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)))
-
static DEFINE_MUTEX(uuid_mutex);
static LIST_HEAD(fs_uuids);
-void btrfs_lock_volumes(void)
-{
- mutex_lock(&uuid_mutex);
-}
-
-void btrfs_unlock_volumes(void)
-{
- mutex_unlock(&uuid_mutex);
-}
-
static void lock_chunks(struct btrfs_root *root)
{
mutex_lock(&root->fs_info->chunk_mutex);
INIT_LIST_HEAD(&device->dev_alloc_list);
mutex_lock(&fs_devices->device_list_mutex);
- list_add(&device->dev_list, &fs_devices->devices);
+ list_add_rcu(&device->dev_list, &fs_devices->devices);
mutex_unlock(&fs_devices->device_list_mutex);
device->fs_devices = fs_devices;
fs_devices->latest_trans = orig->latest_trans;
memcpy(fs_devices->fsid, orig->fsid, sizeof(fs_devices->fsid));
- mutex_lock(&orig->device_list_mutex);
+ /* We have held the volume lock, it is safe to get the devices. */
list_for_each_entry(orig_dev, &orig->devices, dev_list) {
device = kzalloc(sizeof(*device), GFP_NOFS);
if (!device)
device->fs_devices = fs_devices;
fs_devices->num_devices++;
}
- mutex_unlock(&orig->device_list_mutex);
return fs_devices;
error:
- mutex_unlock(&orig->device_list_mutex);
free_fs_devices(fs_devices);
return ERR_PTR(-ENOMEM);
}
mutex_lock(&uuid_mutex);
again:
- mutex_lock(&fs_devices->device_list_mutex);
+ /* This is the initialized path, it is safe to release the devices. */
list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) {
if (device->in_fs_metadata)
continue;
kfree(device->name);
kfree(device);
}
- mutex_unlock(&fs_devices->device_list_mutex);
if (fs_devices->seed) {
fs_devices = fs_devices->seed;
return 0;
}
+static void __free_device(struct work_struct *work)
+{
+ struct btrfs_device *device;
+
+ device = container_of(work, struct btrfs_device, rcu_work);
+
+ if (device->bdev)
+ blkdev_put(device->bdev, device->mode);
+
+ kfree(device->name);
+ kfree(device);
+}
+
+static void free_device(struct rcu_head *head)
+{
+ struct btrfs_device *device;
+
+ device = container_of(head, struct btrfs_device, rcu);
+
+ INIT_WORK(&device->rcu_work, __free_device);
+ schedule_work(&device->rcu_work);
+}
+
static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
{
struct btrfs_device *device;
if (--fs_devices->opened > 0)
return 0;
+ mutex_lock(&fs_devices->device_list_mutex);
list_for_each_entry(device, &fs_devices->devices, dev_list) {
- if (device->bdev) {
- blkdev_put(device->bdev, device->mode);
+ struct btrfs_device *new_device;
+
+ if (device->bdev)
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;
+ new_device = kmalloc(sizeof(*new_device), GFP_NOFS);
+ BUG_ON(!new_device);
+ memcpy(new_device, device, sizeof(*new_device));
+ new_device->name = kstrdup(device->name, GFP_NOFS);
+ BUG_ON(device->name && !new_device->name);
+ new_device->bdev = NULL;
+ new_device->writeable = 0;
+ new_device->in_fs_metadata = 0;
+ list_replace_rcu(&device->dev_list, &new_device->dev_list);
+
+ call_rcu(&device->rcu, free_device);
}
+ mutex_unlock(&fs_devices->device_list_mutex);
+
WARN_ON(fs_devices->open_devices);
WARN_ON(fs_devices->rw_devices);
fs_devices->opened = 0;
list_add(&device->dev_alloc_list,
&fs_devices->alloc_list);
}
+ brelse(bh);
continue;
error_brelse:
transid = btrfs_super_generation(disk_super);
if (disk_super->label[0])
printk(KERN_INFO "device label %s ", disk_super->label);
- else {
- /* FIXME, make a readl uuid parser */
- printk(KERN_INFO "device fsid %llx-%llx ",
- *(unsigned long long *)disk_super->fsid,
- *(unsigned long long *)(disk_super->fsid + 8));
- }
+ else
+ printk(KERN_INFO "device fsid %pU ", disk_super->fsid);
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);
/* we don't want to overwrite the superblock on the drive,
* so we make sure to start at an offset of at least 1MB
*/
- search_start = 1024 * 1024;
-
- if (root->fs_info->alloc_start + num_bytes <= search_end)
- search_start = max(root->fs_info->alloc_start, search_start);
+ search_start = max(root->fs_info->alloc_start, 1024ull * 1024);
max_hole_start = search_start;
max_hole_size = 0;
if (ret > 0) {
ret = btrfs_previous_item(root, path, key.objectid,
BTRFS_DEV_EXTENT_KEY);
- BUG_ON(ret);
+ if (ret)
+ goto out;
leaf = path->nodes[0];
btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
extent = btrfs_item_ptr(leaf, path->slots[0],
struct btrfs_dev_extent);
BUG_ON(found_key.offset > start || found_key.offset +
btrfs_dev_extent_length(leaf, extent) < start);
- ret = 0;
} else if (ret == 0) {
leaf = path->nodes[0];
extent = btrfs_item_ptr(leaf, path->slots[0],
if (device->bytes_used > 0)
device->bytes_used -= btrfs_dev_extent_length(leaf, extent);
ret = btrfs_del_item(trans, root, path);
- BUG_ON(ret);
+out:
btrfs_free_path(path);
return ret;
}
struct btrfs_key found_key;
path = btrfs_alloc_path();
- BUG_ON(!path);
+ if (!path)
+ return -ENOMEM;
key.objectid = objectid;
key.offset = (u64)-1;
struct block_device *bdev;
struct buffer_head *bh = NULL;
struct btrfs_super_block *disk_super;
+ struct btrfs_fs_devices *cur_devices;
u64 all_avail;
u64 devid;
u64 num_devices;
u8 *dev_uuid;
int ret = 0;
+ bool clear_super = false;
mutex_lock(&uuid_mutex);
mutex_lock(&root->fs_info->volume_mutex);
device = NULL;
devices = &root->fs_info->fs_devices->devices;
- mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
+ /*
+ * It is safe to read the devices since the volume_mutex
+ * is held.
+ */
list_for_each_entry(tmp, devices, dev_list) {
if (tmp->in_fs_metadata && !tmp->bdev) {
device = tmp;
break;
}
}
- mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
bdev = NULL;
bh = NULL;
disk_super = NULL;
}
if (device->writeable) {
+ lock_chunks(root);
list_del_init(&device->dev_alloc_list);
+ unlock_chunks(root);
root->fs_info->fs_devices->rw_devices--;
+ clear_super = true;
}
ret = btrfs_shrink_device(device, 0);
goto error_undo;
device->in_fs_metadata = 0;
+ btrfs_scrub_cancel_dev(root, device);
/*
* the device list mutex makes sure that we don't change
* the device list while someone else is writing out all
* the device supers.
*/
+
+ cur_devices = device->fs_devices;
mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
- list_del_init(&device->dev_list);
- mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+ list_del_rcu(&device->dev_list);
device->fs_devices->num_devices--;
if (device->bdev == root->fs_info->fs_devices->latest_bdev)
root->fs_info->fs_devices->latest_bdev = next_device->bdev;
- if (device->bdev) {
- blkdev_put(device->bdev, device->mode);
- device->bdev = NULL;
+ if (device->bdev)
device->fs_devices->open_devices--;
- }
+
+ call_rcu(&device->rcu, free_device);
+ mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
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) {
+ if (cur_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)
+ if (fs_devices->seed == cur_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);
+ fs_devices->seed = cur_devices->seed;
+ cur_devices->seed = NULL;
+ lock_chunks(root);
+ __btrfs_close_devices(cur_devices);
+ unlock_chunks(root);
+ free_fs_devices(cur_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) {
+ if (clear_super) {
/* make sure this device isn't detected as part of
* the FS anymore
*/
sync_dirty_buffer(bh);
}
- kfree(device->name);
- kfree(device);
ret = 0;
error_brelse:
return ret;
error_undo:
if (device->writeable) {
+ lock_chunks(root);
list_add(&device->dev_alloc_list,
&root->fs_info->fs_devices->alloc_list);
+ unlock_chunks(root);
root->fs_info->fs_devices->rw_devices++;
}
goto error_brelse;
INIT_LIST_HEAD(&seed_devices->devices);
INIT_LIST_HEAD(&seed_devices->alloc_list);
mutex_init(&seed_devices->device_list_mutex);
- list_splice_init(&fs_devices->devices, &seed_devices->devices);
+
+ mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
+ list_splice_init_rcu(&fs_devices->devices, &seed_devices->devices,
+ synchronize_rcu);
+ mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+
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;
* half setup
*/
mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
- list_add(&device->dev_list, &root->fs_info->fs_devices->devices);
+ list_add_rcu(&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++;
BUG_ON(ret);
ret = btrfs_del_item(trans, root, path);
- BUG_ON(ret);
btrfs_free_path(path);
- return 0;
+ return ret;
}
static int btrfs_del_sys_chunk(struct btrfs_root *root, u64 chunk_objectid, u64
/* step two, relocate all the chunks */
path = btrfs_alloc_path();
- BUG_ON(!path);
-
+ if (!path) {
+ ret = -ENOMEM;
+ goto error;
+ }
key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
key.offset = (u64)-1;
key.type = BTRFS_CHUNK_ITEM_KEY;
chunk_root->root_key.objectid,
found_key.objectid,
found_key.offset);
- BUG_ON(ret && ret != -ENOSPC);
+ if (ret && ret != -ENOSPC)
+ goto error;
key.offset = found_key.offset - 1;
}
ret = 0;
return 0;
}
-static noinline u64 chunk_bytes_by_type(u64 type, u64 calc_size,
- int num_stripes, int sub_stripes)
+/*
+ * sort the devices in descending order by max_avail, total_avail
+ */
+static int btrfs_cmp_device_info(const void *a, const void *b)
{
- if (type & (BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_DUP))
- return calc_size;
- else if (type & BTRFS_BLOCK_GROUP_RAID10)
- return calc_size * (num_stripes / sub_stripes);
- else
- return calc_size * num_stripes;
-}
+ const struct btrfs_device_info *di_a = a;
+ const struct btrfs_device_info *di_b = b;
-/* Used to sort the devices by max_avail(descending sort) */
-int btrfs_cmp_device_free_bytes(const void *dev_info1, const void *dev_info2)
-{
- if (((struct btrfs_device_info *)dev_info1)->max_avail >
- ((struct btrfs_device_info *)dev_info2)->max_avail)
+ if (di_a->max_avail > di_b->max_avail)
return -1;
- else if (((struct btrfs_device_info *)dev_info1)->max_avail <
- ((struct btrfs_device_info *)dev_info2)->max_avail)
+ if (di_a->max_avail < di_b->max_avail)
return 1;
- else
- return 0;
+ if (di_a->total_avail > di_b->total_avail)
+ return -1;
+ if (di_a->total_avail < di_b->total_avail)
+ return 1;
+ return 0;
}
-static int __btrfs_calc_nstripes(struct btrfs_fs_devices *fs_devices, u64 type,
- int *num_stripes, int *min_stripes,
- int *sub_stripes)
+static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
+ struct btrfs_root *extent_root,
+ struct map_lookup **map_ret,
+ u64 *num_bytes_out, u64 *stripe_size_out,
+ u64 start, u64 type)
{
- *num_stripes = 1;
- *min_stripes = 1;
- *sub_stripes = 0;
+ struct btrfs_fs_info *info = extent_root->fs_info;
+ 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 extent_map *em;
+ struct btrfs_device_info *devices_info = NULL;
+ u64 total_avail;
+ int num_stripes; /* total number of stripes to allocate */
+ int sub_stripes; /* sub_stripes info for map */
+ int dev_stripes; /* stripes per dev */
+ int devs_max; /* max devs to use */
+ int devs_min; /* min devs needed */
+ int devs_increment; /* ndevs has to be a multiple of this */
+ int ncopies; /* how many copies to data has */
+ int ret;
+ u64 max_stripe_size;
+ u64 max_chunk_size;
+ u64 stripe_size;
+ u64 num_bytes;
+ int ndevs;
+ int i;
+ int j;
- if (type & (BTRFS_BLOCK_GROUP_RAID0)) {
- *num_stripes = fs_devices->rw_devices;
- *min_stripes = 2;
- }
- if (type & (BTRFS_BLOCK_GROUP_DUP)) {
- *num_stripes = 2;
- *min_stripes = 2;
- }
- if (type & (BTRFS_BLOCK_GROUP_RAID1)) {
- if (fs_devices->rw_devices < 2)
- return -ENOSPC;
- *num_stripes = 2;
- *min_stripes = 2;
- }
- if (type & (BTRFS_BLOCK_GROUP_RAID10)) {
- *num_stripes = fs_devices->rw_devices;
- if (*num_stripes < 4)
- return -ENOSPC;
- *num_stripes &= ~(u32)1;
- *sub_stripes = 2;
- *min_stripes = 4;
+ if ((type & BTRFS_BLOCK_GROUP_RAID1) &&
+ (type & BTRFS_BLOCK_GROUP_DUP)) {
+ WARN_ON(1);
+ type &= ~BTRFS_BLOCK_GROUP_DUP;
}
- return 0;
-}
+ if (list_empty(&fs_devices->alloc_list))
+ return -ENOSPC;
-static u64 __btrfs_calc_stripe_size(struct btrfs_fs_devices *fs_devices,
- u64 proposed_size, u64 type,
- int num_stripes, int small_stripe)
-{
- int min_stripe_size = 1 * 1024 * 1024;
- u64 calc_size = proposed_size;
- u64 max_chunk_size = calc_size;
- int ncopies = 1;
+ sub_stripes = 1;
+ dev_stripes = 1;
+ devs_increment = 1;
+ ncopies = 1;
+ devs_max = 0; /* 0 == as many as possible */
+ devs_min = 1;
- if (type & (BTRFS_BLOCK_GROUP_RAID1 |
- BTRFS_BLOCK_GROUP_DUP |
- BTRFS_BLOCK_GROUP_RAID10))
+ /*
+ * define the properties of each RAID type.
+ * FIXME: move this to a global table and use it in all RAID
+ * calculation code
+ */
+ if (type & (BTRFS_BLOCK_GROUP_DUP)) {
+ dev_stripes = 2;
+ ncopies = 2;
+ devs_max = 1;
+ } else if (type & (BTRFS_BLOCK_GROUP_RAID0)) {
+ devs_min = 2;
+ } else if (type & (BTRFS_BLOCK_GROUP_RAID1)) {
+ devs_increment = 2;
ncopies = 2;
+ devs_max = 2;
+ devs_min = 2;
+ } else if (type & (BTRFS_BLOCK_GROUP_RAID10)) {
+ sub_stripes = 2;
+ devs_increment = 2;
+ ncopies = 2;
+ devs_min = 4;
+ } else {
+ devs_max = 1;
+ }
if (type & BTRFS_BLOCK_GROUP_DATA) {
- max_chunk_size = 10 * calc_size;
- min_stripe_size = 64 * 1024 * 1024;
+ max_stripe_size = 1024 * 1024 * 1024;
+ max_chunk_size = 10 * max_stripe_size;
} else if (type & BTRFS_BLOCK_GROUP_METADATA) {
- max_chunk_size = 256 * 1024 * 1024;
- min_stripe_size = 32 * 1024 * 1024;
+ max_stripe_size = 256 * 1024 * 1024;
+ max_chunk_size = max_stripe_size;
} else if (type & BTRFS_BLOCK_GROUP_SYSTEM) {
- calc_size = 8 * 1024 * 1024;
- max_chunk_size = calc_size * 2;
- min_stripe_size = 1 * 1024 * 1024;
+ max_stripe_size = 8 * 1024 * 1024;
+ max_chunk_size = 2 * max_stripe_size;
+ } else {
+ printk(KERN_ERR "btrfs: invalid chunk type 0x%llx requested\n",
+ type);
+ BUG_ON(1);
}
/* 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);
- if (calc_size * num_stripes > max_chunk_size * ncopies) {
- calc_size = max_chunk_size * ncopies;
- do_div(calc_size, num_stripes);
- do_div(calc_size, BTRFS_STRIPE_LEN);
- calc_size *= BTRFS_STRIPE_LEN;
- }
+ devices_info = kzalloc(sizeof(*devices_info) * fs_devices->rw_devices,
+ GFP_NOFS);
+ if (!devices_info)
+ return -ENOMEM;
- /* we don't want tiny stripes */
- if (!small_stripe)
- calc_size = max_t(u64, min_stripe_size, calc_size);
+ cur = fs_devices->alloc_list.next;
/*
- * we're about to do_div by the BTRFS_STRIPE_LEN so lets make sure
- * we end up with something bigger than a stripe
+ * in the first pass through the devices list, we gather information
+ * about the available holes on each device.
*/
- calc_size = max_t(u64, calc_size, BTRFS_STRIPE_LEN);
-
- do_div(calc_size, BTRFS_STRIPE_LEN);
- calc_size *= BTRFS_STRIPE_LEN;
-
- return calc_size;
-}
-
-static struct map_lookup *__shrink_map_lookup_stripes(struct map_lookup *map,
- int num_stripes)
-{
- struct map_lookup *new;
- size_t len = map_lookup_size(num_stripes);
-
- BUG_ON(map->num_stripes < num_stripes);
-
- if (map->num_stripes == num_stripes)
- return map;
-
- new = kmalloc(len, GFP_NOFS);
- if (!new) {
- /* just change map->num_stripes */
- map->num_stripes = num_stripes;
- return map;
- }
+ ndevs = 0;
+ while (cur != &fs_devices->alloc_list) {
+ struct btrfs_device *device;
+ u64 max_avail;
+ u64 dev_offset;
- memcpy(new, map, len);
- new->num_stripes = num_stripes;
- kfree(map);
- return new;
-}
-
-/*
- * helper to allocate device space from btrfs_device_info, in which we stored
- * max free space information of every device. It is used when we can not
- * allocate chunks by default size.
- *
- * By this helper, we can allocate a new chunk as larger as possible.
- */
-static int __btrfs_alloc_tiny_space(struct btrfs_trans_handle *trans,
- struct btrfs_fs_devices *fs_devices,
- struct btrfs_device_info *devices,
- int nr_device, u64 type,
- struct map_lookup **map_lookup,
- int min_stripes, u64 *stripe_size)
-{
- int i, index, sort_again = 0;
- int min_devices = min_stripes;
- u64 max_avail, min_free;
- struct map_lookup *map = *map_lookup;
- int ret;
+ device = list_entry(cur, struct btrfs_device, dev_alloc_list);
- if (nr_device < min_stripes)
- return -ENOSPC;
+ cur = cur->next;
- btrfs_descending_sort_devices(devices, nr_device);
+ if (!device->writeable) {
+ printk(KERN_ERR
+ "btrfs: read-only device in alloc_list\n");
+ WARN_ON(1);
+ continue;
+ }
- max_avail = devices[0].max_avail;
- if (!max_avail)
- return -ENOSPC;
+ if (!device->in_fs_metadata)
+ continue;
- for (i = 0; i < nr_device; i++) {
- /*
- * if dev_offset = 0, it means the free space of this device
- * is less than what we need, and we didn't search max avail
- * extent on this device, so do it now.
+ if (device->total_bytes > device->bytes_used)
+ total_avail = device->total_bytes - device->bytes_used;
+ else
+ total_avail = 0;
+ /* avail is off by max(alloc_start, 1MB), but that is the same
+ * for all devices, so it doesn't hurt the sorting later on
*/
- if (!devices[i].dev_offset) {
- ret = find_free_dev_extent(trans, devices[i].dev,
- max_avail,
- &devices[i].dev_offset,
- &devices[i].max_avail);
- if (ret != 0 && ret != -ENOSPC)
- return ret;
- sort_again = 1;
- }
- }
-
- /* we update the max avail free extent of each devices, sort again */
- if (sort_again)
- btrfs_descending_sort_devices(devices, nr_device);
- if (type & BTRFS_BLOCK_GROUP_DUP)
- min_devices = 1;
+ ret = find_free_dev_extent(trans, device,
+ max_stripe_size * dev_stripes,
+ &dev_offset, &max_avail);
+ if (ret && ret != -ENOSPC)
+ goto error;
- if (!devices[min_devices - 1].max_avail)
- return -ENOSPC;
+ if (ret == 0)
+ max_avail = max_stripe_size * dev_stripes;
- max_avail = devices[min_devices - 1].max_avail;
- if (type & BTRFS_BLOCK_GROUP_DUP)
- do_div(max_avail, 2);
+ if (max_avail < BTRFS_STRIPE_LEN * dev_stripes)
+ continue;
- max_avail = __btrfs_calc_stripe_size(fs_devices, max_avail, type,
- min_stripes, 1);
- if (type & BTRFS_BLOCK_GROUP_DUP)
- min_free = max_avail * 2;
- else
- min_free = max_avail;
+ devices_info[ndevs].dev_offset = dev_offset;
+ devices_info[ndevs].max_avail = max_avail;
+ devices_info[ndevs].total_avail = total_avail;
+ devices_info[ndevs].dev = device;
+ ++ndevs;
+ }
- if (min_free > devices[min_devices - 1].max_avail)
- return -ENOSPC;
+ /*
+ * now sort the devices by hole size / available space
+ */
+ sort(devices_info, ndevs, sizeof(struct btrfs_device_info),
+ btrfs_cmp_device_info, NULL);
- map = __shrink_map_lookup_stripes(map, min_stripes);
- *stripe_size = max_avail;
+ /* round down to number of usable stripes */
+ ndevs -= ndevs % devs_increment;
- index = 0;
- for (i = 0; i < min_stripes; i++) {
- map->stripes[i].dev = devices[index].dev;
- map->stripes[i].physical = devices[index].dev_offset;
- if (type & BTRFS_BLOCK_GROUP_DUP) {
- i++;
- map->stripes[i].dev = devices[index].dev;
- map->stripes[i].physical = devices[index].dev_offset +
- max_avail;
- }
- index++;
+ if (ndevs < devs_increment * sub_stripes || ndevs < devs_min) {
+ ret = -ENOSPC;
+ goto error;
}
- *map_lookup = map;
-
- return 0;
-}
-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)
-{
- struct btrfs_fs_info *info = extent_root->fs_info;
- struct btrfs_device *device = NULL;
- struct btrfs_fs_devices *fs_devices = info->fs_devices;
- struct list_head *cur;
- struct map_lookup *map;
- struct extent_map_tree *em_tree;
- struct extent_map *em;
- struct btrfs_device_info *devices_info;
- struct list_head private_devs;
- u64 calc_size = 1024 * 1024 * 1024;
- u64 min_free;
- u64 avail;
- u64 dev_offset;
- int num_stripes;
- int min_stripes;
- int sub_stripes;
- int min_devices; /* the min number of devices we need */
- int i;
- int ret;
- int index;
+ if (devs_max && ndevs > devs_max)
+ ndevs = devs_max;
+ /*
+ * the primary goal is to maximize the number of stripes, so use as many
+ * devices as possible, even if the stripes are not maximum sized.
+ */
+ stripe_size = devices_info[ndevs-1].max_avail;
+ num_stripes = ndevs * dev_stripes;
- if ((type & BTRFS_BLOCK_GROUP_RAID1) &&
- (type & BTRFS_BLOCK_GROUP_DUP)) {
- WARN_ON(1);
- type &= ~BTRFS_BLOCK_GROUP_DUP;
+ if (stripe_size * num_stripes > max_chunk_size * ncopies) {
+ stripe_size = max_chunk_size * ncopies;
+ do_div(stripe_size, num_stripes);
}
- if (list_empty(&fs_devices->alloc_list))
- return -ENOSPC;
- ret = __btrfs_calc_nstripes(fs_devices, type, &num_stripes,
- &min_stripes, &sub_stripes);
- if (ret)
- return ret;
-
- devices_info = kzalloc(sizeof(*devices_info) * fs_devices->rw_devices,
- GFP_NOFS);
- if (!devices_info)
- return -ENOMEM;
+ do_div(stripe_size, dev_stripes);
+ do_div(stripe_size, BTRFS_STRIPE_LEN);
+ stripe_size *= BTRFS_STRIPE_LEN;
map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
if (!map) {
}
map->num_stripes = num_stripes;
- cur = fs_devices->alloc_list.next;
- index = 0;
- i = 0;
-
- calc_size = __btrfs_calc_stripe_size(fs_devices, calc_size, type,
- num_stripes, 0);
-
- if (type & BTRFS_BLOCK_GROUP_DUP) {
- min_free = calc_size * 2;
- min_devices = 1;
- } else {
- min_free = calc_size;
- min_devices = min_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
- avail = 0;
- cur = cur->next;
-
- if (device->in_fs_metadata && avail >= min_free) {
- ret = find_free_dev_extent(trans, device, min_free,
- &devices_info[i].dev_offset,
- &devices_info[i].max_avail);
- if (ret == 0) {
- list_move_tail(&device->dev_alloc_list,
- &private_devs);
- map->stripes[index].dev = device;
- map->stripes[index].physical =
- devices_info[i].dev_offset;
- index++;
- if (type & BTRFS_BLOCK_GROUP_DUP) {
- map->stripes[index].dev = device;
- map->stripes[index].physical =
- devices_info[i].dev_offset +
- calc_size;
- index++;
- }
- } else if (ret != -ENOSPC)
- goto error;
-
- devices_info[i].dev = device;
- i++;
- } else if (device->in_fs_metadata &&
- avail >= BTRFS_STRIPE_LEN) {
- devices_info[i].dev = device;
- devices_info[i].max_avail = avail;
- i++;
- }
-
- if (cur == &fs_devices->alloc_list)
- break;
- }
-
- list_splice(&private_devs, &fs_devices->alloc_list);
- if (index < num_stripes) {
- if (index >= min_stripes) {
- num_stripes = index;
- if (type & (BTRFS_BLOCK_GROUP_RAID10)) {
- num_stripes /= sub_stripes;
- num_stripes *= sub_stripes;
- }
-
- map = __shrink_map_lookup_stripes(map, num_stripes);
- } else if (i >= min_devices) {
- ret = __btrfs_alloc_tiny_space(trans, fs_devices,
- devices_info, i, type,
- &map, min_stripes,
- &calc_size);
- if (ret)
- goto error;
- } else {
- ret = -ENOSPC;
- goto error;
+ for (i = 0; i < ndevs; ++i) {
+ for (j = 0; j < dev_stripes; ++j) {
+ int s = i * dev_stripes + j;
+ map->stripes[s].dev = devices_info[i].dev;
+ map->stripes[s].physical = devices_info[i].dev_offset +
+ j * stripe_size;
}
}
map->sector_size = extent_root->sectorsize;
map->sub_stripes = sub_stripes;
*map_ret = map;
- *stripe_size = calc_size;
- *num_bytes = chunk_bytes_by_type(type, calc_size,
- map->num_stripes, sub_stripes);
+ num_bytes = stripe_size * (num_stripes / ncopies);
- trace_btrfs_chunk_alloc(info->chunk_root, map, start, *num_bytes);
+ *stripe_size_out = stripe_size;
+ *num_bytes_out = num_bytes;
+
+ trace_btrfs_chunk_alloc(info->chunk_root, map, start, num_bytes);
em = alloc_extent_map();
if (!em) {
}
em->bdev = (struct block_device *)map;
em->start = start;
- em->len = *num_bytes;
+ em->len = num_bytes;
em->block_start = 0;
em->block_len = em->len;
ret = btrfs_make_block_group(trans, extent_root, 0, type,
BTRFS_FIRST_CHUNK_TREE_OBJECTID,
- start, *num_bytes);
+ start, num_bytes);
BUG_ON(ret);
- index = 0;
- while (index < map->num_stripes) {
- device = map->stripes[index].dev;
- dev_offset = map->stripes[index].physical;
+ for (i = 0; i < map->num_stripes; ++i) {
+ struct btrfs_device *device;
+ u64 dev_offset;
+
+ device = map->stripes[i].dev;
+ dev_offset = map->stripes[i].physical;
ret = btrfs_alloc_dev_extent(trans, device,
info->chunk_root->root_key.objectid,
BTRFS_FIRST_CHUNK_TREE_OBJECTID,
- start, dev_offset, calc_size);
+ start, dev_offset, stripe_size);
BUG_ON(ret);
- index++;
}
kfree(devices_info);
ret = find_next_chunk(fs_info->chunk_root,
BTRFS_FIRST_CHUNK_TREE_OBJECTID, &chunk_offset);
- BUG_ON(ret);
+ if (ret)
+ return ret;
alloc_profile = BTRFS_BLOCK_GROUP_METADATA |
(fs_info->metadata_alloc_profile &
return ret;
}
-int btrfs_read_super_device(struct btrfs_root *root, struct extent_buffer *buf)
-{
- struct btrfs_dev_item *dev_item;
-
- dev_item = (struct btrfs_dev_item *)offsetof(struct btrfs_super_block,
- dev_item);
- return read_one_dev(root, buf, dev_item);
-}
-
int btrfs_read_sys_array(struct btrfs_root *root)
{
struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
if (!sb)
return -ENOMEM;
btrfs_set_buffer_uptodate(sb);
- btrfs_set_buffer_lockdep_class(sb, 0);
+ btrfs_set_buffer_lockdep_class(root->root_key.objectid, sb, 0);
write_extent_buffer(sb, super_copy, 0, BTRFS_SUPER_INFO_SIZE);
array_size = btrfs_super_sys_array_size(super_copy);
Code Review / linux-2.6.git / blobdiff