Merge rsync://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6

Conflicts:

	include/linux/kernel.h

8 files changed, 160 insertions, 214 deletions

diff --git a/block/Kconfig.iosched b/block/Kconfig.iosched
index f3b7753aac9..48d090e266f 100644
--- a/block/Kconfig.iosched
+++ b/block/Kconfig.iosched
@@ -40,7 +40,7 @@ config IOSCHED_CFQ
 
 choice
 	prompt "Default I/O scheduler"
-	default DEFAULT_AS
+	default DEFAULT_CFQ
 	help
 	  Select the I/O scheduler which will be used by default for all
 	  block devices.
diff --git a/block/as-iosched.c b/block/as-iosched.c
index a7caf35ca0c..5da56d48fbd 100644
--- a/block/as-iosched.c
+++ b/block/as-iosched.c
@@ -10,7 +10,6 @@
 #include <linux/blkdev.h>
 #include <linux/elevator.h>
 #include <linux/bio.h>
-#include <linux/config.h>
 #include <linux/module.h>
 #include <linux/slab.h>
 #include <linux/init.h>
@@ -96,7 +95,7 @@ struct as_data {
 
 	struct as_rq *next_arq[2];	/* next in sort order */
 	sector_t last_sector[2];	/* last REQ_SYNC & REQ_ASYNC sectors */
-	struct list_head *hash;		/* request hash */
+	struct hlist_head *hash;	/* request hash */
 
 	unsigned long exit_prob;	/* probability a task will exit while
 					   being waited on */
@@ -165,8 +164,7 @@ struct as_rq {
 	/*
 	 * request hash, key is the ending offset (for back merge lookup)
 	 */
-	struct list_head hash;
-	unsigned int on_hash;
+	struct hlist_node hash;
 
 	/*
 	 * expire fifo
@@ -282,17 +280,15 @@ static const int as_hash_shift = 6;
 #define AS_HASH_FN(sec)		(hash_long(AS_HASH_BLOCK((sec)), as_hash_shift))
 #define AS_HASH_ENTRIES		(1 << as_hash_shift)
 #define rq_hash_key(rq)		((rq)->sector + (rq)->nr_sectors)
-#define list_entry_hash(ptr)	list_entry((ptr), struct as_rq, hash)
 
 static inline void __as_del_arq_hash(struct as_rq *arq)
 {
-	arq->on_hash = 0;
-	list_del_init(&arq->hash);
+	hlist_del_init(&arq->hash);
 }
 
 static inline void as_del_arq_hash(struct as_rq *arq)
 {
-	if (arq->on_hash)
+	if (!hlist_unhashed(&arq->hash))
 		__as_del_arq_hash(arq);
 }
 
@@ -300,10 +296,9 @@ static void as_add_arq_hash(struct as_data *ad, struct as_rq *arq)
 {
 	struct request *rq = arq->request;
 
-	BUG_ON(arq->on_hash);
+	BUG_ON(!hlist_unhashed(&arq->hash));
 
-	arq->on_hash = 1;
-	list_add(&arq->hash, &ad->hash[AS_HASH_FN(rq_hash_key(rq))]);
+	hlist_add_head(&arq->hash, &ad->hash[AS_HASH_FN(rq_hash_key(rq))]);
 }
 
 /*
@@ -312,31 +307,29 @@ static void as_add_arq_hash(struct as_data *ad, struct as_rq *arq)
 static inline void as_hot_arq_hash(struct as_data *ad, struct as_rq *arq)
 {
 	struct request *rq = arq->request;
-	struct list_head *head = &ad->hash[AS_HASH_FN(rq_hash_key(rq))];
+	struct hlist_head *head = &ad->hash[AS_HASH_FN(rq_hash_key(rq))];
 
-	if (!arq->on_hash) {
+	if (hlist_unhashed(&arq->hash)) {
 		WARN_ON(1);
 		return;
 	}
 
-	if (arq->hash.prev != head) {
-		list_del(&arq->hash);
-		list_add(&arq->hash, head);
+	if (&arq->hash != head->first) {
+		hlist_del(&arq->hash);
+		hlist_add_head(&arq->hash, head);
 	}
 }
 
 static struct request *as_find_arq_hash(struct as_data *ad, sector_t offset)
 {
-	struct list_head *hash_list = &ad->hash[AS_HASH_FN(offset)];
-	struct list_head *entry, *next = hash_list->next;
+	struct hlist_head *hash_list = &ad->hash[AS_HASH_FN(offset)];
+	struct hlist_node *entry, *next;
+	struct as_rq *arq;
 
-	while ((entry = next) != hash_list) {
-		struct as_rq *arq = list_entry_hash(entry);
+	hlist_for_each_entry_safe(arq, entry, next, hash_list, hash) {
 		struct request *__rq = arq->request;
 
-		next = entry->next;
-
-		BUG_ON(!arq->on_hash);
+		BUG_ON(hlist_unhashed(&arq->hash));
 
 		if (!rq_mergeable(__rq)) {
 			as_del_arq_hash(arq);
@@ -353,10 +346,6 @@ static struct request *as_find_arq_hash(struct as_data *ad, sector_t offset)
 /*
  * rb tree support functions
  */
-#define RB_NONE		(2)
-#define RB_EMPTY(root)	((root)->rb_node == NULL)
-#define ON_RB(node)	((node)->rb_color != RB_NONE)
-#define RB_CLEAR(node)	((node)->rb_color = RB_NONE)
 #define rb_entry_arq(node)	rb_entry((node), struct as_rq, rb_node)
 #define ARQ_RB_ROOT(ad, arq)	(&(ad)->sort_list[(arq)->is_sync])
 #define rq_rb_key(rq)		(rq)->sector
@@ -425,13 +414,13 @@ static void as_add_arq_rb(struct as_data *ad, struct as_rq *arq)
 
 static inline void as_del_arq_rb(struct as_data *ad, struct as_rq *arq)
 {
-	if (!ON_RB(&arq->rb_node)) {
+	if (!RB_EMPTY_NODE(&arq->rb_node)) {
 		WARN_ON(1);
 		return;
 	}
 
 	rb_erase(&arq->rb_node, ARQ_RB_ROOT(ad, arq));
-	RB_CLEAR(&arq->rb_node);
+	RB_CLEAR_NODE(&arq->rb_node);
 }
 
 static struct request *
@@ -552,7 +541,7 @@ static struct as_rq *as_find_next_arq(struct as_data *ad, struct as_rq *last)
 	struct rb_node *rbprev = rb_prev(&last->rb_node);
 	struct as_rq *arq_next, *arq_prev;
 
-	BUG_ON(!ON_RB(&last->rb_node));
+	BUG_ON(!RB_EMPTY_NODE(&last->rb_node));
 
 	if (rbprev)
 		arq_prev = rb_entry_arq(rbprev);
@@ -902,7 +891,7 @@ static int as_can_break_anticipation(struct as_data *ad, struct as_rq *arq)
 }
 
 /*
- * as_can_anticipate indicates weather we should either run arq
+ * as_can_anticipate indicates whether we should either run arq
  * or keep anticipating a better request.
  */
 static int as_can_anticipate(struct as_data *ad, struct as_rq *arq)
@@ -1129,7 +1118,7 @@ static void as_move_to_dispatch(struct as_data *ad, struct as_rq *arq)
 	struct request *rq = arq->request;
 	const int data_dir = arq->is_sync;
 
-	BUG_ON(!ON_RB(&arq->rb_node));
+	BUG_ON(!RB_EMPTY_NODE(&arq->rb_node));
 
 	as_antic_stop(ad);
 	ad->antic_status = ANTIC_OFF;
@@ -1254,7 +1243,7 @@ static int as_dispatch_request(request_queue_t *q, int force)
 	 */
 
 	if (reads) {
-		BUG_ON(RB_EMPTY(&ad->sort_list[REQ_SYNC]));
+		BUG_ON(RB_EMPTY_ROOT(&ad->sort_list[REQ_SYNC]));
 
 		if (writes && ad->batch_data_dir == REQ_SYNC)
 			/*
@@ -1278,7 +1267,7 @@ static int as_dispatch_request(request_queue_t *q, int force)
 
 	if (writes) {
 dispatch_writes:
-		BUG_ON(RB_EMPTY(&ad->sort_list[REQ_ASYNC]));
+		BUG_ON(RB_EMPTY_ROOT(&ad->sort_list[REQ_ASYNC]));
 
 		if (ad->batch_data_dir == REQ_SYNC) {
 			ad->changed_batch = 1;
@@ -1346,7 +1335,7 @@ static void as_add_request(request_queue_t *q, struct request *rq)
 	arq->state = AS_RQ_NEW;
 
 	if (rq_data_dir(arq->request) == READ
-			|| current->flags&PF_SYNCWRITE)
+			|| (arq->request->flags & REQ_RW_SYNC))
 		arq->is_sync = 1;
 	else
 		arq->is_sync = 0;
@@ -1598,12 +1587,11 @@ static int as_set_request(request_queue_t *q, struct request *rq,
 
 	if (arq) {
 		memset(arq, 0, sizeof(*arq));
-		RB_CLEAR(&arq->rb_node);
+		RB_CLEAR_NODE(&arq->rb_node);
 		arq->request = rq;
 		arq->state = AS_RQ_PRESCHED;
 		arq->io_context = NULL;
-		INIT_LIST_HEAD(&arq->hash);
-		arq->on_hash = 0;
+		INIT_HLIST_NODE(&arq->hash);
 		INIT_LIST_HEAD(&arq->fifo);
 		rq->elevator_private = arq;
 		return 0;
@@ -1663,7 +1651,7 @@ static void *as_init_queue(request_queue_t *q, elevator_t *e)
 
 	ad->q = q; /* Identify what queue the data belongs to */
 
-	ad->hash = kmalloc_node(sizeof(struct list_head)*AS_HASH_ENTRIES,
+	ad->hash = kmalloc_node(sizeof(struct hlist_head)*AS_HASH_ENTRIES,
 				GFP_KERNEL, q->node);
 	if (!ad->hash) {
 		kfree(ad);
@@ -1685,7 +1673,7 @@ static void *as_init_queue(request_queue_t *q, elevator_t *e)
 	INIT_WORK(&ad->antic_work, as_work_handler, q);
 
 	for (i = 0; i < AS_HASH_ENTRIES; i++)
-		INIT_LIST_HEAD(&ad->hash[i]);
+		INIT_HLIST_HEAD(&ad->hash[i]);
 
 	INIT_LIST_HEAD(&ad->fifo_list[REQ_SYNC]);
 	INIT_LIST_HEAD(&ad->fifo_list[REQ_ASYNC]);
diff --git a/block/blktrace.c b/block/blktrace.c
index 36f3a172275..92925e7d9e6 100644
--- a/block/blktrace.c
+++ b/block/blktrace.c
@@ -15,7 +15,6 @@
  * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  *
  */
-#include <linux/config.h>
 #include <linux/kernel.h>
 #include <linux/blkdev.h>
 #include <linux/blktrace_api.h>
diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
index 052b1748762..102ebc2c5c3 100644
--- a/block/cfq-iosched.c
+++ b/block/cfq-iosched.c
@@ -6,7 +6,6 @@
  *
  *  Copyright (C) 2003 Jens Axboe <axboe@suse.de>
  */
-#include <linux/config.h>
 #include <linux/module.h>
 #include <linux/blkdev.h>
 #include <linux/elevator.h>
@@ -26,7 +25,7 @@ static const int cfq_back_penalty = 2;		/* penalty of a backwards seek */
 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 / 70;
+static int cfq_slice_idle = HZ / 125;
 
 #define CFQ_IDLE_GRACE		(HZ / 10)
 #define CFQ_SLICE_SCALE		(5)
@@ -60,16 +59,6 @@ static DEFINE_SPINLOCK(cfq_exit_lock);
 /*
  * rb-tree defines
  */
-#define RB_NONE			(2)
-#define RB_EMPTY(node)		((node)->rb_node == NULL)
-#define RB_CLEAR_COLOR(node)	(node)->rb_color = RB_NONE
-#define RB_CLEAR(node)		do {	\
-	(node)->rb_parent = NULL;	\
-	RB_CLEAR_COLOR((node));		\
-	(node)->rb_right = NULL;	\
-	(node)->rb_left = NULL;		\
-} while (0)
-#define RB_CLEAR_ROOT(root)	((root)->rb_node = NULL)
 #define rb_entry_crq(node)	rb_entry((node), struct cfq_rq, rb_node)
 #define rq_rb_key(rq)		(rq)->sector
 
@@ -128,8 +117,6 @@ struct cfq_data {
 	 */
 	struct hlist_head *crq_hash;
 
-	unsigned int max_queued;
-
 	mempool_t *crq_pool;
 
 	int rq_in_driver;
@@ -284,8 +271,6 @@ static struct cfq_queue *cfq_find_cfq_hash(struct cfq_data *, unsigned int, unsi
 static void cfq_dispatch_insert(request_queue_t *, struct cfq_rq *);
 static struct cfq_queue *cfq_get_queue(struct cfq_data *cfqd, unsigned int key, struct task_struct *tsk, gfp_t gfp_mask);
 
-#define process_sync(tsk)	((tsk)->flags & PF_SYNCWRITE)
-
 /*
  * lots of deadline iosched dupes, can be abstracted later...
  */
@@ -341,7 +326,7 @@ static int cfq_queue_empty(request_queue_t *q)
 
 static inline pid_t cfq_queue_pid(struct task_struct *task, int rw)
 {
-	if (rw == READ || process_sync(task))
+	if (rw == READ || rw == WRITE_SYNC)
 		return task->pid;
 
 	return CFQ_KEY_ASYNC;
@@ -567,9 +552,8 @@ static inline void cfq_del_crq_rb(struct cfq_rq *crq)
 	cfq_update_next_crq(crq);
 
 	rb_erase(&crq->rb_node, &cfqq->sort_list);
-	RB_CLEAR_COLOR(&crq->rb_node);
 
-	if (cfq_cfqq_on_rr(cfqq) && RB_EMPTY(&cfqq->sort_list))
+	if (cfq_cfqq_on_rr(cfqq) && RB_EMPTY_ROOT(&cfqq->sort_list))
 		cfq_del_cfqq_rr(cfqd, cfqq);
 }
 
@@ -916,13 +900,15 @@ static struct cfq_queue *cfq_set_active_queue(struct cfq_data *cfqd)
 	return cfqq;
 }
 
+#define CIC_SEEKY(cic) ((cic)->seek_mean > (128 * 1024))
+
 static int cfq_arm_slice_timer(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 
 {
 	struct cfq_io_context *cic;
 	unsigned long sl;
 
-	WARN_ON(!RB_EMPTY(&cfqq->sort_list));
+	WARN_ON(!RB_EMPTY_ROOT(&cfqq->sort_list));
 	WARN_ON(cfqq != cfqd->active_queue);
 
 	/*
@@ -949,7 +935,7 @@ static int cfq_arm_slice_timer(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 	 * fair distribution of slice time for a process doing back-to-back
 	 * seeks. so allow a little bit of time for him to submit a new rq
 	 */
-	if (sample_valid(cic->seek_samples) && cic->seek_mean > 131072)
+	if (sample_valid(cic->seek_samples) && CIC_SEEKY(cic))
 		sl = 2;
 
 	mod_timer(&cfqd->idle_slice_timer, jiffies + sl);
@@ -960,11 +946,15 @@ static void cfq_dispatch_insert(request_queue_t *q, struct cfq_rq *crq)
 {
 	struct cfq_data *cfqd = q->elevator->elevator_data;
 	struct cfq_queue *cfqq = crq->cfq_queue;
+	struct request *rq;
 
 	cfqq->next_crq = cfq_find_next_crq(cfqd, cfqq, crq);
 	cfq_remove_request(crq->request);
 	cfqq->on_dispatch[cfq_crq_is_sync(crq)]++;
 	elv_dispatch_sort(q, crq->request);
+
+	rq = list_entry(q->queue_head.prev, struct request, queuelist);
+	cfqd->last_sector = rq->sector + rq->nr_sectors;
 }
 
 /*
@@ -1046,10 +1036,12 @@ static struct cfq_queue *cfq_select_queue(struct cfq_data *cfqd)
 	 * if queue has requests, dispatch one. if not, check if
 	 * enough slice is left to wait for one
 	 */
-	if (!RB_EMPTY(&cfqq->sort_list))
+	if (!RB_EMPTY_ROOT(&cfqq->sort_list))
 		goto keep_queue;
-	else if (cfq_cfqq_class_sync(cfqq) &&
-		 time_before(now, cfqq->slice_end)) {
+	else if (cfq_cfqq_dispatched(cfqq)) {
+		cfqq = NULL;
+		goto keep_queue;
+	} else if (cfq_cfqq_class_sync(cfqq)) {
 		if (cfq_arm_slice_timer(cfqd, cfqq))
 			return NULL;
 	}
@@ -1068,7 +1060,7 @@ __cfq_dispatch_requests(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 {
 	int dispatched = 0;
 
-	BUG_ON(RB_EMPTY(&cfqq->sort_list));
+	BUG_ON(RB_EMPTY_ROOT(&cfqq->sort_list));
 
 	do {
 		struct cfq_rq *crq;
@@ -1092,14 +1084,13 @@ __cfq_dispatch_requests(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 			cfqd->active_cic = crq->io_context;
 		}
 
-		if (RB_EMPTY(&cfqq->sort_list))
+		if (RB_EMPTY_ROOT(&cfqq->sort_list))
 			break;
 
 	} while (dispatched < max_dispatch);
 
 	/*
-	 * if slice end isn't set yet, set it. if at least one request was
-	 * sync, use the sync time slice value
+	 * if slice end isn't set yet, set it.
 	 */
 	if (!cfqq->slice_end)
 		cfq_set_prio_slice(cfqd, cfqq);
@@ -1110,7 +1101,8 @@ __cfq_dispatch_requests(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 	 */
 	if ((!cfq_cfqq_sync(cfqq) &&
 	    cfqd->dispatch_slice >= cfq_prio_to_maxrq(cfqd, cfqq)) ||
-	    cfq_class_idle(cfqq))
+	    cfq_class_idle(cfqq) ||
+	    !cfq_cfqq_idle_window(cfqq))
 		cfq_slice_expired(cfqd, 0);
 
 	return dispatched;
@@ -1119,10 +1111,11 @@ __cfq_dispatch_requests(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 static int
 cfq_forced_dispatch_cfqqs(struct list_head *list)
 {
-	int dispatched = 0;
 	struct cfq_queue *cfqq, *next;
 	struct cfq_rq *crq;
+	int dispatched;
 
+	dispatched = 0;
 	list_for_each_entry_safe(cfqq, next, list, cfq_list) {
 		while ((crq = cfqq->next_crq)) {
 			cfq_dispatch_insert(cfqq->cfqd->queue, crq);
@@ -1130,6 +1123,7 @@ cfq_forced_dispatch_cfqqs(struct list_head *list)
 		}
 		BUG_ON(!list_empty(&cfqq->fifo));
 	}
+
 	return dispatched;
 }
 
@@ -1156,7 +1150,8 @@ static int
 cfq_dispatch_requests(request_queue_t *q, int force)
 {
 	struct cfq_data *cfqd = q->elevator->elevator_data;
-	struct cfq_queue *cfqq;
+	struct cfq_queue *cfqq, *prev_cfqq;
+	int dispatched;
 
 	if (!cfqd->busy_queues)
 		return 0;
@@ -1164,10 +1159,17 @@ cfq_dispatch_requests(request_queue_t *q, int force)
 	if (unlikely(force))
 		return cfq_forced_dispatch(cfqd);
 
-	cfqq = cfq_select_queue(cfqd);
-	if (cfqq) {
+	dispatched = 0;
+	prev_cfqq = NULL;
+	while ((cfqq = cfq_select_queue(cfqd)) != NULL) {
 		int max_dispatch;
 
+		/*
+		 * Don't repeat dispatch from the previous queue.
+		 */
+		if (prev_cfqq == cfqq)
+			break;
+
 		cfq_clear_cfqq_must_dispatch(cfqq);
 		cfq_clear_cfqq_wait_request(cfqq);
 		del_timer(&cfqd->idle_slice_timer);
@@ -1176,10 +1178,19 @@ cfq_dispatch_requests(request_queue_t *q, int force)
 		if (cfq_class_idle(cfqq))
 			max_dispatch = 1;
 
-		return __cfq_dispatch_requests(cfqd, cfqq, max_dispatch);
+		dispatched += __cfq_dispatch_requests(cfqd, cfqq, max_dispatch);
+
+		/*
+		 * If the dispatch cfqq has idling enabled and is still
+		 * the active queue, break out.
+		 */
+		if (cfq_cfqq_idle_window(cfqq) && cfqd->active_queue)
+			break;
+
+		prev_cfqq = cfqq;
 	}
 
-	return 0;
+	return dispatched;
 }
 
 /*
@@ -1324,7 +1335,6 @@ cfq_alloc_io_context(struct cfq_data *cfqd, gfp_t gfp_mask)
 
 	if (cic) {
 		memset(cic, 0, sizeof(*cic));
-		RB_CLEAR_COLOR(&cic->rb_node);
 		cic->last_end_request = jiffies;
 		INIT_LIST_HEAD(&cic->queue_list);
 		cic->dtor = cfq_free_io_context;
@@ -1386,25 +1396,28 @@ static inline void changed_ioprio(struct cfq_io_context *cic)
 {
 	struct cfq_data *cfqd = cic->key;
 	struct cfq_queue *cfqq;
-	if (cfqd) {
-		spin_lock(cfqd->queue->queue_lock);
-		cfqq = cic->cfqq[ASYNC];
-		if (cfqq) {
-			struct cfq_queue *new_cfqq;
-			new_cfqq = cfq_get_queue(cfqd, CFQ_KEY_ASYNC,
-						cic->ioc->task, GFP_ATOMIC);
-			if (new_cfqq) {
-				cic->cfqq[ASYNC] = new_cfqq;
-				cfq_put_queue(cfqq);
-			}
-		}
-		cfqq = cic->cfqq[SYNC];
-		if (cfqq) {
-			cfq_mark_cfqq_prio_changed(cfqq);
-			cfq_init_prio_data(cfqq);
+
+	if (unlikely(!cfqd))
+		return;
+
+	spin_lock(cfqd->queue->queue_lock);
+
+	cfqq = cic->cfqq[ASYNC];
+	if (cfqq) {
+		struct cfq_queue *new_cfqq;
+		new_cfqq = cfq_get_queue(cfqd, CFQ_KEY_ASYNC, cic->ioc->task,
+					 GFP_ATOMIC);
+		if (new_cfqq) {
+			cic->cfqq[ASYNC] = new_cfqq;
+			cfq_put_queue(cfqq);
 		}
-		spin_unlock(cfqd->queue->queue_lock);
 	}
+
+	cfqq = cic->cfqq[SYNC];
+	if (cfqq)
+		cfq_mark_cfqq_prio_changed(cfqq);
+
+	spin_unlock(cfqd->queue->queue_lock);
 }
 
 /*
@@ -1461,7 +1474,6 @@ retry:
 
 		INIT_HLIST_NODE(&cfqq->cfq_hash);
 		INIT_LIST_HEAD(&cfqq->cfq_list);
-		RB_CLEAR_ROOT(&cfqq->sort_list);
 		INIT_LIST_HEAD(&cfqq->fifo);
 
 		cfqq->key = key;
@@ -1473,8 +1485,7 @@ retry:
 		 * set ->slice_left to allow preemption for a new process
 		 */
 		cfqq->slice_left = 2 * cfqd->cfq_slice_idle;
-		if (!cfqd->hw_tag)
-			cfq_mark_cfqq_idle_window(cfqq);
+		cfq_mark_cfqq_idle_window(cfqq);
 		cfq_mark_cfqq_prio_changed(cfqq);
 		cfq_init_prio_data(cfqq);
 	}
@@ -1665,7 +1676,8 @@ cfq_update_idle_window(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 {
 	int enable_idle = cfq_cfqq_idle_window(cfqq);
 
-	if (!cic->ioc->task || !cfqd->cfq_slice_idle || cfqd->hw_tag)
+	if (!cic->ioc->task || !cfqd->cfq_slice_idle ||
+	    (cfqd->hw_tag && CIC_SEEKY(cic)))
 		enable_idle = 0;
 	else if (sample_valid(cic->ttime_samples)) {
 		if (cic->ttime_mean > cfqd->cfq_slice_idle)
@@ -1695,7 +1707,7 @@ cfq_should_preempt(struct cfq_data *cfqd, struct cfq_queue *new_cfqq,
 		return 0;
 
 	if (!cfqq)
-		return 1;
+		return 0;
 
 	if (cfq_class_idle(cfqq))
 		return 1;
@@ -1727,7 +1739,7 @@ static void cfq_preempt_queue(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 		cfqq->slice_left = cfq_prio_to_slice(cfqd, cfqq) / 2;
 
 	cfqq->slice_end = cfqq->slice_left + jiffies;
-	__cfq_slice_expired(cfqd, cfqq, 1);
+	cfq_slice_expired(cfqd, 1);
 	__cfq_set_active_queue(cfqd, cfqq);
 }
 
@@ -1752,11 +1764,7 @@ static void
 cfq_crq_enqueued(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 		 struct cfq_rq *crq)
 {
-	struct cfq_io_context *cic;
-
-	cfqq->next_crq = cfq_choose_req(cfqd, cfqq->next_crq, crq);
-
-	cic = crq->io_context;
+	struct cfq_io_context *cic = crq->io_context;
 
 	/*
 	 * we never wait for an async request and we don't allow preemption
@@ -1846,11 +1854,23 @@ static void cfq_completed_request(request_queue_t *q, struct request *rq)
 			cfqq->service_last = now;
 			cfq_resort_rr_list(cfqq, 0);
 		}
-		cfq_schedule_dispatch(cfqd);
 	}
 
-	if (cfq_crq_is_sync(crq))
+	if (sync)
 		crq->io_context->last_end_request = now;
+
+	/*
+	 * If this is the active queue, check if it needs to be expired,
+	 * or if we want to idle in case it has no pending requests.
+	 */
+	if (cfqd->active_queue == cfqq) {
+		if (time_after(now, cfqq->slice_end))
+			cfq_slice_expired(cfqd, 0);
+		else if (sync && RB_EMPTY_ROOT(&cfqq->sort_list)) {
+			if (!cfq_arm_slice_timer(cfqd, cfqq))
+				cfq_schedule_dispatch(cfqd);
+		}
+	}
 }
 
 static struct request *
@@ -1917,7 +1937,6 @@ static inline int
 __cfq_may_queue(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 		struct task_struct *task, int rw)
 {
-#if 1
 	if ((cfq_cfqq_wait_request(cfqq) || cfq_cfqq_must_alloc(cfqq)) &&
 	    !cfq_cfqq_must_alloc_slice(cfqq)) {
 		cfq_mark_cfqq_must_alloc_slice(cfqq);
@@ -1925,39 +1944,6 @@ __cfq_may_queue(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 	}
 
 	return ELV_MQUEUE_MAY;
-#else
-	if (!cfqq || task->flags & PF_MEMALLOC)
-		return ELV_MQUEUE_MAY;
-	if (!cfqq->allocated[rw] || cfq_cfqq_must_alloc(cfqq)) {
-		if (cfq_cfqq_wait_request(cfqq))
-			return ELV_MQUEUE_MUST;
-
-		/*
-		 * only allow 1 ELV_MQUEUE_MUST per slice, otherwise we
-		 * can quickly flood the queue with writes from a single task
-		 */
-		if (rw == READ || !cfq_cfqq_must_alloc_slice(cfqq)) {
-			cfq_mark_cfqq_must_alloc_slice(cfqq);
-			return ELV_MQUEUE_MUST;
-		}
-
-		return ELV_MQUEUE_MAY;
-	}
-	if (cfq_class_idle(cfqq))
-		return ELV_MQUEUE_NO;
-	if (cfqq->allocated[rw] >= cfqd->max_queued) {
-		struct io_context *ioc = get_io_context(GFP_ATOMIC);
-		int ret = ELV_MQUEUE_NO;
-
-		if (ioc && ioc->nr_batch_requests)
-			ret = ELV_MQUEUE_MAY;
-
-		put_io_context(ioc);
-		return ret;
-	}
-
-	return ELV_MQUEUE_MAY;
-#endif
 }
 
 static int cfq_may_queue(request_queue_t *q, int rw, struct bio *bio)
@@ -1986,16 +1972,13 @@ static int cfq_may_queue(request_queue_t *q, int rw, struct bio *bio)
 static void cfq_check_waiters(request_queue_t *q, struct cfq_queue *cfqq)
 {
 	struct cfq_data *cfqd = q->elevator->elevator_data;
-	struct request_list *rl = &q->rq;
 
-	if (cfqq->allocated[READ] <= cfqd->max_queued || cfqd->rq_starved) {
+	if (unlikely(cfqd->rq_starved)) {
+		struct request_list *rl = &q->rq;
+
 		smp_mb();
 		if (waitqueue_active(&rl->wait[READ]))
 			wake_up(&rl->wait[READ]);
-	}
-
-	if (cfqq->allocated[WRITE] <= cfqd->max_queued || cfqd->rq_starved) {
-		smp_mb();
 		if (waitqueue_active(&rl->wait[WRITE]))
 			wake_up(&rl->wait[WRITE]);
 	}
@@ -2069,7 +2052,7 @@ cfq_set_request(request_queue_t *q, struct request *rq, struct bio *bio,
 
 	crq = mempool_alloc(cfqd->crq_pool, gfp_mask);
 	if (crq) {
-		RB_CLEAR(&crq->rb_node);
+		RB_CLEAR_NODE(&crq->rb_node);
 		crq->rb_key = 0;
 		crq->request = rq;
 		INIT_HLIST_NODE(&crq->hash);
@@ -2155,16 +2138,13 @@ static void cfq_idle_slice_timer(unsigned long data)
 		 * only expire and reinvoke request handler, if there are
 		 * other queues with pending requests
 		 */
-		if (!cfqd->busy_queues) {
-			cfqd->idle_slice_timer.expires = min(now + cfqd->cfq_slice_idle, cfqq->slice_end);
-			add_timer(&cfqd->idle_slice_timer);
+		if (!cfqd->busy_queues)
 			goto out_cont;
-		}
 
 		/*
 		 * not expired and it has a request pending, let it dispatch
 		 */
-		if (!RB_EMPTY(&cfqq->sort_list)) {
+		if (!RB_EMPTY_ROOT(&cfqq->sort_list)) {
 			cfq_mark_cfqq_must_dispatch(cfqq);
 			goto out_kick;
 		}
@@ -2285,9 +2265,6 @@ static void *cfq_init_queue(request_queue_t *q, elevator_t *e)
 
 	cfqd->queue = q;
 
-	cfqd->max_queued = q->nr_requests / 4;
-	q->nr_batching = cfq_queued;
-
 	init_timer(&cfqd->idle_slice_timer);
 	cfqd->idle_slice_timer.function = cfq_idle_slice_timer;
 	cfqd->idle_slice_timer.data = (unsigned long) cfqd;
diff --git a/block/deadline-iosched.c b/block/deadline-iosched.c
index 3bd0415a982..c7ca9f0b649 100644
--- a/block/deadline-iosched.c
+++ b/block/deadline-iosched.c
@@ -8,7 +8,6 @@
 #include <linux/blkdev.h>
 #include <linux/elevator.h>
 #include <linux/bio.h>
-#include <linux/config.h>
 #include <linux/module.h>
 #include <linux/slab.h>
 #include <linux/init.h>
@@ -30,8 +29,7 @@ static const int deadline_hash_shift = 5;
 #define DL_HASH_FN(sec)		(hash_long(DL_HASH_BLOCK((sec)), deadline_hash_shift))
 #define DL_HASH_ENTRIES		(1 << deadline_hash_shift)
 #define rq_hash_key(rq)		((rq)->sector + (rq)->nr_sectors)
-#define list_entry_hash(ptr)	list_entry((ptr), struct deadline_rq, hash)
-#define ON_HASH(drq)		(drq)->on_hash
+#define ON_HASH(drq)		(!hlist_unhashed(&(drq)->hash))
 
 struct deadline_data {
 	/*
@@ -48,7 +46,7 @@ struct deadline_data {
 	 * next in sort order. read, write or both are NULL
 	 */
 	struct deadline_rq *next_drq[2];
-	struct list_head *hash;		/* request hash */
+	struct hlist_head *hash;	/* request hash */
 	unsigned int batching;		/* number of sequential requests made */
 	sector_t last_sector;		/* head position */
 	unsigned int starved;		/* times reads have starved writes */
@@ -79,8 +77,7 @@ struct deadline_rq {
 	/*
 	 * request hash, key is the ending offset (for back merge lookup)
 	 */
-	struct list_head hash;
-	char on_hash;
+	struct hlist_node hash;
 
 	/*
 	 * expire fifo
@@ -100,8 +97,7 @@ static kmem_cache_t *drq_pool;
  */
 static inline void __deadline_del_drq_hash(struct deadline_rq *drq)
 {
-	drq->on_hash = 0;
-	list_del_init(&drq->hash);
+	hlist_del_init(&drq->hash);
 }
 
 static inline void deadline_del_drq_hash(struct deadline_rq *drq)
@@ -117,8 +113,7 @@ deadline_add_drq_hash(struct deadline_data *dd, struct deadline_rq *drq)
 
 	BUG_ON(ON_HASH(drq));
 
-	drq->on_hash = 1;
-	list_add(&drq->hash, &dd->hash[DL_HASH_FN(rq_hash_key(rq))]);
+	hlist_add_head(&drq->hash, &dd->hash[DL_HASH_FN(rq_hash_key(rq))]);
 }
 
 /*
@@ -128,26 +123,24 @@ static inline void
 deadline_hot_drq_hash(struct deadline_data *dd, struct deadline_rq *drq)
 {
 	struct request *rq = drq->request;
-	struct list_head *head = &dd->hash[DL_HASH_FN(rq_hash_key(rq))];
+	struct hlist_head *head = &dd->hash[DL_HASH_FN(rq_hash_key(rq))];
 
-	if (ON_HASH(drq) && drq->hash.prev != head) {
-		list_del(&drq->hash);
-		list_add(&drq->hash, head);
+	if (ON_HASH(drq) && &drq->hash != head->first) {
+		hlist_del(&drq->hash);
+		hlist_add_head(&drq->hash, head);
 	}
 }
 
 static struct request *
 deadline_find_drq_hash(struct deadline_data *dd, sector_t offset)
 {
-	struct list_head *hash_list = &dd->hash[DL_HASH_FN(offset)];
-	struct list_head *entry, *next = hash_list->next;
+	struct hlist_head *hash_list = &dd->hash[DL_HASH_FN(offset)];
+	struct hlist_node *entry, *next;
+	struct deadline_rq *drq;
 
-	while ((entry = next) != hash_list) {
-		struct deadline_rq *drq = list_entry_hash(entry);
+	hlist_for_each_entry_safe(drq, entry, next, hash_list, hash) {
 		struct request *__rq = drq->request;
 
-		next = entry->next;
-		
 		BUG_ON(!ON_HASH(drq));
 
 		if (!rq_mergeable(__rq)) {
@@ -165,10 +158,6 @@ deadline_find_drq_hash(struct deadline_data *dd, sector_t offset)
 /*
  * rb tree support functions
  */
-#define RB_NONE		(2)
-#define RB_EMPTY(root)	((root)->rb_node == NULL)
-#define ON_RB(node)	((node)->rb_color != RB_NONE)
-#define RB_CLEAR(node)	((node)->rb_color = RB_NONE)
 #define rb_entry_drq(node)	rb_entry((node), struct deadline_rq, rb_node)
 #define DRQ_RB_ROOT(dd, drq)	(&(dd)->sort_list[rq_data_dir((drq)->request)])
 #define rq_rb_key(rq)		(rq)->sector
@@ -227,9 +216,9 @@ deadline_del_drq_rb(struct deadline_data *dd, struct deadline_rq *drq)
 			dd->next_drq[data_dir] = rb_entry_drq(rbnext);
 	}
 
-	BUG_ON(!ON_RB(&drq->rb_node));
+	BUG_ON(!RB_EMPTY_NODE(&drq->rb_node));
 	rb_erase(&drq->rb_node, DRQ_RB_ROOT(dd, drq));
-	RB_CLEAR(&drq->rb_node);
+	RB_CLEAR_NODE(&drq->rb_node);
 }
 
 static struct request *
@@ -503,7 +492,7 @@ static int deadline_dispatch_requests(request_queue_t *q, int force)
 	 */
 
 	if (reads) {
-		BUG_ON(RB_EMPTY(&dd->sort_list[READ]));
+		BUG_ON(RB_EMPTY_ROOT(&dd->sort_list[READ]));
 
 		if (writes && (dd->starved++ >= dd->writes_starved))
 			goto dispatch_writes;
@@ -519,7 +508,7 @@ static int deadline_dispatch_requests(request_queue_t *q, int force)
 
 	if (writes) {
 dispatch_writes:
-		BUG_ON(RB_EMPTY(&dd->sort_list[WRITE]));
+		BUG_ON(RB_EMPTY_ROOT(&dd->sort_list[WRITE]));
 
 		dd->starved = 0;
 
@@ -626,7 +615,7 @@ static void *deadline_init_queue(request_queue_t *q, elevator_t *e)
 		return NULL;
 	memset(dd, 0, sizeof(*dd));
 
-	dd->hash = kmalloc_node(sizeof(struct list_head)*DL_HASH_ENTRIES,
+	dd->hash = kmalloc_node(sizeof(struct hlist_head)*DL_HASH_ENTRIES,
 				GFP_KERNEL, q->node);
 	if (!dd->hash) {
 		kfree(dd);
@@ -642,7 +631,7 @@ static void *deadline_init_queue(request_queue_t *q, elevator_t *e)
 	}
 
 	for (i = 0; i < DL_HASH_ENTRIES; i++)
-		INIT_LIST_HEAD(&dd->hash[i]);
+		INIT_HLIST_HEAD(&dd->hash[i]);
 
 	INIT_LIST_HEAD(&dd->fifo_list[READ]);
 	INIT_LIST_HEAD(&dd->fifo_list[WRITE]);
@@ -675,11 +664,10 @@ deadline_set_request(request_queue_t *q, struct request *rq, struct bio *bio,
 	drq = mempool_alloc(dd->drq_pool, gfp_mask);
 	if (drq) {
 		memset(drq, 0, sizeof(*drq));
-		RB_CLEAR(&drq->rb_node);
+		RB_CLEAR_NODE(&drq->rb_node);
 		drq->request = rq;
 
-		INIT_LIST_HEAD(&drq->hash);
-		drq->on_hash = 0;
+		INIT_HLIST_NODE(&drq->hash);
 
 		INIT_LIST_HEAD(&drq->fifo);
 
diff --git a/block/elevator.c b/block/elevator.c
index a0afdd317ce..bc7baeec0d1 100644
--- a/block/elevator.c
+++ b/block/elevator.c
@@ -27,7 +27,6 @@
 #include <linux/blkdev.h>
 #include <linux/elevator.h>
 #include <linux/bio.h>
-#include <linux/config.h>
 #include <linux/module.h>
 #include <linux/slab.h>
 #include <linux/init.h>
@@ -850,12 +849,9 @@ fail_register:
 	 * one again (along with re-adding the sysfs dir)
 	 */
 	elevator_exit(e);
-	e = NULL;
 	q->elevator = old_elevator;
 	elv_register_queue(q);
 	clear_bit(QUEUE_FLAG_ELVSWITCH, &q->queue_flags);
-	if (e)
-		kobject_put(&e->kobj);
 	return 0;
 }
 
diff --git a/block/genhd.c b/block/genhd.c
index 5a8d3bf02f1..25d1f42568c 100644
--- a/block/genhd.c
+++ b/block/genhd.c
@@ -2,7 +2,6 @@
  *  gendisk handling
  */
 
-#include <linux/config.h>
 #include <linux/module.h>
 #include <linux/fs.h>
 #include <linux/genhd.h>
@@ -17,8 +16,7 @@
 #include <linux/buffer_head.h>
 #include <linux/mutex.h>
 
-static struct subsystem block_subsys;
-
+struct subsystem block_subsys;
 static DEFINE_MUTEX(block_subsys_lock);
 
 /*
@@ -511,9 +509,7 @@ static struct kset_uevent_ops block_uevent_ops = {
 	.uevent		= block_uevent,
 };
 
-/* declare block_subsys. */
-static decl_subsys(block, &ktype_block, &block_uevent_ops);
-
+decl_subsys(block, &ktype_block, &block_uevent_ops);
 
 /*
  * aggregate disk stat collector.  Uses the same stats that the sysfs
diff --git a/block/ll_rw_blk.c b/block/ll_rw_blk.c
index 7eb36c53f4b..5813d63c20a 100644
--- a/block/ll_rw_blk.c
+++ b/block/ll_rw_blk.c
@@ -10,7 +10,6 @@
 /*
  * This handles all read/write requests to block devices
  */
-#include <linux/config.h>
 #include <linux/kernel.h>
 #include <linux/module.h>
 #include <linux/backing-dev.h>
@@ -638,7 +637,7 @@ void blk_queue_bounce_limit(request_queue_t *q, u64 dma_addr)
 	/* Assume anything <= 4GB can be handled by IOMMU.
 	   Actually some IOMMUs can handle everything, but I don't
 	   know of a way to test this here. */
-	if (bounce_pfn < (0xffffffff>>PAGE_SHIFT))
+	if (bounce_pfn < (min_t(u64,0xffffffff,BLK_BOUNCE_HIGH) >> PAGE_SHIFT))
 		dma = 1;
 	q->bounce_pfn = max_low_pfn;
 #else
@@ -1663,6 +1662,8 @@ static void blk_unplug_timeout(unsigned long data)
  **/
 void blk_start_queue(request_queue_t *q)
 {
+	WARN_ON(!irqs_disabled());
+
 	clear_bit(QUEUE_FLAG_STOPPED, &q->queue_flags);
 
 	/*
@@ -1878,7 +1879,8 @@ EXPORT_SYMBOL(blk_alloc_queue_node);
  *    get dealt with eventually.
  *
  *    The queue spin lock must be held while manipulating the requests on the
- *    request queue.
+ *    request queue; this lock will be taken also from interrupt context, so irq
+ *    disabling is needed for it.
  *
  *    Function returns a pointer to the initialized request queue, or NULL if
  *    it didn't succeed.
@@ -2742,7 +2744,7 @@ static int attempt_merge(request_queue_t *q, struct request *req,
 		return 0;
 
 	/*
-	 * not contigious
+	 * not contiguous
 	 */
 	if (req->sector + req->nr_sectors != next->sector)
 		return 0;
@@ -2824,6 +2826,9 @@ static void init_request_from_bio(struct request *req, struct bio *bio)
 	if (unlikely(bio_barrier(bio)))
 		req->flags |= (REQ_HARDBARRIER | REQ_NOMERGE);
 
+	if (bio_sync(bio))
+		req->flags |= REQ_RW_SYNC;
+
 	req->errors = 0;
 	req->hard_sector = req->sector = bio->bi_sector;
 	req->hard_nr_sectors = req->nr_sectors = bio_sectors(bio);
@@ -3111,9 +3116,9 @@ void submit_bio(int rw, struct bio *bio)
 	BIO_BUG_ON(!bio->bi_io_vec);
 	bio->bi_rw |= rw;
 	if (rw & WRITE)
-		mod_page_state(pgpgout, count);
+		count_vm_events(PGPGOUT, count);
 	else
-		mod_page_state(pgpgin, count);
+		count_vm_events(PGPGIN, count);
 
 	if (unlikely(block_dump)) {
 		char b[BDEVNAME_SIZE];
@@ -3359,12 +3364,11 @@ EXPORT_SYMBOL(end_that_request_chunk);
  */
 static void blk_done_softirq(struct softirq_action *h)
 {
-	struct list_head *cpu_list;
-	LIST_HEAD(local_list);
+	struct list_head *cpu_list, local_list;
 
 	local_irq_disable();
 	cpu_list = &__get_cpu_var(blk_cpu_done);
-	list_splice_init(cpu_list, &local_list);
+	list_replace_init(cpu_list, &local_list);
 	local_irq_enable();
 
 	while (!list_empty(&local_list)) {
@@ -3398,7 +3402,7 @@ static int blk_cpu_notify(struct notifier_block *self, unsigned long action,
 }
 
 
-static struct notifier_block blk_cpu_notifier = {
+static struct notifier_block __devinitdata blk_cpu_notifier = {
 	.notifier_call	= blk_cpu_notify,
 };
 
@@ -3410,7 +3414,7 @@ static struct notifier_block blk_cpu_notifier = {
  *
  * Description:
  *     Ends all I/O on a request. It does not handle partial completions,
- *     unless the driver actually implements this in its completionc callback
+ *     unless the driver actually implements this in its completion callback
  *     through requeueing. Theh actual completion happens out-of-order,
  *     through a softirq handler. The user must have registered a completion
  *     callback through blk_queue_softirq_done().
@@ -3536,9 +3540,7 @@ int __init blk_dev_init(void)
 		INIT_LIST_HEAD(&per_cpu(blk_cpu_done, i));
 
 	open_softirq(BLOCK_SOFTIRQ, blk_done_softirq, NULL);
-#ifdef CONFIG_HOTPLUG_CPU
-	register_cpu_notifier(&blk_cpu_notifier);
-#endif
+	register_hotcpu_notifier(&blk_cpu_notifier);
 
 	blk_max_low_pfn = max_low_pfn;
 	blk_max_pfn = max_pfn;