workqueue: move nr_running into worker_pool
authorTejun Heo <tj@kernel.org>
Thu, 24 Jan 2013 19:39:44 +0000 (11:39 -0800)
committerTejun Heo <tj@kernel.org>
Thu, 7 Feb 2013 21:14:20 +0000 (13:14 -0800)
As nr_running is likely to be accessed from other CPUs during
try_to_wake_up(), it was kept outside worker_pool; however, while less
frequent, other fields in worker_pool are accessed from other CPUs
for, e.g., non-reentrancy check.  Also, with recent pool related
changes, accessing nr_running matching the worker_pool isn't as simple
as it used to be.

Move nr_running inside worker_pool.  Keep it aligned to cacheline and
define CPU pools using DEFINE_PER_CPU_SHARED_ALIGNED().  This should
give at least the same cacheline behavior.

get_pool_nr_running() is replaced with direct pool->nr_running
accesses.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Joonsoo Kim <js1304@gmail.com>
kernel/workqueue.c

index e2dd61861fbd17a150f3be5f470c782327fc0073..91ce7a984c226cca1796d0074eec12d6b4a4bb56 100644 (file)
@@ -144,6 +144,13 @@ struct worker_pool {
 
        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;
 
 /*
@@ -417,23 +424,12 @@ static LIST_HEAD(workqueues);
 static bool workqueue_freezing;                /* W: have wqs started freezing? */
 
 /*
- * The CPU standard worker pools.  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.
- */
-static DEFINE_PER_CPU(struct worker_pool [NR_STD_WORKER_POOLS],
-                     cpu_std_worker_pools);
-static DEFINE_PER_CPU_SHARED_ALIGNED(atomic_t [NR_STD_WORKER_POOLS],
-                                    cpu_std_pool_nr_running);
-
-/*
- * Standard worker pools and nr_running counter for unbound CPU.  The pools
- * have POOL_DISASSOCIATED set, and all workers have WORKER_UNBOUND set.
+ * 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_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];
-static atomic_t unbound_std_pool_nr_running[NR_STD_WORKER_POOLS] = {
-       [0 ... NR_STD_WORKER_POOLS - 1] = ATOMIC_INIT(0),       /* always 0 */
-};
 
 /* idr of all pools */
 static DEFINE_MUTEX(worker_pool_idr_mutex);
@@ -483,17 +479,6 @@ static struct worker_pool *get_std_worker_pool(int cpu, bool highpri)
        return &pools[highpri];
 }
 
-static atomic_t *get_pool_nr_running(struct worker_pool *pool)
-{
-       int cpu = pool->cpu;
-       int idx = std_worker_pool_pri(pool);
-
-       if (cpu != WORK_CPU_UNBOUND)
-               return &per_cpu(cpu_std_pool_nr_running, cpu)[idx];
-       else
-               return &unbound_std_pool_nr_running[idx];
-}
-
 static struct cpu_workqueue_struct *get_cwq(unsigned int cpu,
                                            struct workqueue_struct *wq)
 {
@@ -654,7 +639,7 @@ static bool work_is_canceling(struct work_struct *work)
 
 static bool __need_more_worker(struct worker_pool *pool)
 {
-       return !atomic_read(get_pool_nr_running(pool));
+       return !atomic_read(&pool->nr_running);
 }
 
 /*
@@ -679,9 +664,8 @@ 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. */
@@ -761,7 +745,7 @@ void wq_worker_waking_up(struct task_struct *task, unsigned int cpu)
 
        if (!(worker->flags & WORKER_NOT_RUNNING)) {
                WARN_ON_ONCE(worker->pool->cpu != cpu);
-               atomic_inc(get_pool_nr_running(worker->pool));
+               atomic_inc(&worker->pool->nr_running);
        }
 }
 
@@ -785,7 +769,6 @@ struct task_struct *wq_worker_sleeping(struct task_struct *task,
 {
        struct worker *worker = kthread_data(task), *to_wakeup = NULL;
        struct worker_pool *pool;
-       atomic_t *nr_running;
 
        /*
         * Rescuers, which may not have all the fields set up like normal
@@ -796,7 +779,6 @@ struct task_struct *wq_worker_sleeping(struct task_struct *task,
                return NULL;
 
        pool = worker->pool;
-       nr_running = get_pool_nr_running(pool);
 
        /* this can only happen on the local cpu */
        BUG_ON(cpu != raw_smp_processor_id());
@@ -812,7 +794,8 @@ struct task_struct *wq_worker_sleeping(struct task_struct *task,
         * 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;
 }
@@ -844,14 +827,12 @@ 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;
@@ -883,7 +864,7 @@ 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);
 }
 
 /**
@@ -1518,7 +1499,7 @@ static void worker_enter_idle(struct worker *worker)
         */
        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));
 }
 
 /**
@@ -3506,7 +3487,7 @@ static void wq_unbind_fn(struct work_struct *work)
         * didn't already.
         */
        for_each_std_worker_pool(pool, cpu)
-               atomic_set(get_pool_nr_running(pool), 0);
+               atomic_set(&pool->nr_running, 0);
 }
 
 /*