simplify cleanup_workqueue_thread()
authorOleg Nesterov <oleg@tv-sign.ru>
Wed, 23 May 2007 20:57:57 +0000 (13:57 -0700)
committerLinus Torvalds <torvalds@woody.linux-foundation.org>
Thu, 24 May 2007 03:14:13 +0000 (20:14 -0700)
cleanup_workqueue_thread() and cwq_should_stop() are overcomplicated.

Convert the code to use kthread_should_stop/kthread_stop as was
suggested by Gautham and Srivatsa.

In particular this patch removes the (unlikely) busy-wait loop from the
exit path, it was a temporary and ugly kludge (if not a bug).

Note: the current code was designed to solve another old problem:
work->func can't share locks with hotplug callbacks.  I think this could
be done, see

http://marc.info/?l=linux-kernel&m=116905366428633

but this needs some more complications to preserve CPU affinity of
cwq->thread during cpu_up().  A freezer-based hotplug looks more
appealing.

[akpm@linux-foundation.org: make it more tolerant of gcc borkenness]
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Zilvinas Valinskas <zilvinas@wilibox.com>
Cc: Gautham R Shenoy <ego@in.ibm.com>
Cc: Srivatsa Vaddagiri <vatsa@in.ibm.com>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
kernel/workqueue.c

index fb56fedd5c0274a3458bd0e7cbc8ae4e0d86cf6d..3bebf73be976592cab56a218e93b7449a5de7df4 100644 (file)
@@ -47,7 +47,6 @@ struct cpu_workqueue_struct {
 
        struct workqueue_struct *wq;
        struct task_struct *thread;
-       int should_stop;
 
        int run_depth;          /* Detect run_workqueue() recursion depth */
 } ____cacheline_aligned;
@@ -71,7 +70,13 @@ static LIST_HEAD(workqueues);
 
 static int singlethread_cpu __read_mostly;
 static cpumask_t cpu_singlethread_map __read_mostly;
-/* optimization, we could use cpu_possible_map */
+/*
+ * _cpu_down() first removes CPU from cpu_online_map, then CPU_DEAD
+ * flushes cwq->worklist. This means that flush_workqueue/wait_on_work
+ * which comes in between can't use for_each_online_cpu(). We could
+ * use cpu_possible_map, the cpumask below is more a documentation
+ * than optimization.
+ */
 static cpumask_t cpu_populated_map __read_mostly;
 
 /* If it's single threaded, it isn't in the list of workqueues. */
@@ -272,24 +277,6 @@ static void run_workqueue(struct cpu_workqueue_struct *cwq)
        spin_unlock_irq(&cwq->lock);
 }
 
-/*
- * NOTE: the caller must not touch *cwq if this func returns true
- */
-static int cwq_should_stop(struct cpu_workqueue_struct *cwq)
-{
-       int should_stop = cwq->should_stop;
-
-       if (unlikely(should_stop)) {
-               spin_lock_irq(&cwq->lock);
-               should_stop = cwq->should_stop && list_empty(&cwq->worklist);
-               if (should_stop)
-                       cwq->thread = NULL;
-               spin_unlock_irq(&cwq->lock);
-       }
-
-       return should_stop;
-}
-
 static int worker_thread(void *__cwq)
 {
        struct cpu_workqueue_struct *cwq = __cwq;
@@ -302,14 +289,15 @@ static int worker_thread(void *__cwq)
 
        for (;;) {
                prepare_to_wait(&cwq->more_work, &wait, TASK_INTERRUPTIBLE);
-               if (!freezing(current) && !cwq->should_stop
-                   && list_empty(&cwq->worklist))
+               if (!freezing(current) &&
+                   !kthread_should_stop() &&
+                   list_empty(&cwq->worklist))
                        schedule();
                finish_wait(&cwq->more_work, &wait);
 
                try_to_freeze();
 
-               if (cwq_should_stop(cwq))
+               if (kthread_should_stop())
                        break;
 
                run_workqueue(cwq);
@@ -340,18 +328,21 @@ static void insert_wq_barrier(struct cpu_workqueue_struct *cwq,
        insert_work(cwq, &barr->work, tail);
 }
 
-static void flush_cpu_workqueue(struct cpu_workqueue_struct *cwq)
+static int flush_cpu_workqueue(struct cpu_workqueue_struct *cwq)
 {
+       int active;
+
        if (cwq->thread == current) {
                /*
                 * Probably keventd trying to flush its own queue. So simply run
                 * it by hand rather than deadlocking.
                 */
                run_workqueue(cwq);
+               active = 1;
        } else {
                struct wq_barrier barr;
-               int active = 0;
 
+               active = 0;
                spin_lock_irq(&cwq->lock);
                if (!list_empty(&cwq->worklist) || cwq->current_work != NULL) {
                        insert_wq_barrier(cwq, &barr, 1);
@@ -362,6 +353,8 @@ static void flush_cpu_workqueue(struct cpu_workqueue_struct *cwq)
                if (active)
                        wait_for_completion(&barr.done);
        }
+
+       return active;
 }
 
 /**
@@ -674,7 +667,6 @@ static int create_workqueue_thread(struct cpu_workqueue_struct *cwq, int cpu)
                return PTR_ERR(p);
 
        cwq->thread = p;
-       cwq->should_stop = 0;
 
        return 0;
 }
@@ -740,29 +732,27 @@ EXPORT_SYMBOL_GPL(__create_workqueue);
 
 static void cleanup_workqueue_thread(struct cpu_workqueue_struct *cwq, int cpu)
 {
-       struct wq_barrier barr;
-       int alive = 0;
-
-       spin_lock_irq(&cwq->lock);
-       if (cwq->thread != NULL) {
-               insert_wq_barrier(cwq, &barr, 1);
-               cwq->should_stop = 1;
-               alive = 1;
-       }
-       spin_unlock_irq(&cwq->lock);
+       /*
+        * Our caller is either destroy_workqueue() or CPU_DEAD,
+        * workqueue_mutex protects cwq->thread
+        */
+       if (cwq->thread == NULL)
+               return;
 
-       if (alive) {
-               wait_for_completion(&barr.done);
+       /*
+        * If the caller is CPU_DEAD the single flush_cpu_workqueue()
+        * is not enough, a concurrent flush_workqueue() can insert a
+        * barrier after us.
+        * When ->worklist becomes empty it is safe to exit because no
+        * more work_structs can be queued on this cwq: flush_workqueue
+        * checks list_empty(), and a "normal" queue_work() can't use
+        * a dead CPU.
+        */
+       while (flush_cpu_workqueue(cwq))
+               ;
 
-               while (unlikely(cwq->thread != NULL))
-                       cpu_relax();
-               /*
-                * Wait until cwq->thread unlocks cwq->lock,
-                * it won't touch *cwq after that.
-                */
-               smp_rmb();
-               spin_unlock_wait(&cwq->lock);
-       }
+       kthread_stop(cwq->thread);
+       cwq->thread = NULL;
 }
 
 /**