1 files changed, 188 insertions, 93 deletions

diff --git a/block/blk-core.c b/block/blk-core.c
index 1f61b74867e4..3c923a7aeb56 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -29,11 +29,13 @@
 #include <linux/fault-inject.h>
 #include <linux/list_sort.h>
 #include <linux/delay.h>
+#include <linux/ratelimit.h>
 
 #define CREATE_TRACE_POINTS
 #include <trace/events/block.h>
 
 #include "blk.h"
+#include "blk-cgroup.h"
 
 EXPORT_TRACEPOINT_SYMBOL_GPL(block_bio_remap);
 EXPORT_TRACEPOINT_SYMBOL_GPL(block_rq_remap);
@@ -280,7 +282,7 @@ EXPORT_SYMBOL(blk_stop_queue);
  *
  *     This function does not cancel any asynchronous activity arising
  *     out of elevator or throttling code. That would require elevaotor_exit()
- *     and blk_throtl_exit() to be called with queue lock initialized.
+ *     and blkcg_exit_queue() to be called with queue lock initialized.
  *
  */
 void blk_sync_queue(struct request_queue *q)
@@ -365,17 +367,23 @@ void blk_drain_queue(struct request_queue *q, bool drain_all)
 
 		spin_lock_irq(q->queue_lock);
 
-		elv_drain_elevator(q);
-		if (drain_all)
-			blk_throtl_drain(q);
+		/*
+		 * The caller might be trying to drain @q before its
+		 * elevator is initialized.
+		 */
+		if (q->elevator)
+			elv_drain_elevator(q);
+
+		blkcg_drain_queue(q);
 
 		/*
 		 * This function might be called on a queue which failed
-		 * driver init after queue creation.  Some drivers
-		 * (e.g. fd) get unhappy in such cases.  Kick queue iff
-		 * dispatch queue has something on it.
+		 * driver init after queue creation or is not yet fully
+		 * active yet.  Some drivers (e.g. fd and loop) get unhappy
+		 * in such cases.  Kick queue iff dispatch queue has
+		 * something on it and @q has request_fn set.
 		 */
-		if (!list_empty(&q->queue_head))
+		if (!list_empty(&q->queue_head) && q->request_fn)
 			__blk_run_queue(q);
 
 		drain |= q->rq.elvpriv;
@@ -403,6 +411,49 @@ void blk_drain_queue(struct request_queue *q, bool drain_all)
 }
 
 /**
+ * blk_queue_bypass_start - enter queue bypass mode
+ * @q: queue of interest
+ *
+ * In bypass mode, only the dispatch FIFO queue of @q is used.  This
+ * function makes @q enter bypass mode and drains all requests which were
+ * throttled or issued before.  On return, it's guaranteed that no request
+ * is being throttled or has ELVPRIV set and blk_queue_bypass() %true
+ * inside queue or RCU read lock.
+ */
+void blk_queue_bypass_start(struct request_queue *q)
+{
+	bool drain;
+
+	spin_lock_irq(q->queue_lock);
+	drain = !q->bypass_depth++;
+	queue_flag_set(QUEUE_FLAG_BYPASS, q);
+	spin_unlock_irq(q->queue_lock);
+
+	if (drain) {
+		blk_drain_queue(q, false);
+		/* ensure blk_queue_bypass() is %true inside RCU read lock */
+		synchronize_rcu();
+	}
+}
+EXPORT_SYMBOL_GPL(blk_queue_bypass_start);
+
+/**
+ * blk_queue_bypass_end - leave queue bypass mode
+ * @q: queue of interest
+ *
+ * Leave bypass mode and restore the normal queueing behavior.
+ */
+void blk_queue_bypass_end(struct request_queue *q)
+{
+	spin_lock_irq(q->queue_lock);
+	if (!--q->bypass_depth)
+		queue_flag_clear(QUEUE_FLAG_BYPASS, q);
+	WARN_ON_ONCE(q->bypass_depth < 0);
+	spin_unlock_irq(q->queue_lock);
+}
+EXPORT_SYMBOL_GPL(blk_queue_bypass_end);
+
+/**
  * blk_cleanup_queue - shutdown a request queue
  * @q: request queue to shutdown
  *
@@ -418,6 +469,19 @@ void blk_cleanup_queue(struct request_queue *q)
 	queue_flag_set_unlocked(QUEUE_FLAG_DEAD, q);
 
 	spin_lock_irq(lock);
+
+	/*
+	 * Dead queue is permanently in bypass mode till released.  Note
+	 * that, unlike blk_queue_bypass_start(), we aren't performing
+	 * synchronize_rcu() after entering bypass mode to avoid the delay
+	 * as some drivers create and destroy a lot of queues while
+	 * probing.  This is still safe because blk_release_queue() will be
+	 * called only after the queue refcnt drops to zero and nothing,
+	 * RCU or not, would be traversing the queue by then.
+	 */
+	q->bypass_depth++;
+	queue_flag_set(QUEUE_FLAG_BYPASS, q);
+
 	queue_flag_set(QUEUE_FLAG_NOMERGES, q);
 	queue_flag_set(QUEUE_FLAG_NOXMERGES, q);
 	queue_flag_set(QUEUE_FLAG_DEAD, q);
@@ -428,13 +492,8 @@ void blk_cleanup_queue(struct request_queue *q)
 	spin_unlock_irq(lock);
 	mutex_unlock(&q->sysfs_lock);
 
-	/*
-	 * Drain all requests queued before DEAD marking.  The caller might
-	 * be trying to tear down @q before its elevator is initialized, in
-	 * which case we don't want to call into draining.
-	 */
-	if (q->elevator)
-		blk_drain_queue(q, true);
+	/* drain all requests queued before DEAD marking */
+	blk_drain_queue(q, true);
 
 	/* @q won't process any more request, flush async actions */
 	del_timer_sync(&q->backing_dev_info.laptop_mode_wb_timer);
@@ -498,14 +557,15 @@ struct request_queue *blk_alloc_queue_node(gfp_t gfp_mask, int node_id)
 	if (err)
 		goto fail_id;
 
-	if (blk_throtl_init(q))
-		goto fail_id;
-
 	setup_timer(&q->backing_dev_info.laptop_mode_wb_timer,
 		    laptop_mode_timer_fn, (unsigned long) q);
 	setup_timer(&q->timeout, blk_rq_timed_out_timer, (unsigned long) q);
+	INIT_LIST_HEAD(&q->queue_head);
 	INIT_LIST_HEAD(&q->timeout_list);
 	INIT_LIST_HEAD(&q->icq_list);
+#ifdef CONFIG_BLK_CGROUP
+	INIT_LIST_HEAD(&q->blkg_list);
+#endif
 	INIT_LIST_HEAD(&q->flush_queue[0]);
 	INIT_LIST_HEAD(&q->flush_queue[1]);
 	INIT_LIST_HEAD(&q->flush_data_in_flight);
@@ -522,6 +582,18 @@ struct request_queue *blk_alloc_queue_node(gfp_t gfp_mask, int node_id)
 	 */
 	q->queue_lock = &q->__queue_lock;
 
+	/*
+	 * A queue starts its life with bypass turned on to avoid
+	 * unnecessary bypass on/off overhead and nasty surprises during
+	 * init.  The initial bypass will be finished at the end of
+	 * blk_init_allocated_queue().
+	 */
+	q->bypass_depth = 1;
+	__set_bit(QUEUE_FLAG_BYPASS, &q->queue_flags);
+
+	if (blkcg_init_queue(q))
+		goto fail_id;
+
 	return q;
 
 fail_id:
@@ -614,15 +686,15 @@ blk_init_allocated_queue(struct request_queue *q, request_fn_proc *rfn,
 
 	q->sg_reserved_size = INT_MAX;
 
-	/*
-	 * all done
-	 */
-	if (!elevator_init(q, NULL)) {
-		blk_queue_congestion_threshold(q);
-		return q;
-	}
+	/* init elevator */
+	if (elevator_init(q, NULL))
+		return NULL;
 
-	return NULL;
+	blk_queue_congestion_threshold(q);
+
+	/* all done, end the initial bypass */
+	blk_queue_bypass_end(q);
+	return q;
 }
 EXPORT_SYMBOL(blk_init_allocated_queue);
 
@@ -648,33 +720,6 @@ static inline void blk_free_request(struct request_queue *q, struct request *rq)
 	mempool_free(rq, q->rq.rq_pool);
 }
 
-static struct request *
-blk_alloc_request(struct request_queue *q, struct io_cq *icq,
-		  unsigned int flags, gfp_t gfp_mask)
-{
-	struct request *rq = mempool_alloc(q->rq.rq_pool, gfp_mask);
-
-	if (!rq)
-		return NULL;
-
-	blk_rq_init(q, rq);
-
-	rq->cmd_flags = flags | REQ_ALLOCED;
-
-	if (flags & REQ_ELVPRIV) {
-		rq->elv.icq = icq;
-		if (unlikely(elv_set_request(q, rq, gfp_mask))) {
-			mempool_free(rq, q->rq.rq_pool);
-			return NULL;
-		}
-		/* @rq->elv.icq holds on to io_context until @rq is freed */
-		if (icq)
-			get_io_context(icq->ioc);
-	}
-
-	return rq;
-}
-
 /*
  * ioc_batching returns true if the ioc is a valid batching request and
  * should be given priority access to a request.
@@ -763,6 +808,22 @@ static bool blk_rq_should_init_elevator(struct bio *bio)
 }
 
 /**
+ * rq_ioc - determine io_context for request allocation
+ * @bio: request being allocated is for this bio (can be %NULL)
+ *
+ * Determine io_context to use for request allocation for @bio.  May return
+ * %NULL if %current->io_context doesn't exist.
+ */
+static struct io_context *rq_ioc(struct bio *bio)
+{
+#ifdef CONFIG_BLK_CGROUP
+	if (bio && bio->bi_ioc)
+		return bio->bi_ioc;
+#endif
+	return current->io_context;
+}
+
+/**
  * get_request - get a free request
  * @q: request_queue to allocate request from
  * @rw_flags: RW and SYNC flags
@@ -779,7 +840,7 @@ static bool blk_rq_should_init_elevator(struct bio *bio)
 static struct request *get_request(struct request_queue *q, int rw_flags,
 				   struct bio *bio, gfp_t gfp_mask)
 {
-	struct request *rq = NULL;
+	struct request *rq;
 	struct request_list *rl = &q->rq;
 	struct elevator_type *et;
 	struct io_context *ioc;
@@ -789,7 +850,7 @@ static struct request *get_request(struct request_queue *q, int rw_flags,
 	int may_queue;
 retry:
 	et = q->elevator->type;
-	ioc = current->io_context;
+	ioc = rq_ioc(bio);
 
 	if (unlikely(blk_queue_dead(q)))
 		return NULL;
@@ -808,7 +869,7 @@ retry:
 			 */
 			if (!ioc && !retried) {
 				spin_unlock_irq(q->queue_lock);
-				create_io_context(current, gfp_mask, q->node);
+				create_io_context(gfp_mask, q->node);
 				spin_lock_irq(q->queue_lock);
 				retried = true;
 				goto retry;
@@ -831,7 +892,7 @@ retry:
 					 * process is not a "batcher", and not
 					 * exempted by the IO scheduler
 					 */
-					goto out;
+					return NULL;
 				}
 			}
 		}
@@ -844,7 +905,7 @@ retry:
 	 * allocated with any setting of ->nr_requests
 	 */
 	if (rl->count[is_sync] >= (3 * q->nr_requests / 2))
-		goto out;
+		return NULL;
 
 	rl->count[is_sync]++;
 	rl->starved[is_sync] = 0;
@@ -859,8 +920,7 @@ retry:
 	 * Also, lookup icq while holding queue_lock.  If it doesn't exist,
 	 * it will be created after releasing queue_lock.
 	 */
-	if (blk_rq_should_init_elevator(bio) &&
-	    !test_bit(QUEUE_FLAG_ELVSWITCH, &q->queue_flags)) {
+	if (blk_rq_should_init_elevator(bio) && !blk_queue_bypass(q)) {
 		rw_flags |= REQ_ELVPRIV;
 		rl->elvpriv++;
 		if (et->icq_cache && ioc)
@@ -871,41 +931,36 @@ retry:
 		rw_flags |= REQ_IO_STAT;
 	spin_unlock_irq(q->queue_lock);
 
-	/* create icq if missing */
-	if ((rw_flags & REQ_ELVPRIV) && unlikely(et->icq_cache && !icq)) {
-		icq = ioc_create_icq(q, gfp_mask);
-		if (!icq)
-			goto fail_icq;
-	}
-
-	rq = blk_alloc_request(q, icq, rw_flags, gfp_mask);
+	/* allocate and init request */
+	rq = mempool_alloc(q->rq.rq_pool, gfp_mask);
+	if (!rq)
+		goto fail_alloc;
 
-fail_icq:
-	if (unlikely(!rq)) {
-		/*
-		 * Allocation failed presumably due to memory. Undo anything
-		 * we might have messed up.
-		 *
-		 * Allocating task should really be put onto the front of the
-		 * wait queue, but this is pretty rare.
-		 */
-		spin_lock_irq(q->queue_lock);
-		freed_request(q, rw_flags);
+	blk_rq_init(q, rq);
+	rq->cmd_flags = rw_flags | REQ_ALLOCED;
+
+	/* init elvpriv */
+	if (rw_flags & REQ_ELVPRIV) {
+		if (unlikely(et->icq_cache && !icq)) {
+			create_io_context(gfp_mask, q->node);
+			ioc = rq_ioc(bio);
+			if (!ioc)
+				goto fail_elvpriv;
+
+			icq = ioc_create_icq(ioc, q, gfp_mask);
+			if (!icq)
+				goto fail_elvpriv;
+		}
 
-		/*
-		 * in the very unlikely event that allocation failed and no
-		 * requests for this direction was pending, mark us starved
-		 * so that freeing of a request in the other direction will
-		 * notice us. another possible fix would be to split the
-		 * rq mempool into READ and WRITE
-		 */
-rq_starved:
-		if (unlikely(rl->count[is_sync] == 0))
-			rl->starved[is_sync] = 1;
+		rq->elv.icq = icq;
+		if (unlikely(elv_set_request(q, rq, bio, gfp_mask)))
+			goto fail_elvpriv;
 
-		goto out;
+		/* @rq->elv.icq holds io_context until @rq is freed */
+		if (icq)
+			get_io_context(icq->ioc);
 	}
-
+out:
 	/*
 	 * ioc may be NULL here, and ioc_batching will be false. That's
 	 * OK, if the queue is under the request limit then requests need
@@ -916,8 +971,48 @@ rq_starved:
 		ioc->nr_batch_requests--;
 
 	trace_block_getrq(q, bio, rw_flags & 1);
-out:
 	return rq;
+
+fail_elvpriv:
+	/*
+	 * elvpriv init failed.  ioc, icq and elvpriv aren't mempool backed
+	 * and may fail indefinitely under memory pressure and thus
+	 * shouldn't stall IO.  Treat this request as !elvpriv.  This will
+	 * disturb iosched and blkcg but weird is bettern than dead.
+	 */
+	printk_ratelimited(KERN_WARNING "%s: request aux data allocation failed, iosched may be disturbed\n",
+			   dev_name(q->backing_dev_info.dev));
+
+	rq->cmd_flags &= ~REQ_ELVPRIV;
+	rq->elv.icq = NULL;
+
+	spin_lock_irq(q->queue_lock);
+	rl->elvpriv--;
+	spin_unlock_irq(q->queue_lock);
+	goto out;
+
+fail_alloc:
+	/*
+	 * Allocation failed presumably due to memory. Undo anything we
+	 * might have messed up.
+	 *
+	 * Allocating task should really be put onto the front of the wait
+	 * queue, but this is pretty rare.
+	 */
+	spin_lock_irq(q->queue_lock);
+	freed_request(q, rw_flags);
+
+	/*
+	 * in the very unlikely event that allocation failed and no
+	 * requests for this direction was pending, mark us starved so that
+	 * freeing of a request in the other direction will notice
+	 * us. another possible fix would be to split the rq mempool into
+	 * READ and WRITE
+	 */
+rq_starved:
+	if (unlikely(rl->count[is_sync] == 0))
+		rl->starved[is_sync] = 1;
+	return NULL;
 }
 
 /**
@@ -961,7 +1056,7 @@ static struct request *get_request_wait(struct request_queue *q, int rw_flags,
 		 * up to a big batch of them for a small period time.
 		 * See ioc_batching, ioc_set_batching
 		 */
-		create_io_context(current, GFP_NOIO, q->node);
+		create_io_context(GFP_NOIO, q->node);
 		ioc_set_batching(q, current->io_context);
 
 		spin_lock_irq(q->queue_lock);