Currently qgroups account for space by intercepting delayed ref updates to fs
trees. It does this by adding sequence numbers to delayed ref updates so that
it can figure out how the tree looked before the update so we can adjust the
counters properly. The problem with this is that it does not allow delayed refs
to be merged, so if you say are defragging an extent with 5k snapshots pointing
to it we will thrash the delayed ref lock because we need to go back and
manually merge these things together. Instead we want to process quota changes
when we know they are going to happen, like when we first allocate an extent, we
free a reference for an extent, we add new references etc. This patch
accomplishes this by only adding qgroup operations for real ref changes. We
only modify the sequence number when we need to lookup roots for bytenrs, this
reduces the amount of churn on the sequence number and allows us to merge
delayed refs as we add them most of the time. This patch encompasses a bunch of
architectural changes
1) qgroup ref operations: instead of tracking qgroup operations through the
delayed refs we simply add new ref operations whenever we notice that we need to
when we've modified the refs themselves.
2) tree mod seq: we no longer have this separation of major/minor counters.
this makes the sequence number stuff much more sane and we can remove some
locking that was needed to protect the counter.
3) delayed ref seq: we now read the tree mod seq number and use that as our
sequence. This means each new delayed ref doesn't have it's own unique
sequence
number, rather whenever we go to lookup backrefs we inc the sequence number so
we can make sure to keep any new operations from screwing up our world view at
that given point. This allows us to merge delayed refs during runtime.
With all of these changes the delayed ref stuff is a little saner and the qgroup
accounting stuff no longer goes negative in some cases like it was before.
Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
---
fs/btrfs/backref.h | 4 +-
fs/btrfs/ctree.c | 45 +--
fs/btrfs/ctree.h | 59 +---
fs/btrfs/delayed-ref.c | 39 ++-
fs/btrfs/delayed-ref.h | 24 +-
fs/btrfs/disk-io.c | 4 +
fs/btrfs/extent-tree.c | 206 +++++++----
fs/btrfs/file.c | 5 +-
fs/btrfs/ioctl.c | 63 +++-
fs/btrfs/qgroup.c | 915 ++++++++++++++++++++++++++++++++++++-------------
fs/btrfs/transaction.c | 52 +--
11 files changed, 937 insertions(+), 479 deletions(-)
diff --git a/fs/btrfs/backref.h b/fs/btrfs/backref.h
index a910b27..94e9442 100644
--- a/fs/btrfs/backref.h
+++ b/fs/btrfs/backref.h
@@ -55,8 +55,8 @@ int iterate_inodes_from_logical(u64 logical, struct
btrfs_fs_info *fs_info,
int paths_from_inode(u64 inum, struct inode_fs_paths *ipath);
int btrfs_find_all_roots(struct btrfs_trans_handle *trans,
- struct btrfs_fs_info *fs_info, u64 bytenr,
- u64 time_seq, struct ulist **roots);
+ struct btrfs_fs_info *fs_info, u64 bytenr,
+ u64 time_seq, struct ulist **roots);
char *btrfs_ref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
u32 name_len, unsigned long name_off,
struct extent_buffer *eb_in, u64 parent,
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index 0e9569c..208a84d 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -354,44 +354,14 @@ static inline void tree_mod_log_write_unlock(struct
btrfs_fs_info *fs_info)
}
/*
- * Increment the upper half of tree_mod_seq, set lower half zero.
- *
- * Must be called with fs_info->tree_mod_seq_lock held.
- */
-static inline u64 btrfs_inc_tree_mod_seq_major(struct btrfs_fs_info *fs_info)
-{
- u64 seq = atomic64_read(&fs_info->tree_mod_seq);
- seq &= 0xffffffff00000000ull;
- seq += 1ull << 32;
- atomic64_set(&fs_info->tree_mod_seq, seq);
- return seq;
-}
-
-/*
- * Increment the lower half of tree_mod_seq.
- *
- * Must be called with fs_info->tree_mod_seq_lock held. The way major
numbers
- * are generated should not technically require a spin lock here. (Rationale:
- * incrementing the minor while incrementing the major seq number is between
its
- * atomic64_read and atomic64_set calls doesn't duplicate sequence numbers,
it
- * just returns a unique sequence number as usual.) We have decided to leave
- * that requirement in here and rethink it once we notice it really imposes a
- * problem on some workload.
+ * Pull a new tree mod seq number for our operation.
*/
-static inline u64 btrfs_inc_tree_mod_seq_minor(struct btrfs_fs_info *fs_info)
+static inline u64 btrfs_inc_tree_mod_seq(struct btrfs_fs_info *fs_info)
{
return atomic64_inc_return(&fs_info->tree_mod_seq);
}
/*
- * return the last minor in the previous major tree_mod_seq number
- */
-u64 btrfs_tree_mod_seq_prev(u64 seq)
-{
- return (seq & 0xffffffff00000000ull) - 1ull;
-}
-
-/*
* This adds a new blocker to the tree mod log's blocker list if the @elem
* passed does not already have a sequence number set. So when a caller expects
* to record tree modifications, it should ensure to set elem->seq to zero
@@ -402,19 +372,16 @@ u64 btrfs_tree_mod_seq_prev(u64 seq)
u64 btrfs_get_tree_mod_seq(struct btrfs_fs_info *fs_info,
struct seq_list *elem)
{
- u64 seq;
-
tree_mod_log_write_lock(fs_info);
spin_lock(&fs_info->tree_mod_seq_lock);
if (!elem->seq) {
- elem->seq = btrfs_inc_tree_mod_seq_major(fs_info);
+ elem->seq = btrfs_inc_tree_mod_seq(fs_info);
list_add_tail(&elem->list, &fs_info->tree_mod_seq_list);
}
- seq = btrfs_inc_tree_mod_seq_minor(fs_info);
spin_unlock(&fs_info->tree_mod_seq_lock);
tree_mod_log_write_unlock(fs_info);
- return seq;
+ return elem->seq;
}
void btrfs_put_tree_mod_seq(struct btrfs_fs_info *fs_info,
@@ -487,9 +454,7 @@ __tree_mod_log_insert(struct btrfs_fs_info *fs_info, struct
tree_mod_elem *tm)
BUG_ON(!tm);
- spin_lock(&fs_info->tree_mod_seq_lock);
- tm->seq = btrfs_inc_tree_mod_seq_minor(fs_info);
- spin_unlock(&fs_info->tree_mod_seq_lock);
+ tm->seq = btrfs_inc_tree_mod_seq(fs_info);
tm_root = &fs_info->tree_mod_log;
new = &tm_root->rb_node;
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 45c4f59..33a1b27 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -1636,7 +1636,10 @@ struct btrfs_fs_info {
/* holds configuration and tracking. Protected by qgroup_lock */
struct rb_root qgroup_tree;
+ struct rb_root qgroup_op_tree;
spinlock_t qgroup_lock;
+ spinlock_t qgroup_op_lock;
+ atomic_t qgroup_op_seq;
/*
* used to avoid frequently calling ulist_alloc()/ulist_free()
@@ -3281,9 +3284,9 @@ int btrfs_reserve_extent(struct btrfs_root *root, u64
num_bytes,
u64 min_alloc_size, u64 empty_size, u64 hint_byte,
struct btrfs_key *ins, int is_data);
int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
- struct extent_buffer *buf, int full_backref, int for_cow);
+ struct extent_buffer *buf, int full_backref, int no_quota);
int btrfs_dec_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
- struct extent_buffer *buf, int full_backref, int for_cow);
+ struct extent_buffer *buf, int full_backref, int no_quota);
int btrfs_set_disk_extent_flags(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 bytenr, u64 num_bytes, u64 flags,
@@ -3291,7 +3294,7 @@ int btrfs_set_disk_extent_flags(struct btrfs_trans_handle
*trans,
int btrfs_free_extent(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 bytenr, u64 num_bytes, u64 parent, u64 root_objectid,
- u64 owner, u64 offset, int for_cow);
+ u64 owner, u64 offset, int no_quota);
int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len);
int btrfs_free_and_pin_reserved_extent(struct btrfs_root *root,
@@ -3303,7 +3306,7 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle
*trans,
int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 bytenr, u64 num_bytes, u64 parent,
- u64 root_objectid, u64 owner, u64 offset, int for_cow);
+ u64 root_objectid, u64 owner, u64 offset, int no_quota);
int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans,
struct btrfs_root *root);
@@ -3391,7 +3394,6 @@ int btrfs_init_space_info(struct btrfs_fs_info *fs_info);
int btrfs_delayed_refs_qgroup_accounting(struct btrfs_trans_handle *trans,
struct btrfs_fs_info *fs_info);
int __get_raid_index(u64 flags);
-
int btrfs_start_nocow_write(struct btrfs_root *root);
void btrfs_end_nocow_write(struct btrfs_root *root);
/* ctree.c */
@@ -3568,7 +3570,6 @@ u64 btrfs_get_tree_mod_seq(struct btrfs_fs_info *fs_info,
struct seq_list *elem);
void btrfs_put_tree_mod_seq(struct btrfs_fs_info *fs_info,
struct seq_list *elem);
-u64 btrfs_tree_mod_seq_prev(u64 seq);
int btrfs_old_root_level(struct btrfs_root *root, u64 time_seq);
/* root-item.c */
@@ -4079,52 +4080,6 @@ void btrfs_reada_detach(void *handle);
int btree_readahead_hook(struct btrfs_root *root, struct extent_buffer *eb,
u64 start, int err);
-/* qgroup.c */
-struct qgroup_update {
- struct list_head list;
- struct btrfs_delayed_ref_node *node;
- struct btrfs_delayed_extent_op *extent_op;
-};
-
-int btrfs_quota_enable(struct btrfs_trans_handle *trans,
- struct btrfs_fs_info *fs_info);
-int btrfs_quota_disable(struct btrfs_trans_handle *trans,
- struct btrfs_fs_info *fs_info);
-int btrfs_qgroup_rescan(struct btrfs_fs_info *fs_info);
-void btrfs_qgroup_rescan_resume(struct btrfs_fs_info *fs_info);
-int btrfs_qgroup_wait_for_completion(struct btrfs_fs_info *fs_info);
-int btrfs_add_qgroup_relation(struct btrfs_trans_handle *trans,
- struct btrfs_fs_info *fs_info, u64 src, u64 dst);
-int btrfs_del_qgroup_relation(struct btrfs_trans_handle *trans,
- struct btrfs_fs_info *fs_info, u64 src, u64 dst);
-int btrfs_create_qgroup(struct btrfs_trans_handle *trans,
- struct btrfs_fs_info *fs_info, u64 qgroupid,
- char *name);
-int btrfs_remove_qgroup(struct btrfs_trans_handle *trans,
- struct btrfs_fs_info *fs_info, u64 qgroupid);
-int btrfs_limit_qgroup(struct btrfs_trans_handle *trans,
- struct btrfs_fs_info *fs_info, u64 qgroupid,
- struct btrfs_qgroup_limit *limit);
-int btrfs_read_qgroup_config(struct btrfs_fs_info *fs_info);
-void btrfs_free_qgroup_config(struct btrfs_fs_info *fs_info);
-struct btrfs_delayed_extent_op;
-int btrfs_qgroup_record_ref(struct btrfs_trans_handle *trans,
- struct btrfs_delayed_ref_node *node,
- struct btrfs_delayed_extent_op *extent_op);
-int btrfs_qgroup_account_ref(struct btrfs_trans_handle *trans,
- struct btrfs_fs_info *fs_info,
- struct btrfs_delayed_ref_node *node,
- struct btrfs_delayed_extent_op *extent_op);
-int btrfs_run_qgroups(struct btrfs_trans_handle *trans,
- struct btrfs_fs_info *fs_info);
-int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans,
- struct btrfs_fs_info *fs_info, u64 srcid, u64 objectid,
- struct btrfs_qgroup_inherit *inherit);
-int btrfs_qgroup_reserve(struct btrfs_root *root, u64 num_bytes);
-void btrfs_qgroup_free(struct btrfs_root *root, u64 num_bytes);
-
-void assert_qgroups_uptodate(struct btrfs_trans_handle *trans);
-
static inline int is_fstree(u64 rootid)
{
if (rootid == BTRFS_FS_TREE_OBJECTID ||
diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c
index 3129964..6d16bea 100644
--- a/fs/btrfs/delayed-ref.c
+++ b/fs/btrfs/delayed-ref.c
@@ -106,6 +106,10 @@ static int comp_entry(struct btrfs_delayed_ref_node *ref2,
return -1;
if (ref1->type > ref2->type)
return 1;
+ if (ref1->no_quota > ref2->no_quota)
+ return 1;
+ if (ref1->no_quota < ref2->no_quota)
+ return -1;
/* merging of sequenced refs is not allowed */
if (compare_seq) {
if (ref1->seq < ref2->seq)
@@ -635,7 +639,7 @@ add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
struct btrfs_delayed_ref_head *head_ref,
struct btrfs_delayed_ref_node *ref, u64 bytenr,
u64 num_bytes, u64 parent, u64 ref_root, int level,
- int action, int for_cow)
+ int action, int no_quota)
{
struct btrfs_delayed_ref_node *existing;
struct btrfs_delayed_tree_ref *full_ref;
@@ -645,6 +649,8 @@ add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
if (action == BTRFS_ADD_DELAYED_EXTENT)
action = BTRFS_ADD_DELAYED_REF;
+ if (is_fstree(ref_root))
+ seq = atomic64_read(&fs_info->tree_mod_seq);
delayed_refs = &trans->transaction->delayed_refs;
/* first set the basic ref node struct up */
@@ -655,9 +661,7 @@ add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
ref->action = action;
ref->is_head = 0;
ref->in_tree = 1;
-
- if (need_ref_seq(for_cow, ref_root))
- seq = btrfs_get_tree_mod_seq(fs_info, &trans->delayed_ref_elem);
+ ref->no_quota = no_quota;
ref->seq = seq;
full_ref = btrfs_delayed_node_to_tree_ref(ref);
@@ -697,7 +701,7 @@ add_delayed_data_ref(struct btrfs_fs_info *fs_info,
struct btrfs_delayed_ref_head *head_ref,
struct btrfs_delayed_ref_node *ref, u64 bytenr,
u64 num_bytes, u64 parent, u64 ref_root, u64 owner,
- u64 offset, int action, int for_cow)
+ u64 offset, int action, int no_quota)
{
struct btrfs_delayed_ref_node *existing;
struct btrfs_delayed_data_ref *full_ref;
@@ -709,6 +713,9 @@ add_delayed_data_ref(struct btrfs_fs_info *fs_info,
delayed_refs = &trans->transaction->delayed_refs;
+ if (is_fstree(ref_root))
+ seq = atomic64_read(&fs_info->tree_mod_seq);
+
/* first set the basic ref node struct up */
atomic_set(&ref->refs, 1);
ref->bytenr = bytenr;
@@ -717,9 +724,7 @@ add_delayed_data_ref(struct btrfs_fs_info *fs_info,
ref->action = action;
ref->is_head = 0;
ref->in_tree = 1;
-
- if (need_ref_seq(for_cow, ref_root))
- seq = btrfs_get_tree_mod_seq(fs_info, &trans->delayed_ref_elem);
+ ref->no_quota = no_quota;
ref->seq = seq;
full_ref = btrfs_delayed_node_to_data_ref(ref);
@@ -762,12 +767,15 @@ int btrfs_add_delayed_tree_ref(struct btrfs_fs_info
*fs_info,
u64 bytenr, u64 num_bytes, u64 parent,
u64 ref_root, int level, int action,
struct btrfs_delayed_extent_op *extent_op,
- int for_cow)
+ int no_quota)
{
struct btrfs_delayed_tree_ref *ref;
struct btrfs_delayed_ref_head *head_ref;
struct btrfs_delayed_ref_root *delayed_refs;
+ if (!is_fstree(ref_root) || !fs_info->quota_enabled)
+ no_quota = 0;
+
BUG_ON(extent_op && extent_op->is_data);
ref = kmem_cache_alloc(btrfs_delayed_tree_ref_cachep, GFP_NOFS);
if (!ref)
@@ -793,10 +801,8 @@ int btrfs_add_delayed_tree_ref(struct btrfs_fs_info
*fs_info,
add_delayed_tree_ref(fs_info, trans, head_ref, &ref->node, bytenr,
num_bytes, parent, ref_root, level, action,
- for_cow);
+ no_quota);
spin_unlock(&delayed_refs->lock);
- if (need_ref_seq(for_cow, ref_root))
- btrfs_qgroup_record_ref(trans, &ref->node, extent_op);
return 0;
}
@@ -810,12 +816,15 @@ int btrfs_add_delayed_data_ref(struct btrfs_fs_info
*fs_info,
u64 parent, u64 ref_root,
u64 owner, u64 offset, int action,
struct btrfs_delayed_extent_op *extent_op,
- int for_cow)
+ int no_quota)
{
struct btrfs_delayed_data_ref *ref;
struct btrfs_delayed_ref_head *head_ref;
struct btrfs_delayed_ref_root *delayed_refs;
+ if (!is_fstree(ref_root) || !fs_info->quota_enabled)
+ no_quota = 0;
+
BUG_ON(extent_op && !extent_op->is_data);
ref = kmem_cache_alloc(btrfs_delayed_data_ref_cachep, GFP_NOFS);
if (!ref)
@@ -841,10 +850,8 @@ int btrfs_add_delayed_data_ref(struct btrfs_fs_info
*fs_info,
add_delayed_data_ref(fs_info, trans, head_ref, &ref->node, bytenr,
num_bytes, parent, ref_root, owner, offset,
- action, for_cow);
+ action, no_quota);
spin_unlock(&delayed_refs->lock);
- if (need_ref_seq(for_cow, ref_root))
- btrfs_qgroup_record_ref(trans, &ref->node, extent_op);
return 0;
}
diff --git a/fs/btrfs/delayed-ref.h b/fs/btrfs/delayed-ref.h
index 4ba9b93..a764e23 100644
--- a/fs/btrfs/delayed-ref.h
+++ b/fs/btrfs/delayed-ref.h
@@ -52,6 +52,7 @@ struct btrfs_delayed_ref_node {
unsigned int action:8;
unsigned int type:8;
+ unsigned int no_quota:1;
/* is this node still in the rbtree? */
unsigned int is_head:1;
unsigned int in_tree:1;
@@ -196,14 +197,14 @@ int btrfs_add_delayed_tree_ref(struct btrfs_fs_info
*fs_info,
u64 bytenr, u64 num_bytes, u64 parent,
u64 ref_root, int level, int action,
struct btrfs_delayed_extent_op *extent_op,
- int for_cow);
+ int no_quota);
int btrfs_add_delayed_data_ref(struct btrfs_fs_info *fs_info,
struct btrfs_trans_handle *trans,
u64 bytenr, u64 num_bytes,
u64 parent, u64 ref_root,
u64 owner, u64 offset, int action,
struct btrfs_delayed_extent_op *extent_op,
- int for_cow);
+ int no_quota);
int btrfs_add_delayed_extent_op(struct btrfs_fs_info *fs_info,
struct btrfs_trans_handle *trans,
u64 bytenr, u64 num_bytes,
@@ -231,25 +232,6 @@ int btrfs_check_delayed_seq(struct btrfs_fs_info *fs_info,
u64 seq);
/*
- * delayed refs with a ref_seq > 0 must be held back during backref walking.
- * this only applies to items in one of the fs-trees. for_cow items never need
- * to be held back, so they won't get a ref_seq number.
- */
-static inline int need_ref_seq(int for_cow, u64 rootid)
-{
- if (for_cow)
- return 0;
-
- if (rootid == BTRFS_FS_TREE_OBJECTID)
- return 1;
-
- if ((s64)rootid >= (s64)BTRFS_FIRST_FREE_OBJECTID)
- return 1;
-
- return 0;
-}
-
-/*
* a node might live in a head or a regular ref, this lets you
* test for the proper type to use.
*/
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 4fc058a..d965f51 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -50,6 +50,7 @@
#include "raid56.h"
#include "sysfs.h"
#include "ref-verify.h"
+#include "qgroup.h"
#ifdef CONFIG_X86
#include <asm/cpufeature.h>
@@ -2225,6 +2226,7 @@ int open_ctree(struct super_block *sb,
spin_lock_init(&fs_info->free_chunk_lock);
spin_lock_init(&fs_info->tree_mod_seq_lock);
spin_lock_init(&fs_info->super_lock);
+ spin_lock_init(&fs_info->qgroup_op_lock);
spin_lock_init(&fs_info->buffer_lock);
rwlock_init(&fs_info->tree_mod_log_lock);
mutex_init(&fs_info->reloc_mutex);
@@ -2250,6 +2252,7 @@ int open_ctree(struct super_block *sb,
atomic_set(&fs_info->async_submit_draining, 0);
atomic_set(&fs_info->nr_async_bios, 0);
atomic_set(&fs_info->defrag_running, 0);
+ atomic_set(&fs_info->qgroup_op_seq, 0);
atomic64_set(&fs_info->tree_mod_seq, 0);
fs_info->sb = sb;
fs_info->max_inline = 8192 * 1024;
@@ -2363,6 +2366,7 @@ int open_ctree(struct super_block *sb,
spin_lock_init(&fs_info->qgroup_lock);
mutex_init(&fs_info->qgroup_ioctl_lock);
fs_info->qgroup_tree = RB_ROOT;
+ fs_info->qgroup_op_tree = RB_ROOT;
INIT_LIST_HEAD(&fs_info->dirty_qgroups);
fs_info->qgroup_seq = 1;
fs_info->quota_enabled = 0;
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 27fc928..887bef2 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -37,6 +37,7 @@
#include "math.h"
#include "sysfs.h"
#include "ref-verify.h"
+#include "qgroup.h"
#undef SCRAMBLE_DELAYED_REFS
@@ -82,7 +83,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle
*trans,
u64 bytenr, u64 num_bytes, u64 parent,
u64 root_objectid, u64 owner_objectid,
u64 owner_offset, int refs_to_drop,
- struct btrfs_delayed_extent_op *extra_op);
+ struct btrfs_delayed_extent_op *extra_op,
+ int no_quota);
static void __run_delayed_extent_op(struct btrfs_delayed_extent_op *extent_op,
struct extent_buffer *leaf,
struct btrfs_extent_item *ei);
@@ -95,7 +97,8 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle
*trans,
struct btrfs_root *root,
u64 parent, u64 root_objectid,
u64 flags, struct btrfs_disk_key *key,
- int level, struct btrfs_key *ins);
+ int level, struct btrfs_key *ins,
+ int no_quota);
static int do_chunk_alloc(struct btrfs_trans_handle *trans,
struct btrfs_root *extent_root, u64 flags,
int force);
@@ -1272,7 +1275,7 @@ fail:
static noinline int remove_extent_data_ref(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_path *path,
- int refs_to_drop)
+ int refs_to_drop, int *last_ref)
{
struct btrfs_key key;
struct btrfs_extent_data_ref *ref1 = NULL;
@@ -1308,6 +1311,7 @@ static noinline int remove_extent_data_ref(struct
btrfs_trans_handle *trans,
if (num_refs == 0) {
ret = btrfs_del_item(trans, root, path);
+ *last_ref = 1;
} else {
if (key.type == BTRFS_EXTENT_DATA_REF_KEY)
btrfs_set_extent_data_ref_count(leaf, ref1, num_refs);
@@ -1765,7 +1769,8 @@ void update_inline_extent_backref(struct btrfs_root *root,
struct btrfs_path *path,
struct btrfs_extent_inline_ref *iref,
int refs_to_mod,
- struct btrfs_delayed_extent_op *extent_op)
+ struct btrfs_delayed_extent_op *extent_op,
+ int *last_ref)
{
struct extent_buffer *leaf;
struct btrfs_extent_item *ei;
@@ -1809,6 +1814,7 @@ void update_inline_extent_backref(struct btrfs_root *root,
else
btrfs_set_shared_data_ref_count(leaf, sref, refs);
} else {
+ *last_ref = 1;
size = btrfs_extent_inline_ref_size(type);
item_size = btrfs_item_size_nr(leaf, path->slots[0]);
ptr = (unsigned long)iref;
@@ -1840,7 +1846,7 @@ int insert_inline_extent_backref(struct btrfs_trans_handle
*trans,
if (ret == 0) {
BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID);
update_inline_extent_backref(root, path, iref,
- refs_to_add, extent_op);
+ refs_to_add, extent_op, NULL);
} else if (ret == -ENOENT) {
setup_inline_extent_backref(root, path, iref, parent,
root_objectid, owner, offset,
@@ -1873,17 +1879,19 @@ static int remove_extent_backref(struct
btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_path *path,
struct btrfs_extent_inline_ref *iref,
- int refs_to_drop, int is_data)
+ int refs_to_drop, int is_data, int *last_ref)
{
int ret = 0;
BUG_ON(!is_data && refs_to_drop != 1);
if (iref) {
update_inline_extent_backref(root, path, iref,
- -refs_to_drop, NULL);
+ -refs_to_drop, NULL, last_ref);
} else if (is_data) {
- ret = remove_extent_data_ref(trans, root, path, refs_to_drop);
+ ret = remove_extent_data_ref(trans, root, path, refs_to_drop,
+ last_ref);
} else {
+ *last_ref = 1;
ret = btrfs_del_item(trans, root, path);
}
return ret;
@@ -1947,7 +1955,8 @@ static int btrfs_discard_extent(struct btrfs_root *root,
u64 bytenr,
int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 bytenr, u64 num_bytes, u64 parent,
- u64 root_objectid, u64 owner, u64 offset, int for_cow)
+ u64 root_objectid, u64 owner, u64 offset,
+ int no_quota)
{
int ret;
struct btrfs_fs_info *fs_info = root->fs_info;
@@ -1962,12 +1971,12 @@ int btrfs_inc_extent_ref(struct btrfs_trans_handle
*trans,
ret = btrfs_add_delayed_tree_ref(fs_info, trans, bytenr,
num_bytes,
parent, root_objectid, (int)owner,
- BTRFS_ADD_DELAYED_REF, NULL, for_cow);
+ BTRFS_ADD_DELAYED_REF, NULL, no_quota);
} else {
ret = btrfs_add_delayed_data_ref(fs_info, trans, bytenr,
num_bytes,
parent, root_objectid, owner, offset,
- BTRFS_ADD_DELAYED_REF, NULL, for_cow);
+ BTRFS_ADD_DELAYED_REF, NULL, no_quota);
}
return ret;
}
@@ -1977,31 +1986,64 @@ static int __btrfs_inc_extent_ref(struct
btrfs_trans_handle *trans,
u64 bytenr, u64 num_bytes,
u64 parent, u64 root_objectid,
u64 owner, u64 offset, int refs_to_add,
+ int no_quota,
struct btrfs_delayed_extent_op *extent_op)
{
+ struct btrfs_fs_info *fs_info = root->fs_info;
struct btrfs_path *path;
struct extent_buffer *leaf;
struct btrfs_extent_item *item;
+ struct btrfs_key key;
u64 refs;
int ret;
+ enum btrfs_qgroup_operation_type type = BTRFS_QGROUP_OPER_ADD_EXCL;
path = btrfs_alloc_path();
if (!path)
return -ENOMEM;
+ if (!is_fstree(root_objectid) || !root->fs_info->quota_enabled)
+ no_quota = 1;
+
path->reada = 1;
path->leave_spinning = 1;
/* this will setup the path even if it fails to insert the back ref */
- ret = insert_inline_extent_backref(trans, root->fs_info->extent_root,
- path, bytenr, num_bytes, parent,
+ ret = insert_inline_extent_backref(trans, fs_info->extent_root, path,
+ bytenr, num_bytes, parent,
root_objectid, owner, offset,
refs_to_add, extent_op);
- if (ret != -EAGAIN)
+ if ((ret < 0 && ret != -EAGAIN) || (!ret && no_quota))
goto out;
+ /*
+ * Ok we were able to insert an inline extent and it appears to be a new
+ * reference, deal with the qgroup accounting.
+ */
+ if (!ret && !no_quota) {
+ ASSERT(root->fs_info->quota_enabled);
+ leaf = path->nodes[0];
+ btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+ item = btrfs_item_ptr(leaf, path->slots[0],
+ struct btrfs_extent_item);
+ if (btrfs_extent_refs(leaf, item) > (u64)refs_to_add)
+ type = BTRFS_QGROUP_OPER_ADD_SHARED;
+ btrfs_release_path(path);
+ ret = btrfs_qgroup_record_ref(trans, fs_info, root_objectid,
+ bytenr, num_bytes, type, 0);
+ goto out;
+ }
+
+ /*
+ * Ok we had -EAGAIN which means we didn't have space to insert and
+ * inline extent ref, so just update the reference count and add a
+ * normal backref.
+ */
leaf = path->nodes[0];
+ btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
refs = btrfs_extent_refs(leaf, item);
+ if (refs)
+ type = BTRFS_QGROUP_OPER_ADD_SHARED;
btrfs_set_extent_refs(leaf, item, refs + refs_to_add);
if (extent_op)
__run_delayed_extent_op(extent_op, leaf, item);
@@ -2009,9 +2051,15 @@ static int __btrfs_inc_extent_ref(struct
btrfs_trans_handle *trans,
btrfs_mark_buffer_dirty(leaf);
btrfs_release_path(path);
+ if (!no_quota) {
+ ret = btrfs_qgroup_record_ref(trans, fs_info, root_objectid,
+ bytenr, num_bytes, type, 0);
+ if (ret)
+ goto out;
+ }
+
path->reada = 1;
path->leave_spinning = 1;
-
/* now insert the actual backref */
ret = insert_extent_backref(trans, root->fs_info->extent_root,
path, bytenr, parent, root_objectid,
@@ -2045,8 +2093,7 @@ static int run_delayed_data_ref(struct btrfs_trans_handle
*trans,
if (node->type == BTRFS_SHARED_DATA_REF_KEY)
parent = ref->parent;
- else
- ref_root = ref->root;
+ ref_root = ref->root;
if (node->action == BTRFS_ADD_DELAYED_REF && insert_reserved) {
if (extent_op)
@@ -2060,13 +2107,13 @@ static int run_delayed_data_ref(struct
btrfs_trans_handle *trans,
node->num_bytes, parent,
ref_root, ref->objectid,
ref->offset, node->ref_mod,
- extent_op);
+ node->no_quota, extent_op);
} else if (node->action == BTRFS_DROP_DELAYED_REF) {
ret = __btrfs_free_extent(trans, root, node->bytenr,
node->num_bytes, parent,
ref_root, ref->objectid,
ref->offset, node->ref_mod,
- extent_op);
+ extent_op, node->no_quota);
} else {
BUG();
}
@@ -2203,8 +2250,7 @@ static int run_delayed_tree_ref(struct btrfs_trans_handle
*trans,
if (node->type == BTRFS_SHARED_BLOCK_REF_KEY)
parent = ref->parent;
- else
- ref_root = ref->root;
+ ref_root = ref->root;
ins.objectid = node->bytenr;
if (skinny_metadata) {
@@ -2222,15 +2268,18 @@ static int run_delayed_tree_ref(struct
btrfs_trans_handle *trans,
parent, ref_root,
extent_op->flags_to_set,
&extent_op->key,
- ref->level, &ins);
+ ref->level, &ins,
+ node->no_quota);
} else if (node->action == BTRFS_ADD_DELAYED_REF) {
ret = __btrfs_inc_extent_ref(trans, root, node->bytenr,
node->num_bytes, parent, ref_root,
- ref->level, 0, 1, extent_op);
+ ref->level, 0, 1, node->no_quota,
+ extent_op);
} else if (node->action == BTRFS_DROP_DELAYED_REF) {
ret = __btrfs_free_extent(trans, root, node->bytenr,
node->num_bytes, parent, ref_root,
- ref->level, 0, 1, extent_op);
+ ref->level, 0, 1, extent_op,
+ node->no_quota);
} else {
BUG();
}
@@ -2578,42 +2627,6 @@ static u64 find_middle(struct rb_root *root)
}
#endif
-int btrfs_delayed_refs_qgroup_accounting(struct btrfs_trans_handle *trans,
- struct btrfs_fs_info *fs_info)
-{
- struct qgroup_update *qgroup_update;
- int ret = 0;
-
- if (list_empty(&trans->qgroup_ref_list) !-
!trans->delayed_ref_elem.seq) {
- /* list without seq or seq without list */
- btrfs_err(fs_info,
- "qgroup accounting update error, list is%s empty, seq is %#x.%x",
- list_empty(&trans->qgroup_ref_list) ? "" : "
not",
- (u32)(trans->delayed_ref_elem.seq >> 32),
- (u32)trans->delayed_ref_elem.seq);
- BUG();
- }
-
- if (!trans->delayed_ref_elem.seq)
- return 0;
-
- while (!list_empty(&trans->qgroup_ref_list)) {
- qgroup_update = list_first_entry(&trans->qgroup_ref_list,
- struct qgroup_update, list);
- list_del(&qgroup_update->list);
- if (!ret)
- ret = btrfs_qgroup_account_ref(
- trans, fs_info, qgroup_update->node,
- qgroup_update->extent_op);
- kfree(qgroup_update);
- }
-
- btrfs_put_tree_mod_seq(fs_info, &trans->delayed_ref_elem);
-
- return ret;
-}
-
static inline u64 heads_to_leaves(struct btrfs_root *root, u64 heads)
{
u64 num_bytes;
@@ -2702,8 +2715,6 @@ int btrfs_run_delayed_refs(struct btrfs_trans_handle
*trans,
if (root == root->fs_info->extent_root)
root = root->fs_info->tree_root;
- btrfs_delayed_refs_qgroup_accounting(trans, root->fs_info);
-
delayed_refs = &trans->transaction->delayed_refs;
if (count == 0) {
count = atomic_read(&delayed_refs->num_entries) * 2;
@@ -2762,6 +2773,9 @@ again:
goto again;
}
out:
+ ret = btrfs_delayed_qgroup_accounting(trans, root->fs_info);
+ if (ret)
+ return ret;
assert_qgroups_uptodate(trans);
return 0;
}
@@ -2968,7 +2982,7 @@ out:
static int __btrfs_mod_ref(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct extent_buffer *buf,
- int full_backref, int inc, int for_cow)
+ int full_backref, int inc, int no_quota)
{
u64 bytenr;
u64 num_bytes;
@@ -3018,7 +3032,7 @@ static int __btrfs_mod_ref(struct btrfs_trans_handle
*trans,
key.offset -= btrfs_file_extent_offset(buf, fi);
ret = process_func(trans, root, bytenr, num_bytes,
parent, ref_root, key.objectid,
- key.offset, for_cow);
+ key.offset, no_quota);
if (ret)
goto fail;
} else {
@@ -3026,7 +3040,7 @@ static int __btrfs_mod_ref(struct btrfs_trans_handle
*trans,
num_bytes = btrfs_level_size(root, level - 1);
ret = process_func(trans, root, bytenr, num_bytes,
parent, ref_root, level - 1, 0,
- for_cow);
+ no_quota);
if (ret)
goto fail;
}
@@ -3037,15 +3051,15 @@ fail:
}
int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
- struct extent_buffer *buf, int full_backref, int for_cow)
+ struct extent_buffer *buf, int full_backref, int no_quota)
{
- return __btrfs_mod_ref(trans, root, buf, full_backref, 1, for_cow);
+ return __btrfs_mod_ref(trans, root, buf, full_backref, 1, no_quota);
}
int btrfs_dec_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
- struct extent_buffer *buf, int full_backref, int for_cow)
+ struct extent_buffer *buf, int full_backref, int no_quota)
{
- return __btrfs_mod_ref(trans, root, buf, full_backref, 0, for_cow);
+ return __btrfs_mod_ref(trans, root, buf, full_backref, 0, no_quota);
}
static int write_one_cache_group(struct btrfs_trans_handle *trans,
@@ -5625,7 +5639,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle
*trans,
u64 bytenr, u64 num_bytes, u64 parent,
u64 root_objectid, u64 owner_objectid,
u64 owner_offset, int refs_to_drop,
- struct btrfs_delayed_extent_op *extent_op)
+ struct btrfs_delayed_extent_op *extent_op,
+ int no_quota)
{
struct btrfs_key key;
struct btrfs_path *path;
@@ -5641,9 +5656,14 @@ static int __btrfs_free_extent(struct btrfs_trans_handle
*trans,
int num_to_del = 1;
u32 item_size;
u64 refs;
+ int last_ref = 0;
+ enum btrfs_qgroup_operation_type type = BTRFS_QGROUP_OPER_SUB_EXCL;
bool skinny_metadata = btrfs_fs_incompat(root->fs_info,
SKINNY_METADATA);
+ if (!info->quota_enabled || !is_fstree(root_objectid))
+ no_quota = 1;
+
path = btrfs_alloc_path();
if (!path)
return -ENOMEM;
@@ -5691,7 +5711,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle
*trans,
BUG_ON(iref);
ret = remove_extent_backref(trans, extent_root, path,
NULL, refs_to_drop,
- is_data);
+ is_data, &last_ref);
if (ret) {
btrfs_abort_transaction(trans, extent_root, ret);
goto out;
@@ -5818,6 +5838,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle
*trans,
refs -= refs_to_drop;
if (refs > 0) {
+ type = BTRFS_QGROUP_OPER_SUB_SHARED;
if (extent_op)
__run_delayed_extent_op(extent_op, leaf, ei);
/*
@@ -5833,7 +5854,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle
*trans,
if (found_extent) {
ret = remove_extent_backref(trans, extent_root, path,
iref, refs_to_drop,
- is_data);
+ is_data, &last_ref);
if (ret) {
btrfs_abort_transaction(trans, extent_root, ret);
goto out;
@@ -5854,6 +5875,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle
*trans,
}
}
+ last_ref = 1;
ret = btrfs_del_items(trans, extent_root, path, path->slots[0],
num_to_del);
if (ret) {
@@ -5876,6 +5898,20 @@ static int __btrfs_free_extent(struct btrfs_trans_handle
*trans,
goto out;
}
}
+ btrfs_release_path(path);
+
+ /* Deal with the quota accounting */
+ if (!ret && last_ref && !no_quota) {
+ int mod_seq = 0;
+
+ if (owner_objectid >= BTRFS_FIRST_FREE_OBJECTID &&
+ type == BTRFS_QGROUP_OPER_SUB_SHARED)
+ mod_seq = 1;
+
+ ret = btrfs_qgroup_record_ref(trans, info, root_objectid,
+ bytenr, num_bytes, type,
+ mod_seq);
+ }
out:
btrfs_free_path(path);
return ret;
@@ -6015,7 +6051,7 @@ out:
/* Can return -ENOMEM */
int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root
*root,
u64 bytenr, u64 num_bytes, u64 parent, u64 root_objectid,
- u64 owner, u64 offset, int for_cow)
+ u64 owner, u64 offset, int no_quota)
{
int ret;
struct btrfs_fs_info *fs_info = root->fs_info;
@@ -6038,13 +6074,13 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
ret = btrfs_add_delayed_tree_ref(fs_info, trans, bytenr,
num_bytes,
parent, root_objectid, (int)owner,
- BTRFS_DROP_DELAYED_REF, NULL, for_cow);
+ BTRFS_DROP_DELAYED_REF, NULL, no_quota);
} else {
ret = btrfs_add_delayed_data_ref(fs_info, trans, bytenr,
num_bytes,
parent, root_objectid, owner,
offset, BTRFS_DROP_DELAYED_REF,
- NULL, for_cow);
+ NULL, no_quota);
}
return ret;
}
@@ -6743,6 +6779,13 @@ static int alloc_reserved_file_extent(struct
btrfs_trans_handle *trans,
btrfs_mark_buffer_dirty(path->nodes[0]);
btrfs_free_path(path);
+ /* Always set parent to 0 here since its exclusive anyway. */
+ ret = btrfs_qgroup_record_ref(trans, fs_info, root_objectid,
+ ins->objectid, ins->offset,
+ BTRFS_QGROUP_OPER_ADD_EXCL, 0);
+ if (ret)
+ return ret;
+
ret = update_block_group(root, ins->objectid, ins->offset, 1);
if (ret) { /* -ENOENT, logic error */
btrfs_err(fs_info, "update block group failed for %llu %llu",
@@ -6757,7 +6800,8 @@ static int alloc_reserved_tree_block(struct
btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 parent, u64 root_objectid,
u64 flags, struct btrfs_disk_key *key,
- int level, struct btrfs_key *ins)
+ int level, struct btrfs_key *ins,
+ int no_quota)
{
int ret;
struct btrfs_fs_info *fs_info = root->fs_info;
@@ -6767,6 +6811,7 @@ static int alloc_reserved_tree_block(struct
btrfs_trans_handle *trans,
struct btrfs_path *path;
struct extent_buffer *leaf;
u32 size = sizeof(*extent_item) + sizeof(*iref);
+ u64 num_bytes = ins->offset;
bool skinny_metadata = btrfs_fs_incompat(root->fs_info,
SKINNY_METADATA);
@@ -6800,6 +6845,7 @@ static int alloc_reserved_tree_block(struct
btrfs_trans_handle *trans,
if (skinny_metadata) {
iref = (struct btrfs_extent_inline_ref *)(extent_item + 1);
+ num_bytes = root->leafsize;
} else {
block_info = (struct btrfs_tree_block_info *)(extent_item + 1);
btrfs_set_tree_block_key(leaf, block_info, key);
@@ -6821,6 +6867,14 @@ static int alloc_reserved_tree_block(struct
btrfs_trans_handle *trans,
btrfs_mark_buffer_dirty(leaf);
btrfs_free_path(path);
+ if (!no_quota) {
+ ret = btrfs_qgroup_record_ref(trans, fs_info, root_objectid,
+ ins->objectid, num_bytes,
+ BTRFS_QGROUP_OPER_ADD_EXCL, 0);
+ if (ret)
+ return ret;
+ }
+
ret = update_block_group(root, ins->objectid, root->leafsize, 1);
if (ret) { /* -ENOENT, logic error */
btrfs_err(fs_info, "update block group failed for %llu %llu",
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 1eee3f7..9a83346 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -40,6 +40,7 @@
#include "tree-log.h"
#include "locking.h"
#include "volumes.h"
+#include "qgroup.h"
static struct kmem_cache *btrfs_inode_defrag_cachep;
/*
@@ -840,7 +841,7 @@ next_slot:
disk_bytenr, num_bytes, 0,
root->root_key.objectid,
new_key.objectid,
- start - extent_offset, 0);
+ start - extent_offset, 1);
BUG_ON(ret); /* -ENOMEM */
}
key.offset = start;
@@ -1196,7 +1197,7 @@ again:
ret = btrfs_inc_extent_ref(trans, root, bytenr, num_bytes, 0,
root->root_key.objectid,
- ino, orig_offset, 0);
+ ino, orig_offset, 1);
BUG_ON(ret); /* -ENOMEM */
if (split == start) {
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 7b001ab..df3020b 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -58,6 +58,7 @@
#include "dev-replace.h"
#include "props.h"
#include "sysfs.h"
+#include "qgroup.h"
#ifdef CONFIG_64BIT
/* If we have a 32-bit userspace and 64-bit kernel, then the UAPI
@@ -2934,6 +2935,41 @@ out:
return ret;
}
+/* Helper to check and see if this root currently has a ref on the given disk
+ * bytenr. If it does then we need to update the quota for this root. This
+ * doesn't do anything if quotas aren't enabled.
+ */
+static int check_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+ u64 disko)
+{
+ struct seq_list tree_mod_seq_elem = {};
+ struct ulist *roots;
+ struct ulist_iterator uiter;
+ struct ulist_node *root_node = NULL;
+ int ret;
+
+ if (!root->fs_info->quota_enabled)
+ return 1;
+
+ btrfs_get_tree_mod_seq(root->fs_info, &tree_mod_seq_elem);
+ ret = btrfs_find_all_roots(trans, root->fs_info, disko,
+ tree_mod_seq_elem.seq, &roots);
+ if (ret < 0)
+ goto out;
+ ret = 0;
+ ULIST_ITER_INIT(&uiter);
+ while ((root_node = ulist_next(roots, &uiter))) {
+ if (root_node->val == root->objectid) {
+ ret = 1;
+ break;
+ }
+ }
+ ulist_free(roots);
+out:
+ btrfs_put_tree_mod_seq(root->fs_info, &tree_mod_seq_elem);
+ return ret;
+}
+
/**
* btrfs_clone() - clone a range from inode file to another
*
@@ -2957,7 +2993,9 @@ static int btrfs_clone(struct inode *src, struct inode
*inode,
u32 nritems;
int slot;
int ret;
+ int no_quota;
u64 len = olen_aligned;
+ u64 last_disko = 0;
ret = -ENOMEM;
buf = vmalloc(btrfs_level_size(root, 0));
@@ -2989,6 +3027,7 @@ static int btrfs_clone(struct inode *src, struct inode
*inode,
nritems = btrfs_header_nritems(path->nodes[0]);
process_slot:
+ no_quota = 1;
if (path->slots[0] >= nritems) {
ret = btrfs_next_leaf(BTRFS_I(src)->root, path);
if (ret < 0)
@@ -3121,6 +3160,28 @@ process_slot:
datao);
btrfs_set_file_extent_num_bytes(leaf, extent,
datal);
+
+ /*
+ * We need to look up the roots that point at
+ * this bytenr and see if the new root does. If
+ * it does not we need to make sure we update
+ * quotas appropriately.
+ */
+ if (disko && root != BTRFS_I(src)->root &&
+ disko != last_disko) {
+ no_quota = check_ref(trans, root,
+ disko);
+ if (no_quota < 0) {
+ btrfs_abort_transaction(trans,
+ root,
+ ret);
+ btrfs_end_transaction(trans,
+ root);
+ ret = no_quota;
+ goto out;
+ }
+ }
+
if (disko) {
inode_add_bytes(inode, datal);
ret = btrfs_inc_extent_ref(trans, root,
@@ -3128,7 +3189,7 @@ process_slot:
root->root_key.objectid,
btrfs_ino(inode),
new_key.offset - datao,
- 0);
+ no_quota);
if (ret) {
btrfs_abort_transaction(trans,
root,
diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c
index 2cf9058..09b8cc8 100644
--- a/fs/btrfs/qgroup.c
+++ b/fs/btrfs/qgroup.c
@@ -32,6 +32,7 @@
#include "ulist.h"
#include "backref.h"
#include "extent_io.h"
+#include "qgroup.h"
/* TODO XXX FIXME
* - subvol delete -> delete when ref goes to 0? delete limits also?
@@ -84,8 +85,8 @@ struct btrfs_qgroup {
/*
* temp variables for accounting operations
*/
- u64 tag;
- u64 refcnt;
+ u64 old_refcnt;
+ u64 new_refcnt;
};
/*
@@ -98,6 +99,9 @@ struct btrfs_qgroup_list {
struct btrfs_qgroup *member;
};
+#define ptr_to_u64(x) ((u64)(uintptr_t)x)
+#define u64_to_ptr(x) ((struct btrfs_qgroup *)(uintptr_t)x)
+
static int
qgroup_rescan_init(struct btrfs_fs_info *fs_info, u64 progress_objectid,
int init_flags);
@@ -1174,33 +1178,198 @@ out:
mutex_unlock(&fs_info->qgroup_ioctl_lock);
return ret;
}
+static int comp_oper(struct btrfs_qgroup_operation *oper1,
+ struct btrfs_qgroup_operation *oper2)
+{
+ if (oper1->bytenr < oper2->bytenr)
+ return -1;
+ if (oper1->bytenr > oper2->bytenr)
+ return 1;
+ if (oper1->seq < oper2->seq)
+ return -1;
+ if (oper1->seq > oper2->seq)
+ return -1;
+ if (oper1->ref_root < oper2->ref_root)
+ return -1;
+ if (oper1->ref_root > oper2->ref_root)
+ return 1;
+ if (oper1->type < oper2->type)
+ return -1;
+ if (oper1->type > oper2->type)
+ return 1;
+ return 0;
+}
+
+static int insert_qgroup_oper(struct btrfs_fs_info *fs_info,
+ struct btrfs_qgroup_operation *oper)
+{
+ struct rb_node **p;
+ struct rb_node *parent = NULL;
+ struct btrfs_qgroup_operation *cur;
+ int cmp;
+
+ spin_lock(&fs_info->qgroup_op_lock);
+ p = &fs_info->qgroup_op_tree.rb_node;
+ while (*p) {
+ parent = *p;
+ cur = rb_entry(parent, struct btrfs_qgroup_operation, n);
+ cmp = comp_oper(cur, oper);
+ if (cmp < 0) {
+ p = &(*p)->rb_right;
+ } else if (cmp) {
+ p = &(*p)->rb_left;
+ } else {
+ spin_unlock(&fs_info->qgroup_op_lock);
+ return -EEXIST;
+ }
+ }
+ rb_link_node(&oper->n, parent, p);
+ rb_insert_color(&oper->n, &fs_info->qgroup_op_tree);
+ spin_unlock(&fs_info->qgroup_op_lock);
+ return 0;
+}
/*
- * btrfs_qgroup_record_ref is called when the ref is added or deleted. it puts
- * the modification into a list that's later used by btrfs_end_transaction
to
- * pass the recorded modifications on to btrfs_qgroup_account_ref.
+ * Record a quota operation for processing later on.
+ * @trans: the transaction we are adding the delayed op to.
+ * @fs_info: the fs_info for this fs.
+ * @ref_root: the root of the reference we are acting on,
+ * @bytenr: the bytenr we are acting on.
+ * @num_bytes: the number of bytes in the reference.
+ * @type: the type of operation this is.
+ * @mod_seq: do we need to get a sequence number for looking up roots.
+ *
+ * We just add it to our trans qgroup_ref_list and carry on and process these
+ * operations in order at some later point. If the reference root isn't a
fs
+ * root then we don't bother with doing anything.
+ *
+ * MUST BE HOLDING THE REF LOCK.
*/
int btrfs_qgroup_record_ref(struct btrfs_trans_handle *trans,
- struct btrfs_delayed_ref_node *node,
- struct btrfs_delayed_extent_op *extent_op)
+ struct btrfs_fs_info *fs_info, u64 ref_root,
+ u64 bytenr, u64 num_bytes,
+ enum btrfs_qgroup_operation_type type, int mod_seq)
{
- struct qgroup_update *u;
+ struct btrfs_qgroup_operation *oper;
+ int ret;
- BUG_ON(!trans->delayed_ref_elem.seq);
- u = kmalloc(sizeof(*u), GFP_NOFS);
- if (!u)
+ if (!is_fstree(ref_root) || !fs_info->quota_enabled)
+ return 0;
+
+ oper = kmalloc(sizeof(*oper), GFP_NOFS);
+ if (!oper)
return -ENOMEM;
- u->node = node;
- u->extent_op = extent_op;
- list_add_tail(&u->list, &trans->qgroup_ref_list);
+ oper->ref_root = ref_root;
+ oper->bytenr = bytenr;
+ oper->num_bytes = num_bytes;
+ oper->type = type;
+ oper->seq = atomic_inc_return(&fs_info->qgroup_op_seq);
+ INIT_LIST_HEAD(&oper->elem.list);
+ oper->elem.seq = 0;
+ ret = insert_qgroup_oper(fs_info, oper);
+ if (ret) {
+ /* Shouldn't happen so have an assert for developers */
+ ASSERT(0);
+ kfree(oper);
+ return ret;
+ }
+ list_add_tail(&oper->list, &trans->qgroup_ref_list);
+
+ if (mod_seq)
+ btrfs_get_tree_mod_seq(fs_info, &oper->elem);
return 0;
}
-static int qgroup_account_ref_step1(struct btrfs_fs_info *fs_info,
- struct ulist *roots, struct ulist *tmp,
- u64 seq)
+/*
+ * The easy accounting, if we are adding/removing the only ref for an extent
+ * then this qgroup and all of the parent qgroups get their refrence and
+ * exclusive counts adjusted.
+ */
+static int qgroup_excl_accounting(struct btrfs_fs_info *fs_info,
+ struct btrfs_qgroup_operation *oper)
+{
+ struct btrfs_qgroup *qgroup;
+ struct ulist *tmp;
+ struct btrfs_qgroup_list *glist;
+ struct ulist_node *unode;
+ struct ulist_iterator uiter;
+ int sign = 0;
+ int ret = 0;
+
+ tmp = ulist_alloc(GFP_NOFS);
+ if (!tmp)
+ return -ENOMEM;
+
+ spin_lock(&fs_info->qgroup_lock);
+ if (!fs_info->quota_root)
+ goto out;
+ qgroup = find_qgroup_rb(fs_info, oper->ref_root);
+ if (!qgroup)
+ goto out;
+ switch (oper->type) {
+ case BTRFS_QGROUP_OPER_ADD_EXCL:
+ sign = 1;
+ break;
+ case BTRFS_QGROUP_OPER_SUB_EXCL:
+ sign = -1;
+ break;
+ default:
+ ASSERT(0);
+ }
+ qgroup->rfer += sign * oper->num_bytes;
+ qgroup->rfer_cmpr += sign * oper->num_bytes;
+
+ WARN_ON(sign < 0 && qgroup->excl < oper->num_bytes);
+ qgroup->excl += sign * oper->num_bytes;
+ qgroup->excl_cmpr += sign * oper->num_bytes;
+
+ qgroup_dirty(fs_info, qgroup);
+
+ /* Get all of the parent groups that contain this qgroup */
+ list_for_each_entry(glist, &qgroup->groups, next_group) {
+ ret = ulist_add(tmp, glist->group->qgroupid,
+ ptr_to_u64(glist->group), GFP_ATOMIC);
+ if (ret < 0)
+ goto out;
+ }
+
+ /* Iterate all of the parents and adjust their reference counts */
+ ULIST_ITER_INIT(&uiter);
+ while ((unode = ulist_next(tmp, &uiter))) {
+ qgroup = u64_to_ptr(unode->aux);
+ qgroup->rfer += sign * oper->num_bytes;
+ qgroup->rfer_cmpr += sign * oper->num_bytes;
+ qgroup->excl += sign * oper->num_bytes;
+ if (sign < 0)
+ WARN_ON(qgroup->excl < oper->num_bytes);
+ qgroup->excl_cmpr += sign * oper->num_bytes;
+ qgroup_dirty(fs_info, qgroup);
+
+ /* Add any parents of the parents */
+ list_for_each_entry(glist, &qgroup->groups, next_group) {
+ ret = ulist_add(tmp, glist->group->qgroupid,
+ ptr_to_u64(glist->group), GFP_ATOMIC);
+ if (ret < 0)
+ goto out;
+ }
+ }
+ ret = 0;
+out:
+ spin_unlock(&fs_info->qgroup_lock);
+ ulist_free(tmp);
+ return ret;
+}
+
+/*
+ * Walk all of the roots that pointed to our bytenr and adjust their refcnts as
+ * properly.
+ */
+static int qgroup_calc_old_refcnt(struct btrfs_fs_info *fs_info,
+ u64 root_to_skip, struct ulist *tmp,
+ struct ulist *roots, struct ulist *qgroups,
+ u64 seq, int *old_roots, int rescan)
{
struct ulist_node *unode;
struct ulist_iterator uiter;
@@ -1211,256 +1380,549 @@ static int qgroup_account_ref_step1(struct
btrfs_fs_info *fs_info,
ULIST_ITER_INIT(&uiter);
while ((unode = ulist_next(roots, &uiter))) {
+ /* We don't count our current root here */
+ if (unode->val == root_to_skip)
+ continue;
qg = find_qgroup_rb(fs_info, unode->val);
if (!qg)
continue;
+ /*
+ * We could have a pending removal of this same ref so we may
+ * not have actually found our ref root when doing
+ * btrfs_find_all_roots, so we need to keep track of how many
+ * old roots we find in case we removed ours and added a
+ * different one at the same time. I don't think this could
+ * happen in practice but that sort of thinking leads to pain
+ * and suffering and to the dark side.
+ */
+ (*old_roots)++;
ulist_reinit(tmp);
- /* XXX id not needed */
- ret = ulist_add(tmp, qg->qgroupid,
- (u64)(uintptr_t)qg, GFP_ATOMIC);
+ ret = ulist_add(qgroups, qg->qgroupid, ptr_to_u64(qg),
+ GFP_ATOMIC);
+ if (ret < 0)
+ return ret;
+ ret = ulist_add(tmp, qg->qgroupid, ptr_to_u64(qg), GFP_ATOMIC);
if (ret < 0)
return ret;
ULIST_ITER_INIT(&tmp_uiter);
while ((tmp_unode = ulist_next(tmp, &tmp_uiter))) {
struct btrfs_qgroup_list *glist;
- qg = (struct btrfs_qgroup *)(uintptr_t)tmp_unode->aux;
- if (qg->refcnt < seq)
- qg->refcnt = seq + 1;
+ qg = u64_to_ptr(tmp_unode->aux);
+ /*
+ * We use this sequence number to keep from having to
+ * run the whole list and 0 out the refcnt every time.
+ * We basically use sequnce as the known 0 count and
+ * then add 1 everytime we see a qgroup. This is how we
+ * get how many of the roots actually point up to the
+ * upper level qgroups in order to determine exclusive
+ * counts.
+ *
+ * For rescan we want to set old_refcnt to seq so our
+ * exclusive calculations end up correct.
+ */
+ if (rescan)
+ qg->old_refcnt = seq;
+ else if (qg->old_refcnt < seq)
+ qg->old_refcnt = seq + 1;
else
- ++qg->refcnt;
+ qg->old_refcnt++;
+ if (qg->new_refcnt < seq)
+ qg->new_refcnt = seq + 1;
+ else
+ qg->new_refcnt++;
list_for_each_entry(glist, &qg->groups, next_group) {
+ ret = ulist_add(qgroups, glist->group->qgroupid,
+ ptr_to_u64(glist->group),
+ GFP_ATOMIC);
+ if (ret < 0)
+ return ret;
ret = ulist_add(tmp, glist->group->qgroupid,
- (u64)(uintptr_t)glist->group,
+ ptr_to_u64(glist->group),
GFP_ATOMIC);
if (ret < 0)
return ret;
}
}
}
+ return 0;
+}
+/*
+ * We need to walk forward in our operation tree and account for any roots that
+ * were deleted after we made this operation.
+ */
+static int qgroup_account_deleted_refs(struct btrfs_fs_info *fs_info,
+ struct btrfs_qgroup_operation *oper,
+ struct ulist *tmp,
+ struct ulist *qgroups, u64 seq,
+ int *old_roots)
+{
+ struct ulist_node *unode;
+ struct ulist_iterator uiter;
+ struct btrfs_qgroup *qg;
+ struct btrfs_qgroup_operation *tmp_oper;
+ struct rb_node *n;
+ int ret;
+
+ ulist_reinit(tmp);
+
+ /*
+ * We only walk forward in the tree since we're only interested in
+ * removals that happened _after_ our operation.
+ */
+ spin_lock(&fs_info->qgroup_op_lock);
+ n = rb_next(&oper->n);
+ spin_unlock(&fs_info->qgroup_op_lock);
+ if (!n)
+ return 0;
+ tmp_oper = rb_entry(n, struct btrfs_qgroup_operation, n);
+ while (tmp_oper->bytenr == oper->bytenr) {
+ /*
+ * If it's not a removal we don't care, additions work out
+ * properly with our refcnt tracking.
+ */
+ if (tmp_oper->type != BTRFS_QGROUP_OPER_SUB_SHARED &&
+ tmp_oper->type != BTRFS_QGROUP_OPER_SUB_EXCL)
+ goto next;
+ qg = find_qgroup_rb(fs_info, tmp_oper->ref_root);
+ if (!qg)
+ goto next;
+ ret = ulist_add(qgroups, qg->qgroupid, ptr_to_u64(qg),
+ GFP_ATOMIC);
+ if (ret) {
+ if (ret < 0)
+ return ret;
+ /*
+ * We only want to increase old_roots if this qgroup is
+ * not already in the list of qgroups. If it is already
+ * there then that means it must have been re-added or
+ * the delete will be discarded because we had an
+ * existing ref that we haven't looked up yet. In this
+ * case we don't want to increase old_roots. So if ret
+ * == 1 then we know that this is the first time we've
+ * seen this qgroup and we can bump the old_roots.
+ */
+ (*old_roots)++;
+ ret = ulist_add(tmp, qg->qgroupid, ptr_to_u64(qg),
+ GFP_ATOMIC);
+ if (ret < 0)
+ return ret;
+ }
+next:
+ spin_lock(&fs_info->qgroup_op_lock);
+ n = rb_next(&tmp_oper->n);
+ spin_unlock(&fs_info->qgroup_op_lock);
+ if (!n)
+ break;
+ tmp_oper = rb_entry(n, struct btrfs_qgroup_operation, n);
+ }
+
+ /* Ok now process the qgroups we found */
+ ULIST_ITER_INIT(&uiter);
+ while ((unode = ulist_next(tmp, &uiter))) {
+ struct btrfs_qgroup_list *glist;
+
+ qg = u64_to_ptr(unode->aux);
+ if (qg->old_refcnt < seq)
+ qg->old_refcnt = seq + 1;
+ else
+ qg->old_refcnt++;
+ if (qg->new_refcnt < seq)
+ qg->new_refcnt = seq + 1;
+ else
+ qg->new_refcnt++;
+ list_for_each_entry(glist, &qg->groups, next_group) {
+ ret = ulist_add(qgroups, glist->group->qgroupid,
+ ptr_to_u64(glist->group), GFP_ATOMIC);
+ if (ret < 0)
+ return ret;
+ ret = ulist_add(tmp, glist->group->qgroupid,
+ ptr_to_u64(glist->group), GFP_ATOMIC);
+ if (ret < 0)
+ return ret;
+ }
+ }
return 0;
}
-static int qgroup_account_ref_step2(struct btrfs_fs_info *fs_info,
- struct ulist *roots, struct ulist *tmp,
- u64 seq, int sgn, u64 num_bytes,
- struct btrfs_qgroup *qgroup)
+/* Add refcnt for the newly added reference. */
+static int qgroup_calc_new_refcnt(struct btrfs_fs_info *fs_info,
+ struct btrfs_qgroup_operation *oper,
+ struct btrfs_qgroup *qgroup,
+ struct ulist *tmp, struct ulist *qgroups,
+ u64 seq)
{
struct ulist_node *unode;
struct ulist_iterator uiter;
struct btrfs_qgroup *qg;
- struct btrfs_qgroup_list *glist;
int ret;
ulist_reinit(tmp);
- ret = ulist_add(tmp, qgroup->qgroupid, (uintptr_t)qgroup, GFP_ATOMIC);
+ ret = ulist_add(qgroups, qgroup->qgroupid, ptr_to_u64(qgroup),
+ GFP_ATOMIC);
+ if (ret < 0)
+ return ret;
+ ret = ulist_add(tmp, qgroup->qgroupid, ptr_to_u64(qgroup),
+ GFP_ATOMIC);
if (ret < 0)
return ret;
-
ULIST_ITER_INIT(&uiter);
while ((unode = ulist_next(tmp, &uiter))) {
- qg = (struct btrfs_qgroup *)(uintptr_t)unode->aux;
- if (qg->refcnt < seq) {
- /* not visited by step 1 */
- qg->rfer += sgn * num_bytes;
- qg->rfer_cmpr += sgn * num_bytes;
- if (roots->nnodes == 0) {
- qg->excl += sgn * num_bytes;
- qg->excl_cmpr += sgn * num_bytes;
- }
- qgroup_dirty(fs_info, qg);
- }
- WARN_ON(qg->tag >= seq);
- qg->tag = seq;
+ struct btrfs_qgroup_list *glist;
+ qg = u64_to_ptr(unode->aux);
+ if (oper->type == BTRFS_QGROUP_OPER_ADD_SHARED) {
+ if (qg->new_refcnt < seq)
+ qg->new_refcnt = seq + 1;
+ else
+ qg->new_refcnt++;
+ } else {
+ if (qg->old_refcnt < seq)
+ qg->old_refcnt = seq + 1;
+ else
+ qg->old_refcnt++;
+ }
list_for_each_entry(glist, &qg->groups, next_group) {
ret = ulist_add(tmp, glist->group->qgroupid,
- (uintptr_t)glist->group, GFP_ATOMIC);
+ ptr_to_u64(glist->group), GFP_ATOMIC);
+ if (ret < 0)
+ return ret;
+ ret = ulist_add(qgroups, glist->group->qgroupid,
+ ptr_to_u64(glist->group), GFP_ATOMIC);
if (ret < 0)
return ret;
}
}
-
return 0;
}
-static int qgroup_account_ref_step3(struct btrfs_fs_info *fs_info,
- struct ulist *roots, struct ulist *tmp,
- u64 seq, int sgn, u64 num_bytes)
+/*
+ * This adjusts the counters for all referenced qgroups if need be.
+ */
+static int qgroup_adjust_counters(struct btrfs_fs_info *fs_info,
+ u64 root_to_skip, u64 num_bytes,
+ struct ulist *qgroups, u64 seq,
+ int old_roots, int new_roots, int rescan)
{
struct ulist_node *unode;
struct ulist_iterator uiter;
struct btrfs_qgroup *qg;
- struct ulist_node *tmp_unode;
- struct ulist_iterator tmp_uiter;
- int ret;
+ u64 cur_new_count, cur_old_count;
ULIST_ITER_INIT(&uiter);
- while ((unode = ulist_next(roots, &uiter))) {
- qg = find_qgroup_rb(fs_info, unode->val);
- if (!qg)
- continue;
+ while ((unode = ulist_next(qgroups, &uiter))) {
+ bool dirty = false;
- ulist_reinit(tmp);
- ret = ulist_add(tmp, qg->qgroupid, (uintptr_t)qg, GFP_ATOMIC);
- if (ret < 0)
- return ret;
+ qg = u64_to_ptr(unode->aux);
+ /*
+ * Wasn't referenced before but is now, add to the reference
+ * counters.
+ */
+ if (qg->old_refcnt <= seq && qg->new_refcnt > seq) {
+ qg->rfer += num_bytes;
+ qg->rfer_cmpr += num_bytes;
+ dirty = true;
+ }
- ULIST_ITER_INIT(&tmp_uiter);
- while ((tmp_unode = ulist_next(tmp, &tmp_uiter))) {
- struct btrfs_qgroup_list *glist;
+ /*
+ * Was referenced before but isn't now, subtract from the
+ * reference counters.
+ */
+ if (qg->old_refcnt > seq && qg->new_refcnt <= seq) {
+ qg->rfer -= num_bytes;
+ qg->rfer_cmpr -= num_bytes;
+ dirty = true;
+ }
- qg = (struct btrfs_qgroup *)(uintptr_t)tmp_unode->aux;
- if (qg->tag == seq)
- continue;
+ if (qg->old_refcnt < seq)
+ cur_old_count = 0;
+ else
+ cur_old_count = qg->old_refcnt - seq;
+ if (qg->new_refcnt < seq)
+ cur_new_count = 0;
+ else
+ cur_new_count = qg->new_refcnt - seq;
- if (qg->refcnt - seq == roots->nnodes) {
- qg->excl -= sgn * num_bytes;
- qg->excl_cmpr -= sgn * num_bytes;
- qgroup_dirty(fs_info, qg);
- }
+ /*
+ * If our refcount was the same as the roots previously but our
+ * new count isn't the same as the number of roots now then we
+ * went from having a exclusive reference on this range to not.
+ */
+ if (old_roots && cur_old_count == old_roots &&
+ (cur_new_count != new_roots || new_roots == 0)) {
+ WARN_ON(cur_new_count != new_roots && new_roots == 0);
+ qg->excl -= num_bytes;
+ qg->excl_cmpr -= num_bytes;
+ dirty = true;
+ }
- list_for_each_entry(glist, &qg->groups, next_group) {
- ret = ulist_add(tmp, glist->group->qgroupid,
- (uintptr_t)glist->group,
- GFP_ATOMIC);
- if (ret < 0)
- return ret;
- }
+ /*
+ * If we didn't reference all the roots before but now we do we
+ * have an exclusive reference to this range.
+ */
+ if ((!old_roots || (old_roots && cur_old_count != old_roots))
+ && cur_new_count == new_roots) {
+ qg->excl += num_bytes;
+ qg->excl_cmpr += num_bytes;
+ dirty = true;
}
- }
+ if (dirty)
+ qgroup_dirty(fs_info, qg);
+ }
return 0;
}
/*
- * btrfs_qgroup_account_ref is called for every ref that is added to or deleted
- * from the fs. First, all roots referencing the extent are searched, and
- * then the space is accounted accordingly to the different roots. The
- * accounting algorithm works in 3 steps documented inline.
+ * If we removed a data extent and there were other references for that bytenr
+ * then we need to lookup all referenced roots to make sure we still don't
+ * reference this bytenr. If we do then we can just discard this operation.
*/
-int btrfs_qgroup_account_ref(struct btrfs_trans_handle *trans,
- struct btrfs_fs_info *fs_info,
- struct btrfs_delayed_ref_node *node,
- struct btrfs_delayed_extent_op *extent_op)
+static int check_existing_refs(struct btrfs_trans_handle *trans,
+ struct btrfs_fs_info *fs_info,
+ struct btrfs_qgroup_operation *oper)
{
- struct btrfs_root *quota_root;
- u64 ref_root;
- struct btrfs_qgroup *qgroup;
struct ulist *roots = NULL;
- u64 seq;
+ struct ulist_node *unode;
+ struct ulist_iterator uiter;
int ret = 0;
- int sgn;
- if (!fs_info->quota_enabled)
- return 0;
-
- BUG_ON(!fs_info->quota_root);
+ ret = btrfs_find_all_roots(trans, fs_info, oper->bytenr,
+ oper->elem.seq, &roots);
+ if (ret < 0)
+ return ret;
+ ret = 0;
- if (node->type == BTRFS_TREE_BLOCK_REF_KEY ||
- node->type == BTRFS_SHARED_BLOCK_REF_KEY) {
- struct btrfs_delayed_tree_ref *ref;
- ref = btrfs_delayed_node_to_tree_ref(node);
- ref_root = ref->root;
- } else if (node->type == BTRFS_EXTENT_DATA_REF_KEY ||
- node->type == BTRFS_SHARED_DATA_REF_KEY) {
- struct btrfs_delayed_data_ref *ref;
- ref = btrfs_delayed_node_to_data_ref(node);
- ref_root = ref->root;
- } else {
- BUG();
+ ULIST_ITER_INIT(&uiter);
+ while ((unode = ulist_next(roots, &uiter))) {
+ if (unode->val == oper->ref_root) {
+ ret = 1;
+ break;
+ }
}
+ ulist_free(roots);
+ btrfs_put_tree_mod_seq(fs_info, &oper->elem);
- if (!is_fstree(ref_root)) {
- /*
- * non-fs-trees are not being accounted
- */
- return 0;
- }
+ return ret;
+}
- switch (node->action) {
- case BTRFS_ADD_DELAYED_REF:
- case BTRFS_ADD_DELAYED_EXTENT:
- sgn = 1;
- seq = btrfs_tree_mod_seq_prev(node->seq);
- break;
- case BTRFS_DROP_DELAYED_REF:
- sgn = -1;
- seq = node->seq;
- break;
- case BTRFS_UPDATE_DELAYED_HEAD:
- return 0;
- default:
- BUG();
- }
+/*
+ * If we share a reference across multiple roots then we may need to adjust
+ * various qgroups referenced and exclusive counters. The basic premise is
this
+ *
+ * 1) We have seq to represent a 0 count. Instead of looping through all of
the
+ * qgroups and resetting their refcount to 0 we just constantly bump this
+ * sequence number to act as the base reference count. This means that if
+ * anybody is equal to or below this sequence they were never referenced. We
+ * jack this sequence up by the number of roots we found each time in order to
+ * make sure we don't have any overlap.
+ *
+ * 2) We first search all the roots that reference the area _except_ the root
+ * we're acting on currently. This makes up the old_refcnt of all the
qgroups
+ * before.
+ *
+ * 3) We walk all of the qgroups referenced by the root we are currently acting
+ * on, and will either adjust old_refcnt in the case of a removal or the
+ * new_refcnt in the case of an addition.
+ *
+ * 4) Finally we walk all the qgroups that are referenced by this range
+ * including the root we are acting on currently. We will adjust the counters
+ * based on the number of roots we had and will have after this operation.
+ *
+ * Take this example as an illustration
+ *
+ * [qgroup 1/0]
+ * / | \
+ * [qg 0/0] [qg 0/1] [qg 0/2]
+ * \ | /
+ * [ extent ]
+ *
+ * Say we are adding a reference that is covered by qg 0/0. The first step
+ * would give a refcnt of 1 to qg 0/1 and 0/2 and a refcnt of 2 to qg 1/0 with
+ * old_roots being 2. Because it is adding new_roots will be 1. We then go
+ * through qg 0/0 which will get the new_refcnt set to 1 and add 1 to qg
1/0's
+ * new_refcnt, bringing it to 3. We then walk through all of the qgroups, we
+ * notice that the old refcnt for qg 0/0 < the new refcnt, so we added a
+ * reference and thus must add the size to the referenced bytes. Everything
+ * else is the same so nothing else changes.
+ */
+static int qgroup_shared_accounting(struct btrfs_trans_handle *trans,
+ struct btrfs_fs_info *fs_info,
+ struct btrfs_qgroup_operation *oper)
+{
+ struct ulist *roots = NULL;
+ struct ulist *qgroups, *tmp;
+ struct btrfs_qgroup *qgroup;
+ struct seq_list elem = {};
+ u64 seq;
+ int old_roots = 0;
+ int new_roots = 0;
+ int ret = 0;
- mutex_lock(&fs_info->qgroup_rescan_lock);
- if (fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN) {
- if (fs_info->qgroup_rescan_progress.objectid <= node->bytenr) {
- mutex_unlock(&fs_info->qgroup_rescan_lock);
+ if (oper->elem.seq) {
+ ret = check_existing_refs(trans, fs_info, oper);
+ if (ret < 0)
+ return ret;
+ if (ret)
return 0;
- }
}
- mutex_unlock(&fs_info->qgroup_rescan_lock);
- /*
- * the delayed ref sequence number we pass depends on the direction of
- * the operation. for add operations, we pass
- * tree_mod_log_prev_seq(node->seq) to skip
- * the delayed ref's current sequence number, because we need the state
- * of the tree before the add operation. for delete operations, we pass
- * (node->seq) to include the delayed ref's current sequence number,
- * because we need the state of the tree after the delete operation.
- */
- ret = btrfs_find_all_roots(trans, fs_info, node->bytenr, seq, &roots);
- if (ret < 0)
- return ret;
-
- spin_lock(&fs_info->qgroup_lock);
+ qgroups = ulist_alloc(GFP_NOFS);
+ if (!qgroups)
+ return -ENOMEM;
- quota_root = fs_info->quota_root;
- if (!quota_root)
- goto unlock;
+ tmp = ulist_alloc(GFP_NOFS);
+ if (!tmp)
+ return -ENOMEM;
- qgroup = find_qgroup_rb(fs_info, ref_root);
+ btrfs_get_tree_mod_seq(fs_info, &elem);
+ ret = btrfs_find_all_roots(trans, fs_info, oper->bytenr, elem.seq,
+ &roots);
+ btrfs_put_tree_mod_seq(fs_info, &elem);
+ if (ret < 0) {
+ ulist_free(qgroups);
+ ulist_free(tmp);
+ return ret;
+ }
+ spin_lock(&fs_info->qgroup_lock);
+ qgroup = find_qgroup_rb(fs_info, oper->ref_root);
if (!qgroup)
- goto unlock;
+ goto out;
+ seq = fs_info->qgroup_seq;
/*
- * step 1: for each old ref, visit all nodes once and inc refcnt
+ * So roots is the list of all the roots currently pointing at the
+ * bytenr, including the ref we are adding if we are adding, or not if
+ * we are removing a ref. So we pass in the ref_root to skip that root
+ * in our calculations. We set old_refnct and new_refcnt cause who the
+ * hell knows what everything looked like before, and it doesn't matter
+ * except...
*/
- ulist_reinit(fs_info->qgroup_ulist);
- seq = fs_info->qgroup_seq;
- fs_info->qgroup_seq += roots->nnodes + 1; /* max refcnt */
+ ret = qgroup_calc_old_refcnt(fs_info, oper->ref_root, tmp, roots, qgroups,
+ seq, &old_roots, 0);
+ if (ret < 0)
+ goto out;
- ret = qgroup_account_ref_step1(fs_info, roots, fs_info->qgroup_ulist,
- seq);
- if (ret)
- goto unlock;
+ /*
+ * Now adjust the refcounts of the qgroups that care about this
+ * reference, either the old_count in the case of removal or new_count
+ * in the case of an addition.
+ */
+ ret = qgroup_calc_new_refcnt(fs_info, oper, qgroup, tmp, qgroups,
+ seq);
+ if (ret < 0)
+ goto out;
/*
- * step 2: walk from the new root
+ * ...in the case of removals. If we had a removal before we got around
+ * to processing this operation then we need to find that guy and count
+ * his references as if they really existed so we don't end up screwing
+ * up the exclusive counts. Then whenever we go to process the delete
+ * everything will be grand and we can account for whatever exclusive
+ * changes need to be made there. We also have to pass in old_roots so
+ * we have an accurate count of the roots as it pertains to this
+ * operations view of the world.
*/
- ret = qgroup_account_ref_step2(fs_info, roots, fs_info->qgroup_ulist,
- seq, sgn, node->num_bytes, qgroup);
- if (ret)
- goto unlock;
+ ret = qgroup_account_deleted_refs(fs_info, oper, tmp, qgroups, seq,
+ &old_roots);
+ if (ret < 0)
+ goto out;
/*
- * step 3: walk again from old refs
+ * We are adding our root, need to adjust up the number of roots,
+ * otherwise old_roots is the number of roots we want.
*/
- ret = qgroup_account_ref_step3(fs_info, roots, fs_info->qgroup_ulist,
- seq, sgn, node->num_bytes);
- if (ret)
- goto unlock;
+ if (oper->type == BTRFS_QGROUP_OPER_ADD_SHARED) {
+ new_roots = old_roots + 1;
+ } else {
+ new_roots = old_roots;
+ old_roots++;
+ }
+ fs_info->qgroup_seq += old_roots + 1;
-unlock:
+
+ /*
+ * And now the magic happens, bless Arne for having a pretty elegant
+ * solution for this.
+ */
+ qgroup_adjust_counters(fs_info, oper->ref_root, oper->num_bytes,
+ qgroups, seq, old_roots, new_roots, 0);
+out:
spin_unlock(&fs_info->qgroup_lock);
+ ulist_free(qgroups);
ulist_free(roots);
+ ulist_free(tmp);
+ return ret;
+}
+
+/*
+ * btrfs_qgroup_account_ref is called for every ref that is added to or deleted
+ * from the fs. First, all roots referencing the extent are searched, and
+ * then the space is accounted accordingly to the different roots. The
+ * accounting algorithm works in 3 steps documented inline.
+ */
+static int btrfs_qgroup_account(struct btrfs_trans_handle *trans,
+ struct btrfs_fs_info *fs_info,
+ struct btrfs_qgroup_operation *oper)
+{
+ int ret = 0;
+
+ if (!fs_info->quota_enabled)
+ return 0;
+
+ BUG_ON(!fs_info->quota_root);
+
+ mutex_lock(&fs_info->qgroup_rescan_lock);
+ if (fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN) {
+ if (fs_info->qgroup_rescan_progress.objectid <= oper->bytenr) {
+ mutex_unlock(&fs_info->qgroup_rescan_lock);
+ return 0;
+ }
+ }
+ mutex_unlock(&fs_info->qgroup_rescan_lock);
+
+ ASSERT(is_fstree(oper->ref_root));
+
+ switch (oper->type) {
+ case BTRFS_QGROUP_OPER_ADD_EXCL:
+ case BTRFS_QGROUP_OPER_SUB_EXCL:
+ ret = qgroup_excl_accounting(fs_info, oper);
+ break;
+ case BTRFS_QGROUP_OPER_ADD_SHARED:
+ case BTRFS_QGROUP_OPER_SUB_SHARED:
+ ret = qgroup_shared_accounting(trans, fs_info, oper);
+ break;
+ default:
+ ASSERT(0);
+ }
+ return ret;
+}
+/*
+ * Needs to be called everytime we run delayed refs, even if there is an error
+ * in order to cleanup outstanding operations.
+ */
+int btrfs_delayed_qgroup_accounting(struct btrfs_trans_handle *trans,
+ struct btrfs_fs_info *fs_info)
+{
+ struct btrfs_qgroup_operation *oper;
+ int ret = 0;
+
+ while (!list_empty(&trans->qgroup_ref_list)) {
+ oper = list_first_entry(&trans->qgroup_ref_list,
+ struct btrfs_qgroup_operation, list);
+ list_del_init(&oper->list);
+ if (!ret || !trans->aborted)
+ ret = btrfs_qgroup_account(trans, fs_info, oper);
+ spin_lock(&fs_info->qgroup_op_lock);
+ rb_erase(&oper->n, &fs_info->qgroup_op_tree);
+ spin_unlock(&fs_info->qgroup_op_lock);
+ btrfs_put_tree_mod_seq(fs_info, &oper->elem);
+ kfree(oper);
+ }
return ret;
}
@@ -1629,8 +2091,16 @@ int btrfs_qgroup_inherit(struct btrfs_trans_handle
*trans,
srcgroup = find_qgroup_rb(fs_info, srcid);
if (!srcgroup)
goto unlock;
- dstgroup->rfer = srcgroup->rfer - level_size;
- dstgroup->rfer_cmpr = srcgroup->rfer_cmpr - level_size;
+
+ /*
+ * We call inherit after we clone the root in order to make sure
+ * our counts don't go crazy, so at this point the only
+ * difference between the two roots should be the root node.
+ */
+ dstgroup->rfer = srcgroup->rfer;
+ dstgroup->rfer_cmpr = srcgroup->rfer_cmpr;
+ dstgroup->excl = level_size;
+ dstgroup->excl_cmpr = level_size;
srcgroup->excl = level_size;
srcgroup->excl_cmpr = level_size;
qgroup_dirty(fs_info, dstgroup);
@@ -1734,7 +2204,7 @@ int btrfs_qgroup_reserve(struct btrfs_root *root, u64
num_bytes)
struct btrfs_qgroup *qg;
struct btrfs_qgroup_list *glist;
- qg = (struct btrfs_qgroup *)(uintptr_t)unode->aux;
+ qg = u64_to_ptr(unode->aux);
if ((qg->lim_flags & BTRFS_QGROUP_LIMIT_MAX_RFER) &&
qg->reserved + (s64)qg->rfer + num_bytes >
@@ -1766,7 +2236,7 @@ int btrfs_qgroup_reserve(struct btrfs_root *root, u64
num_bytes)
while ((unode = ulist_next(fs_info->qgroup_ulist, &uiter))) {
struct btrfs_qgroup *qg;
- qg = (struct btrfs_qgroup *)(uintptr_t)unode->aux;
+ qg = u64_to_ptr(unode->aux);
qg->reserved += num_bytes;
}
@@ -1812,7 +2282,7 @@ void btrfs_qgroup_free(struct btrfs_root *root, u64
num_bytes)
struct btrfs_qgroup *qg;
struct btrfs_qgroup_list *glist;
- qg = (struct btrfs_qgroup *)(uintptr_t)unode->aux;
+ qg = u64_to_ptr(unode->aux);
qg->reserved -= num_bytes;
@@ -1848,15 +2318,15 @@ void assert_qgroups_uptodate(struct btrfs_trans_handle
*trans)
*/
static int
qgroup_rescan_leaf(struct btrfs_fs_info *fs_info, struct btrfs_path *path,
- struct btrfs_trans_handle *trans, struct ulist *tmp,
- struct extent_buffer *scratch_leaf)
+ struct btrfs_trans_handle *trans, struct ulist *qgroups,
+ struct ulist *tmp, struct extent_buffer *scratch_leaf)
{
struct btrfs_key found;
struct ulist *roots = NULL;
- struct ulist_node *unode;
- struct ulist_iterator uiter;
struct seq_list tree_mod_seq_elem = {};
+ u64 num_bytes;
u64 seq;
+ int new_roots;
int slot;
int ret;
@@ -1897,8 +2367,6 @@ qgroup_rescan_leaf(struct btrfs_fs_info *fs_info, struct
btrfs_path *path,
mutex_unlock(&fs_info->qgroup_rescan_lock);
for (; slot < btrfs_header_nritems(scratch_leaf); ++slot) {
- u64 num_bytes;
-
btrfs_item_key_to_cpu(scratch_leaf, &found, slot);
if (found.type != BTRFS_EXTENT_ITEM_KEY &&
found.type != BTRFS_METADATA_ITEM_KEY)
@@ -1908,76 +2376,34 @@ qgroup_rescan_leaf(struct btrfs_fs_info *fs_info, struct
btrfs_path *path,
else
num_bytes = found.offset;
- ret = btrfs_find_all_roots(trans, fs_info, found.objectid,
- tree_mod_seq_elem.seq, &roots);
+ ulist_reinit(qgroups);
+ ret = btrfs_find_all_roots(NULL, fs_info, found.objectid, 0,
+ &roots);
if (ret < 0)
goto out;
spin_lock(&fs_info->qgroup_lock);
seq = fs_info->qgroup_seq;
fs_info->qgroup_seq += roots->nnodes + 1; /* max refcnt */
- ret = qgroup_account_ref_step1(fs_info, roots, tmp, seq);
- if (ret) {
+ new_roots = 0;
+ ret = qgroup_calc_old_refcnt(fs_info, 0, tmp, roots, qgroups,
+ seq, &new_roots, 1);
+ if (ret < 0) {
spin_unlock(&fs_info->qgroup_lock);
ulist_free(roots);
goto out;
}
- /*
- * step2 of btrfs_qgroup_account_ref works from a single root,
- * we're doing all at once here.
- */
- ulist_reinit(tmp);
- ULIST_ITER_INIT(&uiter);
- while ((unode = ulist_next(roots, &uiter))) {
- struct btrfs_qgroup *qg;
-
- qg = find_qgroup_rb(fs_info, unode->val);
- if (!qg)
- continue;
-
- ret = ulist_add(tmp, qg->qgroupid, (uintptr_t)qg,
- GFP_ATOMIC);
- if (ret < 0) {
- spin_unlock(&fs_info->qgroup_lock);
- ulist_free(roots);
- goto out;
- }
- }
-
- /* this loop is similar to step 2 of btrfs_qgroup_account_ref */
- ULIST_ITER_INIT(&uiter);
- while ((unode = ulist_next(tmp, &uiter))) {
- struct btrfs_qgroup *qg;
- struct btrfs_qgroup_list *glist;
-
- qg = (struct btrfs_qgroup *)(uintptr_t) unode->aux;
- qg->rfer += num_bytes;
- qg->rfer_cmpr += num_bytes;
- WARN_ON(qg->tag >= seq);
- if (qg->refcnt - seq == roots->nnodes) {
- qg->excl += num_bytes;
- qg->excl_cmpr += num_bytes;
- }
- qgroup_dirty(fs_info, qg);
-
- list_for_each_entry(glist, &qg->groups, next_group) {
- ret = ulist_add(tmp, glist->group->qgroupid,
- (uintptr_t)glist->group,
- GFP_ATOMIC);
- if (ret < 0) {
- spin_unlock(&fs_info->qgroup_lock);
- ulist_free(roots);
- goto out;
- }
- }
+ ret = qgroup_adjust_counters(fs_info, 0, num_bytes, qgroups,
+ seq, 0, new_roots, 1);
+ if (ret < 0) {
+ spin_unlock(&fs_info->qgroup_lock);
+ ulist_free(roots);
+ goto out;
}
-
spin_unlock(&fs_info->qgroup_lock);
ulist_free(roots);
- ret = 0;
}
-
out:
btrfs_put_tree_mod_seq(fs_info, &tree_mod_seq_elem);
@@ -1990,13 +2416,16 @@ static void btrfs_qgroup_rescan_worker(struct btrfs_work
*work)
qgroup_rescan_work);
struct btrfs_path *path;
struct btrfs_trans_handle *trans = NULL;
- struct ulist *tmp = NULL;
+ struct ulist *tmp = NULL, *qgroups = NULL;
struct extent_buffer *scratch_leaf = NULL;
int err = -ENOMEM;
path = btrfs_alloc_path();
if (!path)
goto out;
+ qgroups = ulist_alloc(GFP_NOFS);
+ if (!qgroups)
+ goto out;
tmp = ulist_alloc(GFP_NOFS);
if (!tmp)
goto out;
@@ -2015,7 +2444,7 @@ static void btrfs_qgroup_rescan_worker(struct btrfs_work
*work)
err = -EINTR;
} else {
err = qgroup_rescan_leaf(fs_info, path, trans,
- tmp, scratch_leaf);
+ qgroups, tmp, scratch_leaf);
}
if (err > 0)
btrfs_commit_transaction(trans, fs_info->fs_root);
@@ -2025,7 +2454,7 @@ static void btrfs_qgroup_rescan_worker(struct btrfs_work
*work)
out:
kfree(scratch_leaf);
- ulist_free(tmp);
+ ulist_free(qgroups);
btrfs_free_path(path);
mutex_lock(&fs_info->qgroup_rescan_lock);
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 7579f6d..c7de11f 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -31,6 +31,7 @@
#include "inode-map.h"
#include "volumes.h"
#include "dev-replace.h"
+#include "qgroup.h"
#define BTRFS_ROOT_TRANS_TAG 0
@@ -702,23 +703,9 @@ static int __btrfs_end_transaction(struct
btrfs_trans_handle *trans,
return 0;
}
- /*
- * do the qgroup accounting as early as possible
- */
- err = btrfs_delayed_refs_qgroup_accounting(trans, info);
-
btrfs_trans_release_metadata(trans, root);
trans->block_rsv = NULL;
- if (trans->qgroup_reserved) {
- /*
- * the same root has to be passed here between start_transaction
- * and end_transaction. Subvolume quota depends on this.
- */
- btrfs_qgroup_free(trans->root, trans->qgroup_reserved);
- trans->qgroup_reserved = 0;
- }
-
if (!list_empty(&trans->new_bgs))
btrfs_create_pending_block_groups(trans, root);
@@ -729,6 +716,15 @@ static int __btrfs_end_transaction(struct
btrfs_trans_handle *trans,
btrfs_run_delayed_refs(trans, root, cur);
}
+ if (trans->qgroup_reserved) {
+ /*
+ * the same root has to be passed here between start_transaction
+ * and end_transaction. Subvolume quota depends on this.
+ */
+ btrfs_qgroup_free(trans->root, trans->qgroup_reserved);
+ trans->qgroup_reserved = 0;
+ }
+
btrfs_trans_release_metadata(trans, root);
trans->block_rsv = NULL;
@@ -1168,12 +1164,6 @@ static noinline int create_pending_snapshot(struct
btrfs_trans_handle *trans,
goto no_free_objectid;
}
- pending->error = btrfs_qgroup_inherit(trans, fs_info,
- root->root_key.objectid,
- objectid, pending->inherit);
- if (pending->error)
- goto no_free_objectid;
-
key.objectid = objectid;
key.offset = (u64)-1;
key.type = BTRFS_ROOT_ITEM_KEY;
@@ -1270,6 +1260,22 @@ static noinline int create_pending_snapshot(struct
btrfs_trans_handle *trans,
goto fail;
}
+ /*
+ * We need to flush delayed refs in order to make sure all of our quota
+ * operations have been done before we call btrfs_qgroup_inherit.
+ */
+ ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1);
+ if (ret) {
+ btrfs_abort_transaction(trans, root, ret);
+ goto fail;
+ }
+
+ pending->error = btrfs_qgroup_inherit(trans, fs_info,
+ root->root_key.objectid,
+ objectid, pending->inherit);
+ if (pending->error)
+ goto no_free_objectid;
+
/* see comments in should_cow_block() */
root->force_cow = 1;
smp_wmb();
@@ -1598,12 +1604,6 @@ static int btrfs_flush_all_pending_stuffs(struct
btrfs_trans_handle *trans,
* them now so that they hinder processing of more delayed refs
* as little as possible.
*/
- if (ret) {
- btrfs_delayed_refs_qgroup_accounting(trans, root->fs_info);
- return ret;
- }
-
- ret = btrfs_delayed_refs_qgroup_accounting(trans, root->fs_info);
if (ret)
return ret;
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html