Btrfs: add new sources for device replace code

This adds a new file to the sources together with the header file
and the changes to ioctl.h and ctree.h that are required by the
new C source file. Additionally, 4 new functions are added to
volume.c that deal with device creation and destruction.

Signed-off-by: Stefan Behrens <sbehrens@giantdisaster.de>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>

fs/btrfs/Makefile

@@ -8,7 +8,7 @@ btrfs-y += super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \
 	   extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \
 	   export.o tree-log.o free-space-cache.o zlib.o lzo.o \
 	   compression.o delayed-ref.o relocation.o delayed-inode.o scrub.o \
-	   reada.o backref.o ulist.o qgroup.o send.o
+	   reada.o backref.o ulist.o qgroup.o send.o dev-replace.o
 
 btrfs-$(CONFIG_BTRFS_FS_POSIX_ACL) += acl.o
 btrfs-$(CONFIG_BTRFS_FS_CHECK_INTEGRITY) += check-integrity.o

fs/btrfs/ctree.h

@@ -142,6 +142,8 @@ struct btrfs_ordered_sum;
 
 #define BTRFS_EMPTY_SUBVOL_DIR_OBJECTID 2
 
+#define BTRFS_DEV_REPLACE_DEVID 0
+
 /*
  * the max metadata block size.  This limit is somewhat artificial,
  * but the memmove costs go through the roof for larger blocks.

fs/btrfs/dev-replace.h

@@ -19,6 +19,24 @@
 #if !defined(__BTRFS_DEV_REPLACE__)
 #define __BTRFS_DEV_REPLACE__
 
+struct btrfs_ioctl_dev_replace_args;
+
+int btrfs_init_dev_replace(struct btrfs_fs_info *fs_info);
+int btrfs_run_dev_replace(struct btrfs_trans_handle *trans,
+			  struct btrfs_fs_info *fs_info);
+void btrfs_after_dev_replace_commit(struct btrfs_fs_info *fs_info);
+int btrfs_dev_replace_start(struct btrfs_root *root,
+			    struct btrfs_ioctl_dev_replace_args *args);
+void btrfs_dev_replace_status(struct btrfs_fs_info *fs_info,
+			      struct btrfs_ioctl_dev_replace_args *args);
+int btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info,
+			     struct btrfs_ioctl_dev_replace_args *args);
+void btrfs_dev_replace_suspend_for_unmount(struct btrfs_fs_info *fs_info);
+int btrfs_resume_dev_replace_async(struct btrfs_fs_info *fs_info);
+int btrfs_dev_replace_is_ongoing(struct btrfs_dev_replace *dev_replace);
+void btrfs_dev_replace_lock(struct btrfs_dev_replace *dev_replace);
+void btrfs_dev_replace_unlock(struct btrfs_dev_replace *dev_replace);
+
 static inline void btrfs_dev_replace_stats_inc(atomic64_t *stat_value)
 {
 	atomic64_inc(stat_value);

fs/btrfs/ioctl.h

@@ -123,6 +123,48 @@ struct btrfs_ioctl_scrub_args {
 	__u64 unused[(1024-32-sizeof(struct btrfs_scrub_progress))/8];
 };
 
+#define BTRFS_IOCTL_DEV_REPLACE_CONT_READING_FROM_SRCDEV_MODE_ALWAYS	0
+#define BTRFS_IOCTL_DEV_REPLACE_CONT_READING_FROM_SRCDEV_MODE_AVOID	1
+struct btrfs_ioctl_dev_replace_start_params {
+	__u64 srcdevid;	/* in, if 0, use srcdev_name instead */
+	__u8 srcdev_name[BTRFS_PATH_NAME_MAX + 1];	/* in */
+	__u8 tgtdev_name[BTRFS_PATH_NAME_MAX + 1];	/* in */
+	__u64 cont_reading_from_srcdev_mode;	/* in, see #define
+						 * above */
+};
+
+#define BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED	0
+#define BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED		1
+#define BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED		2
+#define BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED		3
+#define BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED		4
+struct btrfs_ioctl_dev_replace_status_params {
+	__u64 replace_state;	/* out, see #define above */
+	__u64 progress_1000;	/* out, 0 <= x <= 1000 */
+	__u64 time_started;	/* out, seconds since 1-Jan-1970 */
+	__u64 time_stopped;	/* out, seconds since 1-Jan-1970 */
+	__u64 num_write_errors;	/* out */
+	__u64 num_uncorrectable_read_errors;	/* out */
+};
+
+#define BTRFS_IOCTL_DEV_REPLACE_CMD_START			0
+#define BTRFS_IOCTL_DEV_REPLACE_CMD_STATUS			1
+#define BTRFS_IOCTL_DEV_REPLACE_CMD_CANCEL			2
+#define BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR			0
+#define BTRFS_IOCTL_DEV_REPLACE_RESULT_NOT_STARTED		1
+#define BTRFS_IOCTL_DEV_REPLACE_RESULT_ALREADY_STARTED		2
+struct btrfs_ioctl_dev_replace_args {
+	__u64 cmd;	/* in */
+	__u64 result;	/* out */
+
+	union {
+		struct btrfs_ioctl_dev_replace_start_params start;
+		struct btrfs_ioctl_dev_replace_status_params status;
+	};	/* in/out */
+
+	__u64 spare[64];
+};
+
 #define BTRFS_DEVICE_PATH_NAME_MAX 1024
 struct btrfs_ioctl_dev_info_args {
 	__u64 devid;				/* in/out */
@@ -453,4 +495,7 @@ struct btrfs_ioctl_send_args {
 			       struct btrfs_ioctl_qgroup_limit_args)
 #define BTRFS_IOC_GET_DEV_STATS _IOWR(BTRFS_IOCTL_MAGIC, 52, \
 				      struct btrfs_ioctl_get_dev_stats)
+#define BTRFS_IOC_DEV_REPLACE _IOWR(BTRFS_IOCTL_MAGIC, 53, \
+				    struct btrfs_ioctl_dev_replace_args)
+
 #endif

fs/btrfs/volumes.c

@@ -1537,6 +1537,53 @@ error_undo:
 	goto error_brelse;
 }
 
+void btrfs_rm_dev_replace_srcdev(struct btrfs_fs_info *fs_info,
+				 struct btrfs_device *srcdev)
+{
+	WARN_ON(!mutex_is_locked(&fs_info->fs_devices->device_list_mutex));
+	list_del_rcu(&srcdev->dev_list);
+	list_del_rcu(&srcdev->dev_alloc_list);
+	fs_info->fs_devices->num_devices--;
+	if (srcdev->missing) {
+		fs_info->fs_devices->missing_devices--;
+		fs_info->fs_devices->rw_devices++;
+	}
+	if (srcdev->can_discard)
+		fs_info->fs_devices->num_can_discard--;
+	if (srcdev->bdev)
+		fs_info->fs_devices->open_devices--;
+
+	call_rcu(&srcdev->rcu, free_device);
+}
+
+void btrfs_destroy_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
+				      struct btrfs_device *tgtdev)
+{
+	struct btrfs_device *next_device;
+
+	WARN_ON(!tgtdev);
+	mutex_lock(&fs_info->fs_devices->device_list_mutex);
+	if (tgtdev->bdev) {
+		btrfs_scratch_superblock(tgtdev);
+		fs_info->fs_devices->open_devices--;
+	}
+	fs_info->fs_devices->num_devices--;
+	if (tgtdev->can_discard)
+		fs_info->fs_devices->num_can_discard++;
+
+	next_device = list_entry(fs_info->fs_devices->devices.next,
+				 struct btrfs_device, dev_list);
+	if (tgtdev->bdev == fs_info->sb->s_bdev)
+		fs_info->sb->s_bdev = next_device->bdev;
+	if (tgtdev->bdev == fs_info->fs_devices->latest_bdev)
+		fs_info->fs_devices->latest_bdev = next_device->bdev;
+	list_del_rcu(&tgtdev->dev_list);
+
+	call_rcu(&tgtdev->rcu, free_device);
+
+	mutex_unlock(&fs_info->fs_devices->device_list_mutex);
+}
+
 int btrfs_find_device_by_path(struct btrfs_root *root, char *device_path,
 			      struct btrfs_device **device)
 {
@@ -1931,6 +1978,98 @@ error:
 	return ret;
 }
 
+int btrfs_init_dev_replace_tgtdev(struct btrfs_root *root, char *device_path,
+				  struct btrfs_device **device_out)
+{
+	struct request_queue *q;
+	struct btrfs_device *device;
+	struct block_device *bdev;
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	struct list_head *devices;
+	struct rcu_string *name;
+	int ret = 0;
+
+	*device_out = NULL;
+	if (fs_info->fs_devices->seeding)
+		return -EINVAL;
+
+	bdev = blkdev_get_by_path(device_path, FMODE_WRITE | FMODE_EXCL,
+				  fs_info->bdev_holder);
+	if (IS_ERR(bdev))
+		return PTR_ERR(bdev);
+
+	filemap_write_and_wait(bdev->bd_inode->i_mapping);
+
+	devices = &fs_info->fs_devices->devices;
+	list_for_each_entry(device, devices, dev_list) {
+		if (device->bdev == bdev) {
+			ret = -EEXIST;
+			goto error;
+		}
+	}
+
+	device = kzalloc(sizeof(*device), GFP_NOFS);
+	if (!device) {
+		ret = -ENOMEM;
+		goto error;
+	}
+
+	name = rcu_string_strdup(device_path, GFP_NOFS);
+	if (!name) {
+		kfree(device);
+		ret = -ENOMEM;
+		goto error;
+	}
+	rcu_assign_pointer(device->name, name);
+
+	q = bdev_get_queue(bdev);
+	if (blk_queue_discard(q))
+		device->can_discard = 1;
+	mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
+	device->writeable = 1;
+	device->work.func = pending_bios_fn;
+	generate_random_uuid(device->uuid);
+	device->devid = BTRFS_DEV_REPLACE_DEVID;
+	spin_lock_init(&device->io_lock);
+	device->generation = 0;
+	device->io_width = root->sectorsize;
+	device->io_align = root->sectorsize;
+	device->sector_size = root->sectorsize;
+	device->total_bytes = i_size_read(bdev->bd_inode);
+	device->disk_total_bytes = device->total_bytes;
+	device->dev_root = fs_info->dev_root;
+	device->bdev = bdev;
+	device->in_fs_metadata = 1;
+	device->is_tgtdev_for_dev_replace = 1;
+	device->mode = FMODE_EXCL;
+	set_blocksize(device->bdev, 4096);
+	device->fs_devices = fs_info->fs_devices;
+	list_add(&device->dev_list, &fs_info->fs_devices->devices);
+	fs_info->fs_devices->num_devices++;
+	fs_info->fs_devices->open_devices++;
+	if (device->can_discard)
+		fs_info->fs_devices->num_can_discard++;
+	mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+
+	*device_out = device;
+	return ret;
+
+error:
+	blkdev_put(bdev, FMODE_EXCL);
+	return ret;
+}
+
+void btrfs_init_dev_replace_tgtdev_for_resume(struct btrfs_fs_info *fs_info,
+					      struct btrfs_device *tgtdev)
+{
+	WARN_ON(fs_info->fs_devices->rw_devices == 0);
+	tgtdev->io_width = fs_info->dev_root->sectorsize;
+	tgtdev->io_align = fs_info->dev_root->sectorsize;
+	tgtdev->sector_size = fs_info->dev_root->sectorsize;
+	tgtdev->dev_root = fs_info->dev_root;
+	tgtdev->in_fs_metadata = 1;
+}
+
 static noinline int btrfs_update_device(struct btrfs_trans_handle *trans,
 					struct btrfs_device *device)
 {

fs/btrfs/volumes.h

@@ -286,6 +286,8 @@ struct btrfs_device *btrfs_find_device(struct btrfs_fs_info *fs_info, u64 devid,
 				       u8 *uuid, u8 *fsid);
 int btrfs_shrink_device(struct btrfs_device *device, u64 new_size);
 int btrfs_init_new_device(struct btrfs_root *root, char *path);
+int btrfs_init_dev_replace_tgtdev(struct btrfs_root *root, char *device_path,
+				  struct btrfs_device **device_out);
 int btrfs_balance(struct btrfs_balance_control *bctl,
 		  struct btrfs_ioctl_balance_args *bargs);
 int btrfs_resume_balance_async(struct btrfs_fs_info *fs_info);
@@ -302,6 +304,12 @@ int btrfs_get_dev_stats(struct btrfs_root *root,
 int btrfs_init_dev_stats(struct btrfs_fs_info *fs_info);
 int btrfs_run_dev_stats(struct btrfs_trans_handle *trans,
 			struct btrfs_fs_info *fs_info);
+void btrfs_rm_dev_replace_srcdev(struct btrfs_fs_info *fs_info,
+				 struct btrfs_device *srcdev);
+void btrfs_destroy_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
+				      struct btrfs_device *tgtdev);
+void btrfs_init_dev_replace_tgtdev_for_resume(struct btrfs_fs_info *fs_info,
+					      struct btrfs_device *tgtdev);
 int btrfs_scratch_superblock(struct btrfs_device *device);
 
 static inline void btrfs_dev_stat_inc(struct btrfs_device *dev,