diff options
Diffstat (limited to 'kernel')
| -rw-r--r-- | kernel/async.c | 14 | ||||
| -rw-r--r-- | kernel/sched/core.c | 2 | ||||
| -rw-r--r-- | kernel/workqueue.c | 1530 | ||||
| -rw-r--r-- | kernel/workqueue_internal.h | 65 | ||||
| -rw-r--r-- | kernel/workqueue_sched.h | 9 |
5 files changed, 790 insertions, 830 deletions
diff --git a/kernel/async.c b/kernel/async.c index 6f34904a0b53..6c68fc3fae7b 100644 --- a/kernel/async.c +++ b/kernel/async.c @@ -57,6 +57,8 @@ asynchronous and synchronous parts of the kernel. #include <linux/slab.h> #include <linux/workqueue.h> +#include "workqueue_internal.h" + static async_cookie_t next_cookie = 1; #define MAX_WORK 32768 @@ -353,3 +355,15 @@ void async_synchronize_cookie(async_cookie_t cookie) async_synchronize_cookie_domain(cookie, &async_running); } EXPORT_SYMBOL_GPL(async_synchronize_cookie); + +/** + * current_is_async - is %current an async worker task? + * + * Returns %true if %current is an async worker task. + */ +bool current_is_async(void) +{ + struct worker *worker = current_wq_worker(); + + return worker && worker->current_func == async_run_entry_fn; +} diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 4a88f1d51563..03d7784b7bd2 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -83,7 +83,7 @@ #endif #include "sched.h" -#include "../workqueue_sched.h" +#include "../workqueue_internal.h" #include "../smpboot.h" #define CREATE_TRACE_POINTS diff --git a/kernel/workqueue.c b/kernel/workqueue.c index fbc6576a83c3..f4feacad3812 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -41,32 +41,31 @@ #include <linux/debug_locks.h> #include <linux/lockdep.h> #include <linux/idr.h> +#include <linux/hashtable.h> -#include "workqueue_sched.h" +#include "workqueue_internal.h" enum { /* - * global_cwq flags + * worker_pool flags * - * A bound gcwq is either associated or disassociated with its CPU. + * A bound pool is either associated or disassociated with its CPU. * While associated (!DISASSOCIATED), all workers are bound to the * CPU and none has %WORKER_UNBOUND set and concurrency management * is in effect. * * While DISASSOCIATED, the cpu may be offline and all workers have * %WORKER_UNBOUND set and concurrency management disabled, and may - * be executing on any CPU. The gcwq behaves as an unbound one. + * be executing on any CPU. The pool behaves as an unbound one. * * Note that DISASSOCIATED can be flipped only while holding - * assoc_mutex of all pools on the gcwq to avoid changing binding - * state while create_worker() is in progress. + * assoc_mutex to avoid changing binding state while + * create_worker() is in progress. */ - GCWQ_DISASSOCIATED = 1 << 0, /* cpu can't serve workers */ - GCWQ_FREEZING = 1 << 1, /* freeze in progress */ - - /* pool flags */ POOL_MANAGE_WORKERS = 1 << 0, /* need to manage workers */ POOL_MANAGING_WORKERS = 1 << 1, /* managing workers */ + POOL_DISASSOCIATED = 1 << 2, /* cpu can't serve workers */ + POOL_FREEZING = 1 << 3, /* freeze in progress */ /* worker flags */ WORKER_STARTED = 1 << 0, /* started */ @@ -79,11 +78,9 @@ enum { WORKER_NOT_RUNNING = WORKER_PREP | WORKER_UNBOUND | WORKER_CPU_INTENSIVE, - NR_WORKER_POOLS = 2, /* # worker pools per gcwq */ + NR_STD_WORKER_POOLS = 2, /* # standard pools per cpu */ BUSY_WORKER_HASH_ORDER = 6, /* 64 pointers */ - BUSY_WORKER_HASH_SIZE = 1 << BUSY_WORKER_HASH_ORDER, - BUSY_WORKER_HASH_MASK = BUSY_WORKER_HASH_SIZE - 1, MAX_IDLE_WORKERS_RATIO = 4, /* 1/4 of busy can be idle */ IDLE_WORKER_TIMEOUT = 300 * HZ, /* keep idle ones for 5 mins */ @@ -111,48 +108,24 @@ enum { * P: Preemption protected. Disabling preemption is enough and should * only be modified and accessed from the local cpu. * - * L: gcwq->lock protected. Access with gcwq->lock held. + * L: pool->lock protected. Access with pool->lock held. * - * X: During normal operation, modification requires gcwq->lock and - * should be done only from local cpu. Either disabling preemption - * on local cpu or grabbing gcwq->lock is enough for read access. - * If GCWQ_DISASSOCIATED is set, it's identical to L. + * X: During normal operation, modification requires pool->lock and should + * be done only from local cpu. Either disabling preemption on local + * cpu or grabbing pool->lock is enough for read access. If + * POOL_DISASSOCIATED is set, it's identical to L. * * F: wq->flush_mutex protected. * * W: workqueue_lock protected. */ -struct global_cwq; -struct worker_pool; - -/* - * The poor guys doing the actual heavy lifting. All on-duty workers - * are either serving the manager role, on idle list or on busy hash. - */ -struct worker { - /* on idle list while idle, on busy hash table while busy */ - union { - struct list_head entry; /* L: while idle */ - struct hlist_node hentry; /* L: while busy */ - }; - - struct work_struct *current_work; /* L: work being processed */ - struct cpu_workqueue_struct *current_cwq; /* L: current_work's cwq */ - struct list_head scheduled; /* L: scheduled works */ - struct task_struct *task; /* I: worker task */ - struct worker_pool *pool; /* I: the associated pool */ - /* 64 bytes boundary on 64bit, 32 on 32bit */ - unsigned long last_active; /* L: last active timestamp */ - unsigned int flags; /* X: flags */ - int id; /* I: worker id */ - - /* for rebinding worker to CPU */ - struct work_struct rebind_work; /* L: for busy worker */ -}; +/* struct worker is defined in workqueue_internal.h */ struct worker_pool { - struct global_cwq *gcwq; /* I: the owning gcwq */ + spinlock_t lock; /* the pool lock */ + unsigned int cpu; /* I: the associated cpu */ + int id; /* I: pool ID */ unsigned int flags; /* X: flags */ struct list_head worklist; /* L: list of pending works */ @@ -165,34 +138,28 @@ struct worker_pool { struct timer_list idle_timer; /* L: worker idle timeout */ struct timer_list mayday_timer; /* L: SOS timer for workers */ - struct mutex assoc_mutex; /* protect GCWQ_DISASSOCIATED */ - struct ida worker_ida; /* L: for worker IDs */ -}; - -/* - * Global per-cpu workqueue. There's one and only one for each cpu - * and all works are queued and processed here regardless of their - * target workqueues. - */ -struct global_cwq { - spinlock_t lock; /* the gcwq lock */ - unsigned int cpu; /* I: the associated cpu */ - unsigned int flags; /* L: GCWQ_* flags */ - - /* workers are chained either in busy_hash or pool idle_list */ - struct hlist_head busy_hash[BUSY_WORKER_HASH_SIZE]; + /* workers are chained either in busy_hash or idle_list */ + DECLARE_HASHTABLE(busy_hash, BUSY_WORKER_HASH_ORDER); /* L: hash of busy workers */ - struct worker_pool pools[NR_WORKER_POOLS]; - /* normal and highpri pools */ + struct mutex assoc_mutex; /* protect POOL_DISASSOCIATED */ + struct ida worker_ida; /* L: for worker IDs */ + + /* + * The current concurrency level. As it's likely to be accessed + * from other CPUs during try_to_wake_up(), put it in a separate + * cacheline. + */ + atomic_t nr_running ____cacheline_aligned_in_smp; } ____cacheline_aligned_in_smp; /* - * The per-CPU workqueue. The lower WORK_STRUCT_FLAG_BITS of - * work_struct->data are used for flags and thus cwqs need to be - * aligned at two's power of the number of flag bits. + * The per-pool workqueue. While queued, the lower WORK_STRUCT_FLAG_BITS + * of work_struct->data are used for flags and the remaining high bits + * point to the pwq; thus, pwqs need to be aligned at two's power of the + * number of flag bits. */ -struct cpu_workqueue_struct { +struct pool_workqueue { struct worker_pool *pool; /* I: the associated pool */ struct workqueue_struct *wq; /* I: the owning workqueue */ int work_color; /* L: current color */ @@ -241,16 +208,16 @@ typedef unsigned long mayday_mask_t; struct workqueue_struct { unsigned int flags; /* W: WQ_* flags */ union { - struct cpu_workqueue_struct __percpu *pcpu; - struct cpu_workqueue_struct *single; + struct pool_workqueue __percpu *pcpu; + struct pool_workqueue *single; unsigned long v; - } cpu_wq; /* I: cwq's */ + } pool_wq; /* I: pwq's */ struct list_head list; /* W: list of all workqueues */ struct mutex flush_mutex; /* protects wq flushing */ int work_color; /* F: current work color */ int flush_color; /* F: current flush color */ - atomic_t nr_cwqs_to_flush; /* flush in progress */ + atomic_t nr_pwqs_to_flush; /* flush in progress */ struct wq_flusher *first_flusher; /* F: first flusher */ struct list_head flusher_queue; /* F: flush waiters */ struct list_head flusher_overflow; /* F: flush overflow list */ @@ -259,7 +226,7 @@ struct workqueue_struct { struct worker *rescuer; /* I: rescue worker */ int nr_drainers; /* W: drain in progress */ - int saved_max_active; /* W: saved cwq max_active */ + int saved_max_active; /* W: saved pwq max_active */ #ifdef CONFIG_LOCKDEP struct lockdep_map lockdep_map; #endif @@ -280,16 +247,15 @@ EXPORT_SYMBOL_GPL(system_freezable_wq); #define CREATE_TRACE_POINTS #include <trace/events/workqueue.h> -#define for_each_worker_pool(pool, gcwq) \ - for ((pool) = &(gcwq)->pools[0]; \ - (pool) < &(gcwq)->pools[NR_WORKER_POOLS]; (pool)++) +#define for_each_std_worker_pool(pool, cpu) \ + for ((pool) = &std_worker_pools(cpu)[0]; \ + (pool) < &std_worker_pools(cpu)[NR_STD_WORKER_POOLS]; (pool)++) -#define for_each_busy_worker(worker, i, pos, gcwq) \ - for (i = 0; i < BUSY_WORKER_HASH_SIZE; i++) \ - hlist_for_each_entry(worker, pos, &gcwq->busy_hash[i], hentry) +#define for_each_busy_worker(worker, i, pos, pool) \ + hash_for_each(pool->busy_hash, i, pos, worker, hentry) -static inline int __next_gcwq_cpu(int cpu, const struct cpumask *mask, - unsigned int sw) +static inline int __next_wq_cpu(int cpu, const struct cpumask *mask, + unsigned int sw) { if (cpu < nr_cpu_ids) { if (sw & 1) { @@ -300,42 +266,42 @@ static inline int __next_gcwq_cpu(int cpu, const struct cpumask *mask, if (sw & 2) return WORK_CPU_UNBOUND; } - return WORK_CPU_NONE; + return WORK_CPU_END; } -static inline int __next_wq_cpu(int cpu, const struct cpumask *mask, - struct workqueue_struct *wq) +static inline int __next_pwq_cpu(int cpu, const struct cpumask *mask, + struct workqueue_struct *wq) { - return __next_gcwq_cpu(cpu, mask, !(wq->flags & WQ_UNBOUND) ? 1 : 2); + return __next_wq_cpu(cpu, mask, !(wq->flags & WQ_UNBOUND) ? 1 : 2); } /* * CPU iterators * - * An extra gcwq is defined for an invalid cpu number + * An extra cpu number is defined using an invalid cpu number * (WORK_CPU_UNBOUND) to host workqueues which are not bound to any - * specific CPU. The following iterators are similar to - * for_each_*_cpu() iterators but also considers the unbound gcwq. + * specific CPU. The following iterators are similar to for_each_*_cpu() + * iterators but also considers the unbound CPU. * - * for_each_gcwq_cpu() : possible CPUs + WORK_CPU_UNBOUND - * for_each_online_gcwq_cpu() : online CPUs + WORK_CPU_UNBOUND - * for_each_cwq_cpu() : possible CPUs for bound workqueues, + * for_each_wq_cpu() : possible CPUs + WORK_CPU_UNBOUND + * for_each_online_wq_cpu() : online CPUs + WORK_CPU_UNBOUND + * for_each_pwq_cpu() : possible CPUs for bound workqueues, * WORK_CPU_UNBOUND for unbound workqueues */ -#define for_each_gcwq_cpu(cpu) \ - for ((cpu) = __next_gcwq_cpu(-1, cpu_possible_mask, 3); \ - (cpu) < WORK_CPU_NONE; \ - (cpu) = __next_gcwq_cpu((cpu), cpu_possible_mask, 3)) +#define for_each_wq_cpu(cpu) \ + for ((cpu) = __next_wq_cpu(-1, cpu_possible_mask, 3); \ + (cpu) < WORK_CPU_END; \ + (cpu) = __next_wq_cpu((cpu), cpu_possible_mask, 3)) -#define for_each_online_gcwq_cpu(cpu) \ - for ((cpu) = __next_gcwq_cpu(-1, cpu_online_mask, 3); \ - (cpu) < WORK_CPU_NONE; \ - (cpu) = __next_gcwq_cpu((cpu), cpu_online_mask, 3)) +#define for_each_online_wq_cpu(cpu) \ + for ((cpu) = __next_wq_cpu(-1, cpu_online_mask, 3); \ + (cpu) < WORK_CPU_END; \ + (cpu) = __next_wq_cpu((cpu), cpu_online_mask, 3)) -#define for_each_cwq_cpu(cpu, wq) \ - for ((cpu) = __next_wq_cpu(-1, cpu_possible_mask, (wq)); \ - (cpu) < WORK_CPU_NONE; \ - (cpu) = __next_wq_cpu((cpu), cpu_possible_mask, (wq))) +#define for_each_pwq_cpu(cpu, wq) \ + for ((cpu) = __next_pwq_cpu(-1, cpu_possible_mask, (wq)); \ + (cpu) < WORK_CPU_END; \ + (cpu) = __next_pwq_cpu((cpu), cpu_possible_mask, (wq))) #ifdef CONFIG_DEBUG_OBJECTS_WORK @@ -459,57 +425,69 @@ static LIST_HEAD(workqueues); static bool workqueue_freezing; /* W: have wqs started freezing? */ /* - * The almighty global cpu workqueues. nr_running is the only field - * which is expected to be used frequently by other cpus via - * try_to_wake_up(). Put it in a separate cacheline. + * The CPU and unbound standard worker pools. The unbound ones have + * POOL_DISASSOCIATED set, and their workers have WORKER_UNBOUND set. */ -static DEFINE_PER_CPU(struct global_cwq, global_cwq); -static DEFINE_PER_CPU_SHARED_ALIGNED(atomic_t, pool_nr_running[NR_WORKER_POOLS]); +static DEFINE_PER_CPU_SHARED_ALIGNED(struct worker_pool [NR_STD_WORKER_POOLS], + cpu_std_worker_pools); +static struct worker_pool unbound_std_worker_pools[NR_STD_WORKER_POOLS]; -/* - * Global cpu workqueue and nr_running counter for unbound gcwq. The - * gcwq is always online, has GCWQ_DISASSOCIATED set, and all its - * workers have WORKER_UNBOUND set. - */ -static struct global_cwq unbound_global_cwq; -static atomic_t unbound_pool_nr_running[NR_WORKER_POOLS] = { - [0 ... NR_WORKER_POOLS - 1] = ATOMIC_INIT(0), /* always 0 */ -}; +/* idr of all pools */ +static DEFINE_MUTEX(worker_pool_idr_mutex); +static DEFINE_IDR(worker_pool_idr); static int worker_thread(void *__worker); -static int worker_pool_pri(struct worker_pool *pool) +static struct worker_pool *std_worker_pools(int cpu) { - return pool - pool->gcwq->pools; + if (cpu != WORK_CPU_UNBOUND) + return per_cpu(cpu_std_worker_pools, cpu); + else + return unbound_std_worker_pools; } -static struct global_cwq *get_gcwq(unsigned int cpu) +static int std_worker_pool_pri(struct worker_pool *pool) { - if (cpu != WORK_CPU_UNBOUND) - return &per_cpu(global_cwq, cpu); - else - return &unbound_global_cwq; + return pool - std_worker_pools(pool->cpu); } -static atomic_t *get_pool_nr_running(struct worker_pool *pool) +/* allocate ID and assign it to @pool */ +static int worker_pool_assign_id(struct worker_pool *pool) { - int cpu = pool->gcwq->cpu; - int idx = worker_pool_pri(pool); + int ret; - if (cpu != WORK_CPU_UNBOUND) - return &per_cpu(pool_nr_running, cpu)[idx]; - else - return &unbound_pool_nr_running[idx]; + mutex_lock(&worker_pool_idr_mutex); + idr_pre_get(&worker_pool_idr, GFP_KERNEL); + ret = idr_get_new(&worker_pool_idr, pool, &pool->id); + mutex_unlock(&worker_pool_idr_mutex); + + return ret; } -static struct cpu_workqueue_struct *get_cwq(unsigned int cpu, - struct workqueue_struct *wq) +/* + * Lookup worker_pool by id. The idr currently is built during boot and + * never modified. Don't worry about locking for now. + */ +static struct worker_pool *worker_pool_by_id(int pool_id) +{ + return idr_find(&worker_pool_idr, pool_id); +} + +static struct worker_pool *get_std_worker_pool(int cpu, bool highpri) +{ + struct worker_pool *pools = std_worker_pools(cpu); + + return &pools[highpri]; +} + +static struct pool_workqueue *get_pwq(unsigned int cpu, + struct workqueue_struct *wq) { if (!(wq->flags & WQ_UNBOUND)) { if (likely(cpu < nr_cpu_ids)) - return per_cpu_ptr(wq->cpu_wq.pcpu, cpu); + return per_cpu_ptr(wq->pool_wq.pcpu, cpu); } else if (likely(cpu == WORK_CPU_UNBOUND)) - return wq->cpu_wq.single; + return wq->pool_wq.single; return NULL; } @@ -530,19 +508,19 @@ static int work_next_color(int color) } /* - * While queued, %WORK_STRUCT_CWQ is set and non flag bits of a work's data - * contain the pointer to the queued cwq. Once execution starts, the flag - * is cleared and the high bits contain OFFQ flags and CPU number. + * While queued, %WORK_STRUCT_PWQ is set and non flag bits of a work's data + * contain the pointer to the queued pwq. Once execution starts, the flag + * is cleared and the high bits contain OFFQ flags and pool ID. * - * set_work_cwq(), set_work_cpu_and_clear_pending(), mark_work_canceling() - * and clear_work_data() can be used to set the cwq, cpu or clear + * set_work_pwq(), set_work_pool_and_clear_pending(), mark_work_canceling() + * and clear_work_data() can be used to set the pwq, pool or clear * work->data. These functions should only be called while the work is * owned - ie. while the PENDING bit is set. * - * get_work_[g]cwq() can be used to obtain the gcwq or cwq corresponding to - * a work. gcwq is available once the work has been queued anywhere after - * initialization until it is sync canceled. cwq is available only while - * the work item is queued. + * get_work_pool() and get_work_pwq() can be used to obtain the pool or pwq + * corresponding to a work. Pool is available once the work has been + * queued anywhere after initialization until it is sync canceled. pwq is + * available only while the work item is queued. * * %WORK_OFFQ_CANCELING is used to mark a work item which is being * canceled. While being canceled, a work item may have its PENDING set @@ -556,16 +534,22 @@ static inline void set_work_data(struct work_struct *work, unsigned long data, atomic_long_set(&work->data, data | flags | work_static(work)); } -static void set_work_cwq(struct work_struct *work, - struct cpu_workqueue_struct *cwq, +static void set_work_pwq(struct work_struct *work, struct pool_workqueue *pwq, unsigned long extra_flags) { - set_work_data(work, (unsigned long)cwq, - WORK_STRUCT_PENDING | WORK_STRUCT_CWQ | extra_flags); + set_work_data(work, (unsigned long)pwq, + WORK_STRUCT_PENDING | WORK_STRUCT_PWQ | extra_flags); } -static void set_work_cpu_and_clear_pending(struct work_struct *work, - unsigned int cpu) +static void set_work_pool_and_keep_pending(struct work_struct *work, + int pool_id) +{ + set_work_data(work, (unsigned long)pool_id << WORK_OFFQ_POOL_SHIFT, + WORK_STRUCT_PENDING); +} + +static void set_work_pool_and_clear_pending(struct work_struct *work, + int pool_id) { /* * The following wmb is paired with the implied mb in @@ -574,67 +558,92 @@ static void set_work_cpu_and_clear_pending(struct work_struct *work, * owner. */ smp_wmb(); - set_work_data(work, (unsigned long)cpu << WORK_OFFQ_CPU_SHIFT, 0); + set_work_data(work, (unsigned long)pool_id << WORK_OFFQ_POOL_SHIFT, 0); } static void clear_work_data(struct work_struct *work) { - smp_wmb(); /* see set_work_cpu_and_clear_pending() */ - set_work_data(work, WORK_STRUCT_NO_CPU, 0); + smp_wmb(); /* see set_work_pool_and_clear_pending() */ + set_work_data(work, WORK_STRUCT_NO_POOL, 0); } -static struct cpu_workqueue_struct *get_work_cwq(struct work_struct *work) +static struct pool_workqueue *get_work_pwq(struct work_struct *work) { unsigned long data = atomic_long_read(&work->data); - if (data & WORK_STRUCT_CWQ) + if (data & WORK_STRUCT_PWQ) return (void *)(data & WORK_STRUCT_WQ_DATA_MASK); else return NULL; } -static struct global_cwq *get_work_gcwq(struct work_struct *work) +/** + * get_work_pool - return the worker_pool a given work was associated with + * @work: the work item of interest + * + * Return the worker_pool @work was last associated with. %NULL if none. + */ +static struct worker_pool *get_work_pool(struct work_struct *work) { unsigned long data = atomic_long_read(&work->data); - unsigned int cpu; + struct worker_pool *pool; + int pool_id; - if (data & WORK_STRUCT_CWQ) - return ((struct cpu_workqueue_struct *) - (data & WORK_STRUCT_WQ_DATA_MASK))->pool->gcwq; + if (data & WORK_STRUCT_PWQ) + return ((struct pool_workqueue *) + (data & WORK_STRUCT_WQ_DATA_MASK))->pool; - cpu = data >> WORK_OFFQ_CPU_SHIFT; - if (cpu == WORK_CPU_NONE) + pool_id = data >> WORK_OFFQ_POOL_SHIFT; + if (pool_id == WORK_OFFQ_POOL_NONE) return NULL; - BUG_ON(cpu >= nr_cpu_ids && cpu != WORK_CPU_UNBOUND); - return get_gcwq(cpu); + pool = worker_pool_by_id(pool_id); + WARN_ON_ONCE(!pool); + return pool; +} + +/** + * get_work_pool_id - return the worker pool ID a given work is associated with + * @work: the work item of interest + * + * Return the worker_pool ID @work was last associated with. + * %WORK_OFFQ_POOL_NONE if none. + */ +static int get_work_pool_id(struct work_struct *work) +{ + unsigned long data = atomic_long_read(&work->data); + + if (data & WORK_STRUCT_PWQ) + return ((struct pool_workqueue *) + (data & WORK_STRUCT_WQ_DATA_MASK))->pool->id; + + return data >> WORK_OFFQ_POOL_SHIFT; } static void mark_work_canceling(struct work_struct *work) { - struct global_cwq *gcwq = get_work_gcwq(work); - unsigned long cpu = gcwq ? gcwq->cpu : WORK_CPU_NONE; + unsigned long pool_id = get_work_pool_id(work); - set_work_data(work, (cpu << WORK_OFFQ_CPU_SHIFT) | WORK_OFFQ_CANCELING, - WORK_STRUCT_PENDING); + pool_id <<= WORK_OFFQ_POOL_SHIFT; + set_work_data(work, pool_id | WORK_OFFQ_CANCELING, WORK_STRUCT_PENDING); } static bool work_is_canceling(struct work_struct *work) { unsigned long data = atomic_long_read(&work->data); - return !(data & WORK_STRUCT_CWQ) && (data & WORK_OFFQ_CANCELING); + return !(data & WORK_STRUCT_PWQ) && (data & WORK_OFFQ_CANCELING); } /* * Policy functions. These define the policies on how the global worker * pools are managed. Unless noted otherwise, these functions assume that - * they're being called with gcwq->lock held. + * they're being called with pool->lock held. */ static bool __need_more_worker(struct worker_pool *pool) { - return !atomic_read(get_pool_nr_running(pool)); + return !atomic_read(&pool->nr_running); } /* @@ -642,7 +651,7 @@ static bool __need_more_worker(struct worker_pool *pool) * running workers. * * Note that, because unbound workers never contribute to nr_running, this - * function will always return %true for unbound gcwq as long as the + * function will always return %true for unbound pools as long as the * worklist isn't empty. */ static bool need_more_worker(struct worker_pool *pool) @@ -659,9 +668,8 @@ static bool may_start_working(struct worker_pool *pool) /* Do I need to keep working? Called from currently running workers. */ static bool keep_working(struct worker_pool *pool) { - atomic_t *nr_running = get_pool_nr_running(pool); - - return !list_empty(&pool->worklist) && atomic_read(nr_running) <= 1; + return !list_empty(&pool->worklist) && + atomic_read(&pool->nr_running) <= 1; } /* Do we need a new worker? Called from manager. */ @@ -714,7 +722,7 @@ static struct worker *first_worker(struct worker_pool *pool) * Wake up the first idle worker of @pool. * * CONTEXT: - * spin_lock_irq(gcwq->lock). + * spin_lock_irq(pool->lock). */ static void wake_up_worker(struct worker_pool *pool) { @@ -740,8 +748,8 @@ void wq_worker_waking_up(struct task_struct *task, unsigned int cpu) struct worker *worker = kthread_data(task); if (!(worker->flags & WORKER_NOT_RUNNING)) { - WARN_ON_ONCE(worker->pool->gcwq->cpu != cpu); - atomic_inc(get_pool_nr_running(worker->pool)); + WARN_ON_ONCE(worker->pool->cpu != cpu); + atomic_inc(&worker->pool->nr_running); } } @@ -764,12 +772,18 @@ struct task_struct *wq_worker_sleeping(struct task_struct *task, unsigned int cpu) { struct worker *worker = kthread_data(task), *to_wakeup = NULL; - struct worker_pool *pool = worker->pool; - atomic_t *nr_running = get_pool_nr_running(pool); + struct worker_pool *pool; + /* + * Rescuers, which may not have all the fields set up like normal + * workers, also reach here, let's not access anything before + * checking NOT_RUNNING. + */ if (worker->flags & WORKER_NOT_RUNNING) return NULL; + pool = worker->pool; + /* this can only happen on the local cpu */ BUG_ON(cpu != raw_smp_processor_id()); @@ -781,10 +795,11 @@ struct task_struct *wq_worker_sleeping(struct task_struct *task, * NOT_RUNNING is clear. This means that we're bound to and * running on the local cpu w/ rq lock held and preemption * disabled, which in turn means that none else could be - * manipulating idle_list, so dereferencing idle_list without gcwq + * manipulating idle_list, so dereferencing idle_list without pool * lock is safe. */ - if (atomic_dec_and_test(nr_running) && !list_empty(&pool->worklist)) + if (atomic_dec_and_test(&pool->nr_running) && + !list_empty(&pool->worklist)) to_wakeup = first_worker(pool); return to_wakeup ? to_wakeup->task : NULL; } @@ -800,7 +815,7 @@ struct task_struct *wq_worker_sleeping(struct task_struct *task, * woken up. * * CONTEXT: - * spin_lock_irq(gcwq->lock) + * spin_lock_irq(pool->lock) */ static inline void worker_set_flags(struct worker *worker, unsigned int flags, bool wakeup) @@ -816,14 +831,12 @@ static inline void worker_set_flags(struct worker *worker, unsigned int flags, */ if ((flags & WORKER_NOT_RUNNING) && !(worker->flags & WORKER_NOT_RUNNING)) { - atomic_t *nr_running = get_pool_nr_running(pool); - if (wakeup) { - if (atomic_dec_and_test(nr_running) && + if (atomic_dec_and_test(&pool->nr_running) && !list_empty(&pool->worklist)) wake_up_worker(pool); } else - atomic_dec(nr_running); + atomic_dec(&pool->nr_running); } worker->flags |= flags; @@ -837,7 +850,7 @@ static inline void worker_set_flags(struct worker *worker, unsigned int flags, * Clear @flags in @worker->flags and adjust nr_running accordingly. * * CONTEXT: - * spin_lock_irq(gcwq->lock) + * spin_lock_irq(pool->lock) */ static inline void worker_clr_flags(struct worker *worker, unsigned int flags) { @@ -855,87 +868,56 @@ static inline void worker_clr_flags(struct worker *worker, unsigned int flags) */ if ((flags & WORKER_NOT_RUNNING) && (oflags & WORKER_NOT_RUNNING)) if (!(worker->flags & WORKER_NOT_RUNNING)) - atomic_inc(get_pool_nr_running(pool)); + atomic_inc(&pool->nr_running); } /** - * busy_worker_head - return the busy hash head for a work - * @gcwq: gcwq of interest - * @work: work to be hashed - * - * Return hash head of @gcwq for @work. - * - * CONTEXT: - * spin_lock_irq(gcwq->lock). - * - * RETURNS: - * Pointer to the hash head. - */ -static struct hlist_head *busy_worker_head(struct global_cwq *gcwq, - struct work_struct *work) -{ - const int base_shift = ilog2(sizeof(struct work_struct)); - unsigned long v = (unsigned long)work; - - /* simple shift and fold hash, do we need something better? */ - v >>= base_shift; - v += v >> BUSY_WORKER_HASH_ORDER; - v &= BUSY_WORKER_HASH_MASK; - - return &gcwq->busy_hash[v]; -} - -/** - * __find_worker_executing_work - find worker which is executing a work - * @gcwq: gcwq of interest - * @bwh: hash head as returned by busy_worker_head() + * find_worker_executing_work - find worker which is executing a work + * @pool: pool of interest * @work: work to find worker for * - * Find a worker which is executing @work on @gcwq. @bwh should be - * the hash head obtained by calling busy_worker_head() with the same - * work. + * Find a worker which is executing @work on @pool by searching + * @pool->busy_hash which is keyed by the address of @work. For a worker + * to match, its current execution should match the address of @work and + * its work function. This is to avoid unwanted dependency between + * unrelated work executions through a work item being recycled while still + * being executed. + * + * This is a bit tricky. A work item may be freed once its execution + * starts and nothing prevents the freed area from being recycled for + * another work item. If the same work item address ends up being reused + * before the original execution finishes, workqueue will identify the + * recycled work item as currently executing and make it wait until the + * current execution finishes, introducing an unwanted dependency. + * + * This function checks the work item address, work function and workqueue + * to avoid false positives. Note that this isn't complete as one may + * construct a work function which can introduce dependency onto itself + * through a recycled work item. Well, if somebody wants to shoot oneself + * in the foot that badly, there's only so much we can do, and if such + * deadlock actually occurs, it should be easy to locate the culprit work + * function. * * CONTEXT: - * spin_lock_irq(gcwq->lock). + * spin_lock_irq(pool->lock). * * RETURNS: * Pointer to worker which is executing @work if found, NULL * otherwise. */ -static struct worker *__find_worker_executing_work(struct global_cwq *gcwq, - struct hlist_head *bwh, - struct work_struct *work) +static struct worker *find_worker_executing_work(struct worker_pool *pool, + struct work_struct *work) { struct worker *worker; struct hlist_node *tmp; - hlist_for_each_entry(worker, tmp, bwh, hentry) - if (worker->current_work == work) + hash_for_each_possible(pool->busy_hash, worker, tmp, hentry, + (unsigned long)work) + if (worker->current_work == work && + worker->current_func == work->func) return worker; - return NULL; -} -/** - * find_worker_executing_work - find worker which is executing a work - * @gcwq: gcwq of interest - * @work: work to find worker for - * - * Find a worker which is executing @work on @gcwq. This function is - * identical to __find_worker_executing_work() except that this - * function calculates @bwh itself. - * - * CONTEXT: - * spin_lock_irq(gcwq->lock). - * - * RETURNS: - * Pointer to worker which is executing @work if found, NULL - * otherwise. - */ -static struct worker *find_worker_executing_work(struct global_cwq *gcwq, - struct work_struct *work) -{ - return __find_worker_executing_work(gcwq, busy_worker_head(gcwq, work), - work); + return NULL; } /** @@ -953,7 +935,7 @@ static struct worker *find_worker_executing_work(struct global_cwq *gcwq, * nested inside outer list_for_each_entry_safe(). * * CONTEXT: - * spin_lock_irq(gcwq->lock). + * spin_lock_irq(pool->lock). */ static void move_linked_works(struct work_struct *work, struct list_head *head, struct work_struct **nextp) @@ -979,67 +961,67 @@ static void move_linked_works(struct work_struct *work, struct list_head *head, *nextp = n; } -static void cwq_activate_delayed_work(struct work_struct *work) +static void pwq_activate_delayed_work(struct work_struct *work) { - struct cpu_workqueue_struct *cwq = get_work_cwq(work); + struct pool_workqueue *pwq = get_work_pwq(work); trace_workqueue_activate_work(work); - move_linked_works(work, &cwq->pool->worklist, NULL); + move_linked_works(work, &pwq->pool->worklist, NULL); __clear_bit(WORK_STRUCT_DELAYED_BIT, work_data_bits(work)); - cwq->nr_active++; + pwq->nr_active++; } -static void cwq_activate_first_delayed(struct cpu_workqueue_struct *cwq) +static void pwq_activate_first_delayed(struct pool_workqueue *pwq) { - struct work_struct *work = list_first_entry(&cwq->delayed_works, + struct work_struct *work = list_first_entry(&pwq->delayed_works, struct work_struct, entry); - cwq_activate_delayed_work(work); + pwq_activate_delayed_work(work); } /** - * cwq_dec_nr_in_flight - decrement cwq's nr_in_flight - * @cwq: cwq of interest + * pwq_dec_nr_in_flight - decrement pwq's nr_in_flight + * @pwq: pwq of interest * @color: color of work which left the queue * * A work either has completed or is removed from pending queue, - * decrement nr_in_flight of its cwq and handle workqueue flushing. + * decrement nr_in_flight of its pwq and handle workqueue flushing. * * CONTEXT: - * spin_lock_irq(gcwq->lock). + * spin_lock_irq(pool->lock). */ -static void cwq_dec_nr_in_flight(struct cpu_workqueue_struct *cwq, int color) +static void pwq_dec_nr_in_flight(struct pool_workqueue *pwq, int color) { /* ignore uncolored works */ if (color == WORK_NO_COLOR) return; - cwq->nr_in_flight[color]--; + pwq->nr_in_flight[color]--; - cwq->nr_active--; - if (!list_empty(&cwq->delayed_works)) { + pwq->nr_active--; + if (!list_empty(&pwq->delayed_works)) { /* one down, submit a delayed one */ - if (cwq->nr_active < cwq->max_active) - cwq_activate_first_delayed(cwq); + if (pwq->nr_active < pwq->max_active) + pwq_activate_first_delayed(pwq); } /* is flush in progress and are we at the flushing tip? */ - if (likely(cwq->flush_color != color)) + if (likely(pwq->flush_color != color)) return; /* are there still in-flight works? */ - if (cwq->nr_in_flight[color]) + if (pwq->nr_in_flight[color]) return; - /* this cwq is done, clear flush_color */ - cwq->flush_color = -1; + /* this pwq is done, clear flush_color */ + pwq->flush_color = -1; /* - * If this was the last cwq, wake up the first flusher. It + * If this was the last pwq, wake up the first flusher. It * will handle the rest. */ - if (atomic_dec_and_test(&cwq->wq->nr_cwqs_to_flush)) - complete(&cwq->wq->first_flusher->done); + if (atomic_dec_and_test(&pwq->wq->nr_pwqs_to_flush)) + complete(&pwq->wq->first_flusher->done); } /** @@ -1070,7 +1052,8 @@ static void cwq_dec_nr_in_flight(struct cpu_workqueue_struct *cwq, int color) static int try_to_grab_pending(struct work_struct *work, bool is_dwork, unsigned long *flags) { - struct global_cwq *gcwq; + struct worker_pool *pool; + struct pool_workqueue *pwq; local_irq_save(*flags); @@ -1095,41 +1078,43 @@ static int try_to_grab_pending(struct work_struct *work, bool is_dwork, * The queueing is in progress, or it is already queued. Try to * steal it from ->worklist without clearing WORK_STRUCT_PENDING. */ - gcwq = get_work_gcwq(work); - if (!gcwq) + pool = get_work_pool(work); + if (!pool) goto fail; - spin_lock(&gcwq->lock); - if (!list_empty(&work->entry)) { + spin_lock(&pool->lock); + /* + * work->data is guaranteed to point to pwq only while the work + * item is queued on pwq->wq, and both updating work->data to point + * to pwq on queueing and to pool on dequeueing are done under + * pwq->pool->lock. This in turn guarantees that, if work->data + * points to pwq which is associated with a locked pool, the work + * item is currently queued on that pool. + */ + pwq = get_work_pwq(work); + if (pwq && pwq->pool == pool) { + debug_work_deactivate(work); + /* - * This work is queued, but perhaps we locked the wrong gcwq. - * In that case we must see the new value after rmb(), see - * insert_work()->wmb(). + * A delayed work item cannot be grabbed directly because + * it might have linked NO_COLOR work items which, if left + * on the delayed_list, will confuse pwq->nr_active + * management later on and cause stall. Make sure the work + * item is activated before grabbing. */ - smp_rmb(); - if (gcwq == get_work_gcwq(work)) { - debug_work_deactivate(work); + if (*work_data_bits(work) & WORK_STRUCT_DELAYED) + pwq_activate_delayed_work(work); - /* - * A delayed work item cannot be grabbed directly - * because it might have linked NO_COLOR work items - * which, if left on the delayed_list, will confuse - * cwq->nr_active management later on and cause - * stall. Make sure the work item is activated - * before grabbing. - */ - if (*work_data_bits(work) & WORK_STRUCT_DELAYED) - cwq_activate_delayed_work(work); + list_del_init(&work->entry); + pwq_dec_nr_in_flight(get_work_pwq(work), get_work_color(work)); - list_del_init(&work->entry); - cwq_dec_nr_in_flight(get_work_cwq(work), - get_work_color(work)); + /* work->data points to pwq iff queued, point to pool */ + set_work_pool_and_keep_pending(work, pool->id); - spin_unlock(&gcwq->lock); - return 1; - } + spin_unlock(&pool->lock); + return 1; } - spin_unlock(&gcwq->lock); + spin_unlock(&pool->lock); fail: local_irq_restore(*flags); if (work_is_canceling(work)) @@ -1139,33 +1124,25 @@ fail: } /** - * insert_work - insert a work into gcwq - * @cwq: cwq @work belongs to + * insert_work - insert a work into a pool + * @pwq: pwq @work belongs to * @work: work to insert * @head: insertion point * @extra_flags: extra WORK_STRUCT_* flags to set * - * Insert @work which belongs to @cwq into @gcwq after @head. - * @extra_flags is or'd to work_struct flags. + * Insert @work which belongs to @pwq after @head. @extra_flags is or'd to + * work_struct flags. * * CONTEXT: - * spin_lock_irq(gcwq->lock). + * spin_lock_irq(pool->lock). */ -static void insert_work(struct cpu_workqueue_struct *cwq, - struct work_struct *work, struct list_head *head, - unsigned int extra_flags) +static void insert_work(struct pool_workqueue *pwq, struct work_struct *work, + struct list_head *head, unsigned int extra_flags) { - struct worker_pool *pool = cwq->pool; + struct worker_pool *pool = pwq->pool; /* we own @work, set data and link */ - set_work_cwq(work, cwq, extra_flags); - - /* - * Ensure that we get the right work->data if we see the - * result of list_add() below, see try_to_grab_pending(). - */ - smp_wmb(); - + set_work_pwq(work, pwq, extra_flags); list_add_tail(&work->entry, head); /* @@ -1181,41 +1158,24 @@ static void insert_work(struct cpu_workqueue_struct *cwq, /* * Test whether @work is being queued from another work executing on the - * same workqueue. This is rather expensive and should only be used from - * cold paths. + * same workqueue. */ static bool is_chained_work(struct workqueue_struct *wq) { - unsigned long flags; - unsigned int cpu; - - for_each_gcwq_cpu(cpu) { - struct global_cwq *gcwq = get_gcwq(cpu); - struct worker *worker; - struct hlist_node *pos; - int i; + struct worker *worker; - spin_lock_irqsave(&gcwq->lock, flags); - for_each_busy_worker(worker, i, pos, gcwq) { - if (worker->task != current) - continue; - spin_unlock_irqrestore(&gcwq->lock, flags); - /* - * I'm @worker, no locking necessary. See if @work - * is headed to the same workqueue. - */ - return worker->current_cwq->wq == wq; - } - spin_unlock_irqrestore(&gcwq->lock, flags); - } - return false; + worker = current_wq_worker(); + /* + * Return %true iff I'm a worker execuing a work item on @wq. If + * I'm @worker, it's safe to dereference it without locking. + */ + return worker && worker->current_pwq->wq == wq; } static void __queue_work(unsigned int cpu, struct workqueue_struct *wq, struct work_struct *work) { - struct global_cwq *gcwq; - struct cpu_workqueue_struct *cwq; + struct pool_workqueue *pwq; struct list_head *worklist; unsigned int work_flags; unsigned int req_cpu = cpu; @@ -1235,9 +1195,9 @@ static void __queue_work(unsigned int cpu, struct workqueue_struct *wq, WARN_ON_ONCE(!is_chained_work(wq))) return; - /* determine gcwq to use */ + /* determine the pwq to use */ if (!(wq->flags & WQ_UNBOUND)) { - struct global_cwq *last_gcwq; + struct worker_pool *last_pool; if (cpu == WORK_CPU_UNBOUND) cpu = raw_smp_processor_id(); @@ -1248,55 +1208,54 @@ static void __queue_work(unsigned int cpu, struct workqueue_struct *wq, * work needs to be queued on that cpu to guarantee * non-reentrancy. */ - gcwq = get_gcwq(cpu); - last_gcwq = get_work_gcwq(work); + pwq = get_pwq(cpu, wq); + last_pool = get_work_pool(work); - if (last_gcwq && last_gcwq != gcwq) { + if (last_pool && last_pool != pwq->pool) { struct worker *worker; - spin_lock(&last_gcwq->lock); + spin_lock(&last_pool->lock); - worker = find_worker_executing_work(last_gcwq, work); + worker = find_worker_executing_work(last_pool, work); - if (worker && worker->current_cwq->wq == wq) - gcwq = last_gcwq; - else { + if (worker && worker->current_pwq->wq == wq) { + pwq = get_pwq(last_pool->cpu, wq); + } else { /* meh... not running there, queue here */ - spin_unlock(&last_gcwq->lock); - spin_lock(&gcwq->lock); + spin_unlock(&last_pool->lock); + spin_lock(&pwq->pool->lock); } } else { - spin_lock(&gcwq->lock); + spin_lock(&pwq->pool->lock); } } else { - gcwq = get_gcwq(WORK_CPU_UNBOUND); - spin_lock(&gcwq->lock); + pwq = get_pwq(WORK_CPU_UNBOUND, wq); + spin_lock(&pwq->pool->lock); } - /* gcwq determined, get cwq and queue */ - cwq = get_cwq(gcwq->cpu, wq); - trace_workqueue_queue_work(req_cpu, cwq, work); + /* pwq determined, queue */ + trace_workqueue_queue_work(req_cpu, pwq, work); if (WARN_ON(!list_empty(&work->entry))) { - spin_unlock(&gcwq->lock); + spin_unlock(&pwq->pool->lock); return; } - cwq->nr_in_flight[cwq->work_color]++; - work_flags = work_color_to_flags(cwq->work_color); + pwq->nr_in_flight[pwq->work_color]++; + work_flags = work_color_to_flags(pwq->work_color); - if (likely(cwq->nr_active < cwq->max_active)) { + if (likely(pwq->nr_active < pwq->max_active)) { trace_workqueue_activate_work(work); - cwq->nr_active++; - worklist = &cwq->pool->worklist; + pwq->nr_active++; + worklist = &pwq->pool->worklist; } else { work_flags |= WORK_STRUCT_DELAYED; - worklist = &cwq->delayed_works; + worklist = &pwq->delayed_works; } - insert_work(cwq, work, worklist, work_flags); + insert_work(pwq, work, worklist, work_flags); - spin_unlock(&gcwq->lock); + spin_unlock(&pwq->pool->lock); } /** @@ -1347,19 +1306,17 @@ EXPORT_SYMBOL_GPL(queue_work); void delayed_work_timer_fn(unsigned long __data) { struct delayed_work *dwork = (struct delayed_work *)__data; - struct cpu_workqueue_struct *cwq = get_work_cwq(&dwork->work); /* should have been called from irqsafe timer with irq already off */ - __queue_work(dwork->cpu, cwq->wq, &dwork->work); + __queue_work(dwork->cpu, dwork->wq, &dwork->work); } -EXPORT_SYMBOL_GPL(delayed_work_timer_fn); +EXPORT_SYMBOL(delayed_work_timer_fn); static void __queue_delayed_work(int cpu, struct workqueue_struct *wq, struct delayed_work *dwork, unsigned long delay) { struct timer_list *timer = &dwork->timer; struct work_struct *work = &dwork->work; - unsigned int lcpu; WARN_ON_ONCE(timer->function != delayed_work_timer_fn || timer->data != (unsigned long)dwork); @@ -1379,30 +1336,7 @@ static void __queue_delayed_work(int cpu, struct workqueue_struct *wq, timer_stats_timer_set_start_info(&dwork->timer); - /* - * This stores cwq for the moment, for the timer_fn. Note that the - * work's gcwq is preserved to allow reentrance detection for - * delayed works. - */ - if (!(wq->flags & WQ_UNBOUND)) { - struct global_cwq *gcwq = get_work_gcwq(work); - - /* - * If we cannot get the last gcwq from @work directly, - * select the last CPU such that it avoids unnecessarily - * triggering non-reentrancy check in __queue_work(). - */ - lcpu = cpu; - if (gcwq) - lcpu = gcwq->cpu; - if (lcpu == WORK_CPU_UNBOUND) - lcpu = raw_smp_processor_id(); - } else { - lcpu = WORK_CPU_UNBOUND; - } - - set_work_cwq(work, get_cwq(lcpu, wq), 0); - + dwork->wq = wq; dwork->cpu = cpu; timer->expires = jiffies + delay; @@ -1519,12 +1453,11 @@ EXPORT_SYMBOL_GPL(mod_delayed_work); * necessary. * * LOCKING: - * spin_lock_irq(gcwq->lock). + * spin_lock_irq(pool->lock). */ static void worker_enter_idle(struct worker *worker) { struct worker_pool *pool = worker->pool; - struct global_cwq *gcwq = pool->gcwq; BUG_ON(worker->flags & WORKER_IDLE); BUG_ON(!list_empty(&worker->entry) && @@ -1542,14 +1475,14 @@ static void worker_enter_idle(struct worker *worker) mod_timer(&pool->idle_timer, jiffies + IDLE_WORKER_TIMEOUT); /* - * Sanity check nr_running. Because gcwq_unbind_fn() releases - * gcwq->lock between setting %WORKER_UNBOUND and zapping + * Sanity check nr_running. Because wq_unbind_fn() releases + * pool->lock between setting %WORKER_UNBOUND and zapping * nr_running, the warning may trigger spuriously. Check iff * unbind is not in progress. */ - WARN_ON_ONCE(!(gcwq->flags & GCWQ_DISASSOCIATED) && + WARN_ON_ONCE(!(pool->flags & POOL_DISASSOCIATED) && pool->nr_workers == pool->nr_idle && - atomic_read(get_pool_nr_running(pool))); + atomic_read(&pool->nr_running)); } /** @@ -1559,7 +1492,7 @@ static void worker_enter_idle(struct worker *worker) * @worker is leaving idle state. Update stats. * * LOCKING: - * spin_lock_irq(gcwq->lock). + * spin_lock_irq(pool->lock). */ static void worker_leave_idle(struct worker *worker) { @@ -1572,7 +1505,7 @@ static void worker_leave_idle(struct worker *worker) } /** - * worker_maybe_bind_and_lock - bind worker to its cpu if possible and lock gcwq + * worker_maybe_bind_and_lock - bind worker to its cpu if possible and lock pool * @worker: self * * Works which are scheduled while the cpu is online must at least be @@ -1584,27 +1517,27 @@ static void worker_leave_idle(struct worker *worker) * themselves to the target cpu and may race with cpu going down or * coming online. kthread_bind() can't be used because it may put the * worker to already dead cpu and set_cpus_allowed_ptr() can't be used - * verbatim as it's best effort and blocking and gcwq may be + * verbatim as it's best effort and blocking and pool may be * [dis]associated in the meantime. * - * This function tries set_cpus_allowed() and locks gcwq and verifies the - * binding against %GCWQ_DISASSOCIATED which is set during + * This function tries set_cpus_allowed() and locks pool and verifies the + * binding against %POOL_DISASSOCIATED which is set during * %CPU_DOWN_PREPARE and cleared during %CPU_ONLINE, so if the worker * enters idle state or fetches works without dropping lock, it can * guarantee the scheduling requirement described in the first paragraph. * * CONTEXT: - * Might sleep. Called without any lock but returns with gcwq->lock + * Might sleep. Called without any lock but returns with pool->lock * held. * * RETURNS: - * %true if the associated gcwq is online (@worker is successfully + * %true if the associated pool is online (@worker is successfully * bound), %false if offline. */ static bool worker_maybe_bind_and_lock(struct worker *worker) -__acquires(&gcwq->lock) +__acquires(&pool->lock) { - struct global_cwq *gcwq = worker->pool->gcwq; + struct worker_pool *pool = worker->pool; struct task_struct *task = worker->task; while (true) { @@ -1612,19 +1545,19 @@ __acquires(&gcwq->lock) * The following call may fail, succeed or succeed * without actually migrating the task to the cpu if * it races with cpu hotunplug operation. Verify - * against GCWQ_DISASSOCIATED. + * against POOL_DISASSOCIATED. */ - if (!(gcwq->flags & GCWQ_DISASSOCIATED)) - set_cpus_allowed_ptr(task, get_cpu_mask(gcwq->cpu)); + if (!(pool->flags & POOL_DISASSOCIATED)) + set_cpus_allowed_ptr(task, get_cpu_mask(pool->cpu)); - spin_lock_irq(&gcwq->lock); - if (gcwq->flags & GCWQ_DISASSOCIATED) + spin_lock_irq(&pool->lock); + if (pool->flags & POOL_DISASSOCIATED) return false; - if (task_cpu(task) == gcwq->cpu && + if (task_cpu(task) == pool->cpu && cpumask_equal(¤t->cpus_allowed, - get_cpu_mask(gcwq->cpu))) + get_cpu_mask(pool->cpu))) return true; - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); /* * We've raced with CPU hot[un]plug. Give it a breather @@ -1643,15 +1576,13 @@ __acquires(&gcwq->lock) */ static void idle_worker_rebind(struct worker *worker) { - struct global_cwq *gcwq = worker->pool->gcwq; - /* CPU may go down again inbetween, clear UNBOUND only on success */ if (worker_maybe_bind_and_lock(worker)) worker_clr_flags(worker, WORKER_UNBOUND); /* rebind complete, become available again */ list_add(&worker->entry, &worker->pool->idle_list); - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&worker->pool->lock); } /* @@ -1663,19 +1594,18 @@ static void idle_worker_rebind(struct worker *worker) static void busy_worker_rebind_fn(struct work_struct *work) { struct worker *worker = container_of(work, struct worker, rebind_work); - struct global_cwq *gcwq = worker->pool->gcwq; if (worker_maybe_bind_and_lock(worker)) worker_clr_flags(worker, WORKER_UNBOUND); - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&worker->pool->lock); } /** - * rebind_workers - rebind all workers of a gcwq to the associated CPU - * @gcwq: gcwq of interest + * rebind_workers - rebind all workers of a pool to the associated CPU + * @pool: pool of interest * - * @gcwq->cpu is coming online. Rebind all workers to the CPU. Rebinding + * @pool->cpu is coming online. Rebind all workers to the CPU. Rebinding * is different for idle and busy ones. * * Idle ones will be removed from the idle_list and woken up. They will @@ -1693,38 +1623,32 @@ static void busy_worker_rebind_fn(struct work_struct *work) * including the manager will not appear on @idle_list until rebind is * complete, making local wake-ups safe. */ -static void rebind_workers(struct global_cwq *gcwq) +static void rebind_workers(struct worker_pool *pool) { - struct worker_pool *pool; struct worker *worker, *n; struct hlist_node *pos; int i; - lockdep_assert_held(&gcwq->lock); - - for_each_worker_pool(pool, gcwq) - lockdep_assert_held(&pool->assoc_mutex); + lockdep_assert_held(&pool->assoc_mutex); + lockdep_assert_held(&pool->lock); /* dequeue and kick idle ones */ - for_each_worker_pool(pool, gcwq) { - list_for_each_entry_safe(worker, n, &pool->idle_list, entry) { - /* - * idle workers should be off @pool->idle_list - * until rebind is complete to avoid receiving - * premature local wake-ups. - */ - list_del_init(&worker->entry); + list_for_each_entry_safe(worker, n, &pool->idle_list, entry) { + /* + * idle workers should be off @pool->idle_list until rebind + * is complete to avoid receiving premature local wake-ups. + */ + list_del_init(&worker->entry); - /* - * worker_thread() will see the above dequeuing - * and call idle_worker_rebind(). - */ - wake_up_process(worker->task); - } + /* + * worker_thread() will see the above dequeuing and call + * idle_worker_rebind(). + */ + wake_up_process(worker->task); } /* rebind busy workers */ - for_each_busy_worker(worker, i, pos, gcwq) { + for_each_busy_worker(worker, i, pos, pool) { struct work_struct *rebind_work = &worker->rebind_work; struct workqueue_struct *wq; @@ -1736,16 +1660,16 @@ static void rebind_workers(struct global_cwq *gcwq) /* * wq doesn't really matter but let's keep @worker->pool - * and @cwq->pool consistent for sanity. + * and @pwq->pool consistent for sanity. */ - if (worker_pool_pri(worker->pool)) + if (std_worker_pool_pri(worker->pool)) wq = system_highpri_wq; else wq = system_wq; - insert_work(get_cwq(gcwq->cpu, wq), rebind_work, - worker->scheduled.next, - work_color_to_flags(WORK_NO_COLOR)); + insert_work(get_pwq(pool->cpu, wq), rebind_work, + worker->scheduled.next, + work_color_to_flags(WORK_NO_COLOR)); } } @@ -1780,19 +1704,18 @@ static struct worker *alloc_worker(void) */ static struct worker *create_worker(struct worker_pool *pool) { - struct global_cwq *gcwq = pool->gcwq; - const char *pri = worker_pool_pri(pool) ? "H" : ""; + const char *pri = std_worker_pool_pri(pool) ? "H" : ""; struct worker *worker = NULL; int id = -1; - spin_lock_irq(&gcwq->lock); + spin_lock_irq(&pool->lock); while (ida_get_new(&pool->worker_ida, &id)) { - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); if (!ida_pre_get(&pool->worker_ida, GFP_KERNEL)) goto fail; - spin_lock_irq(&gcwq->lock); + spin_lock_irq(&pool->lock); } - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); worker = alloc_worker(); if (!worker) @@ -1801,30 +1724,30 @@ static struct worker *create_worker(struct worker_pool *pool) worker->pool = pool; worker->id = id; - if (gcwq->cpu != WORK_CPU_UNBOUND) + if (pool->cpu != WORK_CPU_UNBOUND) worker->task = kthread_create_on_node(worker_thread, - worker, cpu_to_node(gcwq->cpu), - "kworker/%u:%d%s", gcwq->cpu, id, pri); + worker, cpu_to_node(pool->cpu), + "kworker/%u:%d%s", pool->cpu, id, pri); else worker->task = kthread_create(worker_thread, worker, "kworker/u:%d%s", id, pri); if (IS_ERR(worker->task)) goto fail; - if (worker_pool_pri(pool)) + if (std_worker_pool_pri(pool)) set_user_nice(worker->task, HIGHPRI_NICE_LEVEL); /* * Determine CPU binding of the new worker depending on - * %GCWQ_DISASSOCIATED. The caller is responsible for ensuring the + * %POOL_DISASSOCIATED. The caller is responsible for ensuring the * flag remains stable across this function. See the comments * above the flag definition for details. * * As an unbound worker may later become a regular one if CPU comes * online, make sure every worker has %PF_THREAD_BOUND set. */ - if (!(gcwq->flags & GCWQ_DISASSOCIATED)) { - kthread_bind(worker->task, gcwq->cpu); + if (!(pool->flags & POOL_DISASSOCIATED)) { + kthread_bind(worker->task, pool->cpu); } else { worker->task->flags |= PF_THREAD_BOUND; worker->flags |= WORKER_UNBOUND; @@ -1833,9 +1756,9 @@ static struct worker *create_worker(struct worker_pool *pool) return worker; fail: if (id >= 0) { - spin_lock_irq(&gcwq->lock); + spin_lock_irq(&pool->lock); ida_remove(&pool->worker_ida, id); - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); } kfree(worker); return NULL; @@ -1845,10 +1768,10 @@ fail: * start_worker - start a newly created worker * @worker: worker to start * - * Make the gcwq aware of @worker and start it. + * Make the pool aware of @worker and start it. * * CONTEXT: - * spin_lock_irq(gcwq->lock). + * spin_lock_irq(pool->lock). */ static void start_worker(struct worker *worker) { @@ -1862,15 +1785,14 @@ static void start_worker(struct worker *worker) * destroy_worker - destroy a workqueue worker * @worker: worker to be destroyed * - * Destroy @worker and adjust @gcwq stats accordingly. + * Destroy @worker and adjust @pool stats accordingly. * * CONTEXT: - * spin_lock_irq(gcwq->lock) which is released and regrabbed. + * spin_lock_irq(pool->lock) which is released and regrabbed. */ static void destroy_worker(struct worker *worker) { struct worker_pool *pool = worker->pool; - struct global_cwq *gcwq = pool->gcwq; int id = worker->id; /* sanity check frenzy */ @@ -1885,21 +1807,20 @@ static void destroy_worker(struct worker *worker) list_del_init(&worker->entry); worker->flags |= WORKER_DIE; - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); kthread_stop(worker->task); kfree(worker); - spin_lock_irq(&gcwq->lock); + spin_lock_irq(&pool->lock); ida_remove(&pool->worker_ida, id); } static void idle_worker_timeout(unsigned long __pool) { struct worker_pool *pool = (void *)__pool; - struct global_cwq *gcwq = pool->gcwq; - spin_lock_irq(&gcwq->lock); + spin_lock_irq(&pool->lock); if (too_many_workers(pool)) { struct worker *worker; @@ -1918,20 +1839,20 @@ static void idle_worker_timeout(unsigned long __pool) } } - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); } static bool send_mayday(struct work_struct *work) { - struct cpu_workqueue_struct *cwq = get_work_cwq(work); - struct workqueue_struct *wq = cwq->wq; + struct pool_workqueue *pwq = get_work_pwq(work); + struct workqueue_struct *wq = pwq->wq; unsigned int cpu; if (!(wq->flags & WQ_RESCUER)) return false; /* mayday mayday mayday */ - cpu = cwq->pool->gcwq->cpu; + cpu = pwq->pool->cpu; /* WORK_CPU_UNBOUND can't be set in cpumask, use cpu 0 instead */ if (cpu == WORK_CPU_UNBOUND) cpu = 0; @@ -1940,13 +1861,12 @@ static bool send_mayday(struct work_struct *work) return true; } -static void gcwq_mayday_timeout(unsigned long __pool) +static void pool_mayday_timeout(unsigned long __pool) { struct worker_pool *pool = (void *)__pool; - struct global_cwq *gcwq = pool->gcwq; struct work_struct *work; - spin_lock_irq(&gcwq->lock); + spin_lock_irq(&pool->lock); if (need_to_create_worker(pool)) { /* @@ -1959,7 +1879,7 @@ static void gcwq_mayday_timeout(unsigned long __pool) send_mayday(work); } - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); mod_timer(&pool->mayday_timer, jiffies + MAYDAY_INTERVAL); } @@ -1978,24 +1898,22 @@ static void gcwq_mayday_timeout(unsigned long __pool) * may_start_working() true. * * LOCKING: - * spin_lock_irq(gcwq->lock) which may be released and regrabbed + * spin_lock_irq(pool->lock) which may be released and regrabbed * multiple times. Does GFP_KERNEL allocations. Called only from * manager. * * RETURNS: - * false if no action was taken and gcwq->lock stayed locked, true + * false if no action was taken and pool->lock stayed locked, true * otherwise. */ static bool maybe_create_worker(struct worker_pool *pool) -__releases(&gcwq->lock) -__acquires(&gcwq->lock) +__releases(&pool->lock) +__acquires(&pool->lock) { - struct global_cwq *gcwq = pool->gcwq; - if (!need_to_create_worker(pool)) return false; restart: - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); /* if we don't make progress in MAYDAY_INITIAL_TIMEOUT, call for help */ mod_timer(&pool->mayday_timer, jiffies + MAYDAY_INITIAL_TIMEOUT); @@ -2006,7 +1924,7 @@ restart: worker = create_worker(pool); if (worker) { del_timer_sync(&pool->mayday_timer); - spin_lock_irq(&gcwq->lock); + spin_lock_irq(&pool->lock); start_worker(worker); BUG_ON(need_to_create_worker(pool)); return true; @@ -2023,7 +1941,7 @@ restart: } del_timer_sync(&pool->mayday_timer); - spin_lock_irq(&gcwq->lock); + spin_lock_irq(&pool->lock); if (need_to_create_worker(pool)) goto restart; return true; @@ -2037,11 +1955,11 @@ restart: * IDLE_WORKER_TIMEOUT. * * LOCKING: - * spin_lock_irq(gcwq->lock) which may be released and regrabbed + * spin_lock_irq(pool->lock) which may be released and regrabbed * multiple times. Called only from manager. * * RETURNS: - * false if no action was taken and gcwq->lock stayed locked, true + * false if no action was taken and pool->lock stayed locked, true * otherwise. */ static bool maybe_destroy_workers(struct worker_pool *pool) @@ -2071,21 +1989,21 @@ static bool maybe_destroy_workers(struct worker_pool *pool) * manage_workers - manage worker pool * @worker: self * - * Assume the manager role and manage gcwq worker pool @worker belongs + * Assume the manager role and manage the worker pool @worker belongs * to. At any given time, there can be only zero or one manager per - * gcwq. The exclusion is handled automatically by this function. + * pool. The exclusion is handled automatically by this function. * * The caller can safely start processing works on false return. On * true return, it's guaranteed that need_to_create_worker() is false * and may_start_working() is true. * * CONTEXT: - * spin_lock_irq(gcwq->lock) which may be released and regrabbed + * spin_lock_irq(pool->lock) which may be released and regrabbed * multiple times. Does GFP_KERNEL allocations. * * RETURNS: - * false if no action was taken and gcwq->lock stayed locked, true if - * some action was taken. + * spin_lock_irq(pool->lock) which may be released and regrabbed + * multiple times. Does GFP_KERNEL allocations. */ static bool manage_workers(struct worker *worker) { @@ -2107,20 +2025,20 @@ static bool manage_workers(struct worker *worker) * manager against CPU hotplug. * * assoc_mutex would always be free unless CPU hotplug is in - * progress. trylock first without dropping @gcwq->lock. + * progress. trylock first without dropping @pool->lock. */ if (unlikely(!mutex_trylock(&pool->assoc_mutex))) { - spin_unlock_irq(&pool->gcwq->lock); + spin_unlock_irq(&pool->lock); mutex_lock(&pool->assoc_mutex); /* * CPU hotplug could have happened while we were waiting * for assoc_mutex. Hotplug itself can't handle us * because manager isn't either on idle or busy list, and - * @gcwq's state and ours could have deviated. + * @pool's state and ours could have deviated. * * As hotplug is now excluded via assoc_mutex, we can * simply try to bind. It will succeed or fail depending - * on @gcwq's current state. Try it and adjust + * on @pool's current state. Try it and adjust * %WORKER_UNBOUND accordingly. */ if (worker_maybe_bind_and_lock(worker)) @@ -2157,18 +2075,15 @@ static bool manage_workers(struct worker *worker) * call this function to process a work. * * CONTEXT: - * spin_lock_irq(gcwq->lock) which is released and regrabbed. + * spin_lock_irq(pool->lock) which is released and regrabbed. */ static void process_one_work(struct worker *worker, struct work_struct *work) -__releases(&gcwq->lock) -__acquires(&gcwq->lock) +__releases(&pool->lock) +__acquires(&pool->lock) { - struct cpu_workqueue_struct *cwq = get_work_cwq(work); + struct pool_workqueue *pwq = get_work_pwq(work); struct worker_pool *pool = worker->pool; - struct global_cwq *gcwq = pool->gcwq; - struct hlist_head *bwh = busy_worker_head(gcwq, work); - bool cpu_intensive = cwq->wq->flags & WQ_CPU_INTENSIVE; - work_func_t f = work->func; + bool cpu_intensive = pwq->wq->flags & WQ_CPU_INTENSIVE; int work_color; struct worker *collision; #ifdef CONFIG_LOCKDEP @@ -2186,11 +2101,11 @@ __acquires(&gcwq->lock) /* * Ensure we're on the correct CPU. DISASSOCIATED test is * necessary to avoid spurious warnings from rescuers servicing the - * unbound or a disassociated gcwq. + * unbound or a disassociated pool. */ WARN_ON_ONCE(!(worker->flags & WORKER_UNBOUND) && - !(gcwq->flags & GCWQ_DISASSOCIATED) && - raw_smp_processor_id() != gcwq->cpu); + !(pool->flags & POOL_DISASSOCIATED) && + raw_smp_processor_id() != pool->cpu); /* * A single work shouldn't be executed concurrently by @@ -2198,7 +2113,7 @@ __acquires(&gcwq->lock) * already processing the work. If so, defer the work to the * currently executing one. */ - collision = __find_worker_executing_work(gcwq, bwh, work); + collision = find_worker_executing_work(pool, work); if (unlikely(collision)) { move_linked_works(work, &collision->scheduled, NULL); return; @@ -2206,9 +2121,10 @@ __acquires(&gcwq->lock) /* claim and dequeue */ debug_work_deactivate(work); - hlist_add_head(&worker->hentry, bwh); + hash_add(pool->busy_hash, &worker->hentry, (unsigned long)work); worker->current_work = work; - worker->current_cwq = cwq; + worker->current_func = work->func; + worker->current_pwq = pwq; work_color = get_work_color(work); list_del_init(&work->entry); @@ -2221,53 +2137,55 @@ __acquires(&gcwq->lock) worker_set_flags(worker, WORKER_CPU_INTENSIVE, true); /* - * Unbound gcwq isn't concurrency managed and work items should be + * Unbound pool isn't concurrency managed and work items should be * executed ASAP. Wake up another worker if necessary. */ if ((worker->flags & WORKER_UNBOUND) && need_more_worker(pool)) wake_up_worker(pool); /* - * Record the last CPU and clear PENDING which should be the last - * update to @work. Also, do this inside @gcwq->lock so that + * Record the last pool and clear PENDING which should be the last + * update to @work. Also, do this inside @pool->lock so that * PENDING and queued state changes happen together while IRQ is * disabled. */ - set_work_cpu_and_clear_pending(work, gcwq->cpu); + set_work_pool_and_clear_pending(work, pool->id); - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); - lock_map_acquire_read(&cwq->wq->lockdep_map); + lock_map_acquire_read(&pwq->wq->lockdep_map); lock_map_acquire(&lockdep_map); trace_workqueue_execute_start(work); - f(work); + worker->current_func(work); /* * While we must be careful to not use "work" after this, the trace * point will only record its address. */ trace_workqueue_execute_end(work); lock_map_release(&lockdep_map); - lock_map_release(&cwq->wq->lockdep_map); + lock_map_release(&pwq->wq->lockdep_map); if (unlikely(in_atomic() || lockdep_depth(current) > 0)) { pr_err("BUG: workqueue leaked lock or atomic: %s/0x%08x/%d\n" " last function: %pf\n", - current->comm, preempt_count(), task_pid_nr(current), f); + current->comm, preempt_count(), task_pid_nr(current), + worker->current_func); debug_show_held_locks(current); dump_stack(); } - spin_lock_irq(&gcwq->lock); + spin_lock_irq(&pool->lock); /* clear cpu intensive status */ if (unlikely(cpu_intensive)) worker_clr_flags(worker, WORKER_CPU_INTENSIVE); /* we're done with it, release */ - hlist_del_init(&worker->hentry); + hash_del(&worker->hentry); worker->current_work = NULL; - worker->current_cwq = NULL; - cwq_dec_nr_in_flight(cwq, work_color); + worker->current_func = NULL; + worker->current_pwq = NULL; + pwq_dec_nr_in_flight(pwq, work_color); } /** @@ -2279,7 +2197,7 @@ __acquires(&gcwq->lock) * fetches a work from the top and executes it. * * CONTEXT: - * spin_lock_irq(gcwq->lock) which may be released and regrabbed + * spin_lock_irq(pool->lock) which may be released and regrabbed * multiple times. */ static void process_scheduled_works(struct worker *worker) @@ -2295,8 +2213,8 @@ static void process_scheduled_works(struct worker *worker) * worker_thread - the worker thread function * @__worker: self * - * The gcwq worker thread function. There's a single dynamic pool of - * these per each cpu. These workers process all works regardless of + * The worker thread function. There are NR_CPU_WORKER_POOLS dynamic pools + * of these per each cpu. These workers process all works regardless of * their specific target workqueue. The only exception is works which * belong to workqueues with a rescuer which will be explained in * rescuer_thread(). @@ -2305,16 +2223,15 @@ static int worker_thread(void *__worker) { struct worker *worker = __worker; struct worker_pool *pool = worker->pool; - struct global_cwq *gcwq = pool->gcwq; /* tell the scheduler that this is a workqueue worker */ worker->task->flags |= PF_WQ_WORKER; woke_up: - spin_lock_irq(&gcwq->lock); + spin_lock_irq(&pool->lock); /* we are off idle list if destruction or rebind is requested */ if (unlikely(list_empty(&worker->entry))) { - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); /* if DIE is set, destruction is requested */ if (worker->flags & WORKER_DIE) { @@ -2373,52 +2290,59 @@ sleep: goto recheck; /* - * gcwq->lock is held and there's no work to process and no - * need to manage, sleep. Workers are woken up only while - * holding gcwq->lock or from local cpu, so setting the - * current state before releasing gcwq->lock is enough to - * prevent losing any event. + * pool->lock is held and there's no work to process and no need to + * manage, sleep. Workers are woken up only while holding + * pool->lock or from local cpu, so setting the current state + * before releasing pool->lock is enough to prevent losing any + * event. */ worker_enter_idle(worker); __set_current_state(TASK_INTERRUPTIBLE); - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); schedule(); goto woke_up; } /** * rescuer_thread - the rescuer thread function - * @__wq: the associated workqueue + * @__rescuer: self * * Workqueue rescuer thread function. There's one rescuer for each * workqueue which has WQ_RESCUER set. * - * Regular work processing on a gcwq may block trying to create a new + * Regular work processing on a pool may block trying to create a new * worker which uses GFP_KERNEL allocation which has slight chance of * developing into deadlock if some works currently on the same queue * need to be processed to satisfy the GFP_KERNEL allocation. This is * the problem rescuer solves. * - * When such condition is possible, the gcwq summons rescuers of all - * workqueues which have works queued on the gcwq and let them process + * When such condition is possible, the pool summons rescuers of all + * workqueues which have works queued on the pool and let them process * those works so that forward progress can be guaranteed. * * This should happen rarely. */ -static int rescuer_thread(void *__wq) +static int rescuer_thread(void *__rescuer) { - struct workqueue_struct *wq = __wq; - struct worker *rescuer = wq->rescuer; + struct worker *rescuer = __rescuer; + struct workqueue_struct *wq = rescuer->rescue_wq; struct list_head *scheduled = &rescuer->scheduled; bool is_unbound = wq->flags & WQ_UNBOUND; unsigned int cpu; set_user_nice(current, RESCUER_NICE_LEVEL); + + /* + * Mark rescuer as worker too. As WORKER_PREP is never cleared, it + * doesn't participate in concurrency management. + */ + rescuer->task->flags |= PF_WQ_WORKER; repeat: set_current_state(TASK_INTERRUPTIBLE); if (kthread_should_stop()) { __set_current_state(TASK_RUNNING); + rescuer->task->flags &= ~PF_WQ_WORKER; return 0; } @@ -2428,9 +2352,8 @@ repeat: */ for_each_mayday_cpu(cpu, wq->mayday_mask) { unsigned int tcpu = is_unbound ? WORK_CPU_UNBOUND : cpu; - struct cpu_workqueue_struct *cwq = get_cwq(tcpu, wq); - struct worker_pool *pool = cwq->pool; - struct global_cwq *gcwq = pool->gcwq; + struct pool_workqueue *pwq = get_pwq(tcpu, wq); + struct worker_pool *pool = pwq->pool; struct work_struct *work, *n; __set_current_state(TASK_RUNNING); @@ -2446,22 +2369,24 @@ repeat: */ BUG_ON(!list_empty(&rescuer->scheduled)); list_for_each_entry_safe(work, n, &pool->worklist, entry) - if (get_work_cwq(work) == cwq) + if (get_work_pwq(work) == pwq) move_linked_works(work, scheduled, &n); process_scheduled_works(rescuer); /* - * Leave this gcwq. If keep_working() is %true, notify a + * Leave this pool. If keep_working() is %true, notify a * regular worker; otherwise, we end up with 0 concurrency * and stalling the execution. */ if (keep_working(pool)) wake_up_worker(pool); - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); } + /* rescuers should never participate in concurrency management */ + WARN_ON_ONCE(!(rescuer->flags & WORKER_NOT_RUNNING)); schedule(); goto repeat; } @@ -2479,7 +2404,7 @@ static void wq_barrier_func(struct work_struct *work) /** * insert_wq_barrier - insert a barrier work - * @cwq: cwq to insert barrier into + * @pwq: pwq to insert barrier into * @barr: wq_barrier to insert * @target: target work to attach @barr to * @worker: worker currently executing @target, NULL if @target is not executing @@ -2496,12 +2421,12 @@ static void wq_barrier_func(struct work_struct *work) * after a work with LINKED flag set. * * Note that when @worker is non-NULL, @target may be modified - * underneath us, so we can't reliably determine cwq from @target. + * underneath us, so we can't reliably determine pwq from @target. * * CONTEXT: - * spin_lock_irq(gcwq->lock). + * spin_lock_irq(pool->lock). */ -static void insert_wq_barrier(struct cpu_workqueue_struct *cwq, +static void insert_wq_barrier(struct pool_workqueue *pwq, struct wq_barrier *barr, struct work_struct *target, struct worker *worker) { @@ -2509,7 +2434,7 @@ static void insert_wq_barrier(struct cpu_workqueue_struct *cwq, unsigned int linked = 0; /* - * debugobject calls are safe here even with gcwq->lock locked + * debugobject calls are safe here even with pool->lock locked * as we know for sure that this will not trigger any of the * checks and call back into the fixup functions where we * might deadlock. @@ -2534,23 +2459,23 @@ static void insert_wq_barrier(struct cpu_workqueue_struct *cwq, } debug_work_activate(&barr->work); - insert_work(cwq, &barr->work, head, + insert_work(pwq, &barr->work, head, work_color_to_flags(WORK_NO_COLOR) | linked); } /** - * flush_workqueue_prep_cwqs - prepare cwqs for workqueue flushing + * flush_workqueue_prep_pwqs - prepare pwqs for workqueue flushing * @wq: workqueue being flushed * @flush_color: new flush color, < 0 for no-op * @work_color: new work color, < 0 for no-op * - * Prepare cwqs for workqueue flushing. + * Prepare pwqs for workqueue flushing. * - * If @flush_color is non-negative, flush_color on all cwqs should be - * -1. If no cwq has in-flight commands at the specified color, all - * cwq->flush_color's stay at -1 and %false is returned. If any cwq - * has in flight commands, its cwq->flush_color is set to - * @flush_color, @wq->nr_cwqs_to_flush is updated accordingly, cwq + * If @flush_color is non-negative, flush_color on all pwqs should be + * -1. If no pwq has in-flight commands at the specified color, all + * pwq->flush_color's stay at -1 and %false is returned. If any pwq + * has in flight commands, its pwq->flush_color is set to + * @flush_color, @wq->nr_pwqs_to_flush is updated accordingly, pwq * wakeup logic is armed and %true is returned. * * The caller should have initialized @wq->first_flusher prior to @@ -2558,7 +2483,7 @@ static void insert_wq_barrier(struct cpu_workqueue_struct *cwq, * @flush_color is negative, no flush color update is done and %false * is returned. * - * If @work_color is non-negative, all cwqs should have the same + * If @work_color is non-negative, all pwqs should have the same * work_color which is previous to @work_color and all will be * advanced to @work_color. * @@ -2569,42 +2494,42 @@ static void insert_wq_barrier(struct cpu_workqueue_struct *cwq, * %true if @flush_color >= 0 and there's something to flush. %false * otherwise. */ -static bool flush_workqueue_prep_cwqs(struct workqueue_struct *wq, +static bool flush_workqueue_prep_pwqs(struct workqueue_struct *wq, int flush_color, int work_color) { bool wait = false; unsigned int cpu; if (flush_color >= 0) { - BUG_ON(atomic_read(&wq->nr_cwqs_to_flush)); - atomic_set(&wq->nr_cwqs_to_flush, 1); + BUG_ON(atomic_read(&wq->nr_pwqs_to_flush)); + atomic_set(&wq->nr_pwqs_to_flush, 1); } - for_each_cwq_cpu(cpu, wq) { - struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq); - struct global_cwq *gcwq = cwq->pool->gcwq; + for_each_pwq_cpu(cpu, wq) { + struct pool_workqueue *pwq = get_pwq(cpu, wq); + struct worker_pool *pool = pwq->pool; - spin_lock_irq(&gcwq->lock); + spin_lock_irq(&pool->lock); if (flush_color >= 0) { - BUG_ON(cwq->flush_color != -1); + BUG_ON(pwq->flush_color != -1); - if (cwq->nr_in_flight[flush_color]) { - cwq->flush_color = flush_color; - atomic_inc(&wq->nr_cwqs_to_flush); + if (pwq->nr_in_flight[flush_color]) { + pwq->flush_color = flush_color; + atomic_inc(&wq->nr_pwqs_to_flush); wait = true; } } if (work_color >= 0) { - BUG_ON(work_color != work_next_color(cwq->work_color)); - cwq->work_color = work_color; + BUG_ON(work_color != work_next_color(pwq->work_color)); + pwq->work_color = work_color; } - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); } - if (flush_color >= 0 && atomic_dec_and_test(&wq->nr_cwqs_to_flush)) + if (flush_color >= 0 && atomic_dec_and_test(&wq->nr_pwqs_to_flush)) complete(&wq->first_flusher->done); return wait; @@ -2655,7 +2580,7 @@ void flush_workqueue(struct workqueue_struct *wq) wq->first_flusher = &this_flusher; - if (!flush_workqueue_prep_cwqs(wq, wq->flush_color, + if (!flush_workqueue_prep_pwqs(wq, wq->flush_color, wq->work_color)) { /* nothing to flush, done */ wq->flush_color = next_color; @@ -2666,7 +2591,7 @@ void flush_workqueue(struct workqueue_struct *wq) /* wait in queue */ BUG_ON(wq->flush_color == this_flusher.flush_color); list_add_tail(&this_flusher.list, &wq->flusher_queue); - flush_workqueue_prep_cwqs(wq, -1, wq->work_color); + flush_workqueue_prep_pwqs(wq, -1, wq->work_color); } } else { /* @@ -2733,7 +2658,7 @@ void flush_workqueue(struct workqueue_struct *wq) list_splice_tail_init(&wq->flusher_overflow, &wq->flusher_queue); - flush_workqueue_prep_cwqs(wq, -1, wq->work_color); + flush_workqueue_prep_pwqs(wq, -1, wq->work_color); } if (list_empty(&wq->flusher_queue)) { @@ -2743,7 +2668,7 @@ void flush_workqueue(struct workqueue_struct *wq) /* * Need to flush more colors. Make the next flusher - * the new first flusher and arm cwqs. + * the new first flusher and arm pwqs. */ BUG_ON(wq->flush_color == wq->work_color); BUG_ON(wq->flush_color != next->flush_color); @@ -2751,7 +2676,7 @@ void flush_workqueue(struct workqueue_struct *wq) list_del_init(&next->list); wq->first_flusher = next; - if (flush_workqueue_prep_cwqs(wq, wq->flush_color, -1)) + if (flush_workqueue_prep_pwqs(wq, wq->flush_color, -1)) break; /* @@ -2794,13 +2719,13 @@ void drain_workqueue(struct workqueue_struct *wq) reflush: flush_workqueue(wq); - for_each_cwq_cpu(cpu, wq) { - struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq); + for_each_pwq_cpu(cpu, wq) { + struct pool_workqueue *pwq = get_pwq(cpu, wq); bool drained; - spin_lock_irq(&cwq->pool->gcwq->lock); - drained = !cwq->nr_active && list_empty(&cwq->delayed_works); - spin_unlock_irq(&cwq->pool->gcwq->lock); + spin_lock_irq(&pwq->pool->lock); + drained = !pwq->nr_active && list_empty(&pwq->delayed_works); + spin_unlock_irq(&pwq->pool->lock); if (drained) continue; @@ -2822,34 +2747,29 @@ EXPORT_SYMBOL_GPL(drain_workqueue); static bool start_flush_work(struct work_struct *work, struct wq_barrier *barr) { struct worker *worker = NULL; - struct global_cwq *gcwq; - struct cpu_workqueue_struct *cwq; + struct worker_pool *pool; + struct pool_workqueue *pwq; might_sleep(); - gcwq = get_work_gcwq(work); - if (!gcwq) + pool = get_work_pool(work); + if (!pool) return false; - spin_lock_irq(&gcwq->lock); - if (!list_empty(&work->entry)) { - /* - * See the comment near try_to_grab_pending()->smp_rmb(). - * If it was re-queued to a different gcwq under us, we - * are not going to wait. - */ - smp_rmb(); - cwq = get_work_cwq(work); - if (unlikely(!cwq || gcwq != cwq->pool->gcwq)) + spin_lock_irq(&pool->lock); + /* see the comment in try_to_grab_pending() with the same code */ + pwq = get_work_pwq(work); + if (pwq) { + if (unlikely(pwq->pool != pool)) goto already_gone; } else { - worker = find_worker_executing_work(gcwq, work); + worker = find_worker_executing_work(pool, work); if (!worker) goto already_gone; - cwq = worker->current_cwq; + pwq = worker->current_pwq; } - insert_wq_barrier(cwq, barr, work, worker); - spin_unlock_irq(&gcwq->lock); + insert_wq_barrier(pwq, barr, work, worker); + spin_unlock_irq(&pool->lock); /* * If @max_active is 1 or rescuer is in use, flushing another work @@ -2857,15 +2777,15 @@ static bool start_flush_work(struct work_struct *work, struct wq_barrier *barr) * flusher is not running on the same workqueue by verifying write * access. */ - if (cwq->wq->saved_max_active == 1 || cwq->wq->flags & WQ_RESCUER) - lock_map_acquire(&cwq->wq->lockdep_map); + if (pwq->wq->saved_max_active == 1 || pwq->wq->flags & WQ_RESCUER) + lock_map_acquire(&pwq->wq->lockdep_map); else - lock_map_acquire_read(&cwq->wq->lockdep_map); - lock_map_release(&cwq->wq->lockdep_map); + lock_map_acquire_read(&pwq->wq->lockdep_map); + lock_map_release(&pwq->wq->lockdep_map); return true; already_gone: - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); return false; } @@ -2961,8 +2881,7 @@ bool flush_delayed_work(struct delayed_work *dwork) { local_irq_disable(); if (del_timer_sync(&dwork->timer)) - __queue_work(dwork->cpu, - get_work_cwq(&dwork->work)->wq, &dwork->work); + __queue_work(dwork->cpu, dwork->wq, &dwork->work); local_irq_enable(); return flush_work(&dwork->work); } @@ -2992,7 +2911,8 @@ bool cancel_delayed_work(struct delayed_work *dwork) if (unlikely(ret < 0)) return false; - set_work_cpu_and_clear_pending(&dwork->work, work_cpu(&dwork->work)); + set_work_pool_and_clear_pending(&dwork->work, + get_work_pool_id(&dwork->work)); local_irq_restore(flags); return ret; } @@ -3171,46 +3091,46 @@ int keventd_up(void) return system_wq != NULL; } -static int alloc_cwqs(struct workqueue_struct *wq) +static int alloc_pwqs(struct workqueue_struct *wq) { /* - * cwqs are forced aligned according to WORK_STRUCT_FLAG_BITS. + * pwqs are forced aligned according to WORK_STRUCT_FLAG_BITS. * Make sure that the alignment isn't lower than that of * unsigned long long. */ - const size_t size = sizeof(struct cpu_workqueue_struct); + const size_t size = sizeof(struct pool_workqueue); const size_t align = max_t(size_t, 1 << WORK_STRUCT_FLAG_BITS, __alignof__(unsigned long long)); if (!(wq->flags & WQ_UNBOUND)) - wq->cpu_wq.pcpu = __alloc_percpu(size, align); + wq->pool_wq.pcpu = __alloc_percpu(size, align); else { void *ptr; /* - * Allocate enough room to align cwq and put an extra + * Allocate enough room to align pwq and put an extra * pointer at the end pointing back to the originally * allocated pointer which will be used for free. */ ptr = kzalloc(size + align + sizeof(void *), GFP_KERNEL); if (ptr) { - wq->cpu_wq.single = PTR_ALIGN(ptr, align); - *(void **)(wq->cpu_wq.single + 1) = ptr; + wq->pool_wq.single = PTR_ALIGN(ptr, align); + *(void **)(wq->pool_wq.single + 1) = ptr; } } /* just in case, make sure it's actually aligned */ - BUG_ON(!IS_ALIGNED(wq->cpu_wq.v, align)); - return wq->cpu_wq.v ? 0 : -ENOMEM; + BUG_ON(!IS_ALIGNED(wq->pool_wq.v, align)); + return wq->pool_wq.v ? 0 : -ENOMEM; } -static void free_cwqs(struct workqueue_struct *wq) +static void free_pwqs(struct workqueue_struct *wq) { if (!(wq->flags & WQ_UNBOUND)) - free_percpu(wq->cpu_wq.pcpu); - else if (wq->cpu_wq.single) { - /* the pointer to free is stored right after the cwq */ - kfree(*(void **)(wq->cpu_wq.single + 1)); + free_percpu(wq->pool_wq.pcpu); + else if (wq->pool_wq.single) { + /* the pointer to free is stored right after the pwq */ + kfree(*(void **)(wq->pool_wq.single + 1)); } } @@ -3264,27 +3184,25 @@ struct workqueue_struct *__alloc_workqueue_key(const char *fmt, wq->flags = flags; wq->saved_max_active = max_active; mutex_init(&wq->flush_mutex); - atomic_set(&wq->nr_cwqs_to_flush, 0); + atomic_set(&wq->nr_pwqs_to_flush, 0); INIT_LIST_HEAD(&wq->flusher_queue); INIT_LIST_HEAD(&wq->flusher_overflow); lockdep_init_map(&wq->lockdep_map, lock_name, key, 0); INIT_LIST_HEAD(&wq->list); - if (alloc_cwqs(wq) < 0) + if (alloc_pwqs(wq) < 0) goto err; - for_each_cwq_cpu(cpu, wq) { - struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq); - struct global_cwq *gcwq = get_gcwq(cpu); - int pool_idx = (bool)(flags & WQ_HIGHPRI); - - BUG_ON((unsigned long)cwq & WORK_STRUCT_FLAG_MASK); - cwq->pool = &gcwq->pools[pool_idx]; - cwq->wq = wq; - cwq->flush_color = -1; - cwq->max_active = max_active; - INIT_LIST_HEAD(&cwq->delayed_works); + for_each_pwq_cpu(cpu, wq) { + struct pool_workqueue *pwq = get_pwq(cpu, wq); + + BUG_ON((unsigned long)pwq & WORK_STRUCT_FLAG_MASK); + pwq->pool = get_std_worker_pool(cpu, flags & WQ_HIGHPRI); + pwq->wq = wq; + pwq->flush_color = -1; + pwq->max_active = max_active; + INIT_LIST_HEAD(&pwq->delayed_works); } if (flags & WQ_RESCUER) { @@ -3297,7 +3215,8 @@ struct workqueue_struct *__alloc_workqueue_key(const char *fmt, if (!rescuer) goto err; - rescuer->task = kthread_create(rescuer_thread, wq, "%s", + rescuer->rescue_wq = wq; + rescuer->task = kthread_create(rescuer_thread, rescuer, "%s", wq->name); if (IS_ERR(rescuer->task)) goto err; @@ -3314,8 +3233,8 @@ struct workqueue_struct *__alloc_workqueue_key(const char *fmt, spin_lock(&workqueue_lock); if (workqueue_freezing && wq->flags & WQ_FREEZABLE) - for_each_cwq_cpu(cpu, wq) - get_cwq(cpu, wq)->max_active = 0; + for_each_pwq_cpu(cpu, wq) + get_pwq(cpu, wq)->max_active = 0; list_add(&wq->list, &workqueues); @@ -3324,7 +3243,7 @@ struct workqueue_struct *__alloc_workqueue_key(const char *fmt, return wq; err: if (wq) { - free_cwqs(wq); + free_pwqs(wq); free_mayday_mask(wq->mayday_mask); kfree(wq->rescuer); kfree(wq); @@ -3355,14 +3274,14 @@ void destroy_workqueue(struct workqueue_struct *wq) spin_unlock(&workqueue_lock); /* sanity check */ - for_each_cwq_cpu(cpu, wq) { - struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq); + for_each_pwq_cpu(cpu, wq) { + struct pool_workqueue *pwq = get_pwq(cpu, wq); int i; for (i = 0; i < WORK_NR_COLORS; i++) - BUG_ON(cwq->nr_in_flight[i]); - BUG_ON(cwq->nr_active); - BUG_ON(!list_empty(&cwq->delayed_works)); + BUG_ON(pwq->nr_in_flight[i]); + BUG_ON(pwq->nr_active); + BUG_ON(!list_empty(&pwq->delayed_works)); } if (wq->flags & WQ_RESCUER) { @@ -3371,29 +3290,29 @@ void destroy_workqueue(struct workqueue_struct *wq) kfree(wq->rescuer); } - free_cwqs(wq); + free_pwqs(wq); kfree(wq); } EXPORT_SYMBOL_GPL(destroy_workqueue); /** - * cwq_set_max_active - adjust max_active of a cwq - * @cwq: target cpu_workqueue_struct + * pwq_set_max_active - adjust max_active of a pwq + * @pwq: target pool_workqueue * @max_active: new max_active value. * - * Set @cwq->max_active to @max_active and activate delayed works if + * Set @pwq->max_active to @max_active and activate delayed works if * increased. * * CONTEXT: - * spin_lock_irq(gcwq->lock). + * spin_lock_irq(pool->lock). */ -static void cwq_set_max_active(struct cpu_workqueue_struct *cwq, int max_active) +static void pwq_set_max_active(struct pool_workqueue *pwq, int max_active) { - cwq->max_active = max_active; + pwq->max_active = max_active; - while (!list_empty(&cwq->delayed_works) && - cwq->nr_active < cwq->max_active) - cwq_activate_first_delayed(cwq); + while (!list_empty(&pwq->delayed_works) && + pwq->nr_active < pwq->max_active) + pwq_activate_first_delayed(pwq); } /** @@ -3416,16 +3335,17 @@ void workqueue_set_max_active(struct workqueue_struct *wq, int max_active) wq->saved_max_active = max_active; - for_each_cwq_cpu(cpu, wq) { - struct global_cwq *gcwq = get_gcwq(cpu); + for_each_pwq_cpu(cpu, wq) { + struct pool_workqueue *pwq = get_pwq(cpu, wq); + struct worker_pool *pool = pwq->pool; - spin_lock_irq(&gcwq->lock); + spin_lock_irq(&pool->lock); if (!(wq->flags & WQ_FREEZABLE) || - !(gcwq->flags & GCWQ_FREEZING)) - cwq_set_max_active(get_cwq(gcwq->cpu, wq), max_active); + !(pool->flags & POOL_FREEZING)) + pwq_set_max_active(pwq, max_active); - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); } spin_unlock(&workqueue_lock); @@ -3446,57 +3366,38 @@ EXPORT_SYMBOL_GPL(workqueue_set_max_active); */ bool workqueue_congested(unsigned int cpu, struct workqueue_struct *wq) { - struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq); + struct pool_workqueue *pwq = get_pwq(cpu, wq); - return !list_empty(&cwq->delayed_works); + return !list_empty(&pwq->delayed_works); } EXPORT_SYMBOL_GPL(workqueue_congested); /** - * work_cpu - return the last known associated cpu for @work - * @work: the work of interest - * - * RETURNS: - * CPU number if @work was ever queued. WORK_CPU_NONE otherwise. - */ -unsigned int work_cpu(struct work_struct *work) -{ - struct global_cwq *gcwq = get_work_gcwq(work); - - return gcwq ? gcwq->cpu : WORK_CPU_NONE; -} -EXPORT_SYMBOL_GPL(work_cpu); - -/** * work_busy - test whether a work is currently pending or running * @work: the work to be tested * * Test whether @work is currently pending or running. There is no * synchronization around this function and the test result is * unreliable and only useful as advisory hints or for debugging. - * Especially for reentrant wqs, the pending state might hide the - * running state. * * RETURNS: * OR'd bitmask of WORK_BUSY_* bits. */ unsigned int work_busy(struct work_struct *work) { - struct global_cwq *gcwq = get_work_gcwq(work); + struct worker_pool *pool = get_work_pool(work); unsigned long flags; unsigned int ret = 0; - if (!gcwq) - return 0; - - spin_lock_irqsave(&gcwq->lock, flags); - if (work_pending(work)) ret |= WORK_BUSY_PENDING; - if (find_worker_executing_work(gcwq, work)) - ret |= WORK_BUSY_RUNNING; - spin_unlock_irqrestore(&gcwq->lock, flags); + if (pool) { + spin_lock_irqsave(&pool->lock, flags); + if (find_worker_executing_work(pool, work)) + ret |= WORK_BUSY_RUNNING; + spin_unlock_irqrestore(&pool->lock, flags); + } return ret; } @@ -3506,65 +3407,49 @@ EXPORT_SYMBOL_GPL(work_busy); * CPU hotplug. * * There are two challenges in supporting CPU hotplug. Firstly, there - * are a lot of assumptions on strong associations among work, cwq and - * gcwq which make migrating pending and scheduled works very + * are a lot of assumptions on strong associations among work, pwq and + * pool which make migrating pending and scheduled works very * difficult to implement without impacting hot paths. Secondly, - * gcwqs serve mix of short, long and very long running works making + * worker pools serve mix of short, long and very long running works making * blocked draining impractical. * - * This is solved by allowing a gcwq to be disassociated from the CPU + * This is solved by allowing the pools to be disassociated from the CPU * running as an unbound one and allowing it to be reattached later if the * cpu comes back online. */ -/* claim manager positions of all pools */ -static void gcwq_claim_assoc_and_lock(struct global_cwq *gcwq) +static void wq_unbind_fn(struct work_struct *work) { - struct worker_pool *pool; - - for_each_worker_pool(pool, gcwq) - mutex_lock_nested(&pool->assoc_mutex, pool - gcwq->pools); - spin_lock_irq(&gcwq->lock); -} - -/* release manager positions */ -static void gcwq_release_assoc_and_unlock(struct global_cwq *gcwq) -{ - struct worker_pool *pool; - - spin_unlock_irq(&gcwq->lock); - for_each_worker_pool(pool, gcwq) - mutex_unlock(&pool->assoc_mutex); -} - -static void gcwq_unbind_fn(struct work_struct *work) -{ - struct global_cwq *gcwq = get_gcwq(smp_processor_id()); + int cpu = smp_processor_id(); struct worker_pool *pool; struct worker *worker; struct hlist_node *pos; int i; - BUG_ON(gcwq->cpu != smp_processor_id()); + for_each_std_worker_pool(pool, cpu) { + BUG_ON(cpu != smp_processor_id()); - gcwq_claim_assoc_and_lock(gcwq); + mutex_lock(&pool->assoc_mutex); + spin_lock_irq(&pool->lock); - /* - * We've claimed all manager positions. Make all workers unbound - * and set DISASSOCIATED. Before this, all workers except for the - * ones which are still executing works from before the last CPU - * down must be on the cpu. After this, they may become diasporas. - */ - for_each_worker_pool(pool, gcwq) + /* + * We've claimed all manager positions. Make all workers + * unbound and set DISASSOCIATED. Before this, all workers + * except for the ones which are still executing works from + * before the last CPU down must be on the cpu. After + * this, they may become diasporas. + */ list_for_each_entry(worker, &pool->idle_list, entry) worker->flags |= WORKER_UNBOUND; - for_each_busy_worker(worker, i, pos, gcwq) - worker->flags |= WORKER_UNBOUND; + for_each_busy_worker(worker, i, pos, pool) + worker->flags |= WORKER_UNBOUND; - gcwq->flags |= GCWQ_DISASSOCIATED; + pool->flags |= POOL_DISASSOCIATED; - gcwq_release_assoc_and_unlock(gcwq); + spin_unlock_irq(&pool->lock); + mutex_unlock(&pool->assoc_mutex); + } /* * Call schedule() so that we cross rq->lock and thus can guarantee @@ -3576,16 +3461,16 @@ static void gcwq_unbind_fn(struct work_struct *work) /* * Sched callbacks are disabled now. Zap nr_running. After this, * nr_running stays zero and need_more_worker() and keep_working() - * are always true as long as the worklist is not empty. @gcwq now - * behaves as unbound (in terms of concurrency management) gcwq - * which is served by workers tied to the CPU. + * are always true as long as the worklist is not empty. Pools on + * @cpu now behave as unbound (in terms of concurrency management) + * pools which are served by workers tied to the CPU. * * On return from this function, the current worker would trigger * unbound chain execution of pending work items if other workers * didn't already. */ - for_each_worker_pool(pool, gcwq) - atomic_set(get_pool_nr_running(pool), 0); + for_each_std_worker_pool(pool, cpu) + atomic_set(&pool->nr_running, 0); } /* @@ -3597,12 +3482,11 @@ static int __cpuinit workqueue_cpu_up_callback(struct notifier_block *nfb, void *hcpu) { unsigned int cpu = (unsigned long)hcpu; - struct global_cwq *gcwq = get_gcwq(cpu); struct worker_pool *pool; switch (action & ~CPU_TASKS_FROZEN) { case CPU_UP_PREPARE: - for_each_worker_pool(pool, gcwq) { + for_each_std_worker_pool(pool, cpu) { struct worker *worker; if (pool->nr_workers) @@ -3612,18 +3496,24 @@ static int __cpuinit workqueue_cpu_up_callback(struct notifier_block *nfb, if (!worker) return NOTIFY_BAD; - spin_lock_irq(&gcwq->lock); + spin_lock_irq(&pool->lock); start_worker(worker); - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); } break; case CPU_DOWN_FAILED: case CPU_ONLINE: - gcwq_claim_assoc_and_lock(gcwq); - gcwq->flags &= ~GCWQ_DISASSOCIATED; - rebind_workers(gcwq); - gcwq_release_assoc_and_unlock(gcwq); + for_each_std_worker_pool(pool, cpu) { + mutex_lock(&pool->assoc_mutex); + spin_lock_irq(&pool->lock); + + pool->flags &= ~POOL_DISASSOCIATED; + rebind_workers(pool); + + spin_unlock_irq(&pool->lock); + mutex_unlock(&pool->assoc_mutex); + } break; } return NOTIFY_OK; @@ -3643,7 +3533,7 @@ static int __cpuinit workqueue_cpu_down_callback(struct notifier_block *nfb, switch (action & ~CPU_TASKS_FROZEN) { case CPU_DOWN_PREPARE: /* unbinding should happen on the local CPU */ - INIT_WORK_ONSTACK(&unbind_work, gcwq_unbind_fn); + INIT_WORK_ONSTACK(&unbind_work, wq_unbind_fn); queue_work_on(cpu, system_highpri_wq, &unbind_work); flush_work(&unbind_work); break; @@ -3696,10 +3586,10 @@ EXPORT_SYMBOL_GPL(work_on_cpu); * * Start freezing workqueues. After this function returns, all freezable * workqueues will queue new works to their frozen_works list instead of - * gcwq->worklist. + * pool->worklist. * * CONTEXT: - * Grabs and releases workqueue_lock and gcwq->lock's. + * Grabs and releases workqueue_lock and pool->lock's. */ void freeze_workqueues_begin(void) { @@ -3710,23 +3600,26 @@ void freeze_workqueues_begin(void) BUG_ON(workqueue_freezing); workqueue_freezing = true; - for_each_gcwq_cpu(cpu) { - struct global_cwq *gcwq = get_gcwq(cpu); + for_each_wq_cpu(cpu) { + struct worker_pool *pool; struct workqueue_struct *wq; - spin_lock_irq(&gcwq->lock); + for_each_std_worker_pool(pool, cpu) { + spin_lock_irq(&pool->lock); - BUG_ON(gcwq->flags & GCWQ_FREEZING); - gcwq->flags |= GCWQ_FREEZING; + WARN_ON_ONCE(pool->flags & POOL_FREEZING); + pool->flags |= POOL_FREEZING; - list_for_each_entry(wq, &workqueues, list) { - struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq); + list_for_each_entry(wq, &workqueues, list) { + struct pool_workqueue *pwq = get_pwq(cpu, wq); - if (cwq && wq->flags & WQ_FREEZABLE) - cwq->max_active = 0; - } + if (pwq && pwq->pool == pool && + (wq->flags & WQ_FREEZABLE)) + pwq->max_active = 0; + } - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); + } } spin_unlock(&workqueue_lock); @@ -3754,20 +3647,20 @@ bool freeze_workqueues_busy(void) BUG_ON(!workqueue_freezing); - for_each_gcwq_cpu(cpu) { + for_each_wq_cpu(cpu) { struct workqueue_struct *wq; /* * nr_active is monotonically decreasing. It's safe * to peek without lock. */ list_for_each_entry(wq, &workqueues, list) { - struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq); + struct pool_workqueue *pwq = get_pwq(cpu, wq); - if (!cwq || !(wq->flags & WQ_FREEZABLE)) + if (!pwq || !(wq->flags & WQ_FREEZABLE)) continue; - BUG_ON(cwq->nr_active < 0); - if (cwq->nr_active) { + BUG_ON(pwq->nr_active < 0); + if (pwq->nr_active) { busy = true; goto out_unlock; } @@ -3782,10 +3675,10 @@ out_unlock: * thaw_workqueues - thaw workqueues * * Thaw workqueues. Normal queueing is restored and all collected - * frozen works are transferred to their respective gcwq worklists. + * frozen works are transferred to their respective pool worklists. * * CONTEXT: - * Grabs and releases workqueue_lock and gcwq->lock's. + * Grabs and releases workqueue_lock and pool->lock's. */ void thaw_workqueues(void) { @@ -3796,30 +3689,31 @@ void thaw_workqueues(void) if (!workqueue_freezing) goto out_unlock; - for_each_gcwq_cpu(cpu) { - struct global_cwq *gcwq = get_gcwq(cpu); + for_each_wq_cpu(cpu) { struct worker_pool *pool; struct workqueue_struct *wq; - spin_lock_irq(&gcwq->lock); + for_each_std_worker_pool(pool, cpu) { + spin_lock_irq(&pool->lock); - BUG_ON(!(gcwq->flags & GCWQ_FREEZING)); - gcwq->flags &= ~GCWQ_FREEZING; + WARN_ON_ONCE(!(pool->flags & POOL_FREEZING)); + pool->flags &= ~POOL_FREEZING; - list_for_each_entry(wq, &workqueues, list) { - struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq); + list_for_each_entry(wq, &workqueues, list) { + struct pool_workqueue *pwq = get_pwq(cpu, wq); - if (!cwq || !(wq->flags & WQ_FREEZABLE)) - continue; + if (!pwq || pwq->pool != pool || + !(wq->flags & WQ_FREEZABLE)) + continue; - /* restore max_active and repopulate worklist */ - cwq_set_max_active(cwq, wq->saved_max_active); - } + /* restore max_active and repopulate worklist */ + pwq_set_max_active(pwq, wq->saved_max_active); + } - for_each_worker_pool(pool, gcwq) wake_up_worker(pool); - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); + } } workqueue_freezing = false; @@ -3831,60 +3725,56 @@ out_unlock: static int __init init_workqueues(void) { unsigned int cpu; - int i; - /* make sure we have enough bits for OFFQ CPU number */ - BUILD_BUG_ON((1LU << (BITS_PER_LONG - WORK_OFFQ_CPU_SHIFT)) < - WORK_CPU_LAST); + /* make sure we have enough bits for OFFQ pool ID */ + BUILD_BUG_ON((1LU << (BITS_PER_LONG - WORK_OFFQ_POOL_SHIFT)) < + WORK_CPU_END * NR_STD_WORKER_POOLS); cpu_notifier(workqueue_cpu_up_callback, CPU_PRI_WORKQUEUE_UP); hotcpu_notifier(workqueue_cpu_down_callback, CPU_PRI_WORKQUEUE_DOWN); - /* initialize gcwqs */ - for_each_gcwq_cpu(cpu) { - struct global_cwq *gcwq = get_gcwq(cpu); + /* initialize CPU pools */ + for_each_wq_cpu(cpu) { struct worker_pool *pool; - spin_lock_init(&gcwq->lock); - gcwq->cpu = cpu; - gcwq->flags |= GCWQ_DISASSOCIATED; - - for (i = 0; i < BUSY_WORKER_HASH_SIZE; i++) - INIT_HLIST_HEAD(&gcwq->busy_hash[i]); - - for_each_worker_pool(pool, gcwq) { - pool->gcwq = gcwq; + for_each_std_worker_pool(pool, cpu) { + spin_lock_init(&pool->lock); + pool->cpu = cpu; + pool->flags |= POOL_DISASSOCIATED; INIT_LIST_HEAD(&pool->worklist); INIT_LIST_HEAD(&pool->idle_list); + hash_init(pool->busy_hash); init_timer_deferrable(&pool->idle_timer); pool->idle_timer.function = idle_worker_timeout; pool->idle_timer.data = (unsigned long)pool; - setup_timer(&pool->mayday_timer, gcwq_mayday_timeout, + setup_timer(&pool->mayday_timer, pool_mayday_timeout, (unsigned long)pool); mutex_init(&pool->assoc_mutex); ida_init(&pool->worker_ida); + + /* alloc pool ID */ + BUG_ON(worker_pool_assign_id(pool)); } } /* create the initial worker */ - for_each_online_gcwq_cpu(cpu) { - struct global_cwq *gcwq = get_gcwq(cpu); + for_each_online_wq_cpu(cpu) { struct worker_pool *pool; - if (cpu != WORK_CPU_UNBOUND) - gcwq->flags &= ~GCWQ_DISASSOCIATED; - - for_each_worker_pool(pool, gcwq) { + for_each_std_worker_pool(pool, cpu) { struct worker *worker; + if (cpu != WORK_CPU_UNBOUND) + pool->flags &= ~POOL_DISASSOCIATED; + worker = create_worker(pool); BUG_ON(!worker); - spin_lock_irq(&gcwq->lock); + spin_lock_irq(&pool->lock); start_worker(worker); - spin_unlock_irq(&gcwq->lock); + spin_unlock_irq(&pool->lock); } } diff --git a/kernel/workqueue_internal.h b/kernel/workqueue_internal.h new file mode 100644 index 000000000000..07650264ec15 --- /dev/null +++ b/kernel/workqueue_internal.h @@ -0,0 +1,65 @@ +/* + * kernel/workqueue_internal.h + * + * Workqueue internal header file. Only to be included by workqueue and + * core kernel subsystems. + */ +#ifndef _KERNEL_WORKQUEUE_INTERNAL_H +#define _KERNEL_WORKQUEUE_INTERNAL_H + +#include <linux/workqueue.h> +#include <linux/kthread.h> + +struct worker_pool; + +/* + * The poor guys doing the actual heavy lifting. All on-duty workers are + * either serving the manager role, on idle list or on busy hash. For + * details on the locking annotation (L, I, X...), refer to workqueue.c. + * + * Only to be used in workqueue and async. + */ +struct worker { + /* on idle list while idle, on busy hash table while busy */ + union { + struct list_head entry; /* L: while idle */ + struct hlist_node hentry; /* L: while busy */ + }; + + struct work_struct *current_work; /* L: work being processed */ + work_func_t current_func; /* L: current_work's fn */ + struct pool_workqueue *current_pwq; /* L: current_work's pwq */ + struct list_head scheduled; /* L: scheduled works */ + struct task_struct *task; /* I: worker task */ + struct worker_pool *pool; /* I: the associated pool */ + /* 64 bytes boundary on 64bit, 32 on 32bit */ + unsigned long last_active; /* L: last active timestamp */ + unsigned int flags; /* X: flags */ + int id; /* I: worker id */ + + /* for rebinding worker to CPU */ + struct work_struct rebind_work; /* L: for busy worker */ + + /* used only by rescuers to point to the target workqueue */ + struct workqueue_struct *rescue_wq; /* I: the workqueue to rescue */ +}; + +/** + * current_wq_worker - return struct worker if %current is a workqueue worker + */ +static inline struct worker *current_wq_worker(void) +{ + if (current->flags & PF_WQ_WORKER) + return kthread_data(current); + return NULL; +} + +/* + * Scheduler hooks for concurrency managed workqueue. Only to be used from + * sched.c and workqueue.c. + */ +void wq_worker_waking_up(struct task_struct *task, unsigned int cpu); +struct task_struct *wq_worker_sleeping(struct task_struct *task, + unsigned int cpu); + +#endif /* _KERNEL_WORKQUEUE_INTERNAL_H */ diff --git a/kernel/workqueue_sched.h b/kernel/workqueue_sched.h deleted file mode 100644 index 2d10fc98dc79..000000000000 --- a/kernel/workqueue_sched.h +++ /dev/null @@ -1,9 +0,0 @@ -/* - * kernel/workqueue_sched.h - * - * Scheduler hooks for concurrency managed workqueue. Only to be - * included from sched.c and workqueue.c. - */ -void wq_worker_waking_up(struct task_struct *task, unsigned int cpu); -struct task_struct *wq_worker_sleeping(struct task_struct *task, - unsigned int cpu); |