[RFC][PATCH 09/16] writeback: do not lose wake-ups in bdi threads

[Artem Bityutskiy <dedekind1@gmail.com>]

From: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>

The bdi threads ('bdi_writeback_thread()') can lose wake-ups if,
for example, 'bdi_queue_work()' is executed while after the bdi
thread finished 'wb_do_writeback()' but before it has called
'schedule_timeout_interruptible()'.

To fix this issue, we have to check whether we have works to
process after we change the task state to 'TASK_INTERRUPTIBLE'.

Also, 'bdi_writeback_thread()' inconsistently handles the cases
when 'dirty_writeback_interval' is zero and non-zero. But there is
no fundamental difference between these cases, so they have to be
handled the same way, which this patch also does.

This patch also removes strange 'list_empty_careful()' call.

Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
---
 fs/fs-writeback.c |   17 +++++++++--------
 1 files changed, 9 insertions(+), 8 deletions(-)

diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
index 3fc5194..f045450 100644
--- a/fs/fs-writeback.c
+++ b/fs/fs-writeback.c
@@ -848,17 +848,18 @@ int bdi_writeback_thread(void *data)
 				break;
 		}
 
-		if (dirty_writeback_interval) {
-			wait_jiffies = msecs_to_jiffies(dirty_writeback_interval * 10);
-			schedule_timeout_interruptible(wait_jiffies);
-		} else {
-			set_current_state(TASK_INTERRUPTIBLE);
-			if (list_empty_careful(&wb->bdi->work_list) &&
-			    !kthread_should_stop())
-				schedule();
+		set_current_state(TASK_INTERRUPTIBLE);
+		if (!list_empty(&bdi->work_list)) {
 			__set_current_state(TASK_RUNNING);
+			continue;
 		}
 
+		if (dirty_writeback_interval) {
+			wait_jiffies = msecs_to_jiffies(dirty_writeback_interval * 10);
+			schedule_timeout(wait_jiffies);
+		} else
+			schedule();
+
 		try_to_freeze();
 	}
 
-- 
1.7.1.1