diff options
Diffstat (limited to 'kernel')
| -rw-r--r-- | kernel/exit.c | 4 | ||||
| -rw-r--r-- | kernel/futex.c | 51 | ||||
| -rw-r--r-- | kernel/irq/chip.c | 3 | ||||
| -rw-r--r-- | kernel/irq/handle.c | 14 | ||||
| -rw-r--r-- | kernel/irq/internals.h | 2 | ||||
| -rw-r--r-- | kernel/irq/manage.c | 86 | ||||
| -rw-r--r-- | kernel/lockdep.c | 8 | ||||
| -rw-r--r-- | kernel/rcu.h | 26 | ||||
| -rw-r--r-- | kernel/rcupdate.c | 5 | ||||
| -rw-r--r-- | kernel/rcutiny.c | 26 | ||||
| -rw-r--r-- | kernel/rcutiny_plugin.h | 77 | ||||
| -rw-r--r-- | kernel/rcutorture.c | 91 | ||||
| -rw-r--r-- | kernel/rcutree.c | 507 | ||||
| -rw-r--r-- | kernel/rcutree.h | 27 | ||||
| -rw-r--r-- | kernel/rcutree_plugin.h | 450 | ||||
| -rw-r--r-- | kernel/rcutree_trace.c | 12 | ||||
| -rw-r--r-- | kernel/softirq.c | 18 | ||||
| -rw-r--r-- | kernel/srcu.c | 33 | ||||
| -rw-r--r-- | kernel/sys.c | 2 | ||||
| -rw-r--r-- | kernel/workqueue.c | 7 |
20 files changed, 981 insertions, 468 deletions
diff --git a/kernel/exit.c b/kernel/exit.c index 4b4042f9bc6a..752d2c0abd19 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -935,8 +935,6 @@ void do_exit(long code) schedule(); } - exit_irq_thread(); - exit_signals(tsk); /* sets PF_EXITING */ /* * tsk->flags are checked in the futex code to protect against @@ -945,6 +943,8 @@ void do_exit(long code) smp_mb(); raw_spin_unlock_wait(&tsk->pi_lock); + exit_irq_thread(); + if (unlikely(in_atomic())) printk(KERN_INFO "note: %s[%d] exited with preempt_count %d\n", current->comm, task_pid_nr(current), diff --git a/kernel/futex.c b/kernel/futex.c index 1614be20173d..72efa1e4359a 100644 --- a/kernel/futex.c +++ b/kernel/futex.c @@ -2628,7 +2628,7 @@ void exit_robust_list(struct task_struct *curr) long do_futex(u32 __user *uaddr, int op, u32 val, ktime_t *timeout, u32 __user *uaddr2, u32 val2, u32 val3) { - int ret = -ENOSYS, cmd = op & FUTEX_CMD_MASK; + int cmd = op & FUTEX_CMD_MASK; unsigned int flags = 0; if (!(op & FUTEX_PRIVATE_FLAG)) @@ -2641,49 +2641,44 @@ long do_futex(u32 __user *uaddr, int op, u32 val, ktime_t *timeout, } switch (cmd) { + case FUTEX_LOCK_PI: + case FUTEX_UNLOCK_PI: + case FUTEX_TRYLOCK_PI: + case FUTEX_WAIT_REQUEUE_PI: + case FUTEX_CMP_REQUEUE_PI: + if (!futex_cmpxchg_enabled) + return -ENOSYS; + } + + switch (cmd) { case FUTEX_WAIT: val3 = FUTEX_BITSET_MATCH_ANY; case FUTEX_WAIT_BITSET: - ret = futex_wait(uaddr, flags, val, timeout, val3); - break; + return futex_wait(uaddr, flags, val, timeout, val3); case FUTEX_WAKE: val3 = FUTEX_BITSET_MATCH_ANY; case FUTEX_WAKE_BITSET: - ret = futex_wake(uaddr, flags, val, val3); - break; + return futex_wake(uaddr, flags, val, val3); case FUTEX_REQUEUE: - ret = futex_requeue(uaddr, flags, uaddr2, val, val2, NULL, 0); - break; + return futex_requeue(uaddr, flags, uaddr2, val, val2, NULL, 0); case FUTEX_CMP_REQUEUE: - ret = futex_requeue(uaddr, flags, uaddr2, val, val2, &val3, 0); - break; + return futex_requeue(uaddr, flags, uaddr2, val, val2, &val3, 0); case FUTEX_WAKE_OP: - ret = futex_wake_op(uaddr, flags, uaddr2, val, val2, val3); - break; + return futex_wake_op(uaddr, flags, uaddr2, val, val2, val3); case FUTEX_LOCK_PI: - if (futex_cmpxchg_enabled) - ret = futex_lock_pi(uaddr, flags, val, timeout, 0); - break; + return futex_lock_pi(uaddr, flags, val, timeout, 0); case FUTEX_UNLOCK_PI: - if (futex_cmpxchg_enabled) - ret = futex_unlock_pi(uaddr, flags); - break; + return futex_unlock_pi(uaddr, flags); case FUTEX_TRYLOCK_PI: - if (futex_cmpxchg_enabled) - ret = futex_lock_pi(uaddr, flags, 0, timeout, 1); - break; + return futex_lock_pi(uaddr, flags, 0, timeout, 1); case FUTEX_WAIT_REQUEUE_PI: val3 = FUTEX_BITSET_MATCH_ANY; - ret = futex_wait_requeue_pi(uaddr, flags, val, timeout, val3, - uaddr2); - break; + return futex_wait_requeue_pi(uaddr, flags, val, timeout, val3, + uaddr2); case FUTEX_CMP_REQUEUE_PI: - ret = futex_requeue(uaddr, flags, uaddr2, val, val2, &val3, 1); - break; - default: - ret = -ENOSYS; + return futex_requeue(uaddr, flags, uaddr2, val, val2, &val3, 1); } - return ret; + return -ENOSYS; } diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index 838687ed802a..6080f6bc8c33 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -63,8 +63,7 @@ int irq_set_irq_type(unsigned int irq, unsigned int type) return -EINVAL; type &= IRQ_TYPE_SENSE_MASK; - if (type != IRQ_TYPE_NONE) - ret = __irq_set_trigger(desc, irq, type); + ret = __irq_set_trigger(desc, irq, type); irq_put_desc_busunlock(desc, flags); return ret; } diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 470d08c82bbe..6ff84e6a954c 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -60,7 +60,7 @@ static void irq_wake_thread(struct irq_desc *desc, struct irqaction *action) * device interrupt, so no irq storm is lurking. If the * RUNTHREAD bit is already set, nothing to do. */ - if (test_bit(IRQTF_DIED, &action->thread_flags) || + if ((action->thread->flags & PF_EXITING) || test_and_set_bit(IRQTF_RUNTHREAD, &action->thread_flags)) return; @@ -110,6 +110,18 @@ static void irq_wake_thread(struct irq_desc *desc, struct irqaction *action) * threads_oneshot untouched and runs the thread another time. */ desc->threads_oneshot |= action->thread_mask; + + /* + * We increment the threads_active counter in case we wake up + * the irq thread. The irq thread decrements the counter when + * it returns from the handler or in the exit path and wakes + * up waiters which are stuck in synchronize_irq() when the + * active count becomes zero. synchronize_irq() is serialized + * against this code (hard irq handler) via IRQS_INPROGRESS + * like the finalize_oneshot() code. See comment above. + */ + atomic_inc(&desc->threads_active); + wake_up_process(action->thread); } diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h index 40378ff877e7..8e5c56b3b7d9 100644 --- a/kernel/irq/internals.h +++ b/kernel/irq/internals.h @@ -20,14 +20,12 @@ extern bool noirqdebug; /* * Bits used by threaded handlers: * IRQTF_RUNTHREAD - signals that the interrupt handler thread should run - * IRQTF_DIED - handler thread died * IRQTF_WARNED - warning "IRQ_WAKE_THREAD w/o thread_fn" has been printed * IRQTF_AFFINITY - irq thread is requested to adjust affinity * IRQTF_FORCED_THREAD - irq action is force threaded */ enum { IRQTF_RUNTHREAD, - IRQTF_DIED, IRQTF_WARNED, IRQTF_AFFINITY, IRQTF_FORCED_THREAD, diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index 0f0d4704ddd8..b0ccd1ac2d6a 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -759,6 +759,13 @@ static irqreturn_t irq_thread_fn(struct irq_desc *desc, return ret; } +static void wake_threads_waitq(struct irq_desc *desc) +{ + if (atomic_dec_and_test(&desc->threads_active) && + waitqueue_active(&desc->wait_for_threads)) + wake_up(&desc->wait_for_threads); +} + /* * Interrupt handler thread */ @@ -771,57 +778,41 @@ static int irq_thread(void *data) struct irq_desc *desc = irq_to_desc(action->irq); irqreturn_t (*handler_fn)(struct irq_desc *desc, struct irqaction *action); - int wake; - if (force_irqthreads & test_bit(IRQTF_FORCED_THREAD, + if (force_irqthreads && test_bit(IRQTF_FORCED_THREAD, &action->thread_flags)) handler_fn = irq_forced_thread_fn; else handler_fn = irq_thread_fn; sched_setscheduler(current, SCHED_FIFO, ¶m); - current->irqaction = action; + current->irq_thread = 1; while (!irq_wait_for_interrupt(action)) { + irqreturn_t action_ret; irq_thread_check_affinity(desc, action); - atomic_inc(&desc->threads_active); + action_ret = handler_fn(desc, action); + if (!noirqdebug) + note_interrupt(action->irq, desc, action_ret); - raw_spin_lock_irq(&desc->lock); - if (unlikely(irqd_irq_disabled(&desc->irq_data))) { - /* - * CHECKME: We might need a dedicated - * IRQ_THREAD_PENDING flag here, which - * retriggers the thread in check_irq_resend() - * but AFAICT IRQS_PENDING should be fine as it - * retriggers the interrupt itself --- tglx - */ - desc->istate |= IRQS_PENDING; - raw_spin_unlock_irq(&desc->lock); - } else { - irqreturn_t action_ret; - - raw_spin_unlock_irq(&desc->lock); - action_ret = handler_fn(desc, action); - if (!noirqdebug) - note_interrupt(action->irq, desc, action_ret); - } - - wake = atomic_dec_and_test(&desc->threads_active); - - if (wake && waitqueue_active(&desc->wait_for_threads)) - wake_up(&desc->wait_for_threads); + wake_threads_waitq(desc); } - /* Prevent a stale desc->threads_oneshot */ - irq_finalize_oneshot(desc, action, true); - /* - * Clear irqaction. Otherwise exit_irq_thread() would make + * This is the regular exit path. __free_irq() is stopping the + * thread via kthread_stop() after calling + * synchronize_irq(). So neither IRQTF_RUNTHREAD nor the + * oneshot mask bit can be set. We cannot verify that as we + * cannot touch the oneshot mask at this point anymore as + * __setup_irq() might have given out currents thread_mask + * again. + * + * Clear irq_thread. Otherwise exit_irq_thread() would make * fuzz about an active irq thread going into nirvana. */ - current->irqaction = NULL; + current->irq_thread = 0; return 0; } @@ -832,27 +823,28 @@ void exit_irq_thread(void) { struct task_struct *tsk = current; struct irq_desc *desc; + struct irqaction *action; - if (!tsk->irqaction) + if (!tsk->irq_thread) return; + action = kthread_data(tsk); + printk(KERN_ERR "exiting task \"%s\" (%d) is an active IRQ thread (irq %d)\n", - tsk->comm ? tsk->comm : "", tsk->pid, tsk->irqaction->irq); + tsk->comm ? tsk->comm : "", tsk->pid, action->irq); - desc = irq_to_desc(tsk->irqaction->irq); + desc = irq_to_desc(action->irq); /* - * Prevent a stale desc->threads_oneshot. Must be called - * before setting the IRQTF_DIED flag. + * If IRQTF_RUNTHREAD is set, we need to decrement + * desc->threads_active and wake possible waiters. */ - irq_finalize_oneshot(desc, tsk->irqaction, true); + if (test_and_clear_bit(IRQTF_RUNTHREAD, &action->thread_flags)) + wake_threads_waitq(desc); - /* - * Set the THREAD DIED flag to prevent further wakeups of the - * soon to be gone threaded handler. - */ - set_bit(IRQTF_DIED, &tsk->irqaction->flags); + /* Prevent a stale desc->threads_oneshot */ + irq_finalize_oneshot(desc, action, true); } static void irq_setup_forced_threading(struct irqaction *new) @@ -1135,8 +1127,7 @@ out_thread: struct task_struct *t = new->thread; new->thread = NULL; - if (likely(!test_bit(IRQTF_DIED, &new->thread_flags))) - kthread_stop(t); + kthread_stop(t); put_task_struct(t); } out_mput: @@ -1246,8 +1237,7 @@ static struct irqaction *__free_irq(unsigned int irq, void *dev_id) #endif if (action->thread) { - if (!test_bit(IRQTF_DIED, &action->thread_flags)) - kthread_stop(action->thread); + kthread_stop(action->thread); put_task_struct(action->thread); } diff --git a/kernel/lockdep.c b/kernel/lockdep.c index 8889f7dd7c46..ea9ee4518c35 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -4176,7 +4176,13 @@ void lockdep_rcu_suspicious(const char *file, const int line, const char *s) printk("-------------------------------\n"); printk("%s:%d %s!\n", file, line, s); printk("\nother info that might help us debug this:\n\n"); - printk("\nrcu_scheduler_active = %d, debug_locks = %d\n", rcu_scheduler_active, debug_locks); + printk("\n%srcu_scheduler_active = %d, debug_locks = %d\n", + !rcu_lockdep_current_cpu_online() + ? "RCU used illegally from offline CPU!\n" + : rcu_is_cpu_idle() + ? "RCU used illegally from idle CPU!\n" + : "", + rcu_scheduler_active, debug_locks); /* * If a CPU is in the RCU-free window in idle (ie: in the section diff --git a/kernel/rcu.h b/kernel/rcu.h index aa88baab5f78..8ba99cdc6515 100644 --- a/kernel/rcu.h +++ b/kernel/rcu.h @@ -33,8 +33,27 @@ * Process-level increment to ->dynticks_nesting field. This allows for * architectures that use half-interrupts and half-exceptions from * process context. + * + * DYNTICK_TASK_NEST_MASK defines a field of width DYNTICK_TASK_NEST_WIDTH + * that counts the number of process-based reasons why RCU cannot + * consider the corresponding CPU to be idle, and DYNTICK_TASK_NEST_VALUE + * is the value used to increment or decrement this field. + * + * The rest of the bits could in principle be used to count interrupts, + * but this would mean that a negative-one value in the interrupt + * field could incorrectly zero out the DYNTICK_TASK_NEST_MASK field. + * We therefore provide a two-bit guard field defined by DYNTICK_TASK_MASK + * that is set to DYNTICK_TASK_FLAG upon initial exit from idle. + * The DYNTICK_TASK_EXIT_IDLE value is thus the combined value used upon + * initial exit from idle. */ -#define DYNTICK_TASK_NESTING (LLONG_MAX / 2 - 1) +#define DYNTICK_TASK_NEST_WIDTH 7 +#define DYNTICK_TASK_NEST_VALUE ((LLONG_MAX >> DYNTICK_TASK_NEST_WIDTH) + 1) +#define DYNTICK_TASK_NEST_MASK (LLONG_MAX - DYNTICK_TASK_NEST_VALUE + 1) +#define DYNTICK_TASK_FLAG ((DYNTICK_TASK_NEST_VALUE / 8) * 2) +#define DYNTICK_TASK_MASK ((DYNTICK_TASK_NEST_VALUE / 8) * 3) +#define DYNTICK_TASK_EXIT_IDLE (DYNTICK_TASK_NEST_VALUE + \ + DYNTICK_TASK_FLAG) /* * debug_rcu_head_queue()/debug_rcu_head_unqueue() are used internally @@ -50,7 +69,6 @@ extern struct debug_obj_descr rcuhead_debug_descr; static inline void debug_rcu_head_queue(struct rcu_head *head) { - WARN_ON_ONCE((unsigned long)head & 0x3); debug_object_activate(head, &rcuhead_debug_descr); debug_object_active_state(head, &rcuhead_debug_descr, STATE_RCU_HEAD_READY, @@ -76,16 +94,18 @@ static inline void debug_rcu_head_unqueue(struct rcu_head *head) extern void kfree(const void *); -static inline void __rcu_reclaim(char *rn, struct rcu_head *head) +static inline bool __rcu_reclaim(char *rn, struct rcu_head *head) { unsigned long offset = (unsigned long)head->func; if (__is_kfree_rcu_offset(offset)) { RCU_TRACE(trace_rcu_invoke_kfree_callback(rn, head, offset)); kfree((void *)head - offset); + return 1; } else { RCU_TRACE(trace_rcu_invoke_callback(rn, head)); head->func(head); + return 0; } } diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index 2bc4e135ff23..a86f1741cc27 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -88,6 +88,9 @@ EXPORT_SYMBOL_GPL(debug_lockdep_rcu_enabled); * section. * * Check debug_lockdep_rcu_enabled() to prevent false positives during boot. + * + * Note that rcu_read_lock() is disallowed if the CPU is either idle or + * offline from an RCU perspective, so check for those as well. */ int rcu_read_lock_bh_held(void) { @@ -95,6 +98,8 @@ int rcu_read_lock_bh_held(void) return 1; if (rcu_is_cpu_idle()) return 0; + if (!rcu_lockdep_current_cpu_online()) + return 0; return in_softirq() || irqs_disabled(); } EXPORT_SYMBOL_GPL(rcu_read_lock_bh_held); diff --git a/kernel/rcutiny.c b/kernel/rcutiny.c index 977296dca0a4..37a5444204d2 100644 --- a/kernel/rcutiny.c +++ b/kernel/rcutiny.c @@ -53,7 +53,7 @@ static void __call_rcu(struct rcu_head *head, #include "rcutiny_plugin.h" -static long long rcu_dynticks_nesting = DYNTICK_TASK_NESTING; +static long long rcu_dynticks_nesting = DYNTICK_TASK_EXIT_IDLE; /* Common code for rcu_idle_enter() and rcu_irq_exit(), see kernel/rcutree.c. */ static void rcu_idle_enter_common(long long oldval) @@ -88,10 +88,16 @@ void rcu_idle_enter(void) local_irq_save(flags); oldval = rcu_dynticks_nesting; - rcu_dynticks_nesting = 0; + WARN_ON_ONCE((rcu_dynticks_nesting & DYNTICK_TASK_NEST_MASK) == 0); + if ((rcu_dynticks_nesting & DYNTICK_TASK_NEST_MASK) == + DYNTICK_TASK_NEST_VALUE) + rcu_dynticks_nesting = 0; + else + rcu_dynticks_nesting -= DYNTICK_TASK_NEST_VALUE; rcu_idle_enter_common(oldval); local_irq_restore(flags); } +EXPORT_SYMBOL_GPL(rcu_idle_enter); /* * Exit an interrupt handler towards idle. @@ -140,11 +146,15 @@ void rcu_idle_exit(void) local_irq_save(flags); oldval = rcu_dynticks_nesting; - WARN_ON_ONCE(oldval != 0); - rcu_dynticks_nesting = DYNTICK_TASK_NESTING; + WARN_ON_ONCE(rcu_dynticks_nesting < 0); + if (rcu_dynticks_nesting & DYNTICK_TASK_NEST_MASK) + rcu_dynticks_nesting += DYNTICK_TASK_NEST_VALUE; + else + rcu_dynticks_nesting = DYNTICK_TASK_EXIT_IDLE; rcu_idle_exit_common(oldval); local_irq_restore(flags); } +EXPORT_SYMBOL_GPL(rcu_idle_exit); /* * Enter an interrupt handler, moving away from idle. @@ -258,7 +268,7 @@ static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp) /* If no RCU callbacks ready to invoke, just return. */ if (&rcp->rcucblist == rcp->donetail) { - RCU_TRACE(trace_rcu_batch_start(rcp->name, 0, -1)); + RCU_TRACE(trace_rcu_batch_start(rcp->name, 0, 0, -1)); RCU_TRACE(trace_rcu_batch_end(rcp->name, 0, ACCESS_ONCE(rcp->rcucblist), need_resched(), @@ -269,7 +279,7 @@ static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp) /* Move the ready-to-invoke callbacks to a local list. */ local_irq_save(flags); - RCU_TRACE(trace_rcu_batch_start(rcp->name, 0, -1)); + RCU_TRACE(trace_rcu_batch_start(rcp->name, 0, rcp->qlen, -1)); list = rcp->rcucblist; rcp->rcucblist = *rcp->donetail; *rcp->donetail = NULL; @@ -319,6 +329,10 @@ static void rcu_process_callbacks(struct softirq_action *unused) */ void synchronize_sched(void) { + rcu_lockdep_assert(!lock_is_held(&rcu_bh_lock_map) && + !lock_is_held(&rcu_lock_map) && + !lock_is_held(&rcu_sched_lock_map), + "Illegal synchronize_sched() in RCU read-side critical section"); cond_resched(); } EXPORT_SYMBOL_GPL(synchronize_sched); diff --git a/kernel/rcutiny_plugin.h b/kernel/rcutiny_plugin.h index 9cb1ae4aabdd..22ecea0dfb62 100644 --- a/kernel/rcutiny_plugin.h +++ b/kernel/rcutiny_plugin.h @@ -132,6 +132,7 @@ static struct rcu_preempt_ctrlblk rcu_preempt_ctrlblk = { RCU_TRACE(.rcb.name = "rcu_preempt") }; +static void rcu_read_unlock_special(struct task_struct *t); static int rcu_preempted_readers_exp(void); static void rcu_report_exp_done(void); @@ -146,6 +147,16 @@ static int rcu_cpu_blocking_cur_gp(void) /* * Check for a running RCU reader. Because there is only one CPU, * there can be but one running RCU reader at a time. ;-) + * + * Returns zero if there are no running readers. Returns a positive + * number if there is at least one reader within its RCU read-side + * critical section. Returns a negative number if an outermost reader + * is in the midst of exiting from its RCU read-side critical section + * + * Returns zero if there are no running readers. Returns a positive + * number if there is at least one reader within its RCU read-side + * critical section. Returns a negative number if an outermost reader + * is in the midst of exiting from its RCU read-side critical section. */ static int rcu_preempt_running_reader(void) { @@ -307,7 +318,6 @@ static int rcu_boost(void) t = container_of(tb, struct task_struct, rcu_node_entry); rt_mutex_init_proxy_locked(&mtx, t); t->rcu_boost_mutex = &mtx; - t->rcu_read_unlock_special |= RCU_READ_UNLOCK_BOOSTED; raw_local_irq_restore(flags); rt_mutex_lock(&mtx); rt_mutex_unlock(&mtx); /* Keep lockdep happy. */ @@ -475,7 +485,7 @@ void rcu_preempt_note_context_switch(void) unsigned long flags; local_irq_save(flags); /* must exclude scheduler_tick(). */ - if (rcu_preempt_running_reader() && + if (rcu_preempt_running_reader() > 0 && (t->rcu_read_unlock_special & RCU_READ_UNLOCK_BLOCKED) == 0) { /* Possibly blocking in an RCU read-side critical section. */ @@ -494,6 +504,13 @@ void rcu_preempt_note_context_switch(void) list_add(&t->rcu_node_entry, &rcu_preempt_ctrlblk.blkd_tasks); if (rcu_cpu_blocking_cur_gp()) rcu_preempt_ctrlblk.gp_tasks = &t->rcu_node_entry; + } else if (rcu_preempt_running_reader() < 0 && + t->rcu_read_unlock_special) { + /* + * Complete exit from RCU read-side critical section on + * behalf of preempted instance of __rcu_read_unlock(). + */ + rcu_read_unlock_special(t); } /* @@ -526,12 +543,15 @@ EXPORT_SYMBOL_GPL(__rcu_read_lock); * notify RCU core processing or task having blocked during the RCU * read-side critical section. */ -static void rcu_read_unlock_special(struct task_struct *t) +static noinline void rcu_read_unlock_special(struct task_struct *t) { int empty; int empty_exp; unsigned long flags; struct list_head *np; +#ifdef CONFIG_RCU_BOOST + struct rt_mutex *rbmp = NULL; +#endif /* #ifdef CONFIG_RCU_BOOST */ int special; /* @@ -552,7 +572,7 @@ static void rcu_read_unlock_special(struct task_struct *t) rcu_preempt_cpu_qs(); /* Hardware IRQ handlers cannot block. */ - if (in_irq()) { + if (in_irq() || in_serving_softirq()) { local_irq_restore(flags); return; } @@ -597,10 +617,10 @@ static void rcu_read_unlock_special(struct task_struct *t) } #ifdef CONFIG_RCU_BOOST /* Unboost self if was boosted. */ - if (special & RCU_READ_UNLOCK_BOOSTED) { - t->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_BOOSTED; - rt_mutex_unlock(t->rcu_boost_mutex); + if (t->rcu_boost_mutex != NULL) { + rbmp = t->rcu_boost_mutex; t->rcu_boost_mutex = NULL; + rt_mutex_unlock(rbmp); } #endif /* #ifdef CONFIG_RCU_BOOST */ local_irq_restore(flags); @@ -618,13 +638,22 @@ void __rcu_read_unlock(void) struct task_struct *t = current; barrier(); /* needed if we ever invoke rcu_read_unlock in rcutiny.c */ - --t->rcu_read_lock_nesting; - barrier(); /* decrement before load of ->rcu_read_unlock_special */ - if (t->rcu_read_lock_nesting == 0 && - unlikely(ACCESS_ONCE(t->rcu_read_unlock_special))) - rcu_read_unlock_special(t); + if (t->rcu_read_lock_nesting != 1) + --t->rcu_read_lock_nesting; + else { + t->rcu_read_lock_nesting = INT_MIN; + barrier(); /* assign before ->rcu_read_unlock_special load */ + if (unlikely(ACCESS_ONCE(t->rcu_read_unlock_special))) + rcu_read_unlock_special(t); + barrier(); /* ->rcu_read_unlock_special load before assign */ + t->rcu_read_lock_nesting = 0; + } #ifdef CONFIG_PROVE_LOCKING - WARN_ON_ONCE(t->rcu_read_lock_nesting < 0); + { + int rrln = ACCESS_ONCE(t->rcu_read_lock_nesting); + + WARN_ON_ONCE(rrln < 0 && rrln > INT_MIN / 2); + } #endif /* #ifdef CONFIG_PROVE_LOCKING */ } EXPORT_SYMBOL_GPL(__rcu_read_unlock); @@ -649,7 +678,7 @@ static void rcu_preempt_check_callbacks(void) invoke_rcu_callbacks(); if (rcu_preempt_gp_in_progress() && rcu_cpu_blocking_cur_gp() && - rcu_preempt_running_reader()) + rcu_preempt_running_reader() > 0) t->rcu_read_unlock_special |= RCU_READ_UNLOCK_NEED_QS; } @@ -706,6 +735,11 @@ EXPORT_SYMBOL_GPL(call_rcu); */ void synchronize_rcu(void) { + rcu_lockdep_assert(!lock_is_held(&rcu_bh_lock_map) && + !lock_is_held(&rcu_lock_map) && + !lock_is_held(&rcu_sched_lock_map), + "Illegal synchronize_rcu() in RCU read-side critical section"); + #ifdef CONFIG_DEBUG_LOCK_ALLOC if (!rcu_scheduler_active) return; @@ -882,7 +916,8 @@ static void rcu_preempt_process_callbacks(void) static void invoke_rcu_callbacks(void) { have_rcu_kthread_work = 1; - wake_up(&rcu_kthread_wq); + if (rcu_kthread_task != NULL) + wake_up(&rcu_kthread_wq); } #ifdef CONFIG_RCU_TRACE @@ -943,12 +978,16 @@ early_initcall(rcu_spawn_kthreads); #else /* #ifdef CONFIG_RCU_BOOST */ +/* Hold off callback invocation until early_initcall() time. */ +static int rcu_scheduler_fully_active __read_mostly; + /* * Start up softirq processing of callbacks. */ void invoke_rcu_callbacks(void) { - raise_softirq(RCU_SOFTIRQ); + if (rcu_scheduler_fully_active) + raise_softirq(RCU_SOFTIRQ); } #ifdef CONFIG_RCU_TRACE @@ -963,10 +1002,14 @@ static bool rcu_is_callbacks_kthread(void) #endif /* #ifdef CONFIG_RCU_TRACE */ -void rcu_init(void) +static int __init rcu_scheduler_really_started(void) { + rcu_scheduler_fully_active = 1; open_softirq(RCU_SOFTIRQ, rcu_process_callbacks); + raise_softirq(RCU_SOFTIRQ); /* Invoke any callbacks from early boot. */ + return 0; } +early_initcall(rcu_scheduler_really_started); #endif /* #else #ifdef CONFIG_RCU_BOOST */ diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index a58ac285fc69..a89b381a8c6e 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -65,7 +65,10 @@ static int fqs_duration; /* Duration of bursts (us), 0 to disable. */ static int fqs_holdoff; /* Hold time within burst (us). */ static int fqs_stutter = 3; /* Wait time between bursts (s). */ static int onoff_interval; /* Wait time between CPU hotplugs, 0=disable. */ +static int onoff_holdoff; /* Seconds after boot before CPU hotplugs. */ static int shutdown_secs; /* Shutdown time (s). <=0 for no shutdown. */ +static int stall_cpu; /* CPU-stall duration (s). 0 for no stall. */ +static int stall_cpu_holdoff = 10; /* Time to wait until stall (s). */ static int test_boost = 1; /* Test RCU prio boost: 0=no, 1=maybe, 2=yes. */ static int test_boost_interval = 7; /* Interval between boost tests, seconds. */ static int test_boost_duration = 4; /* Duration of each boost test, seconds. */ @@ -95,8 +98,14 @@ module_param(fqs_stutter, int, 0444); MODULE_PARM_DESC(fqs_stutter, "Wait time between fqs bursts (s)"); module_param(onoff_interval, int, 0444); MODULE_PARM_DESC(onoff_interval, "Time between CPU hotplugs (s), 0=disable"); +module_param(onoff_holdoff, int, 0444); +MODULE_PARM_DESC(onoff_holdoff, "Time after boot before CPU hotplugs (s)"); module_param(shutdown_secs, int, 0444); MODULE_PARM_DESC(shutdown_secs, "Shutdown time (s), zero to disable."); +module_param(stall_cpu, int, 0444); +MODULE_PARM_DESC(stall_cpu, "Stall duration (s), zero to disable."); +module_param(stall_cpu_holdoff, int, 0444); +MODULE_PARM_DESC(stall_cpu_holdoff, "Time to wait before starting stall (s)."); module_param(test_boost, int, 0444); MODULE_PARM_DESC(test_boost, "Test RCU prio boost: 0=no, 1=maybe, 2=yes."); module_param(test_boost_interval, int, 0444); @@ -129,6 +138,7 @@ static struct task_struct *shutdown_task; #ifdef CONFIG_HOTPLUG_CPU static struct task_struct *onoff_task; #endif /* #ifdef CONFIG_HOTPLUG_CPU */ +static struct task_struct *stall_task; #define RCU_TORTURE_PIPE_LEN 10 @@ -990,12 +1000,12 @@ static void rcu_torture_timer(unsigned long unused) rcu_read_lock_bh_held() || rcu_read_lock_sched_held() || srcu_read_lock_held(&srcu_ctl)); - do_trace_rcu_torture_read(cur_ops->name, &p->rtort_rcu); if (p == NULL) { /* Leave because rcu_torture_writer is not yet underway */ cur_ops->readunlock(idx); return; } + do_trace_rcu_torture_read(cur_ops->name, &p->rtort_rcu); if (p->rtort_mbtest == 0) atomic_inc(&n_rcu_torture_mberror); spin_lock(&rand_lock); @@ -1053,13 +1063,13 @@ rcu_torture_reader(void *arg) rcu_read_lock_bh_held() || rcu_read_lock_sched_held() || srcu_read_lock_held(&srcu_ctl)); - do_trace_rcu_torture_read(cur_ops->name, &p->rtort_rcu); if (p == NULL) { /* Wait for rcu_torture_writer to get underway */ cur_ops->readunlock(idx); schedule_timeout_interruptible(HZ); continue; } + do_trace_rcu_torture_read(cur_ops->name, &p->rtort_rcu); if (p->rtort_mbtest == 0) atomic_inc(&n_rcu_torture_mberror); cur_ops->read_delay(&rand); @@ -1300,13 +1310,13 @@ rcu_torture_print_module_parms(struct rcu_torture_ops *cur_ops, char *tag) "fqs_duration=%d fqs_holdoff=%d fqs_stutter=%d " "test_boost=%d/%d test_boost_interval=%d " "test_boost_duration=%d shutdown_secs=%d " - "onoff_interval=%d\n", + "onoff_interval=%d onoff_holdoff=%d\n", torture_type, tag, nrealreaders, nfakewriters, stat_interval, verbose, test_no_idle_hz, shuffle_interval, stutter, irqreader, fqs_duration, fqs_holdoff, fqs_stutter, test_boost, cur_ops->can_boost, test_boost_interval, test_boost_duration, shutdown_secs, - onoff_interval); + onoff_interval, onoff_holdoff); } static struct notifier_block rcutorture_shutdown_nb = { @@ -1410,6 +1420,11 @@ rcu_torture_onoff(void *arg) for_each_online_cpu(cpu) maxcpu = cpu; WARN_ON(maxcpu < 0); + if (onoff_holdoff > 0) { + VERBOSE_PRINTK_STRING("rcu_torture_onoff begin holdoff"); + schedule_timeout_interruptible(onoff_holdoff * HZ); + VERBOSE_PRINTK_STRING("rcu_torture_onoff end holdoff"); + } while (!kthread_should_stop()) { cpu = (rcu_random(&rand) >> 4) % (maxcpu + 1); if (cpu_online(cpu) && cpu_is_hotpluggable(cpu)) { @@ -1450,12 +1465,15 @@ rcu_torture_onoff(void *arg) static int __cpuinit rcu_torture_onoff_init(void) { + int ret; + if (onoff_interval <= 0) return 0; onoff_task = kthread_run(rcu_torture_onoff, NULL, "rcu_torture_onoff"); if (IS_ERR(onoff_task)) { + ret = PTR_ERR(onoff_task); onoff_task = NULL; - return PTR_ERR(onoff_task); + return ret; } return 0; } @@ -1481,6 +1499,63 @@ static void rcu_torture_onoff_cleanup(void) #endif /* #else #ifdef CONFIG_HOTPLUG_CPU */ +/* + * CPU-stall kthread. It waits as specified by stall_cpu_holdoff, then + * induces a CPU stall for the time specified by stall_cpu. + */ +static int __cpuinit rcu_torture_stall(void *args) +{ + unsigned long stop_at; + + VERBOSE_PRINTK_STRING("rcu_torture_stall task started"); + if (stall_cpu_holdoff > 0) { + VERBOSE_PRINTK_STRING("rcu_torture_stall begin holdoff"); + schedule_timeout_interruptible(stall_cpu_holdoff * HZ); + VERBOSE_PRINTK_STRING("rcu_torture_stall end holdoff"); + } + if (!kthread_should_stop()) { + stop_at = get_seconds() + stall_cpu; + /* RCU CPU stall is expected behavior in following code. */ + printk(KERN_ALERT "rcu_torture_stall start.\n"); + rcu_read_lock(); + preempt_disable(); + while (ULONG_CMP_LT(get_seconds(), stop_at)) + continue; /* Induce RCU CPU stall warning. */ + preempt_enable(); + rcu_read_unlock(); + printk(KERN_ALERT "rcu_torture_stall end.\n"); + } + rcutorture_shutdown_absorb("rcu_torture_stall"); + while (!kthread_should_stop()) + schedule_timeout_interruptible(10 * HZ); + return 0; +} + +/* Spawn CPU-stall kthread, if stall_cpu specified. */ +static int __init rcu_torture_stall_init(void) +{ + int ret; + + if (stall_cpu <= 0) + return 0; + stall_task = kthread_run(rcu_torture_stall, NULL, "rcu_torture_stall"); + if (IS_ERR(stall_task)) { + ret = PTR_ERR(stall_task); + stall_task = NULL; + return ret; + } + return 0; +} + +/* Clean up after the CPU-stall kthread, if one was spawned. */ +static void rcu_torture_stall_cleanup(void) +{ + if (stall_task == NULL) + return; + VERBOSE_PRINTK_STRING("Stopping rcu_torture_stall_task."); + kthread_stop(stall_task); +} + static int rcutorture_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) { @@ -1523,6 +1598,7 @@ rcu_torture_cleanup(void) fullstop = FULLSTOP_RMMOD; mutex_unlock(&fullstop_mutex); unregister_reboot_notifier(&rcutorture_shutdown_nb); + rcu_torture_stall_cleanup(); if (stutter_task) { VERBOSE_PRINTK_STRING("Stopping rcu_torture_stutter task"); kthread_stop(stutter_task); @@ -1602,6 +1678,10 @@ rcu_torture_cleanup(void) cur_ops->cleanup(); if (atomic_read(&n_rcu_torture_error)) rcu_torture_print_module_parms(cur_ops, "End of test: FAILURE"); + else if (n_online_successes != n_online_attempts || + n_offline_successes != n_offline_attempts) + rcu_torture_print_module_parms(cur_ops, + "End of test: RCU_HOTPLUG"); else rcu_torture_print_module_parms(cur_ops, "End of test: SUCCESS"); } @@ -1819,6 +1899,7 @@ rcu_torture_init(void) } rcu_torture_onoff_init(); register_reboot_notifier(&rcutorture_shutdown_nb); + rcu_torture_stall_init(); rcutorture_record_test_transition(); mutex_unlock(&fullstop_mutex); return 0; diff --git a/kernel/rcutree.c b/kernel/rcutree.c index 6c4a6722abfd..1050d6d3922c 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -50,6 +50,8 @@ #include <linux/wait.h> #include <linux/kthread.h> #include <linux/prefetch.h> +#include <linux/delay.h> +#include <linux/stop_machine.h> #include "rcutree.h" #include <trace/events/rcu.h> @@ -196,7 +198,7 @@ void rcu_note_context_switch(int cpu) EXPORT_SYMBOL_GPL(rcu_note_context_switch); DEFINE_PER_CPU(struct rcu_dynticks, rcu_dynticks) = { - .dynticks_nesting = DYNTICK_TASK_NESTING, + .dynticks_nesting = DYNTICK_TASK_EXIT_IDLE, .dynticks = ATOMIC_INIT(1), }; @@ -208,8 +210,11 @@ module_param(blimit, int, 0); module_param(qhimark, int, 0); module_param(qlowmark, int, 0); -int rcu_cpu_stall_suppress __read_mostly; +int rcu_cpu_stall_suppress __read_mostly; /* 1 = suppress stall warnings. */ +int rcu_cpu_stall_timeout __read_mostly = CONFIG_RCU_CPU_STALL_TIMEOUT; + module_param(rcu_cpu_stall_suppress, int, 0644); +module_param(rcu_cpu_stall_timeout, int, 0644); static void force_quiescent_state(struct rcu_state *rsp, int relaxed); static int rcu_pending(int cpu); @@ -301,8 +306,6 @@ static struct rcu_node *rcu_get_root(struct rcu_state *rsp) return &rsp->node[0]; } -#ifdef CONFIG_SMP - /* * If the specified CPU is offline, tell the caller that it is in * a quiescent state. Otherwise, whack it with a reschedule IPI. @@ -317,30 +320,21 @@ static struct rcu_node *rcu_get_root(struct rcu_state *rsp) static int rcu_implicit_offline_qs(struct rcu_data *rdp) { /* - * If the CPU is offline, it is in a quiescent state. We can - * trust its state not to change because interrupts are disabled. + * If the CPU is offline for more than a jiffy, it is in a quiescent + * state. We can trust its state not to change because interrupts + * are disabled. The reason for the jiffy's worth of slack is to + * handle CPUs initializing on the way up and finding their way + * to the idle loop on the way down. */ - if (cpu_is_offline(rdp->cpu)) { + if (cpu_is_offline(rdp->cpu) && + ULONG_CMP_LT(rdp->rsp->gp_start + 2, jiffies)) { trace_rcu_fqs(rdp->rsp->name, rdp->gpnum, rdp->cpu, "ofl"); rdp->offline_fqs++; return 1; } - - /* - * The CPU is online, so send it a reschedule IPI. This forces - * it through the scheduler, and (inefficiently) also handles cases - * where idle loops fail to inform RCU about the CPU being idle. - */ - if (rdp->cpu != smp_processor_id()) - smp_send_reschedule(rdp->cpu); - else - set_need_resched(); - rdp->resched_ipi++; return 0; } -#endif /* #ifdef CONFIG_SMP */ - /* * rcu_idle_enter_common - inform RCU that current CPU is moving towards idle * @@ -366,6 +360,17 @@ static void rcu_idle_enter_common(struct rcu_dynticks *rdtp, long long oldval) atomic_inc(&rdtp->dynticks); smp_mb__after_atomic_inc(); /* Force ordering with next sojourn. */ WARN_ON_ONCE(atomic_read(&rdtp->dynticks) & 0x1); + + /* + * The idle task is not permitted to enter the idle loop while + * in an RCU read-side critical section. + */ + rcu_lockdep_assert(!lock_is_held(&rcu_lock_map), + "Illegal idle entry in RCU read-side critical section."); + rcu_lockdep_assert(!lock_is_held(&rcu_bh_lock_map), + "Illegal idle entry in RCU-bh read-side critical section."); + rcu_lockdep_assert(!lock_is_held(&rcu_sched_lock_map), + "Illegal idle entry in RCU-sched read-side critical section."); } /** @@ -389,10 +394,15 @@ void rcu_idle_enter(void) local_irq_save(flags); rdtp = &__get_cpu_var(rcu_dynticks); oldval = rdtp->dynticks_nesting; - rdtp->dynticks_nesting = 0; + WARN_ON_ONCE((oldval & DYNTICK_TASK_NEST_MASK) == 0); + if ((oldval & DYNTICK_TASK_NEST_MASK) == DYNTICK_TASK_NEST_VALUE) + rdtp->dynticks_nesting = 0; + else + rdtp->dynticks_nesting -= DYNTICK_TASK_NEST_VALUE; rcu_idle_enter_common(rdtp, oldval); local_irq_restore(flags); } +EXPORT_SYMBOL_GPL(rcu_idle_enter); /** * rcu_irq_exit - inform RCU that current CPU is exiting irq towards idle @@ -462,7 +472,7 @@ static void rcu_idle_exit_common(struct rcu_dynticks *rdtp, long long oldval) * Exit idle mode, in other words, -enter- the mode in which RCU * read-side critical sections can occur. * - * We crowbar the ->dynticks_nesting field to DYNTICK_TASK_NESTING to + * We crowbar the ->dynticks_nesting field to DYNTICK_TASK_NEST to * allow for the possibility of usermode upcalls messing up our count * of interrupt nesting level during the busy period that is just * now starting. @@ -476,11 +486,15 @@ void rcu_idle_exit(void) local_irq_save(flags); rdtp = &__get_cpu_var(rcu_dynticks); oldval = rdtp->dynticks_nesting; - WARN_ON_ONCE(oldval != 0); - rdtp->dynticks_nesting = DYNTICK_TASK_NESTING; + WARN_ON_ONCE(oldval < 0); + if (oldval & DYNTICK_TASK_NEST_MASK) + rdtp->dynticks_nesting += DYNTICK_TASK_NEST_VALUE; + else + rdtp->dynticks_nesting = DYNTICK_TASK_EXIT_IDLE; rcu_idle_exit_common(rdtp, oldval); local_irq_restore(flags); } +EXPORT_SYMBOL_GPL(rcu_idle_exit); /** * rcu_irq_enter - inform RCU that current CPU is entering irq away from idle @@ -581,6 +595,49 @@ int rcu_is_cpu_idle(void) } EXPORT_SYMBOL(rcu_is_cpu_idle); +#ifdef CONFIG_HOTPLUG_CPU + +/* + * Is the current CPU online? Disable preemption to avoid false positives + * that could otherwise happen due to the current CPU number being sampled, + * this task being preempted, its old CPU being taken offline, resuming + * on some other CPU, then determining that its old CPU is now offline. + * It is OK to use RCU on an offline processor during initial boot, hence + * the check for rcu_scheduler_fully_active. Note also that it is OK + * for a CPU coming online to use RCU for one jiffy prior to marking itself + * online in the cpu_online_mask. Similarly, it is OK for a CPU going + * offline to continue to use RCU for one jiffy after marking itself + * offline in the cpu_online_mask. This leniency is necessary given the + * non-atomic nature of the online and offline processing, for example, + * the fact that a CPU enters the scheduler after completing the CPU_DYING + * notifiers. + * + * This is also why RCU internally marks CPUs online during the + * CPU_UP_PREPARE phase and offline during the CPU_DEAD phase. + * + * Disable checking if in an NMI handler because we cannot safely report + * errors from NMI handlers anyway. + */ +bool rcu_lockdep_current_cpu_online(void) +{ + struct rcu_data *rdp; + struct rcu_node *rnp; + bool ret; + + if (in_nmi()) + return 1; + preempt_disable(); + rdp = &__get_cpu_var(rcu_sched_data); + rnp = rdp->mynode; + ret = (rdp->grpmask & rnp->qsmaskinit) || + !rcu_scheduler_fully_active; + preempt_enable(); + return ret; +} +EXPORT_SYMBOL_GPL(rcu_lockdep_current_cpu_online); + +#endif /* #ifdef CONFIG_HOTPLUG_CPU */ + #endif /* #ifdef CONFIG_PROVE_RCU */ /** @@ -595,8 +652,6 @@ int rcu_is_cpu_rrupt_from_idle(void) return __get_cpu_var(rcu_dynticks).dynticks_nesting <= 1; } -#ifdef CONFIG_SMP - /* * Snapshot the specified CPU's dynticks counter so that we can later * credit them with an implicit quiescent state. Return 1 if this CPU @@ -640,12 +695,28 @@ static int rcu_implicit_dynticks_qs(struct rcu_data *rdp) return rcu_implicit_offline_qs(rdp); } -#endif /* #ifdef CONFIG_SMP */ +static int jiffies_till_stall_check(void) +{ + int till_stall_check = ACCESS_ONCE(rcu_cpu_stall_timeout); + + /* + * Limit check must be consistent with the Kconfig limits + * for CONFIG_RCU_CPU_STALL_TIMEOUT. + */ + if (till_stall_check < 3) { + ACCESS_ONCE(rcu_cpu_stall_timeout) = 3; + till_stall_check = 3; + } else if (till_stall_check > 300) { + ACCESS_ONCE(rcu_cpu_stall_timeout) = 300; + till_stall_check = 300; + } + return till_stall_check * HZ + RCU_STALL_DELAY_DELTA; +} static void record_gp_stall_check_time(struct rcu_state *rsp) { rsp->gp_start = jiffies; - rsp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_CHECK; + rsp->jiffies_stall = jiffies + jiffies_till_stall_check(); } static void print_other_cpu_stall(struct rcu_state *rsp) @@ -664,13 +735,7 @@ static void print_other_cpu_stall(struct rcu_state *rsp) raw_spin_unlock_irqrestore(&rnp->lock, flags); return; } - rsp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_RECHECK; - - /* - * Now rat on any tasks that got kicked up to the root rcu_node - * due to CPU offlining. - */ - ndetected = rcu_print_task_stall(rnp); + rsp->jiffies_stall = jiffies + 3 * jiffies_till_stall_check() + 3; raw_spin_unlock_irqrestore(&rnp->lock, flags); /* @@ -678,8 +743,9 @@ static void print_other_cpu_stall(struct rcu_state *rsp) * See Documentation/RCU/stallwarn.txt for info on how to debug * RCU CPU stall warnings. */ - printk(KERN_ERR "INFO: %s detected stalls on CPUs/tasks: {", + printk(KERN_ERR "INFO: %s detected stalls on CPUs/tasks:", rsp->name); + print_cpu_stall_info_begin(); rcu_for_each_leaf_node(rsp, rnp) { raw_spin_lock_irqsave(&rnp->lock, flags); ndetected += rcu_print_task_stall(rnp); @@ -688,11 +754,22 @@ static void print_other_cpu_stall(struct rcu_state *rsp) continue; for (cpu = 0; cpu <= rnp->grphi - rnp->grplo; cpu++) if (rnp->qsmask & (1UL << cpu)) { - printk(" %d", rnp->grplo + cpu); + print_cpu_stall_info(rsp, rnp->grplo + cpu); ndetected++; } } - printk("} (detected by %d, t=%ld jiffies)\n", + + /* + * Now rat on any tasks that got kicked up to the root rcu_node + * due to CPU offlining. + */ + rnp = rcu_get_root(rsp); + raw_spin_lock_irqsave(&rnp->lock, flags); + ndetected = rcu_print_task_stall(rnp); + raw_spin_unlock_irqrestore(&rnp->lock, flags); + + print_cpu_stall_info_end(); + printk(KERN_CONT "(detected by %d, t=%ld jiffies)\n", smp_processor_id(), (long)(jiffies - rsp->gp_start)); if (ndetected == 0) printk(KERN_ERR "INFO: Stall ended before state dump start\n"); @@ -716,15 +793,18 @@ static void print_cpu_stall(struct rcu_state *rsp) * See Documentation/RCU/stallwarn.txt for info on how to debug * RCU CPU stall warnings. */ - printk(KERN_ERR "INFO: %s detected stall on CPU %d (t=%lu jiffies)\n", - rsp->name, smp_processor_id(), jiffies - rsp->gp_start); + printk(KERN_ERR "INFO: %s self-detected stall on CPU", rsp->name); + print_cpu_stall_info_begin(); + print_cpu_stall_info(rsp, smp_processor_id()); + print_cpu_stall_info_end(); + printk(KERN_CONT " (t=%lu jiffies)\n", jiffies - rsp->gp_start); if (!trigger_all_cpu_backtrace()) dump_stack(); raw_spin_lock_irqsave(&rnp->lock, flags); if (ULONG_CMP_GE(jiffies, rsp->jiffies_stall)) - rsp->jiffies_stall = - jiffies + RCU_SECONDS_TILL_STALL_RECHECK; + rsp->jiffies_stall = jiffies + + 3 * jiffies_till_stall_check() + 3; raw_spin_unlock_irqrestore(&rnp->lock, flags); set_need_resched(); /* kick ourselves to get things going. */ @@ -807,6 +887,7 @@ static void __note_new_gpnum(struct rcu_state *rsp, struct rcu_node *rnp, struct rdp->passed_quiesce = 0; } else rdp->qs_pending = 0; + zero_cpu_stall_ticks(rdp); } } @@ -943,6 +1024,10 @@ rcu_start_gp_per_cpu(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_dat * in preparation for detecting the next grace period. The caller must hold * the root node's ->lock, which is released before return. Hard irqs must * be disabled. + * + * Note that it is legal for a dying CPU (which is marked as offline) to + * invoke this function. This can happen when the dying CPU reports its + * quiescent state. */ static void rcu_start_gp(struct rcu_state *rsp, unsigned long flags) @@ -980,26 +1065,8 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags) rsp->fqs_state = RCU_GP_INIT; /* Hold off force_quiescent_state. */ rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS; record_gp_stall_check_time(rsp); - - /* Special-case the common single-level case. */ - if (NUM_RCU_NODES == 1) { - rcu_preempt_check_blocked_tasks(rnp); - rnp->qsmask = rnp->qsmaskinit; - rnp->gpnum = rsp->gpnum; - rnp->completed = rsp->completed; - rsp->fqs_state = RCU_SIGNAL_INIT; /* force_quiescent_state OK */ - rcu_start_gp_per_cpu(rsp, rnp, rdp); - rcu_preempt_boost_start_gp(rnp); - trace_rcu_grace_period_init(rsp->name, rnp->gpnum, - rnp->level, rnp->grplo, - rnp->grphi, rnp->qsmask); - raw_spin_unlock_irqrestore(&rnp->lock, flags); - return; - } - raw_spin_unlock(&rnp->lock); /* leave irqs disabled. */ - /* Exclude any concurrent CPU-hotplug operations. */ raw_spin_lock(&rsp->onofflock); /* irqs already disabled. */ @@ -1245,53 +1312,115 @@ rcu_check_quiescent_state(struct rcu_state *rsp, struct rcu_data *rdp) /* * Move a dying CPU's RCU callbacks to online CPU's callback list. - * Synchronization is not required because this function executes - * in stop_machine() context. + * Also record a quiescent state for this CPU for the current grace period. + * Synchronization and interrupt disabling are not required because + * this function executes in stop_machine() context. Therefore, cleanup + * operations that might block must be done later from the CPU_DEAD + * notifier. + * + * Note that the outgoing CPU's bit has already been cleared in the + * cpu_online_mask. This allows us to randomly pick a callback + * destination from the bits set in that mask. */ -static void rcu_send_cbs_to_online(struct rcu_state *rsp) +static void rcu_cleanup_dying_cpu(struct rcu_state *rsp) { int i; - /* current DYING CPU is cleared in the cpu_online_mask */ + unsigned long mask; int receive_cpu = cpumask_any(cpu_online_mask); struct rcu_data *rdp = this_cpu_ptr(rsp->rda); struct rcu_data *receive_rdp = per_cpu_ptr(rsp->rda, receive_cpu); + RCU_TRACE(struct rcu_node *rnp = rdp->mynode); /* For dying CPU. */ + + /* First, adjust the counts. */ + if (rdp->nxtlist != NULL) { + receive_rdp->qlen_lazy += rdp->qlen_lazy; + receive_rdp->qlen += rdp->qlen; + rdp->qlen_lazy = 0; + rdp->qlen = 0; + } - if (rdp->nxtlist == NULL) - return; /* irqs disabled, so comparison is stable. */ + /* + * Next, move ready-to-invoke callbacks to be invoked on some + * other CPU. These will not be required to pass through another + * grace period: They are done, regardless of CPU. + */ + if (rdp->nxtlist != NULL && + rdp->nxttail[RCU_DONE_TAIL] != &rdp->nxtlist) { + struct rcu_head *oldhead; + struct rcu_head **oldtail; + struct rcu_head **newtail; + + oldhead = rdp->nxtlist; + oldtail = receive_rdp->nxttail[RCU_DONE_TAIL]; + rdp->nxtlist = *rdp->nxttail[RCU_DONE_TAIL]; + *rdp->nxttail[RCU_DONE_TAIL] = *oldtail; + *receive_rdp->nxttail[RCU_DONE_TAIL] = oldhead; + newtail = rdp->nxttail[RCU_DONE_TAIL]; + for (i = RCU_DONE_TAIL; i < RCU_NEXT_SIZE; i++) { + if (receive_rdp->nxttail[i] == oldtail) + receive_rdp->nxttail[i] = newtail; + if (rdp->nxttail[i] == newtail) + rdp->nxttail[i] = &rdp->nxtlist; + } + } - *receive_rdp->nxttail[RCU_NEXT_TAIL] = rdp->nxtlist; - receive_rdp->nxttail[RCU_NEXT_TAIL] = rdp->nxttail[RCU_NEXT_TAIL]; - receive_rdp->qlen += rdp->qlen; - receive_rdp->n_cbs_adopted += rdp->qlen; - rdp->n_cbs_orphaned += rdp->qlen; + /* + * Finally, put the rest of the callbacks at the end of the list. + * The ones that made it partway through get to start over: We + * cannot assume that grace periods are synchronized across CPUs. + * (We could splice RCU_WAIT_TAIL into RCU_NEXT_READY_TAIL, but + * this does not seem compelling. Not yet, anyway.) + */ + if (rdp->nxtlist != NULL) { + *receive_rdp->nxttail[RCU_NEXT_TAIL] = rdp->nxtlist; + receive_rdp->nxttail[RCU_NEXT_TAIL] = + rdp->nxttail[RCU_NEXT_TAIL]; + receive_rdp->n_cbs_adopted += rdp->qlen; + rdp->n_cbs_orphaned += rdp->qlen; + + rdp->nxtlist = NULL; + for (i = 0; i < RCU_NEXT_SIZE; i++) + rdp->nxttail[i] = &rdp->nxtlist; + } - rdp->nxtlist = NULL; - for (i = 0; i < RCU_NEXT_SIZE; i++) - rdp->nxttail[i] = &rdp->nxtlist; - rdp->qlen = 0; + /* + * Record a quiescent state for the dying CPU. This is safe + * only because we have already cleared out the callbacks. + * (Otherwise, the RCU core might try to schedule the invocation + * of callbacks on this now-offline CPU, which would be bad.) + */ + mask = rdp->grpmask; /* rnp->grplo is constant. */ + trace_rcu_grace_period(rsp->name, + rnp->gpnum + 1 - !!(rnp->qsmask & mask), + "cpuofl"); + rcu_report_qs_rdp(smp_processor_id(), rsp, rdp, rsp->gpnum); + /* Note that rcu_report_qs_rdp() might call trace_rcu_grace_period(). */ } /* - * Remove the outgoing CPU from the bitmasks in the rcu_node hierarchy - * and move all callbacks from the outgoing CPU to the current one. + * The CPU has been completely removed, and some other CPU is reporting + * this fact from process context. Do the remainder of the cleanup. * There can only be one CPU hotplug operation at a time, so no other * CPU can be attempting to update rcu_cpu_kthread_task. */ -static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp) +static void rcu_cleanup_dead_cpu(int cpu, struct rcu_state *rsp) { unsigned long flags; unsigned long mask; int need_report = 0; struct rcu_data *rdp = per_cpu_ptr(rsp->rda, cpu); - struct rcu_node *rnp; + struct rcu_node *rnp = rdp->mynode; /* Outgoing CPU's rnp. */ + /* Adjust any no-longer-needed kthreads. */ rcu_stop_cpu_kthread(cpu); + rcu_node_kthread_setaffinity(rnp, -1); + + /* Remove the dying CPU from the bitmasks in the rcu_node hierarchy. */ /* Exclude any attempts to start a new grace period. */ 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 { raw_spin_lock(&rnp->lock); /* irqs already disabled. */ @@ -1299,20 +1428,11 @@ static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp) if (rnp->qsmaskinit != 0) { if (rnp != rdp->mynode) raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ - else - trace_rcu_grace_period(rsp->name, - rnp->gpnum + 1 - - !!(rnp->qsmask & mask), - "cpuofl"); break; } - if (rnp == rdp->mynode) { - trace_rcu_grace_period(rsp->name, - rnp->gpnum + 1 - - !!(rnp->qsmask & mask), - "cpuofl"); + if (rnp == rdp->mynode) need_report = rcu_preempt_offline_tasks(rsp, rnp, rdp); - } else + else raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ mask = rnp->grpmask; rnp = rnp->parent; @@ -1332,29 +1452,15 @@ static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp) raw_spin_unlock_irqrestore(&rnp->lock, flags); if (need_report & RCU_OFL_TASKS_EXP_GP) rcu_report_exp_rnp(rsp, rnp, true); - rcu_node_kthread_setaffinity(rnp, -1); -} - -/* - * Remove the specified CPU from the RCU hierarchy and move any pending - * callbacks that it might have to the current CPU. This code assumes - * that at least one CPU in the system will remain running at all times. - * Any attempt to offline -all- CPUs is likely to strand RCU callbacks. - */ -static void rcu_offline_cpu(int cpu) -{ - __rcu_offline_cpu(cpu, &rcu_sched_state); - __rcu_offline_cpu(cpu, &rcu_bh_state); - rcu_preempt_offline_cpu(cpu); } #else /* #ifdef CONFIG_HOTPLUG_CPU */ -static void rcu_send_cbs_to_online(struct rcu_state *rsp) +static void rcu_cleanup_dying_cpu(struct rcu_state *rsp) { } -static void rcu_offline_cpu(int cpu) +static void rcu_cleanup_dead_cpu(int cpu, struct rcu_state *rsp) { } @@ -1368,11 +1474,11 @@ static void rcu_do_batch(struct rcu_state *rsp, struct rcu_data *rdp) { unsigned long flags; struct rcu_head *next, *list, **tail; - int bl, count; + int bl, count, count_lazy; /* If no callbacks are ready, just return.*/ if (!cpu_has_callbacks_ready_to_invoke(rdp)) { - trace_rcu_batch_start(rsp->name, 0, 0); + trace_rcu_batch_start(rsp->name, rdp->qlen_lazy, rdp->qlen, 0); trace_rcu_batch_end(rsp->name, 0, !!ACCESS_ONCE(rdp->nxtlist), need_resched(), is_idle_task(current), rcu_is_callbacks_kthread()); @@ -1384,8 +1490,9 @@ static void rcu_do_batch(struct rcu_state *rsp, struct rcu_data *rdp) * races with call_rcu() from interrupt handlers. */ local_irq_save(flags); + WARN_ON_ONCE(cpu_is_offline(smp_processor_id())); bl = rdp->blimit; - trace_rcu_batch_start(rsp->name, rdp->qlen, bl); + trace_rcu_batch_start(rsp->name, rdp->qlen_lazy, rdp->qlen, bl); list = rdp->nxtlist; rdp->nxtlist = *rdp->nxttail[RCU_DONE_TAIL]; *rdp->nxttail[RCU_DONE_TAIL] = NULL; @@ -1396,12 +1503,13 @@ static void rcu_do_batch(struct rcu_state *rsp, struct rcu_data *rdp) local_irq_restore(flags); /* Invoke callbacks. */ - count = 0; + count = count_lazy = 0; while (list) { next = list->next; prefetch(next); debug_rcu_head_unqueue(list); - __rcu_reclaim(rsp->name, list); + if (__rcu_reclaim(rsp->name, list)) + count_lazy++; list = next; /* Stop only if limit reached and CPU has something to do. */ if (++count >= bl && @@ -1416,6 +1524,7 @@ static void rcu_do_batch(struct rcu_state *rsp, struct rcu_data *rdp) rcu_is_callbacks_kthread()); /* Update count, and requeue any remaining callbacks. */ + rdp->qlen_lazy -= count_lazy; rdp->qlen -= count; rdp->n_cbs_invoked += count; if (list != NULL) { @@ -1458,6 +1567,7 @@ static void rcu_do_batch(struct rcu_state *rsp, struct rcu_data *rdp) void rcu_check_callbacks(int cpu, int user) { trace_rcu_utilization("Start scheduler-tick"); + increment_cpu_stall_ticks(); if (user || rcu_is_cpu_rrupt_from_idle()) { /* @@ -1492,8 +1602,6 @@ void rcu_check_callbacks(int cpu, int user) trace_rcu_utilization("End scheduler-tick"); } -#ifdef CONFIG_SMP - /* * Scan the leaf rcu_node structures, processing dyntick state for any that * have not yet encountered a quiescent state, using the function specified. @@ -1616,15 +1724,6 @@ unlock_fqs_ret: trace_rcu_utilization("End fqs"); } -#else /* #ifdef CONFIG_SMP */ - -static void force_quiescent_state(struct rcu_state *rsp, int relaxed) -{ - set_need_resched(); -} - -#endif /* #else #ifdef CONFIG_SMP */ - /* * This does the RCU core processing work for the specified rcu_state * and rcu_data structures. This may be called only from the CPU to @@ -1702,11 +1801,12 @@ static void invoke_rcu_core(void) static void __call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu), - struct rcu_state *rsp) + struct rcu_state *rsp, bool lazy) { unsigned long flags; struct rcu_data *rdp; + WARN_ON_ONCE((unsigned long)head & 0x3); /* Misaligned rcu_head! */ debug_rcu_head_queue(head); head->func = func; head->next = NULL; @@ -1720,18 +1820,21 @@ __call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu), * a quiescent state betweentimes. */ local_irq_save(flags); + WARN_ON_ONCE(cpu_is_offline(smp_processor_id())); rdp = this_cpu_ptr(rsp->rda); /* Add the callback to our list. */ *rdp->nxttail[RCU_NEXT_TAIL] = head; rdp->nxttail[RCU_NEXT_TAIL] = &head->next; rdp->qlen++; + if (lazy) + rdp->qlen_lazy++; if (__is_kfree_rcu_offset((unsigned long)func)) trace_rcu_kfree_callback(rsp->name, head, (unsigned long)func, - rdp->qlen); + rdp->qlen_lazy, rdp->qlen); else - trace_rcu_callback(rsp->name, head, rdp->qlen); + trace_rcu_callback(rsp->name, head, rdp->qlen_lazy, rdp->qlen); /* If interrupts were disabled, don't dive into RCU core. */ if (irqs_disabled_flags(flags)) { @@ -1778,16 +1881,16 @@ __call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu), */ void call_rcu_sched(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) { - __call_rcu(head, func, &rcu_sched_state); + __call_rcu(head, func, &rcu_sched_state, 0); } EXPORT_SYMBOL_GPL(call_rcu_sched); /* - * Queue an RCU for invocation after a quicker grace period. + * Queue an RCU callback for invocation after a quicker grace period. */ void call_rcu_bh(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) { - __call_rcu(head, func, &rcu_bh_state); + __call_rcu(head, func, &rcu_bh_state, 0); } EXPORT_SYMBOL_GPL(call_rcu_bh); @@ -1816,6 +1919,10 @@ EXPORT_SYMBOL_GPL(call_rcu_bh); */ void synchronize_sched(void) { + rcu_lockdep_assert(!lock_is_held(&rcu_bh_lock_map) && + !lock_is_held(&rcu_lock_map) && + !lock_is_held(&rcu_sched_lock_map), + "Illegal synchronize_sched() in RCU-sched read-side critical section"); if (rcu_blocking_is_gp()) return; wait_rcu_gp(call_rcu_sched); @@ -1833,12 +1940,137 @@ EXPORT_SYMBOL_GPL(synchronize_sched); */ void synchronize_rcu_bh(void) { + rcu_lockdep_assert(!lock_is_held(&rcu_bh_lock_map) && + !lock_is_held(&rcu_lock_map) && + !lock_is_held(&rcu_sched_lock_map), + "Illegal synchronize_rcu_bh() in RCU-bh read-side critical section"); if (rcu_blocking_is_gp()) return; wait_rcu_gp(call_rcu_bh); } EXPORT_SYMBOL_GPL(synchronize_rcu_bh); +static atomic_t sync_sched_expedited_started = ATOMIC_INIT(0); +static atomic_t sync_sched_expedited_done = ATOMIC_INIT(0); + +static int synchronize_sched_expedited_cpu_stop(void *data) +{ + /* + * There must be a full memory barrier on each affected CPU + * between the time that try_stop_cpus() is called and the + * time that it returns. + * + * In the current initial implementation of cpu_stop, the + * above condition is already met when the control reaches + * this point and the following smp_mb() is not strictly + * necessary. Do smp_mb() anyway for documentation and + * robustness against future implementation changes. + */ + smp_mb(); /* See above comment block. */ + return 0; +} + +/** + * synchronize_sched_expedited - Brute-force RCU-sched grace period + * + * Wait for an RCU-sched grace period to elapse, but use a "big hammer" + * approach to force the grace period to end quickly. This consumes + * significant time on all CPUs and is unfriendly to real-time workloads, + * so is thus not recommended for any sort of common-case code. In fact, + * if you are using synchronize_sched_expedited() in a loop, please + * restructure your code to batch your updates, and then use a single + * synchronize_sched() instead. + * + * Note that it is illegal to call this function while holding any lock + * that is acquired by a CPU-hotplug notifier. And yes, it is also illegal + * to call this function from a CPU-hotplug notifier. Failing to observe + * these restriction will result in deadlock. + * + * This implementation can be thought of as an application of ticket + * locking to RCU, with sync_sched_expedited_started and + * sync_sched_expedited_done taking on the roles of the halves + * of the ticket-lock word. Each task atomically increments + * sync_sched_expedited_started upon entry, snapshotting the old value, + * then attempts to stop all the CPUs. If this succeeds, then each + * CPU will have executed a context switch, resulting in an RCU-sched + * grace period. We are then done, so we use atomic_cmpxchg() to + * update sync_sched_expedited_done to match our snapshot -- but + * only if someone else has not already advanced past our snapshot. + * + * On the other hand, if try_stop_cpus() fails, we check the value + * of sync_sched_expedited_done. If it has advanced past our + * initial snapshot, then someone else must have forced a grace period + * some time after we took our snapshot. In this case, our work is + * done for us, and we can simply return. Otherwise, we try again, + * but keep our initial snapshot for purposes of checking for someone + * doing our work for us. + * + * If we fail too many times in a row, we fall back to synchronize_sched(). + */ +void synchronize_sched_expedited(void) +{ + int firstsnap, s, snap, trycount = 0; + + /* Note that atomic_inc_return() implies full memory barrier. */ + firstsnap = snap = atomic_inc_return(&sync_sched_expedited_started); + get_online_cpus(); + WARN_ON_ONCE(cpu_is_offline(raw_smp_processor_id())); + + /* + * Each pass through the following loop attempts to force a + * context switch on each CPU. + */ + while (try_stop_cpus(cpu_online_mask, + synchronize_sched_expedited_cpu_stop, + NULL) == -EAGAIN) { + put_online_cpus(); + + /* No joy, try again later. Or just synchronize_sched(). */ + if (trycount++ < 10) + udelay(trycount * num_online_cpus()); + else { + synchronize_sched(); + return; + } + + /* Check to see if someone else did our work for us. */ + s = atomic_read(&sync_sched_expedited_done); + if (UINT_CMP_GE((unsigned)s, (unsigned)firstsnap)) { + smp_mb(); /* ensure test happens before caller kfree */ + return; + } + + /* + * Refetching sync_sched_expedited_started allows later + * callers to piggyback on our grace period. We subtract + * 1 to get the same token that the last incrementer got. + * We retry after they started, so our grace period works + * for them, and they started after our first try, so their + * grace period works for us. + */ + get_online_cpus(); + snap = atomic_read(&sync_sched_expedited_started); + smp_mb(); /* ensure read is before try_stop_cpus(). */ + } + + /* + * Everyone up to our most recent fetch is covered by our grace + * period. Update the counter, but only if our work is still + * relevant -- which it won't be if someone who started later + * than we did beat us to the punch. + */ + do { + s = atomic_read(&sync_sched_expedited_done); + if (UINT_CMP_GE((unsigned)s, (unsigned)snap)) { + smp_mb(); /* ensure test happens before caller kfree */ + break; + } + } while (atomic_cmpxchg(&sync_sched_expedited_done, s, snap) != s); + + put_online_cpus(); +} +EXPORT_SYMBOL_GPL(synchronize_sched_expedited); + /* * Check to see if there is any immediate RCU-related work to be done * by the current CPU, for the specified type of RCU, returning 1 if so. @@ -1932,7 +2164,7 @@ static int rcu_cpu_has_callbacks(int cpu) /* RCU callbacks either ready or pending? */ return per_cpu(rcu_sched_data, cpu).nxtlist || per_cpu(rcu_bh_data, cpu).nxtlist || - rcu_preempt_needs_cpu(cpu); + rcu_preempt_cpu_has_callbacks(cpu); } static DEFINE_PER_CPU(struct rcu_head, rcu_barrier_head) = {NULL}; @@ -2027,9 +2259,10 @@ rcu_boot_init_percpu_data(int cpu, struct rcu_state *rsp) rdp->nxtlist = NULL; for (i = 0; i < RCU_NEXT_SIZE; i++) rdp->nxttail[i] = &rdp->nxtlist; + rdp->qlen_lazy = 0; rdp->qlen = 0; rdp->dynticks = &per_cpu(rcu_dynticks, cpu); - WARN_ON_ONCE(rdp->dynticks->dynticks_nesting != DYNTICK_TASK_NESTING); + WARN_ON_ONCE(rdp->dynticks->dynticks_nesting != DYNTICK_TASK_EXIT_IDLE); WARN_ON_ONCE(atomic_read(&rdp->dynticks->dynticks) != 1); rdp->cpu = cpu; rdp->rsp = rsp; @@ -2057,7 +2290,7 @@ rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptible) rdp->qlen_last_fqs_check = 0; rdp->n_force_qs_snap = rsp->n_force_qs; rdp->blimit = blimit; - rdp->dynticks->dynticks_nesting = DYNTICK_TASK_NESTING; + rdp->dynticks->dynticks_nesting = DYNTICK_TASK_EXIT_IDLE; atomic_set(&rdp->dynticks->dynticks, (atomic_read(&rdp->dynticks->dynticks) & ~0x1) + 1); rcu_prepare_for_idle_init(cpu); @@ -2139,16 +2372,18 @@ static int __cpuinit rcu_cpu_notify(struct notifier_block *self, * touch any data without introducing corruption. We send the * dying CPU's callbacks to an arbitrarily chosen online CPU. */ - rcu_send_cbs_to_online(&rcu_bh_state); - rcu_send_cbs_to_online(&rcu_sched_state); - rcu_preempt_send_cbs_to_online(); + rcu_cleanup_dying_cpu(&rcu_bh_state); + rcu_cleanup_dying_cpu(&rcu_sched_state); + rcu_preempt_cleanup_dying_cpu(); rcu_cleanup_after_idle(cpu); break; case CPU_DEAD: case CPU_DEAD_FROZEN: case CPU_UP_CANCELED: case CPU_UP_CANCELED_FROZEN: - rcu_offline_cpu(cpu); + rcu_cleanup_dead_cpu(cpu, &rcu_bh_state); + rcu_cleanup_dead_cpu(cpu, &rcu_sched_state); + rcu_preempt_cleanup_dead_cpu(cpu); break; default: break; diff --git a/kernel/rcutree.h b/kernel/rcutree.h index fddff92d6676..cdd1be0a4072 100644 --- a/kernel/rcutree.h +++ b/kernel/rcutree.h @@ -239,6 +239,12 @@ struct rcu_data { bool preemptible; /* Preemptible RCU? */ struct rcu_node *mynode; /* This CPU's leaf of hierarchy */ unsigned long grpmask; /* Mask to apply to leaf qsmask. */ +#ifdef CONFIG_RCU_CPU_STALL_INFO + unsigned long ticks_this_gp; /* The number of scheduling-clock */ + /* ticks this CPU has handled */ + /* during and after the last grace */ + /* period it is aware of. */ +#endif /* #ifdef CONFIG_RCU_CPU_STALL_INFO */ /* 2) batch handling */ /* @@ -265,7 +271,8 @@ struct rcu_data { */ struct rcu_head *nxtlist; struct rcu_head **nxttail[RCU_NEXT_SIZE]; - long qlen; /* # of queued callbacks */ + long qlen_lazy; /* # of lazy queued callbacks */ + long qlen; /* # of queued callbacks, incl lazy */ long qlen_last_fqs_check; /* qlen at last check for QS forcing */ unsigned long n_cbs_invoked; /* count of RCU cbs invoked. */ @@ -282,7 +289,6 @@ struct rcu_data { /* 4) reasons this CPU needed to be kicked by force_quiescent_state */ unsigned long dynticks_fqs; /* Kicked due to dynticks idle. */ unsigned long offline_fqs; /* Kicked due to being offline. */ - unsigned long resched_ipi; /* Sent a resched IPI. */ /* 5) __rcu_pending() statistics. */ unsigned long n_rcu_pending; /* rcu_pending() calls since boot. */ @@ -313,12 +319,6 @@ struct rcu_data { #else #define RCU_STALL_DELAY_DELTA 0 #endif - -#define RCU_SECONDS_TILL_STALL_CHECK (CONFIG_RCU_CPU_STALL_TIMEOUT * HZ + \ - RCU_STALL_DELAY_DELTA) - /* for rsp->jiffies_stall */ -#define RCU_SECONDS_TILL_STALL_RECHECK (3 * RCU_SECONDS_TILL_STALL_CHECK + 30) - /* for rsp->jiffies_stall */ #define RCU_STALL_RAT_DELAY 2 /* Allow other CPUs time */ /* to take at least one */ /* scheduling clock irq */ @@ -438,8 +438,8 @@ static void rcu_preempt_check_blocked_tasks(struct rcu_node *rnp); static int rcu_preempt_offline_tasks(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp); -static void rcu_preempt_offline_cpu(int cpu); #endif /* #ifdef CONFIG_HOTPLUG_CPU */ +static void rcu_preempt_cleanup_dead_cpu(int cpu); static void rcu_preempt_check_callbacks(int cpu); static void rcu_preempt_process_callbacks(void); void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu)); @@ -448,9 +448,9 @@ static void rcu_report_exp_rnp(struct rcu_state *rsp, struct rcu_node *rnp, bool wake); #endif /* #if defined(CONFIG_HOTPLUG_CPU) || defined(CONFIG_TREE_PREEMPT_RCU) */ static int rcu_preempt_pending(int cpu); -static int rcu_preempt_needs_cpu(int cpu); +static int rcu_preempt_cpu_has_callbacks(int cpu); static void __cpuinit rcu_preempt_init_percpu_data(int cpu); -static void rcu_preempt_send_cbs_to_online(void); +static void rcu_preempt_cleanup_dying_cpu(void); static void __init __rcu_init_preempt(void); static void rcu_initiate_boost(struct rcu_node *rnp, unsigned long flags); static void rcu_preempt_boost_start_gp(struct rcu_node *rnp); @@ -471,5 +471,10 @@ static void __cpuinit rcu_prepare_kthreads(int cpu); static void rcu_prepare_for_idle_init(int cpu); static void rcu_cleanup_after_idle(int cpu); static void rcu_prepare_for_idle(int cpu); +static void print_cpu_stall_info_begin(void); +static void print_cpu_stall_info(struct rcu_state *rsp, int cpu); +static void print_cpu_stall_info_end(void); +static void zero_cpu_stall_ticks(struct rcu_data *rdp); +static void increment_cpu_stall_ticks(void); #endif /* #ifndef RCU_TREE_NONCORE */ diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 8bb35d73e1f9..c023464816be 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -25,7 +25,6 @@ */ #include <linux/delay.h> -#include <linux/stop_machine.h> #define RCU_KTHREAD_PRIO 1 @@ -63,7 +62,10 @@ static void __init rcu_bootup_announce_oddness(void) printk(KERN_INFO "\tRCU torture testing starts during boot.\n"); #endif #if defined(CONFIG_TREE_PREEMPT_RCU) && !defined(CONFIG_RCU_CPU_STALL_VERBOSE) - printk(KERN_INFO "\tVerbose stalled-CPUs detection is disabled.\n"); + printk(KERN_INFO "\tDump stacks of tasks blocking RCU-preempt GP.\n"); +#endif +#if defined(CONFIG_RCU_CPU_STALL_INFO) + printk(KERN_INFO "\tAdditional per-CPU info printed with stalls.\n"); #endif #if NUM_RCU_LVL_4 != 0 printk(KERN_INFO "\tExperimental four-level hierarchy is enabled.\n"); @@ -490,6 +492,31 @@ static void rcu_print_detail_task_stall(struct rcu_state *rsp) #endif /* #else #ifdef CONFIG_RCU_CPU_STALL_VERBOSE */ +#ifdef CONFIG_RCU_CPU_STALL_INFO + +static void rcu_print_task_stall_begin(struct rcu_node *rnp) +{ + printk(KERN_ERR "\tTasks blocked on level-%d rcu_node (CPUs %d-%d):", + rnp->level, rnp->grplo, rnp->grphi); +} + +static void rcu_print_task_stall_end(void) +{ + printk(KERN_CONT "\n"); +} + +#else /* #ifdef CONFIG_RCU_CPU_STALL_INFO */ + +static void rcu_print_task_stall_begin(struct rcu_node *rnp) +{ +} + +static void rcu_print_task_stall_end(void) +{ +} + +#endif /* #else #ifdef CONFIG_RCU_CPU_STALL_INFO */ + /* * Scan the current list of tasks blocked within RCU read-side critical * sections, printing out the tid of each. @@ -501,12 +528,14 @@ static int rcu_print_task_stall(struct rcu_node *rnp) if (!rcu_preempt_blocked_readers_cgp(rnp)) return 0; + rcu_print_task_stall_begin(rnp); t = list_entry(rnp->gp_tasks, struct task_struct, rcu_node_entry); list_for_each_entry_continue(t, &rnp->blkd_tasks, rcu_node_entry) { - printk(" P%d", t->pid); + printk(KERN_CONT " P%d", t->pid); ndetected++; } + rcu_print_task_stall_end(); return ndetected; } @@ -581,7 +610,7 @@ static int rcu_preempt_offline_tasks(struct rcu_state *rsp, * absolutely necessary, but this is a good performance/complexity * tradeoff. */ - if (rcu_preempt_blocked_readers_cgp(rnp)) + if (rcu_preempt_blocked_readers_cgp(rnp) && rnp->qsmask == 0) retval |= RCU_OFL_TASKS_NORM_GP; if (rcu_preempted_readers_exp(rnp)) retval |= RCU_OFL_TASKS_EXP_GP; @@ -618,16 +647,16 @@ static int rcu_preempt_offline_tasks(struct rcu_state *rsp, return retval; } +#endif /* #ifdef CONFIG_HOTPLUG_CPU */ + /* * Do CPU-offline processing for preemptible RCU. */ -static void rcu_preempt_offline_cpu(int cpu) +static void rcu_preempt_cleanup_dead_cpu(int cpu) { - __rcu_offline_cpu(cpu, &rcu_preempt_state); + rcu_cleanup_dead_cpu(cpu, &rcu_preempt_state); } -#endif /* #ifdef CONFIG_HOTPLUG_CPU */ - /* * Check for a quiescent state from the current CPU. When a task blocks, * the task is recorded in the corresponding CPU's rcu_node structure, @@ -671,10 +700,24 @@ static void rcu_preempt_do_callbacks(void) */ void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) { - __call_rcu(head, func, &rcu_preempt_state); + __call_rcu(head, func, &rcu_preempt_state, 0); } EXPORT_SYMBOL_GPL(call_rcu); +/* + * Queue an RCU callback for lazy invocation after a grace period. + * This will likely be later named something like "call_rcu_lazy()", + * but this change will require some way of tagging the lazy RCU + * callbacks in the list of pending callbacks. Until then, this + * function may only be called from __kfree_rcu(). + */ +void kfree_call_rcu(struct rcu_head *head, + void (*func)(struct rcu_head *rcu)) +{ + __call_rcu(head, func, &rcu_preempt_state, 1); +} +EXPORT_SYMBOL_GPL(kfree_call_rcu); + /** * synchronize_rcu - wait until a grace period has elapsed. * @@ -688,6 +731,10 @@ EXPORT_SYMBOL_GPL(call_rcu); */ void synchronize_rcu(void) { + rcu_lockdep_assert(!lock_is_held(&rcu_bh_lock_map) && + !lock_is_held(&rcu_lock_map) && + !lock_is_held(&rcu_sched_lock_map), + "Illegal synchronize_rcu() in RCU read-side critical section"); if (!rcu_scheduler_active) return; wait_rcu_gp(call_rcu); @@ -788,10 +835,22 @@ sync_rcu_preempt_exp_init(struct rcu_state *rsp, struct rcu_node *rnp) rcu_report_exp_rnp(rsp, rnp, false); /* Don't wake self. */ } -/* - * Wait for an rcu-preempt grace period, but expedite it. The basic idea - * is to invoke synchronize_sched_expedited() to push all the tasks to - * the ->blkd_tasks lists and wait for this list to drain. +/** + * synchronize_rcu_expedited - Brute-force RCU grace period + * + * Wait for an RCU-preempt grace period, but expedite it. The basic + * idea is to invoke synchronize_sched_expedited() to push all the tasks to + * the ->blkd_tasks lists and wait for this list to drain. This consumes + * significant time on all CPUs and is unfriendly to real-time workloads, + * so is thus not recommended for any sort of common-case code. + * In fact, if you are using synchronize_rcu_expedited() in a loop, + * please restructure your code to batch your updates, and then Use a + * single synchronize_rcu() instead. + * + * Note that it is illegal to call this function while holding any lock + * that is acquired by a CPU-hotplug notifier. And yes, it is also illegal + * to call this function from a CPU-hotplug notifier. Failing to observe + * these restriction will result in deadlock. */ void synchronize_rcu_expedited(void) { @@ -869,9 +928,9 @@ static int rcu_preempt_pending(int cpu) } /* - * Does preemptible RCU need the CPU to stay out of dynticks mode? + * Does preemptible RCU have callbacks on this CPU? */ -static int rcu_preempt_needs_cpu(int cpu) +static int rcu_preempt_cpu_has_callbacks(int cpu) { return !!per_cpu(rcu_preempt_data, cpu).nxtlist; } @@ -894,11 +953,12 @@ static void __cpuinit rcu_preempt_init_percpu_data(int cpu) } /* - * Move preemptible RCU's callbacks from dying CPU to other online CPU. + * Move preemptible RCU's callbacks from dying CPU to other online CPU + * and record a quiescent state. */ -static void rcu_preempt_send_cbs_to_online(void) +static void rcu_preempt_cleanup_dying_cpu(void) { - rcu_send_cbs_to_online(&rcu_preempt_state); + rcu_cleanup_dying_cpu(&rcu_preempt_state); } /* @@ -1034,16 +1094,16 @@ static int rcu_preempt_offline_tasks(struct rcu_state *rsp, return 0; } +#endif /* #ifdef CONFIG_HOTPLUG_CPU */ + /* * Because preemptible RCU does not exist, it never needs CPU-offline * processing. */ -static void rcu_preempt_offline_cpu(int cpu) +static void rcu_preempt_cleanup_dead_cpu(int cpu) { } -#endif /* #ifdef CONFIG_HOTPLUG_CPU */ - /* * Because preemptible RCU does not exist, it never has any callbacks * to check. @@ -1061,6 +1121,22 @@ static void rcu_preempt_process_callbacks(void) } /* + * Queue an RCU callback for lazy invocation after a grace period. + * This will likely be later named something like "call_rcu_lazy()", + * but this change will require some way of tagging the lazy RCU + * callbacks in the list of pending callbacks. Until then, this + * function may only be called from __kfree_rcu(). + * + * Because there is no preemptible RCU, we use RCU-sched instead. + */ +void kfree_call_rcu(struct rcu_head *head, + void (*func)(struct rcu_head *rcu)) +{ + __call_rcu(head, func, &rcu_sched_state, 1); +} +EXPORT_SYMBOL_GPL(kfree_call_rcu); + +/* * Wait for an rcu-preempt grace period, but make it happen quickly. * But because preemptible RCU does not exist, map to rcu-sched. */ @@ -1093,9 +1169,9 @@ static int rcu_preempt_pending(int cpu) } /* - * Because preemptible RCU does not exist, it never needs any CPU. + * Because preemptible RCU does not exist, it never has callbacks */ -static int rcu_preempt_needs_cpu(int cpu) +static int rcu_preempt_cpu_has_callbacks(int cpu) { return 0; } @@ -1119,9 +1195,9 @@ static void __cpuinit rcu_preempt_init_percpu_data(int cpu) } /* - * Because there is no preemptible RCU, there are no callbacks to move. + * Because there is no preemptible RCU, there is no cleanup to do. */ -static void rcu_preempt_send_cbs_to_online(void) +static void rcu_preempt_cleanup_dying_cpu(void) { } @@ -1823,132 +1899,6 @@ static void __cpuinit rcu_prepare_kthreads(int cpu) #endif /* #else #ifdef CONFIG_RCU_BOOST */ -#ifndef CONFIG_SMP - -void synchronize_sched_expedited(void) -{ - cond_resched(); -} -EXPORT_SYMBOL_GPL(synchronize_sched_expedited); - -#else /* #ifndef CONFIG_SMP */ - -static atomic_t sync_sched_expedited_started = ATOMIC_INIT(0); -static atomic_t sync_sched_expedited_done = ATOMIC_INIT(0); - -static int synchronize_sched_expedited_cpu_stop(void *data) -{ - /* - * There must be a full memory barrier on each affected CPU - * between the time that try_stop_cpus() is called and the - * time that it returns. - * - * In the current initial implementation of cpu_stop, the - * above condition is already met when the control reaches - * this point and the following smp_mb() is not strictly - * necessary. Do smp_mb() anyway for documentation and - * robustness against future implementation changes. - */ - smp_mb(); /* See above comment block. */ - return 0; -} - -/* - * Wait for an rcu-sched grace period to elapse, but use "big hammer" - * approach to force grace period to end quickly. This consumes - * significant time on all CPUs, and is thus not recommended for - * any sort of common-case code. - * - * Note that it is illegal to call this function while holding any - * lock that is acquired by a CPU-hotplug notifier. Failing to - * observe this restriction will result in deadlock. - * - * This implementation can be thought of as an application of ticket - * locking to RCU, with sync_sched_expedited_started and - * sync_sched_expedited_done taking on the roles of the halves - * of the ticket-lock word. Each task atomically increments - * sync_sched_expedited_started upon entry, snapshotting the old value, - * then attempts to stop all the CPUs. If this succeeds, then each - * CPU will have executed a context switch, resulting in an RCU-sched - * grace period. We are then done, so we use atomic_cmpxchg() to - * update sync_sched_expedited_done to match our snapshot -- but - * only if someone else has not already advanced past our snapshot. - * - * On the other hand, if try_stop_cpus() fails, we check the value - * of sync_sched_expedited_done. If it has advanced past our - * initial snapshot, then someone else must have forced a grace period - * some time after we took our snapshot. In this case, our work is - * done for us, and we can simply return. Otherwise, we try again, - * but keep our initial snapshot for purposes of checking for someone - * doing our work for us. - * - * If we fail too many times in a row, we fall back to synchronize_sched(). - */ -void synchronize_sched_expedited(void) -{ - int firstsnap, s, snap, trycount = 0; - - /* Note that atomic_inc_return() implies full memory barrier. */ - firstsnap = snap = atomic_inc_return(&sync_sched_expedited_started); - get_online_cpus(); - - /* - * Each pass through the following loop attempts to force a - * context switch on each CPU. - */ - while (try_stop_cpus(cpu_online_mask, - synchronize_sched_expedited_cpu_stop, - NULL) == -EAGAIN) { - put_online_cpus(); - - /* No joy, try again later. Or just synchronize_sched(). */ - if (trycount++ < 10) - udelay(trycount * num_online_cpus()); - else { - synchronize_sched(); - return; - } - - /* Check to see if someone else did our work for us. */ - s = atomic_read(&sync_sched_expedited_done); - if (UINT_CMP_GE((unsigned)s, (unsigned)firstsnap)) { - smp_mb(); /* ensure test happens before caller kfree */ - return; - } - - /* - * Refetching sync_sched_expedited_started allows later - * callers to piggyback on our grace period. We subtract - * 1 to get the same token that the last incrementer got. - * We retry after they started, so our grace period works - * for them, and they started after our first try, so their - * grace period works for us. - */ - get_online_cpus(); - snap = atomic_read(&sync_sched_expedited_started); - smp_mb(); /* ensure read is before try_stop_cpus(). */ - } - - /* - * Everyone up to our most recent fetch is covered by our grace - * period. Update the counter, but only if our work is still - * relevant -- which it won't be if someone who started later - * than we did beat us to the punch. - */ - do { - s = atomic_read(&sync_sched_expedited_done); - if (UINT_CMP_GE((unsigned)s, (unsigned)snap)) { - smp_mb(); /* ensure test happens before caller kfree */ - break; - } - } while (atomic_cmpxchg(&sync_sched_expedited_done, s, snap) != s); - - put_online_cpus(); -} -EXPORT_SYMBOL_GPL(synchronize_sched_expedited); - -#endif /* #else #ifndef CONFIG_SMP */ - #if !defined(CONFIG_RCU_FAST_NO_HZ) /* @@ -1981,7 +1931,7 @@ static void rcu_cleanup_after_idle(int cpu) } /* - * Do the idle-entry grace-period work, which, because CONFIG_RCU_FAST_NO_HZ=y, + * Do the idle-entry grace-period work, which, because CONFIG_RCU_FAST_NO_HZ=n, * is nothing. */ static void rcu_prepare_for_idle(int cpu) @@ -2015,6 +1965,9 @@ static void rcu_prepare_for_idle(int cpu) * number, be warned: Setting RCU_IDLE_GP_DELAY too high can hang your * system. And if you are -that- concerned about energy efficiency, * just power the system down and be done with it! + * RCU_IDLE_LAZY_GP_DELAY gives the number of jiffies that a CPU is + * permitted to sleep in dyntick-idle mode with only lazy RCU + * callbacks pending. Setting this too high can OOM your system. * * The values below work well in practice. If future workloads require * adjustment, they can be converted into kernel config parameters, though @@ -2023,11 +1976,13 @@ static void rcu_prepare_for_idle(int cpu) #define RCU_IDLE_FLUSHES 5 /* Number of dyntick-idle tries. */ #define RCU_IDLE_OPT_FLUSHES 3 /* Optional dyntick-idle tries. */ #define RCU_IDLE_GP_DELAY 6 /* Roughly one grace period. */ +#define RCU_IDLE_LAZY_GP_DELAY (6 * HZ) /* Roughly six seconds. */ static DEFINE_PER_CPU(int, rcu_dyntick_drain); static DEFINE_PER_CPU(unsigned long, rcu_dyntick_holdoff); static DEFINE_PER_CPU(struct hrtimer, rcu_idle_gp_timer); -static ktime_t rcu_idle_gp_wait; +static ktime_t rcu_idle_gp_wait; /* If some non-lazy callbacks. */ +static ktime_t rcu_idle_lazy_gp_wait; /* If only lazy callbacks. */ /* * Allow the CPU to enter dyntick-idle mode if either: (1) There are no @@ -2048,6 +2003,48 @@ int rcu_needs_cpu(int cpu) } /* + * Does the specified flavor of RCU have non-lazy callbacks pending on + * the specified CPU? Both RCU flavor and CPU are specified by the + * rcu_data structure. + */ +static bool __rcu_cpu_has_nonlazy_callbacks(struct rcu_data *rdp) +{ + return rdp->qlen != rdp->qlen_lazy; +} + +#ifdef CONFIG_TREE_PREEMPT_RCU + +/* + * Are there non-lazy RCU-preempt callbacks? (There cannot be if there + * is no RCU-preempt in the kernel.) + */ +static bool rcu_preempt_cpu_has_nonlazy_callbacks(int cpu) +{ + struct rcu_data *rdp = &per_cpu(rcu_preempt_data, cpu); + + return __rcu_cpu_has_nonlazy_callbacks(rdp); +} + +#else /* #ifdef CONFIG_TREE_PREEMPT_RCU */ + +static bool rcu_preempt_cpu_has_nonlazy_callbacks(int cpu) +{ + return 0; +} + +#endif /* else #ifdef CONFIG_TREE_PREEMPT_RCU */ + +/* + * Does any flavor of RCU have non-lazy callbacks on the specified CPU? + */ +static bool rcu_cpu_has_nonlazy_callbacks(int cpu) +{ + return __rcu_cpu_has_nonlazy_callbacks(&per_cpu(rcu_sched_data, cpu)) || + __rcu_cpu_has_nonlazy_callbacks(&per_cpu(rcu_bh_data, cpu)) || + rcu_preempt_cpu_has_nonlazy_callbacks(cpu); +} + +/* * Timer handler used to force CPU to start pushing its remaining RCU * callbacks in the case where it entered dyntick-idle mode with callbacks * pending. The hander doesn't really need to do anything because the @@ -2074,6 +2071,8 @@ static void rcu_prepare_for_idle_init(int cpu) unsigned int upj = jiffies_to_usecs(RCU_IDLE_GP_DELAY); rcu_idle_gp_wait = ns_to_ktime(upj * (u64)1000); + upj = jiffies_to_usecs(RCU_IDLE_LAZY_GP_DELAY); + rcu_idle_lazy_gp_wait = ns_to_ktime(upj * (u64)1000); firsttime = 0; } } @@ -2109,10 +2108,6 @@ static void rcu_cleanup_after_idle(int cpu) */ static void rcu_prepare_for_idle(int cpu) { - unsigned long flags; - - local_irq_save(flags); - /* * If there are no callbacks on this CPU, enter dyntick-idle mode. * Also reset state to avoid prejudicing later attempts. @@ -2120,7 +2115,6 @@ static void rcu_prepare_for_idle(int cpu) if (!rcu_cpu_has_callbacks(cpu)) { per_cpu(rcu_dyntick_holdoff, cpu) = jiffies - 1; per_cpu(rcu_dyntick_drain, cpu) = 0; - local_irq_restore(flags); trace_rcu_prep_idle("No callbacks"); return; } @@ -2130,7 +2124,6 @@ static void rcu_prepare_for_idle(int cpu) * refrained from disabling the scheduling-clock tick. */ if (per_cpu(rcu_dyntick_holdoff, cpu) == jiffies) { - local_irq_restore(flags); trace_rcu_prep_idle("In holdoff"); return; } @@ -2140,18 +2133,22 @@ static void rcu_prepare_for_idle(int cpu) /* First time through, initialize the counter. */ per_cpu(rcu_dyntick_drain, cpu) = RCU_IDLE_FLUSHES; } else if (per_cpu(rcu_dyntick_drain, cpu) <= RCU_IDLE_OPT_FLUSHES && - !rcu_pending(cpu)) { + !rcu_pending(cpu) && + !local_softirq_pending()) { /* Can we go dyntick-idle despite still having callbacks? */ trace_rcu_prep_idle("Dyntick with callbacks"); per_cpu(rcu_dyntick_drain, cpu) = 0; - per_cpu(rcu_dyntick_holdoff, cpu) = jiffies - 1; - hrtimer_start(&per_cpu(rcu_idle_gp_timer, cpu), - rcu_idle_gp_wait, HRTIMER_MODE_REL); + per_cpu(rcu_dyntick_holdoff, cpu) = jiffies; + if (rcu_cpu_has_nonlazy_callbacks(cpu)) + hrtimer_start(&per_cpu(rcu_idle_gp_timer, cpu), + rcu_idle_gp_wait, HRTIMER_MODE_REL); + else + hrtimer_start(&per_cpu(rcu_idle_gp_timer, cpu), + rcu_idle_lazy_gp_wait, HRTIMER_MODE_REL); return; /* Nothing more to do immediately. */ } 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; - local_irq_restore(flags); trace_rcu_prep_idle("Begin holdoff"); invoke_rcu_core(); /* Force the CPU out of dyntick-idle. */ return; @@ -2163,23 +2160,17 @@ static void rcu_prepare_for_idle(int cpu) */ #ifdef CONFIG_TREE_PREEMPT_RCU if (per_cpu(rcu_preempt_data, cpu).nxtlist) { - local_irq_restore(flags); rcu_preempt_qs(cpu); force_quiescent_state(&rcu_preempt_state, 0); - local_irq_save(flags); } #endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */ if (per_cpu(rcu_sched_data, cpu).nxtlist) { - local_irq_restore(flags); rcu_sched_qs(cpu); force_quiescent_state(&rcu_sched_state, 0); - local_irq_save(flags); } if (per_cpu(rcu_bh_data, cpu).nxtlist) { - local_irq_restore(flags); rcu_bh_qs(cpu); force_quiescent_state(&rcu_bh_state, 0); - local_irq_save(flags); } /* @@ -2187,13 +2178,124 @@ static void rcu_prepare_for_idle(int cpu) * So try forcing the callbacks through the grace period. */ if (rcu_cpu_has_callbacks(cpu)) { - local_irq_restore(flags); trace_rcu_prep_idle("More callbacks"); invoke_rcu_core(); - } else { - local_irq_restore(flags); + } else trace_rcu_prep_idle("Callbacks drained"); - } } #endif /* #else #if !defined(CONFIG_RCU_FAST_NO_HZ) */ + +#ifdef CONFIG_RCU_CPU_STALL_INFO + +#ifdef CONFIG_RCU_FAST_NO_HZ + +static void print_cpu_stall_fast_no_hz(char *cp, int cpu) +{ + struct hrtimer *hrtp = &per_cpu(rcu_idle_gp_timer, cpu); + + sprintf(cp, "drain=%d %c timer=%lld", + per_cpu(rcu_dyntick_drain, cpu), + per_cpu(rcu_dyntick_holdoff, cpu) == jiffies ? 'H' : '.', + hrtimer_active(hrtp) + ? ktime_to_us(hrtimer_get_remaining(hrtp)) + : -1); +} + +#else /* #ifdef CONFIG_RCU_FAST_NO_HZ */ + +static void print_cpu_stall_fast_no_hz(char *cp, int cpu) +{ +} + +#endif /* #else #ifdef CONFIG_RCU_FAST_NO_HZ */ + +/* Initiate the stall-info list. */ +static void print_cpu_stall_info_begin(void) +{ + printk(KERN_CONT "\n"); +} + +/* + * Print out diagnostic information for the specified stalled CPU. + * + * If the specified CPU is aware of the current RCU grace period + * (flavor specified by rsp), then print the number of scheduling + * clock interrupts the CPU has taken during the time that it has + * been aware. Otherwise, print the number of RCU grace periods + * that this CPU is ignorant of, for example, "1" if the CPU was + * aware of the previous grace period. + * + * Also print out idle and (if CONFIG_RCU_FAST_NO_HZ) idle-entry info. + */ +static void print_cpu_stall_info(struct rcu_state *rsp, int cpu) +{ + char fast_no_hz[72]; + struct rcu_data *rdp = per_cpu_ptr(rsp->rda, cpu); + struct rcu_dynticks *rdtp = rdp->dynticks; + char *ticks_title; + unsigned long ticks_value; + + if (rsp->gpnum == rdp->gpnum) { + ticks_title = "ticks this GP"; + ticks_value = rdp->ticks_this_gp; + } else { + ticks_title = "GPs behind"; + ticks_value = rsp->gpnum - rdp->gpnum; + } + print_cpu_stall_fast_no_hz(fast_no_hz, cpu); + printk(KERN_ERR "\t%d: (%lu %s) idle=%03x/%llx/%d %s\n", + cpu, ticks_value, ticks_title, + atomic_read(&rdtp->dynticks) & 0xfff, + rdtp->dynticks_nesting, rdtp->dynticks_nmi_nesting, + fast_no_hz); +} + +/* Terminate the stall-info list. */ +static void print_cpu_stall_info_end(void) +{ + printk(KERN_ERR "\t"); +} + +/* Zero ->ticks_this_gp for all flavors of RCU. */ +static void zero_cpu_stall_ticks(struct rcu_data *rdp) +{ + rdp->ticks_this_gp = 0; +} + +/* Increment ->ticks_this_gp for all flavors of RCU. */ +static void increment_cpu_stall_ticks(void) +{ + __get_cpu_var(rcu_sched_data).ticks_this_gp++; + __get_cpu_var(rcu_bh_data).ticks_this_gp++; +#ifdef CONFIG_TREE_PREEMPT_RCU + __get_cpu_var(rcu_preempt_data).ticks_this_gp++; +#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */ +} + +#else /* #ifdef CONFIG_RCU_CPU_STALL_INFO */ + +static void print_cpu_stall_info_begin(void) +{ + printk(KERN_CONT " {"); +} + +static void print_cpu_stall_info(struct rcu_state *rsp, int cpu) +{ + printk(KERN_CONT " %d", cpu); +} + +static void print_cpu_stall_info_end(void) +{ + printk(KERN_CONT "} "); +} + +static void zero_cpu_stall_ticks(struct rcu_data *rdp) +{ +} + +static void increment_cpu_stall_ticks(void) +{ +} + +#endif /* #else #ifdef CONFIG_RCU_CPU_STALL_INFO */ diff --git a/kernel/rcutree_trace.c b/kernel/rcutree_trace.c index 654cfe67f0d1..ed459edeff43 100644 --- a/kernel/rcutree_trace.c +++ b/kernel/rcutree_trace.c @@ -72,9 +72,9 @@ static void print_one_rcu_data(struct seq_file *m, struct rcu_data *rdp) rdp->dynticks->dynticks_nesting, rdp->dynticks->dynticks_nmi_nesting, rdp->dynticks_fqs); - seq_printf(m, " of=%lu ri=%lu", rdp->offline_fqs, rdp->resched_ipi); - seq_printf(m, " ql=%ld qs=%c%c%c%c", - rdp->qlen, + seq_printf(m, " of=%lu", rdp->offline_fqs); + seq_printf(m, " ql=%ld/%ld qs=%c%c%c%c", + rdp->qlen_lazy, rdp->qlen, ".N"[rdp->nxttail[RCU_NEXT_READY_TAIL] != rdp->nxttail[RCU_NEXT_TAIL]], ".R"[rdp->nxttail[RCU_WAIT_TAIL] != @@ -144,8 +144,8 @@ static void print_one_rcu_data_csv(struct seq_file *m, struct rcu_data *rdp) rdp->dynticks->dynticks_nesting, rdp->dynticks->dynticks_nmi_nesting, rdp->dynticks_fqs); - seq_printf(m, ",%lu,%lu", rdp->offline_fqs, rdp->resched_ipi); - seq_printf(m, ",%ld,\"%c%c%c%c\"", rdp->qlen, + seq_printf(m, ",%lu", rdp->offline_fqs); + seq_printf(m, ",%ld,%ld,\"%c%c%c%c\"", rdp->qlen_lazy, rdp->qlen, ".N"[rdp->nxttail[RCU_NEXT_READY_TAIL] != rdp->nxttail[RCU_NEXT_TAIL]], ".R"[rdp->nxttail[RCU_WAIT_TAIL] != @@ -168,7 +168,7 @@ static int show_rcudata_csv(struct seq_file *m, void *unused) { seq_puts(m, "\"CPU\",\"Online?\",\"c\",\"g\",\"pq\",\"pgp\",\"pq\","); seq_puts(m, "\"dt\",\"dt nesting\",\"dt NMI nesting\",\"df\","); - seq_puts(m, "\"of\",\"ri\",\"ql\",\"qs\""); + seq_puts(m, "\"of\",\"qll\",\"ql\",\"qs\""); #ifdef CONFIG_RCU_BOOST seq_puts(m, "\"kt\",\"ktl\""); #endif /* #ifdef CONFIG_RCU_BOOST */ diff --git a/kernel/softirq.c b/kernel/softirq.c index 06d40993594a..8afc6a8d4d7c 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -310,31 +310,21 @@ void irq_enter(void) __irq_enter(); } -#ifdef __ARCH_IRQ_EXIT_IRQS_DISABLED static inline void invoke_softirq(void) { - if (!force_irqthreads) + if (!force_irqthreads) { +#ifdef __ARCH_IRQ_EXIT_IRQS_DISABLED __do_softirq(); - else { - __local_bh_disable((unsigned long)__builtin_return_address(0), - SOFTIRQ_OFFSET); - wakeup_softirqd(); - __local_bh_enable(SOFTIRQ_OFFSET); - } -} #else -static inline void invoke_softirq(void) -{ - if (!force_irqthreads) do_softirq(); - else { +#endif + } else { __local_bh_disable((unsigned long)__builtin_return_address(0), SOFTIRQ_OFFSET); wakeup_softirqd(); __local_bh_enable(SOFTIRQ_OFFSET); } } -#endif /* * Exit an interrupt context. Process softirqs if needed and possible: diff --git a/kernel/srcu.c b/kernel/srcu.c index 0febf61e1aa3..ba35f3a4a1f4 100644 --- a/kernel/srcu.c +++ b/kernel/srcu.c @@ -172,6 +172,12 @@ static void __synchronize_srcu(struct srcu_struct *sp, void (*sync_func)(void)) { int idx; + rcu_lockdep_assert(!lock_is_held(&sp->dep_map) && + !lock_is_held(&rcu_bh_lock_map) && + !lock_is_held(&rcu_lock_map) && + !lock_is_held(&rcu_sched_lock_map), + "Illegal synchronize_srcu() in same-type SRCU (or RCU) read-side critical section"); + idx = sp->completed; mutex_lock(&sp->mutex); @@ -280,19 +286,26 @@ void synchronize_srcu(struct srcu_struct *sp) EXPORT_SYMBOL_GPL(synchronize_srcu); /** - * synchronize_srcu_expedited - like synchronize_srcu, but less patient + * synchronize_srcu_expedited - Brute-force SRCU grace period * @sp: srcu_struct with which to synchronize. * - * Flip the completed counter, and wait for the old count to drain to zero. - * As with classic RCU, the updater must use some separate means of - * synchronizing concurrent updates. Can block; must be called from - * process context. + * Wait for an SRCU grace period to elapse, but use a "big hammer" + * approach to force the grace period to end quickly. This consumes + * significant time on all CPUs and is unfriendly to real-time workloads, + * so is thus not recommended for any sort of common-case code. In fact, + * if you are using synchronize_srcu_expedited() in a loop, please + * restructure your code to batch your updates, and then use a single + * synchronize_srcu() instead. * - * Note that it is illegal to call synchronize_srcu_expedited() - * from the corresponding SRCU read-side critical section; doing so - * will result in deadlock. However, it is perfectly legal to call - * synchronize_srcu_expedited() on one srcu_struct from some other - * srcu_struct's read-side critical section. + * Note that it is illegal to call this function while holding any lock + * that is acquired by a CPU-hotplug notifier. And yes, it is also illegal + * to call this function from a CPU-hotplug notifier. Failing to observe + * these restriction will result in deadlock. It is also illegal to call + * synchronize_srcu_expedited() from the corresponding SRCU read-side + * critical section; doing so will result in deadlock. However, it is + * perfectly legal to call synchronize_srcu_expedited() on one srcu_struct + * from some other srcu_struct's read-side critical section, as long as + * the resulting graph of srcu_structs is acyclic. */ void synchronize_srcu_expedited(struct srcu_struct *sp) { diff --git a/kernel/sys.c b/kernel/sys.c index 40701538fbd1..888d227fd195 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -1706,7 +1706,7 @@ static int prctl_set_mm(int opt, unsigned long addr, if (arg4 | arg5) return -EINVAL; - if (!capable(CAP_SYS_ADMIN)) + if (!capable(CAP_SYS_RESOURCE)) return -EPERM; if (addr >= TASK_SIZE) diff --git a/kernel/workqueue.c b/kernel/workqueue.c index bec7b5b53e03..f2c5638bb5ab 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -253,11 +253,13 @@ struct workqueue_struct *system_long_wq __read_mostly; struct workqueue_struct *system_nrt_wq __read_mostly; struct workqueue_struct *system_unbound_wq __read_mostly; struct workqueue_struct *system_freezable_wq __read_mostly; +struct workqueue_struct *system_nrt_freezable_wq __read_mostly; EXPORT_SYMBOL_GPL(system_wq); EXPORT_SYMBOL_GPL(system_long_wq); EXPORT_SYMBOL_GPL(system_nrt_wq); EXPORT_SYMBOL_GPL(system_unbound_wq); EXPORT_SYMBOL_GPL(system_freezable_wq); +EXPORT_SYMBOL_GPL(system_nrt_freezable_wq); #define CREATE_TRACE_POINTS #include <trace/events/workqueue.h> @@ -3833,8 +3835,11 @@ static int __init init_workqueues(void) WQ_UNBOUND_MAX_ACTIVE); system_freezable_wq = alloc_workqueue("events_freezable", WQ_FREEZABLE, 0); + system_nrt_freezable_wq = alloc_workqueue("events_nrt_freezable", + WQ_NON_REENTRANT | WQ_FREEZABLE, 0); BUG_ON(!system_wq || !system_long_wq || !system_nrt_wq || - !system_unbound_wq || !system_freezable_wq); + !system_unbound_wq || !system_freezable_wq || + !system_nrt_freezable_wq); return 0; } early_initcall(init_workqueues); |