X-Git-Url: http://ftp.safe.ca/?a=blobdiff_plain;f=block%2Fcfq-iosched.c;h=0e3814b662af3f7f1bd7c2f074d95150535ab132;hb=06d2188644c85c56d243efab914f368d1d23c4a3;hp=baef5fc7cff8b5ca3d184e3d48efdd4be7376203;hpb=0a31bd5f2bbb6473ef9d24f0063ca91cfa678b64;p=safe%2Fjmp%2Flinux-2.6

diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
index baef5fc..0e3814b 100644
--- a/block/cfq-iosched.c
+++ b/block/cfq-iosched.c
@@ -11,24 +11,27 @@
 #include <linux/elevator.h>
 #include <linux/rbtree.h>
 #include <linux/ioprio.h>
+#include <linux/blktrace_api.h>
 
 /*
  * tunables
  */
-static const int cfq_quantum = 4;		/* max queue in one round of service */
+/* max queue in one round of service */
+static const int cfq_quantum = 4;
 static const int cfq_fifo_expire[2] = { HZ / 4, HZ / 8 };
-static const int cfq_back_max = 16 * 1024;	/* maximum backwards seek, in KiB */
-static const int cfq_back_penalty = 2;		/* penalty of a backwards seek */
-
+/* maximum backwards seek, in KiB */
+static const int cfq_back_max = 16 * 1024;
+/* penalty of a backwards seek */
+static const int cfq_back_penalty = 2;
 static const int cfq_slice_sync = HZ / 10;
 static int cfq_slice_async = HZ / 25;
 static const int cfq_slice_async_rq = 2;
 static int cfq_slice_idle = HZ / 125;
 
 /*
- * grace period before allowing idle class to get disk access
+ * offset from end of service tree
  */
-#define CFQ_IDLE_GRACE		(HZ / 10)
+#define CFQ_IDLE_DELAY		(HZ / 5)
 
 /*
  * below this threshold, we consider thinktime immediate
@@ -36,23 +39,23 @@ static int cfq_slice_idle = HZ / 125;
 #define CFQ_MIN_TT		(2)
 
 #define CFQ_SLICE_SCALE		(5)
+#define CFQ_HW_QUEUE_MIN	(5)
 
-#define RQ_CIC(rq)		((struct cfq_io_context*)(rq)->elevator_private)
-#define RQ_CFQQ(rq)		((rq)->elevator_private2)
+#define RQ_CIC(rq)		\
+	((struct cfq_io_context *) (rq)->elevator_private)
+#define RQ_CFQQ(rq)		(struct cfq_queue *) ((rq)->elevator_private2)
 
 static struct kmem_cache *cfq_pool;
 static struct kmem_cache *cfq_ioc_pool;
 
 static DEFINE_PER_CPU(unsigned long, ioc_count);
 static struct completion *ioc_gone;
+static DEFINE_SPINLOCK(ioc_gone_lock);
 
 #define CFQ_PRIO_LISTS		IOPRIO_BE_NR
 #define cfq_class_idle(cfqq)	((cfqq)->ioprio_class == IOPRIO_CLASS_IDLE)
 #define cfq_class_rt(cfqq)	((cfqq)->ioprio_class == IOPRIO_CLASS_RT)
 
-#define ASYNC			(0)
-#define SYNC			(1)
-
 #define sample_valid(samples)	((samples) > 80)
 
 /*
@@ -68,20 +71,80 @@ struct cfq_rb_root {
 #define CFQ_RB_ROOT	(struct cfq_rb_root) { RB_ROOT, NULL, }
 
 /*
+ * Per process-grouping structure
+ */
+struct cfq_queue {
+	/* reference count */
+	atomic_t ref;
+	/* various state flags, see below */
+	unsigned int flags;
+	/* parent cfq_data */
+	struct cfq_data *cfqd;
+	/* service_tree member */
+	struct rb_node rb_node;
+	/* service_tree key */
+	unsigned long rb_key;
+	/* prio tree member */
+	struct rb_node p_node;
+	/* prio tree root we belong to, if any */
+	struct rb_root *p_root;
+	/* sorted list of pending requests */
+	struct rb_root sort_list;
+	/* if fifo isn't expired, next request to serve */
+	struct request *next_rq;
+	/* requests queued in sort_list */
+	int queued[2];
+	/* currently allocated requests */
+	int allocated[2];
+	/* fifo list of requests in sort_list */
+	struct list_head fifo;
+
+	unsigned long slice_end;
+	long slice_resid;
+	unsigned int slice_dispatch;
+
+	/* pending metadata requests */
+	int meta_pending;
+	/* number of requests that are on the dispatch list or inside driver */
+	int dispatched;
+
+	/* io prio of this group */
+	unsigned short ioprio, org_ioprio;
+	unsigned short ioprio_class, org_ioprio_class;
+
+	pid_t pid;
+};
+
+/*
  * Per block device queue structure
  */
 struct cfq_data {
-	request_queue_t *queue;
+	struct request_queue *queue;
 
 	/*
 	 * rr list of queues with requests and the count of them
 	 */
 	struct cfq_rb_root service_tree;
+
+	/*
+	 * Each priority tree is sorted by next_request position.  These
+	 * trees are used when determining if two or more queues are
+	 * interleaving requests (see cfq_close_cooperator).
+	 */
+	struct rb_root prio_trees[CFQ_PRIO_LISTS];
+
 	unsigned int busy_queues;
 
-	int rq_in_driver;
+	int rq_in_driver[2];
 	int sync_flight;
+
+	/*
+	 * queue-depth detection
+	 */
+	int rq_queued;
 	int hw_tag;
+	int hw_tag_samples;
+	int rq_in_driver_peak;
 
 	/*
 	 * idle window management
@@ -92,10 +155,13 @@ struct cfq_data {
 	struct cfq_queue *active_queue;
 	struct cfq_io_context *active_cic;
 
-	struct timer_list idle_class_timer;
+	/*
+	 * async queue for each priority case
+	 */
+	struct cfq_queue *async_cfqq[2][IOPRIO_BE_NR];
+	struct cfq_queue *async_idle_cfqq;
 
 	sector_t last_position;
-	unsigned long last_end_request;
 
 	/*
 	 * tunables, see top of file
@@ -110,98 +176,67 @@ struct cfq_data {
 
 	struct list_head cic_list;
 
-	sector_t new_seek_mean;
-	u64 new_seek_total;
-};
-
-/*
- * Per process-grouping structure
- */
-struct cfq_queue {
-	/* reference count */
-	atomic_t ref;
-	/* parent cfq_data */
-	struct cfq_data *cfqd;
-	/* service_tree member */
-	struct rb_node rb_node;
-	/* service_tree key */
-	unsigned long rb_key;
-	/* sorted list of pending requests */
-	struct rb_root sort_list;
-	/* if fifo isn't expired, next request to serve */
-	struct request *next_rq;
-	/* requests queued in sort_list */
-	int queued[2];
-	/* currently allocated requests */
-	int allocated[2];
-	/* pending metadata requests */
-	int meta_pending;
-	/* fifo list of requests in sort_list */
-	struct list_head fifo;
-
-	unsigned long slice_end;
-	long slice_resid;
-
-	/* number of requests that are on the dispatch list or inside driver */
-	int dispatched;
-
-	/* io prio of this group */
-	unsigned short ioprio, org_ioprio;
-	unsigned short ioprio_class, org_ioprio_class;
-
-	/* various state flags, see below */
-	unsigned int flags;
-
-	sector_t last_request_pos;
+	/*
+	 * Fallback dummy cfqq for extreme OOM conditions
+	 */
+	struct cfq_queue oom_cfqq;
 };
 
 enum cfqq_state_flags {
 	CFQ_CFQQ_FLAG_on_rr = 0,	/* on round-robin busy list */
 	CFQ_CFQQ_FLAG_wait_request,	/* waiting for a request */
-	CFQ_CFQQ_FLAG_must_alloc,	/* must be allowed rq alloc */
+	CFQ_CFQQ_FLAG_must_dispatch,	/* must be allowed a dispatch */
 	CFQ_CFQQ_FLAG_must_alloc_slice,	/* per-slice must_alloc flag */
-	CFQ_CFQQ_FLAG_must_dispatch,	/* must dispatch, even if expired */
 	CFQ_CFQQ_FLAG_fifo_expire,	/* FIFO checked in this slice */
 	CFQ_CFQQ_FLAG_idle_window,	/* slice idling enabled */
 	CFQ_CFQQ_FLAG_prio_changed,	/* task priority has changed */
-	CFQ_CFQQ_FLAG_queue_new,	/* queue never been serviced */
 	CFQ_CFQQ_FLAG_slice_new,	/* no requests dispatched in slice */
 	CFQ_CFQQ_FLAG_sync,		/* synchronous queue */
+	CFQ_CFQQ_FLAG_coop,		/* has done a coop jump of the queue */
 };
 
 #define CFQ_CFQQ_FNS(name)						\
 static inline void cfq_mark_cfqq_##name(struct cfq_queue *cfqq)		\
 {									\
-	cfqq->flags |= (1 << CFQ_CFQQ_FLAG_##name);			\
+	(cfqq)->flags |= (1 << CFQ_CFQQ_FLAG_##name);			\
 }									\
 static inline void cfq_clear_cfqq_##name(struct cfq_queue *cfqq)	\
 {									\
-	cfqq->flags &= ~(1 << CFQ_CFQQ_FLAG_##name);			\
+	(cfqq)->flags &= ~(1 << CFQ_CFQQ_FLAG_##name);			\
 }									\
 static inline int cfq_cfqq_##name(const struct cfq_queue *cfqq)		\
 {									\
-	return (cfqq->flags & (1 << CFQ_CFQQ_FLAG_##name)) != 0;	\
+	return ((cfqq)->flags & (1 << CFQ_CFQQ_FLAG_##name)) != 0;	\
 }
 
 CFQ_CFQQ_FNS(on_rr);
 CFQ_CFQQ_FNS(wait_request);
-CFQ_CFQQ_FNS(must_alloc);
-CFQ_CFQQ_FNS(must_alloc_slice);
 CFQ_CFQQ_FNS(must_dispatch);
+CFQ_CFQQ_FNS(must_alloc_slice);
 CFQ_CFQQ_FNS(fifo_expire);
 CFQ_CFQQ_FNS(idle_window);
 CFQ_CFQQ_FNS(prio_changed);
-CFQ_CFQQ_FNS(queue_new);
 CFQ_CFQQ_FNS(slice_new);
 CFQ_CFQQ_FNS(sync);
+CFQ_CFQQ_FNS(coop);
 #undef CFQ_CFQQ_FNS
 
-static void cfq_dispatch_insert(request_queue_t *, struct request *);
+#define cfq_log_cfqq(cfqd, cfqq, fmt, args...)	\
+	blk_add_trace_msg((cfqd)->queue, "cfq%d " fmt, (cfqq)->pid, ##args)
+#define cfq_log(cfqd, fmt, args...)	\
+	blk_add_trace_msg((cfqd)->queue, "cfq " fmt, ##args)
+
+static void cfq_dispatch_insert(struct request_queue *, struct request *);
 static struct cfq_queue *cfq_get_queue(struct cfq_data *, int,
-				       struct task_struct *, gfp_t);
-static struct cfq_io_context *cfq_cic_rb_lookup(struct cfq_data *,
+				       struct io_context *, gfp_t);
+static struct cfq_io_context *cfq_cic_lookup(struct cfq_data *,
 						struct io_context *);
 
+static inline int rq_in_driver(struct cfq_data *cfqd)
+{
+	return cfqd->rq_in_driver[0] + cfqd->rq_in_driver[1];
+}
+
 static inline struct cfq_queue *cic_to_cfqq(struct cfq_io_context *cic,
 					    int is_sync)
 {
@@ -220,7 +255,7 @@ static inline void cic_set_cfqq(struct cfq_io_context *cic,
  */
 static inline int cfq_bio_sync(struct bio *bio)
 {
-	if (bio_data_dir(bio) == READ || bio_sync(bio))
+	if (bio_data_dir(bio) == READ || bio_rw_flagged(bio, BIO_RW_SYNCIO))
 		return 1;
 
 	return 0;
@@ -232,11 +267,13 @@ static inline int cfq_bio_sync(struct bio *bio)
  */
 static inline void cfq_schedule_dispatch(struct cfq_data *cfqd)
 {
-	if (cfqd->busy_queues)
-		kblockd_schedule_work(&cfqd->unplug_work);
+	if (cfqd->busy_queues) {
+		cfq_log(cfqd, "schedule dispatch");
+		kblockd_schedule_work(cfqd->queue, &cfqd->unplug_work);
+	}
 }
 
-static int cfq_queue_empty(request_queue_t *q)
+static int cfq_queue_empty(struct request_queue *q)
 {
 	struct cfq_data *cfqd = q->elevator->elevator_data;
 
@@ -268,6 +305,7 @@ static inline void
 cfq_set_prio_slice(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 {
 	cfqq->slice_end = cfq_prio_to_slice(cfqd, cfqq) + jiffies;
+	cfq_log_cfqq(cfqd, cfqq, "set_slice=%lu", cfqq->slice_end - jiffies);
 }
 
 /*
@@ -313,8 +351,8 @@ cfq_choose_req(struct cfq_data *cfqd, struct request *rq1, struct request *rq2)
 	else if (rq_is_meta(rq2) && !rq_is_meta(rq1))
 		return rq2;
 
-	s1 = rq1->sector;
-	s2 = rq2->sector;
+	s1 = blk_rq_pos(rq1);
+	s2 = blk_rq_pos(rq2);
 
 	last = cfqd->last_position;
 
@@ -383,21 +421,28 @@ cfq_choose_req(struct cfq_data *cfqd, struct request *rq1, struct request *rq2)
 /*
  * The below is leftmost cache rbtree addon
  */
-static struct rb_node *cfq_rb_first(struct cfq_rb_root *root)
+static struct cfq_queue *cfq_rb_first(struct cfq_rb_root *root)
 {
 	if (!root->left)
 		root->left = rb_first(&root->rb);
 
-	return root->left;
+	if (root->left)
+		return rb_entry(root->left, struct cfq_queue, rb_node);
+
+	return NULL;
+}
+
+static void rb_erase_init(struct rb_node *n, struct rb_root *root)
+{
+	rb_erase(n, root);
+	RB_CLEAR_NODE(n);
 }
 
 static void cfq_rb_erase(struct rb_node *n, struct cfq_rb_root *root)
 {
 	if (root->left == n)
 		root->left = NULL;
-
-	rb_erase(n, &root->rb);
-	RB_CLEAR_NODE(n);
+	rb_erase_init(n, &root->rb);
 }
 
 /*
@@ -442,15 +487,23 @@ static unsigned long cfq_slice_offset(struct cfq_data *cfqd,
  * requests waiting to be processed. It is sorted in the order that
  * we will service the queues.
  */
-static void cfq_service_tree_add(struct cfq_data *cfqd,
-				    struct cfq_queue *cfqq, int add_front)
+static void cfq_service_tree_add(struct cfq_data *cfqd, struct cfq_queue *cfqq,
+				 int add_front)
 {
-	struct rb_node **p = &cfqd->service_tree.rb.rb_node;
-	struct rb_node *parent = NULL;
+	struct rb_node **p, *parent;
+	struct cfq_queue *__cfqq;
 	unsigned long rb_key;
 	int left;
 
-	if (!add_front) {
+	if (cfq_class_idle(cfqq)) {
+		rb_key = CFQ_IDLE_DELAY;
+		parent = rb_last(&cfqd->service_tree.rb);
+		if (parent && parent != &cfqq->rb_node) {
+			__cfqq = rb_entry(parent, struct cfq_queue, rb_node);
+			rb_key += __cfqq->rb_key;
+		} else
+			rb_key += jiffies;
+	} else if (!add_front) {
 		rb_key = cfq_slice_offset(cfqd, cfqq) + jiffies;
 		rb_key += cfqq->slice_resid;
 		cfqq->slice_resid = 0;
@@ -468,8 +521,9 @@ static void cfq_service_tree_add(struct cfq_data *cfqd,
 	}
 
 	left = 1;
+	parent = NULL;
+	p = &cfqd->service_tree.rb.rb_node;
 	while (*p) {
-		struct cfq_queue *__cfqq;
 		struct rb_node **n;
 
 		parent = *p;
@@ -507,6 +561,67 @@ static void cfq_service_tree_add(struct cfq_data *cfqd,
 	rb_insert_color(&cfqq->rb_node, &cfqd->service_tree.rb);
 }
 
+static struct cfq_queue *
+cfq_prio_tree_lookup(struct cfq_data *cfqd, struct rb_root *root,
+		     sector_t sector, struct rb_node **ret_parent,
+		     struct rb_node ***rb_link)
+{
+	struct rb_node **p, *parent;
+	struct cfq_queue *cfqq = NULL;
+
+	parent = NULL;
+	p = &root->rb_node;
+	while (*p) {
+		struct rb_node **n;
+
+		parent = *p;
+		cfqq = rb_entry(parent, struct cfq_queue, p_node);
+
+		/*
+		 * Sort strictly based on sector.  Smallest to the left,
+		 * largest to the right.
+		 */
+		if (sector > blk_rq_pos(cfqq->next_rq))
+			n = &(*p)->rb_right;
+		else if (sector < blk_rq_pos(cfqq->next_rq))
+			n = &(*p)->rb_left;
+		else
+			break;
+		p = n;
+		cfqq = NULL;
+	}
+
+	*ret_parent = parent;
+	if (rb_link)
+		*rb_link = p;
+	return cfqq;
+}
+
+static void cfq_prio_tree_add(struct cfq_data *cfqd, struct cfq_queue *cfqq)
+{
+	struct rb_node **p, *parent;
+	struct cfq_queue *__cfqq;
+
+	if (cfqq->p_root) {
+		rb_erase(&cfqq->p_node, cfqq->p_root);
+		cfqq->p_root = NULL;
+	}
+
+	if (cfq_class_idle(cfqq))
+		return;
+	if (!cfqq->next_rq)
+		return;
+
+	cfqq->p_root = &cfqd->prio_trees[cfqq->org_ioprio];
+	__cfqq = cfq_prio_tree_lookup(cfqd, cfqq->p_root,
+				      blk_rq_pos(cfqq->next_rq), &parent, &p);
+	if (!__cfqq) {
+		rb_link_node(&cfqq->p_node, parent, p);
+		rb_insert_color(&cfqq->p_node, cfqq->p_root);
+	} else
+		cfqq->p_root = NULL;
+}
+
 /*
  * Update cfqq's position in the service tree.
  */
@@ -515,17 +630,19 @@ static void cfq_resort_rr_list(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 	/*
 	 * Resorting requires the cfqq to be on the RR list already.
 	 */
-	if (cfq_cfqq_on_rr(cfqq))
+	if (cfq_cfqq_on_rr(cfqq)) {
 		cfq_service_tree_add(cfqd, cfqq, 0);
+		cfq_prio_tree_add(cfqd, cfqq);
+	}
 }
 
 /*
  * add to busy list of queues for service, trying to be fair in ordering
  * the pending list according to last request service
  */
-static inline void
-cfq_add_cfqq_rr(struct cfq_data *cfqd, struct cfq_queue *cfqq)
+static void cfq_add_cfqq_rr(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 {
+	cfq_log_cfqq(cfqd, cfqq, "add_to_rr");
 	BUG_ON(cfq_cfqq_on_rr(cfqq));
 	cfq_mark_cfqq_on_rr(cfqq);
 	cfqd->busy_queues++;
@@ -537,14 +654,18 @@ cfq_add_cfqq_rr(struct cfq_data *cfqd, struct cfq_queue *cfqq)
  * Called when the cfqq no longer has requests pending, remove it from
  * the service tree.
  */
-static inline void
-cfq_del_cfqq_rr(struct cfq_data *cfqd, struct cfq_queue *cfqq)
+static void cfq_del_cfqq_rr(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 {
+	cfq_log_cfqq(cfqd, cfqq, "del_from_rr");
 	BUG_ON(!cfq_cfqq_on_rr(cfqq));
 	cfq_clear_cfqq_on_rr(cfqq);
 
 	if (!RB_EMPTY_NODE(&cfqq->rb_node))
 		cfq_rb_erase(&cfqq->rb_node, &cfqd->service_tree);
+	if (cfqq->p_root) {
+		rb_erase(&cfqq->p_node, cfqq->p_root);
+		cfqq->p_root = NULL;
+	}
 
 	BUG_ON(!cfqd->busy_queues);
 	cfqd->busy_queues--;
@@ -553,7 +674,7 @@ cfq_del_cfqq_rr(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 /*
  * rb tree support functions
  */
-static inline void cfq_del_rq_rb(struct request *rq)
+static void cfq_del_rq_rb(struct request *rq)
 {
 	struct cfq_queue *cfqq = RQ_CFQQ(rq);
 	struct cfq_data *cfqd = cfqq->cfqd;
@@ -572,7 +693,7 @@ static void cfq_add_rq_rb(struct request *rq)
 {
 	struct cfq_queue *cfqq = RQ_CFQQ(rq);
 	struct cfq_data *cfqd = cfqq->cfqd;
-	struct request *__alias;
+	struct request *__alias, *prev;
 
 	cfqq->queued[rq_is_sync(rq)]++;
 
@@ -589,12 +710,19 @@ static void cfq_add_rq_rb(struct request *rq)
 	/*
 	 * check if this request is a better next-serve candidate
 	 */
+	prev = cfqq->next_rq;
 	cfqq->next_rq = cfq_choose_req(cfqd, cfqq->next_rq, rq);
+
+	/*
+	 * adjust priority tree position, if ->next_rq changes
+	 */
+	if (prev != cfqq->next_rq)
+		cfq_prio_tree_add(cfqd, cfqq);
+
 	BUG_ON(!cfqq->next_rq);
 }
 
-static inline void
-cfq_reposition_rq_rb(struct cfq_queue *cfqq, struct request *rq)
+static void cfq_reposition_rq_rb(struct cfq_queue *cfqq, struct request *rq)
 {
 	elv_rb_del(&cfqq->sort_list, rq);
 	cfqq->queued[rq_is_sync(rq)]--;
@@ -608,7 +736,7 @@ cfq_find_rq_fmerge(struct cfq_data *cfqd, struct bio *bio)
 	struct cfq_io_context *cic;
 	struct cfq_queue *cfqq;
 
-	cic = cfq_cic_rb_lookup(cfqd, tsk->io_context);
+	cic = cfq_cic_lookup(cfqd, tsk->io_context);
 	if (!cic)
 		return NULL;
 
@@ -622,30 +750,26 @@ cfq_find_rq_fmerge(struct cfq_data *cfqd, struct bio *bio)
 	return NULL;
 }
 
-static void cfq_activate_request(request_queue_t *q, struct request *rq)
+static void cfq_activate_request(struct request_queue *q, struct request *rq)
 {
 	struct cfq_data *cfqd = q->elevator->elevator_data;
 
-	cfqd->rq_in_driver++;
-
-	/*
-	 * If the depth is larger 1, it really could be queueing. But lets
-	 * make the mark a little higher - idling could still be good for
-	 * low queueing, and a low queueing number could also just indicate
-	 * a SCSI mid layer like behaviour where limit+1 is often seen.
-	 */
-	if (!cfqd->hw_tag && cfqd->rq_in_driver > 4)
-		cfqd->hw_tag = 1;
+	cfqd->rq_in_driver[rq_is_sync(rq)]++;
+	cfq_log_cfqq(cfqd, RQ_CFQQ(rq), "activate rq, drv=%d",
+						rq_in_driver(cfqd));
 
-	cfqd->last_position = rq->hard_sector + rq->hard_nr_sectors;
+	cfqd->last_position = blk_rq_pos(rq) + blk_rq_sectors(rq);
 }
 
-static void cfq_deactivate_request(request_queue_t *q, struct request *rq)
+static void cfq_deactivate_request(struct request_queue *q, struct request *rq)
 {
 	struct cfq_data *cfqd = q->elevator->elevator_data;
+	const int sync = rq_is_sync(rq);
 
-	WARN_ON(!cfqd->rq_in_driver);
-	cfqd->rq_in_driver--;
+	WARN_ON(!cfqd->rq_in_driver[sync]);
+	cfqd->rq_in_driver[sync]--;
+	cfq_log_cfqq(cfqd, RQ_CFQQ(rq), "deactivate rq, drv=%d",
+						rq_in_driver(cfqd));
 }
 
 static void cfq_remove_request(struct request *rq)
@@ -658,13 +782,15 @@ static void cfq_remove_request(struct request *rq)
 	list_del_init(&rq->queuelist);
 	cfq_del_rq_rb(rq);
 
+	cfqq->cfqd->rq_queued--;
 	if (rq_is_meta(rq)) {
 		WARN_ON(!cfqq->meta_pending);
 		cfqq->meta_pending--;
 	}
 }
 
-static int cfq_merge(request_queue_t *q, struct request **req, struct bio *bio)
+static int cfq_merge(struct request_queue *q, struct request **req,
+		     struct bio *bio)
 {
 	struct cfq_data *cfqd = q->elevator->elevator_data;
 	struct request *__rq;
@@ -678,7 +804,7 @@ static int cfq_merge(request_queue_t *q, struct request **req, struct bio *bio)
 	return ELEVATOR_NO_MERGE;
 }
 
-static void cfq_merged_request(request_queue_t *q, struct request *req,
+static void cfq_merged_request(struct request_queue *q, struct request *req,
 			       int type)
 {
 	if (type == ELEVATOR_FRONT_MERGE) {
@@ -689,7 +815,7 @@ static void cfq_merged_request(request_queue_t *q, struct request *req,
 }
 
 static void
-cfq_merged_requests(request_queue_t *q, struct request *rq,
+cfq_merged_requests(struct request_queue *q, struct request *rq,
 		    struct request *next)
 {
 	/*
@@ -702,7 +828,7 @@ cfq_merged_requests(request_queue_t *q, struct request *rq,
 	cfq_remove_request(next);
 }
 
-static int cfq_allow_merge(request_queue_t *q, struct request *rq,
+static int cfq_allow_merge(struct request_queue *q, struct request *rq,
 			   struct bio *bio)
 {
 	struct cfq_data *cfqd = q->elevator->elevator_data;
@@ -719,7 +845,7 @@ static int cfq_allow_merge(request_queue_t *q, struct request *rq,
 	 * Lookup the cfqq that this bio will be queued with. Allow
 	 * merge only if rq is queued there.
 	 */
-	cic = cfq_cic_rb_lookup(cfqd, current->io_context);
+	cic = cfq_cic_lookup(cfqd, current->io_context);
 	if (!cic)
 		return 0;
 
@@ -730,20 +856,21 @@ static int cfq_allow_merge(request_queue_t *q, struct request *rq,
 	return 0;
 }
 
-static inline void
-__cfq_set_active_queue(struct cfq_data *cfqd, struct cfq_queue *cfqq)
+static void __cfq_set_active_queue(struct cfq_data *cfqd,
+				   struct cfq_queue *cfqq)
 {
 	if (cfqq) {
-		/*
-		 * stop potential idle class queues waiting service
-		 */
-		del_timer(&cfqd->idle_class_timer);
-
+		cfq_log_cfqq(cfqd, cfqq, "set_active");
 		cfqq->slice_end = 0;
+		cfqq->slice_dispatch = 0;
+
+		cfq_clear_cfqq_wait_request(cfqq);
+		cfq_clear_cfqq_must_dispatch(cfqq);
 		cfq_clear_cfqq_must_alloc_slice(cfqq);
 		cfq_clear_cfqq_fifo_expire(cfqq);
 		cfq_mark_cfqq_slice_new(cfqq);
-		cfq_clear_cfqq_queue_new(cfqq);
+
+		del_timer(&cfqd->idle_slice_timer);
 	}
 
 	cfqd->active_queue = cfqq;
@@ -756,17 +883,20 @@ static void
 __cfq_slice_expired(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 		    int timed_out)
 {
+	cfq_log_cfqq(cfqd, cfqq, "slice expired t=%d", timed_out);
+
 	if (cfq_cfqq_wait_request(cfqq))
 		del_timer(&cfqd->idle_slice_timer);
 
-	cfq_clear_cfqq_must_dispatch(cfqq);
 	cfq_clear_cfqq_wait_request(cfqq);
 
 	/*
 	 * store what was left of this slice, if the queue idled/timed out
 	 */
-	if (timed_out && !cfq_cfqq_slice_new(cfqq))
+	if (timed_out && !cfq_cfqq_slice_new(cfqq)) {
 		cfqq->slice_resid = cfqq->slice_end - jiffies;
+		cfq_log_cfqq(cfqd, cfqq, "resid=%ld", cfqq->slice_resid);
+	}
 
 	cfq_resort_rr_list(cfqd, cfqq);
 
@@ -793,42 +923,24 @@ static inline void cfq_slice_expired(struct cfq_data *cfqd, int timed_out)
  */
 static struct cfq_queue *cfq_get_next_queue(struct cfq_data *cfqd)
 {
-	struct cfq_queue *cfqq;
-	struct rb_node *n;
-
 	if (RB_EMPTY_ROOT(&cfqd->service_tree.rb))
 		return NULL;
 
-	n = cfq_rb_first(&cfqd->service_tree);
-	cfqq = rb_entry(n, struct cfq_queue, rb_node);
-
-	if (cfq_class_idle(cfqq)) {
-		unsigned long end;
-
-		/*
-		 * if we have idle queues and no rt or be queues had
-		 * pending requests, either allow immediate service if
-		 * the grace period has passed or arm the idle grace
-		 * timer
-		 */
-		end = cfqd->last_end_request + CFQ_IDLE_GRACE;
-		if (time_before(jiffies, end)) {
-			mod_timer(&cfqd->idle_class_timer, end);
-			cfqq = NULL;
-		}
-	}
-
-	return cfqq;
+	return cfq_rb_first(&cfqd->service_tree);
 }
 
 /*
  * Get and set a new active queue for service.
  */
-static struct cfq_queue *cfq_set_active_queue(struct cfq_data *cfqd)
+static struct cfq_queue *cfq_set_active_queue(struct cfq_data *cfqd,
+					      struct cfq_queue *cfqq)
 {
-	struct cfq_queue *cfqq;
+	if (!cfqq) {
+		cfqq = cfq_get_next_queue(cfqd);
+		if (cfqq)
+			cfq_clear_cfqq_coop(cfqq);
+	}
 
-	cfqq = cfq_get_next_queue(cfqd);
 	__cfq_set_active_queue(cfqd, cfqq);
 	return cfqq;
 }
@@ -836,34 +948,106 @@ static struct cfq_queue *cfq_set_active_queue(struct cfq_data *cfqd)
 static inline sector_t cfq_dist_from_last(struct cfq_data *cfqd,
 					  struct request *rq)
 {
-	if (rq->sector >= cfqd->last_position)
-		return rq->sector - cfqd->last_position;
+	if (blk_rq_pos(rq) >= cfqd->last_position)
+		return blk_rq_pos(rq) - cfqd->last_position;
 	else
-		return cfqd->last_position - rq->sector;
+		return cfqd->last_position - blk_rq_pos(rq);
 }
 
+#define CIC_SEEK_THR	8 * 1024
+#define CIC_SEEKY(cic)	((cic)->seek_mean > CIC_SEEK_THR)
+
 static inline int cfq_rq_close(struct cfq_data *cfqd, struct request *rq)
 {
 	struct cfq_io_context *cic = cfqd->active_cic;
+	sector_t sdist = cic->seek_mean;
 
 	if (!sample_valid(cic->seek_samples))
-		return 0;
+		sdist = CIC_SEEK_THR;
 
-	return cfq_dist_from_last(cfqd, rq) <= cic->seek_mean;
+	return cfq_dist_from_last(cfqd, rq) <= sdist;
 }
 
-static int cfq_close_cooperator(struct cfq_data *cfq_data,
-				struct cfq_queue *cfqq)
+static struct cfq_queue *cfqq_close(struct cfq_data *cfqd,
+				    struct cfq_queue *cur_cfqq)
 {
+	struct rb_root *root = &cfqd->prio_trees[cur_cfqq->org_ioprio];
+	struct rb_node *parent, *node;
+	struct cfq_queue *__cfqq;
+	sector_t sector = cfqd->last_position;
+
+	if (RB_EMPTY_ROOT(root))
+		return NULL;
+
+	/*
+	 * First, if we find a request starting at the end of the last
+	 * request, choose it.
+	 */
+	__cfqq = cfq_prio_tree_lookup(cfqd, root, sector, &parent, NULL);
+	if (__cfqq)
+		return __cfqq;
+
+	/*
+	 * If the exact sector wasn't found, the parent of the NULL leaf
+	 * will contain the closest sector.
+	 */
+	__cfqq = rb_entry(parent, struct cfq_queue, p_node);
+	if (cfq_rq_close(cfqd, __cfqq->next_rq))
+		return __cfqq;
+
+	if (blk_rq_pos(__cfqq->next_rq) < sector)
+		node = rb_next(&__cfqq->p_node);
+	else
+		node = rb_prev(&__cfqq->p_node);
+	if (!node)
+		return NULL;
+
+	__cfqq = rb_entry(node, struct cfq_queue, p_node);
+	if (cfq_rq_close(cfqd, __cfqq->next_rq))
+		return __cfqq;
+
+	return NULL;
+}
+
+/*
+ * cfqd - obvious
+ * cur_cfqq - passed in so that we don't decide that the current queue is
+ * 	      closely cooperating with itself.
+ *
+ * So, basically we're assuming that that cur_cfqq has dispatched at least
+ * one request, and that cfqd->last_position reflects a position on the disk
+ * associated with the I/O issued by cur_cfqq.  I'm not sure this is a valid
+ * assumption.
+ */
+static struct cfq_queue *cfq_close_cooperator(struct cfq_data *cfqd,
+					      struct cfq_queue *cur_cfqq,
+					      int probe)
+{
+	struct cfq_queue *cfqq;
+
+	/*
+	 * A valid cfq_io_context is necessary to compare requests against
+	 * the seek_mean of the current cfqq.
+	 */
+	if (!cfqd->active_cic)
+		return NULL;
+
 	/*
 	 * We should notice if some of the queues are cooperating, eg
 	 * working closely on the same area of the disk. In that case,
 	 * we can group them together and don't waste time idling.
 	 */
-	return 0;
-}
+	cfqq = cfqq_close(cfqd, cur_cfqq);
+	if (!cfqq)
+		return NULL;
+
+	if (cfq_cfqq_coop(cfqq))
+		return NULL;
 
-#define CIC_SEEKY(cic) ((cic)->seek_mean > (8 * 1024))
+	if (!probe)
+		cfq_mark_cfqq_coop(cfqq);
+	return cfqq;
+}
 
 static void cfq_arm_slice_timer(struct cfq_data *cfqd)
 {
@@ -871,6 +1055,14 @@ static void cfq_arm_slice_timer(struct cfq_data *cfqd)
 	struct cfq_io_context *cic;
 	unsigned long sl;
 
+	/*
+	 * SSD device without seek penalty, disable idling. But only do so
+	 * for devices that support queuing, otherwise we still have a problem
+	 * with sync vs async workloads.
+	 */
+	if (blk_queue_nonrot(cfqd->queue) && cfqd->hw_tag)
+		return;
+
 	WARN_ON(!RB_EMPTY_ROOT(&cfqq->sort_list));
 	WARN_ON(cfq_cfqq_slice_new(cfqq));
 
@@ -881,20 +1073,18 @@ static void cfq_arm_slice_timer(struct cfq_data *cfqd)
 		return;
 
 	/*
-	 * task has exited, don't wait
+	 * still requests with the driver, don't idle
 	 */
-	cic = cfqd->active_cic;
-	if (!cic || !cic->ioc->task)
+	if (rq_in_driver(cfqd))
 		return;
 
 	/*
-	 * See if this prio level has a good candidate
+	 * task has exited, don't wait
 	 */
-	if (cfq_close_cooperator(cfqd, cfqq) &&
-	    (sample_valid(cic->ttime_samples) && cic->ttime_mean > 2))
+	cic = cfqd->active_cic;
+	if (!cic || !atomic_read(&cic->ioc->nr_tasks))
 		return;
 
-	cfq_mark_cfqq_must_dispatch(cfqq);
 	cfq_mark_cfqq_wait_request(cfqq);
 
 	/*
@@ -907,16 +1097,20 @@ static void cfq_arm_slice_timer(struct cfq_data *cfqd)
 		sl = min(sl, msecs_to_jiffies(CFQ_MIN_TT));
 
 	mod_timer(&cfqd->idle_slice_timer, jiffies + sl);
+	cfq_log_cfqq(cfqd, cfqq, "arm_idle: %lu", sl);
 }
 
 /*
  * Move request from internal lists to the request queue dispatch list.
  */
-static void cfq_dispatch_insert(request_queue_t *q, struct request *rq)
+static void cfq_dispatch_insert(struct request_queue *q, struct request *rq)
 {
 	struct cfq_data *cfqd = q->elevator->elevator_data;
 	struct cfq_queue *cfqq = RQ_CFQQ(rq);
 
+	cfq_log_cfqq(cfqd, cfqq, "dispatch_insert");
+
+	cfqq->next_rq = cfq_find_next_rq(cfqd, cfqq, rq);
 	cfq_remove_request(rq);
 	cfqq->dispatched++;
 	elv_dispatch_sort(q, rq);
@@ -928,7 +1122,7 @@ static void cfq_dispatch_insert(request_queue_t *q, struct request *rq)
 /*
  * return expired entry, or NULL to just start from scratch in rbtree
  */
-static inline struct request *cfq_check_fifo(struct cfq_queue *cfqq)
+static struct request *cfq_check_fifo(struct cfq_queue *cfqq)
 {
 	struct cfq_data *cfqd = cfqq->cfqd;
 	struct request *rq;
@@ -946,8 +1140,9 @@ static inline struct request *cfq_check_fifo(struct cfq_queue *cfqq)
 	rq = rq_entry_fifo(cfqq->fifo.next);
 
 	if (time_before(jiffies, rq->start_time + cfqd->cfq_fifo_expire[fifo]))
-		return NULL;
+		rq = NULL;
 
+	cfq_log_cfqq(cfqd, cfqq, "fifo=%p", rq);
 	return rq;
 }
 
@@ -967,7 +1162,7 @@ cfq_prio_to_maxrq(struct cfq_data *cfqd, struct cfq_queue *cfqq)
  */
 static struct cfq_queue *cfq_select_queue(struct cfq_data *cfqd)
 {
-	struct cfq_queue *cfqq;
+	struct cfq_queue *cfqq, *new_cfqq = NULL;
 
 	cfqq = cfqd->active_queue;
 	if (!cfqq)
@@ -976,7 +1171,7 @@ static struct cfq_queue *cfq_select_queue(struct cfq_data *cfqd)
 	/*
 	 * The active queue has run out of time, expire it and select new.
 	 */
-	if (cfq_slice_used(cfqq))
+	if (cfq_slice_used(cfqq) && !cfq_cfqq_must_dispatch(cfqq))
 		goto expire;
 
 	/*
@@ -987,6 +1182,16 @@ static struct cfq_queue *cfq_select_queue(struct cfq_data *cfqd)
 		goto keep_queue;
 
 	/*
+	 * If another queue has a request waiting within our mean seek
+	 * distance, let it run.  The expire code will check for close
+	 * cooperators and put the close queue at the front of the service
+	 * tree.
+	 */
+	new_cfqq = cfq_close_cooperator(cfqd, cfqq, 0);
+	if (new_cfqq)
+		goto expire;
+
+	/*
 	 * No requests pending. If the active queue still has requests in
 	 * flight or is idling for a new request, allow either of these
 	 * conditions to happen (or time out) before selecting a new queue.
@@ -1000,64 +1205,12 @@ static struct cfq_queue *cfq_select_queue(struct cfq_data *cfqd)
 expire:
 	cfq_slice_expired(cfqd, 0);
 new_queue:
-	cfqq = cfq_set_active_queue(cfqd);
+	cfqq = cfq_set_active_queue(cfqd, new_cfqq);
 keep_queue:
 	return cfqq;
 }
 
-/*
- * Dispatch some requests from cfqq, moving them to the request queue
- * dispatch list.
- */
-static int
-__cfq_dispatch_requests(struct cfq_data *cfqd, struct cfq_queue *cfqq,
-			int max_dispatch)
-{
-	int dispatched = 0;
-
-	BUG_ON(RB_EMPTY_ROOT(&cfqq->sort_list));
-
-	do {
-		struct request *rq;
-
-		/*
-		 * follow expired path, else get first next available
-		 */
-		if ((rq = cfq_check_fifo(cfqq)) == NULL)
-			rq = cfqq->next_rq;
-
-		/*
-		 * finally, insert request into driver dispatch list
-		 */
-		cfq_dispatch_insert(cfqd->queue, rq);
-
-		dispatched++;
-
-		if (!cfqd->active_cic) {
-			atomic_inc(&RQ_CIC(rq)->ioc->refcount);
-			cfqd->active_cic = RQ_CIC(rq);
-		}
-
-		if (RB_EMPTY_ROOT(&cfqq->sort_list))
-			break;
-
-	} while (dispatched < max_dispatch);
-
-	/*
-	 * expire an async queue immediately if it has used up its slice. idle
-	 * queue always expire after 1 dispatch round.
-	 */
-	if (cfqd->busy_queues > 1 && ((!cfq_cfqq_sync(cfqq) &&
-	    dispatched >= cfq_prio_to_maxrq(cfqd, cfqq)) ||
-	    cfq_class_idle(cfqq))) {
-		cfqq->slice_end = jiffies + 1;
-		cfq_slice_expired(cfqd, 0);
-	}
-
-	return dispatched;
-}
-
-static inline int __cfq_forced_dispatch_cfqq(struct cfq_queue *cfqq)
+static int __cfq_forced_dispatch_cfqq(struct cfq_queue *cfqq)
 {
 	int dispatched = 0;
 
@@ -1076,27 +1229,59 @@ static inline int __cfq_forced_dispatch_cfqq(struct cfq_queue *cfqq)
  */
 static int cfq_forced_dispatch(struct cfq_data *cfqd)
 {
+	struct cfq_queue *cfqq;
 	int dispatched = 0;
-	struct rb_node *n;
-
-	while ((n = cfq_rb_first(&cfqd->service_tree)) != NULL) {
-		struct cfq_queue *cfqq = rb_entry(n, struct cfq_queue, rb_node);
 
+	while ((cfqq = cfq_rb_first(&cfqd->service_tree)) != NULL)
 		dispatched += __cfq_forced_dispatch_cfqq(cfqq);
-	}
 
 	cfq_slice_expired(cfqd, 0);
 
 	BUG_ON(cfqd->busy_queues);
 
+	cfq_log(cfqd, "forced_dispatch=%d", dispatched);
 	return dispatched;
 }
 
-static int cfq_dispatch_requests(request_queue_t *q, int force)
+/*
+ * Dispatch a request from cfqq, moving them to the request queue
+ * dispatch list.
+ */
+static void cfq_dispatch_request(struct cfq_data *cfqd, struct cfq_queue *cfqq)
+{
+	struct request *rq;
+
+	BUG_ON(RB_EMPTY_ROOT(&cfqq->sort_list));
+
+	/*
+	 * follow expired path, else get first next available
+	 */
+	rq = cfq_check_fifo(cfqq);
+	if (!rq)
+		rq = cfqq->next_rq;
+
+	/*
+	 * insert request into driver dispatch list
+	 */
+	cfq_dispatch_insert(cfqd->queue, rq);
+
+	if (!cfqd->active_cic) {
+		struct cfq_io_context *cic = RQ_CIC(rq);
+
+		atomic_long_inc(&cic->ioc->refcount);
+		cfqd->active_cic = cic;
+	}
+}
+
+/*
+ * Find the cfqq that we need to service and move a request from that to the
+ * dispatch list
+ */
+static int cfq_dispatch_requests(struct request_queue *q, int force)
 {
 	struct cfq_data *cfqd = q->elevator->elevator_data;
 	struct cfq_queue *cfqq;
-	int dispatched;
+	unsigned int max_dispatch;
 
 	if (!cfqd->busy_queues)
 		return 0;
@@ -1104,32 +1289,69 @@ static int cfq_dispatch_requests(request_queue_t *q, int force)
 	if (unlikely(force))
 		return cfq_forced_dispatch(cfqd);
 
-	dispatched = 0;
-	while ((cfqq = cfq_select_queue(cfqd)) != NULL) {
-		int max_dispatch;
+	cfqq = cfq_select_queue(cfqd);
+	if (!cfqq)
+		return 0;
+
+	/*
+	 * Drain async requests before we start sync IO
+	 */
+	if (cfq_cfqq_idle_window(cfqq) && cfqd->rq_in_driver[BLK_RW_ASYNC])
+		return 0;
+
+	/*
+	 * If this is an async queue and we have sync IO in flight, let it wait
+	 */
+	if (cfqd->sync_flight && !cfq_cfqq_sync(cfqq))
+		return 0;
+
+	max_dispatch = cfqd->cfq_quantum;
+	if (cfq_class_idle(cfqq))
+		max_dispatch = 1;
 
-		max_dispatch = cfqd->cfq_quantum;
+	/*
+	 * Does this cfqq already have too much IO in flight?
+	 */
+	if (cfqq->dispatched >= max_dispatch) {
+		/*
+		 * idle queue must always only have a single IO in flight
+		 */
 		if (cfq_class_idle(cfqq))
-			max_dispatch = 1;
+			return 0;
 
-		if (cfqq->dispatched >= max_dispatch) {
-			if (cfqd->busy_queues > 1)
-				break;
-			if (cfqq->dispatched >= 4 * max_dispatch)
-				break;
-		}
+		/*
+		 * We have other queues, don't allow more IO from this one
+		 */
+		if (cfqd->busy_queues > 1)
+			return 0;
 
-		if (cfqd->sync_flight && !cfq_cfqq_sync(cfqq))
-			break;
+		/*
+		 * we are the only queue, allow up to 4 times of 'quantum'
+		 */
+		if (cfqq->dispatched >= 4 * max_dispatch)
+			return 0;
+	}
 
-		cfq_clear_cfqq_must_dispatch(cfqq);
-		cfq_clear_cfqq_wait_request(cfqq);
-		del_timer(&cfqd->idle_slice_timer);
+	/*
+	 * Dispatch a request from this cfqq
+	 */
+	cfq_dispatch_request(cfqd, cfqq);
+	cfqq->slice_dispatch++;
+	cfq_clear_cfqq_must_dispatch(cfqq);
 
-		dispatched += __cfq_dispatch_requests(cfqd, cfqq, max_dispatch);
+	/*
+	 * expire an async queue immediately if it has used up its slice. idle
+	 * queue always expire after 1 dispatch round.
+	 */
+	if (cfqd->busy_queues > 1 && ((!cfq_cfqq_sync(cfqq) &&
+	    cfqq->slice_dispatch >= cfq_prio_to_maxrq(cfqd, cfqq)) ||
+	    cfq_class_idle(cfqq))) {
+		cfqq->slice_end = jiffies + 1;
+		cfq_slice_expired(cfqd, 0);
 	}
 
-	return dispatched;
+	cfq_log_cfqq(cfqd, cfqq, "dispatched a request");
+	return 1;
 }
 
 /*
@@ -1147,6 +1369,7 @@ static void cfq_put_queue(struct cfq_queue *cfqq)
 	if (!atomic_dec_and_test(&cfqq->ref))
 		return;
 
+	cfq_log_cfqq(cfqd, cfqq, "put_queue");
 	BUG_ON(rb_first(&cfqq->sort_list));
 	BUG_ON(cfqq->allocated[READ] + cfqq->allocated[WRITE]);
 	BUG_ON(cfq_cfqq_on_rr(cfqq));
@@ -1159,25 +1382,89 @@ static void cfq_put_queue(struct cfq_queue *cfqq)
 	kmem_cache_free(cfq_pool, cfqq);
 }
 
-static void cfq_free_io_context(struct io_context *ioc)
+/*
+ * Must always be called with the rcu_read_lock() held
+ */
+static void
+__call_for_each_cic(struct io_context *ioc,
+		    void (*func)(struct io_context *, struct cfq_io_context *))
+{
+	struct cfq_io_context *cic;
+	struct hlist_node *n;
+
+	hlist_for_each_entry_rcu(cic, n, &ioc->cic_list, cic_list)
+		func(ioc, cic);
+}
+
+/*
+ * Call func for each cic attached to this ioc.
+ */
+static void
+call_for_each_cic(struct io_context *ioc,
+		  void (*func)(struct io_context *, struct cfq_io_context *))
+{
+	rcu_read_lock();
+	__call_for_each_cic(ioc, func);
+	rcu_read_unlock();
+}
+
+static void cfq_cic_free_rcu(struct rcu_head *head)
 {
-	struct cfq_io_context *__cic;
-	struct rb_node *n;
-	int freed = 0;
+	struct cfq_io_context *cic;
+
+	cic = container_of(head, struct cfq_io_context, rcu_head);
 
-	ioc->ioc_data = NULL;
+	kmem_cache_free(cfq_ioc_pool, cic);
+	elv_ioc_count_dec(ioc_count);
 
-	while ((n = rb_first(&ioc->cic_root)) != NULL) {
-		__cic = rb_entry(n, struct cfq_io_context, rb_node);
-		rb_erase(&__cic->rb_node, &ioc->cic_root);
-		kmem_cache_free(cfq_ioc_pool, __cic);
-		freed++;
+	if (ioc_gone) {
+		/*
+		 * CFQ scheduler is exiting, grab exit lock and check
+		 * the pending io context count. If it hits zero,
+		 * complete ioc_gone and set it back to NULL
+		 */
+		spin_lock(&ioc_gone_lock);
+		if (ioc_gone && !elv_ioc_count_read(ioc_count)) {
+			complete(ioc_gone);
+			ioc_gone = NULL;
+		}
+		spin_unlock(&ioc_gone_lock);
 	}
+}
+
+static void cfq_cic_free(struct cfq_io_context *cic)
+{
+	call_rcu(&cic->rcu_head, cfq_cic_free_rcu);
+}
+
+static void cic_free_func(struct io_context *ioc, struct cfq_io_context *cic)
+{
+	unsigned long flags;
 
-	elv_ioc_count_mod(ioc_count, -freed);
+	BUG_ON(!cic->dead_key);
 
-	if (ioc_gone && !elv_ioc_count_read(ioc_count))
-		complete(ioc_gone);
+	spin_lock_irqsave(&ioc->lock, flags);
+	radix_tree_delete(&ioc->radix_root, cic->dead_key);
+	hlist_del_rcu(&cic->cic_list);
+	spin_unlock_irqrestore(&ioc->lock, flags);
+
+	cfq_cic_free(cic);
+}
+
+/*
+ * Must be called with rcu_read_lock() held or preemption otherwise disabled.
+ * Only two callers of this - ->dtor() which is called with the rcu_read_lock(),
+ * and ->trim() which is called with the task lock held
+ */
+static void cfq_free_io_context(struct io_context *ioc)
+{
+	/*
+	 * ioc->refcount is zero here, or we are called from elv_unregister(),
+	 * so no more cic's are allowed to be linked into this ioc.  So it
+	 * should be ok to iterate over the known list, we will see all cic's
+	 * since no new ones are added.
+	 */
+	__call_for_each_cic(ioc, cic_free_func);
 }
 
 static void cfq_exit_cfqq(struct cfq_data *cfqd, struct cfq_queue *cfqq)
@@ -1193,31 +1480,51 @@ static void cfq_exit_cfqq(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 static void __cfq_exit_single_io_context(struct cfq_data *cfqd,
 					 struct cfq_io_context *cic)
 {
+	struct io_context *ioc = cic->ioc;
+
 	list_del_init(&cic->queue_list);
+
+	/*
+	 * Make sure key == NULL is seen for dead queues
+	 */
 	smp_wmb();
+	cic->dead_key = (unsigned long) cic->key;
 	cic->key = NULL;
 
-	if (cic->cfqq[ASYNC]) {
-		cfq_exit_cfqq(cfqd, cic->cfqq[ASYNC]);
-		cic->cfqq[ASYNC] = NULL;
+	if (ioc->ioc_data == cic)
+		rcu_assign_pointer(ioc->ioc_data, NULL);
+
+	if (cic->cfqq[BLK_RW_ASYNC]) {
+		cfq_exit_cfqq(cfqd, cic->cfqq[BLK_RW_ASYNC]);
+		cic->cfqq[BLK_RW_ASYNC] = NULL;
 	}
 
-	if (cic->cfqq[SYNC]) {
-		cfq_exit_cfqq(cfqd, cic->cfqq[SYNC]);
-		cic->cfqq[SYNC] = NULL;
+	if (cic->cfqq[BLK_RW_SYNC]) {
+		cfq_exit_cfqq(cfqd, cic->cfqq[BLK_RW_SYNC]);
+		cic->cfqq[BLK_RW_SYNC] = NULL;
 	}
 }
 
-static void cfq_exit_single_io_context(struct cfq_io_context *cic)
+static void cfq_exit_single_io_context(struct io_context *ioc,
+				       struct cfq_io_context *cic)
 {
 	struct cfq_data *cfqd = cic->key;
 
 	if (cfqd) {
-		request_queue_t *q = cfqd->queue;
+		struct request_queue *q = cfqd->queue;
+		unsigned long flags;
 
-		spin_lock_irq(q->queue_lock);
-		__cfq_exit_single_io_context(cfqd, cic);
-		spin_unlock_irq(q->queue_lock);
+		spin_lock_irqsave(q->queue_lock, flags);
+
+		/*
+		 * Ensure we get a fresh copy of the ->key to prevent
+		 * race between exiting task and queue
+		 */
+		smp_read_barrier_depends();
+		if (cic->key)
+			__cfq_exit_single_io_context(cfqd, cic);
+
+		spin_unlock_irqrestore(q->queue_lock, flags);
 	}
 }
 
@@ -1227,21 +1534,7 @@ static void cfq_exit_single_io_context(struct cfq_io_context *cic)
  */
 static void cfq_exit_io_context(struct io_context *ioc)
 {
-	struct cfq_io_context *__cic;
-	struct rb_node *n;
-
-	ioc->ioc_data = NULL;
-
-	/*
-	 * put the reference this task is holding to the various queues
-	 */
-	n = rb_first(&ioc->cic_root);
-	while (n != NULL) {
-		__cic = rb_entry(n, struct cfq_io_context, rb_node);
-
-		cfq_exit_single_io_context(__cic);
-		n = rb_next(n);
-	}
+	call_for_each_cic(ioc, cfq_exit_single_io_context);
 }
 
 static struct cfq_io_context *
@@ -1249,11 +1542,12 @@ cfq_alloc_io_context(struct cfq_data *cfqd, gfp_t gfp_mask)
 {
 	struct cfq_io_context *cic;
 
-	cic = kmem_cache_alloc_node(cfq_ioc_pool, gfp_mask, cfqd->queue->node);
+	cic = kmem_cache_alloc_node(cfq_ioc_pool, gfp_mask | __GFP_ZERO,
+							cfqd->queue->node);
 	if (cic) {
-		memset(cic, 0, sizeof(*cic));
 		cic->last_end_request = jiffies;
 		INIT_LIST_HEAD(&cic->queue_list);
+		INIT_HLIST_NODE(&cic->cic_list);
 		cic->dtor = cfq_free_io_context;
 		cic->exit = cfq_exit_io_context;
 		elv_ioc_count_inc(ioc_count);
@@ -1262,7 +1556,7 @@ cfq_alloc_io_context(struct cfq_data *cfqd, gfp_t gfp_mask)
 	return cic;
 }
 
-static void cfq_init_prio_data(struct cfq_queue *cfqq)
+static void cfq_init_prio_data(struct cfq_queue *cfqq, struct io_context *ioc)
 {
 	struct task_struct *tsk = current;
 	int ioprio_class;
@@ -1270,30 +1564,30 @@ static void cfq_init_prio_data(struct cfq_queue *cfqq)
 	if (!cfq_cfqq_prio_changed(cfqq))
 		return;
 
-	ioprio_class = IOPRIO_PRIO_CLASS(tsk->ioprio);
+	ioprio_class = IOPRIO_PRIO_CLASS(ioc->ioprio);
 	switch (ioprio_class) {
-		default:
-			printk(KERN_ERR "cfq: bad prio %x\n", ioprio_class);
-		case IOPRIO_CLASS_NONE:
-			/*
-			 * no prio set, place us in the middle of the BE classes
-			 */
-			cfqq->ioprio = task_nice_ioprio(tsk);
-			cfqq->ioprio_class = IOPRIO_CLASS_BE;
-			break;
-		case IOPRIO_CLASS_RT:
-			cfqq->ioprio = task_ioprio(tsk);
-			cfqq->ioprio_class = IOPRIO_CLASS_RT;
-			break;
-		case IOPRIO_CLASS_BE:
-			cfqq->ioprio = task_ioprio(tsk);
-			cfqq->ioprio_class = IOPRIO_CLASS_BE;
-			break;
-		case IOPRIO_CLASS_IDLE:
-			cfqq->ioprio_class = IOPRIO_CLASS_IDLE;
-			cfqq->ioprio = 7;
-			cfq_clear_cfqq_idle_window(cfqq);
-			break;
+	default:
+		printk(KERN_ERR "cfq: bad prio %x\n", ioprio_class);
+	case IOPRIO_CLASS_NONE:
+		/*
+		 * no prio set, inherit CPU scheduling settings
+		 */
+		cfqq->ioprio = task_nice_ioprio(tsk);
+		cfqq->ioprio_class = task_nice_ioclass(tsk);
+		break;
+	case IOPRIO_CLASS_RT:
+		cfqq->ioprio = task_ioprio(ioc);
+		cfqq->ioprio_class = IOPRIO_CLASS_RT;
+		break;
+	case IOPRIO_CLASS_BE:
+		cfqq->ioprio = task_ioprio(ioc);
+		cfqq->ioprio_class = IOPRIO_CLASS_BE;
+		break;
+	case IOPRIO_CLASS_IDLE:
+		cfqq->ioprio_class = IOPRIO_CLASS_IDLE;
+		cfqq->ioprio = 7;
+		cfq_clear_cfqq_idle_window(cfqq);
+		break;
 	}
 
 	/*
@@ -1305,7 +1599,7 @@ static void cfq_init_prio_data(struct cfq_queue *cfqq)
 	cfq_clear_cfqq_prio_changed(cfqq);
 }
 
-static inline void changed_ioprio(struct cfq_io_context *cic)
+static void changed_ioprio(struct io_context *ioc, struct cfq_io_context *cic)
 {
 	struct cfq_data *cfqd = cic->key;
 	struct cfq_queue *cfqq;
@@ -1316,18 +1610,18 @@ static inline void changed_ioprio(struct cfq_io_context *cic)
 
 	spin_lock_irqsave(cfqd->queue->queue_lock, flags);
 
-	cfqq = cic->cfqq[ASYNC];
+	cfqq = cic->cfqq[BLK_RW_ASYNC];
 	if (cfqq) {
 		struct cfq_queue *new_cfqq;
-		new_cfqq = cfq_get_queue(cfqd, ASYNC, cic->ioc->task,
-					 GFP_ATOMIC);
+		new_cfqq = cfq_get_queue(cfqd, BLK_RW_ASYNC, cic->ioc,
+						GFP_ATOMIC);
 		if (new_cfqq) {
-			cic->cfqq[ASYNC] = new_cfqq;
+			cic->cfqq[BLK_RW_ASYNC] = new_cfqq;
 			cfq_put_queue(cfqq);
 		}
 	}
 
-	cfqq = cic->cfqq[SYNC];
+	cfqq = cic->cfqq[BLK_RW_SYNC];
 	if (cfqq)
 		cfq_mark_cfqq_prio_changed(cfqq);
 
@@ -1336,78 +1630,120 @@ static inline void changed_ioprio(struct cfq_io_context *cic)
 
 static void cfq_ioc_set_ioprio(struct io_context *ioc)
 {
-	struct cfq_io_context *cic;
-	struct rb_node *n;
-
+	call_for_each_cic(ioc, changed_ioprio);
 	ioc->ioprio_changed = 0;
+}
 
-	n = rb_first(&ioc->cic_root);
-	while (n != NULL) {
-		cic = rb_entry(n, struct cfq_io_context, rb_node);
+static void cfq_init_cfqq(struct cfq_data *cfqd, struct cfq_queue *cfqq,
+			  pid_t pid, int is_sync)
+{
+	RB_CLEAR_NODE(&cfqq->rb_node);
+	RB_CLEAR_NODE(&cfqq->p_node);
+	INIT_LIST_HEAD(&cfqq->fifo);
+
+	atomic_set(&cfqq->ref, 0);
+	cfqq->cfqd = cfqd;
 
-		changed_ioprio(cic);
-		n = rb_next(n);
+	cfq_mark_cfqq_prio_changed(cfqq);
+
+	if (is_sync) {
+		if (!cfq_class_idle(cfqq))
+			cfq_mark_cfqq_idle_window(cfqq);
+		cfq_mark_cfqq_sync(cfqq);
 	}
+	cfqq->pid = pid;
 }
 
 static struct cfq_queue *
-cfq_get_queue(struct cfq_data *cfqd, int is_sync, struct task_struct *tsk,
-	      gfp_t gfp_mask)
+cfq_find_alloc_queue(struct cfq_data *cfqd, int is_sync,
+		     struct io_context *ioc, gfp_t gfp_mask)
 {
 	struct cfq_queue *cfqq, *new_cfqq = NULL;
 	struct cfq_io_context *cic;
 
 retry:
-	cic = cfq_cic_rb_lookup(cfqd, tsk->io_context);
+	cic = cfq_cic_lookup(cfqd, ioc);
 	/* cic always exists here */
 	cfqq = cic_to_cfqq(cic, is_sync);
 
-	if (!cfqq) {
+	/*
+	 * Always try a new alloc if we fell back to the OOM cfqq
+	 * originally, since it should just be a temporary situation.
+	 */
+	if (!cfqq || cfqq == &cfqd->oom_cfqq) {
+		cfqq = NULL;
 		if (new_cfqq) {
 			cfqq = new_cfqq;
 			new_cfqq = NULL;
 		} else if (gfp_mask & __GFP_WAIT) {
-			/*
-			 * Inform the allocator of the fact that we will
-			 * just repeat this allocation if it fails, to allow
-			 * the allocator to do whatever it needs to attempt to
-			 * free memory.
-			 */
 			spin_unlock_irq(cfqd->queue->queue_lock);
-			new_cfqq = kmem_cache_alloc_node(cfq_pool, gfp_mask|__GFP_NOFAIL, cfqd->queue->node);
+			new_cfqq = kmem_cache_alloc_node(cfq_pool,
+					gfp_mask | __GFP_ZERO,
+					cfqd->queue->node);
 			spin_lock_irq(cfqd->queue->queue_lock);
-			goto retry;
+			if (new_cfqq)
+				goto retry;
 		} else {
-			cfqq = kmem_cache_alloc_node(cfq_pool, gfp_mask, cfqd->queue->node);
-			if (!cfqq)
-				goto out;
+			cfqq = kmem_cache_alloc_node(cfq_pool,
+					gfp_mask | __GFP_ZERO,
+					cfqd->queue->node);
 		}
 
-		memset(cfqq, 0, sizeof(*cfqq));
+		if (cfqq) {
+			cfq_init_cfqq(cfqd, cfqq, current->pid, is_sync);
+			cfq_init_prio_data(cfqq, ioc);
+			cfq_log_cfqq(cfqd, cfqq, "alloced");
+		} else
+			cfqq = &cfqd->oom_cfqq;
+	}
 
-		RB_CLEAR_NODE(&cfqq->rb_node);
-		INIT_LIST_HEAD(&cfqq->fifo);
+	if (new_cfqq)
+		kmem_cache_free(cfq_pool, new_cfqq);
 
-		atomic_set(&cfqq->ref, 0);
-		cfqq->cfqd = cfqd;
+	return cfqq;
+}
 
-		if (is_sync) {
-			cfq_mark_cfqq_idle_window(cfqq);
-			cfq_mark_cfqq_sync(cfqq);
-		}
+static struct cfq_queue **
+cfq_async_queue_prio(struct cfq_data *cfqd, int ioprio_class, int ioprio)
+{
+	switch (ioprio_class) {
+	case IOPRIO_CLASS_RT:
+		return &cfqd->async_cfqq[0][ioprio];
+	case IOPRIO_CLASS_BE:
+		return &cfqd->async_cfqq[1][ioprio];
+	case IOPRIO_CLASS_IDLE:
+		return &cfqd->async_idle_cfqq;
+	default:
+		BUG();
+	}
+}
 
-		cfq_mark_cfqq_prio_changed(cfqq);
-		cfq_mark_cfqq_queue_new(cfqq);
+static struct cfq_queue *
+cfq_get_queue(struct cfq_data *cfqd, int is_sync, struct io_context *ioc,
+	      gfp_t gfp_mask)
+{
+	const int ioprio = task_ioprio(ioc);
+	const int ioprio_class = task_ioprio_class(ioc);
+	struct cfq_queue **async_cfqq = NULL;
+	struct cfq_queue *cfqq = NULL;
 
-		cfq_init_prio_data(cfqq);
+	if (!is_sync) {
+		async_cfqq = cfq_async_queue_prio(cfqd, ioprio_class, ioprio);
+		cfqq = *async_cfqq;
 	}
 
-	if (new_cfqq)
-		kmem_cache_free(cfq_pool, new_cfqq);
+	if (!cfqq)
+		cfqq = cfq_find_alloc_queue(cfqd, is_sync, ioc, gfp_mask);
+
+	/*
+	 * pin the queue now that it's allocated, scheduler exit will prune it
+	 */
+	if (!is_sync && !(*async_cfqq)) {
+		atomic_inc(&cfqq->ref);
+		*async_cfqq = cfqq;
+	}
 
 	atomic_inc(&cfqq->ref);
-out:
-	WARN_ON((gfp_mask & __GFP_WAIT) && !cfqq);
 	return cfqq;
 }
 
@@ -1415,99 +1751,103 @@ out:
  * We drop cfq io contexts lazily, so we may find a dead one.
  */
 static void
-cfq_drop_dead_cic(struct io_context *ioc, struct cfq_io_context *cic)
+cfq_drop_dead_cic(struct cfq_data *cfqd, struct io_context *ioc,
+		  struct cfq_io_context *cic)
 {
+	unsigned long flags;
+
 	WARN_ON(!list_empty(&cic->queue_list));
 
-	if (ioc->ioc_data == cic)
-		ioc->ioc_data = NULL;
+	spin_lock_irqsave(&ioc->lock, flags);
 
-	rb_erase(&cic->rb_node, &ioc->cic_root);
-	kmem_cache_free(cfq_ioc_pool, cic);
-	elv_ioc_count_dec(ioc_count);
+	BUG_ON(ioc->ioc_data == cic);
+
+	radix_tree_delete(&ioc->radix_root, (unsigned long) cfqd);
+	hlist_del_rcu(&cic->cic_list);
+	spin_unlock_irqrestore(&ioc->lock, flags);
+
+	cfq_cic_free(cic);
 }
 
 static struct cfq_io_context *
-cfq_cic_rb_lookup(struct cfq_data *cfqd, struct io_context *ioc)
+cfq_cic_lookup(struct cfq_data *cfqd, struct io_context *ioc)
 {
-	struct rb_node *n;
 	struct cfq_io_context *cic;
-	void *k, *key = cfqd;
+	unsigned long flags;
+	void *k;
 
 	if (unlikely(!ioc))
 		return NULL;
 
+	rcu_read_lock();
+
 	/*
 	 * we maintain a last-hit cache, to avoid browsing over the tree
 	 */
-	cic = ioc->ioc_data;
-	if (cic && cic->key == cfqd)
+	cic = rcu_dereference(ioc->ioc_data);
+	if (cic && cic->key == cfqd) {
+		rcu_read_unlock();
 		return cic;
+	}
 
-restart:
-	n = ioc->cic_root.rb_node;
-	while (n) {
-		cic = rb_entry(n, struct cfq_io_context, rb_node);
+	do {
+		cic = radix_tree_lookup(&ioc->radix_root, (unsigned long) cfqd);
+		rcu_read_unlock();
+		if (!cic)
+			break;
 		/* ->key must be copied to avoid race with cfq_exit_queue() */
 		k = cic->key;
 		if (unlikely(!k)) {
-			cfq_drop_dead_cic(ioc, cic);
-			goto restart;
+			cfq_drop_dead_cic(cfqd, ioc, cic);
+			rcu_read_lock();
+			continue;
 		}
 
-		if (key < k)
-			n = n->rb_left;
-		else if (key > k)
-			n = n->rb_right;
-		else {
-			ioc->ioc_data = cic;
-			return cic;
-		}
-	}
+		spin_lock_irqsave(&ioc->lock, flags);
+		rcu_assign_pointer(ioc->ioc_data, cic);
+		spin_unlock_irqrestore(&ioc->lock, flags);
+		break;
+	} while (1);
 
-	return NULL;
+	return cic;
 }
 
-static inline void
-cfq_cic_link(struct cfq_data *cfqd, struct io_context *ioc,
-	     struct cfq_io_context *cic)
+/*
+ * Add cic into ioc, using cfqd as the search key. This enables us to lookup
+ * the process specific cfq io context when entered from the block layer.
+ * Also adds the cic to a per-cfqd list, used when this queue is removed.
+ */
+static int cfq_cic_link(struct cfq_data *cfqd, struct io_context *ioc,
+			struct cfq_io_context *cic, gfp_t gfp_mask)
 {
-	struct rb_node **p;
-	struct rb_node *parent;
-	struct cfq_io_context *__cic;
 	unsigned long flags;
-	void *k;
+	int ret;
 
-	cic->ioc = ioc;
-	cic->key = cfqd;
+	ret = radix_tree_preload(gfp_mask);
+	if (!ret) {
+		cic->ioc = ioc;
+		cic->key = cfqd;
 
-restart:
-	parent = NULL;
-	p = &ioc->cic_root.rb_node;
-	while (*p) {
-		parent = *p;
-		__cic = rb_entry(parent, struct cfq_io_context, rb_node);
-		/* ->key must be copied to avoid race with cfq_exit_queue() */
-		k = __cic->key;
-		if (unlikely(!k)) {
-			cfq_drop_dead_cic(ioc, __cic);
-			goto restart;
-		}
+		spin_lock_irqsave(&ioc->lock, flags);
+		ret = radix_tree_insert(&ioc->radix_root,
+						(unsigned long) cfqd, cic);
+		if (!ret)
+			hlist_add_head_rcu(&cic->cic_list, &ioc->cic_list);
+		spin_unlock_irqrestore(&ioc->lock, flags);
 
-		if (cic->key < k)
-			p = &(*p)->rb_left;
-		else if (cic->key > k)
-			p = &(*p)->rb_right;
-		else
-			BUG();
+		radix_tree_preload_end();
+
+		if (!ret) {
+			spin_lock_irqsave(cfqd->queue->queue_lock, flags);
+			list_add(&cic->queue_list, &cfqd->cic_list);
+			spin_unlock_irqrestore(cfqd->queue->queue_lock, flags);
+		}
 	}
 
-	rb_link_node(&cic->rb_node, parent, p);
-	rb_insert_color(&cic->rb_node, &ioc->cic_root);
+	if (ret)
+		printk(KERN_ERR "cfq: cic link failed!\n");
 
-	spin_lock_irqsave(cfqd->queue->queue_lock, flags);
-	list_add(&cic->queue_list, &cfqd->cic_list);
-	spin_unlock_irqrestore(cfqd->queue->queue_lock, flags);
+	return ret;
 }
 
 /*
@@ -1527,7 +1867,7 @@ cfq_get_io_context(struct cfq_data *cfqd, gfp_t gfp_mask)
 	if (!ioc)
 		return NULL;
 
-	cic = cfq_cic_rb_lookup(cfqd, ioc);
+	cic = cfq_cic_lookup(cfqd, ioc);
 	if (cic)
 		goto out;
 
@@ -1535,13 +1875,17 @@ cfq_get_io_context(struct cfq_data *cfqd, gfp_t gfp_mask)
 	if (cic == NULL)
 		goto err;
 
-	cfq_cic_link(cfqd, ioc, cic);
+	if (cfq_cic_link(cfqd, ioc, cic, gfp_mask))
+		goto err_free;
+
 out:
 	smp_read_barrier_depends();
 	if (unlikely(ioc->ioprio_changed))
 		cfq_ioc_set_ioprio(ioc);
 
 	return cic;
+err_free:
+	cfq_cic_free(cic);
 err:
 	put_io_context(ioc);
 	return NULL;
@@ -1565,15 +1909,12 @@ cfq_update_io_seektime(struct cfq_data *cfqd, struct cfq_io_context *cic,
 	sector_t sdist;
 	u64 total;
 
-	if (cic->last_request_pos < rq->sector)
-		sdist = rq->sector - cic->last_request_pos;
+	if (!cic->last_request_pos)
+		sdist = 0;
+	else if (cic->last_request_pos < blk_rq_pos(rq))
+		sdist = blk_rq_pos(rq) - cic->last_request_pos;
 	else
-		sdist = cic->last_request_pos - rq->sector;
-
-	if (!cic->seek_samples) {
-		cfqd->new_seek_total = (7*cic->seek_total + (u64)256*sdist) / 8;
-		cfqd->new_seek_mean = cfqd->new_seek_total / 256;
-	}
+		sdist = cic->last_request_pos - blk_rq_pos(rq);
 
 	/*
 	 * Don't allow the seek distance to get too large from the
@@ -1599,14 +1940,17 @@ static void
 cfq_update_idle_window(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 		       struct cfq_io_context *cic)
 {
-	int enable_idle;
+	int old_idle, enable_idle;
 
-	if (!cfq_cfqq_sync(cfqq))
+	/*
+	 * Don't idle for async or idle io prio class
+	 */
+	if (!cfq_cfqq_sync(cfqq) || cfq_class_idle(cfqq))
 		return;
 
-	enable_idle = cfq_cfqq_idle_window(cfqq);
+	enable_idle = old_idle = cfq_cfqq_idle_window(cfqq);
 
-	if (!cic->ioc->task || !cfqd->cfq_slice_idle ||
+	if (!atomic_read(&cic->ioc->nr_tasks) || !cfqd->cfq_slice_idle ||
 	    (cfqd->hw_tag && CIC_SEEKY(cic)))
 		enable_idle = 0;
 	else if (sample_valid(cic->ttime_samples)) {
@@ -1616,10 +1960,13 @@ cfq_update_idle_window(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 			enable_idle = 1;
 	}
 
-	if (enable_idle)
-		cfq_mark_cfqq_idle_window(cfqq);
-	else
-		cfq_clear_cfqq_idle_window(cfqq);
+	if (old_idle != enable_idle) {
+		cfq_log_cfqq(cfqd, cfqq, "idle=%d", enable_idle);
+		if (enable_idle)
+			cfq_mark_cfqq_idle_window(cfqq);
+		else
+			cfq_clear_cfqq_idle_window(cfqq);
+	}
 }
 
 /*
@@ -1659,6 +2006,12 @@ cfq_should_preempt(struct cfq_data *cfqd, struct cfq_queue *new_cfqq,
 	if (rq_is_meta(rq) && !cfqq->meta_pending)
 		return 1;
 
+	/*
+	 * Allow an RT request to pre-empt an ongoing non-RT cfqq timeslice.
+	 */
+	if (cfq_class_rt(new_cfqq) && !cfq_class_rt(cfqq))
+		return 1;
+
 	if (!cfqd->active_cic || !cfq_cfqq_wait_request(cfqq))
 		return 0;
 
@@ -1678,6 +2031,7 @@ cfq_should_preempt(struct cfq_data *cfqd, struct cfq_queue *new_cfqq,
  */
 static void cfq_preempt_queue(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 {
+	cfq_log_cfqq(cfqd, cfqq, "preempt");
 	cfq_slice_expired(cfqd, 1);
 
 	/*
@@ -1702,6 +2056,7 @@ cfq_rq_enqueued(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 {
 	struct cfq_io_context *cic = RQ_CIC(rq);
 
+	cfqd->rq_queued++;
 	if (rq_is_meta(rq))
 		cfqq->meta_pending++;
 
@@ -1709,38 +2064,46 @@ cfq_rq_enqueued(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 	cfq_update_io_seektime(cfqd, cic, rq);
 	cfq_update_idle_window(cfqd, cfqq, cic);
 
-	cic->last_request_pos = rq->sector + rq->nr_sectors;
-	cfqq->last_request_pos = cic->last_request_pos;
+	cic->last_request_pos = blk_rq_pos(rq) + blk_rq_sectors(rq);
 
 	if (cfqq == cfqd->active_queue) {
 		/*
-		 * if we are waiting for a request for this queue, let it rip
-		 * immediately and flag that we must not expire this queue
-		 * just now
+		 * Remember that we saw a request from this process, but
+		 * don't start queuing just yet. Otherwise we risk seeing lots
+		 * of tiny requests, because we disrupt the normal plugging
+		 * and merging. If the request is already larger than a single
+		 * page, let it rip immediately. For that case we assume that
+		 * merging is already done. Ditto for a busy system that
+		 * has other work pending, don't risk delaying until the
+		 * idle timer unplug to continue working.
 		 */
 		if (cfq_cfqq_wait_request(cfqq)) {
+			if (blk_rq_bytes(rq) > PAGE_CACHE_SIZE ||
+			    cfqd->busy_queues > 1) {
+				del_timer(&cfqd->idle_slice_timer);
+			__blk_run_queue(cfqd->queue);
+			}
 			cfq_mark_cfqq_must_dispatch(cfqq);
-			del_timer(&cfqd->idle_slice_timer);
-			blk_start_queueing(cfqd->queue);
 		}
 	} else if (cfq_should_preempt(cfqd, cfqq, rq)) {
 		/*
 		 * not the active queue - expire current slice if it is
 		 * idle and has expired it's mean thinktime or this new queue
-		 * has some old slice time left and is of higher priority
+		 * has some old slice time left and is of higher priority or
+		 * this new queue is RT and the current one is BE
 		 */
 		cfq_preempt_queue(cfqd, cfqq);
-		cfq_mark_cfqq_must_dispatch(cfqq);
-		blk_start_queueing(cfqd->queue);
+		__blk_run_queue(cfqd->queue);
 	}
 }
 
-static void cfq_insert_request(request_queue_t *q, struct request *rq)
+static void cfq_insert_request(struct request_queue *q, struct request *rq)
 {
 	struct cfq_data *cfqd = q->elevator->elevator_data;
 	struct cfq_queue *cfqq = RQ_CFQQ(rq);
 
-	cfq_init_prio_data(cfqq);
+	cfq_log_cfqq(cfqd, cfqq, "insert_request");
+	cfq_init_prio_data(cfqq, RQ_CIC(rq)->ioc);
 
 	cfq_add_rq_rb(rq);
 
@@ -1749,7 +2112,32 @@ static void cfq_insert_request(request_queue_t *q, struct request *rq)
 	cfq_rq_enqueued(cfqd, cfqq, rq);
 }
 
-static void cfq_completed_request(request_queue_t *q, struct request *rq)
+/*
+ * Update hw_tag based on peak queue depth over 50 samples under
+ * sufficient load.
+ */
+static void cfq_update_hw_tag(struct cfq_data *cfqd)
+{
+	if (rq_in_driver(cfqd) > cfqd->rq_in_driver_peak)
+		cfqd->rq_in_driver_peak = rq_in_driver(cfqd);
+
+	if (cfqd->rq_queued <= CFQ_HW_QUEUE_MIN &&
+	    rq_in_driver(cfqd) <= CFQ_HW_QUEUE_MIN)
+		return;
+
+	if (cfqd->hw_tag_samples++ < 50)
+		return;
+
+	if (cfqd->rq_in_driver_peak >= CFQ_HW_QUEUE_MIN)
+		cfqd->hw_tag = 1;
+	else
+		cfqd->hw_tag = 0;
+
+	cfqd->hw_tag_samples = 0;
+	cfqd->rq_in_driver_peak = 0;
+}
+
+static void cfq_completed_request(struct request_queue *q, struct request *rq)
 {
 	struct cfq_queue *cfqq = RQ_CFQQ(rq);
 	struct cfq_data *cfqd = cfqq->cfqd;
@@ -1757,18 +2145,18 @@ static void cfq_completed_request(request_queue_t *q, struct request *rq)
 	unsigned long now;
 
 	now = jiffies;
+	cfq_log_cfqq(cfqd, cfqq, "complete");
+
+	cfq_update_hw_tag(cfqd);
 
-	WARN_ON(!cfqd->rq_in_driver);
+	WARN_ON(!cfqd->rq_in_driver[sync]);
 	WARN_ON(!cfqq->dispatched);
-	cfqd->rq_in_driver--;
+	cfqd->rq_in_driver[sync]--;
 	cfqq->dispatched--;
 
 	if (cfq_cfqq_sync(cfqq))
 		cfqd->sync_flight--;
 
-	if (!cfq_class_idle(cfqq))
-		cfqd->last_end_request = now;
-
 	if (sync)
 		RQ_CIC(rq)->last_end_request = now;
 
@@ -1777,17 +2165,27 @@ static void cfq_completed_request(request_queue_t *q, struct request *rq)
 	 * or if we want to idle in case it has no pending requests.
 	 */
 	if (cfqd->active_queue == cfqq) {
+		const bool cfqq_empty = RB_EMPTY_ROOT(&cfqq->sort_list);
+
 		if (cfq_cfqq_slice_new(cfqq)) {
 			cfq_set_prio_slice(cfqd, cfqq);
 			cfq_clear_cfqq_slice_new(cfqq);
 		}
-		if (cfq_slice_used(cfqq))
+		/*
+		 * If there are no requests waiting in this queue, and
+		 * there are other queues ready to issue requests, AND
+		 * those other queues are issuing requests within our
+		 * mean seek distance, give them a chance to run instead
+		 * of idling.
+		 */
+		if (cfq_slice_used(cfqq) || cfq_class_idle(cfqq))
 			cfq_slice_expired(cfqd, 1);
-		else if (sync && RB_EMPTY_ROOT(&cfqq->sort_list))
+		else if (cfqq_empty && !cfq_close_cooperator(cfqd, cfqq, 1) &&
+			 sync && !rq_noidle(rq))
 			cfq_arm_slice_timer(cfqd);
 	}
 
-	if (!cfqd->rq_in_driver)
+	if (!rq_in_driver(cfqd))
 		cfq_schedule_dispatch(cfqd);
 }
 
@@ -1819,8 +2217,7 @@ static void cfq_prio_boost(struct cfq_queue *cfqq)
 
 static inline int __cfq_may_queue(struct cfq_queue *cfqq)
 {
-	if ((cfq_cfqq_wait_request(cfqq) || cfq_cfqq_must_alloc(cfqq)) &&
-	    !cfq_cfqq_must_alloc_slice(cfqq)) {
+	if (cfq_cfqq_wait_request(cfqq) && !cfq_cfqq_must_alloc_slice(cfqq)) {
 		cfq_mark_cfqq_must_alloc_slice(cfqq);
 		return ELV_MQUEUE_MUST;
 	}
@@ -1828,7 +2225,7 @@ static inline int __cfq_may_queue(struct cfq_queue *cfqq)
 	return ELV_MQUEUE_MAY;
 }
 
-static int cfq_may_queue(request_queue_t *q, int rw)
+static int cfq_may_queue(struct request_queue *q, int rw)
 {
 	struct cfq_data *cfqd = q->elevator->elevator_data;
 	struct task_struct *tsk = current;
@@ -1841,13 +2238,13 @@ static int cfq_may_queue(request_queue_t *q, int rw)
 	 * so just lookup a possibly existing queue, or return 'may queue'
 	 * if that fails
 	 */
-	cic = cfq_cic_rb_lookup(cfqd, tsk->io_context);
+	cic = cfq_cic_lookup(cfqd, tsk->io_context);
 	if (!cic)
 		return ELV_MQUEUE_MAY;
 
-	cfqq = cic_to_cfqq(cic, rw & REQ_RW_SYNC);
+	cfqq = cic_to_cfqq(cic, rw_is_sync(rw));
 	if (cfqq) {
-		cfq_init_prio_data(cfqq);
+		cfq_init_prio_data(cfqq, cic->ioc);
 		cfq_prio_boost(cfqq);
 
 		return __cfq_may_queue(cfqq);
@@ -1882,10 +2279,9 @@ static void cfq_put_request(struct request *rq)
  * Allocate cfq data structures associated with this request.
  */
 static int
-cfq_set_request(request_queue_t *q, struct request *rq, gfp_t gfp_mask)
+cfq_set_request(struct request_queue *q, struct request *rq, gfp_t gfp_mask)
 {
 	struct cfq_data *cfqd = q->elevator->elevator_data;
-	struct task_struct *tsk = current;
 	struct cfq_io_context *cic;
 	const int rw = rq_data_dir(rq);
 	const int is_sync = rq_is_sync(rq);
@@ -1902,17 +2298,12 @@ cfq_set_request(request_queue_t *q, struct request *rq, gfp_t gfp_mask)
 		goto queue_fail;
 
 	cfqq = cic_to_cfqq(cic, is_sync);
-	if (!cfqq) {
-		cfqq = cfq_get_queue(cfqd, is_sync, tsk, gfp_mask);
-
-		if (!cfqq)
-			goto queue_fail;
-
+	if (!cfqq || cfqq == &cfqd->oom_cfqq) {
+		cfqq = cfq_get_queue(cfqd, is_sync, cic->ioc, gfp_mask);
 		cic_set_cfqq(cic, cfqq, is_sync);
 	}
 
 	cfqq->allocated[rw]++;
-	cfq_clear_cfqq_must_alloc(cfqq);
 	atomic_inc(&cfqq->ref);
 
 	spin_unlock_irqrestore(q->queue_lock, flags);
@@ -1927,6 +2318,7 @@ queue_fail:
 
 	cfq_schedule_dispatch(cfqd);
 	spin_unlock_irqrestore(q->queue_lock, flags);
+	cfq_log(cfqd, "set_request fail");
 	return 1;
 }
 
@@ -1934,12 +2326,11 @@ static void cfq_kick_queue(struct work_struct *work)
 {
 	struct cfq_data *cfqd =
 		container_of(work, struct cfq_data, unplug_work);
-	request_queue_t *q = cfqd->queue;
-	unsigned long flags;
+	struct request_queue *q = cfqd->queue;
 
-	spin_lock_irqsave(q->queue_lock, flags);
-	blk_start_queueing(q);
-	spin_unlock_irqrestore(q->queue_lock, flags);
+	spin_lock_irq(q->queue_lock);
+	__blk_run_queue(cfqd->queue);
+	spin_unlock_irq(q->queue_lock);
 }
 
 /*
@@ -1952,12 +2343,21 @@ static void cfq_idle_slice_timer(unsigned long data)
 	unsigned long flags;
 	int timed_out = 1;
 
+	cfq_log(cfqd, "idle timer fired");
+
 	spin_lock_irqsave(cfqd->queue->queue_lock, flags);
 
-	if ((cfqq = cfqd->active_queue) != NULL) {
+	cfqq = cfqd->active_queue;
+	if (cfqq) {
 		timed_out = 0;
 
 		/*
+		 * We saw a request before the queue expired, let it through
+		 */
+		if (cfq_cfqq_must_dispatch(cfqq))
+			goto out_kick;
+
+		/*
 		 * expired
 		 */
 		if (cfq_slice_used(cfqq))
@@ -1973,10 +2373,8 @@ static void cfq_idle_slice_timer(unsigned long data)
 		/*
 		 * not expired and it has a request pending, let it dispatch
 		 */
-		if (!RB_EMPTY_ROOT(&cfqq->sort_list)) {
-			cfq_mark_cfqq_must_dispatch(cfqq);
+		if (!RB_EMPTY_ROOT(&cfqq->sort_list))
 			goto out_kick;
-		}
 	}
 expire:
 	cfq_slice_expired(cfqd, timed_out);
@@ -1986,39 +2384,31 @@ out_cont:
 	spin_unlock_irqrestore(cfqd->queue->queue_lock, flags);
 }
 
-/*
- * Timer running if an idle class queue is waiting for service
- */
-static void cfq_idle_class_timer(unsigned long data)
+static void cfq_shutdown_timer_wq(struct cfq_data *cfqd)
 {
-	struct cfq_data *cfqd = (struct cfq_data *) data;
-	unsigned long flags, end;
-
-	spin_lock_irqsave(cfqd->queue->queue_lock, flags);
-
-	/*
-	 * race with a non-idle queue, reset timer
-	 */
-	end = cfqd->last_end_request + CFQ_IDLE_GRACE;
-	if (!time_after_eq(jiffies, end))
-		mod_timer(&cfqd->idle_class_timer, end);
-	else
-		cfq_schedule_dispatch(cfqd);
-
-	spin_unlock_irqrestore(cfqd->queue->queue_lock, flags);
+	del_timer_sync(&cfqd->idle_slice_timer);
+	cancel_work_sync(&cfqd->unplug_work);
 }
 
-static void cfq_shutdown_timer_wq(struct cfq_data *cfqd)
+static void cfq_put_async_queues(struct cfq_data *cfqd)
 {
-	del_timer_sync(&cfqd->idle_slice_timer);
-	del_timer_sync(&cfqd->idle_class_timer);
-	blk_sync_queue(cfqd->queue);
+	int i;
+
+	for (i = 0; i < IOPRIO_BE_NR; i++) {
+		if (cfqd->async_cfqq[0][i])
+			cfq_put_queue(cfqd->async_cfqq[0][i]);
+		if (cfqd->async_cfqq[1][i])
+			cfq_put_queue(cfqd->async_cfqq[1][i]);
+	}
+
+	if (cfqd->async_idle_cfqq)
+		cfq_put_queue(cfqd->async_idle_cfqq);
 }
 
-static void cfq_exit_queue(elevator_t *e)
+static void cfq_exit_queue(struct elevator_queue *e)
 {
 	struct cfq_data *cfqd = e->elevator_data;
-	request_queue_t *q = cfqd->queue;
+	struct request_queue *q = cfqd->queue;
 
 	cfq_shutdown_timer_wq(cfqd);
 
@@ -2035,6 +2425,8 @@ static void cfq_exit_queue(elevator_t *e)
 		__cfq_exit_single_io_context(cfqd, cic);
 	}
 
+	cfq_put_async_queues(cfqd);
+
 	spin_unlock_irq(q->queue_lock);
 
 	cfq_shutdown_timer_wq(cfqd);
@@ -2042,17 +2434,33 @@ static void cfq_exit_queue(elevator_t *e)
 	kfree(cfqd);
 }
 
-static void *cfq_init_queue(request_queue_t *q)
+static void *cfq_init_queue(struct request_queue *q)
 {
 	struct cfq_data *cfqd;
+	int i;
 
-	cfqd = kmalloc_node(sizeof(*cfqd), GFP_KERNEL, q->node);
+	cfqd = kmalloc_node(sizeof(*cfqd), GFP_KERNEL | __GFP_ZERO, q->node);
 	if (!cfqd)
 		return NULL;
 
-	memset(cfqd, 0, sizeof(*cfqd));
-
 	cfqd->service_tree = CFQ_RB_ROOT;
+
+	/*
+	 * Not strictly needed (since RB_ROOT just clears the node and we
+	 * zeroed cfqd on alloc), but better be safe in case someone decides
+	 * to add magic to the rb code
+	 */
+	for (i = 0; i < CFQ_PRIO_LISTS; i++)
+		cfqd->prio_trees[i] = RB_ROOT;
+
+	/*
+	 * Our fallback cfqq if cfq_find_alloc_queue() runs into OOM issues.
+	 * Grab a permanent reference to it, so that the normal code flow
+	 * will not attempt to free it.
+	 */
+	cfq_init_cfqq(cfqd, &cfqd->oom_cfqq, 1, 0);
+	atomic_inc(&cfqd->oom_cfqq.ref);
+
 	INIT_LIST_HEAD(&cfqd->cic_list);
 
 	cfqd->queue = q;
@@ -2061,10 +2469,6 @@ static void *cfq_init_queue(request_queue_t *q)
 	cfqd->idle_slice_timer.function = cfq_idle_slice_timer;
 	cfqd->idle_slice_timer.data = (unsigned long) cfqd;
 
-	init_timer(&cfqd->idle_class_timer);
-	cfqd->idle_class_timer.function = cfq_idle_class_timer;
-	cfqd->idle_class_timer.data = (unsigned long) cfqd;
-
 	INIT_WORK(&cfqd->unplug_work, cfq_kick_queue);
 
 	cfqd->cfq_quantum = cfq_quantum;
@@ -2076,12 +2480,17 @@ static void *cfq_init_queue(request_queue_t *q)
 	cfqd->cfq_slice[1] = cfq_slice_sync;
 	cfqd->cfq_slice_async_rq = cfq_slice_async_rq;
 	cfqd->cfq_slice_idle = cfq_slice_idle;
+	cfqd->hw_tag = 1;
 
 	return cfqd;
 }
 
 static void cfq_slab_kill(void)
 {
+	/*
+	 * Caller already ensured that pending RCU callbacks are completed,
+	 * so we should have no busy allocations at this point.
+	 */
 	if (cfq_pool)
 		kmem_cache_destroy(cfq_pool);
 	if (cfq_ioc_pool)
@@ -2123,7 +2532,7 @@ cfq_var_store(unsigned int *var, const char *page, size_t count)
 }
 
 #define SHOW_FUNCTION(__FUNC, __VAR, __CONV)				\
-static ssize_t __FUNC(elevator_t *e, char *page)			\
+static ssize_t __FUNC(struct elevator_queue *e, char *page)		\
 {									\
 	struct cfq_data *cfqd = e->elevator_data;			\
 	unsigned int __data = __VAR;					\
@@ -2143,7 +2552,7 @@ SHOW_FUNCTION(cfq_slice_async_rq_show, cfqd->cfq_slice_async_rq, 0);
 #undef SHOW_FUNCTION
 
 #define STORE_FUNCTION(__FUNC, __PTR, MIN, MAX, __CONV)			\
-static ssize_t __FUNC(elevator_t *e, const char *page, size_t count)	\
+static ssize_t __FUNC(struct elevator_queue *e, const char *page, size_t count)	\
 {									\
 	struct cfq_data *cfqd = e->elevator_data;			\
 	unsigned int __data;						\
@@ -2159,14 +2568,18 @@ static ssize_t __FUNC(elevator_t *e, const char *page, size_t count)	\
 	return ret;							\
 }
 STORE_FUNCTION(cfq_quantum_store, &cfqd->cfq_quantum, 1, UINT_MAX, 0);
-STORE_FUNCTION(cfq_fifo_expire_sync_store, &cfqd->cfq_fifo_expire[1], 1, UINT_MAX, 1);
-STORE_FUNCTION(cfq_fifo_expire_async_store, &cfqd->cfq_fifo_expire[0], 1, UINT_MAX, 1);
+STORE_FUNCTION(cfq_fifo_expire_sync_store, &cfqd->cfq_fifo_expire[1], 1,
+		UINT_MAX, 1);
+STORE_FUNCTION(cfq_fifo_expire_async_store, &cfqd->cfq_fifo_expire[0], 1,
+		UINT_MAX, 1);
 STORE_FUNCTION(cfq_back_seek_max_store, &cfqd->cfq_back_max, 0, UINT_MAX, 0);
-STORE_FUNCTION(cfq_back_seek_penalty_store, &cfqd->cfq_back_penalty, 1, UINT_MAX, 0);
+STORE_FUNCTION(cfq_back_seek_penalty_store, &cfqd->cfq_back_penalty, 1,
+		UINT_MAX, 0);
 STORE_FUNCTION(cfq_slice_idle_store, &cfqd->cfq_slice_idle, 0, UINT_MAX, 1);
 STORE_FUNCTION(cfq_slice_sync_store, &cfqd->cfq_slice[1], 1, UINT_MAX, 1);
 STORE_FUNCTION(cfq_slice_async_store, &cfqd->cfq_slice[0], 1, UINT_MAX, 1);
-STORE_FUNCTION(cfq_slice_async_rq_store, &cfqd->cfq_slice_async_rq, 1, UINT_MAX, 0);
+STORE_FUNCTION(cfq_slice_async_rq_store, &cfqd->cfq_slice_async_rq, 1,
+		UINT_MAX, 0);
 #undef STORE_FUNCTION
 
 #define CFQ_ATTR(name) \
@@ -2213,8 +2626,6 @@ static struct elevator_type iosched_cfq = {
 
 static int __init cfq_init(void)
 {
-	int ret;
-
 	/*
 	 * could be 0 on HZ < 1000 setups
 	 */
@@ -2226,11 +2637,9 @@ static int __init cfq_init(void)
 	if (cfq_slab_setup())
 		return -ENOMEM;
 
-	ret = elv_register(&iosched_cfq);
-	if (ret)
-		cfq_slab_kill();
+	elv_register(&iosched_cfq);
 
-	return ret;
+	return 0;
 }
 
 static void __exit cfq_exit(void)
@@ -2240,9 +2649,13 @@ static void __exit cfq_exit(void)
 	ioc_gone = &all_gone;
 	/* ioc_gone's update must be visible before reading ioc_count */
 	smp_wmb();
+
+	/*
+	 * this also protects us from entering cfq_slab_kill() with
+	 * pending RCU callbacks
+	 */
 	if (elv_ioc_count_read(ioc_count))
-		wait_for_completion(ioc_gone);
-	synchronize_rcu();
+		wait_for_completion(&all_gone);
 	cfq_slab_kill();
 }