import PULS_20160108
[GitHub/mt8127/android_kernel_alcatel_ttab.git] / mm / backing-dev.c
index 41733c5dc820af44282d382ebcc3c938eca912ac..eea1a9dfac38fc84a56a72f8161ae1fe7ddb8260 100644 (file)
@@ -31,13 +31,14 @@ EXPORT_SYMBOL_GPL(noop_backing_dev_info);
 static struct class *bdi_class;
 
 /*
- * bdi_lock protects updates to bdi_list and bdi_pending_list, as well as
- * reader side protection for bdi_pending_list. bdi_list has RCU reader side
+ * bdi_lock protects updates to bdi_list. bdi_list has RCU reader side
  * locking.
  */
 DEFINE_SPINLOCK(bdi_lock);
 LIST_HEAD(bdi_list);
-LIST_HEAD(bdi_pending_list);
+
+/* bdi_wq serves all asynchronous writeback tasks */
+struct workqueue_struct *bdi_wq;
 
 void bdi_lock_two(struct bdi_writeback *wb1, struct bdi_writeback *wb2)
 {
@@ -257,6 +258,11 @@ static int __init default_bdi_init(void)
 {
        int err;
 
+       bdi_wq = alloc_workqueue("writeback", WQ_MEM_RECLAIM | WQ_FREEZABLE |
+                                             WQ_UNBOUND | WQ_SYSFS, 0);
+       if (!bdi_wq)
+               return -ENOMEM;
+
        err = bdi_init(&default_backing_dev_info);
        if (!err)
                bdi_register(&default_backing_dev_info, NULL, "default");
@@ -271,26 +277,6 @@ int bdi_has_dirty_io(struct backing_dev_info *bdi)
        return wb_has_dirty_io(&bdi->wb);
 }
 
-static void wakeup_timer_fn(unsigned long data)
-{
-       struct backing_dev_info *bdi = (struct backing_dev_info *)data;
-
-       spin_lock_bh(&bdi->wb_lock);
-       if (bdi->wb.task) {
-               trace_writeback_wake_thread(bdi);
-               wake_up_process(bdi->wb.task);
-       } else if (bdi->dev) {
-               /*
-                * When bdi tasks are inactive for long time, they are killed.
-                * In this case we have to wake-up the forker thread which
-                * should create and run the bdi thread.
-                */
-               trace_writeback_wake_forker_thread(bdi);
-               wake_up_process(default_backing_dev_info.wb.task);
-       }
-       spin_unlock_bh(&bdi->wb_lock);
-}
-
 /*
  * This function is used when the first inode for this bdi is marked dirty. It
  * wakes-up the corresponding bdi thread which should then take care of the
@@ -301,182 +287,19 @@ static void wakeup_timer_fn(unsigned long data)
  * Note, we wouldn't bother setting up the timer, but this function is on the
  * fast-path (used by '__mark_inode_dirty()'), so we save few context switches
  * by delaying the wake-up.
+ *
+ * We have to be careful not to postpone flush work if it is scheduled for
+ * earlier. Thus we use queue_delayed_work().
  */
 void bdi_wakeup_thread_delayed(struct backing_dev_info *bdi)
 {
        unsigned long timeout;
 
        timeout = msecs_to_jiffies(dirty_writeback_interval * 10);
-       mod_timer(&bdi->wb.wakeup_timer, jiffies + timeout);
-}
-
-/*
- * Calculate the longest interval (jiffies) bdi threads are allowed to be
- * inactive.
- */
-static unsigned long bdi_longest_inactive(void)
-{
-       unsigned long interval;
-
-       interval = msecs_to_jiffies(dirty_writeback_interval * 10);
-       return max(5UL * 60 * HZ, interval);
-}
-
-/*
- * Clear pending bit and wakeup anybody waiting for flusher thread creation or
- * shutdown
- */
-static void bdi_clear_pending(struct backing_dev_info *bdi)
-{
-       clear_bit(BDI_pending, &bdi->state);
-       smp_mb__after_clear_bit();
-       wake_up_bit(&bdi->state, BDI_pending);
-}
-
-static int bdi_forker_thread(void *ptr)
-{
-       struct bdi_writeback *me = ptr;
-
-       current->flags |= PF_SWAPWRITE;
-       set_freezable();
-
-       /*
-        * Our parent may run at a different priority, just set us to normal
-        */
-       set_user_nice(current, 0);
-
-       for (;;) {
-               struct task_struct *task = NULL;
-               struct backing_dev_info *bdi;
-               enum {
-                       NO_ACTION,   /* Nothing to do */
-                       FORK_THREAD, /* Fork bdi thread */
-                       KILL_THREAD, /* Kill inactive bdi thread */
-               } action = NO_ACTION;
-
-               /*
-                * Temporary measure, we want to make sure we don't see
-                * dirty data on the default backing_dev_info
-                */
-               if (wb_has_dirty_io(me) || !list_empty(&me->bdi->work_list)) {
-                       del_timer(&me->wakeup_timer);
-                       wb_do_writeback(me, 0);
-               }
-
-               spin_lock_bh(&bdi_lock);
-               /*
-                * In the following loop we are going to check whether we have
-                * some work to do without any synchronization with tasks
-                * waking us up to do work for them. Set the task state here
-                * so that we don't miss wakeups after verifying conditions.
-                */
-               set_current_state(TASK_INTERRUPTIBLE);
-
-               list_for_each_entry(bdi, &bdi_list, bdi_list) {
-                       bool have_dirty_io;
-
-                       if (!bdi_cap_writeback_dirty(bdi) ||
-                            bdi_cap_flush_forker(bdi))
-                               continue;
-
-                       WARN(!test_bit(BDI_registered, &bdi->state),
-                            "bdi %p/%s is not registered!\n", bdi, bdi->name);
-
-                       have_dirty_io = !list_empty(&bdi->work_list) ||
-                                       wb_has_dirty_io(&bdi->wb);
-
-                       /*
-                        * If the bdi has work to do, but the thread does not
-                        * exist - create it.
-                        */
-                       if (!bdi->wb.task && have_dirty_io) {
-                               /*
-                                * Set the pending bit - if someone will try to
-                                * unregister this bdi - it'll wait on this bit.
-                                */
-                               set_bit(BDI_pending, &bdi->state);
-                               action = FORK_THREAD;
-                               break;
-                       }
-
-                       spin_lock(&bdi->wb_lock);
-
-                       /*
-                        * If there is no work to do and the bdi thread was
-                        * inactive long enough - kill it. The wb_lock is taken
-                        * to make sure no-one adds more work to this bdi and
-                        * wakes the bdi thread up.
-                        */
-                       if (bdi->wb.task && !have_dirty_io &&
-                           time_after(jiffies, bdi->wb.last_active +
-                                               bdi_longest_inactive())) {
-                               task = bdi->wb.task;
-                               bdi->wb.task = NULL;
-                               spin_unlock(&bdi->wb_lock);
-                               set_bit(BDI_pending, &bdi->state);
-                               action = KILL_THREAD;
-                               break;
-                       }
-                       spin_unlock(&bdi->wb_lock);
-               }
-               spin_unlock_bh(&bdi_lock);
-
-               /* Keep working if default bdi still has things to do */
-               if (!list_empty(&me->bdi->work_list))
-                       __set_current_state(TASK_RUNNING);
-
-               switch (action) {
-               case FORK_THREAD:
-                       __set_current_state(TASK_RUNNING);
-                       task = kthread_create(bdi_writeback_thread, &bdi->wb,
-                                             "flush-%s", dev_name(bdi->dev));
-                       if (IS_ERR(task)) {
-                               /*
-                                * If thread creation fails, force writeout of
-                                * the bdi from the thread. Hopefully 1024 is
-                                * large enough for efficient IO.
-                                */
-                               writeback_inodes_wb(&bdi->wb, 1024,
-                                                   WB_REASON_FORKER_THREAD);
-                       } else {
-                               /*
-                                * The spinlock makes sure we do not lose
-                                * wake-ups when racing with 'bdi_queue_work()'.
-                                * And as soon as the bdi thread is visible, we
-                                * can start it.
-                                */
-                               spin_lock_bh(&bdi->wb_lock);
-                               bdi->wb.task = task;
-                               spin_unlock_bh(&bdi->wb_lock);
-                               wake_up_process(task);
-                       }
-                       bdi_clear_pending(bdi);
-                       break;
-
-               case KILL_THREAD:
-                       __set_current_state(TASK_RUNNING);
-                       kthread_stop(task);
-                       bdi_clear_pending(bdi);
-                       break;
-
-               case NO_ACTION:
-                       if (!wb_has_dirty_io(me) || !dirty_writeback_interval)
-                               /*
-                                * There are no dirty data. The only thing we
-                                * should now care about is checking for
-                                * inactive bdi threads and killing them. Thus,
-                                * let's sleep for longer time, save energy and
-                                * be friendly for battery-driven devices.
-                                */
-                               schedule_timeout(bdi_longest_inactive());
-                       else
-                               schedule_timeout(msecs_to_jiffies(dirty_writeback_interval * 10));
-                       try_to_freeze();
-                       break;
-               }
-       }
-
-       return 0;
+       spin_lock_bh(&bdi->wb_lock);
+       if (test_bit(BDI_registered, &bdi->state))
+               queue_delayed_work(bdi_wq, &bdi->wb.dwork, timeout);
+       spin_unlock_bh(&bdi->wb_lock);
 }
 
 /*
@@ -508,20 +331,6 @@ int bdi_register(struct backing_dev_info *bdi, struct device *parent,
 
        bdi->dev = dev;
 
-       /*
-        * Just start the forker thread for our default backing_dev_info,
-        * and add other bdi's to the list. They will get a thread created
-        * on-demand when they need it.
-        */
-       if (bdi_cap_flush_forker(bdi)) {
-               struct bdi_writeback *wb = &bdi->wb;
-
-               wb->task = kthread_run(bdi_forker_thread, wb, "bdi-%s",
-                                               dev_name(dev));
-               if (IS_ERR(wb->task))
-                       return PTR_ERR(wb->task);
-       }
-
        bdi_debug_register(bdi, dev_name(dev));
        set_bit(BDI_registered, &bdi->state);
 
@@ -545,8 +354,6 @@ EXPORT_SYMBOL(bdi_register_dev);
  */
 static void bdi_wb_shutdown(struct backing_dev_info *bdi)
 {
-       struct task_struct *task;
-
        if (!bdi_cap_writeback_dirty(bdi))
                return;
 
@@ -555,23 +362,26 @@ static void bdi_wb_shutdown(struct backing_dev_info *bdi)
         */
        bdi_remove_from_list(bdi);
 
+       /* Make sure nobody queues further work */
+       spin_lock_bh(&bdi->wb_lock);
+       clear_bit(BDI_registered, &bdi->state);
+       spin_unlock_bh(&bdi->wb_lock);
+
        /*
-        * If setup is pending, wait for that to complete first
+        * Drain work list and shutdown the delayed_work.  At this point,
+        * @bdi->bdi_list is empty telling bdi_Writeback_workfn() that @bdi
+        * is dying and its work_list needs to be drained no matter what.
         */
-       wait_on_bit(&bdi->state, BDI_pending, bdi_sched_wait,
-                       TASK_UNINTERRUPTIBLE);
+       mod_delayed_work(bdi_wq, &bdi->wb.dwork, 0);
+       flush_delayed_work(&bdi->wb.dwork);
+       WARN_ON(!list_empty(&bdi->work_list));
 
        /*
-        * Finally, kill the kernel thread. We don't need to be RCU
-        * safe anymore, since the bdi is gone from visibility.
+        * This shouldn't be necessary unless @bdi for some reason has
+        * unflushed dirty IO after work_list is drained.  Do it anyway
+        * just in case.
         */
-       spin_lock_bh(&bdi->wb_lock);
-       task = bdi->wb.task;
-       bdi->wb.task = NULL;
-       spin_unlock_bh(&bdi->wb_lock);
-
-       if (task)
-               kthread_stop(task);
+       cancel_delayed_work_sync(&bdi->wb.dwork);
 }
 
 /*
@@ -597,10 +407,8 @@ void bdi_unregister(struct backing_dev_info *bdi)
                bdi_set_min_ratio(bdi, 0);
                trace_writeback_bdi_unregister(bdi);
                bdi_prune_sb(bdi);
-               del_timer_sync(&bdi->wb.wakeup_timer);
 
-               if (!bdi_cap_flush_forker(bdi))
-                       bdi_wb_shutdown(bdi);
+               bdi_wb_shutdown(bdi);
                bdi_debug_unregister(bdi);
 
                spin_lock_bh(&bdi->wb_lock);
@@ -622,7 +430,7 @@ static void bdi_wb_init(struct bdi_writeback *wb, struct backing_dev_info *bdi)
        INIT_LIST_HEAD(&wb->b_io);
        INIT_LIST_HEAD(&wb->b_more_io);
        spin_lock_init(&wb->list_lock);
-       setup_timer(&wb->wakeup_timer, wakeup_timer_fn, (unsigned long)bdi);
+       INIT_DELAYED_WORK(&wb->dwork, bdi_writeback_workfn);
 }
 
 /*
@@ -695,12 +503,11 @@ void bdi_destroy(struct backing_dev_info *bdi)
        bdi_unregister(bdi);
 
        /*
-        * If bdi_unregister() had already been called earlier, the
-        * wakeup_timer could still be armed because bdi_prune_sb()
-        * can race with the bdi_wakeup_thread_delayed() calls from
-        * __mark_inode_dirty().
+        * If bdi_unregister() had already been called earlier, the dwork
+        * could still be pending because bdi_prune_sb() can race with the
+        * bdi_wakeup_thread_delayed() calls from __mark_inode_dirty().
         */
-       del_timer_sync(&bdi->wb.wakeup_timer);
+       cancel_delayed_work_sync(&bdi->wb.dwork);
 
        for (i = 0; i < NR_BDI_STAT_ITEMS; i++)
                percpu_counter_destroy(&bdi->bdi_stat[i]);