7 files changed, 93 insertions, 39 deletions

diff --git a/block/blk-core.c b/block/blk-core.c
index 2987fe47b5ee..1905aaba49fb 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -482,6 +482,7 @@ struct request_queue *blk_alloc_queue_node(gfp_t gfp_mask, int node_id)
         kobject_init(&q->kobj, &blk_queue_ktype);
 
         mutex_init(&q->sysfs_lock);
+        spin_lock_init(&q->__queue_lock);
 
         return q;
 }
@@ -544,10 +545,8 @@ blk_init_queue_node(request_fn_proc *rfn, spinlock_t *lock, int node_id)
          * if caller didn't supply a lock, they get per-queue locking with
          * our embedded lock
          */
-        if (!lock) {
-                spin_lock_init(&q->__queue_lock);
+        if (!lock)
                 lock = &q->__queue_lock;
-        }
 
         q->request_fn           = rfn;
         q->prep_rq_fn           = NULL;
@@ -807,35 +806,32 @@ static struct request *get_request_wait(struct request_queue *q, int rw_flags,
         rq = get_request(q, rw_flags, bio, GFP_NOIO);
         while (!rq) {
                 DEFINE_WAIT(wait);
+                struct io_context *ioc;
                 struct request_list *rl = &q->rq;
 
                 prepare_to_wait_exclusive(&rl->wait[rw], &wait,
                                 TASK_UNINTERRUPTIBLE);
 
-                rq = get_request(q, rw_flags, bio, GFP_NOIO);
-
-                if (!rq) {
-                        struct io_context *ioc;
-
-                        blk_add_trace_generic(q, bio, rw, BLK_TA_SLEEPRQ);
+                blk_add_trace_generic(q, bio, rw, BLK_TA_SLEEPRQ);
 
-                        __generic_unplug_device(q);
-                        spin_unlock_irq(q->queue_lock);
-                        io_schedule();
+                __generic_unplug_device(q);
+                spin_unlock_irq(q->queue_lock);
+                io_schedule();
 
-                        /*
-                         * After sleeping, we become a "batching" process and
-                         * will be able to allocate at least one request, and
-                         * up to a big batch of them for a small period time.
-                         * See ioc_batching, ioc_set_batching
-                         */
-                        ioc = current_io_context(GFP_NOIO, q->node);
-                        ioc_set_batching(q, ioc);
+                /*
+                 * After sleeping, we become a "batching" process and
+                 * will be able to allocate at least one request, and
+                 * up to a big batch of them for a small period time.
+                 * See ioc_batching, ioc_set_batching
+                 */
+                ioc = current_io_context(GFP_NOIO, q->node);
+                ioc_set_batching(q, ioc);
 
-                        spin_lock_irq(q->queue_lock);
-                }
+                spin_lock_irq(q->queue_lock);
                 finish_wait(&rl->wait[rw], &wait);
-        }
+
+                rq = get_request(q, rw_flags, bio, GFP_NOIO);
+        };
 
         return rq;
 }
diff --git a/block/blk-settings.c b/block/blk-settings.c
index bb93d4c32775..8dd86418f35d 100644
--- a/block/blk-settings.c
+++ b/block/blk-settings.c
@@ -286,8 +286,14 @@ void blk_queue_stack_limits(struct request_queue *t, struct request_queue *b)
         t->max_hw_segments = min(t->max_hw_segments, b->max_hw_segments);
         t->max_segment_size = min(t->max_segment_size, b->max_segment_size);
         t->hardsect_size = max(t->hardsect_size, b->hardsect_size);
-        if (!test_bit(QUEUE_FLAG_CLUSTER, &b->queue_flags))
+        if (!t->queue_lock)
+                WARN_ON_ONCE(1);
+        else if (!test_bit(QUEUE_FLAG_CLUSTER, &b->queue_flags)) {
+                unsigned long flags;
+                spin_lock_irqsave(t->queue_lock, flags);
                 queue_flag_clear(QUEUE_FLAG_CLUSTER, t);
+                spin_unlock_irqrestore(t->queue_lock, flags);
+        }
 }
 EXPORT_SYMBOL(blk_queue_stack_limits);
 
diff --git a/block/blktrace.c b/block/blktrace.c
index 568588cd16b2..8d3a27780260 100644
--- a/block/blktrace.c
+++ b/block/blktrace.c
@@ -75,6 +75,24 @@ static void trace_note_time(struct blk_trace *bt)
         local_irq_restore(flags);
 }
 
+void __trace_note_message(struct blk_trace *bt, const char *fmt, ...)
+{
+        int n;
+        va_list args;
+        unsigned long flags;
+        char *buf;
+
+        local_irq_save(flags);
+        buf = per_cpu_ptr(bt->msg_data, smp_processor_id());
+        va_start(args, fmt);
+        n = vscnprintf(buf, BLK_TN_MAX_MSG, fmt, args);
+        va_end(args);
+
+        trace_note(bt, 0, BLK_TN_MESSAGE, buf, n);
+        local_irq_restore(flags);
+}
+EXPORT_SYMBOL_GPL(__trace_note_message);
+
 static int act_log_check(struct blk_trace *bt, u32 what, sector_t sector,
                          pid_t pid)
 {
@@ -141,10 +159,7 @@ void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes,
         /*
          * A word about the locking here - we disable interrupts to reserve
          * some space in the relay per-cpu buffer, to prevent an irq
-         * from coming in and stepping on our toes. Once reserved, it's
-         * enough to get preemption disabled to prevent read of this data
-         * before we are through filling it. get_cpu()/put_cpu() does this
-         * for us
+         * from coming in and stepping on our toes.
          */
         local_irq_save(flags);
 
@@ -232,6 +247,7 @@ static void blk_trace_cleanup(struct blk_trace *bt)
         debugfs_remove(bt->dropped_file);
         blk_remove_tree(bt->dir);
         free_percpu(bt->sequence);
+        free_percpu(bt->msg_data);
         kfree(bt);
 }
 
@@ -346,6 +362,10 @@ int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev,
         if (!bt->sequence)
                 goto err;
 
+        bt->msg_data = __alloc_percpu(BLK_TN_MAX_MSG);
+        if (!bt->msg_data)
+                goto err;
+
         ret = -ENOENT;
         dir = blk_create_tree(buts->name);
         if (!dir)
@@ -392,6 +412,7 @@ err:
                 if (bt->dropped_file)
                         debugfs_remove(bt->dropped_file);
                 free_percpu(bt->sequence);
+                free_percpu(bt->msg_data);
                 if (bt->rchan)
                         relay_close(bt->rchan);
                 kfree(bt);
@@ -476,7 +497,7 @@ int blk_trace_ioctl(struct block_device *bdev, unsigned cmd, char __user *arg)
 
         switch (cmd) {
         case BLKTRACESETUP:
-                strcpy(b, bdevname(bdev, b));
+                bdevname(bdev, b);
                 ret = blk_trace_setup(q, b, bdev->bd_dev, arg);
                 break;
         case BLKTRACESTART:
diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
index b399c62936e0..d01b411c72f0 100644
--- a/block/cfq-iosched.c
+++ b/block/cfq-iosched.c
@@ -124,6 +124,8 @@ struct cfq_data {
 struct cfq_queue {
         /* reference count */
         atomic_t ref;
+        /* various state flags, see below */
+        unsigned int flags;
         /* parent cfq_data */
         struct cfq_data *cfqd;
         /* service_tree member */
@@ -138,14 +140,14 @@ struct cfq_queue {
         int queued[2];
         /* currently allocated requests */
         int allocated[2];
-        /* pending metadata requests */
-        int meta_pending;
         /* fifo list of requests in sort_list */
         struct list_head fifo;
 
         unsigned long slice_end;
         long slice_resid;
 
+        /* pending metadata requests */
+        int meta_pending;
         /* number of requests that are on the dispatch list or inside driver */
         int dispatched;
 
@@ -153,8 +155,6 @@ struct cfq_queue {
         unsigned short ioprio, org_ioprio;
         unsigned short ioprio_class, org_ioprio_class;
 
-        /* various state flags, see below */
-        unsigned int flags;
 };
 
 enum cfqq_state_flags {
@@ -1142,6 +1142,9 @@ static void cfq_put_queue(struct cfq_queue *cfqq)
         kmem_cache_free(cfq_pool, cfqq);
 }
 
+/*
+ * Must always be called with the rcu_read_lock() held
+ */
 static void
 __call_for_each_cic(struct io_context *ioc,
                     void (*func)(struct io_context *, struct cfq_io_context *))
@@ -1197,6 +1200,11 @@ static void cic_free_func(struct io_context *ioc, struct cfq_io_context *cic)
         cfq_cic_free(cic);
 }
 
+/*
+ * Must be called with rcu_read_lock() held or preemption otherwise disabled.
+ * Only two callers of this - ->dtor() which is called with the rcu_read_lock(),
+ * and ->trim() which is called with the task lock held
+ */
 static void cfq_free_io_context(struct io_context *ioc)
 {
         /*
@@ -1502,20 +1510,24 @@ static struct cfq_io_context *
 cfq_cic_lookup(struct cfq_data *cfqd, struct io_context *ioc)
 {
         struct cfq_io_context *cic;
+        unsigned long flags;
         void *k;
 
         if (unlikely(!ioc))
                 return NULL;
 
+        rcu_read_lock();
+
         /*
          * we maintain a last-hit cache, to avoid browsing over the tree
          */
         cic = rcu_dereference(ioc->ioc_data);
-        if (cic && cic->key == cfqd)
+        if (cic && cic->key == cfqd) {
+                rcu_read_unlock();
                 return cic;
+        }
 
         do {
-                rcu_read_lock();
                 cic = radix_tree_lookup(&ioc->radix_root, (unsigned long) cfqd);
                 rcu_read_unlock();
                 if (!cic)
@@ -1524,10 +1536,13 @@ cfq_cic_lookup(struct cfq_data *cfqd, struct io_context *ioc)
                 k = cic->key;
                 if (unlikely(!k)) {
                         cfq_drop_dead_cic(cfqd, ioc, cic);
+                        rcu_read_lock();
                         continue;
                 }
 
+                spin_lock_irqsave(&ioc->lock, flags);
                 rcu_assign_pointer(ioc->ioc_data, cic);
+                spin_unlock_irqrestore(&ioc->lock, flags);
                 break;
         } while (1);
 
@@ -2134,6 +2149,10 @@ static void *cfq_init_queue(struct request_queue *q)
 
 static void cfq_slab_kill(void)
 {
+        /*
+         * Caller already ensured that pending RCU callbacks are completed,
+         * so we should have no busy allocations at this point.
+         */
         if (cfq_pool)
                 kmem_cache_destroy(cfq_pool);
         if (cfq_ioc_pool)
@@ -2292,6 +2311,11 @@ static void __exit cfq_exit(void)
         ioc_gone = &all_gone;
         /* ioc_gone's update must be visible before reading ioc_count */
         smp_wmb();
+
+        /*
+         * this also protects us from entering cfq_slab_kill() with
+         * pending RCU callbacks
+         */
         if (elv_ioc_count_read(ioc_count))
                 wait_for_completion(ioc_gone);
         cfq_slab_kill();
diff --git a/block/compat_ioctl.c b/block/compat_ioctl.c
index c70d0b6f666f..c23177e4623f 100644
--- a/block/compat_ioctl.c
+++ b/block/compat_ioctl.c
@@ -555,7 +555,7 @@ static int compat_blk_trace_setup(struct block_device *bdev, char __user *arg)
         if (copy_from_user(&cbuts, arg, sizeof(cbuts)))
                 return -EFAULT;
 
-        strcpy(b, bdevname(bdev, b));
+        bdevname(bdev, b);
 
         buts = (struct blk_user_trace_setup) {
                 .act_mask = cbuts.act_mask,
diff --git a/block/elevator.c b/block/elevator.c
index 980f8ae147b4..902dd1344d56 100644
--- a/block/elevator.c
+++ b/block/elevator.c
@@ -1110,6 +1110,8 @@ static int elevator_switch(struct request_queue *q, struct elevator_type *new_e)
         queue_flag_clear(QUEUE_FLAG_ELVSWITCH, q);
         spin_unlock_irq(q->queue_lock);
 
+        blk_add_trace_msg(q, "elv switch: %s", e->elevator_type->elevator_name);
+
         return 1;
 
 fail_register:
diff --git a/block/genhd.c b/block/genhd.c
index fda9c7a63c29..b922d4801c87 100644
--- a/block/genhd.c
+++ b/block/genhd.c
@@ -653,15 +653,21 @@ void genhd_media_change_notify(struct gendisk *disk)
 EXPORT_SYMBOL_GPL(genhd_media_change_notify);
 #endif  /*  0  */
 
-dev_t blk_lookup_devt(const char *name)
+dev_t blk_lookup_devt(const char *name, int part)
 {
         struct device *dev;
         dev_t devt = MKDEV(0, 0);
 
         mutex_lock(&block_class_lock);
         list_for_each_entry(dev, &block_class.devices, node) {
+                if (dev->type != &disk_type)
+                        continue;
                 if (strcmp(dev->bus_id, name) == 0) {
-                        devt = dev->devt;
+                        struct gendisk *disk = dev_to_disk(dev);
+
+                        if (part < disk->minors)
+                                devt = MKDEV(MAJOR(dev->devt),
+                                             MINOR(dev->devt) + part);
                         break;
                 }
         }
@@ -669,7 +675,6 @@ dev_t blk_lookup_devt(const char *name)
 
         return devt;
 }
-
 EXPORT_SYMBOL(blk_lookup_devt);
 
 struct gendisk *alloc_disk(int minors)