(perf_trace_buf); - - if (!trace_buf) - goto end; - - raw_data = per_cpu_ptr(trace_buf, __cpu); - - /* Zero dead bytes from alignment to avoid buffer leak to userspace */ - *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; - entry = (struct kretprobe_trace_entry *)raw_data; - ent = &entry->ent; + entry = ftrace_perf_buf_prepare(size, call->id, &rctx, &irq_flags); + if (!entry) + return; - tracing_generic_entry_update(ent, irq_flags, pc); - ent->type = call->id; entry->nargs = tp->nr_args; entry->func = (unsigned long)tp->rp.kp.addr; entry->ret_ip = (unsigned long)ri->ret_addr; for (i = 0; i < tp->nr_args; i++) entry->args[i] = call_fetch(&tp->args[i].fetch, regs); - perf_tp_event(call->id, entry->ret_ip, 1, entry, size); - -end: - perf_swevent_put_recursion_context(rctx); -end_recursion: - local_irq_restore(irq_flags); - return 0; + ftrace_perf_buf_submit(entry, size, rctx, entry->ret_ip, 1, irq_flags); } static int probe_profile_enable(struct ftrace_event_call *call) @@ -1408,7 +1299,7 @@ static void probe_profile_disable(struct ftrace_event_call *call) disable_kprobe(&tp->rp.kp); } } -#endif /* CONFIG_EVENT_PROFILE */ +#endif /* CONFIG_PERF_EVENTS */ static __kprobes @@ -1418,10 +1309,10 @@ int kprobe_dispatcher(struct kprobe *kp, struct pt_regs *regs) if (tp->flags & TP_FLAG_TRACE) kprobe_trace_func(kp, regs); -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS if (tp->flags & TP_FLAG_PROFILE) kprobe_profile_func(kp, regs); -#endif /* CONFIG_EVENT_PROFILE */ +#endif return 0; /* We don't tweek kernel, so just return 0 */ } @@ -1432,10 +1323,10 @@ int kretprobe_dispatcher(struct kretprobe_instance *ri, struct pt_regs *regs) if (tp->flags & TP_FLAG_TRACE) kretprobe_trace_func(ri, regs); -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS if (tp->flags & TP_FLAG_PROFILE) kretprobe_profile_func(ri, regs); -#endif /* CONFIG_EVENT_PROFILE */ +#endif return 0; /* We don't tweek kernel, so just return 0 */ } @@ -1448,23 +1339,25 @@ static int register_probe_event(struct trace_probe *tp) if (probe_is_return(tp)) { tp->event.trace = print_kretprobe_event; call->raw_init = probe_event_raw_init; - call->show_format = kretprobe_event_show_format; call->define_fields = kretprobe_event_define_fields; } else { tp->event.trace = print_kprobe_event; call->raw_init = probe_event_raw_init; - call->show_format = kprobe_event_show_format; call->define_fields = kprobe_event_define_fields; } + if (set_print_fmt(tp) < 0) + return -ENOMEM; call->event = &tp->event; call->id = register_ftrace_event(&tp->event); - if (!call->id) + if (!call->id) { + kfree(call->print_fmt); return -ENODEV; + } call->enabled = 0; call->regfunc = probe_event_enable; call->unregfunc = probe_event_disable; -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS call->profile_enable = probe_profile_enable; call->profile_disable = probe_profile_disable; #endif @@ -1472,6 +1365,7 @@ static int register_probe_event(struct trace_probe *tp) ret = trace_add_event_call(call); if (ret) { pr_info("Failed to register kprobe event: %s\n", call->name); + kfree(call->print_fmt); unregister_ftrace_event(&tp->event); } return ret; @@ -1481,6 +1375,7 @@ static void unregister_probe_event(struct trace_probe *tp) { /* tp->event is unregistered in trace_remove_event_call() */ trace_remove_event_call(&tp->call); + kfree(tp->call.print_fmt); } /* Make a debugfs interface for controling probe points */ @@ -1523,28 +1418,67 @@ static int kprobe_trace_selftest_target(int a1, int a2, int a3, static __init int kprobe_trace_self_tests_init(void) { - int ret; + int ret, warn = 0; int (*target)(int, int, int, int, int, int); + struct trace_probe *tp; target = kprobe_trace_selftest_target; pr_info("Testing kprobe tracing: "); ret = command_trace_probe("p:testprobe kprobe_trace_selftest_target " - "$arg1 $arg2 $arg3 $arg4 $stack $stack0"); - if (WARN_ON_ONCE(ret)) - pr_warning("error enabling function entry\n"); + "$stack $stack0 +0($stack)"); + if (WARN_ON_ONCE(ret)) { + pr_warning("error on probing function entry.\n"); + warn++; + } else { + /* Enable trace point */ + tp = find_probe_event("testprobe", KPROBE_EVENT_SYSTEM); + if (WARN_ON_ONCE(tp == NULL)) { + pr_warning("error on getting new probe.\n"); + warn++; + } else + probe_event_enable(&tp->call); + } ret = command_trace_probe("r:testprobe2 kprobe_trace_selftest_target " "$retval"); - if (WARN_ON_ONCE(ret)) - pr_warning("error enabling function return\n"); + if (WARN_ON_ONCE(ret)) { + pr_warning("error on probing function return.\n"); + warn++; + } else { + /* Enable trace point */ + tp = find_probe_event("testprobe2", KPROBE_EVENT_SYSTEM); + if (WARN_ON_ONCE(tp == NULL)) { + pr_warning("error on getting new probe.\n"); + warn++; + } else + probe_event_enable(&tp->call); + } + + if (warn) + goto end; ret = target(1, 2, 3, 4, 5, 6); - cleanup_all_probes(); + ret = command_trace_probe("-:testprobe"); + if (WARN_ON_ONCE(ret)) { + pr_warning("error on deleting a probe.\n"); + warn++; + } - pr_cont("OK\n"); + ret = command_trace_probe("-:testprobe2"); + if (WARN_ON_ONCE(ret)) { + pr_warning("error on deleting a probe.\n"); + warn++; + } + +end: + cleanup_all_probes(); + if (warn) + pr_cont("NG: Some tests are failed. Please check them.\n"); + else + pr_cont("OK\n"); return 0; } diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 75289f372dd2..cba47d7935cc 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -143,70 +143,65 @@ extern char *__bad_type_size(void); #type, #name, offsetof(typeof(trace), name), \ sizeof(trace.name), is_signed_type(type) -int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s) +static +int __set_enter_print_fmt(struct syscall_metadata *entry, char *buf, int len) { int i; - int ret; - struct syscall_metadata *entry = call->data; - struct syscall_trace_enter trace; - int offset = offsetof(struct syscall_trace_enter, args); + int pos = 0; - ret = trace_seq_printf(s, "\tfield:%s %s;\toffset:%zu;\tsize:%zu;" - "\tsigned:%u;\n", - SYSCALL_FIELD(int, nr)); - if (!ret) - return 0; + /* When len=0, we just calculate the needed length */ +#define LEN_OR_ZERO (len ? len - pos : 0) + pos += snprintf(buf + pos, LEN_OR_ZERO, "\""); for (i = 0; i < entry->nb_args; i++) { - ret = trace_seq_printf(s, "\tfield:%s %s;", entry->types[i], - entry->args[i]); - if (!ret) - return 0; - ret = trace_seq_printf(s, "\toffset:%d;\tsize:%zu;" - "\tsigned:%u;\n", offset, - sizeof(unsigned long), - is_signed_type(unsigned long)); - if (!ret) - return 0; - offset += sizeof(unsigned long); + pos += snprintf(buf + pos, LEN_OR_ZERO, "%s: 0x%%0%zulx%s", + entry->args[i], sizeof(unsigned long), + i == entry->nb_args - 1 ? "" : ", "); } + pos += snprintf(buf + pos, LEN_OR_ZERO, "\""); - trace_seq_puts(s, "\nprint fmt: \""); for (i = 0; i < entry->nb_args; i++) { - ret = trace_seq_printf(s, "%s: 0x%%0%zulx%s", entry->args[i], - sizeof(unsigned long), - i == entry->nb_args - 1 ? "" : ", "); - if (!ret) - return 0; + pos += snprintf(buf + pos, LEN_OR_ZERO, + ", ((unsigned long)(REC->%s))", entry->args[i]); } - trace_seq_putc(s, '"'); - for (i = 0; i < entry->nb_args; i++) { - ret = trace_seq_printf(s, ", ((unsigned long)(REC->%s))", - entry->args[i]); - if (!ret) - return 0; - } +#undef LEN_OR_ZERO - return trace_seq_putc(s, '\n'); + /* return the length of print_fmt */ + return pos; } -int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s) +static int set_syscall_print_fmt(struct ftrace_event_call *call) { - int ret; - struct syscall_trace_exit trace; + char *print_fmt; + int len; + struct syscall_metadata *entry = call->data; - ret = trace_seq_printf(s, - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;" - "\tsigned:%u;\n" - "\tfield:%s %s;\toffset:%zu;\tsize:%zu;" - "\tsigned:%u;\n", - SYSCALL_FIELD(int, nr), - SYSCALL_FIELD(long, ret)); - if (!ret) + if (entry->enter_event != call) { + call->print_fmt = "\"0x%lx\", REC->ret"; return 0; + } + + /* First: called with 0 length to calculate the needed length */ + len = __set_enter_print_fmt(entry, NULL, 0); + + print_fmt = kmalloc(len + 1, GFP_KERNEL); + if (!print_fmt) + return -ENOMEM; + + /* Second: actually write the @print_fmt */ + __set_enter_print_fmt(entry, print_fmt, len + 1); + call->print_fmt = print_fmt; - return trace_seq_printf(s, "\nprint fmt: \"0x%%lx\", REC->ret\n"); + return 0; +} + +static void free_syscall_print_fmt(struct ftrace_event_call *call) +{ + struct syscall_metadata *entry = call->data; + + if (entry->enter_event == call) + kfree(call->print_fmt); } int syscall_enter_define_fields(struct ftrace_event_call *call) @@ -386,12 +381,22 @@ int init_syscall_trace(struct ftrace_event_call *call) { int id; - id = register_ftrace_event(call->event); - if (!id) - return -ENODEV; - call->id = id; - INIT_LIST_HEAD(&call->fields); - return 0; + if (set_syscall_print_fmt(call) < 0) + return -ENOMEM; + + id = trace_event_raw_init(call); + + if (id < 0) { + free_syscall_print_fmt(call); + return id; + } + + return id; +} + +unsigned long __init arch_syscall_addr(int nr) +{ + return (unsigned long)sys_call_table[nr]; } int __init init_ftrace_syscalls(void) @@ -421,7 +426,7 @@ int __init init_ftrace_syscalls(void) } core_initcall(init_ftrace_syscalls); -#ifdef CONFIG_EVENT_PROFILE +#ifdef CONFIG_PERF_EVENTS static DECLARE_BITMAP(enabled_prof_enter_syscalls, NR_syscalls); static DECLARE_BITMAP(enabled_prof_exit_syscalls, NR_syscalls); @@ -433,12 +438,9 @@ static void prof_syscall_enter(struct pt_regs *regs, long id) struct syscall_metadata *sys_data; struct syscall_trace_enter *rec; unsigned long flags; - char *trace_buf; - char *raw_data; int syscall_nr; int rctx; int size; - int cpu; syscall_nr = syscall_get_nr(current, regs); if (!test_bit(syscall_nr, enabled_prof_enter_syscalls)) @@ -457,37 +459,15 @@ static void prof_syscall_enter(struct pt_regs *regs, long id) "profile buffer not large enough")) return; - /* Protect the per cpu buffer, begin the rcu read side */ - local_irq_save(flags); - - rctx = perf_swevent_get_recursion_context(); - if (rctx < 0) - goto end_recursion; - - cpu = smp_processor_id(); - - trace_buf = rcu_dereference(perf_trace_buf); - - if (!trace_buf) - goto end; - - raw_data = per_cpu_ptr(trace_buf, cpu); - - /* zero the dead bytes from align to not leak stack to user */ - *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; + rec = (struct syscall_trace_enter *)ftrace_perf_buf_prepare(size, + sys_data->enter_event->id, &rctx, &flags); + if (!rec) + return; - rec = (struct syscall_trace_enter *) raw_data; - tracing_generic_entry_update(&rec->ent, 0, 0); - rec->ent.type = sys_data->enter_event->id; rec->nr = syscall_nr; syscall_get_arguments(current, regs, 0, sys_data->nb_args, (unsigned long *)&rec->args); - perf_tp_event(sys_data->enter_event->id, 0, 1, rec, size); - -end: - perf_swevent_put_recursion_context(rctx); -end_recursion: - local_irq_restore(flags); + ftrace_perf_buf_submit(rec, size, rctx, 0, 1, flags); } int prof_sysenter_enable(struct ftrace_event_call *call) @@ -531,11 +511,8 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret) struct syscall_trace_exit *rec; unsigned long flags; int syscall_nr; - char *trace_buf; - char *raw_data; int rctx; int size; - int cpu; syscall_nr = syscall_get_nr(current, regs); if (!test_bit(syscall_nr, enabled_prof_exit_syscalls)) @@ -557,38 +534,15 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret) "exit event has grown above profile buffer size")) return; - /* Protect the per cpu buffer, begin the rcu read side */ - local_irq_save(flags); - - rctx = perf_swevent_get_recursion_context(); - if (rctx < 0) - goto end_recursion; - - cpu = smp_processor_id(); - - trace_buf = rcu_dereference(perf_trace_buf); - - if (!trace_buf) - goto end; - - raw_data = per_cpu_ptr(trace_buf, cpu); - - /* zero the dead bytes from align to not leak stack to user */ - *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; - - rec = (struct syscall_trace_exit *)raw_data; + rec = (struct syscall_trace_exit *)ftrace_perf_buf_prepare(size, + sys_data->exit_event->id, &rctx, &flags); + if (!rec) + return; - tracing_generic_entry_update(&rec->ent, 0, 0); - rec->ent.type = sys_data->exit_event->id; rec->nr = syscall_nr; rec->ret = syscall_get_return_value(current, regs); - perf_tp_event(sys_data->exit_event->id, 0, 1, rec, size); - -end: - perf_swevent_put_recursion_context(rctx); -end_recursion: - local_irq_restore(flags); + ftrace_perf_buf_submit(rec, size, rctx, 0, 1, flags); } int prof_sysexit_enable(struct ftrace_event_call *call) @@ -603,7 +557,7 @@ int prof_sysexit_enable(struct ftrace_event_call *call) ret = register_trace_sys_exit(prof_syscall_exit); if (ret) { pr_info("event trace: Could not activate" - "syscall entry trace point"); + "syscall exit trace point"); } else { set_bit(num, enabled_prof_exit_syscalls); sys_prof_refcount_exit++; @@ -626,6 +580,5 @@ void prof_sysexit_disable(struct ftrace_event_call *call) mutex_unlock(&syscall_trace_lock); } -#endif - +#endif /* CONFIG_PERF_EVENTS */ diff --git a/kernel/user.c b/kernel/user.c index 46d0165ca70c..766467b3bcb7 100644 --- a/kernel/user.c +++ b/kernel/user.c @@ -56,9 +56,6 @@ struct user_struct root_user = { .sigpending = ATOMIC_INIT(0), .locked_shm = 0, .user_ns = &init_user_ns, -#ifdef CONFIG_USER_SCHED - .tg = &init_task_group, -#endif }; /* @@ -75,268 +72,6 @@ static void uid_hash_remove(struct user_struct *up) put_user_ns(up->user_ns); } -#ifdef CONFIG_USER_SCHED - -static void sched_destroy_user(struct user_struct *up) -{ - sched_destroy_group(up->tg); -} - -static int sched_create_user(struct user_struct *up) -{ - int rc = 0; - - up->tg = sched_create_group(&root_task_group); - if (IS_ERR(up->tg)) - rc = -ENOMEM; - - set_tg_uid(up); - - return rc; -} - -#else /* CONFIG_USER_SCHED */ - -static void sched_destroy_user(struct user_struct *up) { } -static int sched_create_user(struct user_struct *up) { return 0; } - -#endif /* CONFIG_USER_SCHED */ - -#if defined(CONFIG_USER_SCHED) && defined(CONFIG_SYSFS) - -static struct user_struct *uid_hash_find(uid_t uid, struct hlist_head *hashent) -{ - struct user_struct *user; - struct hlist_node *h; - - hlist_for_each_entry(user, h, hashent, uidhash_node) { - if (user->uid == uid) { - /* possibly resurrect an "almost deleted" object */ - if (atomic_inc_return(&user->__count) == 1) - cancel_delayed_work(&user->work); - return user; - } - } - - return NULL; -} - -static struct kset *uids_kset; /* represents the /sys/kernel/uids/ directory */ -static DEFINE_MUTEX(uids_mutex); - -static inline void uids_mutex_lock(void) -{ - mutex_lock(&uids_mutex); -} - -static inline void uids_mutex_unlock(void) -{ - mutex_unlock(&uids_mutex); -} - -/* uid directory attributes */ -#ifdef CONFIG_FAIR_GROUP_SCHED -static ssize_t cpu_shares_show(struct kobject *kobj, - struct kobj_attribute *attr, - char *buf) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - - return sprintf(buf, "%lu\n", sched_group_shares(up->tg)); -} - -static ssize_t cpu_shares_store(struct kobject *kobj, - struct kobj_attribute *attr, - const char *buf, size_t size) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - unsigned long shares; - int rc; - - sscanf(buf, "%lu", &shares); - - rc = sched_group_set_shares(up->tg, shares); - - return (rc ? rc : size); -} - -static struct kobj_attribute cpu_share_attr = - __ATTR(cpu_share, 0644, cpu_shares_show, cpu_shares_store); -#endif - -#ifdef CONFIG_RT_GROUP_SCHED -static ssize_t cpu_rt_runtime_show(struct kobject *kobj, - struct kobj_attribute *attr, - char *buf) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - - return sprintf(buf, "%ld\n", sched_group_rt_runtime(up->tg)); -} - -static ssize_t cpu_rt_runtime_store(struct kobject *kobj, - struct kobj_attribute *attr, - const char *buf, size_t size) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - unsigned long rt_runtime; - int rc; - - sscanf(buf, "%ld", &rt_runtime); - - rc = sched_group_set_rt_runtime(up->tg, rt_runtime); - - return (rc ? rc : size); -} - -static struct kobj_attribute cpu_rt_runtime_attr = - __ATTR(cpu_rt_runtime, 0644, cpu_rt_runtime_show, cpu_rt_runtime_store); - -static ssize_t cpu_rt_period_show(struct kobject *kobj, - struct kobj_attribute *attr, - char *buf) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - - return sprintf(buf, "%lu\n", sched_group_rt_period(up->tg)); -} - -static ssize_t cpu_rt_period_store(struct kobject *kobj, - struct kobj_attribute *attr, - const char *buf, size_t size) -{ - struct user_struct *up = container_of(kobj, struct user_struct, kobj); - unsigned long rt_period; - int rc; - - sscanf(buf, "%lu", &rt_period); - - rc = sched_group_set_rt_period(up->tg, rt_period); - - return (rc ? rc : size); -} - -static struct kobj_attribute cpu_rt_period_attr = - __ATTR(cpu_rt_period, 0644, cpu_rt_period_show, cpu_rt_period_store); -#endif - -/* default attributes per uid directory */ -static struct attribute *uids_attributes[] = { -#ifdef CONFIG_FAIR_GROUP_SCHED - &cpu_share_attr.attr, -#endif -#ifdef CONFIG_RT_GROUP_SCHED - &cpu_rt_runtime_attr.attr, - &cpu_rt_period_attr.attr, -#endif - NULL -}; - -/* the lifetime of user_struct is not managed by the core (now) */ -static void uids_release(struct kobject *kobj) -{ - return; -} - -static struct kobj_type uids_ktype = { - .sysfs_ops = &kobj_sysfs_ops, - .default_attrs = uids_attributes, - .release = uids_release, -}; - -/* - * Create /sys/kernel/uids/<uid>/cpu_share file for this user - * We do not create this file for users in a user namespace (until - * sysfs tagging is implemented). - * - * See Documentation/scheduler/sched-design-CFS.txt for ramifications. - */ -static int uids_user_create(struct user_struct *up) -{ - struct kobject *kobj = &up->kobj; - int error; - - memset(kobj, 0, sizeof(struct kobject)); - if (up->user_ns != &init_user_ns) - return 0; - kobj->kset = uids_kset; - error = kobject_init_and_add(kobj, &uids_ktype, NULL, "%d", up->uid); - if (error) { - kobject_put(kobj); - goto done; - } - - kobject_uevent(kobj, KOBJ_ADD); -done: - return error; -} - -/* create these entries in sysfs: - * "/sys/kernel/uids" directory - * "/sys/kernel/uids/0" directory (for root user) - * "/sys/kernel/uids/0/cpu_share" file (for root user) - */ -int __init uids_sysfs_init(void) -{ - uids_kset = kset_create_and_add("uids", NULL, kernel_kobj); - if (!uids_kset) - return -ENOMEM; - - return uids_user_create(&root_user); -} - -/* delayed work function to remove sysfs directory for a user and free up - * corresponding structures. - */ -static void cleanup_user_struct(struct work_struct *w) -{ - struct user_struct *up = container_of(w, struct user_struct, work.work); - unsigned long flags; - int remove_user = 0; - - /* Make uid_hash_remove() + sysfs_remove_file() + kobject_del() - * atomic. - */ - uids_mutex_lock(); - - spin_lock_irqsave(&uidhash_lock, flags); - if (atomic_read(&up->__count) == 0) { - uid_hash_remove(up); - remove_user = 1; - } - spin_unlock_irqrestore(&uidhash_lock, flags); - - if (!remove_user) - goto done; - - if (up->user_ns == &init_user_ns) { - kobject_uevent(&up->kobj, KOBJ_REMOVE); - kobject_del(&up->kobj); - kobject_put(&up->kobj); - } - - sched_destroy_user(up); - key_put(up->uid_keyring); - key_put(up->session_keyring); - kmem_cache_free(uid_cachep, up); - -done: - uids_mutex_unlock(); -} - -/* IRQs are disabled and uidhash_lock is held upon function entry. - * IRQ state (as stored in flags) is restored and uidhash_lock released - * upon function exit. - */ -static void free_user(struct user_struct *up, unsigned long flags) -{ - INIT_DELAYED_WORK(&up->work, cleanup_user_struct); - schedule_delayed_work(&up->work, msecs_to_jiffies(1000)); - spin_unlock_irqrestore(&uidhash_lock, flags); -} - -#else /* CONFIG_USER_SCHED && CONFIG_SYSFS */ - static struct user_struct *uid_hash_find(uid_t uid, struct hlist_head *hashent) { struct user_struct *user; @@ -352,11 +87,6 @@ static struct user_struct *uid_hash_find(uid_t uid, struct hlist_head *hashent) return NULL; } -int uids_sysfs_init(void) { return 0; } -static inline int uids_user_create(struct user_struct *up) { return 0; } -static inline void uids_mutex_lock(void) { } -static inline void uids_mutex_unlock(void) { } - /* IRQs are disabled and uidhash_lock is held upon function entry. * IRQ state (as stored in flags) is restored and uidhash_lock released * upon function exit. @@ -365,32 +95,11 @@ static void free_user(struct user_struct *up, unsigned long flags) { uid_hash_remove(up); spin_unlock_irqrestore(&uidhash_lock, flags); - sched_destroy_user(up); key_put(up->uid_keyring); key_put(up->session_keyring); kmem_cache_free(uid_cachep, up); } -#endif - -#if defined(CONFIG_RT_GROUP_SCHED) && defined(CONFIG_USER_SCHED) -/* - * We need to check if a setuid can take place. This function should be called - * before successfully completing the setuid. - */ -int task_can_switch_user(struct user_struct *up, struct task_struct *tsk) -{ - - return sched_rt_can_attach(up->tg, tsk); - -} -#else -int task_can_switch_user(struct user_struct *up, struct task_struct *tsk) -{ - return 1; -} -#endif - /* * Locate the user_struct for the passed UID. If found, take a ref on it. The * caller must undo that ref with free_uid(). @@ -431,8 +140,6 @@ struct user_struct *alloc_uid(struct user_namespace *ns, uid_t uid) /* Make uid_hash_find() + uids_user_create() + uid_hash_insert() * atomic. */ - uids_mutex_lock(); - spin_lock_irq(&uidhash_lock); up = uid_hash_find(uid, hashent); spin_unlock_irq(&uidhash_lock); @@ -445,14 +152,8 @@ struct user_struct *alloc_uid(struct user_namespace *ns, uid_t uid) new->uid = uid; atomic_set(&new->__count, 1); - if (sched_create_user(new) < 0) - goto out_free_user; - new->user_ns = get_user_ns(ns); - if (uids_user_create(new)) - goto out_destoy_sched; - /* * Before adding this, check whether we raced * on adding the same user already.. @@ -475,17 +176,11 @@ struct user_struct *alloc_uid(struct user_namespace *ns, uid_t uid) spin_unlock_irq(&uidhash_lock); } - uids_mutex_unlock(); - return up; -out_destoy_sched: - sched_destroy_user(new); put_user_ns(new->user_ns); -out_free_user: kmem_cache_free(uid_cachep, new); out_unlock: - uids_mutex_unlock(); return NULL; } |