Merge branch 'master' into for-2.6.35

Conflicts:
	fs/ext3/fsync.c

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>

1 files changed, 217 insertions, 118 deletions

diff --git a/drivers/mtd/mtd_blkdevs.c b/drivers/mtd/mtd_blkdevs.c
index c82e09bbc5fd..03e19c1965cc 100644
--- a/drivers/mtd/mtd_blkdevs.c
+++ b/drivers/mtd/mtd_blkdevs.c
@@ -14,7 +14,6 @@
 #include <linux/mtd/mtd.h>
 #include <linux/blkdev.h>
 #include <linux/blkpg.h>
-#include <linux/freezer.h>
 #include <linux/spinlock.h>
 #include <linux/hdreg.h>
 #include <linux/init.h>
@@ -25,12 +24,42 @@
 #include "mtdcore.h"
 
 static LIST_HEAD(blktrans_majors);
+static DEFINE_MUTEX(blktrans_ref_mutex);
+
+void blktrans_dev_release(struct kref *kref)
+{
+	struct mtd_blktrans_dev *dev =
+		container_of(kref, struct mtd_blktrans_dev, ref);
+
+	dev->disk->private_data = NULL;
+	blk_cleanup_queue(dev->rq);
+	put_disk(dev->disk);
+	list_del(&dev->list);
+	kfree(dev);
+}
+
+static struct mtd_blktrans_dev *blktrans_dev_get(struct gendisk *disk)
+{
+	struct mtd_blktrans_dev *dev;
+
+	mutex_lock(&blktrans_ref_mutex);
+	dev = disk->private_data;
+
+	if (!dev)
+		goto unlock;
+	kref_get(&dev->ref);
+unlock:
+	mutex_unlock(&blktrans_ref_mutex);
+	return dev;
+}
+
+void blktrans_dev_put(struct mtd_blktrans_dev *dev)
+{
+	mutex_lock(&blktrans_ref_mutex);
+	kref_put(&dev->ref, blktrans_dev_release);
+	mutex_unlock(&blktrans_ref_mutex);
+}
 
-struct mtd_blkcore_priv {
-	struct task_struct *thread;
-	struct request_queue *rq;
-	spinlock_t queue_lock;
-};
 
 static int do_blktrans_request(struct mtd_blktrans_ops *tr,
 			       struct mtd_blktrans_dev *dev,
@@ -61,7 +90,6 @@ static int do_blktrans_request(struct mtd_blktrans_ops *tr,
 				return -EIO;
 		rq_flush_dcache_pages(req);
 		return 0;
-
 	case WRITE:
 		if (!tr->writesect)
 			return -EIO;
@@ -71,7 +99,6 @@ static int do_blktrans_request(struct mtd_blktrans_ops *tr,
 			if (tr->writesect(dev, block, buf))
 				return -EIO;
 		return 0;
-
 	default:
 		printk(KERN_NOTICE "Unknown request %u\n", rq_data_dir(req));
 		return -EIO;
@@ -80,14 +107,13 @@ static int do_blktrans_request(struct mtd_blktrans_ops *tr,
 
 static int mtd_blktrans_thread(void *arg)
 {
-	struct mtd_blktrans_ops *tr = arg;
-	struct request_queue *rq = tr->blkcore_priv->rq;
+	struct mtd_blktrans_dev *dev = arg;
+	struct request_queue *rq = dev->rq;
 	struct request *req = NULL;
 
 	spin_lock_irq(rq->queue_lock);
 
 	while (!kthread_should_stop()) {
-		struct mtd_blktrans_dev *dev;
 		int res;
 
 		if (!req && !(req = blk_fetch_request(rq))) {
@@ -98,13 +124,10 @@ static int mtd_blktrans_thread(void *arg)
 			continue;
 		}
 
-		dev = req->rq_disk->private_data;
-		tr = dev->tr;
-
 		spin_unlock_irq(rq->queue_lock);
 
 		mutex_lock(&dev->lock);
-		res = do_blktrans_request(tr, dev, req);
+		res = do_blktrans_request(dev->tr, dev, req);
 		mutex_unlock(&dev->lock);
 
 		spin_lock_irq(rq->queue_lock);
@@ -123,81 +146,112 @@ static int mtd_blktrans_thread(void *arg)
 
 static void mtd_blktrans_request(struct request_queue *rq)
 {
-	struct mtd_blktrans_ops *tr = rq->queuedata;
-	wake_up_process(tr->blkcore_priv->thread);
-}
+	struct mtd_blktrans_dev *dev;
+	struct request *req = NULL;
+
+	dev = rq->queuedata;
 
+	if (!dev)
+		while ((req = blk_fetch_request(rq)) != NULL)
+			__blk_end_request_all(req, -ENODEV);
+	else
+		wake_up_process(dev->thread);
+}
 
 static int blktrans_open(struct block_device *bdev, fmode_t mode)
 {
-	struct mtd_blktrans_dev *dev = bdev->bd_disk->private_data;
-	struct mtd_blktrans_ops *tr = dev->tr;
-	int ret = -ENODEV;
-
-	if (!get_mtd_device(NULL, dev->mtd->index))
-		goto out;
-
-	if (!try_module_get(tr->owner))
-		goto out_tr;
-
-	/* FIXME: Locking. A hot pluggable device can go away
-	   (del_mtd_device can be called for it) without its module
-	   being unloaded. */
-	dev->mtd->usecount++;
-
-	ret = 0;
-	if (tr->open && (ret = tr->open(dev))) {
-		dev->mtd->usecount--;
-		put_mtd_device(dev->mtd);
-	out_tr:
-		module_put(tr->owner);
+	struct mtd_blktrans_dev *dev = blktrans_dev_get(bdev->bd_disk);
+	int ret;
+
+	if (!dev)
+		return -ERESTARTSYS;
+
+	mutex_lock(&dev->lock);
+
+	if (!dev->mtd) {
+		ret = -ENXIO;
+		goto unlock;
 	}
- out:
+
+	ret = !dev->open++ && dev->tr->open ? dev->tr->open(dev) : 0;
+
+	/* Take another reference on the device so it won't go away till
+		last release */
+	if (!ret)
+		kref_get(&dev->ref);
+unlock:
+	mutex_unlock(&dev->lock);
+	blktrans_dev_put(dev);
 	return ret;
 }
 
 static int blktrans_release(struct gendisk *disk, fmode_t mode)
 {
-	struct mtd_blktrans_dev *dev = disk->private_data;
-	struct mtd_blktrans_ops *tr = dev->tr;
-	int ret = 0;
+	struct mtd_blktrans_dev *dev = blktrans_dev_get(disk);
+	int ret = -ENXIO;
 
-	if (tr->release)
-		ret = tr->release(dev);
+	if (!dev)
+		return ret;
 
-	if (!ret) {
-		dev->mtd->usecount--;
-		put_mtd_device(dev->mtd);
-		module_put(tr->owner);
-	}
+	mutex_lock(&dev->lock);
+
+	/* Release one reference, we sure its not the last one here*/
+	kref_put(&dev->ref, blktrans_dev_release);
 
+	if (!dev->mtd)
+		goto unlock;
+
+	ret = !--dev->open && dev->tr->release ? dev->tr->release(dev) : 0;
+unlock:
+	mutex_unlock(&dev->lock);
+	blktrans_dev_put(dev);
 	return ret;
 }
 
 static int blktrans_getgeo(struct block_device *bdev, struct hd_geometry *geo)
 {
-	struct mtd_blktrans_dev *dev = bdev->bd_disk->private_data;
+	struct mtd_blktrans_dev *dev = blktrans_dev_get(bdev->bd_disk);
+	int ret = -ENXIO;
+
+	if (!dev)
+		return ret;
+
+	mutex_lock(&dev->lock);
+
+	if (!dev->mtd)
+		goto unlock;
 
-	if (dev->tr->getgeo)
-		return dev->tr->getgeo(dev, geo);
-	return -ENOTTY;
+	ret = dev->tr->getgeo ? dev->tr->getgeo(dev, geo) : 0;
+unlock:
+	mutex_unlock(&dev->lock);
+	blktrans_dev_put(dev);
+	return ret;
 }
 
 static int blktrans_ioctl(struct block_device *bdev, fmode_t mode,
 			      unsigned int cmd, unsigned long arg)
 {
-	struct mtd_blktrans_dev *dev = bdev->bd_disk->private_data;
-	struct mtd_blktrans_ops *tr = dev->tr;
+	struct mtd_blktrans_dev *dev = blktrans_dev_get(bdev->bd_disk);
+	int ret = -ENXIO;
+
+	if (!dev)
+		return ret;
+
+	mutex_lock(&dev->lock);
+
+	if (!dev->mtd)
+		goto unlock;
 
 	switch (cmd) {
 	case BLKFLSBUF:
-		if (tr->flush)
-			return tr->flush(dev);
-		/* The core code did the work, we had nothing to do. */
-		return 0;
+		ret = dev->tr->flush ? dev->tr->flush(dev) : 0;
 	default:
-		return -ENOTTY;
+		ret = -ENOTTY;
 	}
+unlock:
+	mutex_unlock(&dev->lock);
+	blktrans_dev_put(dev);
+	return ret;
 }
 
 static const struct block_device_operations mtd_blktrans_ops = {
@@ -214,12 +268,14 @@ int add_mtd_blktrans_dev(struct mtd_blktrans_dev *new)
 	struct mtd_blktrans_dev *d;
 	int last_devnum = -1;
 	struct gendisk *gd;
+	int ret;
 
 	if (mutex_trylock(&mtd_table_mutex)) {
 		mutex_unlock(&mtd_table_mutex);
 		BUG();
 	}
 
+	mutex_lock(&blktrans_ref_mutex);
 	list_for_each_entry(d, &tr->devs, list) {
 		if (new->devnum == -1) {
 			/* Use first free number */
@@ -231,6 +287,7 @@ int add_mtd_blktrans_dev(struct mtd_blktrans_dev *new)
 			}
 		} else if (d->devnum == new->devnum) {
 			/* Required number taken */
+			mutex_unlock(&blktrans_ref_mutex);
 			return -EBUSY;
 		} else if (d->devnum > new->devnum) {
 			/* Required number was free */
@@ -239,24 +296,38 @@ int add_mtd_blktrans_dev(struct mtd_blktrans_dev *new)
 		}
 		last_devnum = d->devnum;
 	}
+
+	ret = -EBUSY;
 	if (new->devnum == -1)
 		new->devnum = last_devnum+1;
 
-	if ((new->devnum << tr->part_bits) > 256) {
-		return -EBUSY;
+	/* Check that the device and any partitions will get valid
+	 * minor numbers and that the disk naming code below can cope
+	 * with this number. */
+	if (new->devnum > (MINORMASK >> tr->part_bits) ||
+	    (tr->part_bits && new->devnum >= 27 * 26)) {
+		mutex_unlock(&blktrans_ref_mutex);
+		goto error1;
 	}
 
 	list_add_tail(&new->list, &tr->devs);
  added:
+	mutex_unlock(&blktrans_ref_mutex);
+
 	mutex_init(&new->lock);
+	kref_init(&new->ref);
 	if (!tr->writesect)
 		new->readonly = 1;
 
+	/* Create gendisk */
+	ret = -ENOMEM;
 	gd = alloc_disk(1 << tr->part_bits);
-	if (!gd) {
-		list_del(&new->list);
-		return -ENOMEM;
-	}
+
+	if (!gd)
+		goto error2;
+
+	new->disk = gd;
+	gd->private_data = new;
 	gd->major = tr->major;
 	gd->first_minor = (new->devnum) << tr->part_bits;
 	gd->fops = &mtd_blktrans_ops;
@@ -274,13 +345,35 @@ int add_mtd_blktrans_dev(struct mtd_blktrans_dev *new)
 		snprintf(gd->disk_name, sizeof(gd->disk_name),
 			 "%s%d", tr->name, new->devnum);
 
-	/* 2.5 has capacity in units of 512 bytes while still
-	   having BLOCK_SIZE_BITS set to 10. Just to keep us amused. */
 	set_capacity(gd, (new->size * tr->blksize) >> 9);
 
-	gd->private_data = new;
-	new->blkcore_priv = gd;
-	gd->queue = tr->blkcore_priv->rq;
+	/* Create the request queue */
+	spin_lock_init(&new->queue_lock);
+	new->rq = blk_init_queue(mtd_blktrans_request, &new->queue_lock);
+
+	if (!new->rq)
+		goto error3;
+
+	new->rq->queuedata = new;
+	blk_queue_logical_block_size(new->rq, tr->blksize);
+
+	if (tr->discard)
+		queue_flag_set_unlocked(QUEUE_FLAG_DISCARD,
+					new->rq);
+
+	gd->queue = new->rq;
+
+	__get_mtd_device(new->mtd);
+	__module_get(tr->owner);
+
+	/* Create processing thread */
+	/* TODO: workqueue ? */
+	new->thread = kthread_run(mtd_blktrans_thread, new,
+			"%s%d", tr->name, new->mtd->index);
+	if (IS_ERR(new->thread)) {
+		ret = PTR_ERR(new->thread);
+		goto error4;
+	}
 	gd->driverfs_dev = &new->mtd->dev;
 
 	if (new->readonly)
@@ -288,21 +381,65 @@ int add_mtd_blktrans_dev(struct mtd_blktrans_dev *new)
 
 	add_disk(gd);
 
+	if (new->disk_attributes) {
+		ret = sysfs_create_group(&disk_to_dev(gd)->kobj,
+					new->disk_attributes);
+		WARN_ON(ret);
+	}
 	return 0;
+error4:
+	module_put(tr->owner);
+	__put_mtd_device(new->mtd);
+	blk_cleanup_queue(new->rq);
+error3:
+	put_disk(new->disk);
+error2:
+	list_del(&new->list);
+error1:
+	kfree(new);
+	return ret;
 }
 
 int del_mtd_blktrans_dev(struct mtd_blktrans_dev *old)
 {
+	unsigned long flags;
+
 	if (mutex_trylock(&mtd_table_mutex)) {
 		mutex_unlock(&mtd_table_mutex);
 		BUG();
 	}
 
-	list_del(&old->list);
+	/* Stop new requests to arrive */
+	del_gendisk(old->disk);
+
+	if (old->disk_attributes)
+		sysfs_remove_group(&disk_to_dev(old->disk)->kobj,
+						old->disk_attributes);
+
+	/* Stop the thread */
+	kthread_stop(old->thread);
+
+	/* Kill current requests */
+	spin_lock_irqsave(&old->queue_lock, flags);
+	old->rq->queuedata = NULL;
+	blk_start_queue(old->rq);
+	spin_unlock_irqrestore(&old->queue_lock, flags);
+
+	/* Ask trans driver for release to the mtd device */
+	mutex_lock(&old->lock);
+	if (old->open && old->tr->release) {
+		old->tr->release(old);
+		old->open = 0;
+	}
+
+	__put_mtd_device(old->mtd);
+	module_put(old->tr->owner);
 
-	del_gendisk(old->blkcore_priv);
-	put_disk(old->blkcore_priv);
+	/* At that point, we don't touch the mtd anymore */
+	old->mtd = NULL;
 
+	mutex_unlock(&old->lock);
+	blktrans_dev_put(old);
 	return 0;
 }
 
@@ -335,7 +472,8 @@ static struct mtd_notifier blktrans_notifier = {
 
 int register_mtd_blktrans(struct mtd_blktrans_ops *tr)
 {
-	int ret, i;
+	struct mtd_info *mtd;
+	int ret;
 
 	/* Register the notifier if/when the first device type is
 	   registered, to prevent the link/init ordering from fucking
@@ -343,9 +481,6 @@ int register_mtd_blktrans(struct mtd_blktrans_ops *tr)
 	if (!blktrans_notifier.list.next)
 		register_mtd_user(&blktrans_notifier);
 
-	tr->blkcore_priv = kzalloc(sizeof(*tr->blkcore_priv), GFP_KERNEL);
-	if (!tr->blkcore_priv)
-		return -ENOMEM;
 
 	mutex_lock(&mtd_table_mutex);
 
@@ -353,49 +488,20 @@ int register_mtd_blktrans(struct mtd_blktrans_ops *tr)
 	if (ret) {
 		printk(KERN_WARNING "Unable to register %s block device on major %d: %d\n",
 		       tr->name, tr->major, ret);
-		kfree(tr->blkcore_priv);
 		mutex_unlock(&mtd_table_mutex);
 		return ret;
 	}
-	spin_lock_init(&tr->blkcore_priv->queue_lock);
-
-	tr->blkcore_priv->rq = blk_init_queue(mtd_blktrans_request, &tr->blkcore_priv->queue_lock);
-	if (!tr->blkcore_priv->rq) {
-		unregister_blkdev(tr->major, tr->name);
-		kfree(tr->blkcore_priv);
-		mutex_unlock(&mtd_table_mutex);
-		return -ENOMEM;
-	}
-
-	tr->blkcore_priv->rq->queuedata = tr;
-	blk_queue_logical_block_size(tr->blkcore_priv->rq, tr->blksize);
-	if (tr->discard)
-		queue_flag_set_unlocked(QUEUE_FLAG_DISCARD,
-					tr->blkcore_priv->rq);
 
 	tr->blkshift = ffs(tr->blksize) - 1;
 
-	tr->blkcore_priv->thread = kthread_run(mtd_blktrans_thread, tr,
-			"%sd", tr->name);
-	if (IS_ERR(tr->blkcore_priv->thread)) {
-		ret = PTR_ERR(tr->blkcore_priv->thread);
-		blk_cleanup_queue(tr->blkcore_priv->rq);
-		unregister_blkdev(tr->major, tr->name);
-		kfree(tr->blkcore_priv);
-		mutex_unlock(&mtd_table_mutex);
-		return ret;
-	}
-
 	INIT_LIST_HEAD(&tr->devs);
 	list_add(&tr->list, &blktrans_majors);
 
-	for (i=0; i<MAX_MTD_DEVICES; i++) {
-		if (mtd_table[i] && mtd_table[i]->type != MTD_ABSENT)
-			tr->add_mtd(tr, mtd_table[i]);
-	}
+	mtd_for_each_device(mtd)
+		if (mtd->type != MTD_ABSENT)
+			tr->add_mtd(tr, mtd);
 
 	mutex_unlock(&mtd_table_mutex);
-
 	return 0;
 }
 
@@ -405,22 +511,15 @@ int deregister_mtd_blktrans(struct mtd_blktrans_ops *tr)
 
 	mutex_lock(&mtd_table_mutex);
 
-	/* Clean up the kernel thread */
-	kthread_stop(tr->blkcore_priv->thread);
-
 	/* Remove it from the list of active majors */
 	list_del(&tr->list);
 
 	list_for_each_entry_safe(dev, next, &tr->devs, list)
 		tr->remove_dev(dev);
 
-	blk_cleanup_queue(tr->blkcore_priv->rq);
 	unregister_blkdev(tr->major, tr->name);
-
 	mutex_unlock(&mtd_table_mutex);
 
-	kfree(tr->blkcore_priv);
-
 	BUG_ON(!list_empty(&tr->devs));
 	return 0;
 }