Merge branch 'for-3.1/core' of git://git.kernel.dk/linux-block

* 'for-3.1/core' of git://git.kernel.dk/linux-block: (24 commits)
  block: strict rq_affinity
  backing-dev: use synchronize_rcu_expedited instead of synchronize_rcu
  block: fix patch import error in max_discard_sectors check
  block: reorder request_queue to remove 64 bit alignment padding
  CFQ: add think time check for group
  CFQ: add think time check for service tree
  CFQ: move think time check variables to a separate struct
  fixlet: Remove fs_excl from struct task.
  cfq: Remove special treatment for metadata rqs.
  block: document blk_plug list access
  block: avoid building too big plug list
  compat_ioctl: fix make headers_check regression
  block: eliminate potential for infinite loop in blkdev_issue_discard
  compat_ioctl: fix warning caused by qemu
  block: flush MEDIA_CHANGE from drivers on close(2)
  blk-throttle: Make total_nr_queued unsigned
  block: Add __attribute__((format(printf...) and fix fallout
  fs/partitions/check.c: make local symbols static
  block:remove some spare spaces in genhd.c
  block:fix the comment error in blkdev.h
  ...

Documentation/block/queue-sysfs.txt

@@ -45,9 +45,13 @@ device.
 
 rq_affinity (RW)
 ----------------
-If this option is enabled, the block layer will migrate request completions
-to the CPU that originally submitted the request. For some workloads
-this provides a significant reduction in CPU cycles due to caching effects.
+If this option is '1', the block layer will migrate request completions to the
+cpu "group" that originally submitted the request. For some workloads this
+provides a significant reduction in CPU cycles due to caching effects.
+
+For storage configurations that need to maximize distribution of completion
+processing setting this option to '2' forces the completion to run on the
+requesting cpu (bypassing the "group" aggregation logic).
 
 scheduler (RW)
 --------------

block/blk-core.c

@@ -1282,10 +1282,8 @@ static int __make_request(struct request_queue *q, struct bio *bio)
 	init_request_from_bio(req, bio);
 
 	if (test_bit(QUEUE_FLAG_SAME_COMP, &q->queue_flags) ||
-	    bio_flagged(bio, BIO_CPU_AFFINE)) {
-		req->cpu = blk_cpu_to_group(get_cpu());
-		put_cpu();
-	}
+	    bio_flagged(bio, BIO_CPU_AFFINE))
+		req->cpu = smp_processor_id();
 
 	plug = current->plug;
 	if (plug) {
@@ -1305,7 +1303,10 @@ static int __make_request(struct request_queue *q, struct bio *bio)
 				plug->should_sort = 1;
 		}
 		list_add_tail(&req->queuelist, &plug->list);
+		plug->count++;
 		drive_stat_acct(req, 1);
+		if (plug->count >= BLK_MAX_REQUEST_COUNT)
+			blk_flush_plug_list(plug, false);
 	} else {
 		spin_lock_irq(q->queue_lock);
 		add_acct_request(q, req, where);
@@ -2629,6 +2630,7 @@ void blk_start_plug(struct blk_plug *plug)
 	INIT_LIST_HEAD(&plug->list);
 	INIT_LIST_HEAD(&plug->cb_list);
 	plug->should_sort = 0;
+	plug->count = 0;
 
 	/*
 	 * If this is a nested plug, don't actually assign it. It will be
@@ -2712,6 +2714,7 @@ void blk_flush_plug_list(struct blk_plug *plug, bool from_schedule)
 		return;
 
 	list_splice_init(&plug->list, &list);
+	plug->count = 0;
 
 	if (plug->should_sort) {
 		list_sort(NULL, &list, plug_rq_cmp);

block/blk-ioc.c

@@ -82,26 +82,26 @@ void exit_io_context(struct task_struct *task)
 
 struct io_context *alloc_io_context(gfp_t gfp_flags, int node)
 {
-	struct io_context *ret;
+	struct io_context *ioc;
 
-	ret = kmem_cache_alloc_node(iocontext_cachep, gfp_flags, node);
-	if (ret) {
-		atomic_long_set(&ret->refcount, 1);
-		atomic_set(&ret->nr_tasks, 1);
-		spin_lock_init(&ret->lock);
-		ret->ioprio_changed = 0;
-		ret->ioprio = 0;
-		ret->last_waited = 0; /* doesn't matter... */
-		ret->nr_batch_requests = 0; /* because this is 0 */
-		INIT_RADIX_TREE(&ret->radix_root, GFP_ATOMIC | __GFP_HIGH);
-		INIT_HLIST_HEAD(&ret->cic_list);
-		ret->ioc_data = NULL;
+	ioc = kmem_cache_alloc_node(iocontext_cachep, gfp_flags, node);
+	if (ioc) {
+		atomic_long_set(&ioc->refcount, 1);
+		atomic_set(&ioc->nr_tasks, 1);
+		spin_lock_init(&ioc->lock);
+		ioc->ioprio_changed = 0;
+		ioc->ioprio = 0;
+		ioc->last_waited = 0; /* doesn't matter... */
+		ioc->nr_batch_requests = 0; /* because this is 0 */
+		INIT_RADIX_TREE(&ioc->radix_root, GFP_ATOMIC | __GFP_HIGH);
+		INIT_HLIST_HEAD(&ioc->cic_list);
+		ioc->ioc_data = NULL;
 #if defined(CONFIG_BLK_CGROUP) || defined(CONFIG_BLK_CGROUP_MODULE)
-		ret->cgroup_changed = 0;
+		ioc->cgroup_changed = 0;
 #endif
 	}
 
-	return ret;
+	return ioc;
 }
 
 /*
@@ -139,19 +139,19 @@ struct io_context *current_io_context(gfp_t gfp_flags, int node)
  */
 struct io_context *get_io_context(gfp_t gfp_flags, int node)
 {
-	struct io_context *ret = NULL;
+	struct io_context *ioc = NULL;
 
 	/*
 	 * Check for unlikely race with exiting task. ioc ref count is
 	 * zero when ioc is being detached.
 	 */
 	do {
-		ret = current_io_context(gfp_flags, node);
-		if (unlikely(!ret))
+		ioc = current_io_context(gfp_flags, node);
+		if (unlikely(!ioc))
 			break;
-	} while (!atomic_long_inc_not_zero(&ret->refcount));
+	} while (!atomic_long_inc_not_zero(&ioc->refcount));
 
-	return ret;
+	return ioc;
 }
 EXPORT_SYMBOL(get_io_context);
 

block/blk-lib.c

@@ -59,7 +59,10 @@ int blkdev_issue_discard(struct block_device *bdev, sector_t sector,
 	 * granularity
 	 */
 	max_discard_sectors = min(q->limits.max_discard_sectors, UINT_MAX >> 9);
-	if (q->limits.discard_granularity) {
+	if (unlikely(!max_discard_sectors)) {
+		/* Avoid infinite loop below. Being cautious never hurts. */
+		return -EOPNOTSUPP;
+	} else if (q->limits.discard_granularity) {
 		unsigned int disc_sects = q->limits.discard_granularity >> 9;
 
 		max_discard_sectors &= ~(disc_sects - 1);

block/blk-softirq.c

@@ -103,22 +103,25 @@ static struct notifier_block __cpuinitdata blk_cpu_notifier = {
 
 void __blk_complete_request(struct request *req)
 {
+	int ccpu, cpu, group_cpu = NR_CPUS;
 	struct request_queue *q = req->q;
 	unsigned long flags;
-	int ccpu, cpu, group_cpu;
 
 	BUG_ON(!q->softirq_done_fn);
 
 	local_irq_save(flags);
 	cpu = smp_processor_id();
-	group_cpu = blk_cpu_to_group(cpu);
 
 	/*
 	 * Select completion CPU
 	 */
-	if (test_bit(QUEUE_FLAG_SAME_COMP, &q->queue_flags) && req->cpu != -1)
+	if (test_bit(QUEUE_FLAG_SAME_COMP, &q->queue_flags) && req->cpu != -1) {
 		ccpu = req->cpu;
-	else
+		if (!test_bit(QUEUE_FLAG_SAME_FORCE, &q->queue_flags)) {
+			ccpu = blk_cpu_to_group(ccpu);
+			group_cpu = blk_cpu_to_group(cpu);
+		}
+	} else
 		ccpu = cpu;
 
 	if (ccpu == cpu || ccpu == group_cpu) {

block/blk-sysfs.c

@@ -244,8 +244,9 @@ static ssize_t queue_nomerges_store(struct request_queue *q, const char *page,
 static ssize_t queue_rq_affinity_show(struct request_queue *q, char *page)
 {
 	bool set = test_bit(QUEUE_FLAG_SAME_COMP, &q->queue_flags);
+	bool force = test_bit(QUEUE_FLAG_SAME_FORCE, &q->queue_flags);
 
-	return queue_var_show(set, page);
+	return queue_var_show(set << force, page);
 }
 
 static ssize_t
@@ -257,10 +258,14 @@ queue_rq_affinity_store(struct request_queue *q, const char *page, size_t count)
 
 	ret = queue_var_store(&val, page, count);
 	spin_lock_irq(q->queue_lock);
-	if (val)
+	if (val) {
 		queue_flag_set(QUEUE_FLAG_SAME_COMP, q);
-	else
-		queue_flag_clear(QUEUE_FLAG_SAME_COMP,  q);
+		if (val == 2)
+			queue_flag_set(QUEUE_FLAG_SAME_FORCE, q);
+	} else {
+		queue_flag_clear(QUEUE_FLAG_SAME_COMP, q);
+		queue_flag_clear(QUEUE_FLAG_SAME_FORCE, q);
+	}
 	spin_unlock_irq(q->queue_lock);
 #endif
 	return ret;

block/blk-throttle.c

@@ -142,9 +142,9 @@ static inline struct throtl_grp *tg_of_blkg(struct blkio_group *blkg)
 	return NULL;
 }
 
-static inline int total_nr_queued(struct throtl_data *td)
+static inline unsigned int total_nr_queued(struct throtl_data *td)
 {
-	return (td->nr_queued[0] + td->nr_queued[1]);
+	return td->nr_queued[0] + td->nr_queued[1];
 }
 
 static inline struct throtl_grp *throtl_ref_get_tg(struct throtl_grp *tg)
@@ -927,7 +927,7 @@ static int throtl_dispatch(struct request_queue *q)
 
 	bio_list_init(&bio_list_on_stack);
 
-	throtl_log(td, "dispatch nr_queued=%d read=%u write=%u",
+	throtl_log(td, "dispatch nr_queued=%u read=%u write=%u",
 			total_nr_queued(td), td->nr_queued[READ],
 			td->nr_queued[WRITE]);
 
@@ -970,7 +970,7 @@ throtl_schedule_delayed_work(struct throtl_data *td, unsigned long delay)
 	struct delayed_work *dwork = &td->throtl_work;
 
 	/* schedule work if limits changed even if no bio is queued */
-	if (total_nr_queued(td) > 0 || td->limits_changed) {
+	if (total_nr_queued(td) || td->limits_changed) {
 		/*
 		 * We might have a work scheduled to be executed in future.
 		 * Cancel that and schedule a new one.

block/cfq-iosched.c

@@ -87,9 +87,10 @@ struct cfq_rb_root {
 	unsigned count;
 	unsigned total_weight;
 	u64 min_vdisktime;
+	struct cfq_ttime ttime;
 };
-#define CFQ_RB_ROOT	(struct cfq_rb_root) { .rb = RB_ROOT, .left = NULL, \
-			.count = 0, .min_vdisktime = 0, }
+#define CFQ_RB_ROOT	(struct cfq_rb_root) { .rb = RB_ROOT, \
+			.ttime = {.last_end_request = jiffies,},}
 
 /*
  * Per process-grouping structure
@@ -129,14 +130,12 @@ struct cfq_queue {
 	unsigned long slice_end;
 	long slice_resid;
 
-	/* 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;
+	unsigned short ioprio_class;
 
 	pid_t pid;
 
@@ -212,6 +211,7 @@ struct cfq_group {
 #endif
 	/* number of requests that are on the dispatch list or inside driver */
 	int dispatched;
+	struct cfq_ttime ttime;
 };
 
 /*
@@ -393,6 +393,18 @@ CFQ_CFQQ_FNS(wait_busy);
 			j++, st = i < IDLE_WORKLOAD ? \
 			&cfqg->service_trees[i][j]: NULL) \
 
+static inline bool cfq_io_thinktime_big(struct cfq_data *cfqd,
+	struct cfq_ttime *ttime, bool group_idle)
+{
+	unsigned long slice;
+	if (!sample_valid(ttime->ttime_samples))
+		return false;
+	if (group_idle)
+		slice = cfqd->cfq_group_idle;
+	else
+		slice = cfqd->cfq_slice_idle;
+	return ttime->ttime_mean > slice;
+}
 
 static inline bool iops_mode(struct cfq_data *cfqd)
 {
@@ -670,9 +682,6 @@ cfq_choose_req(struct cfq_data *cfqd, struct request *rq1, struct request *rq2,
 	if (rq_is_sync(rq1) != rq_is_sync(rq2))
 		return rq_is_sync(rq1) ? rq1 : rq2;
 
-	if ((rq1->cmd_flags ^ rq2->cmd_flags) & REQ_META)
-		return rq1->cmd_flags & REQ_META ? rq1 : rq2;
-
 	s1 = blk_rq_pos(rq1);
 	s2 = blk_rq_pos(rq2);
 
@@ -1005,8 +1014,8 @@ static inline struct cfq_group *cfqg_of_blkg(struct blkio_group *blkg)
 	return NULL;
 }
 
-void cfq_update_blkio_group_weight(void *key, struct blkio_group *blkg,
-					unsigned int weight)
+static void cfq_update_blkio_group_weight(void *key, struct blkio_group *blkg,
+					  unsigned int weight)
 {
 	struct cfq_group *cfqg = cfqg_of_blkg(blkg);
 	cfqg->new_weight = weight;
@@ -1059,6 +1068,8 @@ static struct cfq_group * cfq_alloc_cfqg(struct cfq_data *cfqd)
 		*st = CFQ_RB_ROOT;
 	RB_CLEAR_NODE(&cfqg->rb_node);
 
+	cfqg->ttime.last_end_request = jiffies;
+
 	/*
 	 * Take the initial reference that will be released on destroy
 	 * This can be thought of a joint reference by cgroup and
@@ -1235,7 +1246,7 @@ static void cfq_release_cfq_groups(struct cfq_data *cfqd)
  * it should not be NULL as even if elevator was exiting, cgroup deltion
  * path got to it first.
  */
-void cfq_unlink_blkio_group(void *key, struct blkio_group *blkg)
+static void cfq_unlink_blkio_group(void *key, struct blkio_group *blkg)
 {
 	unsigned long  flags;
 	struct cfq_data *cfqd = key;
@@ -1502,16 +1513,11 @@ 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, *prev;
+	struct request *prev;
 
 	cfqq->queued[rq_is_sync(rq)]++;
 
-	/*
-	 * looks a little odd, but the first insert might return an alias.
-	 * if that happens, put the alias on the dispatch list
-	 */
-	while ((__alias = elv_rb_add(&cfqq->sort_list, rq)) != NULL)
-		cfq_dispatch_insert(cfqd->queue, __alias);
+	elv_rb_add(&cfqq->sort_list, rq);
 
 	if (!cfq_cfqq_on_rr(cfqq))
 		cfq_add_cfqq_rr(cfqd, cfqq);
@@ -1598,10 +1604,6 @@ static void cfq_remove_request(struct request *rq)
 	cfqq->cfqd->rq_queued--;
 	cfq_blkiocg_update_io_remove_stats(&(RQ_CFQG(rq))->blkg,
 					rq_data_dir(rq), rq_is_sync(rq));
-	if (rq->cmd_flags & REQ_META) {
-		WARN_ON(!cfqq->meta_pending);
-		cfqq->meta_pending--;
-	}
 }
 
 static int cfq_merge(struct request_queue *q, struct request **req,
@@ -1969,7 +1971,8 @@ static bool cfq_should_idle(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 	 * Otherwise, we do only if they are the last ones
 	 * in their service tree.
 	 */
-	if (service_tree->count == 1 && cfq_cfqq_sync(cfqq))
+	if (service_tree->count == 1 && cfq_cfqq_sync(cfqq) &&
+	   !cfq_io_thinktime_big(cfqd, &service_tree->ttime, false))
 		return true;
 	cfq_log_cfqq(cfqd, cfqq, "Not idling. st->count:%d",
 			service_tree->count);
@@ -2022,10 +2025,10 @@ static void cfq_arm_slice_timer(struct cfq_data *cfqd)
 	 * slice, then don't idle. This avoids overrunning the allotted
 	 * time slice.
 	 */
-	if (sample_valid(cic->ttime_samples) &&
-	    (cfqq->slice_end - jiffies < cic->ttime_mean)) {
+	if (sample_valid(cic->ttime.ttime_samples) &&
+	    (cfqq->slice_end - jiffies < cic->ttime.ttime_mean)) {
 		cfq_log_cfqq(cfqd, cfqq, "Not idling. think_time:%lu",
-			     cic->ttime_mean);
+			     cic->ttime.ttime_mean);
 		return;
 	}
 
@@ -2381,8 +2384,9 @@ static struct cfq_queue *cfq_select_queue(struct cfq_data *cfqd)
 	 * this group, wait for requests to complete.
 	 */
 check_group_idle:
-	if (cfqd->cfq_group_idle && cfqq->cfqg->nr_cfqq == 1
-	    && cfqq->cfqg->dispatched) {
+	if (cfqd->cfq_group_idle && cfqq->cfqg->nr_cfqq == 1 &&
+	    cfqq->cfqg->dispatched &&
+	    !cfq_io_thinktime_big(cfqd, &cfqq->cfqg->ttime, true)) {
 		cfqq = NULL;
 		goto keep_queue;
 	}
@@ -2833,7 +2837,7 @@ cfq_alloc_io_context(struct cfq_data *cfqd, gfp_t gfp_mask)
 	cic = kmem_cache_alloc_node(cfq_ioc_pool, gfp_mask | __GFP_ZERO,
 							cfqd->queue->node);
 	if (cic) {
-		cic->last_end_request = jiffies;
+		cic->ttime.last_end_request = jiffies;
 		INIT_LIST_HEAD(&cic->queue_list);
 		INIT_HLIST_NODE(&cic->cic_list);
 		cic->dtor = cfq_free_io_context;
@@ -2883,7 +2887,6 @@ static void cfq_init_prio_data(struct cfq_queue *cfqq, struct io_context *ioc)
 	 * elevate the priority of this queue
 	 */
 	cfqq->org_ioprio = cfqq->ioprio;
-	cfqq->org_ioprio_class = cfqq->ioprio_class;
 	cfq_clear_cfqq_prio_changed(cfqq);
 }
 
@@ -3221,14 +3224,28 @@ cfq_get_io_context(struct cfq_data *cfqd, gfp_t gfp_mask)
 }
 
 static void
-cfq_update_io_thinktime(struct cfq_data *cfqd, struct cfq_io_context *cic)
+__cfq_update_io_thinktime(struct cfq_ttime *ttime, unsigned long slice_idle)
 {
-	unsigned long elapsed = jiffies - cic->last_end_request;
-	unsigned long ttime = min(elapsed, 2UL * cfqd->cfq_slice_idle);
+	unsigned long elapsed = jiffies - ttime->last_end_request;
+	elapsed = min(elapsed, 2UL * slice_idle);
 
-	cic->ttime_samples = (7*cic->ttime_samples + 256) / 8;
-	cic->ttime_total = (7*cic->ttime_total + 256*ttime) / 8;
-	cic->ttime_mean = (cic->ttime_total + 128) / cic->ttime_samples;
+	ttime->ttime_samples = (7*ttime->ttime_samples + 256) / 8;
+	ttime->ttime_total = (7*ttime->ttime_total + 256*elapsed) / 8;
+	ttime->ttime_mean = (ttime->ttime_total + 128) / ttime->ttime_samples;
+}
+
+static void
+cfq_update_io_thinktime(struct cfq_data *cfqd, struct cfq_queue *cfqq,
+	struct cfq_io_context *cic)
+{
+	if (cfq_cfqq_sync(cfqq)) {
+		__cfq_update_io_thinktime(&cic->ttime, cfqd->cfq_slice_idle);
+		__cfq_update_io_thinktime(&cfqq->service_tree->ttime,
+			cfqd->cfq_slice_idle);
+	}
+#ifdef CONFIG_CFQ_GROUP_IOSCHED
+	__cfq_update_io_thinktime(&cfqq->cfqg->ttime, cfqd->cfq_group_idle);
+#endif
 }
 
 static void
@@ -3277,8 +3294,8 @@ cfq_update_idle_window(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 	else if (!atomic_read(&cic->ioc->nr_tasks) || !cfqd->cfq_slice_idle ||
 	    (!cfq_cfqq_deep(cfqq) && CFQQ_SEEKY(cfqq)))
 		enable_idle = 0;
-	else if (sample_valid(cic->ttime_samples)) {
-		if (cic->ttime_mean > cfqd->cfq_slice_idle)
+	else if (sample_valid(cic->ttime.ttime_samples)) {
+		if (cic->ttime.ttime_mean > cfqd->cfq_slice_idle)
 			enable_idle = 0;
 		else
 			enable_idle = 1;
@@ -3339,13 +3356,6 @@ cfq_should_preempt(struct cfq_data *cfqd, struct cfq_queue *new_cfqq,
 	    RB_EMPTY_ROOT(&cfqq->sort_list))
 		return true;
 
-	/*
-	 * So both queues are sync. Let the new request get disk time if
-	 * it's a metadata request and the current queue is doing regular IO.
-	 */
-	if ((rq->cmd_flags & REQ_META) && !cfqq->meta_pending)
-		return true;
-
 	/*
 	 * Allow an RT request to pre-empt an ongoing non-RT cfqq timeslice.
 	 */
@@ -3410,10 +3420,8 @@ cfq_rq_enqueued(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 	struct cfq_io_context *cic = RQ_CIC(rq);
 
 	cfqd->rq_queued++;
-	if (rq->cmd_flags & REQ_META)
-		cfqq->meta_pending++;
 
-	cfq_update_io_thinktime(cfqd, cic);
+	cfq_update_io_thinktime(cfqd, cfqq, cic);
 	cfq_update_io_seektime(cfqd, cfqq, rq);
 	cfq_update_idle_window(cfqd, cfqq, cic);
 
@@ -3520,12 +3528,16 @@ static bool cfq_should_wait_busy(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 	if (cfqq->cfqg->nr_cfqq > 1)
 		return false;
 
+	/* the only queue in the group, but think time is big */
+	if (cfq_io_thinktime_big(cfqd, &cfqq->cfqg->ttime, true))
+		return false;
+
 	if (cfq_slice_used(cfqq))
 		return true;
 
 	/* if slice left is less than think time, wait busy */
-	if (cic && sample_valid(cic->ttime_samples)
-	    && (cfqq->slice_end - jiffies < cic->ttime_mean))
+	if (cic && sample_valid(cic->ttime.ttime_samples)
+	    && (cfqq->slice_end - jiffies < cic->ttime.ttime_mean))
 		return true;
 
 	/*
@@ -3566,11 +3578,24 @@ static void cfq_completed_request(struct request_queue *q, struct request *rq)
 	cfqd->rq_in_flight[cfq_cfqq_sync(cfqq)]--;
 
 	if (sync) {
-		RQ_CIC(rq)->last_end_request = now;
+		struct cfq_rb_root *service_tree;
+
+		RQ_CIC(rq)->ttime.last_end_request = now;
+
+		if (cfq_cfqq_on_rr(cfqq))
+			service_tree = cfqq->service_tree;
+		else
+			service_tree = service_tree_for(cfqq->cfqg,
+				cfqq_prio(cfqq), cfqq_type(cfqq));
+		service_tree->ttime.last_end_request = now;
 		if (!time_after(rq->start_time + cfqd->cfq_fifo_expire[1], now))
 			cfqd->last_delayed_sync = now;
 	}
 
+#ifdef CONFIG_CFQ_GROUP_IOSCHED
+	cfqq->cfqg->ttime.last_end_request = now;
+#endif
+
 	/*
 	 * 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.
@@ -3616,30 +3641,6 @@ static void cfq_completed_request(struct request_queue *q, struct request *rq)
 		cfq_schedule_dispatch(cfqd);
 }
 
-/*
- * we temporarily boost lower priority queues if they are holding fs exclusive
- * resources. they are boosted to normal prio (CLASS_BE/4)
- */
-static void cfq_prio_boost(struct cfq_queue *cfqq)
-{
-	if (has_fs_excl()) {
-		/*
-		 * boost idle prio on transactions that would lock out other
-		 * users of the filesystem
-		 */
-		if (cfq_class_idle(cfqq))
-			cfqq->ioprio_class = IOPRIO_CLASS_BE;
-		if (cfqq->ioprio > IOPRIO_NORM)
-			cfqq->ioprio = IOPRIO_NORM;
-	} else {
-		/*
-		 * unboost the queue (if needed)
-		 */
-		cfqq->ioprio_class = cfqq->org_ioprio_class;
-		cfqq->ioprio = cfqq->org_ioprio;
-	}
-}
-
 static inline int __cfq_may_queue(struct cfq_queue *cfqq)
 {
 	if (cfq_cfqq_wait_request(cfqq) && !cfq_cfqq_must_alloc_slice(cfqq)) {
@@ -3670,7 +3671,6 @@ static int cfq_may_queue(struct request_queue *q, int rw)
 	cfqq = cic_to_cfqq(cic, rw_is_sync(rw));
 	if (cfqq) {
 		cfq_init_prio_data(cfqq, cic->ioc);
-		cfq_prio_boost(cfqq);
 
 		return __cfq_may_queue(cfqq);
 	}

block/compat_ioctl.c

@@ -208,19 +208,6 @@ static int compat_blkpg_ioctl(struct block_device *bdev, fmode_t mode,
 #define BLKBSZSET_32		_IOW(0x12, 113, int)
 #define BLKGETSIZE64_32		_IOR(0x12, 114, int)
 
-struct compat_floppy_struct {
-	compat_uint_t	size;
-	compat_uint_t	sect;
-	compat_uint_t	head;
-	compat_uint_t	track;
-	compat_uint_t	stretch;
-	unsigned char	gap;
-	unsigned char	rate;
-	unsigned char	spec1;
-	unsigned char	fmt_gap;
-	const compat_caddr_t name;
-};
-
 struct compat_floppy_drive_params {
 	char		cmos;
 	compat_ulong_t	max_dtr;
@@ -288,7 +275,6 @@ struct compat_floppy_write_errors {
 
 #define FDSETPRM32 _IOW(2, 0x42, struct compat_floppy_struct)
 #define FDDEFPRM32 _IOW(2, 0x43, struct compat_floppy_struct)
-#define FDGETPRM32 _IOR(2, 0x04, struct compat_floppy_struct)
 #define FDSETDRVPRM32 _IOW(2, 0x90, struct compat_floppy_drive_params)
 #define FDGETDRVPRM32 _IOR(2, 0x11, struct compat_floppy_drive_params)
 #define FDGETDRVSTAT32 _IOR(2, 0x12, struct compat_floppy_drive_struct)

block/deadline-iosched.c

@@ -77,10 +77,8 @@ static void
 deadline_add_rq_rb(struct deadline_data *dd, struct request *rq)
 {
 	struct rb_root *root = deadline_rb_root(dd, rq);
-	struct request *__alias;
 
-	while (unlikely(__alias = elv_rb_add(root, rq)))
-		deadline_move_request(dd, __alias);
+	elv_rb_add(root, rq);
 }
 
 static inline void

block/elevator.c

@@ -353,7 +353,7 @@ static struct request *elv_rqhash_find(struct request_queue *q, sector_t offset)
  * RB-tree support functions for inserting/lookup/removal of requests
  * in a sorted RB tree.
  */
-struct request *elv_rb_add(struct rb_root *root, struct request *rq)
+void elv_rb_add(struct rb_root *root, struct request *rq)
 {
 	struct rb_node **p = &root->rb_node;
 	struct rb_node *parent = NULL;
@@ -365,15 +365,12 @@ struct request *elv_rb_add(struct rb_root *root, struct request *rq)
 
 		if (blk_rq_pos(rq) < blk_rq_pos(__rq))
 			p = &(*p)->rb_left;
-		else if (blk_rq_pos(rq) > blk_rq_pos(__rq))
+		else if (blk_rq_pos(rq) >= blk_rq_pos(__rq))
 			p = &(*p)->rb_right;
-		else
-			return __rq;
 	}
 
 	rb_link_node(&rq->rb_node, parent, p);
 	rb_insert_color(&rq->rb_node, root);
-	return NULL;
 }
 EXPORT_SYMBOL(elv_rb_add);
 

block/genhd.c

@@ -602,7 +602,7 @@ void add_disk(struct gendisk *disk)
 	disk->major = MAJOR(devt);
 	disk->first_minor = MINOR(devt);
 
-	/* Register BDI before referencing it from bdev */ 
+	/* Register BDI before referencing it from bdev */
 	bdi = &disk->queue->backing_dev_info;
 	bdi_register_dev(bdi, disk_devt(disk));
 
@@ -1140,7 +1140,7 @@ static int diskstats_show(struct seq_file *seqf, void *v)
 				"wsect wuse running use aveq"
 				"\n\n");
 	*/
- 
+
 	disk_part_iter_init(&piter, gp, DISK_PITER_INCL_EMPTY_PART0);
 	while ((hd = disk_part_iter_next(&piter))) {
 		cpu = part_stat_lock();
@@ -1164,7 +1164,7 @@ static int diskstats_show(struct seq_file *seqf, void *v)
 			);
 	}
 	disk_part_iter_exit(&piter);
- 
+
 	return 0;
 }
 
@@ -1492,30 +1492,32 @@ void disk_unblock_events(struct gendisk *disk)
 }
 
 /**
- * disk_check_events - schedule immediate event checking
- * @disk: disk to check events for
+ * disk_flush_events - schedule immediate event checking and flushing
+ * @disk: disk to check and flush events for
+ * @mask: events to flush
  *
- * Schedule immediate event checking on @disk if not blocked.
+ * Schedule immediate event checking on @disk if not blocked.  Events in
+ * @mask are scheduled to be cleared from the driver.  Note that this
+ * doesn't clear the events from @disk->ev.
  *
  * CONTEXT:
- * Don't care.  Safe to call from irq context.
+ * If @mask is non-zero must be called with bdev->bd_mutex held.
  */
-void disk_check_events(struct gendisk *disk)
+void disk_flush_events(struct gendisk *disk, unsigned int mask)
 {
 	struct disk_events *ev = disk->ev;
-	unsigned long flags;
 
 	if (!ev)
 		return;
 
-	spin_lock_irqsave(&ev->lock, flags);
+	spin_lock_irq(&ev->lock);
+	ev->clearing |= mask;
 	if (!ev->block) {
 		cancel_delayed_work(&ev->dwork);
 		queue_delayed_work(system_nrt_wq, &ev->dwork, 0);
 	}
-	spin_unlock_irqrestore(&ev->lock, flags);
+	spin_unlock_irq(&ev->lock);
 }
-EXPORT_SYMBOL_GPL(disk_check_events);
 
 /**
  * disk_clear_events - synchronously check, clear and return pending events
@@ -1705,7 +1707,7 @@ static int disk_events_set_dfl_poll_msecs(const char *val,
 	mutex_lock(&disk_events_mutex);
 
 	list_for_each_entry(ev, &disk_events, node)
-		disk_check_events(ev->disk);
+		disk_flush_events(ev->disk, 0);
 
 	mutex_unlock(&disk_events_mutex);
 

fs/block_dev.c

@@ -1448,6 +1448,8 @@ static int __blkdev_put(struct block_device *bdev, fmode_t mode, int for_part)
 
 int blkdev_put(struct block_device *bdev, fmode_t mode)
 {
+	mutex_lock(&bdev->bd_mutex);
+
 	if (mode & FMODE_EXCL) {
 		bool bdev_free;
 
@@ -1456,7 +1458,6 @@ int blkdev_put(struct block_device *bdev, fmode_t mode)
 		 * are protected with bdev_lock.  bd_mutex is to
 		 * synchronize disk_holder unlinking.
 		 */
-		mutex_lock(&bdev->bd_mutex);
 		spin_lock(&bdev_lock);
 
 		WARN_ON_ONCE(--bdev->bd_holders < 0);
@@ -1474,17 +1475,21 @@ int blkdev_put(struct block_device *bdev, fmode_t mode)
 		 * If this was the last claim, remove holder link and
 		 * unblock evpoll if it was a write holder.
 		 */
-		if (bdev_free) {
-			if (bdev->bd_write_holder) {
-				disk_unblock_events(bdev->bd_disk);
-				disk_check_events(bdev->bd_disk);
-				bdev->bd_write_holder = false;
-			}
+		if (bdev_free && bdev->bd_write_holder) {
+			disk_unblock_events(bdev->bd_disk);
+			bdev->bd_write_holder = false;
 		}
-
-		mutex_unlock(&bdev->bd_mutex);
 	}
 
+	/*
+	 * Trigger event checking and tell drivers to flush MEDIA_CHANGE
+	 * event.  This is to ensure detection of media removal commanded
+	 * from userland - e.g. eject(1).
+	 */
+	disk_flush_events(bdev->bd_disk, DISK_EVENT_MEDIA_CHANGE);
+
+	mutex_unlock(&bdev->bd_mutex);
+
 	return __blkdev_put(bdev, mode, 0);
 }
 EXPORT_SYMBOL(blkdev_put);

fs/compat_ioctl.c

@@ -68,6 +68,8 @@
 
 #ifdef CONFIG_BLOCK
 #include <linux/loop.h>
+#include <linux/cdrom.h>
+#include <linux/fd.h>
 #include <scsi/scsi.h>
 #include <scsi/scsi_ioctl.h>
 #include <scsi/sg.h>
@@ -944,6 +946,9 @@ COMPATIBLE_IOCTL(FIOQSIZE)
 IGNORE_IOCTL(LOOP_CLR_FD)
 /* md calls this on random blockdevs */
 IGNORE_IOCTL(RAID_VERSION)
+/* qemu/qemu-img might call these two on plain files for probing */
+IGNORE_IOCTL(CDROM_DRIVE_STATUS)
+IGNORE_IOCTL(FDGETPRM32)
 /* SG stuff */
 COMPATIBLE_IOCTL(SG_SET_TIMEOUT)
 COMPATIBLE_IOCTL(SG_GET_TIMEOUT)

fs/partitions/check.c

@@ -237,22 +237,22 @@ ssize_t part_size_show(struct device *dev,
 	return sprintf(buf, "%llu\n",(unsigned long long)p->nr_sects);
 }
 
-ssize_t part_ro_show(struct device *dev,
-		       struct device_attribute *attr, char *buf)
+static ssize_t part_ro_show(struct device *dev,
+			    struct device_attribute *attr, char *buf)
 {
 	struct hd_struct *p = dev_to_part(dev);
 	return sprintf(buf, "%d\n", p->policy ? 1 : 0);
 }
 
-ssize_t part_alignment_offset_show(struct device *dev,
-				   struct device_attribute *attr, char *buf)
+static ssize_t part_alignment_offset_show(struct device *dev,
+					  struct device_attribute *attr, char *buf)
 {
 	struct hd_struct *p = dev_to_part(dev);
 	return sprintf(buf, "%llu\n", (unsigned long long)p->alignment_offset);
 }
 
-ssize_t part_discard_alignment_show(struct device *dev,
-				   struct device_attribute *attr, char *buf)
+static ssize_t part_discard_alignment_show(struct device *dev,
+					   struct device_attribute *attr, char *buf)
 {
 	struct hd_struct *p = dev_to_part(dev);
 	return sprintf(buf, "%u\n", p->discard_alignment);

fs/reiserfs/journal.c

@@ -678,23 +678,19 @@ struct buffer_chunk {
 static void write_chunk(struct buffer_chunk *chunk)
 {
 	int i;
-	get_fs_excl();
 	for (i = 0; i < chunk->nr; i++) {
 		submit_logged_buffer(chunk->bh[i]);
 	}
 	chunk->nr = 0;
-	put_fs_excl();
 }
 
 static void write_ordered_chunk(struct buffer_chunk *chunk)
 {
 	int i;
-	get_fs_excl();
 	for (i = 0; i < chunk->nr; i++) {
 		submit_ordered_buffer(chunk->bh[i]);
 	}
 	chunk->nr = 0;
-	put_fs_excl();
 }
 
 static int add_to_chunk(struct buffer_chunk *chunk, struct buffer_head *bh,
@@ -986,8 +982,6 @@ static int flush_commit_list(struct super_block *s,
 		return 0;
 	}
 
-	get_fs_excl();
-
 	/* before we can put our commit blocks on disk, we have to make sure everyone older than
 	 ** us is on disk too
 	 */
@@ -1145,7 +1139,6 @@ static int flush_commit_list(struct super_block *s,
 	if (retval)
 		reiserfs_abort(s, retval, "Journal write error in %s",
 			       __func__);
-	put_fs_excl();
 	return retval;
 }
 
@@ -1374,8 +1367,6 @@ static int flush_journal_list(struct super_block *s,
 		return 0;
 	}
 
-	get_fs_excl();
-
 	/* if all the work is already done, get out of here */
 	if (atomic_read(&(jl->j_nonzerolen)) <= 0 &&
 	    atomic_read(&(jl->j_commit_left)) <= 0) {
@@ -1597,7 +1588,6 @@ static int flush_journal_list(struct super_block *s,
 	put_journal_list(s, jl);
 	if (flushall)
 		mutex_unlock(&journal->j_flush_mutex);
-	put_fs_excl();
 	return err;
 }
 
@@ -3108,7 +3098,6 @@ static int do_journal_begin_r(struct reiserfs_transaction_handle *th,
 	th->t_trans_id = journal->j_trans_id;
 	unlock_journal(sb);
 	INIT_LIST_HEAD(&th->t_list);
-	get_fs_excl();
 	return 0;
 
       out_fail:
@@ -3964,7 +3953,6 @@ static int do_journal_end(struct reiserfs_transaction_handle *th,
 	flush = flags & FLUSH_ALL;
 	wait_on_commit = flags & WAIT;
 
-	put_fs_excl();
 	current->journal_info = th->t_handle_save;
 	reiserfs_check_lock_depth(sb, "journal end");
 	if (journal->j_len == 0) {
@@ -4316,4 +4304,3 @@ void reiserfs_abort_journal(struct super_block *sb, int errno)
 	dump_stack();
 #endif
 }
-

fs/super.c

@@ -351,13 +351,11 @@ bool grab_super_passive(struct super_block *sb)
  */
 void lock_super(struct super_block * sb)
 {
-	get_fs_excl();
 	mutex_lock(&sb->s_lock);
 }
 
 void unlock_super(struct super_block * sb)
 {
-	put_fs_excl();
 	mutex_unlock(&sb->s_lock);
 }
 
@@ -385,7 +383,6 @@ void generic_shutdown_super(struct super_block *sb)
 	if (sb->s_root) {
 		shrink_dcache_for_umount(sb);
 		sync_filesystem(sb);
-		get_fs_excl();
 		sb->s_flags &= ~MS_ACTIVE;
 
 		fsnotify_unmount_inodes(&sb->s_inodes);
@@ -400,7 +397,6 @@ void generic_shutdown_super(struct super_block *sb)
 			   "Self-destruct in 5 seconds.  Have a nice day...\n",
 			   sb->s_id);
 		}
-		put_fs_excl();
 	}
 	spin_lock(&sb_lock);
 	/* should be initialized for __put_super_and_need_restart() */

include/linux/blkdev.h

@@ -73,7 +73,7 @@ enum rq_cmd_type_bits {
 
 /*
  * try to put the fields that are referenced together in the same cacheline.
- * if you modify this structure, be sure to check block/blk-core.c:rq_init()
+ * if you modify this structure, be sure to check block/blk-core.c:blk_rq_init()
  * as well!
  */
 struct request {
@@ -260,8 +260,7 @@ struct queue_limits {
 	unsigned char		discard_zeroes_data;
 };
 
-struct request_queue
-{
+struct request_queue {
 	/*
 	 * Together with queue_head for cacheline sharing
 	 */
@@ -304,14 +303,14 @@ struct request_queue
 	void			*queuedata;
 
 	/*
-	 * queue needs bounce pages for pages above this limit
+	 * various queue flags, see QUEUE_* below
 	 */
-	gfp_t			bounce_gfp;
+	unsigned long		queue_flags;
 
 	/*
-	 * various queue flags, see QUEUE_* below
+	 * queue needs bounce pages for pages above this limit
 	 */
-	unsigned long		queue_flags;
+	gfp_t			bounce_gfp;
 
 	/*
 	 * protects queue structures from reentrancy. ->__queue_lock should
@@ -334,8 +333,8 @@ struct request_queue
 	unsigned int		nr_congestion_off;
 	unsigned int		nr_batching;
 
-	void			*dma_drain_buffer;
 	unsigned int		dma_drain_size;
+	void			*dma_drain_buffer;
 	unsigned int		dma_pad_mask;
 	unsigned int		dma_alignment;
 
@@ -393,7 +392,7 @@ struct request_queue
 #define QUEUE_FLAG_ELVSWITCH	6	/* don't use elevator, just do FIFO */
 #define QUEUE_FLAG_BIDI		7	/* queue supports bidi requests */
 #define QUEUE_FLAG_NOMERGES     8	/* disable merge attempts */
-#define QUEUE_FLAG_SAME_COMP	9	/* force complete on same CPU */
+#define QUEUE_FLAG_SAME_COMP	9	/* complete on same CPU-group */
 #define QUEUE_FLAG_FAIL_IO     10	/* fake timeout */
 #define QUEUE_FLAG_STACKABLE   11	/* supports request stacking */
 #define QUEUE_FLAG_NONROT      12	/* non-rotational device (SSD) */
@@ -403,6 +402,7 @@ struct request_queue
 #define QUEUE_FLAG_NOXMERGES   15	/* No extended merges */
 #define QUEUE_FLAG_ADD_RANDOM  16	/* Contributes to random pool */
 #define QUEUE_FLAG_SECDISCARD  17	/* supports SECDISCARD */
+#define QUEUE_FLAG_SAME_FORCE  18	/* force complete on same CPU */
 
 #define QUEUE_FLAG_DEFAULT	((1 << QUEUE_FLAG_IO_STAT) |		\
 				 (1 << QUEUE_FLAG_STACKABLE)	|	\
@@ -857,12 +857,21 @@ struct request_queue *blk_alloc_queue(gfp_t);
 struct request_queue *blk_alloc_queue_node(gfp_t, int);
 extern void blk_put_queue(struct request_queue *);
 
+/*
+ * Note: Code in between changing the blk_plug list/cb_list or element of such
+ * lists is preemptable, but such code can't do sleep (or be very careful),
+ * otherwise data is corrupted. For details, please check schedule() where
+ * blk_schedule_flush_plug() is called.
+ */
 struct blk_plug {
 	unsigned long magic;
 	struct list_head list;
 	struct list_head cb_list;
 	unsigned int should_sort;
+	unsigned int count;
 };
+#define BLK_MAX_REQUEST_COUNT 16
+
 struct blk_plug_cb {
 	struct list_head list;
 	void (*callback)(struct blk_plug_cb *);

include/linux/elevator.h

@@ -146,7 +146,7 @@ extern struct request *elv_rb_latter_request(struct request_queue *, struct requ
 /*
  * rb support functions.
  */
-extern struct request *elv_rb_add(struct rb_root *, struct request *);
+extern void elv_rb_add(struct rb_root *, struct request *);
 extern void elv_rb_del(struct rb_root *, struct request *);
 extern struct request *elv_rb_find(struct rb_root *, sector_t);
 

include/linux/fd.h

@@ -377,4 +377,26 @@ struct floppy_raw_cmd {
 #define FDEJECT _IO(2, 0x5a)
 /* eject the disk */
 
+
+#ifdef __KERNEL__
+#ifdef CONFIG_COMPAT
+#include <linux/compat.h>
+
+struct compat_floppy_struct {
+	compat_uint_t	size;
+	compat_uint_t	sect;
+	compat_uint_t	head;
+	compat_uint_t	track;
+	compat_uint_t	stretch;
+	unsigned char	gap;
+	unsigned char	rate;
+	unsigned char	spec1;
+	unsigned char	fmt_gap;
+	const compat_caddr_t name;
+};
+
+#define FDGETPRM32 _IOR(2, 0x04, struct compat_floppy_struct)
+#endif
+#endif
+
 #endif

include/linux/fs.h

@@ -1469,10 +1469,6 @@ enum {
 #define vfs_check_frozen(sb, level) \
 	wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
 
-#define get_fs_excl() atomic_inc(&current->fs_excl)
-#define put_fs_excl() atomic_dec(&current->fs_excl)
-#define has_fs_excl() atomic_read(&current->fs_excl)
-
 /*
  * until VFS tracks user namespaces for inodes, just make all files
  * belong to init_user_ns

include/linux/genhd.h

@@ -420,7 +420,7 @@ static inline int get_disk_ro(struct gendisk *disk)
 
 extern void disk_block_events(struct gendisk *disk);
 extern void disk_unblock_events(struct gendisk *disk);
-extern void disk_check_events(struct gendisk *disk);
+extern void disk_flush_events(struct gendisk *disk, unsigned int mask);
 extern unsigned int disk_clear_events(struct gendisk *disk, unsigned int mask);
 
 /* drivers/char/random.c */

include/linux/init_task.h

@@ -176,7 +176,6 @@ extern struct cred init_cred;
 	.alloc_lock	= __SPIN_LOCK_UNLOCKED(tsk.alloc_lock),		\
 	.journal_info	= NULL,						\
 	.cpu_timers	= INIT_CPU_TIMERS(tsk.cpu_timers),		\
-	.fs_excl	= ATOMIC_INIT(0),				\
 	.pi_lock	= __RAW_SPIN_LOCK_UNLOCKED(tsk.pi_lock),	\
 	.timer_slack_ns = 50000, /* 50 usec default slack */		\
 	.pids = {							\

include/linux/iocontext.h

@@ -5,6 +5,14 @@
 #include <linux/rcupdate.h>
 
 struct cfq_queue;
+struct cfq_ttime {
+	unsigned long last_end_request;
+
+	unsigned long ttime_total;
+	unsigned long ttime_samples;
+	unsigned long ttime_mean;
+};
+
 struct cfq_io_context {
 	void *key;
 
@@ -12,11 +20,7 @@ struct cfq_io_context {
 
 	struct io_context *ioc;
 
-	unsigned long last_end_request;
-
-	unsigned long ttime_total;
-	unsigned long ttime_samples;
-	unsigned long ttime_mean;
+	struct cfq_ttime ttime;
 
 	struct list_head queue_list;
 	struct hlist_node cic_list;

include/linux/sched.h

@@ -1512,7 +1512,6 @@ struct task_struct {
 	short il_next;
 	short pref_node_fork;
 #endif
-	atomic_t fs_excl;	/* holding fs exclusive resources */
 	struct rcu_head rcu;
 
 	/*

kernel/exit.c

@@ -898,7 +898,6 @@ NORET_TYPE void do_exit(long code)
 
 	profile_task_exit(tsk);
 
-	WARN_ON(atomic_read(&tsk->fs_excl));
 	WARN_ON(blk_needs_flush_plug(tsk));
 
 	if (unlikely(in_interrupt()))

kernel/fork.c

@@ -290,7 +290,6 @@ static struct task_struct *dup_task_struct(struct task_struct *orig)
 
 	/* One for us, one for whoever does the "release_task()" (usually parent) */
 	atomic_set(&tsk->usage,2);
-	atomic_set(&tsk->fs_excl, 0);
 #ifdef CONFIG_BLK_DEV_IO_TRACE
 	tsk->btrace_seq = 0;
 #endif

mm/backing-dev.c

@@ -505,7 +505,7 @@ static void bdi_remove_from_list(struct backing_dev_info *bdi)
 	list_del_rcu(&bdi->bdi_list);
 	spin_unlock_bh(&bdi_lock);
 
-	synchronize_rcu();
+	synchronize_rcu_expedited();
 }
 
 int bdi_register(struct backing_dev_info *bdi, struct device *parent,