diff options
Diffstat (limited to 'kernel')
| -rw-r--r-- | kernel/locking/mutex-debug.c | 4 | ||||
| -rw-r--r-- | kernel/rcu/srcutiny.c | 120 | ||||
| -rw-r--r-- | kernel/rcu/srcutree.c | 489 | ||||
| -rw-r--r-- | kernel/rcu/sync.c | 13 | ||||
| -rw-r--r-- | kernel/rcu/tree.c | 50 | ||||
| -rw-r--r-- | kernel/rcu/tree_exp.h | 10 | ||||
| -rw-r--r-- | kernel/rcu/tree_plugin.h | 46 | ||||
| -rw-r--r-- | kernel/rcu/update.c | 3 |
8 files changed, 382 insertions, 353 deletions
diff --git a/kernel/locking/mutex-debug.c b/kernel/locking/mutex-debug.c index 9aa713629387..771d4ca96dda 100644 --- a/kernel/locking/mutex-debug.c +++ b/kernel/locking/mutex-debug.c @@ -36,7 +36,7 @@ void debug_mutex_lock_common(struct mutex *lock, struct mutex_waiter *waiter) void debug_mutex_wake_waiter(struct mutex *lock, struct mutex_waiter *waiter) { - SMP_DEBUG_LOCKS_WARN_ON(!spin_is_locked(&lock->wait_lock)); + lockdep_assert_held(&lock->wait_lock); DEBUG_LOCKS_WARN_ON(list_empty(&lock->wait_list)); DEBUG_LOCKS_WARN_ON(waiter->magic != waiter); DEBUG_LOCKS_WARN_ON(list_empty(&waiter->list)); @@ -51,7 +51,7 @@ void debug_mutex_free_waiter(struct mutex_waiter *waiter) void debug_mutex_add_waiter(struct mutex *lock, struct mutex_waiter *waiter, struct task_struct *task) { - SMP_DEBUG_LOCKS_WARN_ON(!spin_is_locked(&lock->wait_lock)); + lockdep_assert_held(&lock->wait_lock); /* Mark the current thread as blocked on the lock: */ task->blocked_on = waiter; diff --git a/kernel/rcu/srcutiny.c b/kernel/rcu/srcutiny.c index b46e6683f8c9..32dfd6522548 100644 --- a/kernel/rcu/srcutiny.c +++ b/kernel/rcu/srcutiny.c @@ -37,30 +37,30 @@ int rcu_scheduler_active __read_mostly; static LIST_HEAD(srcu_boot_list); static bool srcu_init_done; -static int init_srcu_struct_fields(struct srcu_struct *sp) +static int init_srcu_struct_fields(struct srcu_struct *ssp) { - sp->srcu_lock_nesting[0] = 0; - sp->srcu_lock_nesting[1] = 0; - init_swait_queue_head(&sp->srcu_wq); - sp->srcu_cb_head = NULL; - sp->srcu_cb_tail = &sp->srcu_cb_head; - sp->srcu_gp_running = false; - sp->srcu_gp_waiting = false; - sp->srcu_idx = 0; - INIT_WORK(&sp->srcu_work, srcu_drive_gp); - INIT_LIST_HEAD(&sp->srcu_work.entry); + ssp->srcu_lock_nesting[0] = 0; + ssp->srcu_lock_nesting[1] = 0; + init_swait_queue_head(&ssp->srcu_wq); + ssp->srcu_cb_head = NULL; + ssp->srcu_cb_tail = &ssp->srcu_cb_head; + ssp->srcu_gp_running = false; + ssp->srcu_gp_waiting = false; + ssp->srcu_idx = 0; + INIT_WORK(&ssp->srcu_work, srcu_drive_gp); + INIT_LIST_HEAD(&ssp->srcu_work.entry); return 0; } #ifdef CONFIG_DEBUG_LOCK_ALLOC -int __init_srcu_struct(struct srcu_struct *sp, const char *name, +int __init_srcu_struct(struct srcu_struct *ssp, const char *name, struct lock_class_key *key) { /* Don't re-initialize a lock while it is held. */ - debug_check_no_locks_freed((void *)sp, sizeof(*sp)); - lockdep_init_map(&sp->dep_map, name, key, 0); - return init_srcu_struct_fields(sp); + debug_check_no_locks_freed((void *)ssp, sizeof(*ssp)); + lockdep_init_map(&ssp->dep_map, name, key, 0); + return init_srcu_struct_fields(ssp); } EXPORT_SYMBOL_GPL(__init_srcu_struct); @@ -68,15 +68,15 @@ EXPORT_SYMBOL_GPL(__init_srcu_struct); /* * init_srcu_struct - initialize a sleep-RCU structure - * @sp: structure to initialize. + * @ssp: structure to initialize. * * Must invoke this on a given srcu_struct before passing that srcu_struct * to any other function. Each srcu_struct represents a separate domain * of SRCU protection. */ -int init_srcu_struct(struct srcu_struct *sp) +int init_srcu_struct(struct srcu_struct *ssp) { - return init_srcu_struct_fields(sp); + return init_srcu_struct_fields(ssp); } EXPORT_SYMBOL_GPL(init_srcu_struct); @@ -84,22 +84,22 @@ EXPORT_SYMBOL_GPL(init_srcu_struct); /* * cleanup_srcu_struct - deconstruct a sleep-RCU structure - * @sp: structure to clean up. + * @ssp: structure to clean up. * * Must invoke this after you are finished using a given srcu_struct that * was initialized via init_srcu_struct(), else you leak memory. */ -void _cleanup_srcu_struct(struct srcu_struct *sp, bool quiesced) +void _cleanup_srcu_struct(struct srcu_struct *ssp, bool quiesced) { - WARN_ON(sp->srcu_lock_nesting[0] || sp->srcu_lock_nesting[1]); + WARN_ON(ssp->srcu_lock_nesting[0] || ssp->srcu_lock_nesting[1]); if (quiesced) - WARN_ON(work_pending(&sp->srcu_work)); + WARN_ON(work_pending(&ssp->srcu_work)); else - flush_work(&sp->srcu_work); - WARN_ON(sp->srcu_gp_running); - WARN_ON(sp->srcu_gp_waiting); - WARN_ON(sp->srcu_cb_head); - WARN_ON(&sp->srcu_cb_head != sp->srcu_cb_tail); + flush_work(&ssp->srcu_work); + WARN_ON(ssp->srcu_gp_running); + WARN_ON(ssp->srcu_gp_waiting); + WARN_ON(ssp->srcu_cb_head); + WARN_ON(&ssp->srcu_cb_head != ssp->srcu_cb_tail); } EXPORT_SYMBOL_GPL(_cleanup_srcu_struct); @@ -107,13 +107,13 @@ EXPORT_SYMBOL_GPL(_cleanup_srcu_struct); * Removes the count for the old reader from the appropriate element of * the srcu_struct. */ -void __srcu_read_unlock(struct srcu_struct *sp, int idx) +void __srcu_read_unlock(struct srcu_struct *ssp, int idx) { - int newval = sp->srcu_lock_nesting[idx] - 1; + int newval = ssp->srcu_lock_nesting[idx] - 1; - WRITE_ONCE(sp->srcu_lock_nesting[idx], newval); - if (!newval && READ_ONCE(sp->srcu_gp_waiting)) - swake_up_one(&sp->srcu_wq); + WRITE_ONCE(ssp->srcu_lock_nesting[idx], newval); + if (!newval && READ_ONCE(ssp->srcu_gp_waiting)) + swake_up_one(&ssp->srcu_wq); } EXPORT_SYMBOL_GPL(__srcu_read_unlock); @@ -127,24 +127,24 @@ void srcu_drive_gp(struct work_struct *wp) int idx; struct rcu_head *lh; struct rcu_head *rhp; - struct srcu_struct *sp; + struct srcu_struct *ssp; - sp = container_of(wp, struct srcu_struct, srcu_work); - if (sp->srcu_gp_running || !READ_ONCE(sp->srcu_cb_head)) + ssp = container_of(wp, struct srcu_struct, srcu_work); + if (ssp->srcu_gp_running || !READ_ONCE(ssp->srcu_cb_head)) return; /* Already running or nothing to do. */ /* Remove recently arrived callbacks and wait for readers. */ - WRITE_ONCE(sp->srcu_gp_running, true); + WRITE_ONCE(ssp->srcu_gp_running, true); local_irq_disable(); - lh = sp->srcu_cb_head; - sp->srcu_cb_head = NULL; - sp->srcu_cb_tail = &sp->srcu_cb_head; + lh = ssp->srcu_cb_head; + ssp->srcu_cb_head = NULL; + ssp->srcu_cb_tail = &ssp->srcu_cb_head; local_irq_enable(); - idx = sp->srcu_idx; - WRITE_ONCE(sp->srcu_idx, !sp->srcu_idx); - WRITE_ONCE(sp->srcu_gp_waiting, true); /* srcu_read_unlock() wakes! */ - swait_event_exclusive(sp->srcu_wq, !READ_ONCE(sp->srcu_lock_nesting[idx])); - WRITE_ONCE(sp->srcu_gp_waiting, false); /* srcu_read_unlock() cheap. */ + idx = ssp->srcu_idx; + WRITE_ONCE(ssp->srcu_idx, !ssp->srcu_idx); + WRITE_ONCE(ssp->srcu_gp_waiting, true); /* srcu_read_unlock() wakes! */ + swait_event_exclusive(ssp->srcu_wq, !READ_ONCE(ssp->srcu_lock_nesting[idx])); + WRITE_ONCE(ssp->srcu_gp_waiting, false); /* srcu_read_unlock() cheap. */ /* Invoke the callbacks we removed above. */ while (lh) { @@ -161,9 +161,9 @@ void srcu_drive_gp(struct work_struct *wp) * at interrupt level, but the ->srcu_gp_running checks will * straighten that out. */ - WRITE_ONCE(sp->srcu_gp_running, false); - if (READ_ONCE(sp->srcu_cb_head)) - schedule_work(&sp->srcu_work); + WRITE_ONCE(ssp->srcu_gp_running, false); + if (READ_ONCE(ssp->srcu_cb_head)) + schedule_work(&ssp->srcu_work); } EXPORT_SYMBOL_GPL(srcu_drive_gp); @@ -171,7 +171,7 @@ EXPORT_SYMBOL_GPL(srcu_drive_gp); * Enqueue an SRCU callback on the specified srcu_struct structure, * initiating grace-period processing if it is not already running. */ -void call_srcu(struct srcu_struct *sp, struct rcu_head *rhp, +void call_srcu(struct srcu_struct *ssp, struct rcu_head *rhp, rcu_callback_t func) { unsigned long flags; @@ -179,14 +179,14 @@ void call_srcu(struct srcu_struct *sp, struct rcu_head *rhp, rhp->func = func; rhp->next = NULL; local_irq_save(flags); - *sp->srcu_cb_tail = rhp; - sp->srcu_cb_tail = &rhp->next; + *ssp->srcu_cb_tail = rhp; + ssp->srcu_cb_tail = &rhp->next; local_irq_restore(flags); - if (!READ_ONCE(sp->srcu_gp_running)) { + if (!READ_ONCE(ssp->srcu_gp_running)) { if (likely(srcu_init_done)) - schedule_work(&sp->srcu_work); - else if (list_empty(&sp->srcu_work.entry)) - list_add(&sp->srcu_work.entry, &srcu_boot_list); + schedule_work(&ssp->srcu_work); + else if (list_empty(&ssp->srcu_work.entry)) + list_add(&ssp->srcu_work.entry, &srcu_boot_list); } } EXPORT_SYMBOL_GPL(call_srcu); @@ -194,13 +194,13 @@ EXPORT_SYMBOL_GPL(call_srcu); /* * synchronize_srcu - wait for prior SRCU read-side critical-section completion */ -void synchronize_srcu(struct srcu_struct *sp) +void synchronize_srcu(struct srcu_struct *ssp) { struct rcu_synchronize rs; init_rcu_head_on_stack(&rs.head); init_completion(&rs.completion); - call_srcu(sp, &rs.head, wakeme_after_rcu); + call_srcu(ssp, &rs.head, wakeme_after_rcu); wait_for_completion(&rs.completion); destroy_rcu_head_on_stack(&rs.head); } @@ -219,13 +219,13 @@ void __init rcu_scheduler_starting(void) */ void __init srcu_init(void) { - struct srcu_struct *sp; + struct srcu_struct *ssp; srcu_init_done = true; while (!list_empty(&srcu_boot_list)) { - sp = list_first_entry(&srcu_boot_list, + ssp = list_first_entry(&srcu_boot_list, struct srcu_struct, srcu_work.entry); - list_del_init(&sp->srcu_work.entry); - schedule_work(&sp->srcu_work); + list_del_init(&ssp->srcu_work.entry); + schedule_work(&ssp->srcu_work); } } diff --git a/kernel/rcu/srcutree.c b/kernel/rcu/srcutree.c index a8846ed7f352..3600d88d8956 100644 --- a/kernel/rcu/srcutree.c +++ b/kernel/rcu/srcutree.c @@ -56,7 +56,7 @@ static LIST_HEAD(srcu_boot_list); static bool __read_mostly srcu_init_done; static void srcu_invoke_callbacks(struct work_struct *work); -static void srcu_reschedule(struct srcu_struct *sp, unsigned long delay); +static void srcu_reschedule(struct srcu_struct *ssp, unsigned long delay); static void process_srcu(struct work_struct *work); /* Wrappers for lock acquisition and release, see raw_spin_lock_rcu_node(). */ @@ -92,7 +92,7 @@ do { \ * srcu_read_unlock() running against them. So if the is_static parameter * is set, don't initialize ->srcu_lock_count[] and ->srcu_unlock_count[]. */ -static void init_srcu_struct_nodes(struct srcu_struct *sp, bool is_static) +static void init_srcu_struct_nodes(struct srcu_struct *ssp, bool is_static) { int cpu; int i; @@ -103,13 +103,13 @@ static void init_srcu_struct_nodes(struct srcu_struct *sp, bool is_static) struct srcu_node *snp_first; /* Work out the overall tree geometry. */ - sp->level[0] = &sp->node[0]; + ssp->level[0] = &ssp->node[0]; for (i = 1; i < rcu_num_lvls; i++) - sp->level[i] = sp->level[i - 1] + num_rcu_lvl[i - 1]; + ssp->level[i] = ssp->level[i - 1] + num_rcu_lvl[i - 1]; rcu_init_levelspread(levelspread, num_rcu_lvl); /* Each pass through this loop initializes one srcu_node structure. */ - srcu_for_each_node_breadth_first(sp, snp) { + srcu_for_each_node_breadth_first(ssp, snp) { spin_lock_init(&ACCESS_PRIVATE(snp, lock)); WARN_ON_ONCE(ARRAY_SIZE(snp->srcu_have_cbs) != ARRAY_SIZE(snp->srcu_data_have_cbs)); @@ -120,17 +120,17 @@ static void init_srcu_struct_nodes(struct srcu_struct *sp, bool is_static) snp->srcu_gp_seq_needed_exp = 0; snp->grplo = -1; snp->grphi = -1; - if (snp == &sp->node[0]) { + if (snp == &ssp->node[0]) { /* Root node, special case. */ snp->srcu_parent = NULL; continue; } /* Non-root node. */ - if (snp == sp->level[level + 1]) + if (snp == ssp->level[level + 1]) level++; - snp->srcu_parent = sp->level[level - 1] + - (snp - sp->level[level]) / + snp->srcu_parent = ssp->level[level - 1] + + (snp - ssp->level[level]) / levelspread[level - 1]; } @@ -141,14 +141,14 @@ static void init_srcu_struct_nodes(struct srcu_struct *sp, bool is_static) WARN_ON_ONCE(ARRAY_SIZE(sdp->srcu_lock_count) != ARRAY_SIZE(sdp->srcu_unlock_count)); level = rcu_num_lvls - 1; - snp_first = sp->level[level]; + snp_first = ssp->level[level]; for_each_possible_cpu(cpu) { - sdp = per_cpu_ptr(sp->sda, cpu); + sdp = per_cpu_ptr(ssp->sda, cpu); spin_lock_init(&ACCESS_PRIVATE(sdp, lock)); rcu_segcblist_init(&sdp->srcu_cblist); sdp->srcu_cblist_invoking = false; - sdp->srcu_gp_seq_needed = sp->srcu_gp_seq; - sdp->srcu_gp_seq_needed_exp = sp->srcu_gp_seq; + sdp->srcu_gp_seq_needed = ssp->srcu_gp_seq; + sdp->srcu_gp_seq_needed_exp = ssp->srcu_gp_seq; sdp->mynode = &snp_first[cpu / levelspread[level]]; for (snp = sdp->mynode; snp != NULL; snp = snp->srcu_parent) { if (snp->grplo < 0) @@ -157,7 +157,7 @@ static void init_srcu_struct_nodes(struct srcu_struct *sp, bool is_static) } sdp->cpu = cpu; INIT_DELAYED_WORK(&sdp->work, srcu_invoke_callbacks); - sdp->sp = sp; + sdp->ssp = ssp; sdp->grpmask = 1 << (cpu - sdp->mynode->grplo); if (is_static) continue; @@ -176,35 +176,35 @@ static void init_srcu_struct_nodes(struct srcu_struct *sp, bool is_static) * parameter is passed through to init_srcu_struct_nodes(), and * also tells us that ->sda has already been wired up to srcu_data. */ -static int init_srcu_struct_fields(struct srcu_struct *sp, bool is_static) +static int init_srcu_struct_fields(struct srcu_struct *ssp, bool is_static) { - mutex_init(&sp->srcu_cb_mutex); - mutex_init(&sp->srcu_gp_mutex); - sp->srcu_idx = 0; - sp->srcu_gp_seq = 0; - sp->srcu_barrier_seq = 0; - mutex_init(&sp->srcu_barrier_mutex); - atomic_set(&sp->srcu_barrier_cpu_cnt, 0); - INIT_DELAYED_WORK(&sp->work, process_srcu); + mutex_init(&ssp->srcu_cb_mutex); + mutex_init(&ssp->srcu_gp_mutex); + ssp->srcu_idx = 0; + ssp->srcu_gp_seq = 0; + ssp->srcu_barrier_seq = 0; + mutex_init(&ssp->srcu_barrier_mutex); + atomic_set(&ssp->srcu_barrier_cpu_cnt, 0); + INIT_DELAYED_WORK(&ssp->work, process_srcu); if (!is_static) - sp->sda = alloc_percpu(struct srcu_data); - init_srcu_struct_nodes(sp, is_static); - sp->srcu_gp_seq_needed_exp = 0; - sp->srcu_last_gp_end = ktime_get_mono_fast_ns(); - smp_store_release(&sp->srcu_gp_seq_needed, 0); /* Init done. */ - return sp->sda ? 0 : -ENOMEM; + ssp->sda = alloc_percpu(struct srcu_data); + init_srcu_struct_nodes(ssp, is_static); + ssp->srcu_gp_seq_needed_exp = 0; + ssp->srcu_last_gp_end = ktime_get_mono_fast_ns(); + smp_store_release(&ssp->srcu_gp_seq_needed, 0); /* Init done. */ + return ssp->sda ? 0 : -ENOMEM; } #ifdef CONFIG_DEBUG_LOCK_ALLOC -int __init_srcu_struct(struct srcu_struct *sp, const char *name, +int __init_srcu_struct(struct srcu_struct *ssp, const char *name, struct lock_class_key *key) { /* Don't re-initialize a lock while it is held. */ - debug_check_no_locks_freed((void *)sp, sizeof(*sp)); - lockdep_init_map(&sp->dep_map, name, key, 0); - spin_lock_init(&ACCESS_PRIVATE(sp, lock)); - return init_srcu_struct_fields(sp, false); + debug_check_no_locks_freed((void *)ssp, sizeof(*ssp)); + lockdep_init_map(&ssp->dep_map, name, key, 0); + spin_lock_init(&ACCESS_PRIVATE(ssp, lock)); + return init_srcu_struct_fields(ssp, false); } EXPORT_SYMBOL_GPL(__init_srcu_struct); @@ -212,16 +212,16 @@ EXPORT_SYMBOL_GPL(__init_srcu_struct); /** * init_srcu_struct - initialize a sleep-RCU structure - * @sp: structure to initialize. + * @ssp: structure to initialize. * * Must invoke this on a given srcu_struct before passing that srcu_struct * to any other function. Each srcu_struct represents a separate domain * of SRCU protection. */ -int init_srcu_struct(struct srcu_struct *sp) +int init_srcu_struct(struct srcu_struct *ssp) { - spin_lock_init(&ACCESS_PRIVATE(sp, lock)); - return init_srcu_struct_fields(sp, false); + spin_lock_init(&ACCESS_PRIVATE(ssp, lock)); + return init_srcu_struct_fields(ssp, false); } EXPORT_SYMBOL_GPL(init_srcu_struct); @@ -231,37 +231,37 @@ EXPORT_SYMBOL_GPL(init_srcu_struct); * First-use initialization of statically allocated srcu_struct * structure. Wiring up the combining tree is more than can be * done with compile-time initialization, so this check is added - * to each update-side SRCU primitive. Use sp->lock, which -is- + * to each update-side SRCU primitive. Use ssp->lock, which -is- * compile-time initialized, to resolve races involving multiple * CPUs trying to garner first-use privileges. */ -static void check_init_srcu_struct(struct srcu_struct *sp) +static void check_init_srcu_struct(struct srcu_struct *ssp) { unsigned long flags; /* The smp_load_acquire() pairs with the smp_store_release(). */ - if (!rcu_seq_state(smp_load_acquire(&sp->srcu_gp_seq_needed))) /*^^^*/ + if (!rcu_seq_state(smp_load_acquire(&ssp->srcu_gp_seq_needed))) /*^^^*/ return; /* Already initialized. */ - spin_lock_irqsave_rcu_node(sp, flags); - if (!rcu_seq_state(sp->srcu_gp_seq_needed)) { - spin_unlock_irqrestore_rcu_node(sp, flags); + spin_lock_irqsave_rcu_node(ssp, flags); + if (!rcu_seq_state(ssp->srcu_gp_seq_needed)) { + spin_unlock_irqrestore_rcu_node(ssp, flags); return; } - init_srcu_struct_fields(sp, true); - spin_unlock_irqrestore_rcu_node(sp, flags); + init_srcu_struct_fields(ssp, true); + spin_unlock_irqrestore_rcu_node(ssp, flags); } /* * Returns approximate total of the readers' ->srcu_lock_count[] values * for the rank of per-CPU counters specified by idx. */ -static unsigned long srcu_readers_lock_idx(struct srcu_struct *sp, int idx) +static unsigned long srcu_readers_lock_idx(struct srcu_struct *ssp, int idx) { int cpu; unsigned long sum = 0; for_each_possible_cpu(cpu) { - struct srcu_data *cpuc = per_cpu_ptr(sp->sda, cpu); + struct srcu_data *cpuc = per_cpu_ptr(ssp->sda, cpu); sum += READ_ONCE(cpuc->srcu_lock_count[idx]); } @@ -272,13 +272,13 @@ static unsigned long srcu_readers_lock_idx(struct srcu_struct *sp, int idx) * Returns approximate total of the readers' ->srcu_unlock_count[] values * for the rank of per-CPU counters specified by idx. */ -static unsigned long srcu_readers_unlock_idx(struct srcu_struct *sp, int idx) +static unsigned long srcu_readers_unlock_idx(struct srcu_struct *ssp, int idx) { int cpu; unsigned long sum = 0; for_each_possible_cpu(cpu) { - struct srcu_data *cpuc = per_cpu_ptr(sp->sda, cpu); + struct srcu_data *cpuc = per_cpu_ptr(ssp->sda, cpu); sum += READ_ONCE(cpuc->srcu_unlock_count[idx]); } @@ -289,11 +289,11 @@ static unsigned long srcu_readers_unlock_idx(struct srcu_struct *sp, int idx) * Return true if the number of pre-existing readers is determined to * be zero. */ -static bool srcu_readers_active_idx_check(struct srcu_struct *sp, int idx) +static bool srcu_readers_active_idx_check(struct srcu_struct *ssp, int idx) { unsigned long unlocks; - unlocks = srcu_readers_unlock_idx(sp, idx); + unlocks = srcu_readers_unlock_idx(ssp, idx); /* * Make sure that a lock is always counted if the corresponding @@ -329,25 +329,25 @@ static bool srcu_readers_active_idx_check(struct srcu_struct *sp, int idx) * of floor(ULONG_MAX/NR_CPUS/2), which should be sufficient, * especially on 64-bit systems. */ - return srcu_readers_lock_idx(sp, idx) == unlocks; + return srcu_readers_lock_idx(ssp, idx) == unlocks; } /** * srcu_readers_active - returns true if there are readers. and false * otherwise - * @sp: which srcu_struct to count active readers (holding srcu_read_lock). + * @ssp: which srcu_struct to count active readers (holding srcu_read_lock). * * Note that this is not an atomic primitive, and can therefore suffer * severe errors when invoked on an active srcu_struct. That said, it * can be useful as an error check at cleanup time. */ -static bool srcu_readers_active(struct srcu_struct *sp) +static bool srcu_readers_active(struct srcu_struct *ssp) { int cpu; unsigned long sum = 0; for_each_possible_cpu(cpu) { - struct srcu_data *cpuc = per_cpu_ptr(sp->sda, cpu); + struct srcu_data *cpuc = per_cpu_ptr(ssp->sda, cpu); sum += READ_ONCE(cpuc->srcu_lock_count[0]); sum += READ_ONCE(cpuc->srcu_lock_count[1]); @@ -363,44 +363,44 @@ static bool srcu_readers_active(struct srcu_struct *sp) * Return grace-period delay, zero if there are expedited grace * periods pending, SRCU_INTERVAL otherwise. */ -static unsigned long srcu_get_delay(struct srcu_struct *sp) +static unsigned long srcu_get_delay(struct srcu_struct *ssp) { - if (ULONG_CMP_LT(READ_ONCE(sp->srcu_gp_seq), - READ_ONCE(sp->srcu_gp_seq_needed_exp))) + if (ULONG_CMP_LT(READ_ONCE(ssp->srcu_gp_seq), + READ_ONCE(ssp->srcu_gp_seq_needed_exp))) return 0; return SRCU_INTERVAL; } /* Helper for cleanup_srcu_struct() and cleanup_srcu_struct_quiesced(). */ -void _cleanup_srcu_struct(struct srcu_struct *sp, bool quiesced) +void _cleanup_srcu_struct(struct srcu_struct *ssp, bool quiesced) { int cpu; - if (WARN_ON(!srcu_get_delay(sp))) + if (WARN_ON(!srcu_get_delay(ssp))) return; /* Just leak it! */ - if (WARN_ON(srcu_readers_active(sp))) + if (WARN_ON(srcu_readers_active(ssp))) return; /* Just leak it! */ if (quiesced) { - if (WARN_ON(delayed_work_pending(&sp->work))) + if (WARN_ON(delayed_work_pending(&ssp->work))) return; /* Just leak it! */ } else { - flush_delayed_work(&sp->work); + flush_delayed_work(&ssp->work); } for_each_possible_cpu(cpu) if (quiesced) { - if (WARN_ON(delayed_work_pending(&per_cpu_ptr(sp->sda, cpu)->work))) + if (WARN_ON(delayed_work_pending(&per_cpu_ptr(ssp->sda, cpu)->work))) return; /* Just leak it! */ } else { - flush_delayed_work(&per_cpu_ptr(sp->sda, cpu)->work); + flush_delayed_work(&per_cpu_ptr(ssp->sda, cpu)->work); } - if (WARN_ON(rcu_seq_state(READ_ONCE(sp->srcu_gp_seq)) != SRCU_STATE_IDLE) || - WARN_ON(srcu_readers_active(sp))) { + if (WARN_ON(rcu_seq_state(READ_ONCE(ssp->srcu_gp_seq)) != SRCU_STATE_IDLE) || + WARN_ON(srcu_readers_active(ssp))) { pr_info("%s: Active srcu_struct %p state: %d\n", - __func__, sp, rcu_seq_state(READ_ONCE(sp->srcu_gp_seq))); + __func__, ssp, rcu_seq_state(READ_ONCE(ssp->srcu_gp_seq))); return; /* Caller forgot to stop doing call_srcu()? */ } - free_percpu(sp->sda); - sp->sda = NULL; + free_percpu(ssp->sda); + ssp->sda = NULL; } EXPORT_SYMBOL_GPL(_cleanup_srcu_struct); @@ -409,12 +409,12 @@ EXPORT_SYMBOL_GPL(_cleanup_srcu_struct); * srcu_struct. * Returns an index that must be passed to the matching srcu_read_unlock(). */ -int __srcu_read_lock(struct srcu_struct *sp) +int __srcu_read_lock(struct srcu_struct *ssp) { int idx; - idx = READ_ONCE(sp->srcu_idx) & 0x1; - this_cpu_inc(sp->sda->srcu_lock_count[idx]); + idx = READ_ONCE(ssp->srcu_idx) & 0x1; + this_cpu_inc(ssp->sda->srcu_lock_count[idx]); smp_mb(); /* B */ /* Avoid leaking the critical section. */ return idx; } @@ -425,10 +425,10 @@ EXPORT_SYMBOL_GPL(__srcu_read_lock); * element of the srcu_struct. Note that this may well be a different * CPU than that which was incremented by the corresponding srcu_read_lock(). */ -void __srcu_read_unlock(struct srcu_struct *sp, int idx) +void __srcu_read_unlock(struct srcu_struct *ssp, int idx) { smp_mb(); /* C */ /* Avoid leaking the critical section. */ - this_cpu_inc(sp->sda->srcu_unlock_count[idx]); + this_cpu_inc(ssp->sda->srcu_unlock_count[idx]); } EXPORT_SYMBOL_GPL(__srcu_read_unlock); @@ -444,20 +444,22 @@ EXPORT_SYMBOL_GPL(__srcu_read_unlock); /* * Start an SRCU grace period. */ -static void srcu_gp_start(struct srcu_struct *sp) +static void srcu_gp_start(struct srcu_struct *ssp) { - struct srcu_data *sdp = this_cpu_ptr(sp->sda); + struct srcu_data *sdp = this_cpu_ptr(ssp->sda); int state; - lockdep_assert_held(&ACCESS_PRIVATE(sp, lock)); - WARN_ON_ONCE(ULONG_CMP_GE(sp->srcu_gp_seq, sp->srcu_gp_seq_needed)); + lockdep_assert_held(&ACCESS_PRIVATE(ssp, lock)); + WARN_ON_ONCE(ULONG_CMP_GE(ssp->srcu_gp_seq, ssp->srcu_gp_seq_needed)); + spin_lock_rcu_node(sdp); /* Interrupts already disabled. */ rcu_segcblist_advance(&sdp->srcu_cblist, - rcu_seq_current(&sp->srcu_gp_seq)); + rcu_seq_current(&ssp->srcu_gp_seq)); (void)rcu_segcblist_accelerate(&sdp->srcu_cblist, - rcu_seq_snap(&sp->srcu_gp_seq)); + rcu_seq_snap(&ssp->srcu_gp_seq)); + spin_unlock_rcu_node(sdp); /* Interrupts remain disabled. */ smp_mb(); /* Order prior store to ->srcu_gp_seq_needed vs. GP start. */ - rcu_seq_start(&sp->srcu_gp_seq); - state = rcu_seq_state(READ_ONCE(sp->srcu_gp_seq)); + rcu_seq_start(&ssp->srcu_gp_seq); + state = rcu_seq_state(READ_ONCE(ssp->srcu_gp_seq)); WARN_ON_ONCE(state != SRCU_STATE_SCAN1); } @@ -511,7 +513,7 @@ static void srcu_schedule_cbs_sdp(struct srcu_data *sdp, unsigned long delay) * just-completed grace period, the one corresponding to idx. If possible, * schedule this invocation on the corresponding CPUs. */ -static void srcu_schedule_cbs_snp(struct srcu_struct *sp, struct srcu_node *snp, +static void srcu_schedule_cbs_snp(struct srcu_struct *ssp, struct srcu_node *snp, unsigned long mask, unsigned long delay) { int cpu; @@ -519,7 +521,7 @@ static void srcu_schedule_cbs_snp(struct srcu_struct *sp, struct srcu_node *snp, for (cpu = snp->grplo; cpu <= snp->grphi; cpu++) { if (!(mask & (1 << (cpu - snp->grplo)))) continue; - srcu_schedule_cbs_sdp(per_cpu_ptr(sp->sda, cpu), delay); + srcu_schedule_cbs_sdp(per_cpu_ptr(ssp->sda, cpu), delay); } } @@ -532,7 +534,7 @@ static void srcu_schedule_cbs_snp(struct srcu_struct *sp, struct srcu_node *snp, * are initiating callback invocation. This allows the ->srcu_have_cbs[] * array to have a finite number of elements. */ -static void srcu_gp_end(struct srcu_struct *sp) +static void srcu_gp_end(struct srcu_struct *ssp) { unsigned long cbdelay; bool cbs; @@ -546,28 +548,28 @@ static void srcu_gp_end(struct srcu_struct *sp) struct srcu_node *snp; /* Prevent more than one additional grace period. */ - mutex_lock(&sp->srcu_cb_mutex); + mutex_lock(&ssp->srcu_cb_mutex); /* End the current grace period. */ - spin_lock_irq_rcu_node(sp); - idx = rcu_seq_state(sp->srcu_gp_seq); + spin_lock_irq_rcu_node(ssp); + idx = rcu_seq_state(ssp->srcu_gp_seq); WARN_ON_ONCE(idx != SRCU_STATE_SCAN2); - cbdelay = srcu_get_delay(sp); - sp->srcu_last_gp_end = ktime_get_mono_fast_ns(); - rcu_seq_end(&sp->srcu_gp_seq); - gpseq = rcu_seq_current(&sp->srcu_gp_seq); - if (ULONG_CMP_LT(sp->srcu_gp_seq_needed_exp, gpseq)) - sp->srcu_gp_seq_needed_exp = gpseq; - spin_unlock_irq_rcu_node(sp); - mutex_unlock(&sp->srcu_gp_mutex); + cbdelay = srcu_get_delay(ssp); + ssp->srcu_last_gp_end = ktime_get_mono_fast_ns(); + rcu_seq_end(&ssp->srcu_gp_seq); + gpseq = rcu_seq_current(&ssp->srcu_gp_seq); + if (ULONG_CMP_LT(ssp->srcu_gp_seq_needed_exp, gpseq)) + ssp->srcu_gp_seq_needed_exp = gpseq; + spin_unlock_irq_rcu_node(ssp); + mutex_unlock(&ssp->srcu_gp_mutex); /* A new grace period can start at this point. But only one. */ /* Initiate callback invocation as needed. */ idx = rcu_seq_ctr(gpseq) % ARRAY_SIZE(snp->srcu_have_cbs); - srcu_for_each_node_breadth_first(sp, snp) { + srcu_for_each_node_breadth_first(ssp, snp) { spin_lock_irq_rcu_node(snp); cbs = false; - last_lvl = snp >= sp->level[rcu_num_lvls - 1]; + last_lvl = snp >= ssp->level[rcu_num_lvls - 1]; if (last_lvl) cbs = snp->srcu_have_cbs[idx] == gpseq; snp->srcu_have_cbs[idx] = gpseq; @@ -578,12 +580,12 @@ static void srcu_gp_end(struct srcu_struct *sp) snp->srcu_data_have_cbs[idx] = 0; spin_unlock_irq_rcu_node(snp); if (cbs) - srcu_schedule_cbs_snp(sp, snp, mask, cbdelay); + srcu_schedule_cbs_snp(ssp, snp, mask, cbdelay); /* Occasionally prevent srcu_data counter wrap. */ if (!(gpseq & counter_wrap_check) && last_lvl) for (cpu = snp->grplo; cpu <= snp->grphi; cpu++) { - sdp = per_cpu_ptr(sp->sda, cpu); + sdp = per_cpu_ptr(ssp->sda, cpu); spin_lock_irqsave_rcu_node(sdp, flags); if (ULONG_CMP_GE(gpseq, sdp->srcu_gp_seq_needed + 100)) @@ -596,18 +598,18 @@ static void srcu_gp_end(struct srcu_struct *sp) } /* Callback initiation done, allow grace periods after next. */ - mutex_unlock(&sp->srcu_cb_mutex); + mutex_unlock(&ssp->srcu_cb_mutex); /* Start a new grace period if needed. */ - spin_lock_irq_rcu_node(sp); - gpseq = rcu_seq_current(&sp->srcu_gp_seq); + spin_lock_irq_rcu_node(ssp); + gpseq = rcu_seq_current(&ssp->srcu_gp_seq); if (!rcu_seq_state(gpseq) && - ULONG_CMP_LT(gpseq, sp->srcu_gp_seq_needed)) { - srcu_gp_start(sp); - spin_unlock_irq_rcu_node(sp); - srcu_reschedule(sp, 0); + ULONG_CMP_LT(gpseq, ssp->srcu_gp_seq_needed)) { + srcu_gp_start(ssp); + spin_unlock_irq_rcu_node(ssp); + srcu_reschedule(ssp, 0); } else { - spin_unlock_irq_rcu_node(sp); + spin_unlock_irq_rcu_node(ssp); } } @@ -618,13 +620,13 @@ static void srcu_gp_end(struct srcu_struct *sp) * but without expediting. To start a completely new grace period, * whether expedited or not, use srcu_funnel_gp_start() instead. */ -static void srcu_funnel_exp_start(struct srcu_struct *sp, struct srcu_node *snp, +static void srcu_funnel_exp_start(struct srcu_struct *ssp, struct srcu_node *snp, unsigned long s) { unsigned long flags; for (; snp != NULL; snp = snp->srcu_parent) { - if (rcu_seq_done(&sp->srcu_gp_seq, s) || + if (rcu_seq_done(&ssp->srcu_gp_seq, s) || ULONG_CMP_GE(READ_ONCE(snp->srcu_gp_seq_needed_exp), s)) return; spin_lock_irqsave_rcu_node(snp, flags); @@ -635,10 +637,10 @@ static void srcu_funnel_exp_start(struct srcu_struct *sp, struct srcu_node *snp, WRITE_ONCE(snp->srcu_gp_seq_needed_exp, s); spin_unlock_irqrestore_rcu_node(snp, flags); } - spin_lock_irqsave_rcu_node(sp, flags); - if (ULONG_CMP_LT(sp->srcu_gp_seq_needed_exp, s)) - sp->srcu_gp_seq_needed_exp = s; - spin_unlock_irqrestore_rcu_node(sp, flags); + spin_lock_irqsave_rcu_node(ssp, flags); + if (ULONG_CMP_LT(ssp->srcu_gp_seq_needed_exp, s)) + ssp->srcu_gp_seq_needed_exp = s; + spin_unlock_irqrestore_rcu_node(ssp, flags); } /* @@ -651,7 +653,7 @@ static void srcu_funnel_exp_start(struct srcu_struct *sp, struct srcu_node *snp, * Note that this function also does the work of srcu_funnel_exp_start(), * in some cases by directly invoking it. */ -static void srcu_funnel_gp_start(struct srcu_struct *sp, struct srcu_data *sdp, +static void srcu_funnel_gp_start(struct srcu_struct *ssp, struct srcu_data *sdp, unsigned long s, bool do_norm) { unsigned long flags; @@ -661,7 +663,7 @@ static void srcu_funnel_gp_start(struct srcu_struct *sp, struct srcu_data *sdp, /* Each pass through the loop does one level of the srcu_node tree. */ for (; snp != NULL; snp = snp->srcu_parent) { - if (rcu_seq_done(&sp->srcu_gp_seq, s) && snp != sdp->mynode) + if (rcu_seq_done(&ssp->srcu_gp_seq, s) && snp != sdp->mynode) return; /* GP already done and CBs recorded. */ spin_lock_irqsave_rcu_node(snp, flags); if (ULONG_CMP_GE(snp->srcu_have_cbs[idx], s)) { @@ -676,7 +678,7 @@ static void srcu_funnel_gp_start(struct srcu_struct *sp, struct srcu_data *sdp, return; } if (!do_norm) - srcu_funnel_exp_start(sp, snp, s); + srcu_funnel_exp_start(ssp, snp, s); return; } snp->srcu_have_cbs[idx] = s; @@ -688,29 +690,29 @@ static void srcu_funnel_gp_start(struct srcu_struct *sp, struct srcu_data *sdp, } /* Top of tree, must ensure the grace period will be started. */ - spin_lock_irqsave_rcu_node(sp, flags); - if (ULONG_CMP_LT(sp->srcu_gp_seq_needed, s)) { + spin_lock_irqsave_rcu_node(ssp, flags); + if (ULONG_CMP_LT(ssp->srcu_gp_seq_needed, s)) { /* * Record need for grace period s. Pair with load * acquire setting up for initialization. */ - smp_store_release(&sp->srcu_gp_seq_needed, s); /*^^^*/ + smp_store_release(&ssp->srcu_gp_seq_needed, s); /*^^^*/ } - if (!do_norm && ULONG_CMP_LT(sp->srcu_gp_seq_needed_exp, s)) - sp->srcu_gp_seq_needed_exp = s; + if (!do_norm && ULONG_CMP_LT(ssp->srcu_gp_seq_needed_exp, s)) + ssp->srcu_gp_seq_needed_exp = s; /* If grace period not already done and none in progress, start it. */ - if (!rcu_seq_done(&sp->srcu_gp_seq, s) && - rcu_seq_state(sp->srcu_gp_seq) == SRCU_STATE_IDLE) { - WARN_ON_ONCE(ULONG_CMP_GE(sp->srcu_gp_seq, sp->srcu_gp_seq_needed)); - srcu_gp_start(sp); + if (!rcu_seq_done(&ssp->srcu_gp_seq, s) && + rcu_seq_state(ssp->srcu_gp_seq) == SRCU_STATE_IDLE) { + WARN_ON_ONCE(ULONG_CMP_GE(ssp->srcu_gp_seq, ssp->srcu_gp_seq_needed)); + srcu_gp_start(ssp); if (likely(srcu_init_done)) - queue_delayed_work(rcu_gp_wq, &sp->work, - srcu_get_delay(sp)); - else if (list_empty(&sp->work.work.entry)) - list_add(&sp->work.work.entry, &srcu_boot_list); + queue_delayed_work(rcu_gp_wq, &ssp->work, + srcu_get_delay(ssp)); + else if (list_empty(&ssp->work.work.entry)) + list_add(&ssp->work.work.entry, &srcu_boot_list); } - spin_unlock_irqrestore_rcu_node(sp, flags); + spin_unlock_irqrestore_rcu_node(ssp, flags); } /* @@ -718,12 +720,12 @@ static void srcu_funnel_gp_start(struct srcu_struct *sp, struct srcu_data *sdp, * loop an additional time if there is an expedited grace period pending. * The caller must ensure that ->srcu_idx is not changed while checking. */ -static bool try_check_zero(struct srcu_struct *sp, int idx, int trycount) +static bool try_check_zero(struct srcu_struct *ssp, int idx, int trycount) { for (;;) { - if (srcu_readers_active_idx_check(sp, idx)) + if (srcu_readers_active_idx_check(ssp, idx)) return true; - if (--trycount + !srcu_get_delay(sp) <= 0) + if (--trycount + !srcu_get_delay(ssp) <= 0) return false; udelay(SRCU_RETRY_CHECK_DELAY); } @@ -734,7 +736,7 @@ static bool try_check_zero(struct srcu_struct *sp, int idx, int trycount) * use the other rank of the ->srcu_(un)lock_count[] arrays. This allows * us to wait for pre-existing readers in a starvation-free manner. */ -static void srcu_flip(struct srcu_struct *sp) +static void srcu_flip(struct srcu_struct *ssp) { /* * Ensure that if this updater saw a given reader's increment @@ -746,7 +748,7 @@ static void srcu_flip(struct srcu_struct *sp) */ smp_mb(); /* E */ /* Pairs with B and C. */ - WRITE_ONCE(sp->srcu_idx, sp->srcu_idx + 1); + WRITE_ONCE(ssp->srcu_idx, ssp->srcu_idx + 1); /* * Ensure that if the updater misses an __srcu_read_unlock() @@ -779,7 +781,7 @@ static void srcu_flip(struct srcu_struct *sp) * negligible when amoritized over that time period, and the extra latency * of a needlessly non-expedited grace period is similarly negligible. */ -static bool srcu_might_be_idle(struct srcu_struct *sp) +static bool srcu_might_be_idle(struct srcu_struct *ssp) { unsigned long curseq; unsigned long flags; @@ -788,7 +790,7 @@ static bool srcu_might_be_idle(struct srcu_struct *sp) /* If the local srcu_data structure has callbacks, not idle. */ local_irq_save(flags); - sdp = this_cpu_ptr(sp->sda); + sdp = this_cpu_ptr(ssp->sda); if (rcu_segcblist_pend_cbs(&sdp->srcu_cblist)) { local_irq_restore(flags); return false; /* Callbacks already present, so not idle. */ @@ -804,17 +806,17 @@ static bool srcu_might_be_idle(struct srcu_struct *sp) /* First, see if enough time has passed since the last GP. */ t = ktime_get_mono_fast_ns(); if (exp_holdoff == 0 || - time_in_range_open(t, sp->srcu_last_gp_end, - sp->srcu_last_gp_end + exp_holdoff)) + time_in_range_open(t, ssp->srcu_last_gp_end, + ssp->srcu_last_gp_end + exp_holdoff)) return false; /* Too soon after last GP. */ /* Next, check for probable idleness. */ - curseq = rcu_seq_current(&sp->srcu_gp_seq); + curseq = rcu_seq_current(&ssp->srcu_gp_seq); smp_mb(); /* Order ->srcu_gp_seq with ->srcu_gp_seq_needed. */ - if (ULONG_CMP_LT(curseq, READ_ONCE(sp->srcu_gp_seq_needed))) + if (ULONG_CMP_LT(curseq, READ_ONCE(ssp->srcu_gp_seq_needed))) return false; /* Grace period in progress, so not idle. */ smp_mb(); /* Order ->srcu_gp_seq with prior access. */ - if (curseq != rcu_seq_current(&sp->srcu_gp_seq)) + if (curseq != rcu_seq_current(&ssp->srcu_gp_seq)) return false; /* GP # changed, so not idle. */ return true; /* With reasonable probability, idle! */ } @@ -854,16 +856,17 @@ static void srcu_leak_callback(struct rcu_head *rhp) * srcu_read_lock(), and srcu_read_unlock() that are all passed the same * srcu_struct structure. */ -void __call_srcu(struct srcu_struct *sp, struct rcu_head *rhp, +void __call_srcu(struct srcu_struct *ssp, struct rcu_head *rhp, rcu_callback_t func, bool do_norm) { unsigned long flags; + int idx; bool needexp = false; bool needgp = false; unsigned long s; struct srcu_data *sdp; - check_init_srcu_struct(sp); + check_init_srcu_struct(ssp); if (debug_rcu_head_queue(rhp)) { /* Probable double call_srcu(), so leak the callback. */ WRITE_ONCE(rhp->func, srcu_leak_callback); @@ -871,13 +874,14 @@ void __call_srcu(struct srcu_struct *sp, struct rcu_head *rhp, return; } rhp->func = func; + idx = srcu_read_lock(ssp); local_irq_save(flags); - sdp = this_cpu_ptr(sp->sda); + sdp = this_cpu_ptr(ssp->sda); spin_lock_rcu_node(sdp); rcu_segcblist_enqueue(&sdp->srcu_cblist, rhp, false); rcu_segcblist_advance(&sdp->srcu_cblist, - rcu_seq_current(&sp->srcu_gp_seq)); - s = rcu_seq_snap(&sp->srcu_gp_seq); + rcu_seq_current(&ssp->srcu_gp_seq)); + s = rcu_seq_snap(&ssp->srcu_gp_seq); (void)rcu_segcblist_accelerate(&sdp->srcu_cblist, s); if (ULONG_CMP_LT(sdp->srcu_gp_seq_needed, s)) { sdp->srcu_gp_seq_needed = s; @@ -889,14 +893,15 @@ void __call_srcu(struct srcu_struct *sp, struct rcu_head *rhp, } spin_unlock_irqrestore_rcu_node(sdp, flags); if (needgp) - srcu_funnel_gp_start(sp, sdp, s, do_norm); + srcu_funnel_gp_start(ssp, sdp, s, do_norm); else if (needexp) - srcu_funnel_exp_start(sp, sdp->mynode, s); + srcu_funnel_exp_start(ssp, sdp->mynode, s); + srcu_read_unlock(ssp, idx); } /** * call_srcu() - Queue a callback for invocation after an SRCU grace period - * @sp: srcu_struct in queue the callback + * @ssp: srcu_struct in queue the callback * @rhp: structure to be used for queueing the SRCU callback. * @func: function to be invoked after the SRCU grace period * @@ -911,21 +916,21 @@ void __call_srcu(struct srcu_struct *sp, struct rcu_head *rhp, * The callback will be invoked from process context, but must nevertheless * be fast and must not block. */ -void call_srcu(struct srcu_struct *sp, struct rcu_head *rhp, +void call_srcu(struct srcu_struct *ssp, struct rcu_head *rhp, rcu_callback_t func) { - __call_srcu(sp, rhp, func, true); + __call_srcu(ssp, rhp, func, true); } EXPORT_SYMBOL_GPL(call_srcu); /* * Helper function for synchronize_srcu() and synchronize_srcu_expedited(). */ -static void __synchronize_srcu(struct srcu_struct *sp, bool do_norm) +static void __synchronize_srcu(struct srcu_struct *ssp, bool do_norm) { struct rcu_synchronize rcu; - RCU_LOCKDEP_WARN(lock_is_held(&sp->dep_map) || + RCU_LOCKDEP_WARN(lock_is_held(&ssp->dep_map) || lock_is_held(&rcu_bh_lock_map) || lock_is_held(&rcu_lock_map) || lock_is_held(&rcu_sched_lock_map), @@ -934,10 +939,10 @@ static void __synchronize_srcu(struct srcu_struct *sp, bool do_norm) if (rcu_scheduler_active == RCU_SCHEDULER_INACTIVE) return; might_sleep(); - check_init_srcu_struct(sp); + check_init_srcu_struct(ssp); init_completion(&rcu.completion); init_rcu_head_on_stack(&rcu.head); - __call_srcu(sp, &rcu.head, wakeme_after_rcu, do_norm); + __call_srcu(ssp, &rcu.head, wakeme_after_rcu, do_norm); wait_for_completion(&rcu.completion); destroy_rcu_head_on_stack(&rcu.head); @@ -953,7 +958,7 @@ static void __synchronize_srcu(struct srcu_struct *sp, bool do_norm) /** * synchronize_srcu_expedited - Brute-force SRCU grace period - * @sp: srcu_struct with which to synchronize. + * @ssp: srcu_struct with which to synchronize. * * Wait for an SRCU grace period to elapse, but be more aggressive about * spinning rather than blocking when waiting. @@ -961,15 +966,15 @@ static void __synchronize_srcu(struct srcu_struct *sp, bool do_norm) * Note that synchronize_srcu_expedited() has the same deadlock and * memory-ordering properties as does synchronize_srcu(). */ -void synchronize_srcu_expedited(struct srcu_struct *sp) +void synchronize_srcu_expedited(struct srcu_struct *ssp) { - __synchronize_srcu(sp, rcu_gp_is_normal()); + __synchronize_srcu(ssp, rcu_gp_is_normal()); } EXPORT_SYMBOL_GPL(synchronize_srcu_expedited); /** * synchronize_srcu - wait for prior SRCU read-side critical-section completion - * @sp: srcu_struct with which to synchronize. + * @ssp: srcu_struct with which to synchronize. * * Wait for the count to drain to zero of both indexes. To avoid the * possible starvation of synchronize_srcu(), it waits for the count of @@ -1011,12 +1016,12 @@ EXPORT_SYMBOL_GPL(synchronize_srcu_expedited); * SRCU must also provide it. Note that detecting idleness is heuristic * and subject to both false positives and negatives. */ -void synchronize_srcu(struct srcu_struct *sp) +void synchronize_srcu(struct srcu_struct *ssp) { - if (srcu_might_be_idle(sp) || rcu_gp_is_expedited()) - synchronize_srcu_expedited(sp); + if (srcu_might_be_idle(ssp) || rcu_gp_is_expedited()) + synchronize_srcu_expedited(ssp); else - __synchronize_srcu(sp, true); + __synchronize_srcu(ssp, true); } EXPORT_SYMBOL_GPL(synchronize_srcu); @@ -1026,36 +1031,36 @@ EXPORT_SYMBOL_GPL(synchronize_srcu); static void srcu_barrier_cb(struct rcu_head *rhp) { struct srcu_data *sdp; - struct srcu_struct *sp; + struct srcu_struct *ssp; sdp = container_of(rhp, struct srcu_data, srcu_barrier_head); - sp = sdp->sp; - if (atomic_dec_and_test(&sp->srcu_barrier_cpu_cnt)) - complete(&sp->srcu_barrier_completion); + ssp = sdp->ssp; + if (atomic_dec_and_test(&ssp->srcu_barrier_cpu_cnt)) + complete(&ssp->srcu_barrier_completion); } /** * srcu_barrier - Wait until all in-flight call_srcu() callbacks complete. - * @sp: srcu_struct on which to wait for in-flight callbacks. + * @ssp: srcu_struct on which to wait for in-flight callbacks. */ -void srcu_barrier(struct srcu_struct *sp) +void srcu_barrier(struct srcu_struct *ssp) { int cpu; struct srcu_data *sdp; - unsigned long s = rcu_seq_snap(&sp->srcu_barrier_seq); + unsigned long s = rcu_seq_snap(&ssp->srcu_barrier_seq); - check_init_srcu_struct(sp); - mutex_lock(&sp->srcu_barrier_mutex); - if (rcu_seq_done(&sp->srcu_barrier_seq, s)) { + check_init_srcu_struct(ssp); + mutex_lock(&ssp->srcu_barrier_mutex); + if (rcu_seq_done(&ssp->srcu_barrier_seq, s)) { smp_mb(); /* Force ordering following return. */ - mutex_unlock(&sp->srcu_barrier_mutex); + mutex_unlock(&ssp->srcu_barrier_mutex); return; /* Someone else did our work for us. */ } - rcu_seq_start(&sp->srcu_barrier_seq); - init_completion(&sp->srcu_barrier_completion); + rcu_seq_start(&ssp->srcu_barrier_seq); + init_completion(&ssp->srcu_barrier_completion); /* Initial count prevents reaching zero until all CBs are posted. */ - atomic_set(&sp->srcu_barrier_cpu_cnt, 1); + atomic_set(&ssp->srcu_barrier_cpu_cnt, 1); /* * Each pass through this loop enqueues a callback, but only @@ -1066,39 +1071,39 @@ void srcu_barrier(struct srcu_struct *sp) * grace period as the last callback already in the queue. */ for_each_possible_cpu(cpu) { - sdp = per_cpu_ptr(sp->sda, cpu); + sdp = per_cpu_ptr(ssp->sda, cpu); spin_lock_irq_rcu_node(sdp); - atomic_inc(&sp->srcu_barrier_cpu_cnt); + atomic_inc(&ssp->srcu_barrier_cpu_cnt); sdp->srcu_barrier_head.func = srcu_barrier_cb; debug_rcu_head_queue(&sdp->srcu_barrier_head); if (!rcu_segcblist_entrain(&sdp->srcu_cblist, &sdp->srcu_barrier_head, 0)) { debug_rcu_head_unqueue(&sdp->srcu_barrier_head); - atomic_dec(&sp->srcu_barrier_cpu_cnt); + atomic_dec(&ssp->srcu_barrier_cpu_cnt); } spin_unlock_irq_rcu_node(sdp); } /* Remove the initial count, at which point reaching zero can happen. */ - if (atomic_dec_and_test(&sp->srcu_barrier_cpu_cnt)) - complete(&sp->srcu_barrier_completion); - wait_for_completion(&sp->srcu_barrier_completion); + if (atomic_dec_and_test(&ssp->srcu_barrier_cpu_cnt)) + complete(&ssp->srcu_barrier_completion); + wait_for_completion(&ssp->srcu_barrier_completion); - rcu_seq_end(&sp->srcu_barrier_seq); - mutex_unlock(&sp->srcu_barrier_mutex); + rcu_seq_end(&ssp->srcu_barrier_seq); + mutex_unlock(&ssp->srcu_barrier_mutex); } EXPORT_SYMBOL_GPL(srcu_barrier); /** * srcu_batches_completed - return batches completed. - * @sp: srcu_struct on which to report batch completion. + * @ssp: srcu_struct on which to report batch completion. * * Report the number of batches, correlated with, but not necessarily * precisely the same as, the number of grace periods that have elapsed. */ -unsigned long srcu_batches_completed(struct srcu_struct *sp) +unsigned long srcu_batches_completed(struct srcu_struct *ssp) { - return sp->srcu_idx; + return ssp->srcu_idx; } EXPORT_SYMBOL_GPL(srcu_batches_completed); @@ -1107,11 +1112,11 @@ EXPORT_SYMBOL_GPL(srcu_batches_completed); * to SRCU_STATE_SCAN2, and invoke srcu_gp_end() when scan has * completed in that state. */ -static void srcu_advance_state(struct srcu_struct *sp) +static void srcu_advance_state(struct srcu_struct *ssp) { int idx; - mutex_lock(&sp->srcu_gp_mutex); + mutex_lock(&ssp->srcu_gp_mutex); /* * Because readers might be delayed for an extended period after @@ -1123,47 +1128,47 @@ static void srcu_advance_state(struct srcu_struct *sp) * The load-acquire ensures that we see the accesses performed * by the prior grace period. */ - idx = rcu_seq_state(smp_load_acquire(&sp->srcu_gp_seq)); /* ^^^ */ + idx = rcu_seq_state(smp_load_acquire(&ssp->srcu_gp_seq)); /* ^^^ */ if (idx == SRCU_STATE_IDLE) { - spin_lock_irq_rcu_node(sp); - if (ULONG_CMP_GE(sp->srcu_gp_seq, sp->srcu_gp_seq_needed)) { - WARN_ON_ONCE(rcu_seq_state(sp->srcu_gp_seq)); - spin_unlock_irq_rcu_node(sp); - mutex_unlock(&sp->srcu_gp_mutex); + spin_lock_irq_rcu_node(ssp); + if (ULONG_CMP_GE(ssp->srcu_gp_seq, ssp->srcu_gp_seq_needed)) { + WARN_ON_ONCE(rcu_seq_state(ssp->srcu_gp_seq)); + spin_unlock_irq_rcu_node(ssp); + mutex_unlock(&ssp->srcu_gp_mutex); return; } - idx = rcu_seq_state(READ_ONCE(sp->srcu_gp_seq)); + idx = rcu_seq_state(READ_ONCE(ssp->srcu_gp_seq)); if (idx == SRCU_STATE_IDLE) - srcu_gp_start(sp); - spin_unlock_irq_rcu_node(sp); + srcu_gp_start(ssp); + spin_unlock_irq_rcu_node(ssp); if (idx != SRCU_STATE_IDLE) { - mutex_unlock(&sp->srcu_gp_mutex); + mutex_unlock(&ssp->srcu_gp_mutex); return; /* Someone else started the grace period. */ } } - if (rcu_seq_state(READ_ONCE(sp->srcu_gp_seq)) == SRCU_STATE_SCAN1) { - idx = 1 ^ (sp->srcu_idx & 1); - if (!try_check_zero(sp, idx, 1)) { - mutex_unlock(&sp->srcu_gp_mutex); + if (rcu_seq_state(READ_ONCE(ssp->srcu_gp_seq)) == SRCU_STATE_SCAN1) { + idx = 1 ^ (ssp->srcu_idx & 1); + if (!try_check_zero(ssp, idx, 1)) { + mutex_unlock(&ssp->srcu_gp_mutex); return; /* readers present, retry later. */ } - srcu_flip(sp); - rcu_seq_set_state(&sp->srcu_gp_seq, SRCU_STATE_SCAN2); + srcu_flip(ssp); + rcu_seq_set_state(&ssp->srcu_gp_seq, SRCU_STATE_SCAN2); } - if (rcu_seq_state(READ_ONCE(sp->srcu_gp_seq)) == SRCU_STATE_SCAN2) { + if (rcu_seq_state(READ_ONCE(ssp->srcu_gp_seq)) == SRCU_STATE_SCAN2) { /* * SRCU read-side critical sections are normally short, * so check at least twice in quick succession after a flip. */ - idx = 1 ^ (sp->srcu_idx & 1); - if (!try_check_zero(sp, idx, 2)) { - mutex_unlock(&sp->srcu_gp_mutex); + idx = 1 ^ (ssp->srcu_idx & 1); + if (!try_check_zero(ssp, idx, 2)) { + mutex_unlock(&ssp->srcu_gp_mutex); return; /* readers present, retry later. */ } - srcu_gp_end(sp); /* Releases ->srcu_gp_mutex. */ + srcu_gp_end(ssp); /* Releases ->srcu_gp_mutex. */ } } @@ -1179,14 +1184,14 @@ static void srcu_invoke_callbacks(struct work_struct *work) struct rcu_cblist ready_cbs; struct rcu_head *rhp; struct srcu_data *sdp; - struct srcu_struct *sp; + struct srcu_struct *ssp; sdp = container_of(work, struct srcu_data, work.work); - sp = sdp->sp; + ssp = sdp->ssp; rcu_cblist_init(&ready_cbs); spin_lock_irq_rcu_node(sdp); rcu_segcblist_advance(&sdp->srcu_cblist, - rcu_seq_current(&sp->srcu_gp_seq)); + rcu_seq_current(&ssp->srcu_gp_seq)); if (sdp->srcu_cblist_invoking || !rcu_segcblist_ready_cbs(&sdp->srcu_cblist)) { spin_unlock_irq_rcu_node(sdp); @@ -1212,7 +1217,7 @@ static void srcu_invoke_callbacks(struct work_struct *work) spin_lock_irq_rcu_node(sdp); rcu_segcblist_insert_count(&sdp->srcu_cblist, &ready_cbs); (void)rcu_segcblist_accelerate(&sdp->srcu_cblist, - rcu_seq_snap(&sp->srcu_gp_seq)); + rcu_seq_snap(&ssp->srcu_gp_seq)); sdp->srcu_cblist_invoking = false; more = rcu_segcblist_ready_cbs(&sdp->srcu_cblist); spin_unlock_irq_rcu_node(sdp); @@ -1224,24 +1229,24 @@ static void srcu_invoke_callbacks(struct work_struct *work) * Finished one round of SRCU grace period. Start another if there are * more SRCU callbacks queued, otherwise put SRCU into not-running state. */ -static void srcu_reschedule(struct srcu_struct *sp, unsigned long delay) +static void srcu_reschedule(struct srcu_struct *ssp, unsigned long delay) { bool pushgp = true; - spin_lock_irq_rcu_node(sp); - if (ULONG_CMP_GE(sp->srcu_gp_seq, sp->srcu_gp_seq_needed)) { - if (!WARN_ON_ONCE(rcu_seq_state(sp->srcu_gp_seq))) { + spin_lock_irq_rcu_node(ssp); + if (ULONG_CMP_GE(ssp->srcu_gp_seq, ssp->srcu_gp_seq_needed)) { + if (!WARN_ON_ONCE(rcu_seq_state(ssp->srcu_gp_seq))) { /* All requests fulfilled, time to go idle. */ pushgp = false; } - } else if (!rcu_seq_state(sp->srcu_gp_seq)) { + } else if (!rcu_seq_state(ssp->srcu_gp_seq)) { /* Outstanding request and no GP. Start one. */ - srcu_gp_start(sp); + srcu_gp_start(ssp); } - spin_unlock_irq_rcu_node(sp); + spin_unlock_irq_rcu_node(ssp); if (pushgp) - queue_delayed_work(rcu_gp_wq, &sp->work, delay); + queue_delayed_work(rcu_gp_wq, &ssp->work, delay); } /* @@ -1249,41 +1254,41 @@ static void srcu_reschedule(struct srcu_struct *sp, unsigned long delay) */ static void process_srcu(struct work_struct *work) { - struct srcu_struct *sp; + struct srcu_struct *ssp; - sp = container_of(work, struct srcu_struct, work.work); + ssp = container_of(work, struct srcu_struct, work.work); - srcu_advance_state(sp); - srcu_reschedule(sp, srcu_get_delay(sp)); + srcu_advance_state(ssp); + srcu_reschedule(ssp, srcu_get_delay(ssp)); } void srcutorture_get_gp_data(enum rcutorture_type test_type, - struct srcu_struct *sp, int *flags, + struct srcu_struct *ssp, int *flags, unsigned long *gp_seq) { if (test_type != SRCU_FLAVOR) return; *flags = 0; - *gp_seq = rcu_seq_current(&sp->srcu_gp_seq); + *gp_seq = rcu_seq_current(&ssp->srcu_gp_seq); } EXPORT_SYMBOL_GPL(srcutorture_get_gp_data); -void srcu_torture_stats_print(struct srcu_struct *sp, char *tt, char *tf) +void srcu_torture_stats_print(struct srcu_struct *ssp, char *tt, char *tf) { int cpu; int idx; unsigned long s0 = 0, s1 = 0; - idx = sp->srcu_idx & 0x1; + idx = ssp->srcu_idx & 0x1; pr_alert("%s%s Tree SRCU g%ld per-CPU(idx=%d):", - tt, tf, rcu_seq_current(&sp->srcu_gp_seq), idx); + tt, tf, rcu_seq_current(&ssp->srcu_gp_seq), idx); for_each_possible_cpu(cpu) { unsigned long l0, l1; unsigned long u0, u1; long c0, c1; struct srcu_data *sdp; - sdp = per_cpu_ptr(sp->sda, cpu); + sdp = per_cpu_ptr(ssp->sda, cpu); u0 = sdp->srcu_unlock_count[!idx]; u1 = sdp->srcu_unlock_count[idx]; @@ -1318,14 +1323,14 @@ early_initcall(srcu_bootup_announce); void __init srcu_init(void) { - struct srcu_struct *sp; + struct srcu_struct *ssp; srcu_init_done = true; while (!list_empty(&srcu_boot_list)) { - sp = list_first_entry(&srcu_boot_list, struct srcu_struct, + ssp = list_first_entry(&srcu_boot_list, struct srcu_struct, work.work.entry); - check_init_srcu_struct(sp); - list_del_init(&sp->work.work.entry); - queue_work(rcu_gp_wq, &sp->work.work); + check_init_srcu_struct(ssp); + list_del_init(&ssp->work.work.entry); + queue_work(rcu_gp_wq, &ssp->work.work); } } diff --git a/kernel/rcu/sync.c b/kernel/rcu/sync.c index 9d570b1892b0..be10036fa621 100644 --- a/kernel/rcu/sync.c +++ b/kernel/rcu/sync.c @@ -125,8 +125,7 @@ void rcu_sync_enter(struct rcu_sync *rsp) rsp->gp_state = GP_PENDING; spin_unlock_irq(&rsp->rss_lock); - BUG_ON(need_wait && need_sync); - + WARN_ON_ONCE(need_wait && need_sync); if (need_sync) { gp_ops[rsp->gp_type].sync(); rsp->gp_state = GP_PASSED; @@ -139,7 +138,7 @@ void rcu_sync_enter(struct rcu_sync *rsp) * Nobody has yet been allowed the 'fast' path and thus we can * avoid doing any sync(). The callback will get 'dropped'. */ - BUG_ON(rsp->gp_state != GP_PASSED); + WARN_ON_ONCE(rsp->gp_state != GP_PASSED); } } @@ -166,8 +165,8 @@ static void rcu_sync_func(struct rcu_head *rhp) struct rcu_sync *rsp = container_of(rhp, struct rcu_sync, cb_head); unsigned long flags; - BUG_ON(rsp->gp_state != GP_PASSED); - BUG_ON(rsp->cb_state == CB_IDLE); + WARN_ON_ONCE(rsp->gp_state != GP_PASSED); + WARN_ON_ONCE(rsp->cb_state == CB_IDLE); spin_lock_irqsave(&rsp->rss_lock, flags); if (rsp->gp_count) { @@ -225,7 +224,7 @@ void rcu_sync_dtor(struct rcu_sync *rsp) { int cb_state; - BUG_ON(rsp->gp_count); + WARN_ON_ONCE(rsp->gp_count); spin_lock_irq(&rsp->rss_lock); if (rsp->cb_state == CB_REPLAY) @@ -235,6 +234,6 @@ void rcu_sync_dtor(struct rcu_sync *rsp) if (cb_state != CB_IDLE) { gp_ops[rsp->gp_type].wait(); - BUG_ON(rsp->cb_state != CB_IDLE); + WARN_ON_ONCE(rsp->cb_state != CB_IDLE); } } diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c index 121f833acd04..6ec3abbe90e2 100644 --- a/kernel/rcu/tree.c +++ b/kernel/rcu/tree.c @@ -500,16 +500,29 @@ void rcu_force_quiescent_state(void) EXPORT_SYMBOL_GPL(rcu_force_quiescent_state); /* + * Convert a ->gp_state value to a character string. + */ +static const char *gp_state_getname(short gs) +{ + if (gs < 0 || gs >= ARRAY_SIZE(gp_state_names)) + return "???"; + return gp_state_names[gs]; +} + +/* * Show the state of the grace-period kthreads. */ void show_rcu_gp_kthreads(void) { int cpu; + unsigned long j; struct rcu_data *rdp; struct rcu_node *rnp; - pr_info("%s: wait state: %d ->state: %#lx\n", rcu_state.name, - rcu_state.gp_state, rcu_state.gp_kthread->state); + j = jiffies - READ_ONCE(rcu_state.gp_activity); + pr_info("%s: wait state: %s(%d) ->state: %#lx delta ->gp_activity %ld\n", + rcu_state.name, gp_state_getname(rcu_state.gp_state), + rcu_state.gp_state, rcu_state.gp_kthread->state, j); rcu_for_each_node_breadth_first(rnp) { if (ULONG_CMP_GE(rcu_state.gp_seq, rnp->gp_seq_needed)) continue; @@ -891,12 +904,12 @@ void rcu_irq_enter_irqson(void) } /** - * rcu_is_watching - see if RCU thinks that the current CPU is idle + * rcu_is_watching - see if RCU thinks that the current CPU is not idle * * Return true if RCU is watching the running CPU, which means that this * CPU can safely enter RCU read-side critical sections. In other words, - * if the current CPU is in its idle loop and is neither in an interrupt - * or NMI handler, return true. + * if the current CPU is not in its idle loop or is in an interrupt or + * NMI handler, return true. */ bool notrace rcu_is_watching(void) { @@ -1143,16 +1156,6 @@ static void record_gp_stall_check_time(void) } /* - * Convert a ->gp_state value to a character string. - */ -static const char *gp_state_getname(short gs) -{ - if (gs < 0 || gs >= ARRAY_SIZE(gp_state_names)) - return "???"; - return gp_state_names[gs]; -} - -/* * Complain about starvation of grace-period kthread. */ static void rcu_check_gp_kthread_starvation(void) @@ -2032,9 +2035,9 @@ static void rcu_gp_cleanup(void) rnp = rcu_get_root(); raw_spin_lock_irq_rcu_node(rnp); /* GP before ->gp_seq update. */ - /* Declare grace period done. */ - rcu_seq_end(&rcu_state.gp_seq); + /* Declare grace period done, trace first to use old GP number. */ trace_rcu_grace_period(rcu_state.name, rcu_state.gp_seq, TPS("end")); + rcu_seq_end(&rcu_state.gp_seq); rcu_state.gp_state = RCU_GP_IDLE; /* Check for GP requests since above loop. */ rdp = this_cpu_ptr(&rcu_data); @@ -2600,10 +2603,10 @@ static void force_quiescent_state(void) * This function checks for grace-period requests that fail to motivate * RCU to come out of its idle mode. */ -static void -rcu_check_gp_start_stall(struct rcu_node *rnp, struct rcu_data *rdp) +void +rcu_check_gp_start_stall(struct rcu_node *rnp, struct rcu_data *rdp, + const unsigned long gpssdelay) { - const unsigned long gpssdelay = rcu_jiffies_till_stall_check() * HZ; unsigned long flags; unsigned long j; struct rcu_node *rnp_root = rcu_get_root(); @@ -2690,7 +2693,7 @@ static __latent_entropy void rcu_process_callbacks(struct softirq_action *unused local_irq_restore(flags); } - rcu_check_gp_start_stall(rnp, rdp); + rcu_check_gp_start_stall(rnp, rdp, rcu_jiffies_till_stall_check()); /* If there are callbacks ready, invoke them. */ if (rcu_segcblist_ready_cbs(&rdp->cblist)) @@ -2826,7 +2829,7 @@ __call_rcu(struct rcu_head *head, rcu_callback_t func, int cpu, bool lazy) * Very early boot, before rcu_init(). Initialize if needed * and then drop through to queue the callback. */ - BUG_ON(cpu != -1); + WARN_ON_ONCE(cpu != -1); WARN_ON_ONCE(!rcu_is_watching()); if (rcu_segcblist_empty(&rdp->cblist)) rcu_segcblist_init(&rdp->cblist); @@ -3485,7 +3488,8 @@ static int __init rcu_spawn_gp_kthread(void) rcu_scheduler_fully_active = 1; t = kthread_create(rcu_gp_kthread, NULL, "%s", rcu_state.name); - BUG_ON(IS_ERR(t)); + if (WARN_ONCE(IS_ERR(t), "%s: Could not start grace-period kthread, OOM is now expected behavior\n", __func__)) + return 0; rnp = rcu_get_root(); raw_spin_lock_irqsave_rcu_node(rnp, flags); rcu_state.gp_kthread = t; diff --git a/kernel/rcu/tree_exp.h b/kernel/rcu/tree_exp.h index 8d18c1014e2b..928fe5893a57 100644 --- a/kernel/rcu/tree_exp.h +++ b/kernel/rcu/tree_exp.h @@ -450,10 +450,12 @@ static void sync_rcu_exp_select_cpus(smp_call_func_t func) } INIT_WORK(&rnp->rew.rew_work, sync_rcu_exp_select_node_cpus); preempt_disable(); - cpu = cpumask_next(rnp->grplo - 1, cpu_online_mask); + cpu = find_next_bit(&rnp->ffmask, BITS_PER_LONG, -1); /* If all offline, queue the work on an unbound CPU. */ - if (unlikely(cpu > rnp->grphi)) + if (unlikely(cpu > rnp->grphi - rnp->grplo)) cpu = WORK_CPU_UNBOUND; + else + cpu += rnp->grplo; queue_work_on(cpu, rcu_par_gp_wq, &rnp->rew.rew_work); preempt_enable(); rnp->exp_need_flush = true; @@ -690,8 +692,10 @@ static void sync_rcu_exp_handler(void *unused) */ if (t->rcu_read_lock_nesting > 0) { raw_spin_lock_irqsave_rcu_node(rnp, flags); - if (rnp->expmask & rdp->grpmask) + if (rnp->expmask & rdp->grpmask) { rdp->deferred_qs = true; + WRITE_ONCE(t->rcu_read_unlock_special.b.exp_hint, true); + } raw_spin_unlock_irqrestore_rcu_node(rnp, flags); } diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h index 05915e536336..605ff3b06098 100644 --- a/kernel/rcu/tree_plugin.h +++ b/kernel/rcu/tree_plugin.h @@ -397,6 +397,11 @@ static int rcu_preempt_blocked_readers_cgp(struct rcu_node *rnp) return rnp->gp_tasks != NULL; } +/* Bias and limit values for ->rcu_read_lock_nesting. */ +#define RCU_NEST_BIAS INT_MAX +#define RCU_NEST_NMAX (-INT_MAX / 2) +#define RCU_NEST_PMAX (INT_MAX / 2) + /* * Preemptible RCU implementation for rcu_read_lock(). * Just increment ->rcu_read_lock_nesting, shared state will be updated @@ -405,6 +410,8 @@ static int rcu_preempt_blocked_readers_cgp(struct rcu_node *rnp) void __rcu_read_lock(void) { current->rcu_read_lock_nesting++; + if (IS_ENABLED(CONFIG_PROVE_LOCKING)) + WARN_ON_ONCE(current->rcu_read_lock_nesting > RCU_NEST_PMAX); barrier(); /* critical section after entry code. */ } EXPORT_SYMBOL_GPL(__rcu_read_lock); @@ -424,20 +431,18 @@ void __rcu_read_unlock(void) --t->rcu_read_lock_nesting; } else { barrier(); /* critical section before exit code. */ - t->rcu_read_lock_nesting = INT_MIN; + t->rcu_read_lock_nesting = -RCU_NEST_BIAS; barrier(); /* assign before ->rcu_read_unlock_special load */ if (unlikely(READ_ONCE(t->rcu_read_unlock_special.s))) rcu_read_unlock_special(t); barrier(); /* ->rcu_read_unlock_special load before assign */ t->rcu_read_lock_nesting = 0; } -#ifdef CONFIG_PROVE_LOCKING - { - int rrln = READ_ONCE(t->rcu_read_lock_nesting); + if (IS_ENABLED(CONFIG_PROVE_LOCKING)) { + int rrln = t->rcu_read_lock_nesting; - WARN_ON_ONCE(rrln < 0 && rrln > INT_MIN / 2); + WARN_ON_ONCE(rrln < 0 && rrln > RCU_NEST_NMAX); } -#endif /* #ifdef CONFIG_PROVE_LOCKING */ } EXPORT_SYMBOL_GPL(__rcu_read_unlock); @@ -597,7 +602,7 @@ rcu_preempt_deferred_qs_irqrestore(struct task_struct *t, unsigned long flags) */ static bool rcu_preempt_need_deferred_qs(struct task_struct *t) { - return (this_cpu_ptr(&rcu_data)->deferred_qs || + return (__this_cpu_read(rcu_data.deferred_qs) || READ_ONCE(t->rcu_read_unlock_special.s)) && t->rcu_read_lock_nesting <= 0; } @@ -617,11 +622,11 @@ static void rcu_preempt_deferred_qs(struct task_struct *t) if (!rcu_preempt_need_deferred_qs(t)) return; if (couldrecurse) - t->rcu_read_lock_nesting -= INT_MIN; + t->rcu_read_lock_nesting -= RCU_NEST_BIAS; local_irq_save(flags); rcu_preempt_deferred_qs_irqrestore(t, flags); if (couldrecurse) - t->rcu_read_lock_nesting += INT_MIN; + t->rcu_read_lock_nesting += RCU_NEST_BIAS; } /* @@ -642,13 +647,21 @@ static void rcu_read_unlock_special(struct task_struct *t) local_irq_save(flags); irqs_were_disabled = irqs_disabled_flags(flags); - if ((preempt_bh_were_disabled || irqs_were_disabled) && - t->rcu_read_unlock_special.b.blocked) { + if (preempt_bh_were_disabled || irqs_were_disabled) { + WRITE_ONCE(t->rcu_read_unlock_special.b.exp_hint, false); /* Need to defer quiescent state until everything is enabled. */ - raise_softirq_irqoff(RCU_SOFTIRQ); + if (irqs_were_disabled) { + /* Enabling irqs does not reschedule, so... */ + raise_softirq_irqoff(RCU_SOFTIRQ); + } else { + /* Enabling BH or preempt does reschedule, so... */ + set_tsk_need_resched(current); + set_preempt_need_resched(); + } local_irq_restore(flags); return; } + WRITE_ONCE(t->rcu_read_unlock_special.b.exp_hint, false); rcu_preempt_deferred_qs_irqrestore(t, flags); } @@ -1464,7 +1477,8 @@ static void __init rcu_spawn_boost_kthreads(void) for_each_possible_cpu(cpu) per_cpu(rcu_cpu_has_work, cpu) = 0; - BUG_ON(smpboot_register_percpu_thread(&rcu_cpu_thread_spec)); + if (WARN_ONCE(smpboot_register_percpu_thread(&rcu_cpu_thread_spec), "%s: Could not start rcub kthread, OOM is now expected behavior\n", __func__)) + return; rcu_for_each_leaf_node(rnp) (void)rcu_spawn_one_boost_kthread(rnp); } @@ -2322,7 +2336,8 @@ static int rcu_nocb_kthread(void *arg) tail = rdp->nocb_follower_tail; rdp->nocb_follower_tail = &rdp->nocb_follower_head; raw_spin_unlock_irqrestore(&rdp->nocb_lock, flags); - BUG_ON(!list); + if (WARN_ON_ONCE(!list)) + continue; trace_rcu_nocb_wake(rcu_state.name, rdp->cpu, TPS("WokeNonEmpty")); /* Each pass through the following loop invokes a callback. */ @@ -2495,7 +2510,8 @@ static void rcu_spawn_one_nocb_kthread(int cpu) /* Spawn the kthread for this CPU. */ t = kthread_run(rcu_nocb_kthread, rdp_spawn, "rcuo%c/%d", rcu_state.abbr, cpu); - BUG_ON(IS_ERR(t)); + if (WARN_ONCE(IS_ERR(t), "%s: Could not start rcuo kthread, OOM is now expected behavior\n", __func__)) + return; WRITE_ONCE(rdp_spawn->nocb_kthread, t); } diff --git a/kernel/rcu/update.c b/kernel/rcu/update.c index c729ca5e6ee2..1971869c4072 100644 --- a/kernel/rcu/update.c +++ b/kernel/rcu/update.c @@ -820,7 +820,8 @@ static int __init rcu_spawn_tasks_kthread(void) struct task_struct *t; t = kthread_run(rcu_tasks_kthread, NULL, "rcu_tasks_kthread"); - BUG_ON(IS_ERR(t)); + if (WARN_ONCE(IS_ERR(t), "%s: Could not start Tasks-RCU grace-period kthread, OOM is now expected behavior\n", __func__)) + return 0; smp_mb(); /* Ensure others see full kthread. */ WRITE_ONCE(rcu_tasks_kthread_ptr, t); return 0; |