Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfs
[pandora-kernel.git] / block / blk-core.c
index 90f22cc..a2e58ee 100644 (file)
@@ -198,26 +198,13 @@ void blk_dump_rq_flags(struct request *rq, char *msg)
 }
 EXPORT_SYMBOL(blk_dump_rq_flags);
 
-/*
- * Make sure that plugs that were pending when this function was entered,
- * are now complete and requests pushed to the queue.
-*/
-static inline void queue_sync_plugs(struct request_queue *q)
-{
-       /*
-        * If the current process is plugged and has barriers submitted,
-        * we will livelock if we don't unplug first.
-        */
-       blk_flush_plug(current);
-}
-
 static void blk_delay_work(struct work_struct *work)
 {
        struct request_queue *q;
 
        q = container_of(work, struct request_queue, delay_work.work);
        spin_lock_irq(q->queue_lock);
-       __blk_run_queue(q, false);
+       __blk_run_queue(q);
        spin_unlock_irq(q->queue_lock);
 }
 
@@ -233,7 +220,8 @@ static void blk_delay_work(struct work_struct *work)
  */
 void blk_delay_queue(struct request_queue *q, unsigned long msecs)
 {
-       schedule_delayed_work(&q->delay_work, msecs_to_jiffies(msecs));
+       queue_delayed_work(kblockd_workqueue, &q->delay_work,
+                               msecs_to_jiffies(msecs));
 }
 EXPORT_SYMBOL(blk_delay_queue);
 
@@ -251,7 +239,7 @@ void blk_start_queue(struct request_queue *q)
        WARN_ON(!irqs_disabled());
 
        queue_flag_clear(QUEUE_FLAG_STOPPED, q);
-       __blk_run_queue(q, false);
+       __blk_run_queue(q);
 }
 EXPORT_SYMBOL(blk_start_queue);
 
@@ -298,37 +286,41 @@ void blk_sync_queue(struct request_queue *q)
 {
        del_timer_sync(&q->timeout);
        cancel_delayed_work_sync(&q->delay_work);
-       queue_sync_plugs(q);
 }
 EXPORT_SYMBOL(blk_sync_queue);
 
 /**
  * __blk_run_queue - run a single device queue
  * @q: The queue to run
- * @force_kblockd: Don't run @q->request_fn directly.  Use kblockd.
  *
  * Description:
  *    See @blk_run_queue. This variant must be called with the queue lock
  *    held and interrupts disabled.
- *
  */
-void __blk_run_queue(struct request_queue *q, bool force_kblockd)
+void __blk_run_queue(struct request_queue *q)
 {
        if (unlikely(blk_queue_stopped(q)))
                return;
 
-       /*
-        * Only recurse once to avoid overrunning the stack, let the unplug
-        * handling reinvoke the handler shortly if we already got there.
-        */
-       if (!force_kblockd && !queue_flag_test_and_set(QUEUE_FLAG_REENTER, q)) {
-               q->request_fn(q);
-               queue_flag_clear(QUEUE_FLAG_REENTER, q);
-       } else
-               queue_delayed_work(kblockd_workqueue, &q->delay_work, 0);
+       q->request_fn(q);
 }
 EXPORT_SYMBOL(__blk_run_queue);
 
+/**
+ * blk_run_queue_async - run a single device queue in workqueue context
+ * @q: The queue to run
+ *
+ * Description:
+ *    Tells kblockd to perform the equivalent of @blk_run_queue on behalf
+ *    of us.
+ */
+void blk_run_queue_async(struct request_queue *q)
+{
+       if (likely(!blk_queue_stopped(q)))
+               queue_delayed_work(kblockd_workqueue, &q->delay_work, 0);
+}
+EXPORT_SYMBOL(blk_run_queue_async);
+
 /**
  * blk_run_queue - run a single device queue
  * @q: The queue to run
@@ -342,7 +334,7 @@ void blk_run_queue(struct request_queue *q)
        unsigned long flags;
 
        spin_lock_irqsave(q->queue_lock, flags);
-       __blk_run_queue(q, false);
+       __blk_run_queue(q);
        spin_unlock_irqrestore(q->queue_lock, flags);
 }
 EXPORT_SYMBOL(blk_run_queue);
@@ -991,7 +983,7 @@ void blk_insert_request(struct request_queue *q, struct request *rq,
                blk_queue_end_tag(q, rq);
 
        add_acct_request(q, rq, where);
-       __blk_run_queue(q, false);
+       __blk_run_queue(q);
        spin_unlock_irqrestore(q->queue_lock, flags);
 }
 EXPORT_SYMBOL(blk_insert_request);
@@ -1311,7 +1303,15 @@ get_rq:
 
        plug = current->plug;
        if (plug) {
-               if (!plug->should_sort && !list_empty(&plug->list)) {
+               /*
+                * If this is the first request added after a plug, fire
+                * of a plug trace. If others have been added before, check
+                * if we have multiple devices in this plug. If so, make a
+                * note to sort the list before dispatch.
+                */
+               if (list_empty(&plug->list))
+                       trace_block_plug(q);
+               else if (!plug->should_sort) {
                        struct request *__rq;
 
                        __rq = list_entry_rq(plug->list.prev);
@@ -1327,7 +1327,7 @@ get_rq:
        } else {
                spin_lock_irq(q->queue_lock);
                add_acct_request(q, req, where);
-               __blk_run_queue(q, false);
+               __blk_run_queue(q);
 out_unlock:
                spin_unlock_irq(q->queue_lock);
        }
@@ -2644,6 +2644,7 @@ void blk_start_plug(struct blk_plug *plug)
 
        plug->magic = PLUG_MAGIC;
        INIT_LIST_HEAD(&plug->list);
+       INIT_LIST_HEAD(&plug->cb_list);
        plug->should_sort = 0;
 
        /*
@@ -2668,33 +2669,93 @@ static int plug_rq_cmp(void *priv, struct list_head *a, struct list_head *b)
        return !(rqa->q <= rqb->q);
 }
 
-static void flush_plug_list(struct blk_plug *plug)
+/*
+ * If 'from_schedule' is true, then postpone the dispatch of requests
+ * until a safe kblockd context. We due this to avoid accidental big
+ * additional stack usage in driver dispatch, in places where the originally
+ * plugger did not intend it.
+ */
+static void queue_unplugged(struct request_queue *q, unsigned int depth,
+                           bool from_schedule)
+       __releases(q->queue_lock)
+{
+       trace_block_unplug(q, depth, !from_schedule);
+
+       /*
+        * If we are punting this to kblockd, then we can safely drop
+        * the queue_lock before waking kblockd (which needs to take
+        * this lock).
+        */
+       if (from_schedule) {
+               spin_unlock(q->queue_lock);
+               blk_run_queue_async(q);
+       } else {
+               __blk_run_queue(q);
+               spin_unlock(q->queue_lock);
+       }
+
+}
+
+static void flush_plug_callbacks(struct blk_plug *plug)
+{
+       LIST_HEAD(callbacks);
+
+       if (list_empty(&plug->cb_list))
+               return;
+
+       list_splice_init(&plug->cb_list, &callbacks);
+
+       while (!list_empty(&callbacks)) {
+               struct blk_plug_cb *cb = list_first_entry(&callbacks,
+                                                         struct blk_plug_cb,
+                                                         list);
+               list_del(&cb->list);
+               cb->callback(cb);
+       }
+}
+
+void blk_flush_plug_list(struct blk_plug *plug, bool from_schedule)
 {
        struct request_queue *q;
        unsigned long flags;
        struct request *rq;
+       LIST_HEAD(list);
+       unsigned int depth;
 
        BUG_ON(plug->magic != PLUG_MAGIC);
 
+       flush_plug_callbacks(plug);
        if (list_empty(&plug->list))
                return;
 
-       if (plug->should_sort)
-               list_sort(NULL, &plug->list, plug_rq_cmp);
+       list_splice_init(&plug->list, &list);
+
+       if (plug->should_sort) {
+               list_sort(NULL, &list, plug_rq_cmp);
+               plug->should_sort = 0;
+       }
 
        q = NULL;
+       depth = 0;
+
+       /*
+        * Save and disable interrupts here, to avoid doing it for every
+        * queue lock we have to take.
+        */
        local_irq_save(flags);
-       while (!list_empty(&plug->list)) {
-               rq = list_entry_rq(plug->list.next);
+       while (!list_empty(&list)) {
+               rq = list_entry_rq(list.next);
                list_del_init(&rq->queuelist);
                BUG_ON(!(rq->cmd_flags & REQ_ON_PLUG));
                BUG_ON(!rq->q);
                if (rq->q != q) {
-                       if (q) {
-                               __blk_run_queue(q, false);
-                               spin_unlock(q->queue_lock);
-                       }
+                       /*
+                        * This drops the queue lock
+                        */
+                       if (q)
+                               queue_unplugged(q, depth, from_schedule);
                        q = rq->q;
+                       depth = 0;
                        spin_lock(q->queue_lock);
                }
                rq->cmd_flags &= ~REQ_ON_PLUG;
@@ -2706,38 +2767,27 @@ static void flush_plug_list(struct blk_plug *plug)
                        __elv_add_request(q, rq, ELEVATOR_INSERT_FLUSH);
                else
                        __elv_add_request(q, rq, ELEVATOR_INSERT_SORT_MERGE);
-       }
 
-       if (q) {
-               __blk_run_queue(q, false);
-               spin_unlock(q->queue_lock);
+               depth++;
        }
 
-       BUG_ON(!list_empty(&plug->list));
-       local_irq_restore(flags);
-}
-
-static void __blk_finish_plug(struct task_struct *tsk, struct blk_plug *plug)
-{
-       flush_plug_list(plug);
+       /*
+        * This drops the queue lock
+        */
+       if (q)
+               queue_unplugged(q, depth, from_schedule);
 
-       if (plug == tsk->plug)
-               tsk->plug = NULL;
+       local_irq_restore(flags);
 }
 
 void blk_finish_plug(struct blk_plug *plug)
 {
-       if (plug)
-               __blk_finish_plug(current, plug);
-}
-EXPORT_SYMBOL(blk_finish_plug);
+       blk_flush_plug_list(plug, false);
 
-void __blk_flush_plug(struct task_struct *tsk, struct blk_plug *plug)
-{
-       __blk_finish_plug(tsk, plug);
-       tsk->plug = plug;
+       if (plug == current->plug)
+               current->plug = NULL;
 }
-EXPORT_SYMBOL(__blk_flush_plug);
+EXPORT_SYMBOL(blk_finish_plug);
 
 int __init blk_dev_init(void)
 {