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

Pull block fixes from Jens Axboe: "Been sitting on this for a while, but lets get this out the door. This fixes various important bugs for 3.3 final, along with a few more trivial ones. Please pull!" * 'for-linus' of git://git.kernel.dk/linux-block: block: fix ioc leak in put_io_context block, sx8: fix pointer math issue getting fw version Block: use a freezable workqueue for disk-event polling drivers/block/DAC960: fix -Wuninitialized warning drivers/block/DAC960: fix DAC960_V2_IOCTL_Opcode_T -Wenum-compare warning block: fix __blkdev_get and add_disk race condition block: Fix setting bio flags in drivers (sd_dif/floppy) block: Fix NULL pointer dereference in sd_revalidate_disk block: exit_io_context() should call elevator_exit_icq_fn() block: simplify ioc_release_fn() block: replace icq->changed with icq->flags

Merge branch 'for-linus' of git://git.kernel.dk/linux-block
Pull block fixes from Jens Axboe: "Been sitting on this for a while, but lets get this out the door. This fixes various important bugs for 3.3 final, along with a few more trivial ones. Please pull!" * 'for-linus' of git://git.kernel.dk/linux-block: block: fix ioc leak in put_io_context block, sx8: fix pointer math issue getting fw version Block: use a freezable workqueue for disk-event polling drivers/block/DAC960: fix -Wuninitialized warning drivers/block/DAC960: fix DAC960_V2_IOCTL_Opcode_T -Wenum-compare warning block: fix __blkdev_get and add_disk race condition block: Fix setting bio flags in drivers (sd_dif/floppy) block: Fix NULL pointer dereference in sd_revalidate_disk block: exit_io_context() should call elevator_exit_icq_fn() block: simplify ioc_release_fn() block: replace icq->changed with icq->flags
Linus Torvalds
2 parents ff398c45b0 ff8c1474cc
Showing 11 changed files Side-by-side Diff
block/blk-ioc.c
block/cfq-iosched.c
block/genhd.c
block/partition-generic.c
drivers/block/DAC960.c
drivers/block/sx8.c
fs/block_dev.c
include/linux/genhd.h
include/linux/iocontext.h
include/linux/workqueue.h
kernel/workqueue.c
@@ -36,11 +36,23 @@
 	kmem_cache_free(icq->__rcu_icq_cache, icq);
 }
  
-/*
- * Exit and free an icq.  Called with both ioc and q locked.
- */
+/* Exit an icq. Called with both ioc and q locked. */
 static void ioc_exit_icq(struct io_cq *icq)
 {
+	struct elevator_type *et = icq->q->elevator->type;
+
+	if (icq->flags & ICQ_EXITED)
+		return;
+
+	if (et->ops.elevator_exit_icq_fn)
+		et->ops.elevator_exit_icq_fn(icq);
+
+	icq->flags |= ICQ_EXITED;
+}
+
+/* Release an icq.  Called with both ioc and q locked. */
+static void ioc_destroy_icq(struct io_cq *icq)
+{
 	struct io_context *ioc = icq->ioc;
 	struct request_queue *q = icq->q;
 	struct elevator_type *et = q->elevator->type;
@@ -60,8 +72,7 @@
 	if (rcu_dereference_raw(ioc->icq_hint) == icq)
 		rcu_assign_pointer(ioc->icq_hint, NULL);
  
-	if (et->ops.elevator_exit_icq_fn)
-		et->ops.elevator_exit_icq_fn(icq);
+	ioc_exit_icq(icq);
  
 	/*
 	 * @icq->q might have gone away by the time RCU callback runs
@@ -79,7 +90,6 @@
 {
 	struct io_context *ioc = container_of(work, struct io_context,
 					      release_work);
-	struct request_queue *last_q = NULL;
 	unsigned long flags;
  
 	/*
  
  
  
@@ -93,44 +103,19 @@
 	while (!hlist_empty(&ioc->icq_list)) {
 		struct io_cq *icq = hlist_entry(ioc->icq_list.first,
 						struct io_cq, ioc_node);
-		struct request_queue *this_q = icq->q;
+		struct request_queue *q = icq->q;
  
-		if (this_q != last_q) {
-			/*
-			 * Need to switch to @this_q.  Once we release
-			 * @ioc->lock, it can go away along with @cic.
-			 * Hold on to it.
-			 */
-			__blk_get_queue(this_q);
-
-			/*
-			 * blk_put_queue() might sleep thanks to kobject
-			 * idiocy.  Always release both locks, put and
-			 * restart.
-			 */
-			if (last_q) {
-				spin_unlock(last_q->queue_lock);
-				spin_unlock_irqrestore(&ioc->lock, flags);
-				blk_put_queue(last_q);
-			} else {
-				spin_unlock_irqrestore(&ioc->lock, flags);
-			}
-
-			last_q = this_q;
-			spin_lock_irqsave(this_q->queue_lock, flags);
-			spin_lock_nested(&ioc->lock, 1);
-			continue;
+		if (spin_trylock(q->queue_lock)) {
+			ioc_destroy_icq(icq);
+			spin_unlock(q->queue_lock);
+		} else {
+			spin_unlock_irqrestore(&ioc->lock, flags);
+			cpu_relax();
+			spin_lock_irqsave_nested(&ioc->lock, flags, 1);
 		}
-		ioc_exit_icq(icq);
 	}
  
-	if (last_q) {
-		spin_unlock(last_q->queue_lock);
-		spin_unlock_irqrestore(&ioc->lock, flags);
-		blk_put_queue(last_q);
-	} else {
-		spin_unlock_irqrestore(&ioc->lock, flags);
-	}
+	spin_unlock_irqrestore(&ioc->lock, flags);
  
 	kmem_cache_free(iocontext_cachep, ioc);
 }
@@ -145,6 +130,7 @@
 void put_io_context(struct io_context *ioc)
 {
 	unsigned long flags;
+	bool free_ioc = false;
  
 	if (ioc == NULL)
 		return;
  
@@ -159,8 +145,13 @@
 		spin_lock_irqsave(&ioc->lock, flags);
 		if (!hlist_empty(&ioc->icq_list))
 			schedule_work(&ioc->release_work);
+		else
+			free_ioc = true;
 		spin_unlock_irqrestore(&ioc->lock, flags);
 	}
+
+	if (free_ioc)
+		kmem_cache_free(iocontext_cachep, ioc);
 }
 EXPORT_SYMBOL(put_io_context);
  
  
@@ -168,13 +159,41 @@
 void exit_io_context(struct task_struct *task)
 {
 	struct io_context *ioc;
+	struct io_cq *icq;
+	struct hlist_node *n;
+	unsigned long flags;
  
 	task_lock(task);
 	ioc = task->io_context;
 	task->io_context = NULL;
 	task_unlock(task);
  
-	atomic_dec(&ioc->nr_tasks);
+	if (!atomic_dec_and_test(&ioc->nr_tasks)) {
+		put_io_context(ioc);
+		return;
+	}
+
+	/*
+	 * Need ioc lock to walk icq_list and q lock to exit icq.  Perform
+	 * reverse double locking.  Read comment in ioc_release_fn() for
+	 * explanation on the nested locking annotation.
+	 */
+retry:
+	spin_lock_irqsave_nested(&ioc->lock, flags, 1);
+	hlist_for_each_entry(icq, n, &ioc->icq_list, ioc_node) {
+		if (icq->flags & ICQ_EXITED)
+			continue;
+		if (spin_trylock(icq->q->queue_lock)) {
+			ioc_exit_icq(icq);
+			spin_unlock(icq->q->queue_lock);
+		} else {
+			spin_unlock_irqrestore(&ioc->lock, flags);
+			cpu_relax();
+			goto retry;
+		}
+	}
+	spin_unlock_irqrestore(&ioc->lock, flags);
+
 	put_io_context(ioc);
 }
  
@@ -194,7 +213,7 @@
 		struct io_context *ioc = icq->ioc;
  
 		spin_lock(&ioc->lock);
-		ioc_exit_icq(icq);
+		ioc_destroy_icq(icq);
 		spin_unlock(&ioc->lock);
 	}
 }
  
@@ -363,13 +382,13 @@
 	return icq;
 }
  
-void ioc_set_changed(struct io_context *ioc, int which)
+void ioc_set_icq_flags(struct io_context *ioc, unsigned int flags)
 {
 	struct io_cq *icq;
 	struct hlist_node *n;
  
 	hlist_for_each_entry(icq, n, &ioc->icq_list, ioc_node)
-		set_bit(which, &icq->changed);
+		icq->flags |= flags;
 }
  
 /**
@@ -387,7 +406,7 @@
  
 	spin_lock_irqsave(&ioc->lock, flags);
 	ioc->ioprio = ioprio;
-	ioc_set_changed(ioc, ICQ_IOPRIO_CHANGED);
+	ioc_set_icq_flags(ioc, ICQ_IOPRIO_CHANGED);
 	spin_unlock_irqrestore(&ioc->lock, flags);
 }
  
  
@@ -404,10 +423,32 @@
 	unsigned long flags;
  
 	spin_lock_irqsave(&ioc->lock, flags);
-	ioc_set_changed(ioc, ICQ_CGROUP_CHANGED);
+	ioc_set_icq_flags(ioc, ICQ_CGROUP_CHANGED);
 	spin_unlock_irqrestore(&ioc->lock, flags);
 }
 EXPORT_SYMBOL(ioc_cgroup_changed);
+
+/**
+ * icq_get_changed - fetch and clear icq changed mask
+ * @icq: icq of interest
+ *
+ * Fetch and clear ICQ_*_CHANGED bits from @icq.  Grabs and releases
+ * @icq->ioc->lock.
+ */
+unsigned icq_get_changed(struct io_cq *icq)
+{
+	unsigned int changed = 0;
+	unsigned long flags;
+
+	if (unlikely(icq->flags & ICQ_CHANGED_MASK)) {
+		spin_lock_irqsave(&icq->ioc->lock, flags);
+		changed = icq->flags & ICQ_CHANGED_MASK;
+		icq->flags &= ~ICQ_CHANGED_MASK;
+		spin_unlock_irqrestore(&icq->ioc->lock, flags);
+	}
+	return changed;
+}
+EXPORT_SYMBOL(icq_get_changed);
  
 static int __init blk_ioc_init(void)
 {
@@ -3470,20 +3470,20 @@
 	const int rw = rq_data_dir(rq);
 	const bool is_sync = rq_is_sync(rq);
 	struct cfq_queue *cfqq;
+	unsigned int changed;
  
 	might_sleep_if(gfp_mask & __GFP_WAIT);
  
 	spin_lock_irq(q->queue_lock);
  
 	/* handle changed notifications */
-	if (unlikely(cic->icq.changed)) {
-		if (test_and_clear_bit(ICQ_IOPRIO_CHANGED, &cic->icq.changed))
-			changed_ioprio(cic);
+	changed = icq_get_changed(&cic->icq);
+	if (unlikely(changed & ICQ_IOPRIO_CHANGED))
+		changed_ioprio(cic);
 #ifdef CONFIG_CFQ_GROUP_IOSCHED
-		if (test_and_clear_bit(ICQ_CGROUP_CHANGED, &cic->icq.changed))
-			changed_cgroup(cic);
+	if (unlikely(changed & ICQ_CGROUP_CHANGED))
+		changed_cgroup(cic);
 #endif
-	}
  
 new_queue:
 	cfqq = cic_to_cfqq(cic, is_sync);
@@ -35,6 +35,7 @@
  
 static struct device_type disk_type;
  
+static void disk_alloc_events(struct gendisk *disk);
 static void disk_add_events(struct gendisk *disk);
 static void disk_del_events(struct gendisk *disk);
 static void disk_release_events(struct gendisk *disk);
@@ -601,6 +602,8 @@
 	disk->major = MAJOR(devt);
 	disk->first_minor = MINOR(devt);
  
+	disk_alloc_events(disk);
+
 	/* Register BDI before referencing it from bdev */
 	bdi = &disk->queue->backing_dev_info;
 	bdi_register_dev(bdi, disk_devt(disk));
  
@@ -1475,9 +1478,9 @@
 	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);
+		queue_delayed_work(system_nrt_freezable_wq, &ev->dwork, 0);
 	else if (intv)
-		queue_delayed_work(system_nrt_wq, &ev->dwork, intv);
+		queue_delayed_work(system_nrt_freezable_wq, &ev->dwork, intv);
 out_unlock:
 	spin_unlock_irqrestore(&ev->lock, flags);
 }
@@ -1521,7 +1524,7 @@
 	ev->clearing |= mask;
 	if (!ev->block) {
 		cancel_delayed_work(&ev->dwork);
-		queue_delayed_work(system_nrt_wq, &ev->dwork, 0);
+		queue_delayed_work(system_nrt_freezable_wq, &ev->dwork, 0);
 	}
 	spin_unlock_irq(&ev->lock);
 }
@@ -1558,7 +1561,7 @@
  
 	/* uncondtionally schedule event check and wait for it to finish */
 	disk_block_events(disk);
-	queue_delayed_work(system_nrt_wq, &ev->dwork, 0);
+	queue_delayed_work(system_nrt_freezable_wq, &ev->dwork, 0);
 	flush_delayed_work(&ev->dwork);
 	__disk_unblock_events(disk, false);
  
@@ -1595,7 +1598,7 @@
  
 	intv = disk_events_poll_jiffies(disk);
 	if (!ev->block && intv)
-		queue_delayed_work(system_nrt_wq, &ev->dwork, intv);
+		queue_delayed_work(system_nrt_freezable_wq, &ev->dwork, intv);
  
 	spin_unlock_irq(&ev->lock);
  
  
@@ -1733,9 +1736,9 @@
 		&disk_events_dfl_poll_msecs, 0644);
  
 /*
- * disk_{add|del|release}_events - initialize and destroy disk_events.
+ * disk_{alloc|add|del|release}_events - initialize and destroy disk_events.
  */
-static void disk_add_events(struct gendisk *disk)
+static void disk_alloc_events(struct gendisk *disk)
 {
 	struct disk_events *ev;
  
@@ -1748,16 +1751,6 @@
 		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);
  
@@ -1766,8 +1759,21 @@
 	ev->poll_msecs = -1;
 	INIT_DELAYED_WORK(&ev->dwork, disk_events_workfn);
  
+	disk->ev = ev;
+}
+
+static void disk_add_events(struct gendisk *disk)
+{
+	if (!disk->ev)
+		return;
+
+	/* FIXME: error handling */
+	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);
+
 	mutex_lock(&disk_events_mutex);
-	list_add_tail(&ev->node, &disk_events);
+	list_add_tail(&disk->ev->node, &disk_events);
 	mutex_unlock(&disk_events_mutex);
  
 	/*
@@ -389,17 +389,11 @@
 	}
 }
  
-int rescan_partitions(struct gendisk *disk, struct block_device *bdev)
+static int drop_partitions(struct gendisk *disk, struct block_device *bdev)
 {
-	struct parsed_partitions *state = NULL;
 	struct disk_part_iter piter;
 	struct hd_struct *part;
-	int p, highest, res;
-rescan:
-	if (state && !IS_ERR(state)) {
-		kfree(state);
-		state = NULL;
-	}
+	int res;
  
 	if (bdev->bd_part_count)
 		return -EBUSY;
@@ -412,6 +406,24 @@
 		delete_partition(disk, part->partno);
 	disk_part_iter_exit(&piter);
  
+	return 0;
+}
+
+int rescan_partitions(struct gendisk *disk, struct block_device *bdev)
+{
+	struct parsed_partitions *state = NULL;
+	struct hd_struct *part;
+	int p, highest, res;
+rescan:
+	if (state && !IS_ERR(state)) {
+		kfree(state);
+		state = NULL;
+	}
+
+	res = drop_partitions(disk, bdev);
+	if (res)
+		return res;
+
 	if (disk->fops->revalidate_disk)
 		disk->fops->revalidate_disk(disk);
 	check_disk_size_change(disk, bdev);
@@ -512,6 +524,26 @@
 #endif
 	}
 	kfree(state);
+	return 0;
+}
+
+int invalidate_partitions(struct gendisk *disk, struct block_device *bdev)
+{
+	int res;
+
+	if (!bdev->bd_invalidated)
+		return 0;
+
+	res = drop_partitions(disk, bdev);
+	if (res)
+		return res;
+
+	set_capacity(disk, 0);
+	check_disk_size_change(disk, bdev);
+	bdev->bd_invalidated = 0;
+	/* tell userspace that the media / partition table may have changed */
+	kobject_uevent(&disk_to_dev(disk)->kobj, KOBJ_CHANGE);
+
 	return 0;
 }
  
@@ -1177,7 +1177,8 @@
   int TimeoutCounter;
   int i;
  
-  
+  memset(&CommandMailbox, 0, sizeof(DAC960_V1_CommandMailbox_T));
+
   if (pci_set_dma_mask(Controller->PCIDevice, DMA_BIT_MASK(32)))
 	return DAC960_Failure(Controller, "DMA mask out of range");
   Controller->BounceBufferLimit = DMA_BIT_MASK(32);
@@ -4627,7 +4628,8 @@
   DAC960_Controller_T *Controller = Command->Controller;
   DAC960_CommandType_T CommandType = Command->CommandType;
   DAC960_V2_CommandMailbox_T *CommandMailbox = &Command->V2.CommandMailbox;
-  DAC960_V2_IOCTL_Opcode_T CommandOpcode = CommandMailbox->Common.IOCTL_Opcode;
+  DAC960_V2_IOCTL_Opcode_T IOCTLOpcode = CommandMailbox->Common.IOCTL_Opcode;
+  DAC960_V2_CommandOpcode_T CommandOpcode = CommandMailbox->SCSI_10.CommandOpcode;
   DAC960_V2_CommandStatus_T CommandStatus = Command->V2.CommandStatus;
  
   if (CommandType == DAC960_ReadCommand ||
@@ -4699,7 +4701,7 @@
     {
       if (Controller->ShutdownMonitoringTimer)
 	      return;
-      if (CommandOpcode == DAC960_V2_GetControllerInfo)
+      if (IOCTLOpcode == DAC960_V2_GetControllerInfo)
 	{
 	  DAC960_V2_ControllerInfo_T *NewControllerInfo =
 	    Controller->V2.NewControllerInformation;
  
@@ -4719,14 +4721,14 @@
 	  memcpy(ControllerInfo, NewControllerInfo,
 		 sizeof(DAC960_V2_ControllerInfo_T));
 	}
-      else if (CommandOpcode == DAC960_V2_GetEvent)
+      else if (IOCTLOpcode == DAC960_V2_GetEvent)
 	{
 	  if (CommandStatus == DAC960_V2_NormalCompletion) {
 	    DAC960_V2_ReportEvent(Controller, Controller->V2.Event);
 	  }
 	  Controller->V2.NextEventSequenceNumber++;
 	}
-      else if (CommandOpcode == DAC960_V2_GetPhysicalDeviceInfoValid &&
+      else if (IOCTLOpcode == DAC960_V2_GetPhysicalDeviceInfoValid &&
 	       CommandStatus == DAC960_V2_NormalCompletion)
 	{
 	  DAC960_V2_PhysicalDeviceInfo_T *NewPhysicalDeviceInfo =
@@ -4915,7 +4917,7 @@
 	  NewPhysicalDeviceInfo->LogicalUnit++;
 	  Controller->V2.PhysicalDeviceIndex++;
 	}
-      else if (CommandOpcode == DAC960_V2_GetPhysicalDeviceInfoValid)
+      else if (IOCTLOpcode == DAC960_V2_GetPhysicalDeviceInfoValid)
 	{
 	  unsigned int DeviceIndex;
 	  for (DeviceIndex = Controller->V2.PhysicalDeviceIndex;
@@ -4938,7 +4940,7 @@
 	    }
 	  Controller->V2.NeedPhysicalDeviceInformation = false;
 	}
-      else if (CommandOpcode == DAC960_V2_GetLogicalDeviceInfoValid &&
+      else if (IOCTLOpcode == DAC960_V2_GetLogicalDeviceInfoValid &&
 	       CommandStatus == DAC960_V2_NormalCompletion)
 	{
 	  DAC960_V2_LogicalDeviceInfo_T *NewLogicalDeviceInfo =
@@ -5065,7 +5067,7 @@
 			 [LogicalDeviceNumber] = true;
 	  NewLogicalDeviceInfo->LogicalDeviceNumber++;
 	}
-      else if (CommandOpcode == DAC960_V2_GetLogicalDeviceInfoValid)
+      else if (IOCTLOpcode == DAC960_V2_GetLogicalDeviceInfoValid)
 	{
 	  int LogicalDriveNumber;
 	  for (LogicalDriveNumber = 0;
@@ -1120,7 +1120,7 @@
 			break;
 		case MISC_GET_FW_VER: {
 			struct carm_fw_ver *ver = (struct carm_fw_ver *)
-				mem + sizeof(struct carm_msg_get_fw_ver);
+				(mem + sizeof(struct carm_msg_get_fw_ver));
 			if (!error) {
 				host->fw_ver = le32_to_cpu(ver->version);
 				host->flags |= (ver->features & FL_FW_VER_MASK);
@@ -1183,8 +1183,12 @@
 			 * The latter is necessary to prevent ghost
 			 * partitions on a removed medium.
 			 */
-			if (bdev->bd_invalidated && (!ret || ret == -ENOMEDIUM))
-				rescan_partitions(disk, bdev);
+			if (bdev->bd_invalidated) {
+				if (!ret)
+					rescan_partitions(disk, bdev);
+				else if (ret == -ENOMEDIUM)
+					invalidate_partitions(disk, bdev);
+			}
 			if (ret)
 				goto out_clear;
 		} else {
@@ -1214,8 +1218,12 @@
 			if (bdev->bd_disk->fops->open)
 				ret = bdev->bd_disk->fops->open(bdev, mode);
 			/* the same as first opener case, read comment there */
-			if (bdev->bd_invalidated && (!ret || ret == -ENOMEDIUM))
-				rescan_partitions(bdev->bd_disk, bdev);
+			if (bdev->bd_invalidated) {
+				if (!ret)
+					rescan_partitions(bdev->bd_disk, bdev);
+				else if (ret == -ENOMEDIUM)
+					invalidate_partitions(bdev->bd_disk, bdev);
+			}
 			if (ret)
 				goto out_unlock_bdev;
 		}
@@ -596,6 +596,7 @@
  
 extern int disk_expand_part_tbl(struct gendisk *disk, int target);
 extern int rescan_partitions(struct gendisk *disk, struct block_device *bdev);
+extern int invalidate_partitions(struct gendisk *disk, struct block_device *bdev);
 extern struct hd_struct * __must_check add_partition(struct gendisk *disk,
 						     int partno, sector_t start,
 						     sector_t len, int flags,
@@ -6,8 +6,11 @@
 #include <linux/workqueue.h>
  
 enum {
-	ICQ_IOPRIO_CHANGED,
-	ICQ_CGROUP_CHANGED,
+	ICQ_IOPRIO_CHANGED	= 1 << 0,
+	ICQ_CGROUP_CHANGED	= 1 << 1,
+	ICQ_EXITED		= 1 << 2,
+
+	ICQ_CHANGED_MASK	= ICQ_IOPRIO_CHANGED | ICQ_CGROUP_CHANGED,
 };
  
 /*
@@ -88,7 +91,7 @@
 		struct rcu_head		__rcu_head;
 	};
  
-	unsigned long		changed;
+	unsigned int		flags;
 };
  
 /*
@@ -139,6 +142,7 @@
 				       gfp_t gfp_flags, int node);
 void ioc_ioprio_changed(struct io_context *ioc, int ioprio);
 void ioc_cgroup_changed(struct io_context *ioc);
+unsigned int icq_get_changed(struct io_cq *icq);
 #else
 struct io_context;
 static inline void put_io_context(struct io_context *ioc) { }
@@ -289,12 +289,16 @@
  *
  * system_freezable_wq is equivalent to system_wq except that it's
  * freezable.
+ *
+ * system_nrt_freezable_wq is equivalent to system_nrt_wq except that
+ * it's freezable.
  */
 extern struct workqueue_struct *system_wq;
 extern struct workqueue_struct *system_long_wq;
 extern struct workqueue_struct *system_nrt_wq;
 extern struct workqueue_struct *system_unbound_wq;
 extern struct workqueue_struct *system_freezable_wq;
+extern struct workqueue_struct *system_nrt_freezable_wq;
  
 extern struct workqueue_struct *
 __alloc_workqueue_key(const char *fmt, unsigned int flags, int max_active,
@@ -253,11 +253,13 @@
 struct workqueue_struct *system_nrt_wq __read_mostly;
 struct workqueue_struct *system_unbound_wq __read_mostly;
 struct workqueue_struct *system_freezable_wq __read_mostly;
+struct workqueue_struct *system_nrt_freezable_wq __read_mostly;
 EXPORT_SYMBOL_GPL(system_wq);
 EXPORT_SYMBOL_GPL(system_long_wq);
 EXPORT_SYMBOL_GPL(system_nrt_wq);
 EXPORT_SYMBOL_GPL(system_unbound_wq);
 EXPORT_SYMBOL_GPL(system_freezable_wq);
+EXPORT_SYMBOL_GPL(system_nrt_freezable_wq);
  
 #define CREATE_TRACE_POINTS
 #include <trace/events/workqueue.h>
  
@@ -3833,8 +3835,11 @@
 					    WQ_UNBOUND_MAX_ACTIVE);
 	system_freezable_wq = alloc_workqueue("events_freezable",
 					      WQ_FREEZABLE, 0);
+	system_nrt_freezable_wq = alloc_workqueue("events_nrt_freezable",
+			WQ_NON_REENTRANT | WQ_FREEZABLE, 0);
 	BUG_ON(!system_wq || !system_long_wq || !system_nrt_wq ||
-	       !system_unbound_wq || !system_freezable_wq);
+	       !system_unbound_wq || !system_freezable_wq ||
+		!system_nrt_freezable_wq);
 	return 0;
 }
 early_initcall(init_workqueues);
...	...	@@ -36,11 +36,23 @@
36	36	kmem_cache_free(icq->__rcu_icq_cache, icq);
37	37	}
38	38
39		-/*
40		- * Exit and free an icq. Called with both ioc and q locked.
41		- */
	39	+/* Exit an icq. Called with both ioc and q locked. */
42	40	static void ioc_exit_icq(struct io_cq *icq)
43	41	{
	42	+ struct elevator_type *et = icq->q->elevator->type;
	43	+
	44	+ if (icq->flags & ICQ_EXITED)
	45	+ return;
	46	+
	47	+ if (et->ops.elevator_exit_icq_fn)
	48	+ et->ops.elevator_exit_icq_fn(icq);
	49	+
	50	+ icq->flags \|= ICQ_EXITED;
	51	+}
	52	+
	53	+/* Release an icq. Called with both ioc and q locked. */
	54	+static void ioc_destroy_icq(struct io_cq *icq)
	55	+{
44	56	struct io_context *ioc = icq->ioc;
45	57	struct request_queue *q = icq->q;
46	58	struct elevator_type *et = q->elevator->type;
...	...	@@ -60,8 +72,7 @@
60	72	if (rcu_dereference_raw(ioc->icq_hint) == icq)
61	73	rcu_assign_pointer(ioc->icq_hint, NULL);
62	74
63		- if (et->ops.elevator_exit_icq_fn)
64		- et->ops.elevator_exit_icq_fn(icq);
	75	+ ioc_exit_icq(icq);
65	76
66	77	/*
67	78	* @icq->q might have gone away by the time RCU callback runs
...	...	@@ -79,7 +90,6 @@
79	90	{
80	91	struct io_context *ioc = container_of(work, struct io_context,
81	92	release_work);
82		- struct request_queue *last_q = NULL;
83	93	unsigned long flags;
84	94
85	95	/*
86	96
87	97
88	98
...	...	@@ -93,44 +103,19 @@
93	103	while (!hlist_empty(&ioc->icq_list)) {
94	104	struct io_cq *icq = hlist_entry(ioc->icq_list.first,
95	105	struct io_cq, ioc_node);
96		- struct request_queue *this_q = icq->q;
	106	+ struct request_queue *q = icq->q;
97	107
98		- if (this_q != last_q) {
99		- /*
100		- * Need to switch to @this_q. Once we release
101		- * @ioc->lock, it can go away along with @cic.
102		- * Hold on to it.
103		- */
104		- __blk_get_queue(this_q);
105		-
106		- /*
107		- * blk_put_queue() might sleep thanks to kobject
108		- * idiocy. Always release both locks, put and
109		- * restart.
110		- */
111		- if (last_q) {
112		- spin_unlock(last_q->queue_lock);
113		- spin_unlock_irqrestore(&ioc->lock, flags);
114		- blk_put_queue(last_q);
115		- } else {
116		- spin_unlock_irqrestore(&ioc->lock, flags);
117		- }
118		-
119		- last_q = this_q;
120		- spin_lock_irqsave(this_q->queue_lock, flags);
121		- spin_lock_nested(&ioc->lock, 1);
122		- continue;
	108	+ if (spin_trylock(q->queue_lock)) {
	109	+ ioc_destroy_icq(icq);
	110	+ spin_unlock(q->queue_lock);
	111	+ } else {
	112	+ spin_unlock_irqrestore(&ioc->lock, flags);
	113	+ cpu_relax();
	114	+ spin_lock_irqsave_nested(&ioc->lock, flags, 1);
123	115	}
124		- ioc_exit_icq(icq);
125	116	}
126	117
127		- if (last_q) {
128		- spin_unlock(last_q->queue_lock);
129		- spin_unlock_irqrestore(&ioc->lock, flags);
130		- blk_put_queue(last_q);
131		- } else {
132		- spin_unlock_irqrestore(&ioc->lock, flags);
133		- }
	118	+ spin_unlock_irqrestore(&ioc->lock, flags);
134	119
135	120	kmem_cache_free(iocontext_cachep, ioc);
136	121	}
...	...	@@ -145,6 +130,7 @@
145	130	void put_io_context(struct io_context *ioc)
146	131	{
147	132	unsigned long flags;
	133	+ bool free_ioc = false;
148	134
149	135	if (ioc == NULL)
150	136	return;
151	137
...	...	@@ -159,8 +145,13 @@
159	145	spin_lock_irqsave(&ioc->lock, flags);
160	146	if (!hlist_empty(&ioc->icq_list))
161	147	schedule_work(&ioc->release_work);
	148	+ else
	149	+ free_ioc = true;
162	150	spin_unlock_irqrestore(&ioc->lock, flags);
163	151	}
	152	+
	153	+ if (free_ioc)
	154	+ kmem_cache_free(iocontext_cachep, ioc);
164	155	}
165	156	EXPORT_SYMBOL(put_io_context);
166	157
167	158
...	...	@@ -168,13 +159,41 @@
168	159	void exit_io_context(struct task_struct *task)
169	160	{
170	161	struct io_context *ioc;
	162	+ struct io_cq *icq;
	163	+ struct hlist_node *n;
	164	+ unsigned long flags;
171	165
172	166	task_lock(task);
173	167	ioc = task->io_context;
174	168	task->io_context = NULL;
175	169	task_unlock(task);
176	170
177		- atomic_dec(&ioc->nr_tasks);
	171	+ if (!atomic_dec_and_test(&ioc->nr_tasks)) {
	172	+ put_io_context(ioc);
	173	+ return;
	174	+ }
	175	+
	176	+ /*
	177	+ * Need ioc lock to walk icq_list and q lock to exit icq. Perform
	178	+ * reverse double locking. Read comment in ioc_release_fn() for
	179	+ * explanation on the nested locking annotation.
	180	+ */
	181	+retry:
	182	+ spin_lock_irqsave_nested(&ioc->lock, flags, 1);
	183	+ hlist_for_each_entry(icq, n, &ioc->icq_list, ioc_node) {
	184	+ if (icq->flags & ICQ_EXITED)
	185	+ continue;
	186	+ if (spin_trylock(icq->q->queue_lock)) {
	187	+ ioc_exit_icq(icq);
	188	+ spin_unlock(icq->q->queue_lock);
	189	+ } else {
	190	+ spin_unlock_irqrestore(&ioc->lock, flags);
	191	+ cpu_relax();
	192	+ goto retry;
	193	+ }
	194	+ }
	195	+ spin_unlock_irqrestore(&ioc->lock, flags);
	196	+
178	197	put_io_context(ioc);
179	198	}
180	199
...	...	@@ -194,7 +213,7 @@
194	213	struct io_context *ioc = icq->ioc;
195	214
196	215	spin_lock(&ioc->lock);
197		- ioc_exit_icq(icq);
	216	+ ioc_destroy_icq(icq);
198	217	spin_unlock(&ioc->lock);
199	218	}
200	219	}
201	220
...	...	@@ -363,13 +382,13 @@
363	382	return icq;
364	383	}
365	384
366		-void ioc_set_changed(struct io_context *ioc, int which)
	385	+void ioc_set_icq_flags(struct io_context *ioc, unsigned int flags)
367	386	{
368	387	struct io_cq *icq;
369	388	struct hlist_node *n;
370	389
371	390	hlist_for_each_entry(icq, n, &ioc->icq_list, ioc_node)
372		- set_bit(which, &icq->changed);
	391	+ icq->flags \|= flags;
373	392	}
374	393
375	394	/**
...	...	@@ -387,7 +406,7 @@
387	406
388	407	spin_lock_irqsave(&ioc->lock, flags);
389	408	ioc->ioprio = ioprio;
390		- ioc_set_changed(ioc, ICQ_IOPRIO_CHANGED);
	409	+ ioc_set_icq_flags(ioc, ICQ_IOPRIO_CHANGED);
391	410	spin_unlock_irqrestore(&ioc->lock, flags);
392	411	}
393	412
394	413
...	...	@@ -404,10 +423,32 @@
404	423	unsigned long flags;
405	424
406	425	spin_lock_irqsave(&ioc->lock, flags);
407		- ioc_set_changed(ioc, ICQ_CGROUP_CHANGED);
	426	+ ioc_set_icq_flags(ioc, ICQ_CGROUP_CHANGED);
408	427	spin_unlock_irqrestore(&ioc->lock, flags);
409	428	}
410	429	EXPORT_SYMBOL(ioc_cgroup_changed);
	430	+
	431	+/**
	432	+ * icq_get_changed - fetch and clear icq changed mask
	433	+ * @icq: icq of interest
	434	+ *
	435	+ * Fetch and clear ICQ_*_CHANGED bits from @icq. Grabs and releases
	436	+ * @icq->ioc->lock.
	437	+ */
	438	+unsigned icq_get_changed(struct io_cq *icq)
	439	+{
	440	+ unsigned int changed = 0;
	441	+ unsigned long flags;
	442	+
	443	+ if (unlikely(icq->flags & ICQ_CHANGED_MASK)) {
	444	+ spin_lock_irqsave(&icq->ioc->lock, flags);
	445	+ changed = icq->flags & ICQ_CHANGED_MASK;
	446	+ icq->flags &= ~ICQ_CHANGED_MASK;
	447	+ spin_unlock_irqrestore(&icq->ioc->lock, flags);
	448	+ }
	449	+ return changed;
	450	+}
	451	+EXPORT_SYMBOL(icq_get_changed);
411	452
412	453	static int __init blk_ioc_init(void)
413	454	{
...	...	@@ -3470,20 +3470,20 @@
3470	3470	const int rw = rq_data_dir(rq);
3471	3471	const bool is_sync = rq_is_sync(rq);
3472	3472	struct cfq_queue *cfqq;
	3473	+ unsigned int changed;
3473	3474
3474	3475	might_sleep_if(gfp_mask & __GFP_WAIT);
3475	3476
3476	3477	spin_lock_irq(q->queue_lock);
3477	3478
3478	3479	/* handle changed notifications */
3479		- if (unlikely(cic->icq.changed)) {
3480		- if (test_and_clear_bit(ICQ_IOPRIO_CHANGED, &cic->icq.changed))
3481		- changed_ioprio(cic);
	3480	+ changed = icq_get_changed(&cic->icq);
	3481	+ if (unlikely(changed & ICQ_IOPRIO_CHANGED))
	3482	+ changed_ioprio(cic);
3482	3483	#ifdef CONFIG_CFQ_GROUP_IOSCHED
3483		- if (test_and_clear_bit(ICQ_CGROUP_CHANGED, &cic->icq.changed))
3484		- changed_cgroup(cic);
	3484	+ if (unlikely(changed & ICQ_CGROUP_CHANGED))
	3485	+ changed_cgroup(cic);
3485	3486	#endif
3486		- }
3487	3487
3488	3488	new_queue:
3489	3489	cfqq = cic_to_cfqq(cic, is_sync);
...	...	@@ -35,6 +35,7 @@
35	35
36	36	static struct device_type disk_type;
37	37
	38	+static void disk_alloc_events(struct gendisk *disk);
38	39	static void disk_add_events(struct gendisk *disk);
39	40	static void disk_del_events(struct gendisk *disk);
40	41	static void disk_release_events(struct gendisk *disk);
...	...	@@ -601,6 +602,8 @@
601	602	disk->major = MAJOR(devt);
602	603	disk->first_minor = MINOR(devt);
603	604
	605	+ disk_alloc_events(disk);
	606	+
604	607	/* Register BDI before referencing it from bdev */
605	608	bdi = &disk->queue->backing_dev_info;
606	609	bdi_register_dev(bdi, disk_devt(disk));
607	610
...	...	@@ -1475,9 +1478,9 @@
1475	1478	intv = disk_events_poll_jiffies(disk);
1476	1479	set_timer_slack(&ev->dwork.timer, intv / 4);
1477	1480	if (check_now)
1478		- queue_delayed_work(system_nrt_wq, &ev->dwork, 0);
	1481	+ queue_delayed_work(system_nrt_freezable_wq, &ev->dwork, 0);
1479	1482	else if (intv)
1480		- queue_delayed_work(system_nrt_wq, &ev->dwork, intv);
	1483	+ queue_delayed_work(system_nrt_freezable_wq, &ev->dwork, intv);
1481	1484	out_unlock:
1482	1485	spin_unlock_irqrestore(&ev->lock, flags);
1483	1486	}
...	...	@@ -1521,7 +1524,7 @@
1521	1524	ev->clearing \|= mask;
1522	1525	if (!ev->block) {
1523	1526	cancel_delayed_work(&ev->dwork);
1524		- queue_delayed_work(system_nrt_wq, &ev->dwork, 0);
	1527	+ queue_delayed_work(system_nrt_freezable_wq, &ev->dwork, 0);
1525	1528	}
1526	1529	spin_unlock_irq(&ev->lock);
1527	1530	}
...	...	@@ -1558,7 +1561,7 @@
1558	1561
1559	1562	/* uncondtionally schedule event check and wait for it to finish */
1560	1563	disk_block_events(disk);
1561		- queue_delayed_work(system_nrt_wq, &ev->dwork, 0);
	1564	+ queue_delayed_work(system_nrt_freezable_wq, &ev->dwork, 0);
1562	1565	flush_delayed_work(&ev->dwork);
1563	1566	__disk_unblock_events(disk, false);
1564	1567
...	...	@@ -1595,7 +1598,7 @@
1595	1598
1596	1599	intv = disk_events_poll_jiffies(disk);
1597	1600	if (!ev->block && intv)
1598		- queue_delayed_work(system_nrt_wq, &ev->dwork, intv);
	1601	+ queue_delayed_work(system_nrt_freezable_wq, &ev->dwork, intv);
1599	1602
1600	1603	spin_unlock_irq(&ev->lock);
1601	1604
1602	1605
...	...	@@ -1733,9 +1736,9 @@
1733	1736	&disk_events_dfl_poll_msecs, 0644);
1734	1737
1735	1738	/*
1736		- * disk_{add\|del\|release}_events - initialize and destroy disk_events.
	1739	+ * disk_{alloc\|add\|del\|release}_events - initialize and destroy disk_events.
1737	1740	*/
1738		-static void disk_add_events(struct gendisk *disk)
	1741	+static void disk_alloc_events(struct gendisk *disk)
1739	1742	{
1740	1743	struct disk_events *ev;
1741	1744
...	...	@@ -1748,16 +1751,6 @@
1748	1751	return;
1749	1752	}
1750	1753
1751		- if (sysfs_create_files(&disk_to_dev(disk)->kobj,
1752		- disk_events_attrs) < 0) {
1753		- pr_warn("%s: failed to create sysfs files for events\n",
1754		- disk->disk_name);
1755		- kfree(ev);
1756		- return;
1757		- }
1758		-
1759		- disk->ev = ev;
1760		-
1761	1754	INIT_LIST_HEAD(&ev->node);
1762	1755	ev->disk = disk;
1763	1756	spin_lock_init(&ev->lock);
1764	1757
...	...	@@ -1766,8 +1759,21 @@
1766	1759	ev->poll_msecs = -1;
1767	1760	INIT_DELAYED_WORK(&ev->dwork, disk_events_workfn);
1768	1761
	1762	+ disk->ev = ev;
	1763	+}
	1764	+
	1765	+static void disk_add_events(struct gendisk *disk)
	1766	+{
	1767	+ if (!disk->ev)
	1768	+ return;
	1769	+
	1770	+ /* FIXME: error handling */
	1771	+ if (sysfs_create_files(&disk_to_dev(disk)->kobj, disk_events_attrs) < 0)
	1772	+ pr_warn("%s: failed to create sysfs files for events\n",
	1773	+ disk->disk_name);
	1774	+
1769	1775	mutex_lock(&disk_events_mutex);
1770		- list_add_tail(&ev->node, &disk_events);
	1776	+ list_add_tail(&disk->ev->node, &disk_events);
1771	1777	mutex_unlock(&disk_events_mutex);
1772	1778
1773	1779	/*
...	...	@@ -389,17 +389,11 @@
389	389	}
390	390	}
391	391
392		-int rescan_partitions(struct gendisk disk, struct block_device bdev)
	392	+static int drop_partitions(struct gendisk disk, struct block_device bdev)
393	393	{
394		- struct parsed_partitions *state = NULL;
395	394	struct disk_part_iter piter;
396	395	struct hd_struct *part;
397		- int p, highest, res;
398		-rescan:
399		- if (state && !IS_ERR(state)) {
400		- kfree(state);
401		- state = NULL;
402		- }
	396	+ int res;
403	397
404	398	if (bdev->bd_part_count)
405	399	return -EBUSY;
...	...	@@ -412,6 +406,24 @@
412	406	delete_partition(disk, part->partno);
413	407	disk_part_iter_exit(&piter);
414	408
	409	+ return 0;
	410	+}
	411	+
	412	+int rescan_partitions(struct gendisk disk, struct block_device bdev)
	413	+{
	414	+ struct parsed_partitions *state = NULL;
	415	+ struct hd_struct *part;
	416	+ int p, highest, res;
	417	+rescan:
	418	+ if (state && !IS_ERR(state)) {
	419	+ kfree(state);
	420	+ state = NULL;
	421	+ }
	422	+
	423	+ res = drop_partitions(disk, bdev);
	424	+ if (res)
	425	+ return res;
	426	+
415	427	if (disk->fops->revalidate_disk)
416	428	disk->fops->revalidate_disk(disk);
417	429	check_disk_size_change(disk, bdev);
...	...	@@ -512,6 +524,26 @@
512	524	#endif
513	525	}
514	526	kfree(state);
	527	+ return 0;
	528	+}
	529	+
	530	+int invalidate_partitions(struct gendisk disk, struct block_device bdev)
	531	+{
	532	+ int res;
	533	+
	534	+ if (!bdev->bd_invalidated)
	535	+ return 0;
	536	+
	537	+ res = drop_partitions(disk, bdev);
	538	+ if (res)
	539	+ return res;
	540	+
	541	+ set_capacity(disk, 0);
	542	+ check_disk_size_change(disk, bdev);
	543	+ bdev->bd_invalidated = 0;
	544	+ /* tell userspace that the media / partition table may have changed */
	545	+ kobject_uevent(&disk_to_dev(disk)->kobj, KOBJ_CHANGE);
	546	+
515	547	return 0;
516	548	}
517	549
...	...	@@ -1177,7 +1177,8 @@
1177	1177	int TimeoutCounter;
1178	1178	int i;
1179	1179
1180		-
	1180	+ memset(&CommandMailbox, 0, sizeof(DAC960_V1_CommandMailbox_T));
	1181	+
1181	1182	if (pci_set_dma_mask(Controller->PCIDevice, DMA_BIT_MASK(32)))
1182	1183	return DAC960_Failure(Controller, "DMA mask out of range");
1183	1184	Controller->BounceBufferLimit = DMA_BIT_MASK(32);
...	...	@@ -4627,7 +4628,8 @@
4627	4628	DAC960_Controller_T *Controller = Command->Controller;
4628	4629	DAC960_CommandType_T CommandType = Command->CommandType;
4629	4630	DAC960_V2_CommandMailbox_T *CommandMailbox = &Command->V2.CommandMailbox;
4630		- DAC960_V2_IOCTL_Opcode_T CommandOpcode = CommandMailbox->Common.IOCTL_Opcode;
	4631	+ DAC960_V2_IOCTL_Opcode_T IOCTLOpcode = CommandMailbox->Common.IOCTL_Opcode;
	4632	+ DAC960_V2_CommandOpcode_T CommandOpcode = CommandMailbox->SCSI_10.CommandOpcode;
4631	4633	DAC960_V2_CommandStatus_T CommandStatus = Command->V2.CommandStatus;
4632	4634
4633	4635	if (CommandType == DAC960_ReadCommand \|\|
...	...	@@ -4699,7 +4701,7 @@
4699	4701	{
4700	4702	if (Controller->ShutdownMonitoringTimer)
4701	4703	return;
4702		- if (CommandOpcode == DAC960_V2_GetControllerInfo)
	4704	+ if (IOCTLOpcode == DAC960_V2_GetControllerInfo)
4703	4705	{
4704	4706	DAC960_V2_ControllerInfo_T *NewControllerInfo =
4705	4707	Controller->V2.NewControllerInformation;
4706	4708
...	...	@@ -4719,14 +4721,14 @@
4719	4721	memcpy(ControllerInfo, NewControllerInfo,
4720	4722	sizeof(DAC960_V2_ControllerInfo_T));
4721	4723	}
4722		- else if (CommandOpcode == DAC960_V2_GetEvent)
	4724	+ else if (IOCTLOpcode == DAC960_V2_GetEvent)
4723	4725	{
4724	4726	if (CommandStatus == DAC960_V2_NormalCompletion) {
4725	4727	DAC960_V2_ReportEvent(Controller, Controller->V2.Event);
4726	4728	}
4727	4729	Controller->V2.NextEventSequenceNumber++;
4728	4730	}
4729		- else if (CommandOpcode == DAC960_V2_GetPhysicalDeviceInfoValid &&
	4731	+ else if (IOCTLOpcode == DAC960_V2_GetPhysicalDeviceInfoValid &&
4730	4732	CommandStatus == DAC960_V2_NormalCompletion)
4731	4733	{
4732	4734	DAC960_V2_PhysicalDeviceInfo_T *NewPhysicalDeviceInfo =
...	...	@@ -4915,7 +4917,7 @@
4915	4917	NewPhysicalDeviceInfo->LogicalUnit++;
4916	4918	Controller->V2.PhysicalDeviceIndex++;
4917	4919	}
4918		- else if (CommandOpcode == DAC960_V2_GetPhysicalDeviceInfoValid)
	4920	+ else if (IOCTLOpcode == DAC960_V2_GetPhysicalDeviceInfoValid)
4919	4921	{
4920	4922	unsigned int DeviceIndex;
4921	4923	for (DeviceIndex = Controller->V2.PhysicalDeviceIndex;
...	...	@@ -4938,7 +4940,7 @@
4938	4940	}
4939	4941	Controller->V2.NeedPhysicalDeviceInformation = false;
4940	4942	}
4941		- else if (CommandOpcode == DAC960_V2_GetLogicalDeviceInfoValid &&
	4943	+ else if (IOCTLOpcode == DAC960_V2_GetLogicalDeviceInfoValid &&
4942	4944	CommandStatus == DAC960_V2_NormalCompletion)
4943	4945	{
4944	4946	DAC960_V2_LogicalDeviceInfo_T *NewLogicalDeviceInfo =
...	...	@@ -5065,7 +5067,7 @@
5065	5067	[LogicalDeviceNumber] = true;
5066	5068	NewLogicalDeviceInfo->LogicalDeviceNumber++;
5067	5069	}
5068		- else if (CommandOpcode == DAC960_V2_GetLogicalDeviceInfoValid)
	5070	+ else if (IOCTLOpcode == DAC960_V2_GetLogicalDeviceInfoValid)
5069	5071	{
5070	5072	int LogicalDriveNumber;
5071	5073	for (LogicalDriveNumber = 0;
...	...	@@ -1120,7 +1120,7 @@
1120	1120	break;
1121	1121	case MISC_GET_FW_VER: {
1122	1122	struct carm_fw_ver ver = (struct carm_fw_ver )
1123		- mem + sizeof(struct carm_msg_get_fw_ver);
	1123	+ (mem + sizeof(struct carm_msg_get_fw_ver));
1124	1124	if (!error) {
1125	1125	host->fw_ver = le32_to_cpu(ver->version);
1126	1126	host->flags \|= (ver->features & FL_FW_VER_MASK);
...	...	@@ -1183,8 +1183,12 @@
1183	1183	* The latter is necessary to prevent ghost
1184	1184	* partitions on a removed medium.
1185	1185	*/
1186		- if (bdev->bd_invalidated && (!ret \|\| ret == -ENOMEDIUM))
1187		- rescan_partitions(disk, bdev);
	1186	+ if (bdev->bd_invalidated) {
	1187	+ if (!ret)
	1188	+ rescan_partitions(disk, bdev);
	1189	+ else if (ret == -ENOMEDIUM)
	1190	+ invalidate_partitions(disk, bdev);
	1191	+ }
1188	1192	if (ret)
1189	1193	goto out_clear;
1190	1194	} else {
...	...	@@ -1214,8 +1218,12 @@
1214	1218	if (bdev->bd_disk->fops->open)
1215	1219	ret = bdev->bd_disk->fops->open(bdev, mode);
1216	1220	/* the same as first opener case, read comment there */
1217		- if (bdev->bd_invalidated && (!ret \|\| ret == -ENOMEDIUM))
1218		- rescan_partitions(bdev->bd_disk, bdev);
	1221	+ if (bdev->bd_invalidated) {
	1222	+ if (!ret)
	1223	+ rescan_partitions(bdev->bd_disk, bdev);
	1224	+ else if (ret == -ENOMEDIUM)
	1225	+ invalidate_partitions(bdev->bd_disk, bdev);
	1226	+ }
1219	1227	if (ret)
1220	1228	goto out_unlock_bdev;
1221	1229	}
...	...	@@ -596,6 +596,7 @@
596	596
597	597	extern int disk_expand_part_tbl(struct gendisk *disk, int target);
598	598	extern int rescan_partitions(struct gendisk disk, struct block_device bdev);
	599	+extern int invalidate_partitions(struct gendisk disk, struct block_device bdev);
599	600	extern struct hd_struct * __must_check add_partition(struct gendisk *disk,
600	601	int partno, sector_t start,
601	602	sector_t len, int flags,
...	...	@@ -6,8 +6,11 @@
6	6	#include <linux/workqueue.h>
7	7
8	8	enum {
9		- ICQ_IOPRIO_CHANGED,
10		- ICQ_CGROUP_CHANGED,
	9	+ ICQ_IOPRIO_CHANGED = 1 << 0,
	10	+ ICQ_CGROUP_CHANGED = 1 << 1,
	11	+ ICQ_EXITED = 1 << 2,
	12	+
	13	+ ICQ_CHANGED_MASK = ICQ_IOPRIO_CHANGED \| ICQ_CGROUP_CHANGED,
11	14	};
12	15
13	16	/*
...	...	@@ -88,7 +91,7 @@
88	91	struct rcu_head __rcu_head;
89	92	};
90	93
91		- unsigned long changed;
	94	+ unsigned int flags;
92	95	};
93	96
94	97	/*
...	...	@@ -139,6 +142,7 @@
139	142	gfp_t gfp_flags, int node);
140	143	void ioc_ioprio_changed(struct io_context *ioc, int ioprio);
141	144	void ioc_cgroup_changed(struct io_context *ioc);
	145	+unsigned int icq_get_changed(struct io_cq *icq);
142	146	#else
143	147	struct io_context;
144	148	static inline void put_io_context(struct io_context *ioc) { }
...	...	@@ -289,12 +289,16 @@
289	289	*
290	290	* system_freezable_wq is equivalent to system_wq except that it's
291	291	* freezable.
	292	+ *
	293	+ * system_nrt_freezable_wq is equivalent to system_nrt_wq except that
	294	+ * it's freezable.
292	295	*/
293	296	extern struct workqueue_struct *system_wq;
294	297	extern struct workqueue_struct *system_long_wq;
295	298	extern struct workqueue_struct *system_nrt_wq;
296	299	extern struct workqueue_struct *system_unbound_wq;
297	300	extern struct workqueue_struct *system_freezable_wq;
	301	+extern struct workqueue_struct *system_nrt_freezable_wq;
298	302
299	303	extern struct workqueue_struct *
300	304	__alloc_workqueue_key(const char *fmt, unsigned int flags, int max_active,