Merge commit 'v2.6.38-rc6' into for-2.6.39/core

Conflicts:
	block/cfq-iosched.c

Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

9 files changed, 650 insertions, 109 deletions

diff --git a/block/Kconfig b/block/Kconfig
index 6c9213ef15a1..60be1e0455da 100644
--- a/block/Kconfig
+++ b/block/Kconfig
@@ -2,7 +2,7 @@
 # Block layer core configuration
 #
 menuconfig BLOCK
-       bool "Enable the block layer" if EMBEDDED
+       bool "Enable the block layer" if EXPERT
        default y
        help
          Provide block layer support for the kernel.
diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index b1febd0f6d2a..455768a3eb9e 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -1452,10 +1452,6 @@ blkiocg_create(struct cgroup_subsys *subsys, struct cgroup *cgroup)
                 goto done;
         }
 
-        /* Currently we do not support hierarchy deeper than two level (0,1) */
-        if (parent != cgroup->top_cgroup)
-                return ERR_PTR(-EPERM);
-
         blkcg = kzalloc(sizeof(*blkcg), GFP_KERNEL);
         if (!blkcg)
                 return ERR_PTR(-ENOMEM);
diff --git a/block/blk-core.c b/block/blk-core.c
index ab4a7696956d..3cc17e6064d6 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -33,7 +33,7 @@
 
 #include "blk.h"
 
-EXPORT_TRACEPOINT_SYMBOL_GPL(block_remap);
+EXPORT_TRACEPOINT_SYMBOL_GPL(block_bio_remap);
 EXPORT_TRACEPOINT_SYMBOL_GPL(block_rq_remap);
 EXPORT_TRACEPOINT_SYMBOL_GPL(block_bio_complete);
 
@@ -64,13 +64,27 @@ static void drive_stat_acct(struct request *rq, int new_io)
                 return;
 
         cpu = part_stat_lock();
-        part = disk_map_sector_rcu(rq->rq_disk, blk_rq_pos(rq));
 
-        if (!new_io)
+        if (!new_io) {
+                part = rq->part;
                 part_stat_inc(cpu, part, merges[rw]);
-        else {
+        } else {
+                part = disk_map_sector_rcu(rq->rq_disk, blk_rq_pos(rq));
+                if (!hd_struct_try_get(part)) {
+                        /*
+                         * The partition is already being removed,
+                         * the request will be accounted on the disk only
+                         *
+                         * We take a reference on disk->part0 although that
+                         * partition will never be deleted, so we can treat
+                         * it as any other partition.
+                         */
+                        part = &rq->rq_disk->part0;
+                        hd_struct_get(part);
+                }
                 part_round_stats(cpu, part);
                 part_inc_in_flight(part, rw);
+                rq->part = part;
         }
 
         part_stat_unlock();
@@ -128,6 +142,7 @@ void blk_rq_init(struct request_queue *q, struct request *rq)
         rq->ref_count = 1;
         rq->start_time = jiffies;
         set_start_time_ns(rq);
+        rq->part = NULL;
 }
 EXPORT_SYMBOL(blk_rq_init);
 
@@ -1342,9 +1357,9 @@ static inline void blk_partition_remap(struct bio *bio)
                 bio->bi_sector += p->start_sect;
                 bio->bi_bdev = bdev->bd_contains;
 
-                trace_block_remap(bdev_get_queue(bio->bi_bdev), bio,
-                                    bdev->bd_dev,
-                                    bio->bi_sector - p->start_sect);
+                trace_block_bio_remap(bdev_get_queue(bio->bi_bdev), bio,
+                                      bdev->bd_dev,
+                                      bio->bi_sector - p->start_sect);
         }
 }
 
@@ -1513,7 +1528,7 @@ static inline void __generic_make_request(struct bio *bio)
                         goto end_io;
 
                 if (old_sector != -1)
-                        trace_block_remap(q, bio, old_dev, old_sector);
+                        trace_block_bio_remap(q, bio, old_dev, old_sector);
 
                 old_sector = bio->bi_sector;
                 old_dev = bio->bi_bdev->bd_dev;
@@ -1789,7 +1804,7 @@ static void blk_account_io_completion(struct request *req, unsigned int bytes)
                 int cpu;
 
                 cpu = part_stat_lock();
-                part = disk_map_sector_rcu(req->rq_disk, blk_rq_pos(req));
+                part = req->part;
                 part_stat_add(cpu, part, sectors[rw], bytes >> 9);
                 part_stat_unlock();
         }
@@ -1809,13 +1824,14 @@ static void blk_account_io_done(struct request *req)
                 int cpu;
 
                 cpu = part_stat_lock();
-                part = disk_map_sector_rcu(req->rq_disk, blk_rq_pos(req));
+                part = req->part;
 
                 part_stat_inc(cpu, part, ios[rw]);
                 part_stat_add(cpu, part, ticks[rw], duration);
                 part_round_stats(cpu, part);
                 part_dec_in_flight(part, rw);
 
+                hd_struct_put(part);
                 part_stat_unlock();
         }
 }
@@ -2619,7 +2635,9 @@ int __init blk_dev_init(void)
         BUILD_BUG_ON(__REQ_NR_BITS > 8 *
                         sizeof(((struct request *)0)->cmd_flags));
 
-        kblockd_workqueue = create_workqueue("kblockd");
+        /* used for unplugging and affects IO latency/throughput - HIGHPRI */
+        kblockd_workqueue = alloc_workqueue("kblockd",
+                                            WQ_MEM_RECLAIM | WQ_HIGHPRI, 0);
         if (!kblockd_workqueue)
                 panic("Failed to create kblockd\n");
 
diff --git a/block/blk-ioc.c b/block/blk-ioc.c
index 3c7a339fe381..b791022beef3 100644
--- a/block/blk-ioc.c
+++ b/block/blk-ioc.c
@@ -64,7 +64,7 @@ static void cfq_exit(struct io_context *ioc)
         rcu_read_unlock();
 }
 
-/* Called by the exitting task */
+/* Called by the exiting task */
 void exit_io_context(struct task_struct *task)
 {
         struct io_context *ioc;
@@ -74,10 +74,9 @@ void exit_io_context(struct task_struct *task)
         task->io_context = NULL;
         task_unlock(task);
 
-        if (atomic_dec_and_test(&ioc->nr_tasks)) {
+        if (atomic_dec_and_test(&ioc->nr_tasks))
                 cfq_exit(ioc);
 
-        }
         put_io_context(ioc);
 }
 
diff --git a/block/blk-merge.c b/block/blk-merge.c
index 74bc4a768f32..ea85e20d5e94 100644
--- a/block/blk-merge.c
+++ b/block/blk-merge.c
@@ -351,11 +351,12 @@ static void blk_account_io_merge(struct request *req)
                 int cpu;
 
                 cpu = part_stat_lock();
-                part = disk_map_sector_rcu(req->rq_disk, blk_rq_pos(req));
+                part = req->part;
 
                 part_round_stats(cpu, part);
                 part_dec_in_flight(part, rq_data_dir(req));
 
+                hd_struct_put(part);
                 part_stat_unlock();
         }
 }
diff --git a/block/blk-throttle.c b/block/blk-throttle.c
index 381b09bb562b..a89043a3caa4 100644
--- a/block/blk-throttle.c
+++ b/block/blk-throttle.c
@@ -168,7 +168,15 @@ static struct throtl_grp * throtl_find_alloc_tg(struct throtl_data *td,
          * tree of blkg (instead of traversing through hash list all
          * the time.
          */
-        tg = tg_of_blkg(blkiocg_lookup_group(blkcg, key));
+
+        /*
+         * This is the common case when there are no blkio cgroups.
+         * Avoid lookup in this case
+         */
+        if (blkcg == &blkio_root_cgroup)
+                tg = &td->root_tg;
+        else
+                tg = tg_of_blkg(blkiocg_lookup_group(blkcg, key));
 
         /* Fill in device details for root group */
         if (tg && !tg->blkg.dev && bdi->dev && dev_name(bdi->dev)) {
diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
index 968455c57e1a..f27ff3efe6cd 100644
--- a/block/cfq-iosched.c
+++ b/block/cfq-iosched.c
@@ -87,7 +87,6 @@ struct cfq_rb_root {
         unsigned count;
         unsigned total_weight;
         u64 min_vdisktime;
-        struct rb_node *active;
 };
 #define CFQ_RB_ROOT     (struct cfq_rb_root) { .rb = RB_ROOT, .left = NULL, \
                         .count = 0, .min_vdisktime = 0, }
@@ -97,7 +96,7 @@ struct cfq_rb_root {
  */
 struct cfq_queue {
         /* reference count */
-        atomic_t ref;
+        int ref;
         /* various state flags, see below */
         unsigned int flags;
         /* parent cfq_data */
@@ -180,7 +179,6 @@ struct cfq_group {
         /* group service_tree key */
         u64 vdisktime;
         unsigned int weight;
-        bool on_st;
 
         /* number of cfqq currently on this group */
         int nr_cfqq;
@@ -209,7 +207,7 @@ struct cfq_group {
         struct blkio_group blkg;
 #ifdef CONFIG_CFQ_GROUP_IOSCHED
         struct hlist_node cfqd_node;
-        atomic_t ref;
+        int ref;
 #endif
         /* number of requests that are on the dispatch list or inside driver */
         int dispatched;
@@ -563,11 +561,6 @@ static void update_min_vdisktime(struct cfq_rb_root *st)
         u64 vdisktime = st->min_vdisktime;
         struct cfq_group *cfqg;
 
-        if (st->active) {
-                cfqg = rb_entry_cfqg(st->active);
-                vdisktime = cfqg->vdisktime;
-        }
-
         if (st->left) {
                 cfqg = rb_entry_cfqg(st->left);
                 vdisktime = min_vdisktime(vdisktime, cfqg->vdisktime);
@@ -605,8 +598,8 @@ cfq_group_slice(struct cfq_data *cfqd, struct cfq_group *cfqg)
         return cfq_target_latency * cfqg->weight / st->total_weight;
 }
 
-static inline void
-cfq_set_prio_slice(struct cfq_data *cfqd, struct cfq_queue *cfqq)
+static inline unsigned
+cfq_scaled_cfqq_slice(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 {
         unsigned slice = cfq_prio_to_slice(cfqd, cfqq);
         if (cfqd->cfq_latency) {
@@ -632,6 +625,14 @@ cfq_set_prio_slice(struct cfq_data *cfqd, struct cfq_queue *cfqq)
                                     low_slice);
                 }
         }
+        return slice;
+}
+
+static inline void
+cfq_set_prio_slice(struct cfq_data *cfqd, struct cfq_queue *cfqq)
+{
+        unsigned slice = cfq_scaled_cfqq_slice(cfqd, cfqq);
+
         cfqq->slice_start = jiffies;
         cfqq->slice_end = jiffies + slice;
         cfqq->allocated_slice = slice;
@@ -646,11 +647,11 @@ cfq_set_prio_slice(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 static inline bool cfq_slice_used(struct cfq_queue *cfqq)
 {
         if (cfq_cfqq_slice_new(cfqq))
-                return 0;
+                return false;
         if (time_before(jiffies, cfqq->slice_end))
-                return 0;
+                return false;
 
-        return 1;
+        return true;
 }
 
 /*
@@ -869,7 +870,7 @@ cfq_group_service_tree_add(struct cfq_data *cfqd, struct cfq_group *cfqg)
         struct rb_node *n;
 
         cfqg->nr_cfqq++;
-        if (cfqg->on_st)
+        if (!RB_EMPTY_NODE(&cfqg->rb_node))
                 return;
 
         /*
@@ -885,7 +886,6 @@ cfq_group_service_tree_add(struct cfq_data *cfqd, struct cfq_group *cfqg)
                 cfqg->vdisktime = st->min_vdisktime;
 
         __cfq_group_service_tree_add(st, cfqg);
-        cfqg->on_st = true;
         st->total_weight += cfqg->weight;
 }
 
@@ -894,9 +894,6 @@ cfq_group_service_tree_del(struct cfq_data *cfqd, struct cfq_group *cfqg)
 {
         struct cfq_rb_root *st = &cfqd->grp_service_tree;
 
-        if (st->active == &cfqg->rb_node)
-                st->active = NULL;
-
         BUG_ON(cfqg->nr_cfqq < 1);
         cfqg->nr_cfqq--;
 
@@ -905,7 +902,6 @@ cfq_group_service_tree_del(struct cfq_data *cfqd, struct cfq_group *cfqg)
                 return;
 
         cfq_log_cfqg(cfqd, cfqg, "del_from_rr group");
-        cfqg->on_st = false;
         st->total_weight -= cfqg->weight;
         if (!RB_EMPTY_NODE(&cfqg->rb_node))
                 cfq_rb_erase(&cfqg->rb_node, st);
@@ -1026,11 +1022,11 @@ cfq_find_alloc_cfqg(struct cfq_data *cfqd, struct cgroup *cgroup, int create)
          * elevator which will be dropped by either elevator exit
          * or cgroup deletion path depending on who is exiting first.
          */
-        atomic_set(&cfqg->ref, 1);
+        cfqg->ref = 1;
 
         /*
          * Add group onto cgroup list. It might happen that bdi->dev is
-         * not initiliazed yet. Initialize this new group without major
+         * not initialized yet. Initialize this new group without major
          * and minor info and this info will be filled in once a new thread
          * comes for IO. See code above.
          */
@@ -1071,7 +1067,7 @@ static struct cfq_group *cfq_get_cfqg(struct cfq_data *cfqd, int create)
 
 static inline struct cfq_group *cfq_ref_get_cfqg(struct cfq_group *cfqg)
 {
-        atomic_inc(&cfqg->ref);
+        cfqg->ref++;
         return cfqg;
 }
 
@@ -1083,7 +1079,7 @@ static void cfq_link_cfqq_cfqg(struct cfq_queue *cfqq, struct cfq_group *cfqg)
 
         cfqq->cfqg = cfqg;
         /* cfqq reference on cfqg */
-        atomic_inc(&cfqq->cfqg->ref);
+        cfqq->cfqg->ref++;
 }
 
 static void cfq_put_cfqg(struct cfq_group *cfqg)
@@ -1091,11 +1087,12 @@ static void cfq_put_cfqg(struct cfq_group *cfqg)
         struct cfq_rb_root *st;
         int i, j;
 
-        BUG_ON(atomic_read(&cfqg->ref) <= 0);
-        if (!atomic_dec_and_test(&cfqg->ref))
+        BUG_ON(cfqg->ref <= 0);
+        cfqg->ref--;
+        if (cfqg->ref)
                 return;
         for_each_cfqg_st(cfqg, i, j, st)
-                BUG_ON(!RB_EMPTY_ROOT(&st->rb) || st->active != NULL);
+                BUG_ON(!RB_EMPTY_ROOT(&st->rb));
         kfree(cfqg);
 }
 
@@ -1200,7 +1197,7 @@ static void cfq_service_tree_add(struct cfq_data *cfqd, struct cfq_queue *cfqq,
                         cfq_group_service_tree_del(cfqd, cfqq->cfqg);
                 cfqq->orig_cfqg = cfqq->cfqg;
                 cfqq->cfqg = &cfqd->root_group;
-                atomic_inc(&cfqd->root_group.ref);
+                cfqd->root_group.ref++;
                 group_changed = 1;
         } else if (!cfqd->cfq_group_isolation
                    && cfqq_type(cfqq) == SYNC_WORKLOAD && cfqq->orig_cfqg) {
@@ -1672,8 +1669,11 @@ __cfq_slice_expired(struct cfq_data *cfqd, struct cfq_queue *cfqq,
         /*
          * store what was left of this slice, if the queue idled/timed out
          */
-        if (timed_out && !cfq_cfqq_slice_new(cfqq)) {
-                cfqq->slice_resid = cfqq->slice_end - jiffies;
+        if (timed_out) {
+                if (cfq_cfqq_slice_new(cfqq))
+                        cfqq->slice_resid = cfq_scaled_cfqq_slice(cfqd, cfqq);
+                else
+                        cfqq->slice_resid = cfqq->slice_end - jiffies;
                 cfq_log_cfqq(cfqd, cfqq, "resid=%ld", cfqq->slice_resid);
         }
 
@@ -1687,9 +1687,6 @@ __cfq_slice_expired(struct cfq_data *cfqd, struct cfq_queue *cfqq,
         if (cfqq == cfqd->active_queue)
                 cfqd->active_queue = NULL;
 
-        if (&cfqq->cfqg->rb_node == cfqd->grp_service_tree.active)
-                cfqd->grp_service_tree.active = NULL;
-
         if (cfqd->active_cic) {
                 put_io_context(cfqd->active_cic->ioc);
                 cfqd->active_cic = NULL;
@@ -1901,10 +1898,10 @@ static bool cfq_should_idle(struct cfq_data *cfqd, struct cfq_queue *cfqq)
          * in their service tree.
          */
         if (service_tree->count == 1 && cfq_cfqq_sync(cfqq))
-                return 1;
+                return true;
         cfq_log_cfqq(cfqd, cfqq, "Not idling. st->count:%d",
                         service_tree->count);
-        return 0;
+        return false;
 }
 
 static void cfq_arm_slice_timer(struct cfq_data *cfqd)
@@ -2040,7 +2037,7 @@ static int cfqq_process_refs(struct cfq_queue *cfqq)
         int process_refs, io_refs;
 
         io_refs = cfqq->allocated[READ] + cfqq->allocated[WRITE];
-        process_refs = atomic_read(&cfqq->ref) - io_refs;
+        process_refs = cfqq->ref - io_refs;
         BUG_ON(process_refs < 0);
         return process_refs;
 }
@@ -2080,10 +2077,10 @@ static void cfq_setup_merge(struct cfq_queue *cfqq, struct cfq_queue *new_cfqq)
          */
         if (new_process_refs >= process_refs) {
                 cfqq->new_cfqq = new_cfqq;
-                atomic_add(process_refs, &new_cfqq->ref);
+                new_cfqq->ref += process_refs;
         } else {
                 new_cfqq->new_cfqq = cfqq;
-                atomic_add(new_process_refs, &cfqq->ref);
+                cfqq->ref += new_process_refs;
         }
 }
 
@@ -2116,12 +2113,7 @@ static void choose_service_tree(struct cfq_data *cfqd, struct cfq_group *cfqg)
         unsigned count;
         struct cfq_rb_root *st;
         unsigned group_slice;
-
-        if (!cfqg) {
-                cfqd->serving_prio = IDLE_WORKLOAD;
-                cfqd->workload_expires = jiffies + 1;
-                return;
-        }
+        enum wl_prio_t original_prio = cfqd->serving_prio;
 
         /* Choose next priority. RT > BE > IDLE */
         if (cfq_group_busy_queues_wl(RT_WORKLOAD, cfqd, cfqg))
@@ -2134,6 +2126,9 @@ static void choose_service_tree(struct cfq_data *cfqd, struct cfq_group *cfqg)
                 return;
         }
 
+        if (original_prio != cfqd->serving_prio)
+                goto new_workload;
+
         /*
          * For RT and BE, we have to choose also the type
          * (SYNC, SYNC_NOIDLE, ASYNC), and to compute a workload
@@ -2148,6 +2143,7 @@ static void choose_service_tree(struct cfq_data *cfqd, struct cfq_group *cfqg)
         if (count && !time_after(jiffies, cfqd->workload_expires))
                 return;
 
+new_workload:
         /* otherwise select new workload type */
         cfqd->serving_type =
                 cfq_choose_wl(cfqd, cfqg, cfqd->serving_prio);
@@ -2199,7 +2195,6 @@ static struct cfq_group *cfq_get_next_cfqg(struct cfq_data *cfqd)
         if (RB_EMPTY_ROOT(&st->rb))
                 return NULL;
         cfqg = cfq_rb_first_group(st);
-        st->active = &cfqg->rb_node;
         update_min_vdisktime(st);
         return cfqg;
 }
@@ -2293,6 +2288,17 @@ static struct cfq_queue *cfq_select_queue(struct cfq_data *cfqd)
                 goto keep_queue;
         }
 
+        /*
+         * This is a deep seek queue, but the device is much faster than
+         * the queue can deliver, don't idle
+         **/
+        if (CFQQ_SEEKY(cfqq) && cfq_cfqq_idle_window(cfqq) &&
+            (cfq_cfqq_slice_new(cfqq) ||
+            (cfqq->slice_end - jiffies > jiffies - cfqq->slice_start))) {
+                cfq_clear_cfqq_deep(cfqq);
+                cfq_clear_cfqq_idle_window(cfqq);
+        }
+
         if (cfqq->dispatched && cfq_should_idle(cfqd, cfqq)) {
                 cfqq = NULL;
                 goto keep_queue;
@@ -2367,12 +2373,12 @@ static inline bool cfq_slice_used_soon(struct cfq_data *cfqd,
 {
         /* the queue hasn't finished any request, can't estimate */
         if (cfq_cfqq_slice_new(cfqq))
-                return 1;
+                return true;
         if (time_after(jiffies + cfqd->cfq_slice_idle * cfqq->dispatched,
                 cfqq->slice_end))
-                return 1;
+                return true;
 
-        return 0;
+        return false;
 }
 
 static bool cfq_may_dispatch(struct cfq_data *cfqd, struct cfq_queue *cfqq)
@@ -2538,9 +2544,10 @@ static void cfq_put_queue(struct cfq_queue *cfqq)
         struct cfq_data *cfqd = cfqq->cfqd;
         struct cfq_group *cfqg, *orig_cfqg;
 
-        BUG_ON(atomic_read(&cfqq->ref) <= 0);
+        BUG_ON(cfqq->ref <= 0);
 
-        if (!atomic_dec_and_test(&cfqq->ref))
+        cfqq->ref--;
+        if (cfqq->ref)
                 return;
 
         cfq_log_cfqq(cfqd, cfqq, "put_queue");
@@ -2843,7 +2850,7 @@ static void cfq_init_cfqq(struct cfq_data *cfqd, struct cfq_queue *cfqq,
         RB_CLEAR_NODE(&cfqq->p_node);
         INIT_LIST_HEAD(&cfqq->fifo);
 
-        atomic_set(&cfqq->ref, 0);
+        cfqq->ref = 0;
         cfqq->cfqd = cfqd;
 
         cfq_mark_cfqq_prio_changed(cfqq);
@@ -2979,11 +2986,11 @@ cfq_get_queue(struct cfq_data *cfqd, bool is_sync, struct io_context *ioc,
          * pin the queue now that it's allocated, scheduler exit will prune it
          */
         if (!is_sync && !(*async_cfqq)) {
-                atomic_inc(&cfqq->ref);
+                cfqq->ref++;
                 *async_cfqq = cfqq;
         }
 
-        atomic_inc(&cfqq->ref);
+        cfqq->ref++;
         return cfqq;
 }
 
@@ -3265,6 +3272,10 @@ cfq_should_preempt(struct cfq_data *cfqd, struct cfq_queue *new_cfqq,
         if (cfq_class_rt(new_cfqq) && !cfq_class_rt(cfqq))
                 return true;
 
+        /* An idle queue should not be idle now for some reason */
+        if (RB_EMPTY_ROOT(&cfqq->sort_list) && !cfq_should_idle(cfqd, cfqq))
+                return true;
+
         if (!cfqd->active_cic || !cfq_cfqq_wait_request(cfqq))
                 return false;
 
@@ -3284,10 +3295,19 @@ 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)
 {
+        struct cfq_queue *old_cfqq = cfqd->active_queue;
+
         cfq_log_cfqq(cfqd, cfqq, "preempt");
         cfq_slice_expired(cfqd, 1);
 
         /*
+         * workload type is changed, don't save slice, otherwise preempt
+         * doesn't happen
+         */
+        if (cfqq_type(old_cfqq) != cfqq_type(cfqq))
+                cfqq->cfqg->saved_workload_slice = 0;
+
+        /*
          * Put the new queue at the front of the of the current list,
          * so we know that it will be selected next.
          */
@@ -3412,6 +3432,10 @@ static bool cfq_should_wait_busy(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 {
         struct cfq_io_context *cic = cfqd->active_cic;
 
+        /* If the queue already has requests, don't wait */
+        if (!RB_EMPTY_ROOT(&cfqq->sort_list))
+                return false;
+
         /* If there are other queues in the group, don't wait */
         if (cfqq->cfqg->nr_cfqq > 1)
                 return false;
@@ -3681,10 +3705,10 @@ new_queue:
         }
 
         cfqq->allocated[rw]++;
-        atomic_inc(&cfqq->ref);
 
         spin_unlock_irqrestore(q->queue_lock, flags);
 
+        cfqq->ref++;
         rq->elevator_private[0] = cic;
         rq->elevator_private[1] = cfqq;
         rq->elevator_private[2] = cfq_ref_get_cfqg(cfqq->cfqg);
@@ -3862,6 +3886,10 @@ static void *cfq_init_queue(struct request_queue *q)
         if (!cfqd)
                 return NULL;
 
+        /*
+         * Don't need take queue_lock in the routine, since we are
+         * initializing the ioscheduler, and nobody is using cfqd
+         */
         cfqd->cic_index = i;
 
         /* Init root service tree */
@@ -3881,7 +3909,7 @@ static void *cfq_init_queue(struct request_queue *q)
          * Take a reference to root group which we never drop. This is just
          * to make sure that cfq_put_cfqg() does not try to kfree root group
          */
-        atomic_set(&cfqg->ref, 1);
+        cfqg->ref = 1;
         rcu_read_lock();
         cfq_blkiocg_add_blkio_group(&blkio_root_cgroup, &cfqg->blkg,
                                         (void *)cfqd, 0);
@@ -3901,7 +3929,7 @@ static void *cfq_init_queue(struct request_queue *q)
          * will not attempt to free it.
          */
         cfq_init_cfqq(cfqd, &cfqd->oom_cfqq, 1, 0);
-        atomic_inc(&cfqd->oom_cfqq.ref);
+        cfqd->oom_cfqq.ref++;
         cfq_link_cfqq_cfqg(&cfqd->oom_cfqq, &cfqd->root_group);
 
         INIT_LIST_HEAD(&cfqd->cic_list);
diff --git a/block/genhd.c b/block/genhd.c
index 5fa2b44a72ff..6a5b772aa201 100644
--- a/block/genhd.c
+++ b/block/genhd.c
@@ -18,6 +18,7 @@
 #include <linux/buffer_head.h>
 #include <linux/mutex.h>
 #include <linux/idr.h>
+#include <linux/log2.h>
 
 #include "blk.h"
 
@@ -35,6 +36,10 @@ static DEFINE_IDR(ext_devt_idr);
 
 static struct device_type disk_type;
 
+static void disk_add_events(struct gendisk *disk);
+static void disk_del_events(struct gendisk *disk);
+static void disk_release_events(struct gendisk *disk);
+
 /**
  * disk_get_part - get partition
  * @disk: disk to look partition from
@@ -239,7 +244,7 @@ static struct blk_major_name {
 } *major_names[BLKDEV_MAJOR_HASH_SIZE];
 
 /* index in the above - for now: assume no multimajor ranges */
-static inline int major_to_index(int major)
+static inline int major_to_index(unsigned major)
 {
         return major % BLKDEV_MAJOR_HASH_SIZE;
 }
@@ -502,6 +507,64 @@ static int exact_lock(dev_t devt, void *data)
         return 0;
 }
 
+void register_disk(struct gendisk *disk)
+{
+        struct device *ddev = disk_to_dev(disk);
+        struct block_device *bdev;
+        struct disk_part_iter piter;
+        struct hd_struct *part;
+        int err;
+
+        ddev->parent = disk->driverfs_dev;
+
+        dev_set_name(ddev, disk->disk_name);
+
+        /* delay uevents, until we scanned partition table */
+        dev_set_uevent_suppress(ddev, 1);
+
+        if (device_add(ddev))
+                return;
+        if (!sysfs_deprecated) {
+                err = sysfs_create_link(block_depr, &ddev->kobj,
+                                        kobject_name(&ddev->kobj));
+                if (err) {
+                        device_del(ddev);
+                        return;
+                }
+        }
+        disk->part0.holder_dir = kobject_create_and_add("holders", &ddev->kobj);
+        disk->slave_dir = kobject_create_and_add("slaves", &ddev->kobj);
+
+        /* No minors to use for partitions */
+        if (!disk_partitionable(disk))
+                goto exit;
+
+        /* No such device (e.g., media were just removed) */
+        if (!get_capacity(disk))
+                goto exit;
+
+        bdev = bdget_disk(disk, 0);
+        if (!bdev)
+                goto exit;
+
+        bdev->bd_invalidated = 1;
+        err = blkdev_get(bdev, FMODE_READ, NULL);
+        if (err < 0)
+                goto exit;
+        blkdev_put(bdev, FMODE_READ);
+
+exit:
+        /* announce disk after possible partitions are created */
+        dev_set_uevent_suppress(ddev, 0);
+        kobject_uevent(&ddev->kobj, KOBJ_ADD);
+
+        /* announce possible partitions */
+        disk_part_iter_init(&piter, disk, 0);
+        while ((part = disk_part_iter_next(&piter)))
+                kobject_uevent(&part_to_dev(part)->kobj, KOBJ_ADD);
+        disk_part_iter_exit(&piter);
+}
+
 /**
  * add_disk - add partitioning information to kernel list
  * @disk: per-device partitioning information
@@ -551,18 +614,48 @@ void add_disk(struct gendisk *disk)
         retval = sysfs_create_link(&disk_to_dev(disk)->kobj, &bdi->dev->kobj,
                                    "bdi");
         WARN_ON(retval);
-}
 
+        disk_add_events(disk);
+}
 EXPORT_SYMBOL(add_disk);
-EXPORT_SYMBOL(del_gendisk);     /* in partitions/check.c */
 
-void unlink_gendisk(struct gendisk *disk)
+void del_gendisk(struct gendisk *disk)
 {
+        struct disk_part_iter piter;
+        struct hd_struct *part;
+
+        disk_del_events(disk);
+
+        /* invalidate stuff */
+        disk_part_iter_init(&piter, disk,
+                             DISK_PITER_INCL_EMPTY | DISK_PITER_REVERSE);
+        while ((part = disk_part_iter_next(&piter))) {
+                invalidate_partition(disk, part->partno);
+                delete_partition(disk, part->partno);
+        }
+        disk_part_iter_exit(&piter);
+
+        invalidate_partition(disk, 0);
+        blk_free_devt(disk_to_dev(disk)->devt);
+        set_capacity(disk, 0);
+        disk->flags &= ~GENHD_FL_UP;
+
         sysfs_remove_link(&disk_to_dev(disk)->kobj, "bdi");
         bdi_unregister(&disk->queue->backing_dev_info);
         blk_unregister_queue(disk);
         blk_unregister_region(disk_devt(disk), disk->minors);
+
+        part_stat_set_all(&disk->part0, 0);
+        disk->part0.stamp = 0;
+
+        kobject_put(disk->part0.holder_dir);
+        kobject_put(disk->slave_dir);
+        disk->driverfs_dev = NULL;
+        if (!sysfs_deprecated)
+                sysfs_remove_link(block_depr, dev_name(disk_to_dev(disk)));
+        device_del(disk_to_dev(disk));
 }
+EXPORT_SYMBOL(del_gendisk);
 
 /**
  * get_gendisk - get partitioning information for a given device
@@ -735,7 +828,7 @@ static void *show_partition_start(struct seq_file *seqf, loff_t *pos)
         static void *p;
 
         p = disk_seqf_start(seqf, pos);
-        if (!IS_ERR(p) && p && !*pos)
+        if (!IS_ERR_OR_NULL(p) && !*pos)
                 seq_puts(seqf, "major minor  #blocks  name\n\n");
         return p;
 }
@@ -1005,6 +1098,7 @@ static void disk_release(struct device *dev)
 {
         struct gendisk *disk = dev_to_disk(dev);
 
+        disk_release_events(disk);
         kfree(disk->random);
         disk_replace_part_tbl(disk, NULL);
         free_part_stats(&disk->part0);
@@ -1110,29 +1204,6 @@ static int __init proc_genhd_init(void)
 module_init(proc_genhd_init);
 #endif /* CONFIG_PROC_FS */
 
-static void media_change_notify_thread(struct work_struct *work)
-{
-        struct gendisk *gd = container_of(work, struct gendisk, async_notify);
-        char event[] = "MEDIA_CHANGE=1";
-        char *envp[] = { event, NULL };
-
-        /*
-         * set enviroment vars to indicate which event this is for
-         * so that user space will know to go check the media status.
-         */
-        kobject_uevent_env(&disk_to_dev(gd)->kobj, KOBJ_CHANGE, envp);
-        put_device(gd->driverfs_dev);
-}
-
-#if 0
-void genhd_media_change_notify(struct gendisk *disk)
-{
-        get_device(disk->driverfs_dev);
-        schedule_work(&disk->async_notify);
-}
-EXPORT_SYMBOL_GPL(genhd_media_change_notify);
-#endif  /*  0  */
-
 dev_t blk_lookup_devt(const char *name, int partno)
 {
         dev_t devt = MKDEV(0, 0);
@@ -1193,13 +1264,13 @@ struct gendisk *alloc_disk_node(int minors, int node_id)
                 }
                 disk->part_tbl->part[0] = &disk->part0;
 
+                hd_ref_init(&disk->part0);
+
                 disk->minors = minors;
                 rand_initialize_disk(disk);
                 disk_to_dev(disk)->class = &block_class;
                 disk_to_dev(disk)->type = &disk_type;
                 device_initialize(disk_to_dev(disk));
-                INIT_WORK(&disk->async_notify,
-                        media_change_notify_thread);
         }
         return disk;
 }
@@ -1291,3 +1362,422 @@ int invalidate_partition(struct gendisk *disk, int partno)
 }
 
 EXPORT_SYMBOL(invalidate_partition);
+
+/*
+ * Disk events - monitor disk events like media change and eject request.
+ */
+struct disk_events {
+        struct list_head        node;           /* all disk_event's */
+        struct gendisk          *disk;          /* the associated disk */
+        spinlock_t              lock;
+
+        int                     block;          /* event blocking depth */
+        unsigned int            pending;        /* events already sent out */
+        unsigned int            clearing;       /* events being cleared */
+
+        long                    poll_msecs;     /* interval, -1 for default */
+        struct delayed_work     dwork;
+};
+
+static const char *disk_events_strs[] = {
+        [ilog2(DISK_EVENT_MEDIA_CHANGE)]        = "media_change",
+        [ilog2(DISK_EVENT_EJECT_REQUEST)]       = "eject_request",
+};
+
+static char *disk_uevents[] = {
+        [ilog2(DISK_EVENT_MEDIA_CHANGE)]        = "DISK_MEDIA_CHANGE=1",
+        [ilog2(DISK_EVENT_EJECT_REQUEST)]       = "DISK_EJECT_REQUEST=1",
+};
+
+/* list of all disk_events */
+static DEFINE_MUTEX(disk_events_mutex);
+static LIST_HEAD(disk_events);
+
+/* disable in-kernel polling by default */
+static unsigned long disk_events_dfl_poll_msecs = 0;
+
+static unsigned long disk_events_poll_jiffies(struct gendisk *disk)
+{
+        struct disk_events *ev = disk->ev;
+        long intv_msecs = 0;
+
+        /*
+         * If device-specific poll interval is set, always use it.  If
+         * the default is being used, poll iff there are events which
+         * can't be monitored asynchronously.
+         */
+        if (ev->poll_msecs >= 0)
+                intv_msecs = ev->poll_msecs;
+        else if (disk->events & ~disk->async_events)
+                intv_msecs = disk_events_dfl_poll_msecs;
+
+        return msecs_to_jiffies(intv_msecs);
+}
+
+static void __disk_block_events(struct gendisk *disk, bool sync)
+{
+        struct disk_events *ev = disk->ev;
+        unsigned long flags;
+        bool cancel;
+
+        spin_lock_irqsave(&ev->lock, flags);
+        cancel = !ev->block++;
+        spin_unlock_irqrestore(&ev->lock, flags);
+
+        if (cancel) {
+                if (sync)
+                        cancel_delayed_work_sync(&disk->ev->dwork);
+                else
+                        cancel_delayed_work(&disk->ev->dwork);
+        }
+}
+
+static void __disk_unblock_events(struct gendisk *disk, bool check_now)
+{
+        struct disk_events *ev = disk->ev;
+        unsigned long intv;
+        unsigned long flags;
+
+        spin_lock_irqsave(&ev->lock, flags);
+
+        if (WARN_ON_ONCE(ev->block <= 0))
+                goto out_unlock;
+
+        if (--ev->block)
+                goto out_unlock;
+
+        /*
+         * Not exactly a latency critical operation, set poll timer
+         * slack to 25% and kick event check.
+         */
+        intv = disk_events_poll_jiffies(disk);
+        set_timer_slack(&ev->dwork.timer, intv / 4);
+        if (check_now)
+                queue_delayed_work(system_nrt_wq, &ev->dwork, 0);
+        else if (intv)
+                queue_delayed_work(system_nrt_wq, &ev->dwork, intv);
+out_unlock:
+        spin_unlock_irqrestore(&ev->lock, flags);
+}
+
+/**
+ * disk_block_events - block and flush disk event checking
+ * @disk: disk to block events for
+ *
+ * On return from this function, it is guaranteed that event checking
+ * isn't in progress and won't happen until unblocked by
+ * disk_unblock_events().  Events blocking is counted and the actual
+ * unblocking happens after the matching number of unblocks are done.
+ *
+ * Note that this intentionally does not block event checking from
+ * disk_clear_events().
+ *
+ * CONTEXT:
+ * Might sleep.
+ */
+void disk_block_events(struct gendisk *disk)
+{
+        if (disk->ev)
+                __disk_block_events(disk, true);
+}
+
+/**
+ * disk_unblock_events - unblock disk event checking
+ * @disk: disk to unblock events for
+ *
+ * Undo disk_block_events().  When the block count reaches zero, it
+ * starts events polling if configured.
+ *
+ * CONTEXT:
+ * Don't care.  Safe to call from irq context.
+ */
+void disk_unblock_events(struct gendisk *disk)
+{
+        if (disk->ev)
+                __disk_unblock_events(disk, true);
+}
+
+/**
+ * disk_check_events - schedule immediate event checking
+ * @disk: disk to check events for
+ *
+ * Schedule immediate event checking on @disk if not blocked.
+ *
+ * CONTEXT:
+ * Don't care.  Safe to call from irq context.
+ */
+void disk_check_events(struct gendisk *disk)
+{
+        if (disk->ev) {
+                __disk_block_events(disk, false);
+                __disk_unblock_events(disk, true);
+        }
+}
+EXPORT_SYMBOL_GPL(disk_check_events);
+
+/**
+ * disk_clear_events - synchronously check, clear and return pending events
+ * @disk: disk to fetch and clear events from
+ * @mask: mask of events to be fetched and clearted
+ *
+ * Disk events are synchronously checked and pending events in @mask
+ * are cleared and returned.  This ignores the block count.
+ *
+ * CONTEXT:
+ * Might sleep.
+ */
+unsigned int disk_clear_events(struct gendisk *disk, unsigned int mask)
+{
+        const struct block_device_operations *bdops = disk->fops;
+        struct disk_events *ev = disk->ev;
+        unsigned int pending;
+
+        if (!ev) {
+                /* for drivers still using the old ->media_changed method */
+                if ((mask & DISK_EVENT_MEDIA_CHANGE) &&
+                    bdops->media_changed && bdops->media_changed(disk))
+                        return DISK_EVENT_MEDIA_CHANGE;
+                return 0;
+        }
+
+        /* tell the workfn about the events being cleared */
+        spin_lock_irq(&ev->lock);
+        ev->clearing |= mask;
+        spin_unlock_irq(&ev->lock);
+
+        /* uncondtionally schedule event check and wait for it to finish */
+        __disk_block_events(disk, true);
+        queue_delayed_work(system_nrt_wq, &ev->dwork, 0);
+        flush_delayed_work(&ev->dwork);
+        __disk_unblock_events(disk, false);
+
+        /* then, fetch and clear pending events */
+        spin_lock_irq(&ev->lock);
+        WARN_ON_ONCE(ev->clearing & mask);      /* cleared by workfn */
+        pending = ev->pending & mask;
+        ev->pending &= ~mask;
+        spin_unlock_irq(&ev->lock);
+
+        return pending;
+}
+
+static void disk_events_workfn(struct work_struct *work)
+{
+        struct delayed_work *dwork = to_delayed_work(work);
+        struct disk_events *ev = container_of(dwork, struct disk_events, dwork);
+        struct gendisk *disk = ev->disk;
+        char *envp[ARRAY_SIZE(disk_uevents) + 1] = { };
+        unsigned int clearing = ev->clearing;
+        unsigned int events;
+        unsigned long intv;
+        int nr_events = 0, i;
+
+        /* check events */
+        events = disk->fops->check_events(disk, clearing);
+
+        /* accumulate pending events and schedule next poll if necessary */
+        spin_lock_irq(&ev->lock);
+
+        events &= ~ev->pending;
+        ev->pending |= events;
+        ev->clearing &= ~clearing;
+
+        intv = disk_events_poll_jiffies(disk);
+        if (!ev->block && intv)
+                queue_delayed_work(system_nrt_wq, &ev->dwork, intv);
+
+        spin_unlock_irq(&ev->lock);
+
+        /* tell userland about new events */
+        for (i = 0; i < ARRAY_SIZE(disk_uevents); i++)
+                if (events & (1 << i))
+                        envp[nr_events++] = disk_uevents[i];
+
+        if (nr_events)
+                kobject_uevent_env(&disk_to_dev(disk)->kobj, KOBJ_CHANGE, envp);
+}
+
+/*
+ * A disk events enabled device has the following sysfs nodes under
+ * its /sys/block/X/ directory.
+ *
+ * events               : list of all supported events
+ * events_async         : list of events which can be detected w/o polling
+ * events_poll_msecs    : polling interval, 0: disable, -1: system default
+ */
+static ssize_t __disk_events_show(unsigned int events, char *buf)
+{
+        const char *delim = "";
+        ssize_t pos = 0;
+        int i;
+
+        for (i = 0; i < ARRAY_SIZE(disk_events_strs); i++)
+                if (events & (1 << i)) {
+                        pos += sprintf(buf + pos, "%s%s",
+                                       delim, disk_events_strs[i]);
+                        delim = " ";
+                }
+        if (pos)
+                pos += sprintf(buf + pos, "\n");
+        return pos;
+}
+
+static ssize_t disk_events_show(struct device *dev,
+                                struct device_attribute *attr, char *buf)
+{
+        struct gendisk *disk = dev_to_disk(dev);
+
+        return __disk_events_show(disk->events, buf);
+}
+
+static ssize_t disk_events_async_show(struct device *dev,
+                                      struct device_attribute *attr, char *buf)
+{
+        struct gendisk *disk = dev_to_disk(dev);
+
+        return __disk_events_show(disk->async_events, buf);
+}
+
+static ssize_t disk_events_poll_msecs_show(struct device *dev,
+                                           struct device_attribute *attr,
+                                           char *buf)
+{
+        struct gendisk *disk = dev_to_disk(dev);
+
+        return sprintf(buf, "%ld\n", disk->ev->poll_msecs);
+}
+
+static ssize_t disk_events_poll_msecs_store(struct device *dev,
+                                            struct device_attribute *attr,
+                                            const char *buf, size_t count)
+{
+        struct gendisk *disk = dev_to_disk(dev);
+        long intv;
+
+        if (!count || !sscanf(buf, "%ld", &intv))
+                return -EINVAL;
+
+        if (intv < 0 && intv != -1)
+                return -EINVAL;
+
+        __disk_block_events(disk, true);
+        disk->ev->poll_msecs = intv;
+        __disk_unblock_events(disk, true);
+
+        return count;
+}
+
+static const DEVICE_ATTR(events, S_IRUGO, disk_events_show, NULL);
+static const DEVICE_ATTR(events_async, S_IRUGO, disk_events_async_show, NULL);
+static const DEVICE_ATTR(events_poll_msecs, S_IRUGO|S_IWUSR,
+                         disk_events_poll_msecs_show,
+                         disk_events_poll_msecs_store);
+
+static const struct attribute *disk_events_attrs[] = {
+        &dev_attr_events.attr,
+        &dev_attr_events_async.attr,
+        &dev_attr_events_poll_msecs.attr,
+        NULL,
+};
+
+/*
+ * The default polling interval can be specified by the kernel
+ * parameter block.events_dfl_poll_msecs which defaults to 0
+ * (disable).  This can also be modified runtime by writing to
+ * /sys/module/block/events_dfl_poll_msecs.
+ */
+static int disk_events_set_dfl_poll_msecs(const char *val,
+                                          const struct kernel_param *kp)
+{
+        struct disk_events *ev;
+        int ret;
+
+        ret = param_set_ulong(val, kp);
+        if (ret < 0)
+                return ret;
+
+        mutex_lock(&disk_events_mutex);
+
+        list_for_each_entry(ev, &disk_events, node)
+                disk_check_events(ev->disk);
+
+        mutex_unlock(&disk_events_mutex);
+
+        return 0;
+}
+
+static const struct kernel_param_ops disk_events_dfl_poll_msecs_param_ops = {
+        .set    = disk_events_set_dfl_poll_msecs,
+        .get    = param_get_ulong,
+};
+
+#undef MODULE_PARAM_PREFIX
+#define MODULE_PARAM_PREFIX     "block."
+
+module_param_cb(events_dfl_poll_msecs, &disk_events_dfl_poll_msecs_param_ops,
+                &disk_events_dfl_poll_msecs, 0644);
+
+/*
+ * disk_{add|del|release}_events - initialize and destroy disk_events.
+ */
+static void disk_add_events(struct gendisk *disk)
+{
+        struct disk_events *ev;
+
+        if (!disk->fops->check_events || !(disk->events | disk->async_events))
+                return;
+
+        ev = kzalloc(sizeof(*ev), GFP_KERNEL);
+        if (!ev) {
+                pr_warn("%s: failed to initialize events\n", disk->disk_name);
+                return;
+        }
+
+        if (sysfs_create_files(&disk_to_dev(disk)->kobj,
+                               disk_events_attrs) < 0) {
+                pr_warn("%s: failed to create sysfs files for events\n",
+                        disk->disk_name);
+                kfree(ev);
+                return;
+        }
+
+        disk->ev = ev;
+
+        INIT_LIST_HEAD(&ev->node);
+        ev->disk = disk;
+        spin_lock_init(&ev->lock);
+        ev->block = 1;
+        ev->poll_msecs = -1;
+        INIT_DELAYED_WORK(&ev->dwork, disk_events_workfn);
+
+        mutex_lock(&disk_events_mutex);
+        list_add_tail(&ev->node, &disk_events);
+        mutex_unlock(&disk_events_mutex);
+
+        /*
+         * Block count is initialized to 1 and the following initial
+         * unblock kicks it into action.
+         */
+        __disk_unblock_events(disk, true);
+}
+
+static void disk_del_events(struct gendisk *disk)
+{
+        if (!disk->ev)
+                return;
+
+        __disk_block_events(disk, true);
+
+        mutex_lock(&disk_events_mutex);
+        list_del_init(&disk->ev->node);
+        mutex_unlock(&disk_events_mutex);
+
+        sysfs_remove_files(&disk_to_dev(disk)->kobj, disk_events_attrs);
+}
+
+static void disk_release_events(struct gendisk *disk)
+{
+        /* the block count should be 1 from disk_del_events() */
+        WARN_ON_ONCE(disk->ev && disk->ev->block != 1);
+        kfree(disk->ev);
+}
diff --git a/block/ioctl.c b/block/ioctl.c
index a9a302eba01e..9049d460fa89 100644
--- a/block/ioctl.c
+++ b/block/ioctl.c
@@ -294,11 +294,12 @@ int blkdev_ioctl(struct block_device *bdev, fmode_t mode, unsigned cmd,
                         return -EINVAL;
                 if (get_user(n, (int __user *) arg))
                         return -EFAULT;
-                if (!(mode & FMODE_EXCL) && bd_claim(bdev, &bdev) < 0)
+                if (!(mode & FMODE_EXCL) &&
+                    blkdev_get(bdev, mode | FMODE_EXCL, &bdev) < 0)
                         return -EBUSY;
                 ret = set_blocksize(bdev, n);
                 if (!(mode & FMODE_EXCL))
-                        bd_release(bdev);
+                        blkdev_put(bdev, mode | FMODE_EXCL);
                 return ret;
         case BLKPG:
                 ret = blkpg_ioctl(bdev, (struct blkpg_ioctl_arg __user *) arg);