#include <linux/blkdev.h>
#include <linux/random.h>
#include <asm/div64.h>
+#include "compat.h"
#include "ctree.h"
#include "extent_map.h"
#include "disk-io.h"
#include "transaction.h"
#include "print-tree.h"
#include "volumes.h"
+#include "async-thread.h"
struct map_lookup {
u64 type;
struct btrfs_bio_stripe stripes[];
};
+static int init_first_rw_device(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_device *device);
+static int btrfs_relocate_sys_chunks(struct btrfs_root *root);
+
#define map_lookup_size(n) (sizeof(struct map_lookup) + \
(sizeof(struct btrfs_bio_stripe) * (n)))
mutex_unlock(&uuid_mutex);
}
+static void lock_chunks(struct btrfs_root *root)
+{
+ mutex_lock(&root->fs_info->chunk_mutex);
+}
+
+static void unlock_chunks(struct btrfs_root *root)
+{
+ mutex_unlock(&root->fs_info->chunk_mutex);
+}
+
+static void free_fs_devices(struct btrfs_fs_devices *fs_devices)
+{
+ struct btrfs_device *device;
+ WARN_ON(fs_devices->opened);
+ while (!list_empty(&fs_devices->devices)) {
+ device = list_entry(fs_devices->devices.next,
+ struct btrfs_device, dev_list);
+ list_del(&device->dev_list);
+ kfree(device->name);
+ kfree(device);
+ }
+ kfree(fs_devices);
+}
+
int btrfs_cleanup_fs_uuids(void)
{
struct btrfs_fs_devices *fs_devices;
- struct list_head *uuid_cur;
- struct list_head *devices_cur;
- struct btrfs_device *dev;
- list_for_each(uuid_cur, &fs_uuids) {
- fs_devices = list_entry(uuid_cur, struct btrfs_fs_devices,
- list);
- while(!list_empty(&fs_devices->devices)) {
- devices_cur = fs_devices->devices.next;
- dev = list_entry(devices_cur, struct btrfs_device,
- dev_list);
- if (dev->bdev) {
- close_bdev_excl(dev->bdev);
- fs_devices->open_devices--;
- }
- list_del(&dev->dev_list);
- kfree(dev->name);
- kfree(dev);
- }
+ while (!list_empty(&fs_uuids)) {
+ fs_devices = list_entry(fs_uuids.next,
+ struct btrfs_fs_devices, list);
+ list_del(&fs_devices->list);
+ free_fs_devices(fs_devices);
}
return 0;
}
-static struct btrfs_device *__find_device(struct list_head *head, u64 devid,
- u8 *uuid)
+static noinline struct btrfs_device *__find_device(struct list_head *head,
+ u64 devid, u8 *uuid)
{
struct btrfs_device *dev;
- struct list_head *cur;
- list_for_each(cur, head) {
- dev = list_entry(cur, struct btrfs_device, dev_list);
+ list_for_each_entry(dev, head, dev_list) {
if (dev->devid == devid &&
(!uuid || !memcmp(dev->uuid, uuid, BTRFS_UUID_SIZE))) {
return dev;
return NULL;
}
-static struct btrfs_fs_devices *find_fsid(u8 *fsid)
+static noinline struct btrfs_fs_devices *find_fsid(u8 *fsid)
{
- struct list_head *cur;
struct btrfs_fs_devices *fs_devices;
- list_for_each(cur, &fs_uuids) {
- fs_devices = list_entry(cur, struct btrfs_fs_devices, list);
+ list_for_each_entry(fs_devices, &fs_uuids, list) {
if (memcmp(fsid, fs_devices->fsid, BTRFS_FSID_SIZE) == 0)
return fs_devices;
}
return NULL;
}
-static int device_list_add(const char *path,
+/*
+ * we try to collect pending bios for a device so we don't get a large
+ * number of procs sending bios down to the same device. This greatly
+ * improves the schedulers ability to collect and merge the bios.
+ *
+ * But, it also turns into a long list of bios to process and that is sure
+ * to eventually make the worker thread block. The solution here is to
+ * make some progress and then put this work struct back at the end of
+ * the list if the block device is congested. This way, multiple devices
+ * can make progress from a single worker thread.
+ */
+static noinline int run_scheduled_bios(struct btrfs_device *device)
+{
+ struct bio *pending;
+ struct backing_dev_info *bdi;
+ struct btrfs_fs_info *fs_info;
+ struct bio *tail;
+ struct bio *cur;
+ int again = 0;
+ unsigned long num_run = 0;
+ unsigned long limit;
+
+ bdi = device->bdev->bd_inode->i_mapping->backing_dev_info;
+ fs_info = device->dev_root->fs_info;
+ limit = btrfs_async_submit_limit(fs_info);
+ limit = limit * 2 / 3;
+
+loop:
+ spin_lock(&device->io_lock);
+
+ /* take all the bios off the list at once and process them
+ * later on (without the lock held). But, remember the
+ * tail and other pointers so the bios can be properly reinserted
+ * into the list if we hit congestion
+ */
+ pending = device->pending_bios;
+ tail = device->pending_bio_tail;
+ WARN_ON(pending && !tail);
+ device->pending_bios = NULL;
+ device->pending_bio_tail = NULL;
+
+ /*
+ * if pending was null this time around, no bios need processing
+ * at all and we can stop. Otherwise it'll loop back up again
+ * and do an additional check so no bios are missed.
+ *
+ * device->running_pending is used to synchronize with the
+ * schedule_bio code.
+ */
+ if (pending) {
+ again = 1;
+ device->running_pending = 1;
+ } else {
+ again = 0;
+ device->running_pending = 0;
+ }
+ spin_unlock(&device->io_lock);
+
+ while (pending) {
+ cur = pending;
+ pending = pending->bi_next;
+ cur->bi_next = NULL;
+ atomic_dec(&fs_info->nr_async_bios);
+
+ if (atomic_read(&fs_info->nr_async_bios) < limit &&
+ waitqueue_active(&fs_info->async_submit_wait))
+ wake_up(&fs_info->async_submit_wait);
+
+ BUG_ON(atomic_read(&cur->bi_cnt) == 0);
+ bio_get(cur);
+ submit_bio(cur->bi_rw, cur);
+ bio_put(cur);
+ num_run++;
+
+ /*
+ * we made progress, there is more work to do and the bdi
+ * is now congested. Back off and let other work structs
+ * run instead
+ */
+ if (pending && bdi_write_congested(bdi) &&
+ fs_info->fs_devices->open_devices > 1) {
+ struct bio *old_head;
+
+ spin_lock(&device->io_lock);
+
+ old_head = device->pending_bios;
+ device->pending_bios = pending;
+ if (device->pending_bio_tail)
+ tail->bi_next = old_head;
+ else
+ device->pending_bio_tail = tail;
+ device->running_pending = 0;
+
+ spin_unlock(&device->io_lock);
+ btrfs_requeue_work(&device->work);
+ goto done;
+ }
+ }
+ if (again)
+ goto loop;
+done:
+ return 0;
+}
+
+static void pending_bios_fn(struct btrfs_work *work)
+{
+ struct btrfs_device *device;
+
+ device = container_of(work, struct btrfs_device, work);
+ run_scheduled_bios(device);
+}
+
+static noinline int device_list_add(const char *path,
struct btrfs_super_block *disk_super,
u64 devid, struct btrfs_fs_devices **fs_devices_ret)
{
fs_devices = find_fsid(disk_super->fsid);
if (!fs_devices) {
- fs_devices = kmalloc(sizeof(*fs_devices), GFP_NOFS);
+ fs_devices = kzalloc(sizeof(*fs_devices), GFP_NOFS);
if (!fs_devices)
return -ENOMEM;
INIT_LIST_HEAD(&fs_devices->devices);
memcpy(fs_devices->fsid, disk_super->fsid, BTRFS_FSID_SIZE);
fs_devices->latest_devid = devid;
fs_devices->latest_trans = found_transid;
- fs_devices->num_devices = 0;
device = NULL;
} else {
device = __find_device(&fs_devices->devices, devid,
disk_super->dev_item.uuid);
}
if (!device) {
+ if (fs_devices->opened)
+ return -EBUSY;
+
device = kzalloc(sizeof(*device), GFP_NOFS);
if (!device) {
/* we can safely leave the fs_devices entry around */
return -ENOMEM;
}
device->devid = devid;
+ device->work.func = pending_bios_fn;
memcpy(device->uuid, disk_super->dev_item.uuid,
BTRFS_UUID_SIZE);
device->barriers = 1;
kfree(device);
return -ENOMEM;
}
+ INIT_LIST_HEAD(&device->dev_alloc_list);
list_add(&device->dev_list, &fs_devices->devices);
- list_add(&device->dev_alloc_list, &fs_devices->alloc_list);
+ device->fs_devices = fs_devices;
fs_devices->num_devices++;
}
return 0;
}
-int btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices)
+static struct btrfs_fs_devices *clone_fs_devices(struct btrfs_fs_devices *orig)
{
- struct list_head *head = &fs_devices->devices;
- struct list_head *cur;
+ struct btrfs_fs_devices *fs_devices;
struct btrfs_device *device;
+ struct btrfs_device *orig_dev;
+
+ fs_devices = kzalloc(sizeof(*fs_devices), GFP_NOFS);
+ if (!fs_devices)
+ return ERR_PTR(-ENOMEM);
+
+ INIT_LIST_HEAD(&fs_devices->devices);
+ INIT_LIST_HEAD(&fs_devices->alloc_list);
+ INIT_LIST_HEAD(&fs_devices->list);
+ fs_devices->latest_devid = orig->latest_devid;
+ fs_devices->latest_trans = orig->latest_trans;
+ memcpy(fs_devices->fsid, orig->fsid, sizeof(fs_devices->fsid));
+
+ list_for_each_entry(orig_dev, &orig->devices, dev_list) {
+ device = kzalloc(sizeof(*device), GFP_NOFS);
+ if (!device)
+ goto error;
+
+ device->name = kstrdup(orig_dev->name, GFP_NOFS);
+ if (!device->name)
+ goto error;
+
+ device->devid = orig_dev->devid;
+ device->work.func = pending_bios_fn;
+ memcpy(device->uuid, orig_dev->uuid, sizeof(device->uuid));
+ device->barriers = 1;
+ spin_lock_init(&device->io_lock);
+ INIT_LIST_HEAD(&device->dev_list);
+ INIT_LIST_HEAD(&device->dev_alloc_list);
+
+ list_add(&device->dev_list, &fs_devices->devices);
+ device->fs_devices = fs_devices;
+ fs_devices->num_devices++;
+ }
+ return fs_devices;
+error:
+ free_fs_devices(fs_devices);
+ return ERR_PTR(-ENOMEM);
+}
+
+int btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices)
+{
+ struct btrfs_device *device, *next;
mutex_lock(&uuid_mutex);
again:
- list_for_each(cur, head) {
- device = list_entry(cur, struct btrfs_device, dev_list);
- if (!device->in_fs_metadata) {
- if (device->bdev) {
- close_bdev_excl(device->bdev);
- fs_devices->open_devices--;
- }
- list_del(&device->dev_list);
- list_del(&device->dev_alloc_list);
- fs_devices->num_devices--;
- kfree(device->name);
- kfree(device);
- goto again;
+ list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) {
+ if (device->in_fs_metadata)
+ continue;
+
+ if (device->bdev) {
+ close_bdev_exclusive(device->bdev, device->mode);
+ device->bdev = NULL;
+ fs_devices->open_devices--;
}
+ if (device->writeable) {
+ list_del_init(&device->dev_alloc_list);
+ device->writeable = 0;
+ fs_devices->rw_devices--;
+ }
+ list_del_init(&device->dev_list);
+ fs_devices->num_devices--;
+ kfree(device->name);
+ kfree(device);
+ }
+
+ if (fs_devices->seed) {
+ fs_devices = fs_devices->seed;
+ goto again;
}
+
mutex_unlock(&uuid_mutex);
return 0;
}
-int btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
+static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
{
- struct list_head *head = &fs_devices->devices;
- struct list_head *cur;
struct btrfs_device *device;
- mutex_lock(&uuid_mutex);
- list_for_each(cur, head) {
- device = list_entry(cur, struct btrfs_device, dev_list);
+ if (--fs_devices->opened > 0)
+ return 0;
+
+ list_for_each_entry(device, &fs_devices->devices, dev_list) {
if (device->bdev) {
- close_bdev_excl(device->bdev);
+ close_bdev_exclusive(device->bdev, device->mode);
fs_devices->open_devices--;
}
+ if (device->writeable) {
+ list_del_init(&device->dev_alloc_list);
+ fs_devices->rw_devices--;
+ }
+
device->bdev = NULL;
+ device->writeable = 0;
device->in_fs_metadata = 0;
}
- fs_devices->mounted = 0;
- mutex_unlock(&uuid_mutex);
+ WARN_ON(fs_devices->open_devices);
+ WARN_ON(fs_devices->rw_devices);
+ fs_devices->opened = 0;
+ fs_devices->seeding = 0;
+
return 0;
}
-int btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
- int flags, void *holder)
+int btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
+{
+ struct btrfs_fs_devices *seed_devices = NULL;
+ int ret;
+
+ mutex_lock(&uuid_mutex);
+ ret = __btrfs_close_devices(fs_devices);
+ if (!fs_devices->opened) {
+ seed_devices = fs_devices->seed;
+ fs_devices->seed = NULL;
+ }
+ mutex_unlock(&uuid_mutex);
+
+ while (seed_devices) {
+ fs_devices = seed_devices;
+ seed_devices = fs_devices->seed;
+ __btrfs_close_devices(fs_devices);
+ free_fs_devices(fs_devices);
+ }
+ return ret;
+}
+
+static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
+ fmode_t flags, void *holder)
{
struct block_device *bdev;
struct list_head *head = &fs_devices->devices;
- struct list_head *cur;
struct btrfs_device *device;
struct block_device *latest_bdev = NULL;
struct buffer_head *bh;
struct btrfs_super_block *disk_super;
u64 latest_devid = 0;
u64 latest_transid = 0;
- u64 transid;
u64 devid;
+ int seeding = 1;
int ret = 0;
- mutex_lock(&uuid_mutex);
- if (fs_devices->mounted)
- goto out;
-
- list_for_each(cur, head) {
- device = list_entry(cur, struct btrfs_device, dev_list);
+ list_for_each_entry(device, head, dev_list) {
if (device->bdev)
continue;
-
if (!device->name)
continue;
- bdev = open_bdev_excl(device->name, flags, holder);
-
+ bdev = open_bdev_exclusive(device->name, flags, holder);
if (IS_ERR(bdev)) {
- printk("open %s failed\n", device->name);
+ printk(KERN_INFO "open %s failed\n", device->name);
goto error;
}
set_blocksize(bdev, 4096);
- bh = __bread(bdev, BTRFS_SUPER_INFO_OFFSET / 4096, 4096);
+ bh = btrfs_read_dev_super(bdev);
if (!bh)
goto error_close;
disk_super = (struct btrfs_super_block *)bh->b_data;
- if (strncmp((char *)(&disk_super->magic), BTRFS_MAGIC,
- sizeof(disk_super->magic)))
- goto error_brelse;
-
devid = le64_to_cpu(disk_super->dev_item.devid);
if (devid != device->devid)
goto error_brelse;
- transid = btrfs_super_generation(disk_super);
- if (transid > latest_transid) {
+ if (memcmp(device->uuid, disk_super->dev_item.uuid,
+ BTRFS_UUID_SIZE))
+ goto error_brelse;
+
+ device->generation = btrfs_super_generation(disk_super);
+ if (!latest_transid || device->generation > latest_transid) {
latest_devid = devid;
- latest_transid = transid;
+ latest_transid = device->generation;
latest_bdev = bdev;
}
+ if (btrfs_super_flags(disk_super) & BTRFS_SUPER_FLAG_SEEDING) {
+ device->writeable = 0;
+ } else {
+ device->writeable = !bdev_read_only(bdev);
+ seeding = 0;
+ }
+
device->bdev = bdev;
device->in_fs_metadata = 0;
+ device->mode = flags;
+
fs_devices->open_devices++;
+ if (device->writeable) {
+ fs_devices->rw_devices++;
+ list_add(&device->dev_alloc_list,
+ &fs_devices->alloc_list);
+ }
continue;
error_brelse:
brelse(bh);
error_close:
- close_bdev_excl(bdev);
+ close_bdev_exclusive(bdev, FMODE_READ);
error:
continue;
}
ret = -EIO;
goto out;
}
- fs_devices->mounted = 1;
+ fs_devices->seeding = seeding;
+ fs_devices->opened = 1;
fs_devices->latest_bdev = latest_bdev;
fs_devices->latest_devid = latest_devid;
fs_devices->latest_trans = latest_transid;
+ fs_devices->total_rw_bytes = 0;
out:
+ return ret;
+}
+
+int btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
+ fmode_t flags, void *holder)
+{
+ int ret;
+
+ mutex_lock(&uuid_mutex);
+ if (fs_devices->opened) {
+ fs_devices->opened++;
+ ret = 0;
+ } else {
+ ret = __btrfs_open_devices(fs_devices, flags, holder);
+ }
mutex_unlock(&uuid_mutex);
return ret;
}
-int btrfs_scan_one_device(const char *path, int flags, void *holder,
+int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
struct btrfs_fs_devices **fs_devices_ret)
{
struct btrfs_super_block *disk_super;
mutex_lock(&uuid_mutex);
- bdev = open_bdev_excl(path, flags, holder);
+ bdev = open_bdev_exclusive(path, flags, holder);
if (IS_ERR(bdev)) {
ret = PTR_ERR(bdev);
ret = set_blocksize(bdev, 4096);
if (ret)
goto error_close;
- bh = __bread(bdev, BTRFS_SUPER_INFO_OFFSET / 4096, 4096);
+ bh = btrfs_read_dev_super(bdev);
if (!bh) {
ret = -EIO;
goto error_close;
}
disk_super = (struct btrfs_super_block *)bh->b_data;
- if (strncmp((char *)(&disk_super->magic), BTRFS_MAGIC,
- sizeof(disk_super->magic))) {
- ret = -EINVAL;
- goto error_brelse;
- }
devid = le64_to_cpu(disk_super->dev_item.devid);
transid = btrfs_super_generation(disk_super);
if (disk_super->label[0])
- printk("device label %s ", disk_super->label);
+ printk(KERN_INFO "device label %s ", disk_super->label);
else {
/* FIXME, make a readl uuid parser */
- printk("device fsid %llx-%llx ",
+ printk(KERN_INFO "device fsid %llx-%llx ",
*(unsigned long long *)disk_super->fsid,
*(unsigned long long *)(disk_super->fsid + 8));
}
- printk("devid %Lu transid %Lu %s\n", devid, transid, path);
+ printk(KERN_CONT "devid %llu transid %llu %s\n",
+ (unsigned long long)devid, (unsigned long long)transid, path);
ret = device_list_add(path, disk_super, devid, fs_devices_ret);
-error_brelse:
brelse(bh);
error_close:
- close_bdev_excl(bdev);
+ close_bdev_exclusive(bdev, flags);
error:
mutex_unlock(&uuid_mutex);
return ret;
* called very infrequently and that a given device has a small number
* of extents
*/
-static int find_free_dev_extent(struct btrfs_trans_handle *trans,
- struct btrfs_device *device,
- struct btrfs_path *path,
- u64 num_bytes, u64 *start)
+static noinline int find_free_dev_extent(struct btrfs_trans_handle *trans,
+ struct btrfs_device *device,
+ u64 num_bytes, u64 *start)
{
struct btrfs_key key;
struct btrfs_root *root = device->dev_root;
struct btrfs_dev_extent *dev_extent = NULL;
+ struct btrfs_path *path;
u64 hole_size = 0;
u64 last_byte = 0;
u64 search_start = 0;
int start_found;
struct extent_buffer *l;
- start_found = 0;
+ path = btrfs_alloc_path();
+ if (!path)
+ return -ENOMEM;
path->reada = 2;
+ start_found = 0;
/* FIXME use last free of some kind */
goto check_pending;
}
}
- if (btrfs_key_type(&key) != BTRFS_DEV_EXTENT_KEY) {
+ if (btrfs_key_type(&key) != BTRFS_DEV_EXTENT_KEY)
goto next;
- }
start_found = 1;
dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent);
/* we have to make sure we didn't find an extent that has already
* been allocated by the map tree or the original allocation
*/
- btrfs_release_path(root, path);
BUG_ON(*start < search_start);
if (*start + num_bytes > search_end) {
goto error;
}
/* check for pending inserts here */
- return 0;
+ ret = 0;
error:
- btrfs_release_path(root, path);
+ btrfs_free_path(path);
return ret;
}
-int btrfs_free_dev_extent(struct btrfs_trans_handle *trans,
+static int btrfs_free_dev_extent(struct btrfs_trans_handle *trans,
struct btrfs_device *device,
u64 start)
{
int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans,
struct btrfs_device *device,
u64 chunk_tree, u64 chunk_objectid,
- u64 chunk_offset,
- u64 num_bytes, u64 *start)
+ u64 chunk_offset, u64 start, u64 num_bytes)
{
int ret;
struct btrfs_path *path;
if (!path)
return -ENOMEM;
- ret = find_free_dev_extent(trans, device, path, num_bytes, start);
- if (ret) {
- goto err;
- }
-
key.objectid = device->devid;
- key.offset = *start;
+ key.offset = start;
key.type = BTRFS_DEV_EXTENT_KEY;
ret = btrfs_insert_empty_item(trans, root, path, &key,
sizeof(*extent));
btrfs_set_dev_extent_length(leaf, extent, num_bytes);
btrfs_mark_buffer_dirty(leaf);
-err:
btrfs_free_path(path);
return ret;
}
-static int find_next_chunk(struct btrfs_root *root, u64 objectid, u64 *offset)
+static noinline int find_next_chunk(struct btrfs_root *root,
+ u64 objectid, u64 *offset)
{
struct btrfs_path *path;
int ret;
return ret;
}
-static int find_next_devid(struct btrfs_root *root, struct btrfs_path *path,
- u64 *objectid)
+static noinline int find_next_devid(struct btrfs_root *root, u64 *objectid)
{
int ret;
struct btrfs_key key;
struct btrfs_key found_key;
+ struct btrfs_path *path;
+
+ root = root->fs_info->chunk_root;
+
+ path = btrfs_alloc_path();
+ if (!path)
+ return -ENOMEM;
key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
key.type = BTRFS_DEV_ITEM_KEY;
}
ret = 0;
error:
- btrfs_release_path(root, path);
+ btrfs_free_path(path);
return ret;
}
struct extent_buffer *leaf;
struct btrfs_key key;
unsigned long ptr;
- u64 free_devid = 0;
root = root->fs_info->chunk_root;
if (!path)
return -ENOMEM;
- ret = find_next_devid(root, path, &free_devid);
- if (ret)
- goto out;
-
key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
key.type = BTRFS_DEV_ITEM_KEY;
- key.offset = free_devid;
+ key.offset = device->devid;
ret = btrfs_insert_empty_item(trans, root, path, &key,
sizeof(*dev_item));
leaf = path->nodes[0];
dev_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dev_item);
- device->devid = free_devid;
btrfs_set_device_id(leaf, dev_item, device->devid);
+ btrfs_set_device_generation(leaf, dev_item, 0);
btrfs_set_device_type(leaf, dev_item, device->type);
btrfs_set_device_io_align(leaf, dev_item, device->io_align);
btrfs_set_device_io_width(leaf, dev_item, device->io_width);
btrfs_set_device_group(leaf, dev_item, 0);
btrfs_set_device_seek_speed(leaf, dev_item, 0);
btrfs_set_device_bandwidth(leaf, dev_item, 0);
+ btrfs_set_device_start_offset(leaf, dev_item, 0);
ptr = (unsigned long)btrfs_device_uuid(dev_item);
write_extent_buffer(leaf, device->uuid, ptr, BTRFS_UUID_SIZE);
+ ptr = (unsigned long)btrfs_device_fsid(dev_item);
+ write_extent_buffer(leaf, root->fs_info->fsid, ptr, BTRFS_UUID_SIZE);
btrfs_mark_buffer_dirty(leaf);
- ret = 0;
+ ret = 0;
out:
btrfs_free_path(path);
return ret;
{
int ret;
struct btrfs_path *path;
- struct block_device *bdev = device->bdev;
- struct btrfs_device *next_dev;
struct btrfs_key key;
- u64 total_bytes;
- struct btrfs_fs_devices *fs_devices;
struct btrfs_trans_handle *trans;
root = root->fs_info->chunk_root;
key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
key.type = BTRFS_DEV_ITEM_KEY;
key.offset = device->devid;
+ lock_chunks(root);
ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
if (ret < 0)
ret = btrfs_del_item(trans, root, path);
if (ret)
goto out;
-
- /*
- * at this point, the device is zero sized. We want to
- * remove it from the devices list and zero out the old super
- */
- list_del_init(&device->dev_list);
- list_del_init(&device->dev_alloc_list);
- fs_devices = root->fs_info->fs_devices;
-
- next_dev = list_entry(fs_devices->devices.next, struct btrfs_device,
- dev_list);
- if (bdev == root->fs_info->sb->s_bdev)
- root->fs_info->sb->s_bdev = next_dev->bdev;
- if (bdev == fs_devices->latest_bdev)
- fs_devices->latest_bdev = next_dev->bdev;
-
- total_bytes = btrfs_super_total_bytes(&root->fs_info->super_copy);
- btrfs_set_super_total_bytes(&root->fs_info->super_copy,
- total_bytes - device->total_bytes);
-
- total_bytes = btrfs_super_num_devices(&root->fs_info->super_copy);
- btrfs_set_super_num_devices(&root->fs_info->super_copy,
- total_bytes - 1);
out:
btrfs_free_path(path);
+ unlock_chunks(root);
btrfs_commit_transaction(trans, root);
return ret;
}
int btrfs_rm_device(struct btrfs_root *root, char *device_path)
{
struct btrfs_device *device;
+ struct btrfs_device *next_device;
struct block_device *bdev;
struct buffer_head *bh = NULL;
struct btrfs_super_block *disk_super;
u64 all_avail;
u64 devid;
+ u64 num_devices;
+ u8 *dev_uuid;
int ret = 0;
- mutex_lock(&root->fs_info->fs_mutex);
mutex_lock(&uuid_mutex);
+ mutex_lock(&root->fs_info->volume_mutex);
all_avail = root->fs_info->avail_data_alloc_bits |
root->fs_info->avail_system_alloc_bits |
root->fs_info->avail_metadata_alloc_bits;
if ((all_avail & BTRFS_BLOCK_GROUP_RAID10) &&
- btrfs_super_num_devices(&root->fs_info->super_copy) <= 4) {
- printk("btrfs: unable to go below four devices on raid10\n");
+ root->fs_info->fs_devices->rw_devices <= 4) {
+ printk(KERN_ERR "btrfs: unable to go below four devices "
+ "on raid10\n");
ret = -EINVAL;
goto out;
}
if ((all_avail & BTRFS_BLOCK_GROUP_RAID1) &&
- btrfs_super_num_devices(&root->fs_info->super_copy) <= 2) {
- printk("btrfs: unable to go below two devices on raid1\n");
+ root->fs_info->fs_devices->rw_devices <= 2) {
+ printk(KERN_ERR "btrfs: unable to go below two "
+ "devices on raid1\n");
ret = -EINVAL;
goto out;
}
if (strcmp(device_path, "missing") == 0) {
- struct list_head *cur;
struct list_head *devices;
struct btrfs_device *tmp;
device = NULL;
devices = &root->fs_info->fs_devices->devices;
- list_for_each(cur, devices) {
- tmp = list_entry(cur, struct btrfs_device, dev_list);
+ list_for_each_entry(tmp, devices, dev_list) {
if (tmp->in_fs_metadata && !tmp->bdev) {
device = tmp;
break;
bh = NULL;
disk_super = NULL;
if (!device) {
- printk("btrfs: no missing devices found to remove\n");
+ printk(KERN_ERR "btrfs: no missing devices found to "
+ "remove\n");
goto out;
}
-
} else {
- bdev = open_bdev_excl(device_path, 0,
+ bdev = open_bdev_exclusive(device_path, FMODE_READ,
root->fs_info->bdev_holder);
if (IS_ERR(bdev)) {
ret = PTR_ERR(bdev);
goto out;
}
- bh = __bread(bdev, BTRFS_SUPER_INFO_OFFSET / 4096, 4096);
+ set_blocksize(bdev, 4096);
+ bh = btrfs_read_dev_super(bdev);
if (!bh) {
ret = -EIO;
goto error_close;
}
disk_super = (struct btrfs_super_block *)bh->b_data;
- if (strncmp((char *)(&disk_super->magic), BTRFS_MAGIC,
- sizeof(disk_super->magic))) {
- ret = -ENOENT;
- goto error_brelse;
- }
- if (memcmp(disk_super->fsid, root->fs_info->fsid,
- BTRFS_FSID_SIZE)) {
- ret = -ENOENT;
- goto error_brelse;
- }
devid = le64_to_cpu(disk_super->dev_item.devid);
- device = btrfs_find_device(root, devid, NULL);
+ dev_uuid = disk_super->dev_item.uuid;
+ device = btrfs_find_device(root, devid, dev_uuid,
+ disk_super->fsid);
if (!device) {
ret = -ENOENT;
goto error_brelse;
}
+ }
+ if (device->writeable && root->fs_info->fs_devices->rw_devices == 1) {
+ printk(KERN_ERR "btrfs: unable to remove the only writeable "
+ "device\n");
+ ret = -EINVAL;
+ goto error_brelse;
+ }
+
+ if (device->writeable) {
+ list_del_init(&device->dev_alloc_list);
+ root->fs_info->fs_devices->rw_devices--;
}
- root->fs_info->fs_devices->num_devices--;
ret = btrfs_shrink_device(device, 0);
if (ret)
goto error_brelse;
-
ret = btrfs_rm_dev_item(root->fs_info->chunk_root, device);
if (ret)
goto error_brelse;
- if (bh) {
+ device->in_fs_metadata = 0;
+ list_del_init(&device->dev_list);
+ device->fs_devices->num_devices--;
+
+ next_device = list_entry(root->fs_info->fs_devices->devices.next,
+ struct btrfs_device, dev_list);
+ if (device->bdev == root->fs_info->sb->s_bdev)
+ root->fs_info->sb->s_bdev = next_device->bdev;
+ if (device->bdev == root->fs_info->fs_devices->latest_bdev)
+ root->fs_info->fs_devices->latest_bdev = next_device->bdev;
+
+ if (device->bdev) {
+ close_bdev_exclusive(device->bdev, device->mode);
+ device->bdev = NULL;
+ device->fs_devices->open_devices--;
+ }
+
+ num_devices = btrfs_super_num_devices(&root->fs_info->super_copy) - 1;
+ btrfs_set_super_num_devices(&root->fs_info->super_copy, num_devices);
+
+ if (device->fs_devices->open_devices == 0) {
+ struct btrfs_fs_devices *fs_devices;
+ fs_devices = root->fs_info->fs_devices;
+ while (fs_devices) {
+ if (fs_devices->seed == device->fs_devices)
+ break;
+ fs_devices = fs_devices->seed;
+ }
+ fs_devices->seed = device->fs_devices->seed;
+ device->fs_devices->seed = NULL;
+ __btrfs_close_devices(device->fs_devices);
+ free_fs_devices(device->fs_devices);
+ }
+
+ /*
+ * at this point, the device is zero sized. We want to
+ * remove it from the devices list and zero out the old super
+ */
+ if (device->writeable) {
/* make sure this device isn't detected as part of
* the FS anymore
*/
memset(&disk_super->magic, 0, sizeof(disk_super->magic));
set_buffer_dirty(bh);
sync_dirty_buffer(bh);
-
- brelse(bh);
}
- if (device->bdev) {
- /* one close for the device struct or super_block */
- close_bdev_excl(device->bdev);
- root->fs_info->fs_devices->open_devices--;
- }
- if (bdev) {
- /* one close for us */
- close_bdev_excl(bdev);
- }
kfree(device->name);
kfree(device);
ret = 0;
- goto out;
error_brelse:
brelse(bh);
error_close:
if (bdev)
- close_bdev_excl(bdev);
+ close_bdev_exclusive(bdev, FMODE_READ);
out:
+ mutex_unlock(&root->fs_info->volume_mutex);
mutex_unlock(&uuid_mutex);
- mutex_unlock(&root->fs_info->fs_mutex);
return ret;
}
-int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
+/*
+ * does all the dirty work required for changing file system's UUID.
+ */
+static int btrfs_prepare_sprout(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root)
{
- struct btrfs_trans_handle *trans;
+ struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
+ struct btrfs_fs_devices *old_devices;
+ struct btrfs_fs_devices *seed_devices;
+ struct btrfs_super_block *disk_super = &root->fs_info->super_copy;
struct btrfs_device *device;
- struct block_device *bdev;
- struct list_head *cur;
- struct list_head *devices;
- u64 total_bytes;
- int ret = 0;
+ u64 super_flags;
+ BUG_ON(!mutex_is_locked(&uuid_mutex));
+ if (!fs_devices->seeding)
+ return -EINVAL;
- bdev = open_bdev_excl(device_path, 0, root->fs_info->bdev_holder);
- if (!bdev) {
- return -EIO;
- }
- mutex_lock(&root->fs_info->fs_mutex);
- trans = btrfs_start_transaction(root, 1);
- devices = &root->fs_info->fs_devices->devices;
- list_for_each(cur, devices) {
- device = list_entry(cur, struct btrfs_device, dev_list);
- if (device->bdev == bdev) {
- ret = -EEXIST;
- goto out;
- }
- }
+ seed_devices = kzalloc(sizeof(*fs_devices), GFP_NOFS);
+ if (!seed_devices)
+ return -ENOMEM;
- device = kzalloc(sizeof(*device), GFP_NOFS);
- if (!device) {
- /* we can safely leave the fs_devices entry around */
- ret = -ENOMEM;
- goto out_close_bdev;
+ old_devices = clone_fs_devices(fs_devices);
+ if (IS_ERR(old_devices)) {
+ kfree(seed_devices);
+ return PTR_ERR(old_devices);
}
- device->barriers = 1;
- generate_random_uuid(device->uuid);
- spin_lock_init(&device->io_lock);
- device->name = kstrdup(device_path, GFP_NOFS);
- if (!device->name) {
- kfree(device);
- goto out_close_bdev;
+ list_add(&old_devices->list, &fs_uuids);
+
+ memcpy(seed_devices, fs_devices, sizeof(*seed_devices));
+ seed_devices->opened = 1;
+ INIT_LIST_HEAD(&seed_devices->devices);
+ INIT_LIST_HEAD(&seed_devices->alloc_list);
+ list_splice_init(&fs_devices->devices, &seed_devices->devices);
+ list_splice_init(&fs_devices->alloc_list, &seed_devices->alloc_list);
+ list_for_each_entry(device, &seed_devices->devices, dev_list) {
+ device->fs_devices = seed_devices;
}
- 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->dev_root = root->fs_info->dev_root;
- device->bdev = bdev;
+
+ fs_devices->seeding = 0;
+ fs_devices->num_devices = 0;
+ fs_devices->open_devices = 0;
+ fs_devices->seed = seed_devices;
+
+ generate_random_uuid(fs_devices->fsid);
+ memcpy(root->fs_info->fsid, fs_devices->fsid, BTRFS_FSID_SIZE);
+ memcpy(disk_super->fsid, fs_devices->fsid, BTRFS_FSID_SIZE);
+ super_flags = btrfs_super_flags(disk_super) &
+ ~BTRFS_SUPER_FLAG_SEEDING;
+ btrfs_set_super_flags(disk_super, super_flags);
+
+ return 0;
+}
+
+/*
+ * strore the expected generation for seed devices in device items.
+ */
+static int btrfs_finish_sprout(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root)
+{
+ struct btrfs_path *path;
+ struct extent_buffer *leaf;
+ struct btrfs_dev_item *dev_item;
+ struct btrfs_device *device;
+ struct btrfs_key key;
+ u8 fs_uuid[BTRFS_UUID_SIZE];
+ u8 dev_uuid[BTRFS_UUID_SIZE];
+ u64 devid;
+ int ret;
+
+ path = btrfs_alloc_path();
+ if (!path)
+ return -ENOMEM;
+
+ root = root->fs_info->chunk_root;
+ key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
+ key.offset = 0;
+ key.type = BTRFS_DEV_ITEM_KEY;
+
+ while (1) {
+ ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
+ if (ret < 0)
+ goto error;
+
+ leaf = path->nodes[0];
+next_slot:
+ if (path->slots[0] >= btrfs_header_nritems(leaf)) {
+ ret = btrfs_next_leaf(root, path);
+ if (ret > 0)
+ break;
+ if (ret < 0)
+ goto error;
+ leaf = path->nodes[0];
+ btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+ btrfs_release_path(root, path);
+ continue;
+ }
+
+ btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+ if (key.objectid != BTRFS_DEV_ITEMS_OBJECTID ||
+ key.type != BTRFS_DEV_ITEM_KEY)
+ break;
+
+ dev_item = btrfs_item_ptr(leaf, path->slots[0],
+ struct btrfs_dev_item);
+ devid = btrfs_device_id(leaf, dev_item);
+ read_extent_buffer(leaf, dev_uuid,
+ (unsigned long)btrfs_device_uuid(dev_item),
+ BTRFS_UUID_SIZE);
+ read_extent_buffer(leaf, fs_uuid,
+ (unsigned long)btrfs_device_fsid(dev_item),
+ BTRFS_UUID_SIZE);
+ device = btrfs_find_device(root, devid, dev_uuid, fs_uuid);
+ BUG_ON(!device);
+
+ if (device->fs_devices->seeding) {
+ btrfs_set_device_generation(leaf, dev_item,
+ device->generation);
+ btrfs_mark_buffer_dirty(leaf);
+ }
+
+ path->slots[0]++;
+ goto next_slot;
+ }
+ ret = 0;
+error:
+ btrfs_free_path(path);
+ return ret;
+}
+
+int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
+{
+ struct btrfs_trans_handle *trans;
+ struct btrfs_device *device;
+ struct block_device *bdev;
+ struct list_head *devices;
+ struct super_block *sb = root->fs_info->sb;
+ u64 total_bytes;
+ int seeding_dev = 0;
+ int ret = 0;
+
+ if ((sb->s_flags & MS_RDONLY) && !root->fs_info->fs_devices->seeding)
+ return -EINVAL;
+
+ bdev = open_bdev_exclusive(device_path, 0, root->fs_info->bdev_holder);
+ if (!bdev)
+ return -EIO;
+
+ if (root->fs_info->fs_devices->seeding) {
+ seeding_dev = 1;
+ down_write(&sb->s_umount);
+ mutex_lock(&uuid_mutex);
+ }
+
+ filemap_write_and_wait(bdev->bd_inode->i_mapping);
+ mutex_lock(&root->fs_info->volume_mutex);
+
+ devices = &root->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) {
+ /* we can safely leave the fs_devices entry around */
+ ret = -ENOMEM;
+ goto error;
+ }
+
+ device->name = kstrdup(device_path, GFP_NOFS);
+ if (!device->name) {
+ kfree(device);
+ ret = -ENOMEM;
+ goto error;
+ }
+
+ ret = find_next_devid(root, &device->devid);
+ if (ret) {
+ kfree(device);
+ goto error;
+ }
+
+ trans = btrfs_start_transaction(root, 1);
+ lock_chunks(root);
+
+ device->barriers = 1;
+ device->writeable = 1;
+ device->work.func = pending_bios_fn;
+ generate_random_uuid(device->uuid);
+ spin_lock_init(&device->io_lock);
+ device->generation = trans->transid;
+ 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->dev_root = root->fs_info->dev_root;
+ device->bdev = bdev;
device->in_fs_metadata = 1;
+ device->mode = 0;
+ set_blocksize(device->bdev, 4096);
- ret = btrfs_add_device(trans, root, device);
- if (ret)
- goto out_close_bdev;
+ if (seeding_dev) {
+ sb->s_flags &= ~MS_RDONLY;
+ ret = btrfs_prepare_sprout(trans, root);
+ BUG_ON(ret);
+ }
+
+ device->fs_devices = root->fs_info->fs_devices;
+ list_add(&device->dev_list, &root->fs_info->fs_devices->devices);
+ list_add(&device->dev_alloc_list,
+ &root->fs_info->fs_devices->alloc_list);
+ root->fs_info->fs_devices->num_devices++;
+ root->fs_info->fs_devices->open_devices++;
+ root->fs_info->fs_devices->rw_devices++;
+ root->fs_info->fs_devices->total_rw_bytes += device->total_bytes;
total_bytes = btrfs_super_total_bytes(&root->fs_info->super_copy);
btrfs_set_super_total_bytes(&root->fs_info->super_copy,
btrfs_set_super_num_devices(&root->fs_info->super_copy,
total_bytes + 1);
- list_add(&device->dev_list, &root->fs_info->fs_devices->devices);
- list_add(&device->dev_alloc_list,
- &root->fs_info->fs_devices->alloc_list);
- root->fs_info->fs_devices->num_devices++;
- root->fs_info->fs_devices->open_devices++;
+ if (seeding_dev) {
+ ret = init_first_rw_device(trans, root, device);
+ BUG_ON(ret);
+ ret = btrfs_finish_sprout(trans, root);
+ BUG_ON(ret);
+ } else {
+ ret = btrfs_add_device(trans, root, device);
+ }
+
+ unlock_chunks(root);
+ btrfs_commit_transaction(trans, root);
+
+ if (seeding_dev) {
+ mutex_unlock(&uuid_mutex);
+ up_write(&sb->s_umount);
+
+ ret = btrfs_relocate_sys_chunks(root);
+ BUG_ON(ret);
+ }
out:
- btrfs_end_transaction(trans, root);
- mutex_unlock(&root->fs_info->fs_mutex);
+ mutex_unlock(&root->fs_info->volume_mutex);
return ret;
-
-out_close_bdev:
- close_bdev_excl(bdev);
+error:
+ close_bdev_exclusive(bdev, 0);
+ if (seeding_dev) {
+ mutex_unlock(&uuid_mutex);
+ up_write(&sb->s_umount);
+ }
goto out;
}
-int btrfs_update_device(struct btrfs_trans_handle *trans,
- struct btrfs_device *device)
+static noinline int btrfs_update_device(struct btrfs_trans_handle *trans,
+ struct btrfs_device *device)
{
int ret;
struct btrfs_path *path;
return ret;
}
-int btrfs_grow_device(struct btrfs_trans_handle *trans,
+static int __btrfs_grow_device(struct btrfs_trans_handle *trans,
struct btrfs_device *device, u64 new_size)
{
struct btrfs_super_block *super_copy =
u64 old_total = btrfs_super_total_bytes(super_copy);
u64 diff = new_size - device->total_bytes;
+ if (!device->writeable)
+ return -EACCES;
+ if (new_size <= device->total_bytes)
+ return -EINVAL;
+
btrfs_set_super_total_bytes(super_copy, old_total + diff);
+ device->fs_devices->total_rw_bytes += diff;
+
+ device->total_bytes = new_size;
return btrfs_update_device(trans, device);
}
+int btrfs_grow_device(struct btrfs_trans_handle *trans,
+ struct btrfs_device *device, u64 new_size)
+{
+ int ret;
+ lock_chunks(device->dev_root);
+ ret = __btrfs_grow_device(trans, device, new_size);
+ unlock_chunks(device->dev_root);
+ return ret;
+}
+
static int btrfs_free_chunk(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 chunk_tree, u64 chunk_objectid,
return 0;
}
-int btrfs_del_sys_chunk(struct btrfs_root *root, u64 chunk_objectid, u64
+static int btrfs_del_sys_chunk(struct btrfs_root *root, u64 chunk_objectid, u64
chunk_offset)
{
struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
return ret;
}
-
-int btrfs_relocate_chunk(struct btrfs_root *root,
+static int btrfs_relocate_chunk(struct btrfs_root *root,
u64 chunk_tree, u64 chunk_objectid,
u64 chunk_offset)
{
int ret;
int i;
- printk("btrfs relocating chunk %llu\n",
+ printk(KERN_INFO "btrfs relocating chunk %llu\n",
(unsigned long long)chunk_offset);
root = root->fs_info->chunk_root;
extent_root = root->fs_info->extent_root;
em_tree = &root->fs_info->mapping_tree.map_tree;
/* step one, relocate all the extents inside this chunk */
- ret = btrfs_shrink_extent_tree(extent_root, chunk_offset);
+ ret = btrfs_relocate_block_group(extent_root, chunk_offset);
BUG_ON(ret);
trans = btrfs_start_transaction(root, 1);
BUG_ON(!trans);
+ lock_chunks(root);
+
/*
* step two, delete the device extents and the
* chunk tree entries
BUG_ON(ret);
}
+ ret = btrfs_remove_block_group(trans, extent_root, chunk_offset);
+ BUG_ON(ret);
+
spin_lock(&em_tree->lock);
remove_extent_mapping(em_tree, em);
+ spin_unlock(&em_tree->lock);
+
kfree(map);
em->bdev = NULL;
/* once for the tree */
free_extent_map(em);
- spin_unlock(&em_tree->lock);
-
/* once for us */
free_extent_map(em);
+ unlock_chunks(root);
btrfs_end_transaction(trans, root);
return 0;
}
+static int btrfs_relocate_sys_chunks(struct btrfs_root *root)
+{
+ struct btrfs_root *chunk_root = root->fs_info->chunk_root;
+ struct btrfs_path *path;
+ struct extent_buffer *leaf;
+ struct btrfs_chunk *chunk;
+ struct btrfs_key key;
+ struct btrfs_key found_key;
+ u64 chunk_tree = chunk_root->root_key.objectid;
+ u64 chunk_type;
+ int ret;
+
+ path = btrfs_alloc_path();
+ if (!path)
+ return -ENOMEM;
+
+ key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
+ key.offset = (u64)-1;
+ key.type = BTRFS_CHUNK_ITEM_KEY;
+
+ while (1) {
+ ret = btrfs_search_slot(NULL, chunk_root, &key, path, 0, 0);
+ if (ret < 0)
+ goto error;
+ BUG_ON(ret == 0);
+
+ ret = btrfs_previous_item(chunk_root, path, key.objectid,
+ key.type);
+ if (ret < 0)
+ goto error;
+ if (ret > 0)
+ break;
+
+ leaf = path->nodes[0];
+ btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+
+ chunk = btrfs_item_ptr(leaf, path->slots[0],
+ struct btrfs_chunk);
+ chunk_type = btrfs_chunk_type(leaf, chunk);
+ btrfs_release_path(chunk_root, path);
+
+ if (chunk_type & BTRFS_BLOCK_GROUP_SYSTEM) {
+ ret = btrfs_relocate_chunk(chunk_root, chunk_tree,
+ found_key.objectid,
+ found_key.offset);
+ BUG_ON(ret);
+ }
+
+ if (found_key.offset == 0)
+ break;
+ key.offset = found_key.offset - 1;
+ }
+ ret = 0;
+error:
+ btrfs_free_path(path);
+ return ret;
+}
+
static u64 div_factor(u64 num, int factor)
{
if (factor == 10)
return num;
}
-
int btrfs_balance(struct btrfs_root *dev_root)
{
int ret;
- struct list_head *cur;
struct list_head *devices = &dev_root->fs_info->fs_devices->devices;
struct btrfs_device *device;
u64 old_size;
struct btrfs_trans_handle *trans;
struct btrfs_key found_key;
+ if (dev_root->fs_info->sb->s_flags & MS_RDONLY)
+ return -EROFS;
+ mutex_lock(&dev_root->fs_info->volume_mutex);
dev_root = dev_root->fs_info->dev_root;
- mutex_lock(&dev_root->fs_info->fs_mutex);
/* step one make some room on all the devices */
- list_for_each(cur, devices) {
- device = list_entry(cur, struct btrfs_device, dev_list);
+ list_for_each_entry(device, devices, dev_list) {
old_size = device->total_bytes;
size_to_free = div_factor(old_size, 1);
size_to_free = min(size_to_free, (u64)1 * 1024 * 1024);
- if (device->total_bytes - device->bytes_used > size_to_free)
+ if (!device->writeable ||
+ device->total_bytes - device->bytes_used > size_to_free)
continue;
ret = btrfs_shrink_device(device, old_size - size_to_free);
key.offset = (u64)-1;
key.type = BTRFS_CHUNK_ITEM_KEY;
- while(1) {
+ while (1) {
ret = btrfs_search_slot(NULL, chunk_root, &key, path, 0, 0);
if (ret < 0)
goto error;
ret = btrfs_previous_item(chunk_root, path, 0,
BTRFS_CHUNK_ITEM_KEY);
- if (ret) {
+ if (ret)
break;
- }
+
btrfs_item_key_to_cpu(path->nodes[0], &found_key,
path->slots[0]);
if (found_key.objectid != key.objectid)
break;
+
chunk = btrfs_item_ptr(path->nodes[0],
path->slots[0],
struct btrfs_chunk);
if (key.offset == 0)
break;
+ btrfs_release_path(chunk_root, path);
ret = btrfs_relocate_chunk(chunk_root,
chunk_root->root_key.objectid,
found_key.objectid,
found_key.offset);
BUG_ON(ret);
- btrfs_release_path(chunk_root, path);
}
ret = 0;
error:
btrfs_free_path(path);
- mutex_unlock(&dev_root->fs_info->fs_mutex);
+ mutex_unlock(&dev_root->fs_info->volume_mutex);
return ret;
}
u64 old_total = btrfs_super_total_bytes(super_copy);
u64 diff = device->total_bytes - new_size;
+ if (new_size >= device->total_bytes)
+ return -EINVAL;
path = btrfs_alloc_path();
if (!path)
path->reada = 2;
+ lock_chunks(root);
+
device->total_bytes = new_size;
+ if (device->writeable)
+ device->fs_devices->total_rw_bytes -= diff;
ret = btrfs_update_device(trans, device);
if (ret) {
+ unlock_chunks(root);
btrfs_end_transaction(trans, root);
goto done;
}
WARN_ON(diff > old_total);
btrfs_set_super_total_bytes(super_copy, old_total - diff);
+ unlock_chunks(root);
btrfs_end_transaction(trans, root);
key.objectid = device->devid;
return ret;
}
-int btrfs_add_system_chunk(struct btrfs_trans_handle *trans,
+static int btrfs_add_system_chunk(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_key *key,
struct btrfs_chunk *chunk, int item_size)
return 0;
}
-static u64 chunk_bytes_by_type(u64 type, u64 calc_size, int num_stripes,
- int sub_stripes)
+static noinline u64 chunk_bytes_by_type(u64 type, u64 calc_size,
+ int num_stripes, int sub_stripes)
{
if (type & (BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_DUP))
return calc_size;
return calc_size * num_stripes;
}
-
-int btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
- struct btrfs_root *extent_root, u64 *start,
- u64 *num_bytes, u64 type)
+static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
+ struct btrfs_root *extent_root,
+ struct map_lookup **map_ret,
+ u64 *num_bytes, u64 *stripe_size,
+ u64 start, u64 type)
{
- u64 dev_offset;
struct btrfs_fs_info *info = extent_root->fs_info;
- struct btrfs_root *chunk_root = extent_root->fs_info->chunk_root;
- struct btrfs_path *path;
- struct btrfs_stripe *stripes;
struct btrfs_device *device = NULL;
- struct btrfs_chunk *chunk;
- struct list_head private_devs;
- struct list_head *dev_list;
+ struct btrfs_fs_devices *fs_devices = info->fs_devices;
struct list_head *cur;
+ struct map_lookup *map = NULL;
struct extent_map_tree *em_tree;
- struct map_lookup *map;
struct extent_map *em;
+ struct list_head private_devs;
int min_stripe_size = 1 * 1024 * 1024;
- u64 physical;
u64 calc_size = 1024 * 1024 * 1024;
u64 max_chunk_size = calc_size;
u64 min_free;
u64 avail;
u64 max_avail = 0;
- u64 percent_max;
+ u64 dev_offset;
int num_stripes = 1;
int min_stripes = 1;
int sub_stripes = 0;
int ret;
int index;
int stripe_len = 64 * 1024;
- struct btrfs_key key;
if ((type & BTRFS_BLOCK_GROUP_RAID1) &&
(type & BTRFS_BLOCK_GROUP_DUP)) {
WARN_ON(1);
type &= ~BTRFS_BLOCK_GROUP_DUP;
}
- dev_list = &extent_root->fs_info->fs_devices->alloc_list;
- if (list_empty(dev_list))
+ if (list_empty(&fs_devices->alloc_list))
return -ENOSPC;
if (type & (BTRFS_BLOCK_GROUP_RAID0)) {
- num_stripes = btrfs_super_num_devices(&info->super_copy);
+ num_stripes = fs_devices->rw_devices;
min_stripes = 2;
}
if (type & (BTRFS_BLOCK_GROUP_DUP)) {
min_stripes = 2;
}
if (type & (BTRFS_BLOCK_GROUP_RAID1)) {
- num_stripes = min_t(u64, 2,
- btrfs_super_num_devices(&info->super_copy));
+ num_stripes = min_t(u64, 2, fs_devices->rw_devices);
if (num_stripes < 2)
return -ENOSPC;
min_stripes = 2;
}
if (type & (BTRFS_BLOCK_GROUP_RAID10)) {
- num_stripes = btrfs_super_num_devices(&info->super_copy);
+ num_stripes = fs_devices->rw_devices;
if (num_stripes < 4)
return -ENOSPC;
num_stripes &= ~(u32)1;
min_stripe_size = 1 * 1024 * 1024;
}
- path = btrfs_alloc_path();
- if (!path)
- return -ENOMEM;
-
- /* we don't want a chunk larger than 10% of the FS */
- percent_max = div_factor(btrfs_super_total_bytes(&info->super_copy), 1);
- max_chunk_size = min(percent_max, max_chunk_size);
+ /* we don't want a chunk larger than 10% of writeable space */
+ max_chunk_size = min(div_factor(fs_devices->total_rw_bytes, 1),
+ max_chunk_size);
again:
+ if (!map || map->num_stripes != num_stripes) {
+ kfree(map);
+ map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
+ if (!map)
+ return -ENOMEM;
+ map->num_stripes = num_stripes;
+ }
+
if (calc_size * num_stripes > max_chunk_size) {
calc_size = max_chunk_size;
do_div(calc_size, num_stripes);
do_div(calc_size, stripe_len);
calc_size *= stripe_len;
- INIT_LIST_HEAD(&private_devs);
- cur = dev_list->next;
+ cur = fs_devices->alloc_list.next;
index = 0;
if (type & BTRFS_BLOCK_GROUP_DUP)
else
min_free = calc_size;
- /* we add 1MB because we never use the first 1MB of the device */
- min_free += 1024 * 1024;
+ /*
+ * we add 1MB because we never use the first 1MB of the device, unless
+ * we've looped, then we are likely allocating the maximum amount of
+ * space left already
+ */
+ if (!looped)
+ min_free += 1024 * 1024;
- /* build a private list of devices we will allocate from */
- while(index < num_stripes) {
+ INIT_LIST_HEAD(&private_devs);
+ while (index < num_stripes) {
device = list_entry(cur, struct btrfs_device, dev_alloc_list);
-
+ BUG_ON(!device->writeable);
if (device->total_bytes > device->bytes_used)
avail = device->total_bytes - device->bytes_used;
else
cur = cur->next;
if (device->in_fs_metadata && avail >= min_free) {
- u64 ignored_start = 0;
- ret = find_free_dev_extent(trans, device, path,
- min_free,
- &ignored_start);
+ ret = find_free_dev_extent(trans, device,
+ min_free, &dev_offset);
if (ret == 0) {
list_move_tail(&device->dev_alloc_list,
&private_devs);
+ map->stripes[index].dev = device;
+ map->stripes[index].physical = dev_offset;
index++;
- if (type & BTRFS_BLOCK_GROUP_DUP)
+ if (type & BTRFS_BLOCK_GROUP_DUP) {
+ map->stripes[index].dev = device;
+ map->stripes[index].physical =
+ dev_offset + calc_size;
index++;
+ }
}
} else if (device->in_fs_metadata && avail > max_avail)
max_avail = avail;
- if (cur == dev_list)
+ if (cur == &fs_devices->alloc_list)
break;
}
+ list_splice(&private_devs, &fs_devices->alloc_list);
if (index < num_stripes) {
- list_splice(&private_devs, dev_list);
if (index >= min_stripes) {
num_stripes = index;
if (type & (BTRFS_BLOCK_GROUP_RAID10)) {
calc_size = max_avail;
goto again;
}
- btrfs_free_path(path);
+ kfree(map);
return -ENOSPC;
}
- key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
- key.type = BTRFS_CHUNK_ITEM_KEY;
- ret = find_next_chunk(chunk_root, BTRFS_FIRST_CHUNK_TREE_OBJECTID,
- &key.offset);
- if (ret) {
- btrfs_free_path(path);
- return ret;
- }
+ map->sector_size = extent_root->sectorsize;
+ map->stripe_len = stripe_len;
+ map->io_align = stripe_len;
+ map->io_width = stripe_len;
+ map->type = type;
+ map->num_stripes = num_stripes;
+ map->sub_stripes = sub_stripes;
- chunk = kmalloc(btrfs_chunk_item_size(num_stripes), GFP_NOFS);
- if (!chunk) {
- btrfs_free_path(path);
- return -ENOMEM;
- }
+ *map_ret = map;
+ *stripe_size = calc_size;
+ *num_bytes = chunk_bytes_by_type(type, calc_size,
+ num_stripes, sub_stripes);
- map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
- if (!map) {
- kfree(chunk);
- btrfs_free_path(path);
+ em = alloc_extent_map(GFP_NOFS);
+ if (!em) {
+ kfree(map);
return -ENOMEM;
}
- btrfs_free_path(path);
- path = NULL;
+ em->bdev = (struct block_device *)map;
+ em->start = start;
+ em->len = *num_bytes;
+ em->block_start = 0;
+ em->block_len = em->len;
- stripes = &chunk->stripe;
- *num_bytes = chunk_bytes_by_type(type, calc_size,
- num_stripes, sub_stripes);
+ em_tree = &extent_root->fs_info->mapping_tree.map_tree;
+ spin_lock(&em_tree->lock);
+ ret = add_extent_mapping(em_tree, em);
+ spin_unlock(&em_tree->lock);
+ BUG_ON(ret);
+ free_extent_map(em);
- index = 0;
- while(index < num_stripes) {
- struct btrfs_stripe *stripe;
- BUG_ON(list_empty(&private_devs));
- cur = private_devs.next;
- device = list_entry(cur, struct btrfs_device, dev_alloc_list);
+ ret = btrfs_make_block_group(trans, extent_root, 0, type,
+ BTRFS_FIRST_CHUNK_TREE_OBJECTID,
+ start, *num_bytes);
+ BUG_ON(ret);
- /* loop over this device again if we're doing a dup group */
- if (!(type & BTRFS_BLOCK_GROUP_DUP) ||
- (index == num_stripes - 1))
- list_move_tail(&device->dev_alloc_list, dev_list);
+ index = 0;
+ while (index < map->num_stripes) {
+ device = map->stripes[index].dev;
+ dev_offset = map->stripes[index].physical;
ret = btrfs_alloc_dev_extent(trans, device,
- info->chunk_root->root_key.objectid,
- BTRFS_FIRST_CHUNK_TREE_OBJECTID, key.offset,
- calc_size, &dev_offset);
+ info->chunk_root->root_key.objectid,
+ BTRFS_FIRST_CHUNK_TREE_OBJECTID,
+ start, dev_offset, calc_size);
BUG_ON(ret);
- device->bytes_used += calc_size;
+ index++;
+ }
+
+ return 0;
+}
+
+static int __finish_chunk_alloc(struct btrfs_trans_handle *trans,
+ struct btrfs_root *extent_root,
+ struct map_lookup *map, u64 chunk_offset,
+ u64 chunk_size, u64 stripe_size)
+{
+ u64 dev_offset;
+ struct btrfs_key key;
+ struct btrfs_root *chunk_root = extent_root->fs_info->chunk_root;
+ struct btrfs_device *device;
+ struct btrfs_chunk *chunk;
+ struct btrfs_stripe *stripe;
+ size_t item_size = btrfs_chunk_item_size(map->num_stripes);
+ int index = 0;
+ int ret;
+
+ chunk = kzalloc(item_size, GFP_NOFS);
+ if (!chunk)
+ return -ENOMEM;
+
+ index = 0;
+ while (index < map->num_stripes) {
+ device = map->stripes[index].dev;
+ device->bytes_used += stripe_size;
ret = btrfs_update_device(trans, device);
BUG_ON(ret);
+ index++;
+ }
+
+ index = 0;
+ stripe = &chunk->stripe;
+ while (index < map->num_stripes) {
+ device = map->stripes[index].dev;
+ dev_offset = map->stripes[index].physical;
- map->stripes[index].dev = device;
- map->stripes[index].physical = dev_offset;
- stripe = stripes + index;
btrfs_set_stack_stripe_devid(stripe, device->devid);
btrfs_set_stack_stripe_offset(stripe, dev_offset);
memcpy(stripe->dev_uuid, device->uuid, BTRFS_UUID_SIZE);
- physical = dev_offset;
+ stripe++;
index++;
}
- BUG_ON(!list_empty(&private_devs));
- /* key was set above */
- btrfs_set_stack_chunk_length(chunk, *num_bytes);
+ btrfs_set_stack_chunk_length(chunk, chunk_size);
btrfs_set_stack_chunk_owner(chunk, extent_root->root_key.objectid);
- btrfs_set_stack_chunk_stripe_len(chunk, stripe_len);
- btrfs_set_stack_chunk_type(chunk, type);
- btrfs_set_stack_chunk_num_stripes(chunk, num_stripes);
- btrfs_set_stack_chunk_io_align(chunk, stripe_len);
- btrfs_set_stack_chunk_io_width(chunk, stripe_len);
+ btrfs_set_stack_chunk_stripe_len(chunk, map->stripe_len);
+ btrfs_set_stack_chunk_type(chunk, map->type);
+ btrfs_set_stack_chunk_num_stripes(chunk, map->num_stripes);
+ btrfs_set_stack_chunk_io_align(chunk, map->stripe_len);
+ btrfs_set_stack_chunk_io_width(chunk, map->stripe_len);
btrfs_set_stack_chunk_sector_size(chunk, extent_root->sectorsize);
- btrfs_set_stack_chunk_sub_stripes(chunk, sub_stripes);
- map->sector_size = extent_root->sectorsize;
- map->stripe_len = stripe_len;
- map->io_align = stripe_len;
- map->io_width = stripe_len;
- map->type = type;
- map->num_stripes = num_stripes;
- map->sub_stripes = sub_stripes;
+ btrfs_set_stack_chunk_sub_stripes(chunk, map->sub_stripes);
- ret = btrfs_insert_item(trans, chunk_root, &key, chunk,
- btrfs_chunk_item_size(num_stripes));
- BUG_ON(ret);
- *start = key.offset;;
+ key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
+ key.type = BTRFS_CHUNK_ITEM_KEY;
+ key.offset = chunk_offset;
- em = alloc_extent_map(GFP_NOFS);
- if (!em)
- return -ENOMEM;
- em->bdev = (struct block_device *)map;
- em->start = key.offset;
- em->len = *num_bytes;
- em->block_start = 0;
+ ret = btrfs_insert_item(trans, chunk_root, &key, chunk, item_size);
+ BUG_ON(ret);
- if (type & BTRFS_BLOCK_GROUP_SYSTEM) {
- ret = btrfs_add_system_chunk(trans, chunk_root, &key,
- chunk, btrfs_chunk_item_size(num_stripes));
+ if (map->type & BTRFS_BLOCK_GROUP_SYSTEM) {
+ ret = btrfs_add_system_chunk(trans, chunk_root, &key, chunk,
+ item_size);
BUG_ON(ret);
}
kfree(chunk);
+ return 0;
+}
- em_tree = &extent_root->fs_info->mapping_tree.map_tree;
- spin_lock(&em_tree->lock);
- ret = add_extent_mapping(em_tree, em);
- spin_unlock(&em_tree->lock);
+/*
+ * Chunk allocation falls into two parts. The first part does works
+ * that make the new allocated chunk useable, but not do any operation
+ * that modifies the chunk tree. The second part does the works that
+ * require modifying the chunk tree. This division is important for the
+ * bootstrap process of adding storage to a seed btrfs.
+ */
+int btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
+ struct btrfs_root *extent_root, u64 type)
+{
+ u64 chunk_offset;
+ u64 chunk_size;
+ u64 stripe_size;
+ struct map_lookup *map;
+ struct btrfs_root *chunk_root = extent_root->fs_info->chunk_root;
+ int ret;
+
+ ret = find_next_chunk(chunk_root, BTRFS_FIRST_CHUNK_TREE_OBJECTID,
+ &chunk_offset);
+ if (ret)
+ return ret;
+
+ ret = __btrfs_alloc_chunk(trans, extent_root, &map, &chunk_size,
+ &stripe_size, chunk_offset, type);
+ if (ret)
+ return ret;
+
+ ret = __finish_chunk_alloc(trans, extent_root, map, chunk_offset,
+ chunk_size, stripe_size);
+ BUG_ON(ret);
+ return 0;
+}
+
+static noinline int init_first_rw_device(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_device *device)
+{
+ u64 chunk_offset;
+ u64 sys_chunk_offset;
+ u64 chunk_size;
+ u64 sys_chunk_size;
+ u64 stripe_size;
+ u64 sys_stripe_size;
+ u64 alloc_profile;
+ struct map_lookup *map;
+ struct map_lookup *sys_map;
+ struct btrfs_fs_info *fs_info = root->fs_info;
+ struct btrfs_root *extent_root = fs_info->extent_root;
+ int ret;
+
+ ret = find_next_chunk(fs_info->chunk_root,
+ BTRFS_FIRST_CHUNK_TREE_OBJECTID, &chunk_offset);
BUG_ON(ret);
+
+ alloc_profile = BTRFS_BLOCK_GROUP_METADATA |
+ (fs_info->metadata_alloc_profile &
+ fs_info->avail_metadata_alloc_bits);
+ alloc_profile = btrfs_reduce_alloc_profile(root, alloc_profile);
+
+ ret = __btrfs_alloc_chunk(trans, extent_root, &map, &chunk_size,
+ &stripe_size, chunk_offset, alloc_profile);
+ BUG_ON(ret);
+
+ sys_chunk_offset = chunk_offset + chunk_size;
+
+ alloc_profile = BTRFS_BLOCK_GROUP_SYSTEM |
+ (fs_info->system_alloc_profile &
+ fs_info->avail_system_alloc_bits);
+ alloc_profile = btrfs_reduce_alloc_profile(root, alloc_profile);
+
+ ret = __btrfs_alloc_chunk(trans, extent_root, &sys_map,
+ &sys_chunk_size, &sys_stripe_size,
+ sys_chunk_offset, alloc_profile);
+ BUG_ON(ret);
+
+ ret = btrfs_add_device(trans, fs_info->chunk_root, device);
+ BUG_ON(ret);
+
+ /*
+ * Modifying chunk tree needs allocating new blocks from both
+ * system block group and metadata block group. So we only can
+ * do operations require modifying the chunk tree after both
+ * block groups were created.
+ */
+ ret = __finish_chunk_alloc(trans, extent_root, map, chunk_offset,
+ chunk_size, stripe_size);
+ BUG_ON(ret);
+
+ ret = __finish_chunk_alloc(trans, extent_root, sys_map,
+ sys_chunk_offset, sys_chunk_size,
+ sys_stripe_size);
+ BUG_ON(ret);
+ return 0;
+}
+
+int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset)
+{
+ struct extent_map *em;
+ struct map_lookup *map;
+ struct btrfs_mapping_tree *map_tree = &root->fs_info->mapping_tree;
+ int readonly = 0;
+ int i;
+
+ spin_lock(&map_tree->map_tree.lock);
+ em = lookup_extent_mapping(&map_tree->map_tree, chunk_offset, 1);
+ spin_unlock(&map_tree->map_tree.lock);
+ if (!em)
+ return 1;
+
+ map = (struct map_lookup *)em->bdev;
+ for (i = 0; i < map->num_stripes; i++) {
+ if (!map->stripes[i].dev->writeable) {
+ readonly = 1;
+ break;
+ }
+ }
free_extent_map(em);
- return ret;
+ return readonly;
}
void btrfs_mapping_init(struct btrfs_mapping_tree *tree)
{
struct extent_map *em;
- while(1) {
+ while (1) {
spin_lock(&tree->map_tree.lock);
em = lookup_extent_mapping(&tree->map_tree, 0, (u64)-1);
if (em)
int max_errors = 0;
struct btrfs_multi_bio *multi = NULL;
- if (multi_ret && !(rw & (1 << BIO_RW))) {
+ if (multi_ret && !(rw & (1 << BIO_RW)))
stripes_allocated = 1;
- }
again:
if (multi_ret) {
multi = kzalloc(btrfs_multi_bio_size(stripes_allocated),
return 0;
if (!em) {
- printk("unable to find logical %Lu len %Lu\n", logical, *length);
+ printk(KERN_CRIT "unable to find logical %llu len %llu\n",
+ (unsigned long long)logical,
+ (unsigned long long)*length);
BUG();
}
device = map->stripes[stripe_index].dev;
if (device->bdev) {
bdi = blk_get_backing_dev_info(device->bdev);
- if (bdi->unplug_io_fn) {
+ if (bdi->unplug_io_fn)
bdi->unplug_io_fn(bdi, unplug_page);
- }
}
} else {
multi->stripes[i].physical =
mirror_num, NULL);
}
+int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree,
+ u64 chunk_start, u64 physical, u64 devid,
+ u64 **logical, int *naddrs, int *stripe_len)
+{
+ struct extent_map_tree *em_tree = &map_tree->map_tree;
+ struct extent_map *em;
+ struct map_lookup *map;
+ u64 *buf;
+ u64 bytenr;
+ u64 length;
+ u64 stripe_nr;
+ int i, j, nr = 0;
+
+ spin_lock(&em_tree->lock);
+ em = lookup_extent_mapping(em_tree, chunk_start, 1);
+ spin_unlock(&em_tree->lock);
+
+ BUG_ON(!em || em->start != chunk_start);
+ map = (struct map_lookup *)em->bdev;
+
+ length = em->len;
+ if (map->type & BTRFS_BLOCK_GROUP_RAID10)
+ do_div(length, map->num_stripes / map->sub_stripes);
+ else if (map->type & BTRFS_BLOCK_GROUP_RAID0)
+ do_div(length, map->num_stripes);
+
+ buf = kzalloc(sizeof(u64) * map->num_stripes, GFP_NOFS);
+ BUG_ON(!buf);
+
+ for (i = 0; i < map->num_stripes; i++) {
+ if (devid && map->stripes[i].dev->devid != devid)
+ continue;
+ if (map->stripes[i].physical > physical ||
+ map->stripes[i].physical + length <= physical)
+ continue;
+
+ stripe_nr = physical - map->stripes[i].physical;
+ do_div(stripe_nr, map->stripe_len);
+
+ if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
+ stripe_nr = stripe_nr * map->num_stripes + i;
+ do_div(stripe_nr, map->sub_stripes);
+ } else if (map->type & BTRFS_BLOCK_GROUP_RAID0) {
+ stripe_nr = stripe_nr * map->num_stripes + i;
+ }
+ bytenr = chunk_start + stripe_nr * map->stripe_len;
+ WARN_ON(nr >= map->num_stripes);
+ for (j = 0; j < nr; j++) {
+ if (buf[j] == bytenr)
+ break;
+ }
+ if (j == nr) {
+ WARN_ON(nr >= map->num_stripes);
+ buf[nr++] = bytenr;
+ }
+ }
+
+ for (i = 0; i > nr; i++) {
+ struct btrfs_multi_bio *multi;
+ struct btrfs_bio_stripe *stripe;
+ int ret;
+
+ length = 1;
+ ret = btrfs_map_block(map_tree, WRITE, buf[i],
+ &length, &multi, 0);
+ BUG_ON(ret);
+
+ stripe = multi->stripes;
+ for (j = 0; j < multi->num_stripes; j++) {
+ if (stripe->physical >= physical &&
+ physical < stripe->physical + length)
+ break;
+ }
+ BUG_ON(j >= multi->num_stripes);
+ kfree(multi);
+ }
+
+ *logical = buf;
+ *naddrs = nr;
+ *stripe_len = map->stripe_len;
+
+ free_extent_map(em);
+ return 0;
+}
+
int btrfs_unplug_page(struct btrfs_mapping_tree *map_tree,
u64 logical, struct page *page)
{
NULL, 0, page);
}
-
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,23)
static void end_bio_multi_stripe(struct bio *bio, int err)
-#else
-static int end_bio_multi_stripe(struct bio *bio,
- unsigned int bytes_done, int err)
-#endif
{
struct btrfs_multi_bio *multi = bio->bi_private;
+ int is_orig_bio = 0;
-#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,23)
- if (bio->bi_size)
- return 1;
-#endif
if (err)
atomic_inc(&multi->error);
+ if (bio == multi->orig_bio)
+ is_orig_bio = 1;
+
if (atomic_dec_and_test(&multi->stripes_pending)) {
+ if (!is_orig_bio) {
+ bio_put(bio);
+ bio = multi->orig_bio;
+ }
bio->bi_private = multi->private;
bio->bi_end_io = multi->end_io;
/* only send an error to the higher layers if it is
}
kfree(multi);
-#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,23)
- bio_endio(bio, bio->bi_size, err);
-#else
bio_endio(bio, err);
-#endif
- } else {
+ } else if (!is_orig_bio) {
bio_put(bio);
}
-#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,23)
+}
+
+struct async_sched {
+ struct bio *bio;
+ int rw;
+ struct btrfs_fs_info *info;
+ struct btrfs_work work;
+};
+
+/*
+ * see run_scheduled_bios for a description of why bios are collected for
+ * async submit.
+ *
+ * This will add one bio to the pending list for a device and make sure
+ * the work struct is scheduled.
+ */
+static noinline int schedule_bio(struct btrfs_root *root,
+ struct btrfs_device *device,
+ int rw, struct bio *bio)
+{
+ int should_queue = 1;
+
+ /* don't bother with additional async steps for reads, right now */
+ if (!(rw & (1 << BIO_RW))) {
+ bio_get(bio);
+ submit_bio(rw, bio);
+ bio_put(bio);
+ return 0;
+ }
+
+ /*
+ * nr_async_bios allows us to reliably return congestion to the
+ * higher layers. Otherwise, the async bio makes it appear we have
+ * made progress against dirty pages when we've really just put it
+ * on a queue for later
+ */
+ atomic_inc(&root->fs_info->nr_async_bios);
+ WARN_ON(bio->bi_next);
+ bio->bi_next = NULL;
+ bio->bi_rw |= rw;
+
+ spin_lock(&device->io_lock);
+
+ if (device->pending_bio_tail)
+ device->pending_bio_tail->bi_next = bio;
+
+ device->pending_bio_tail = bio;
+ if (!device->pending_bios)
+ device->pending_bios = bio;
+ if (device->running_pending)
+ should_queue = 0;
+
+ spin_unlock(&device->io_lock);
+
+ if (should_queue)
+ btrfs_queue_worker(&root->fs_info->submit_workers,
+ &device->work);
return 0;
-#endif
}
int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio,
- int mirror_num)
+ int mirror_num, int async_submit)
{
struct btrfs_mapping_tree *map_tree;
struct btrfs_device *dev;
struct bio *first_bio = bio;
- u64 logical = bio->bi_sector << 9;
+ u64 logical = (u64)bio->bi_sector << 9;
u64 length = 0;
u64 map_length;
struct btrfs_multi_bio *multi = NULL;
total_devs = multi->num_stripes;
if (map_length < length) {
- printk("mapping failed logical %Lu bio len %Lu "
- "len %Lu\n", logical, length, map_length);
+ printk(KERN_CRIT "mapping failed logical %llu bio len %llu "
+ "len %llu\n", (unsigned long long)logical,
+ (unsigned long long)length,
+ (unsigned long long)map_length);
BUG();
}
multi->end_io = first_bio->bi_end_io;
multi->private = first_bio->bi_private;
+ multi->orig_bio = first_bio;
atomic_set(&multi->stripes_pending, multi->num_stripes);
- while(dev_nr < total_devs) {
+ while (dev_nr < total_devs) {
if (total_devs > 1) {
if (dev_nr < total_devs - 1) {
bio = bio_clone(first_bio, GFP_NOFS);
}
bio->bi_sector = multi->stripes[dev_nr].physical >> 9;
dev = multi->stripes[dev_nr].dev;
+ BUG_ON(rw == WRITE && !dev->writeable);
if (dev && dev->bdev) {
bio->bi_bdev = dev->bdev;
- spin_lock(&dev->io_lock);
- dev->total_ios++;
- spin_unlock(&dev->io_lock);
- submit_bio(rw, bio);
+ if (async_submit)
+ schedule_bio(root, dev, rw, bio);
+ else
+ submit_bio(rw, bio);
} else {
bio->bi_bdev = root->fs_info->fs_devices->latest_bdev;
bio->bi_sector = logical >> 9;
-#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,23)
- bio_endio(bio, bio->bi_size, -EIO);
-#else
bio_endio(bio, -EIO);
-#endif
}
dev_nr++;
}
}
struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid,
- u8 *uuid)
+ u8 *uuid, u8 *fsid)
{
- struct list_head *head = &root->fs_info->fs_devices->devices;
-
- return __find_device(head, devid, uuid);
+ struct btrfs_device *device;
+ struct btrfs_fs_devices *cur_devices;
+
+ cur_devices = root->fs_info->fs_devices;
+ while (cur_devices) {
+ if (!fsid ||
+ !memcmp(cur_devices->fsid, fsid, BTRFS_UUID_SIZE)) {
+ device = __find_device(&cur_devices->devices,
+ devid, uuid);
+ if (device)
+ return device;
+ }
+ cur_devices = cur_devices->seed;
+ }
+ return NULL;
}
static struct btrfs_device *add_missing_dev(struct btrfs_root *root,
struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
device = kzalloc(sizeof(*device), GFP_NOFS);
+ if (!device)
+ return NULL;
list_add(&device->dev_list,
&fs_devices->devices);
- list_add(&device->dev_alloc_list,
- &fs_devices->alloc_list);
device->barriers = 1;
device->dev_root = root->fs_info->dev_root;
device->devid = devid;
+ device->work.func = pending_bios_fn;
+ device->fs_devices = fs_devices;
fs_devices->num_devices++;
spin_lock_init(&device->io_lock);
+ INIT_LIST_HEAD(&device->dev_alloc_list);
memcpy(device->uuid, dev_uuid, BTRFS_UUID_SIZE);
return device;
}
-
static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key,
struct extent_buffer *leaf,
struct btrfs_chunk *chunk)
em->start = logical;
em->len = length;
em->block_start = 0;
+ em->block_len = em->len;
map->num_stripes = num_stripes;
map->io_width = btrfs_chunk_io_width(leaf, chunk);
read_extent_buffer(leaf, uuid, (unsigned long)
btrfs_stripe_dev_uuid_nr(chunk, i),
BTRFS_UUID_SIZE);
- map->stripes[i].dev = btrfs_find_device(root, devid, uuid);
-
+ map->stripes[i].dev = btrfs_find_device(root, devid, uuid,
+ NULL);
if (!map->stripes[i].dev && !btrfs_test_opt(root, DEGRADED)) {
kfree(map);
free_extent_map(em);
return 0;
}
+static int open_seed_devices(struct btrfs_root *root, u8 *fsid)
+{
+ struct btrfs_fs_devices *fs_devices;
+ int ret;
+
+ mutex_lock(&uuid_mutex);
+
+ fs_devices = root->fs_info->fs_devices->seed;
+ while (fs_devices) {
+ if (!memcmp(fs_devices->fsid, fsid, BTRFS_UUID_SIZE)) {
+ ret = 0;
+ goto out;
+ }
+ fs_devices = fs_devices->seed;
+ }
+
+ fs_devices = find_fsid(fsid);
+ if (!fs_devices) {
+ ret = -ENOENT;
+ goto out;
+ }
+
+ fs_devices = clone_fs_devices(fs_devices);
+ if (IS_ERR(fs_devices)) {
+ ret = PTR_ERR(fs_devices);
+ goto out;
+ }
+
+ ret = __btrfs_open_devices(fs_devices, FMODE_READ,
+ root->fs_info->bdev_holder);
+ if (ret)
+ goto out;
+
+ if (!fs_devices->seeding) {
+ __btrfs_close_devices(fs_devices);
+ free_fs_devices(fs_devices);
+ ret = -EINVAL;
+ goto out;
+ }
+
+ fs_devices->seed = root->fs_info->fs_devices->seed;
+ root->fs_info->fs_devices->seed = fs_devices;
+out:
+ mutex_unlock(&uuid_mutex);
+ return ret;
+}
+
static int read_one_dev(struct btrfs_root *root,
struct extent_buffer *leaf,
struct btrfs_dev_item *dev_item)
struct btrfs_device *device;
u64 devid;
int ret;
+ u8 fs_uuid[BTRFS_UUID_SIZE];
u8 dev_uuid[BTRFS_UUID_SIZE];
devid = btrfs_device_id(leaf, dev_item);
read_extent_buffer(leaf, dev_uuid,
(unsigned long)btrfs_device_uuid(dev_item),
BTRFS_UUID_SIZE);
- device = btrfs_find_device(root, devid, dev_uuid);
- if (!device) {
- printk("warning devid %Lu missing\n", devid);
- device = add_missing_dev(root, devid, dev_uuid);
- if (!device)
- return -ENOMEM;
+ read_extent_buffer(leaf, fs_uuid,
+ (unsigned long)btrfs_device_fsid(dev_item),
+ BTRFS_UUID_SIZE);
+
+ if (memcmp(fs_uuid, root->fs_info->fsid, BTRFS_UUID_SIZE)) {
+ ret = open_seed_devices(root, fs_uuid);
+ if (ret && !btrfs_test_opt(root, DEGRADED))
+ return ret;
+ }
+
+ device = btrfs_find_device(root, devid, dev_uuid, fs_uuid);
+ if (!device || !device->bdev) {
+ if (!btrfs_test_opt(root, DEGRADED))
+ return -EIO;
+
+ if (!device) {
+ printk(KERN_WARNING "warning devid %llu missing\n",
+ (unsigned long long)devid);
+ device = add_missing_dev(root, devid, dev_uuid);
+ if (!device)
+ return -ENOMEM;
+ }
+ }
+
+ if (device->fs_devices != root->fs_info->fs_devices) {
+ BUG_ON(device->writeable);
+ if (device->generation !=
+ btrfs_device_generation(leaf, dev_item))
+ return -EINVAL;
}
fill_device_from_item(leaf, dev_item, device);
device->dev_root = root->fs_info->dev_root;
device->in_fs_metadata = 1;
+ if (device->writeable)
+ device->fs_devices->total_rw_bytes += device->total_bytes;
ret = 0;
-#if 0
- ret = btrfs_open_device(device);
- if (ret) {
- kfree(device);
- }
-#endif
return ret;
}
key.type = 0;
again:
ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
- while(1) {
+ while (1) {
leaf = path->nodes[0];
slot = path->slots[0];
if (slot >= btrfs_header_nritems(leaf)) {
dev_item = btrfs_item_ptr(leaf, slot,
struct btrfs_dev_item);
ret = read_one_dev(root, leaf, dev_item);
- BUG_ON(ret);
+ if (ret)
+ goto error;
}
} else if (found_key.type == BTRFS_CHUNK_ITEM_KEY) {
struct btrfs_chunk *chunk;
chunk = btrfs_item_ptr(leaf, slot, struct btrfs_chunk);
ret = read_one_chunk(root, &found_key, leaf, chunk);
+ if (ret)
+ goto error;
}
path->slots[0]++;
}
btrfs_release_path(root, path);
goto again;
}
-
- btrfs_free_path(path);
ret = 0;
error:
+ btrfs_free_path(path);
return ret;
}
-
git://ftp.safe.ca / safe / jmp / linux-2.6 / blobdiff