#include <linux/kthread.h>
#include <linux/slab.h>
#include <linux/ratelimit.h>
+#include <linux/percpu_counter.h>
#include "compat.h"
#include "hash.h"
#include "ctree.h"
/* need to make sure the commit_root doesn't disappear */
down_read(&fs_info->extent_commit_sem);
+next:
ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
if (ret < 0)
goto err;
continue;
}
+ if (key.objectid < last) {
+ key.objectid = last;
+ key.offset = 0;
+ key.type = BTRFS_EXTENT_ITEM_KEY;
+
+ caching_ctl->progress = last;
+ btrfs_release_path(path);
+ goto next;
+ }
+
if (key.objectid < block_group->key.objectid) {
path->slots[0]++;
continue;
goto out_free;
if (ret > 0 && metadata && key.type == BTRFS_METADATA_ITEM_KEY) {
- key.type = BTRFS_EXTENT_ITEM_KEY;
- key.offset = root->leafsize;
- btrfs_release_path(path);
- goto again;
+ metadata = 0;
+ if (path->slots[0]) {
+ path->slots[0]--;
+ btrfs_item_key_to_cpu(path->nodes[0], &key,
+ path->slots[0]);
+ if (key.objectid == bytenr &&
+ key.type == BTRFS_EXTENT_ITEM_KEY &&
+ key.offset == root->leafsize)
+ ret = 0;
+ }
+ if (ret) {
+ key.objectid = bytenr;
+ key.type = BTRFS_EXTENT_ITEM_KEY;
+ key.offset = root->leafsize;
+ btrfs_release_path(path);
+ goto again;
+ }
}
if (ret == 0) {
return 0;
}
+static inline u64 heads_to_leaves(struct btrfs_root *root, u64 heads)
+{
+ u64 num_bytes;
+
+ num_bytes = heads * (sizeof(struct btrfs_extent_item) +
+ sizeof(struct btrfs_extent_inline_ref));
+ if (!btrfs_fs_incompat(root->fs_info, SKINNY_METADATA))
+ num_bytes += heads * sizeof(struct btrfs_tree_block_info);
+
+ /*
+ * We don't ever fill up leaves all the way so multiply by 2 just to be
+ * closer to what we're really going to want to ouse.
+ */
+ return div64_u64(num_bytes, BTRFS_LEAF_DATA_SIZE(root));
+}
+
+int btrfs_should_throttle_delayed_refs(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root)
+{
+ struct btrfs_block_rsv *global_rsv;
+ u64 num_heads = trans->transaction->delayed_refs.num_heads_ready;
+ u64 num_bytes;
+ int ret = 0;
+
+ num_bytes = btrfs_calc_trans_metadata_size(root, 1);
+ num_heads = heads_to_leaves(root, num_heads);
+ if (num_heads > 1)
+ num_bytes += (num_heads - 1) * root->leafsize;
+ num_bytes <<= 1;
+ global_rsv = &root->fs_info->global_block_rsv;
+
+ /*
+ * If we can't allocate any more chunks lets make sure we have _lots_ of
+ * wiggle room since running delayed refs can create more delayed refs.
+ */
+ if (global_rsv->space_info->full)
+ num_bytes <<= 1;
+
+ spin_lock(&global_rsv->lock);
+ if (global_rsv->reserved <= num_bytes)
+ ret = 1;
+ spin_unlock(&global_rsv->lock);
+ return ret;
+}
+
/*
* this starts processing the delayed reference count updates and
* extent insertions we have queued up so far. count can be
old = atomic_cmpxchg(&delayed_refs->procs_running_refs, 0, 1);
if (old) {
DEFINE_WAIT(__wait);
- if (delayed_refs->num_entries < 16348)
+ if (delayed_refs->flushing ||
+ !btrfs_should_throttle_delayed_refs(trans, root))
return 0;
prepare_to_wait(&delayed_refs->wait, &__wait,
while (1) {
if (!(run_all || run_most) &&
- delayed_refs->num_heads_ready < 64)
+ !btrfs_should_throttle_delayed_refs(trans, root))
break;
/*
spin_unlock(&delayed_refs->lock);
btrfs_abort_transaction(trans, root, ret);
atomic_dec(&delayed_refs->procs_running_refs);
+ wake_up(&delayed_refs->wait);
return ret;
}
struct btrfs_space_info *found;
int i;
int factor;
+ int ret;
if (flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 |
BTRFS_BLOCK_GROUP_RAID10))
if (!found)
return -ENOMEM;
+ ret = percpu_counter_init(&found->total_bytes_pinned, 0);
+ if (ret) {
+ kfree(found);
+ return ret;
+ }
+
for (i = 0; i < BTRFS_NR_RAID_TYPES; i++)
INIT_LIST_HEAD(&found->block_groups[i]);
init_rwsem(&found->groups_sem);
}
/*
- * If we have less pinned bytes than we want to allocate then
- * don't bother committing the transaction, it won't help us.
+ * If we don't have enough pinned space to deal with this
+ * allocation don't bother committing the transaction.
*/
- if (data_sinfo->bytes_pinned < bytes)
+ if (percpu_counter_compare(&data_sinfo->total_bytes_pinned,
+ bytes) < 0)
committed = 1;
spin_unlock(&data_sinfo->lock);
if (!committed &&
!atomic_read(&root->fs_info->open_ioctl_trans)) {
committed = 1;
+
trans = btrfs_join_transaction(root);
if (IS_ERR(trans))
return PTR_ERR(trans);
data_sinfo = root->fs_info->data_sinfo;
spin_lock(&data_sinfo->lock);
+ WARN_ON(data_sinfo->bytes_may_use < bytes);
data_sinfo->bytes_may_use -= bytes;
trace_btrfs_space_reservation(root->fs_info, "space_info",
data_sinfo->flags, bytes, 0);
unsigned long nr_pages)
{
struct super_block *sb = root->fs_info->sb;
- int started;
- /* If we can not start writeback, just sync all the delalloc file. */
- started = try_to_writeback_inodes_sb_nr(sb, nr_pages,
- WB_REASON_FS_FREE_SPACE);
- if (!started) {
+ if (down_read_trylock(&sb->s_umount)) {
+ writeback_inodes_sb_nr(sb, nr_pages, WB_REASON_FS_FREE_SPACE);
+ up_read(&sb->s_umount);
+ } else {
/*
* We needn't worry the filesystem going from r/w to r/o though
* we don't acquire ->s_umount mutex, because the filesystem
* the filesystem is readonly(all dirty pages are written to
* the disk).
*/
- btrfs_start_delalloc_inodes(root, 0);
+ btrfs_start_all_delalloc_inodes(root->fs_info, 0);
if (!current->journal_info)
- btrfs_wait_ordered_extents(root, 0);
+ btrfs_wait_all_ordered_extents(root->fs_info, 0);
}
}
if (delalloc_bytes == 0) {
if (trans)
return;
- btrfs_wait_ordered_extents(root, 0);
+ btrfs_wait_all_ordered_extents(root->fs_info, 0);
return;
}
loops++;
if (wait_ordered && !trans) {
- btrfs_wait_ordered_extents(root, 0);
+ btrfs_wait_all_ordered_extents(root->fs_info, 0);
} else {
time_left = schedule_timeout_killable(1);
if (time_left)
/* See if there is enough pinned space to make this reservation */
spin_lock(&space_info->lock);
- if (space_info->bytes_pinned >= bytes) {
+ if (percpu_counter_compare(&space_info->total_bytes_pinned,
+ bytes) >= 0) {
spin_unlock(&space_info->lock);
goto commit;
}
spin_lock(&space_info->lock);
spin_lock(&delayed_rsv->lock);
- if (space_info->bytes_pinned + delayed_rsv->size < bytes) {
+ if (percpu_counter_compare(&space_info->total_bytes_pinned,
+ bytes - delayed_rsv->size) >= 0) {
spin_unlock(&delayed_rsv->lock);
spin_unlock(&space_info->lock);
return -ENOSPC;
spin_unlock(&block_rsv->lock);
}
+int btrfs_cond_migrate_bytes(struct btrfs_fs_info *fs_info,
+ struct btrfs_block_rsv *dest, u64 num_bytes,
+ int min_factor)
+{
+ struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv;
+ u64 min_bytes;
+
+ if (global_rsv->space_info != dest->space_info)
+ return -ENOSPC;
+
+ spin_lock(&global_rsv->lock);
+ min_bytes = div_factor(global_rsv->size, min_factor);
+ if (global_rsv->reserved < min_bytes + num_bytes) {
+ spin_unlock(&global_rsv->lock);
+ return -ENOSPC;
+ }
+ global_rsv->reserved -= num_bytes;
+ if (global_rsv->reserved < global_rsv->size)
+ global_rsv->full = 0;
+ spin_unlock(&global_rsv->lock);
+
+ block_rsv_add_bytes(dest, num_bytes, 1);
+ return 0;
+}
+
static void block_rsv_release_bytes(struct btrfs_fs_info *fs_info,
struct btrfs_block_rsv *block_rsv,
struct btrfs_block_rsv *dest, u64 num_bytes)
int btrfs_subvolume_reserve_metadata(struct btrfs_root *root,
struct btrfs_block_rsv *rsv,
int items,
- u64 *qgroup_reserved)
+ u64 *qgroup_reserved,
+ bool use_global_rsv)
{
u64 num_bytes;
int ret;
+ struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv;
if (root->fs_info->quota_enabled) {
/* One for parent inode, two for dir entries */
BTRFS_BLOCK_GROUP_METADATA);
ret = btrfs_block_rsv_add(root, rsv, num_bytes,
BTRFS_RESERVE_FLUSH_ALL);
+
+ if (ret == -ENOSPC && use_global_rsv)
+ ret = btrfs_block_rsv_migrate(global_rsv, rsv, num_bytes);
+
if (ret) {
if (*qgroup_reserved)
btrfs_qgroup_free(root, *qgroup_reserved);
int factor;
/* block accounting for super block */
- spin_lock(&info->delalloc_lock);
+ spin_lock(&info->delalloc_root_lock);
old_val = btrfs_super_bytes_used(info->super_copy);
if (alloc)
old_val += num_bytes;
else
old_val -= num_bytes;
btrfs_set_super_bytes_used(info->super_copy, old_val);
- spin_unlock(&info->delalloc_lock);
+ spin_unlock(&info->delalloc_root_lock);
while (total) {
cache = btrfs_lookup_block_group(info, bytenr);
return ret;
}
+static int __exclude_logged_extent(struct btrfs_root *root, u64 start, u64 num_bytes)
+{
+ int ret;
+ struct btrfs_block_group_cache *block_group;
+ struct btrfs_caching_control *caching_ctl;
+
+ block_group = btrfs_lookup_block_group(root->fs_info, start);
+ if (!block_group)
+ return -EINVAL;
+
+ cache_block_group(block_group, 0);
+ caching_ctl = get_caching_control(block_group);
+
+ if (!caching_ctl) {
+ /* Logic error */
+ BUG_ON(!block_group_cache_done(block_group));
+ ret = btrfs_remove_free_space(block_group, start, num_bytes);
+ } else {
+ mutex_lock(&caching_ctl->mutex);
+
+ if (start >= caching_ctl->progress) {
+ ret = add_excluded_extent(root, start, num_bytes);
+ } else if (start + num_bytes <= caching_ctl->progress) {
+ ret = btrfs_remove_free_space(block_group,
+ start, num_bytes);
+ } else {
+ num_bytes = caching_ctl->progress - start;
+ ret = btrfs_remove_free_space(block_group,
+ start, num_bytes);
+ if (ret)
+ goto out_lock;
+
+ num_bytes = (start + num_bytes) -
+ caching_ctl->progress;
+ start = caching_ctl->progress;
+ ret = add_excluded_extent(root, start, num_bytes);
+ }
+out_lock:
+ mutex_unlock(&caching_ctl->mutex);
+ put_caching_control(caching_ctl);
+ }
+ btrfs_put_block_group(block_group);
+ return ret;
+}
+
+int btrfs_exclude_logged_extents(struct btrfs_root *log,
+ struct extent_buffer *eb)
+{
+ struct btrfs_file_extent_item *item;
+ struct btrfs_key key;
+ int found_type;
+ int i;
+
+ if (!btrfs_fs_incompat(log->fs_info, MIXED_GROUPS))
+ return 0;
+
+ for (i = 0; i < btrfs_header_nritems(eb); i++) {
+ btrfs_item_key_to_cpu(eb, &key, i);
+ if (key.type != BTRFS_EXTENT_DATA_KEY)
+ continue;
+ item = btrfs_item_ptr(eb, i, struct btrfs_file_extent_item);
+ found_type = btrfs_file_extent_type(eb, item);
+ if (found_type == BTRFS_FILE_EXTENT_INLINE)
+ continue;
+ if (btrfs_file_extent_disk_bytenr(eb, item) == 0)
+ continue;
+ key.objectid = btrfs_file_extent_disk_bytenr(eb, item);
+ key.offset = btrfs_file_extent_disk_num_bytes(eb, item);
+ __exclude_logged_extent(log, key.objectid, key.offset);
+ }
+
+ return 0;
+}
+
/**
* btrfs_update_reserved_bytes - update the block_group and space info counters
* @cache: The cache we are manipulating
struct btrfs_caching_control *next;
struct btrfs_caching_control *caching_ctl;
struct btrfs_block_group_cache *cache;
+ struct btrfs_space_info *space_info;
down_write(&fs_info->extent_commit_sem);
up_write(&fs_info->extent_commit_sem);
+ list_for_each_entry_rcu(space_info, &fs_info->space_info, list)
+ percpu_counter_set(&space_info->total_bytes_pinned, 0);
+
update_global_block_rsv(fs_info);
}
return 0;
}
+static void add_pinned_bytes(struct btrfs_fs_info *fs_info, u64 num_bytes,
+ u64 owner, u64 root_objectid)
+{
+ struct btrfs_space_info *space_info;
+ u64 flags;
+
+ if (owner < BTRFS_FIRST_FREE_OBJECTID) {
+ if (root_objectid == BTRFS_CHUNK_TREE_OBJECTID)
+ flags = BTRFS_BLOCK_GROUP_SYSTEM;
+ else
+ flags = BTRFS_BLOCK_GROUP_METADATA;
+ } else {
+ flags = BTRFS_BLOCK_GROUP_DATA;
+ }
+
+ space_info = __find_space_info(fs_info, flags);
+ BUG_ON(!space_info); /* Logic bug */
+ percpu_counter_add(&space_info->total_bytes_pinned, num_bytes);
+}
+
+
static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 bytenr, u64 num_bytes, u64 parent,
goto out;
}
}
+ add_pinned_bytes(root->fs_info, -num_bytes, owner_objectid,
+ root_objectid);
} else {
if (found_extent) {
BUG_ON(is_data && refs_to_drop !=
u64 parent, int last_ref)
{
struct btrfs_block_group_cache *cache = NULL;
+ int pin = 1;
int ret;
if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) {
btrfs_add_free_space(cache, buf->start, buf->len);
btrfs_update_reserved_bytes(cache, buf->len, RESERVE_FREE);
+ pin = 0;
}
out:
+ if (pin)
+ add_pinned_bytes(root->fs_info, buf->len,
+ btrfs_header_level(buf),
+ root->root_key.objectid);
+
/*
* Deleting the buffer, clear the corrupt flag since it doesn't matter
* anymore.
int ret;
struct btrfs_fs_info *fs_info = root->fs_info;
+ add_pinned_bytes(root->fs_info, num_bytes, owner, root_objectid);
+
/*
* tree log blocks never actually go into the extent allocation
* tree, just update pinning info and exit early.
{
int ret;
struct btrfs_block_group_cache *block_group;
- struct btrfs_caching_control *caching_ctl;
- u64 start = ins->objectid;
- u64 num_bytes = ins->offset;
- block_group = btrfs_lookup_block_group(root->fs_info, ins->objectid);
- cache_block_group(block_group, 0);
- caching_ctl = get_caching_control(block_group);
-
- if (!caching_ctl) {
- BUG_ON(!block_group_cache_done(block_group));
- ret = btrfs_remove_free_space(block_group, start, num_bytes);
- if (ret)
- goto out;
- } else {
- mutex_lock(&caching_ctl->mutex);
-
- if (start >= caching_ctl->progress) {
- ret = add_excluded_extent(root, start, num_bytes);
- } else if (start + num_bytes <= caching_ctl->progress) {
- ret = btrfs_remove_free_space(block_group,
- start, num_bytes);
- } else {
- num_bytes = caching_ctl->progress - start;
- ret = btrfs_remove_free_space(block_group,
- start, num_bytes);
- if (ret)
- goto out_lock;
-
- start = caching_ctl->progress;
- num_bytes = ins->objectid + ins->offset -
- caching_ctl->progress;
- ret = add_excluded_extent(root, start, num_bytes);
- }
-out_lock:
- mutex_unlock(&caching_ctl->mutex);
- put_caching_control(caching_ctl);
+ /*
+ * Mixed block groups will exclude before processing the log so we only
+ * need to do the exlude dance if this fs isn't mixed.
+ */
+ if (!btrfs_fs_incompat(root->fs_info, MIXED_GROUPS)) {
+ ret = __exclude_logged_extent(root, ins->objectid, ins->offset);
if (ret)
- goto out;
+ return ret;
}
+ block_group = btrfs_lookup_block_group(root->fs_info, ins->objectid);
+ if (!block_group)
+ return -EINVAL;
+
ret = btrfs_update_reserved_bytes(block_group, ins->offset,
RESERVE_ALLOC_NO_ACCOUNT);
BUG_ON(ret); /* logic error */
ret = alloc_reserved_file_extent(trans, root, 0, root_objectid,
0, owner, offset, ins, 1);
-out:
btrfs_put_block_group(block_group);
return ret;
}
next = btrfs_find_create_tree_block(root, bytenr, blocksize);
if (!next)
return -ENOMEM;
+ btrfs_set_buffer_lockdep_class(root->root_key.objectid, next,
+ level - 1);
reada = 1;
}
btrfs_tree_lock(next);
int err = 0;
int ret;
int level;
+ bool root_dropped = false;
path = btrfs_alloc_path();
if (!path) {
while (1) {
btrfs_tree_lock(path->nodes[level]);
btrfs_set_lock_blocking(path->nodes[level]);
+ path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING;
ret = btrfs_lookup_extent_info(trans, root,
path->nodes[level]->start,
break;
btrfs_tree_unlock(path->nodes[level]);
+ path->locks[level] = 0;
WARN_ON(wc->refs[level] != 1);
level--;
}
wc->reada_count = BTRFS_NODEPTRS_PER_BLOCK(root);
while (1) {
- if (!for_reloc && btrfs_fs_closing(root->fs_info)) {
- pr_debug("btrfs: drop snapshot early exit\n");
- err = -EAGAIN;
- goto out_end_trans;
- }
ret = walk_down_tree(trans, root, path, wc);
if (ret < 0) {
}
BUG_ON(wc->level == 0);
- if (btrfs_should_end_transaction(trans, tree_root)) {
+ if (btrfs_should_end_transaction(trans, tree_root) ||
+ (!for_reloc && btrfs_need_cleaner_sleep(root))) {
ret = btrfs_update_root(trans, tree_root,
&root->root_key,
root_item);
}
btrfs_end_transaction_throttle(trans, tree_root);
+ if (!for_reloc && btrfs_need_cleaner_sleep(root)) {
+ pr_debug("btrfs: drop snapshot early exit\n");
+ err = -EAGAIN;
+ goto out_free;
+ }
+
trans = btrfs_start_transaction(tree_root, 0);
if (IS_ERR(trans)) {
err = PTR_ERR(trans);
}
if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID) {
- ret = btrfs_find_last_root(tree_root, root->root_key.objectid,
- NULL, NULL);
+ ret = btrfs_find_root(tree_root, &root->root_key, path,
+ NULL, NULL);
if (ret < 0) {
btrfs_abort_transaction(trans, tree_root, ret);
err = ret;
}
if (root->in_radix) {
- btrfs_free_fs_root(tree_root->fs_info, root);
+ btrfs_drop_and_free_fs_root(tree_root->fs_info, root);
} else {
free_extent_buffer(root->node);
free_extent_buffer(root->commit_root);
- kfree(root);
+ btrfs_put_fs_root(root);
}
+ root_dropped = true;
out_end_trans:
btrfs_end_transaction_throttle(trans, tree_root);
out_free:
kfree(wc);
btrfs_free_path(path);
out:
+ /*
+ * So if we need to stop dropping the snapshot for whatever reason we
+ * need to make sure to add it back to the dead root list so that we
+ * keep trying to do the work later. This also cleans up roots if we
+ * don't have it in the radix (like when we recover after a power fail
+ * or unmount) so we don't leak memory.
+ */
+ if (root_dropped == false)
+ btrfs_add_dead_root(root);
if (err)
btrfs_std_error(root->fs_info, err);
return err;
struct btrfs_space_info *space_info;
struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
struct btrfs_device *device;
+ struct btrfs_trans_handle *trans;
u64 min_free;
u64 dev_min = 1;
u64 dev_nr = 0;
do_div(min_free, dev_min);
}
+ /* We need to do this so that we can look at pending chunks */
+ trans = btrfs_join_transaction(root);
+ if (IS_ERR(trans)) {
+ ret = PTR_ERR(trans);
+ goto out;
+ }
+
mutex_lock(&root->fs_info->chunk_mutex);
list_for_each_entry(device, &fs_devices->alloc_list, dev_alloc_list) {
u64 dev_offset;
*/
if (device->total_bytes > device->bytes_used + min_free &&
!device->is_tgtdev_for_dev_replace) {
- ret = find_free_dev_extent(device, min_free,
+ ret = find_free_dev_extent(trans, device, min_free,
&dev_offset, NULL);
if (!ret)
dev_nr++;
}
}
mutex_unlock(&root->fs_info->chunk_mutex);
+ btrfs_end_transaction(trans, root);
out:
btrfs_put_block_group(block_group);
return ret;
dump_space_info(space_info, 0, 0);
}
}
+ percpu_counter_destroy(&space_info->total_bytes_pinned);
list_del(&space_info->list);
kfree(space_info);
}
* avoid allocating from un-mirrored block group if there are
* mirrored block groups.
*/
- list_for_each_entry(cache, &space_info->block_groups[3], list)
+ list_for_each_entry(cache,
+ &space_info->block_groups[BTRFS_RAID_RAID0],
+ list)
set_block_group_ro(cache, 1);
- list_for_each_entry(cache, &space_info->block_groups[4], list)
+ list_for_each_entry(cache,
+ &space_info->block_groups[BTRFS_RAID_SINGLE],
+ list)
set_block_group_ro(cache, 1);
}
sizeof(item));
if (ret)
btrfs_abort_transaction(trans, extent_root, ret);
+ ret = btrfs_finish_chunk_alloc(trans, extent_root,
+ key.objectid, key.offset);
+ if (ret)
+ btrfs_abort_transaction(trans, extent_root, ret);
}
}
if (end - start >= range->minlen) {
if (!block_group_cache_done(cache)) {
ret = cache_block_group(cache, 0);
- if (!ret)
- wait_block_group_cache_done(cache);
+ if (ret) {
+ btrfs_put_block_group(cache);
+ break;
+ }
+ ret = wait_block_group_cache_done(cache);
+ if (ret) {
+ btrfs_put_block_group(cache);
+ break;
+ }
}
ret = btrfs_trim_block_group(cache,
&group_trimmed,
projects / deliverable / linux.git / blobdiff