Commit a08489c5 authored by Linus Torvalds's avatar Linus Torvalds

Merge branch 'for-3.6' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq

Pull workqueue changes from Tejun Heo:
 "There are three major changes.

   - WQ_HIGHPRI has been reimplemented so that high priority work items
     are served by worker threads with -20 nice value from dedicated
     highpri worker pools.

   - CPU hotplug support has been reimplemented such that idle workers
     are kept across CPU hotplug events.  This makes CPU hotplug cheaper
     (for PM) and makes the code simpler.

   - flush_kthread_work() has been reimplemented so that a work item can
     be freed while executing.  This removes an annoying behavior
     difference between kthread_worker and workqueue."

* 'for-3.6' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq:
  workqueue: fix spurious CPU locality WARN from process_one_work()
  kthread_worker: reimplement flush_kthread_work() to allow freeing the work item being executed
  kthread_worker: reorganize to prepare for flush_kthread_work() reimplementation
  workqueue: simplify CPU hotplug code
  workqueue: remove CPU offline trustee
  workqueue: don't butcher idle workers on an offline CPU
  workqueue: reimplement CPU online rebinding to handle idle workers
  workqueue: drop @bind from create_worker()
  workqueue: use mutex for global_cwq manager exclusion
  workqueue: ROGUE workers are UNBOUND workers
  workqueue: drop CPU_DYING notifier operation
  workqueue: perform cpu down operations from low priority cpu_notifier()
  workqueue: reimplement WQ_HIGHPRI using a separate worker_pool
  workqueue: introduce NR_WORKER_POOLS and for_each_worker_pool()
  workqueue: separate out worker_pool flags
  workqueue: use @pool instead of @gcwq or @cpu where applicable
  workqueue: factor out worker_pool from global_cwq
  workqueue: don't use WQ_HIGHPRI for unbound workqueues
parents 08d9329c 6fec10a1
...@@ -89,25 +89,28 @@ called thread-pools. ...@@ -89,25 +89,28 @@ called thread-pools.
The cmwq design differentiates between the user-facing workqueues that The cmwq design differentiates between the user-facing workqueues that
subsystems and drivers queue work items on and the backend mechanism subsystems and drivers queue work items on and the backend mechanism
which manages thread-pool and processes the queued work items. which manages thread-pools and processes the queued work items.
The backend is called gcwq. There is one gcwq for each possible CPU The backend is called gcwq. There is one gcwq for each possible CPU
and one gcwq to serve work items queued on unbound workqueues. and one gcwq to serve work items queued on unbound workqueues. Each
gcwq has two thread-pools - one for normal work items and the other
for high priority ones.
Subsystems and drivers can create and queue work items through special Subsystems and drivers can create and queue work items through special
workqueue API functions as they see fit. They can influence some workqueue API functions as they see fit. They can influence some
aspects of the way the work items are executed by setting flags on the aspects of the way the work items are executed by setting flags on the
workqueue they are putting the work item on. These flags include workqueue they are putting the work item on. These flags include
things like CPU locality, reentrancy, concurrency limits and more. To things like CPU locality, reentrancy, concurrency limits, priority and
get a detailed overview refer to the API description of more. To get a detailed overview refer to the API description of
alloc_workqueue() below. alloc_workqueue() below.
When a work item is queued to a workqueue, the target gcwq is When a work item is queued to a workqueue, the target gcwq and
determined according to the queue parameters and workqueue attributes thread-pool is determined according to the queue parameters and
and appended on the shared worklist of the gcwq. For example, unless workqueue attributes and appended on the shared worklist of the
specifically overridden, a work item of a bound workqueue will be thread-pool. For example, unless specifically overridden, a work item
queued on the worklist of exactly that gcwq that is associated to the of a bound workqueue will be queued on the worklist of either normal
CPU the issuer is running on. or highpri thread-pool of the gcwq that is associated to the CPU the
issuer is running on.
For any worker pool implementation, managing the concurrency level For any worker pool implementation, managing the concurrency level
(how many execution contexts are active) is an important issue. cmwq (how many execution contexts are active) is an important issue. cmwq
...@@ -115,26 +118,26 @@ tries to keep the concurrency at a minimal but sufficient level. ...@@ -115,26 +118,26 @@ tries to keep the concurrency at a minimal but sufficient level.
Minimal to save resources and sufficient in that the system is used at Minimal to save resources and sufficient in that the system is used at
its full capacity. its full capacity.
Each gcwq bound to an actual CPU implements concurrency management by Each thread-pool bound to an actual CPU implements concurrency
hooking into the scheduler. The gcwq is notified whenever an active management by hooking into the scheduler. The thread-pool is notified
worker wakes up or sleeps and keeps track of the number of the whenever an active worker wakes up or sleeps and keeps track of the
currently runnable workers. Generally, work items are not expected to number of the currently runnable workers. Generally, work items are
hog a CPU and consume many cycles. That means maintaining just enough not expected to hog a CPU and consume many cycles. That means
concurrency to prevent work processing from stalling should be maintaining just enough concurrency to prevent work processing from
optimal. As long as there are one or more runnable workers on the stalling should be optimal. As long as there are one or more runnable
CPU, the gcwq doesn't start execution of a new work, but, when the workers on the CPU, the thread-pool doesn't start execution of a new
last running worker goes to sleep, it immediately schedules a new work, but, when the last running worker goes to sleep, it immediately
worker so that the CPU doesn't sit idle while there are pending work schedules a new worker so that the CPU doesn't sit idle while there
items. This allows using a minimal number of workers without losing are pending work items. This allows using a minimal number of workers
execution bandwidth. without losing execution bandwidth.
Keeping idle workers around doesn't cost other than the memory space Keeping idle workers around doesn't cost other than the memory space
for kthreads, so cmwq holds onto idle ones for a while before killing for kthreads, so cmwq holds onto idle ones for a while before killing
them. them.
For an unbound wq, the above concurrency management doesn't apply and For an unbound wq, the above concurrency management doesn't apply and
the gcwq for the pseudo unbound CPU tries to start executing all work the thread-pools for the pseudo unbound CPU try to start executing all
items as soon as possible. The responsibility of regulating work items as soon as possible. The responsibility of regulating
concurrency level is on the users. There is also a flag to mark a concurrency level is on the users. There is also a flag to mark a
bound wq to ignore the concurrency management. Please refer to the bound wq to ignore the concurrency management. Please refer to the
API section for details. API section for details.
...@@ -205,31 +208,22 @@ resources, scheduled and executed. ...@@ -205,31 +208,22 @@ resources, scheduled and executed.
WQ_HIGHPRI WQ_HIGHPRI
Work items of a highpri wq are queued at the head of the Work items of a highpri wq are queued to the highpri
worklist of the target gcwq and start execution regardless of thread-pool of the target gcwq. Highpri thread-pools are
the current concurrency level. In other words, highpri work served by worker threads with elevated nice level.
items will always start execution as soon as execution
resource is available.
Ordering among highpri work items is preserved - a highpri Note that normal and highpri thread-pools don't interact with
work item queued after another highpri work item will start each other. Each maintain its separate pool of workers and
execution after the earlier highpri work item starts. implements concurrency management among its workers.
Although highpri work items are not held back by other
runnable work items, they still contribute to the concurrency
level. Highpri work items in runnable state will prevent
non-highpri work items from starting execution.
This flag is meaningless for unbound wq.
WQ_CPU_INTENSIVE WQ_CPU_INTENSIVE
Work items of a CPU intensive wq do not contribute to the Work items of a CPU intensive wq do not contribute to the
concurrency level. In other words, runnable CPU intensive concurrency level. In other words, runnable CPU intensive
work items will not prevent other work items from starting work items will not prevent other work items in the same
execution. This is useful for bound work items which are thread-pool from starting execution. This is useful for bound
expected to hog CPU cycles so that their execution is work items which are expected to hog CPU cycles so that their
regulated by the system scheduler. execution is regulated by the system scheduler.
Although CPU intensive work items don't contribute to the Although CPU intensive work items don't contribute to the
concurrency level, start of their executions is still concurrency level, start of their executions is still
...@@ -239,14 +233,6 @@ resources, scheduled and executed. ...@@ -239,14 +233,6 @@ resources, scheduled and executed.
This flag is meaningless for unbound wq. This flag is meaningless for unbound wq.
WQ_HIGHPRI | WQ_CPU_INTENSIVE
This combination makes the wq avoid interaction with
concurrency management completely and behave as a simple
per-CPU execution context provider. Work items queued on a
highpri CPU-intensive wq start execution as soon as resources
are available and don't affect execution of other work items.
@max_active: @max_active:
@max_active determines the maximum number of execution contexts per @max_active determines the maximum number of execution contexts per
...@@ -328,20 +314,7 @@ If @max_active == 2, ...@@ -328,20 +314,7 @@ If @max_active == 2,
35 w2 wakes up and finishes 35 w2 wakes up and finishes
Now, let's assume w1 and w2 are queued to a different wq q1 which has Now, let's assume w1 and w2 are queued to a different wq q1 which has
WQ_HIGHPRI set, WQ_CPU_INTENSIVE set,
TIME IN MSECS EVENT
0 w1 and w2 start and burn CPU
5 w1 sleeps
10 w2 sleeps
10 w0 starts and burns CPU
15 w0 sleeps
15 w1 wakes up and finishes
20 w2 wakes up and finishes
25 w0 wakes up and burns CPU
30 w0 finishes
If q1 has WQ_CPU_INTENSIVE set,
TIME IN MSECS EVENT TIME IN MSECS EVENT
0 w0 starts and burns CPU 0 w0 starts and burns CPU
......
...@@ -73,8 +73,9 @@ enum { ...@@ -73,8 +73,9 @@ enum {
/* migration should happen before other stuff but after perf */ /* migration should happen before other stuff but after perf */
CPU_PRI_PERF = 20, CPU_PRI_PERF = 20,
CPU_PRI_MIGRATION = 10, CPU_PRI_MIGRATION = 10,
/* prepare workqueues for other notifiers */ /* bring up workqueues before normal notifiers and down after */
CPU_PRI_WORKQUEUE = 5, CPU_PRI_WORKQUEUE_UP = 5,
CPU_PRI_WORKQUEUE_DOWN = -5,
}; };
#define CPU_ONLINE 0x0002 /* CPU (unsigned)v is up */ #define CPU_ONLINE 0x0002 /* CPU (unsigned)v is up */
......
...@@ -49,8 +49,6 @@ extern int tsk_fork_get_node(struct task_struct *tsk); ...@@ -49,8 +49,6 @@ extern int tsk_fork_get_node(struct task_struct *tsk);
* can be queued and flushed using queue/flush_kthread_work() * can be queued and flushed using queue/flush_kthread_work()
* respectively. Queued kthread_works are processed by a kthread * respectively. Queued kthread_works are processed by a kthread
* running kthread_worker_fn(). * running kthread_worker_fn().
*
* A kthread_work can't be freed while it is executing.
*/ */
struct kthread_work; struct kthread_work;
typedef void (*kthread_work_func_t)(struct kthread_work *work); typedef void (*kthread_work_func_t)(struct kthread_work *work);
...@@ -59,15 +57,14 @@ struct kthread_worker { ...@@ -59,15 +57,14 @@ struct kthread_worker {
spinlock_t lock; spinlock_t lock;
struct list_head work_list; struct list_head work_list;
struct task_struct *task; struct task_struct *task;
struct kthread_work *current_work;
}; };
struct kthread_work { struct kthread_work {
struct list_head node; struct list_head node;
kthread_work_func_t func; kthread_work_func_t func;
wait_queue_head_t done; wait_queue_head_t done;
atomic_t flushing; struct kthread_worker *worker;
int queue_seq;
int done_seq;
}; };
#define KTHREAD_WORKER_INIT(worker) { \ #define KTHREAD_WORKER_INIT(worker) { \
...@@ -79,7 +76,6 @@ struct kthread_work { ...@@ -79,7 +76,6 @@ struct kthread_work {
.node = LIST_HEAD_INIT((work).node), \ .node = LIST_HEAD_INIT((work).node), \
.func = (fn), \ .func = (fn), \
.done = __WAIT_QUEUE_HEAD_INITIALIZER((work).done), \ .done = __WAIT_QUEUE_HEAD_INITIALIZER((work).done), \
.flushing = ATOMIC_INIT(0), \
} }
#define DEFINE_KTHREAD_WORKER(worker) \ #define DEFINE_KTHREAD_WORKER(worker) \
......
...@@ -54,7 +54,7 @@ TRACE_EVENT(workqueue_queue_work, ...@@ -54,7 +54,7 @@ TRACE_EVENT(workqueue_queue_work,
__entry->function = work->func; __entry->function = work->func;
__entry->workqueue = cwq->wq; __entry->workqueue = cwq->wq;
__entry->req_cpu = req_cpu; __entry->req_cpu = req_cpu;
__entry->cpu = cwq->gcwq->cpu; __entry->cpu = cwq->pool->gcwq->cpu;
), ),
TP_printk("work struct=%p function=%pf workqueue=%p req_cpu=%u cpu=%u", TP_printk("work struct=%p function=%pf workqueue=%p req_cpu=%u cpu=%u",
......
...@@ -360,16 +360,12 @@ int kthread_worker_fn(void *worker_ptr) ...@@ -360,16 +360,12 @@ int kthread_worker_fn(void *worker_ptr)
struct kthread_work, node); struct kthread_work, node);
list_del_init(&work->node); list_del_init(&work->node);
} }
worker->current_work = work;
spin_unlock_irq(&worker->lock); spin_unlock_irq(&worker->lock);
if (work) { if (work) {
__set_current_state(TASK_RUNNING); __set_current_state(TASK_RUNNING);
work->func(work); work->func(work);
smp_wmb(); /* wmb worker-b0 paired with flush-b1 */
work->done_seq = work->queue_seq;
smp_mb(); /* mb worker-b1 paired with flush-b0 */
if (atomic_read(&work->flushing))
wake_up_all(&work->done);
} else if (!freezing(current)) } else if (!freezing(current))
schedule(); schedule();
...@@ -378,6 +374,19 @@ int kthread_worker_fn(void *worker_ptr) ...@@ -378,6 +374,19 @@ int kthread_worker_fn(void *worker_ptr)
} }
EXPORT_SYMBOL_GPL(kthread_worker_fn); EXPORT_SYMBOL_GPL(kthread_worker_fn);
/* insert @work before @pos in @worker */
static void insert_kthread_work(struct kthread_worker *worker,
struct kthread_work *work,
struct list_head *pos)
{
lockdep_assert_held(&worker->lock);
list_add_tail(&work->node, pos);
work->worker = worker;
if (likely(worker->task))
wake_up_process(worker->task);
}
/** /**
* queue_kthread_work - queue a kthread_work * queue_kthread_work - queue a kthread_work
* @worker: target kthread_worker * @worker: target kthread_worker
...@@ -395,10 +404,7 @@ bool queue_kthread_work(struct kthread_worker *worker, ...@@ -395,10 +404,7 @@ bool queue_kthread_work(struct kthread_worker *worker,
spin_lock_irqsave(&worker->lock, flags); spin_lock_irqsave(&worker->lock, flags);
if (list_empty(&work->node)) { if (list_empty(&work->node)) {
list_add_tail(&work->node, &worker->work_list); insert_kthread_work(worker, work, &worker->work_list);
work->queue_seq++;
if (likely(worker->task))
wake_up_process(worker->task);
ret = true; ret = true;
} }
spin_unlock_irqrestore(&worker->lock, flags); spin_unlock_irqrestore(&worker->lock, flags);
...@@ -406,6 +412,18 @@ bool queue_kthread_work(struct kthread_worker *worker, ...@@ -406,6 +412,18 @@ bool queue_kthread_work(struct kthread_worker *worker,
} }
EXPORT_SYMBOL_GPL(queue_kthread_work); EXPORT_SYMBOL_GPL(queue_kthread_work);
struct kthread_flush_work {
struct kthread_work work;
struct completion done;
};
static void kthread_flush_work_fn(struct kthread_work *work)
{
struct kthread_flush_work *fwork =
container_of(work, struct kthread_flush_work, work);
complete(&fwork->done);
}
/** /**
* flush_kthread_work - flush a kthread_work * flush_kthread_work - flush a kthread_work
* @work: work to flush * @work: work to flush
...@@ -414,39 +432,37 @@ EXPORT_SYMBOL_GPL(queue_kthread_work); ...@@ -414,39 +432,37 @@ EXPORT_SYMBOL_GPL(queue_kthread_work);
*/ */
void flush_kthread_work(struct kthread_work *work) void flush_kthread_work(struct kthread_work *work)
{ {
int seq = work->queue_seq; struct kthread_flush_work fwork = {
KTHREAD_WORK_INIT(fwork.work, kthread_flush_work_fn),
atomic_inc(&work->flushing); COMPLETION_INITIALIZER_ONSTACK(fwork.done),
};
struct kthread_worker *worker;
bool noop = false;
/* retry:
* mb flush-b0 paired with worker-b1, to make sure either worker = work->worker;
* worker sees the above increment or we see done_seq update. if (!worker)
*/ return;
smp_mb__after_atomic_inc();
/* A - B <= 0 tests whether B is in front of A regardless of overflow */ spin_lock_irq(&worker->lock);
wait_event(work->done, seq - work->done_seq <= 0); if (work->worker != worker) {
atomic_dec(&work->flushing); spin_unlock_irq(&worker->lock);
goto retry;
}
/* if (!list_empty(&work->node))
* rmb flush-b1 paired with worker-b0, to make sure our caller insert_kthread_work(worker, &fwork.work, work->node.next);
* sees every change made by work->func(). else if (worker->current_work == work)
*/ insert_kthread_work(worker, &fwork.work, worker->work_list.next);
smp_mb__after_atomic_dec(); else
} noop = true;
EXPORT_SYMBOL_GPL(flush_kthread_work);
struct kthread_flush_work { spin_unlock_irq(&worker->lock);
struct kthread_work work;
struct completion done;
};
static void kthread_flush_work_fn(struct kthread_work *work) if (!noop)
{ wait_for_completion(&fwork.done);
struct kthread_flush_work *fwork =
container_of(work, struct kthread_flush_work, work);
complete(&fwork->done);
} }
EXPORT_SYMBOL_GPL(flush_kthread_work);
/** /**
* flush_kthread_worker - flush all current works on a kthread_worker * flush_kthread_worker - flush all current works on a kthread_worker
......
This diff is collapsed.
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment