Home Home > GIT Browse > openSUSE-42.3
summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorTakashi Iwai <tiwai@suse.de>2019-05-17 16:15:48 +0200
committerTakashi Iwai <tiwai@suse.de>2019-05-17 16:15:48 +0200
commit4e93bd37874648fdb14109ecd3a79983ee9b5e7e (patch)
tree5e7a5f3d6a6206b861be219cb624a9e24add892c
parentb1bb959e5425f351e25cc9ae93f0d29b4afe3ba1 (diff)
parentba7f8deee3398a233d3c1e48410e7cf13b40940f (diff)
Merge branch 'users/wqu/SLE12-SP3/for-next' into SLE12-SP3
Pull btrfs fixes from Qu Wenruo (bsc#1134813) suse-commit: d88cfc3d427ce6e74df1be4eb26c4c09cf830de7
-rw-r--r--fs/btrfs/backref.c4
-rw-r--r--fs/btrfs/delayed-ref.c374
-rw-r--r--fs/btrfs/delayed-ref.h49
-rw-r--r--fs/btrfs/disk-io.c11
-rw-r--r--fs/btrfs/extent-tree.c89
-rw-r--r--include/trace/events/btrfs.h13
6 files changed, 234 insertions, 306 deletions
diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
index 42833b0565be..b8ee2bed52f0 100644
--- a/fs/btrfs/backref.c
+++ b/fs/btrfs/backref.c
@@ -1163,7 +1163,7 @@ again:
head = btrfs_find_delayed_ref_head(delayed_refs, bytenr);
if (head) {
if (!mutex_trylock(&head->mutex)) {
- atomic_inc(&head->node.refs);
+ atomic_inc(&head->refs);
spin_unlock(&delayed_refs->lock);
btrfs_release_path(path);
@@ -1174,7 +1174,7 @@ again:
*/
mutex_lock(&head->mutex);
mutex_unlock(&head->mutex);
- btrfs_put_delayed_ref(&head->node);
+ btrfs_put_delayed_ref_head(head);
goto again;
}
spin_unlock(&delayed_refs->lock);
diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c
index 11d1a79929f5..2a74307e2655 100644
--- a/fs/btrfs/delayed-ref.c
+++ b/fs/btrfs/delayed-ref.c
@@ -124,15 +124,15 @@ static struct btrfs_delayed_ref_head *htree_insert(struct rb_root *root,
u64 bytenr;
ins = rb_entry(node, struct btrfs_delayed_ref_head, href_node);
- bytenr = ins->node.bytenr;
+ bytenr = ins->bytenr;
while (*p) {
parent_node = *p;
entry = rb_entry(parent_node, struct btrfs_delayed_ref_head,
href_node);
- if (bytenr < entry->node.bytenr)
+ if (bytenr < entry->bytenr)
p = &(*p)->rb_left;
- else if (bytenr > entry->node.bytenr)
+ else if (bytenr > entry->bytenr)
p = &(*p)->rb_right;
else
return entry;
@@ -189,15 +189,15 @@ find_ref_head(struct rb_root *root, u64 bytenr,
while (n) {
entry = rb_entry(n, struct btrfs_delayed_ref_head, href_node);
- if (bytenr < entry->node.bytenr)
+ if (bytenr < entry->bytenr)
n = n->rb_left;
- else if (bytenr > entry->node.bytenr)
+ else if (bytenr > entry->bytenr)
n = n->rb_right;
else
return entry;
}
if (entry && return_bigger) {
- if (bytenr > entry->node.bytenr) {
+ if (bytenr > entry->bytenr) {
n = rb_next(&entry->href_node);
if (!n)
n = rb_first(root);
@@ -220,17 +220,17 @@ int btrfs_delayed_ref_lock(struct btrfs_trans_handle *trans,
if (mutex_trylock(&head->mutex))
return 0;
- atomic_inc(&head->node.refs);
+ atomic_inc(&head->refs);
spin_unlock(&delayed_refs->lock);
mutex_lock(&head->mutex);
spin_lock(&delayed_refs->lock);
- if (!head->node.in_tree) {
+ if (RB_EMPTY_NODE(&head->href_node)) {
mutex_unlock(&head->mutex);
- btrfs_put_delayed_ref(&head->node);
+ btrfs_put_delayed_ref_head(head);
return -EAGAIN;
}
- btrfs_put_delayed_ref(&head->node);
+ btrfs_put_delayed_ref_head(head);
return 0;
}
@@ -239,16 +239,11 @@ static inline void drop_delayed_ref(struct btrfs_trans_handle *trans,
struct btrfs_delayed_ref_head *head,
struct btrfs_delayed_ref_node *ref)
{
- if (btrfs_delayed_ref_is_head(ref)) {
- head = btrfs_delayed_node_to_head(ref);
- rb_erase(&head->href_node, &delayed_refs->href_root);
- } else {
- assert_spin_locked(&head->lock);
- rb_erase(&ref->ref_node, &head->ref_tree);
- RB_CLEAR_NODE(&ref->ref_node);
- if (!list_empty(&ref->add_list))
- list_del(&ref->add_list);
- }
+ assert_spin_locked(&head->lock);
+ rb_erase(&ref->ref_node, &head->ref_tree);
+ RB_CLEAR_NODE(&ref->ref_node);
+ if (!list_empty(&ref->add_list))
+ list_del(&ref->add_list);
ref->in_tree = 0;
btrfs_put_delayed_ref(ref);
atomic_dec(&delayed_refs->num_entries);
@@ -409,8 +404,8 @@ again:
head->processing = 1;
WARN_ON(delayed_refs->num_heads_ready == 0);
delayed_refs->num_heads_ready--;
- delayed_refs->run_delayed_start = head->node.bytenr +
- head->node.num_bytes;
+ delayed_refs->run_delayed_start = head->bytenr +
+ head->num_bytes;
return head;
}
@@ -478,20 +473,16 @@ inserted:
*/
static noinline void
update_existing_head_ref(struct btrfs_delayed_ref_root *delayed_refs,
- struct btrfs_delayed_ref_node *existing,
- struct btrfs_delayed_ref_node *update,
+ struct btrfs_delayed_ref_head *existing,
+ struct btrfs_delayed_ref_head *update,
int *old_ref_mod_ret)
{
- struct btrfs_delayed_ref_head *existing_ref;
- struct btrfs_delayed_ref_head *ref;
int old_ref_mod;
- existing_ref = btrfs_delayed_node_to_head(existing);
- ref = btrfs_delayed_node_to_head(update);
- BUG_ON(existing_ref->is_data != ref->is_data);
+ BUG_ON(existing->is_data != update->is_data);
- spin_lock(&existing_ref->lock);
- if (ref->must_insert_reserved) {
+ spin_lock(&existing->lock);
+ if (update->must_insert_reserved) {
/* if the extent was freed and then
* reallocated before the delayed ref
* entries were processed, we can end up
@@ -499,7 +490,7 @@ update_existing_head_ref(struct btrfs_delayed_ref_root *delayed_refs,
* the must_insert_reserved flag set.
* Set it again here
*/
- existing_ref->must_insert_reserved = ref->must_insert_reserved;
+ existing->must_insert_reserved = update->must_insert_reserved;
/*
* update the num_bytes so we make sure the accounting
@@ -509,22 +500,22 @@ update_existing_head_ref(struct btrfs_delayed_ref_root *delayed_refs,
}
- if (ref->extent_op) {
- if (!existing_ref->extent_op) {
- existing_ref->extent_op = ref->extent_op;
+ if (update->extent_op) {
+ if (!existing->extent_op) {
+ existing->extent_op = update->extent_op;
} else {
- if (ref->extent_op->update_key) {
- memcpy(&existing_ref->extent_op->key,
- &ref->extent_op->key,
- sizeof(ref->extent_op->key));
- existing_ref->extent_op->update_key = true;
+ if (update->extent_op->update_key) {
+ memcpy(&existing->extent_op->key,
+ &update->extent_op->key,
+ sizeof(update->extent_op->key));
+ existing->extent_op->update_key = true;
}
- if (ref->extent_op->update_flags) {
- existing_ref->extent_op->flags_to_set |=
- ref->extent_op->flags_to_set;
- existing_ref->extent_op->update_flags = true;
+ if (update->extent_op->update_flags) {
+ existing->extent_op->flags_to_set |=
+ update->extent_op->flags_to_set;
+ existing->extent_op->update_flags = true;
}
- btrfs_free_delayed_extent_op(ref->extent_op);
+ btrfs_free_delayed_extent_op(update->extent_op);
}
}
/*
@@ -532,43 +523,31 @@ update_existing_head_ref(struct btrfs_delayed_ref_root *delayed_refs,
* only need the lock for this case cause we could be processing it
* currently, for refs we just added we know we're a-ok.
*/
- old_ref_mod = existing_ref->total_ref_mod;
+ old_ref_mod = existing->total_ref_mod;
if (old_ref_mod_ret)
*old_ref_mod_ret = old_ref_mod;
existing->ref_mod += update->ref_mod;
- existing_ref->total_ref_mod += update->ref_mod;
+ existing->total_ref_mod += update->ref_mod;
/*
* If we are going to from a positive ref mod to a negative or vice
* versa we need to make sure to adjust pending_csums accordingly.
*/
- if (existing_ref->is_data) {
- if (existing_ref->total_ref_mod >= 0 && old_ref_mod < 0)
+ if (existing->is_data) {
+ if (existing->total_ref_mod >= 0 && old_ref_mod < 0)
delayed_refs->pending_csums -= existing->num_bytes;
- if (existing_ref->total_ref_mod < 0 && old_ref_mod >= 0)
+ if (existing->total_ref_mod < 0 && old_ref_mod >= 0)
delayed_refs->pending_csums += existing->num_bytes;
}
- spin_unlock(&existing_ref->lock);
+ spin_unlock(&existing->lock);
}
-/*
- * helper function to actually insert a head node into the rbtree.
- * this does all the dirty work in terms of maintaining the correct
- * overall modification count.
- */
-static noinline struct btrfs_delayed_ref_head *
-add_delayed_ref_head(struct btrfs_fs_info *fs_info,
- struct btrfs_trans_handle *trans,
- struct btrfs_delayed_ref_node *ref,
- struct btrfs_qgroup_extent_record *qrecord,
- u64 bytenr, u64 num_bytes, u64 ref_root, u64 reserved,
- int action, int is_data, int is_system, int *old_ref_mod,
- int *new_ref_mod)
-
+static void init_delayed_ref_head(struct btrfs_delayed_ref_head *head_ref,
+ struct btrfs_qgroup_extent_record *qrecord,
+ u64 bytenr, u64 num_bytes, u64 ref_root,
+ u64 reserved, int action, bool is_data,
+ bool is_system)
{
- struct btrfs_delayed_ref_head *existing;
- struct btrfs_delayed_ref_head *head_ref = NULL;
- struct btrfs_delayed_ref_root *delayed_refs;
int count_mod = 1;
int must_insert_reserved = 0;
@@ -576,7 +555,7 @@ add_delayed_ref_head(struct btrfs_fs_info *fs_info,
BUG_ON(!is_data && reserved);
/*
- * the head node stores the sum of all the mods, so dropping a ref
+ * The head node stores the sum of all the mods, so dropping a ref
* should drop the sum in the head node by one.
*/
if (action == BTRFS_UPDATE_DELAYED_HEAD)
@@ -585,12 +564,11 @@ add_delayed_ref_head(struct btrfs_fs_info *fs_info,
count_mod = -1;
/*
- * BTRFS_ADD_DELAYED_EXTENT means that we need to update
- * the reserved accounting when the extent is finally added, or
- * if a later modification deletes the delayed ref without ever
- * inserting the extent into the extent allocation tree.
- * ref->must_insert_reserved is the flag used to record
- * that accounting mods are required.
+ * BTRFS_ADD_DELAYED_EXTENT means that we need to update the reserved
+ * accounting when the extent is finally added, or if a later
+ * modification deletes the delayed ref without ever inserting the
+ * extent into the extent allocation tree. ref->must_insert_reserved
+ * is the flag used to record that accounting mods are required.
*
* Once we record must_insert_reserved, switch the action to
* BTRFS_ADD_DELAYED_REF because other special casing is not required.
@@ -600,29 +578,21 @@ add_delayed_ref_head(struct btrfs_fs_info *fs_info,
else
must_insert_reserved = 0;
- delayed_refs = &trans->transaction->delayed_refs;
-
- /* first set the basic ref node struct up */
- atomic_set(&ref->refs, 1);
- ref->bytenr = bytenr;
- ref->num_bytes = num_bytes;
- ref->ref_mod = count_mod;
- ref->type = 0;
- ref->action = 0;
- ref->is_head = 1;
- ref->in_tree = 1;
- ref->seq = 0;
-
- head_ref = btrfs_delayed_node_to_head(ref);
+ atomic_set(&head_ref->refs, 1);
+ head_ref->bytenr = bytenr;
+ head_ref->num_bytes = num_bytes;
+ head_ref->ref_mod = count_mod;
head_ref->must_insert_reserved = must_insert_reserved;
head_ref->is_data = is_data;
head_ref->is_system = is_system;
head_ref->ref_tree = RB_ROOT;
INIT_LIST_HEAD(&head_ref->ref_add_list);
+ RB_CLEAR_NODE(&head_ref->href_node);
head_ref->processing = 0;
head_ref->total_ref_mod = count_mod;
+ spin_lock_init(&head_ref->lock);
+ mutex_init(&head_ref->mutex);
- /* Record qgroup extent info if provided */
if (qrecord) {
if (ref_root && reserved) {
qrecord->data_rsv = reserved;
@@ -632,21 +602,43 @@ add_delayed_ref_head(struct btrfs_fs_info *fs_info,
qrecord->bytenr = bytenr;
qrecord->num_bytes = num_bytes;
qrecord->old_roots = NULL;
+ }
+}
+
+/*
+ * helper function to actually insert a head node into the rbtree.
+ * this does all the dirty work in terms of maintaining the correct
+ * overall modification count.
+ */
+static noinline struct btrfs_delayed_ref_head *
+add_delayed_ref_head(struct btrfs_fs_info *fs_info,
+ struct btrfs_trans_handle *trans,
+ struct btrfs_delayed_ref_head *head_ref,
+ struct btrfs_qgroup_extent_record *qrecord,
+ int action, int *old_ref_mod,
+ int *new_ref_mod)
+
+{
+ struct btrfs_delayed_ref_head *existing;
+ struct btrfs_delayed_ref_root *delayed_refs;
+
+ delayed_refs = &trans->transaction->delayed_refs;
+
+ mutex_init(&head_ref->mutex);
+ /* Record qgroup extent info if provided */
+ if (qrecord) {
if(btrfs_qgroup_trace_extent_nolock(fs_info,
delayed_refs, qrecord))
kfree(qrecord);
}
- spin_lock_init(&head_ref->lock);
- mutex_init(&head_ref->mutex);
-
- trace_add_delayed_ref_head(fs_info, ref, head_ref, action);
+ trace_add_delayed_ref_head(fs_info, head_ref, action);
existing = htree_insert(&delayed_refs->href_root,
&head_ref->href_node);
if (existing) {
- update_existing_head_ref(delayed_refs, &existing->node, ref,
+ update_existing_head_ref(delayed_refs, existing, head_ref,
old_ref_mod);
/*
* we've updated the existing ref, free the newly
@@ -657,8 +649,8 @@ add_delayed_ref_head(struct btrfs_fs_info *fs_info,
} else {
if (old_ref_mod)
*old_ref_mod = 0;
- if (is_data && count_mod < 0)
- delayed_refs->pending_csums += num_bytes;
+ if (head_ref->is_data && head_ref->ref_mod < 0)
+ delayed_refs->pending_csums += head_ref->num_bytes;
delayed_refs->num_heads++;
delayed_refs->num_heads_ready++;
atomic_inc(&delayed_refs->num_entries);
@@ -666,90 +658,48 @@ add_delayed_ref_head(struct btrfs_fs_info *fs_info,
}
if (new_ref_mod)
*new_ref_mod = head_ref->total_ref_mod;
- return head_ref;
-}
-
-/*
- * helper to insert a delayed tree ref into the rbtree.
- */
-static noinline void
-add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
- struct btrfs_trans_handle *trans,
- 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)
-{
- struct btrfs_delayed_tree_ref *full_ref;
- struct btrfs_delayed_ref_root *delayed_refs;
- u64 seq = 0;
- int ret;
-
- 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 */
- atomic_set(&ref->refs, 1);
- ref->bytenr = bytenr;
- ref->num_bytes = num_bytes;
- ref->ref_mod = 1;
- ref->action = action;
- ref->is_head = 0;
- ref->in_tree = 1;
- ref->seq = seq;
- RB_CLEAR_NODE(&ref->ref_node);
- INIT_LIST_HEAD(&ref->add_list);
-
- full_ref = btrfs_delayed_node_to_tree_ref(ref);
- full_ref->parent = parent;
- full_ref->root = ref_root;
- if (parent)
- ref->type = BTRFS_SHARED_BLOCK_REF_KEY;
- else
- ref->type = BTRFS_TREE_BLOCK_REF_KEY;
- full_ref->level = level;
-
- trace_add_delayed_tree_ref(fs_info, ref, full_ref, action);
- ret = insert_delayed_ref(trans, delayed_refs, head_ref, ref);
-
- /*
- * XXX: memory should be freed at the same level allocated.
- * But bad practice is anywhere... Follow it now. Need cleanup.
- */
- if (ret > 0)
- kmem_cache_free(btrfs_delayed_tree_ref_cachep, full_ref);
+ return head_ref;
}
/*
- * helper to insert a delayed data ref into the rbtree.
+ * init_delayed_ref_common - Initialize the structure which represents a
+ * modification to a an extent.
+ *
+ * @fs_info: Internal to the mounted filesystem mount structure.
+ *
+ * @ref: The structure which is going to be initialized.
+ *
+ * @bytenr: The logical address of the extent for which a modification is
+ * going to be recorded.
+ *
+ * @num_bytes: Size of the extent whose modification is being recorded.
+ *
+ * @ref_root: The id of the root where this modification has originated, this
+ * can be either one of the well-known metadata trees or the
+ * subvolume id which references this extent.
+ *
+ * @action: Can be one of BTRFS_ADD_DELAYED_REF/BTRFS_DROP_DELAYED_REF or
+ * BTRFS_ADD_DELAYED_EXTENT
+ *
+ * @ref_type: Holds the type of the extent which is being recorded, can be
+ * one of BTRFS_SHARED_BLOCK_REF_KEY/BTRFS_TREE_BLOCK_REF_KEY
+ * when recording a metadata extent or BTRFS_SHARED_DATA_REF_KEY/
+ * BTRFS_EXTENT_DATA_REF_KEY when recording data extent
*/
-static noinline void
-add_delayed_data_ref(struct btrfs_fs_info *fs_info,
- struct btrfs_trans_handle *trans,
- 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)
+static void init_delayed_ref_common(struct btrfs_fs_info *fs_info,
+ struct btrfs_delayed_ref_node *ref,
+ u64 bytenr, u64 num_bytes, u64 ref_root,
+ int action, u8 ref_type)
{
- struct btrfs_delayed_data_ref *full_ref;
- struct btrfs_delayed_ref_root *delayed_refs;
u64 seq = 0;
- int ret;
if (action == BTRFS_ADD_DELAYED_EXTENT)
action = BTRFS_ADD_DELAYED_REF;
- 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;
ref->num_bytes = num_bytes;
@@ -758,25 +708,9 @@ add_delayed_data_ref(struct btrfs_fs_info *fs_info,
ref->is_head = 0;
ref->in_tree = 1;
ref->seq = seq;
+ ref->type = ref_type;
RB_CLEAR_NODE(&ref->ref_node);
INIT_LIST_HEAD(&ref->add_list);
-
- full_ref = btrfs_delayed_node_to_data_ref(ref);
- full_ref->parent = parent;
- full_ref->root = ref_root;
- if (parent)
- ref->type = BTRFS_SHARED_DATA_REF_KEY;
- else
- ref->type = BTRFS_EXTENT_DATA_REF_KEY;
-
- full_ref->objectid = owner;
- full_ref->offset = offset;
-
- trace_add_delayed_data_ref(fs_info, ref, full_ref, action);
-
- ret = insert_delayed_ref(trans, delayed_refs, head_ref, ref);
- if (ret > 0)
- kmem_cache_free(btrfs_delayed_data_ref_cachep, full_ref);
}
/*
@@ -795,13 +729,25 @@ int btrfs_add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
struct btrfs_delayed_ref_head *head_ref;
struct btrfs_delayed_ref_root *delayed_refs;
struct btrfs_qgroup_extent_record *record = NULL;
- int is_system = (ref_root == BTRFS_CHUNK_TREE_OBJECTID);
+ bool is_system = (ref_root == BTRFS_CHUNK_TREE_OBJECTID);
+ int ret;
+ u8 ref_type;
BUG_ON(extent_op && extent_op->is_data);
ref = kmem_cache_alloc(btrfs_delayed_tree_ref_cachep, GFP_NOFS);
if (!ref)
return -ENOMEM;
+ if (parent)
+ ref_type = BTRFS_SHARED_BLOCK_REF_KEY;
+ else
+ ref_type = BTRFS_TREE_BLOCK_REF_KEY;
+ init_delayed_ref_common(fs_info, &ref->node, bytenr, num_bytes,
+ ref_root, action, ref_type);
+ ref->root = ref_root;
+ ref->parent = parent;
+ ref->level = level;
+
head_ref = kmem_cache_alloc(btrfs_delayed_ref_head_cachep, GFP_NOFS);
if (!head_ref)
goto free_ref;
@@ -813,6 +759,8 @@ int btrfs_add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
goto free_head_ref;
}
+ init_delayed_ref_head(head_ref, record, bytenr, num_bytes,
+ ref_root, 0, action, false, is_system);
head_ref->extent_op = extent_op;
delayed_refs = &trans->transaction->delayed_refs;
@@ -822,14 +770,18 @@ int btrfs_add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
* insert both the head node and the new ref without dropping
* the spin lock
*/
- head_ref = add_delayed_ref_head(fs_info, trans, &head_ref->node, record,
- bytenr, num_bytes, 0, 0, action, 0,
- is_system, old_ref_mod, new_ref_mod);
+ head_ref = add_delayed_ref_head(fs_info, trans, head_ref, record,
+ action, old_ref_mod, new_ref_mod);
- add_delayed_tree_ref(fs_info, trans, head_ref, &ref->node, bytenr,
- num_bytes, parent, ref_root, level, action);
+ ret = insert_delayed_ref(trans, delayed_refs, head_ref, &ref->node);
spin_unlock(&delayed_refs->lock);
+ trace_add_delayed_tree_ref(fs_info, &ref->node, ref,
+ action == BTRFS_ADD_DELAYED_EXTENT ?
+ BTRFS_ADD_DELAYED_REF : action);
+ if (ret > 0)
+ kmem_cache_free(btrfs_delayed_tree_ref_cachep, ref);
+
return 0;
free_head_ref:
@@ -855,12 +807,25 @@ int btrfs_add_delayed_data_ref(struct btrfs_fs_info *fs_info,
struct btrfs_delayed_ref_head *head_ref;
struct btrfs_delayed_ref_root *delayed_refs;
struct btrfs_qgroup_extent_record *record = NULL;
+ int ret;
+ u8 ref_type;
BUG_ON(extent_op && !extent_op->is_data);
ref = kmem_cache_alloc(btrfs_delayed_data_ref_cachep, GFP_NOFS);
if (!ref)
return -ENOMEM;
+ if (parent)
+ ref_type = BTRFS_SHARED_DATA_REF_KEY;
+ else
+ ref_type = BTRFS_EXTENT_DATA_REF_KEY;
+ init_delayed_ref_common(fs_info, &ref->node, bytenr, num_bytes,
+ ref_root, action, ref_type);
+ ref->root = ref_root;
+ ref->parent = parent;
+ ref->objectid = owner;
+ ref->offset = offset;
+
head_ref = kmem_cache_alloc(btrfs_delayed_ref_head_cachep, GFP_NOFS);
if (!head_ref) {
kmem_cache_free(btrfs_delayed_data_ref_cachep, ref);
@@ -878,6 +843,8 @@ int btrfs_add_delayed_data_ref(struct btrfs_fs_info *fs_info,
}
}
+ init_delayed_ref_head(head_ref, record, bytenr, num_bytes, ref_root,
+ reserved, action, true, false);
head_ref->extent_op = extent_op;
delayed_refs = &trans->transaction->delayed_refs;
@@ -887,15 +854,17 @@ int btrfs_add_delayed_data_ref(struct btrfs_fs_info *fs_info,
* insert both the head node and the new ref without dropping
* the spin lock
*/
- head_ref = add_delayed_ref_head(fs_info, trans, &head_ref->node, record,
- bytenr, num_bytes, ref_root, reserved,
- action, 1, 0, old_ref_mod, new_ref_mod);
+ head_ref = add_delayed_ref_head(fs_info, trans, head_ref, record,
+ action, old_ref_mod, new_ref_mod);
- add_delayed_data_ref(fs_info, trans, head_ref, &ref->node, bytenr,
- num_bytes, parent, ref_root, owner, offset,
- action);
+ ret = insert_delayed_ref(trans, delayed_refs, head_ref, &ref->node);
spin_unlock(&delayed_refs->lock);
+ trace_add_delayed_data_ref(trans->fs_info, &ref->node, ref,
+ action == BTRFS_ADD_DELAYED_EXTENT ?
+ BTRFS_ADD_DELAYED_REF : action);
+ if (ret > 0)
+ kmem_cache_free(btrfs_delayed_data_ref_cachep, ref);
return 0;
}
@@ -911,19 +880,16 @@ int btrfs_add_delayed_extent_op(struct btrfs_fs_info *fs_info,
if (!head_ref)
return -ENOMEM;
+ init_delayed_ref_head(head_ref, NULL, bytenr, num_bytes, 0, 0,
+ BTRFS_UPDATE_DELAYED_HEAD, extent_op->is_data,
+ false);
head_ref->extent_op = extent_op;
delayed_refs = &trans->transaction->delayed_refs;
spin_lock(&delayed_refs->lock);
- /*
- * extent_ops just modify the flags of an extent and they don't result
- * in ref count changes, hence it's safe to pass false/0 for is_system
- * argument
- */
- add_delayed_ref_head(fs_info, trans, &head_ref->node, NULL, bytenr,
- num_bytes, 0, 0, BTRFS_UPDATE_DELAYED_HEAD,
- extent_op->is_data, 0, NULL, NULL);
+ add_delayed_ref_head(fs_info, trans, head_ref, NULL,
+ BTRFS_UPDATE_DELAYED_HEAD, NULL, NULL);
spin_unlock(&delayed_refs->lock);
return 0;
diff --git a/fs/btrfs/delayed-ref.h b/fs/btrfs/delayed-ref.h
index 13e915dc90bc..ec349bddf0ff 100644
--- a/fs/btrfs/delayed-ref.h
+++ b/fs/btrfs/delayed-ref.h
@@ -24,15 +24,6 @@
#define BTRFS_ADD_DELAYED_EXTENT 3 /* record a full extent allocation */
#define BTRFS_UPDATE_DELAYED_HEAD 4 /* not changing ref count on head ref */
-/*
- * XXX: Qu: I really hate the design that ref_head and tree/data ref shares the
- * same ref_node structure.
- * Ref_head is in a higher logic level than tree/data ref, and duplicated
- * bytenr/num_bytes in ref_node is really a waste or memory, they should be
- * referred from ref_head.
- * This gets more disgusting after we use list to store tree/data ref in
- * ref_head. Must clean this mess up later.
- */
struct btrfs_delayed_ref_node {
/*
* ref_head use rb tree, stored in ref_root->href.
@@ -94,8 +85,9 @@ struct btrfs_delayed_extent_op {
* reference count modifications we've queued up.
*/
struct btrfs_delayed_ref_head {
- struct btrfs_delayed_ref_node node;
-
+ u64 bytenr;
+ u64 num_bytes;
+ atomic_t refs;
/*
* the mutex is held while running the refs, and it is also
* held when checking the sum of reference modifications.
@@ -119,6 +111,14 @@ struct btrfs_delayed_ref_head {
int total_ref_mod;
/*
+ * This is the current outstanding mod references for this bytenr. This
+ * is used with lookup_extent_info to get an accurate reference count
+ * for a bytenr, so it is adjusted as delayed refs are run so that any
+ * on disk reference count + ref_mod is accurate.
+ */
+ int ref_mod;
+
+ /*
* when a new extent is allocated, it is just reserved in memory
* The actual extent isn't inserted into the extent allocation tree
* until the delayed ref is processed. must_insert_reserved is
@@ -227,15 +227,18 @@ static inline void btrfs_put_delayed_ref(struct btrfs_delayed_ref_node *ref)
case BTRFS_SHARED_DATA_REF_KEY:
kmem_cache_free(btrfs_delayed_data_ref_cachep, ref);
break;
- case 0:
- kmem_cache_free(btrfs_delayed_ref_head_cachep, ref);
- break;
default:
BUG();
}
}
}
+static inline void btrfs_put_delayed_ref_head(struct btrfs_delayed_ref_head *head)
+{
+ if (atomic_dec_and_test(&head->refs))
+ kmem_cache_free(btrfs_delayed_ref_head_cachep, head);
+}
+
int btrfs_add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
struct btrfs_trans_handle *trans,
u64 bytenr, u64 num_bytes, u64 parent,
@@ -280,35 +283,17 @@ int btrfs_check_delayed_seq(struct btrfs_fs_info *fs_info,
u64 seq);
/*
- * a node might live in a head or a regular ref, this lets you
- * test for the proper type to use.
- */
-static int btrfs_delayed_ref_is_head(struct btrfs_delayed_ref_node *node)
-{
- return node->is_head;
-}
-
-/*
* helper functions to cast a node into its container
*/
static inline struct btrfs_delayed_tree_ref *
btrfs_delayed_node_to_tree_ref(struct btrfs_delayed_ref_node *node)
{
- WARN_ON(btrfs_delayed_ref_is_head(node));
return container_of(node, struct btrfs_delayed_tree_ref, node);
}
static inline struct btrfs_delayed_data_ref *
btrfs_delayed_node_to_data_ref(struct btrfs_delayed_ref_node *node)
{
- WARN_ON(btrfs_delayed_ref_is_head(node));
return container_of(node, struct btrfs_delayed_data_ref, node);
}
-
-static inline struct btrfs_delayed_ref_head *
-btrfs_delayed_node_to_head(struct btrfs_delayed_ref_node *node)
-{
- WARN_ON(!btrfs_delayed_ref_is_head(node));
- return container_of(node, struct btrfs_delayed_ref_head, node);
-}
#endif
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 7da000f89806..c8029bda308c 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -4268,12 +4268,12 @@ static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
head = rb_entry(node, struct btrfs_delayed_ref_head,
href_node);
if (!mutex_trylock(&head->mutex)) {
- atomic_inc(&head->node.refs);
+ atomic_inc(&head->refs);
spin_unlock(&delayed_refs->lock);
mutex_lock(&head->mutex);
mutex_unlock(&head->mutex);
- btrfs_put_delayed_ref(&head->node);
+ btrfs_put_delayed_ref_head(head);
spin_lock(&delayed_refs->lock);
continue;
}
@@ -4296,16 +4296,15 @@ static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
if (head->processing == 0)
delayed_refs->num_heads_ready--;
atomic_dec(&delayed_refs->num_entries);
- head->node.in_tree = 0;
rb_erase(&head->href_node, &delayed_refs->href_root);
spin_unlock(&head->lock);
spin_unlock(&delayed_refs->lock);
mutex_unlock(&head->mutex);
if (pin_bytes)
- btrfs_pin_extent(root, head->node.bytenr,
- head->node.num_bytes, 1);
- btrfs_put_delayed_ref(&head->node);
+ btrfs_pin_extent(root, head->bytenr,
+ head->num_bytes, 1);
+ btrfs_put_delayed_ref_head(head);
cond_resched();
spin_lock(&delayed_refs->lock);
}
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 84f8a2409050..b016c070141e 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -913,7 +913,7 @@ search_again:
head = btrfs_find_delayed_ref_head(delayed_refs, bytenr);
if (head) {
if (!mutex_trylock(&head->mutex)) {
- atomic_inc(&head->node.refs);
+ atomic_inc(&head->refs);
spin_unlock(&delayed_refs->lock);
btrfs_release_path(path);
@@ -924,7 +924,7 @@ search_again:
*/
mutex_lock(&head->mutex);
mutex_unlock(&head->mutex);
- btrfs_put_delayed_ref(&head->node);
+ btrfs_put_delayed_ref_head(head);
goto search_again;
}
spin_lock(&head->lock);
@@ -933,7 +933,7 @@ search_again:
else
BUG_ON(num_refs == 0);
- num_refs += head->node.ref_mod;
+ num_refs += head->ref_mod;
spin_unlock(&head->lock);
mutex_unlock(&head->mutex);
}
@@ -2250,7 +2250,7 @@ static void __run_delayed_extent_op(struct btrfs_delayed_extent_op *extent_op,
static int run_delayed_extent_op(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
- struct btrfs_delayed_ref_node *node,
+ struct btrfs_delayed_ref_head *head,
struct btrfs_delayed_extent_op *extent_op)
{
struct btrfs_key key;
@@ -2272,14 +2272,14 @@ static int run_delayed_extent_op(struct btrfs_trans_handle *trans,
if (!path)
return -ENOMEM;
- key.objectid = node->bytenr;
+ key.objectid = head->bytenr;
if (metadata) {
key.type = BTRFS_METADATA_ITEM_KEY;
key.offset = extent_op->level;
} else {
key.type = BTRFS_EXTENT_ITEM_KEY;
- key.offset = node->num_bytes;
+ key.offset = head->num_bytes;
}
again:
@@ -2297,17 +2297,17 @@ again:
path->slots[0]--;
btrfs_item_key_to_cpu(path->nodes[0], &key,
path->slots[0]);
- if (key.objectid == node->bytenr &&
+ if (key.objectid == head->bytenr &&
key.type == BTRFS_EXTENT_ITEM_KEY &&
- key.offset == node->num_bytes)
+ key.offset == head->num_bytes)
ret = 0;
}
if (ret > 0) {
btrfs_release_path(path);
metadata = 0;
- key.objectid = node->bytenr;
- key.offset = node->num_bytes;
+ key.objectid = head->bytenr;
+ key.offset = head->num_bytes;
key.type = BTRFS_EXTENT_ITEM_KEY;
goto again;
}
@@ -2478,7 +2478,7 @@ static int cleanup_extent_op(struct btrfs_trans_handle *trans,
return 0;
}
spin_unlock(&head->lock);
- ret = run_delayed_extent_op(trans, fs_info->extent_root, &head->node,
+ ret = run_delayed_extent_op(trans, fs_info->extent_root, head,
extent_op);
btrfs_free_delayed_extent_op(extent_op);
return ret ? ret : 1;
@@ -2514,20 +2514,19 @@ static int cleanup_ref_head(struct btrfs_trans_handle *trans,
spin_unlock(&delayed_refs->lock);
return 1;
}
- head->node.in_tree = 0;
delayed_refs->num_heads--;
rb_erase(&head->href_node, &delayed_refs->href_root);
+ RB_CLEAR_NODE(&head->href_node);
spin_unlock(&delayed_refs->lock);
spin_unlock(&head->lock);
atomic_dec(&delayed_refs->num_entries);
- trace_run_delayed_ref_head(fs_info, &head->node, head,
- head->node.action);
+ trace_run_delayed_ref_head(fs_info, head, 0);
if (head->total_ref_mod < 0) {
struct btrfs_space_info *space_info;
u64 flags;
-
+
if (head->is_data)
flags = BTRFS_BLOCK_GROUP_DATA;
else if (head->is_system)
@@ -2537,26 +2536,27 @@ static int cleanup_ref_head(struct btrfs_trans_handle *trans,
space_info = __find_space_info(fs_info, flags);
ASSERT(space_info);
percpu_counter_add(&space_info->total_bytes_pinned,
- -head->node.num_bytes);
+ -head->num_bytes);
+
if (head->is_data) {
spin_lock(&delayed_refs->lock);
- delayed_refs->pending_csums -= head->node.num_bytes;
+ delayed_refs->pending_csums -= head->num_bytes;
spin_unlock(&delayed_refs->lock);
}
}
if (head->must_insert_reserved) {
- btrfs_pin_extent(fs_info->extent_root, head->node.bytenr,
- head->node.num_bytes, 1);
+ btrfs_pin_extent(fs_info->extent_root, head->bytenr,
+ head->num_bytes, 1);
if (head->is_data) {
ret = btrfs_del_csums(trans, fs_info->csum_root,
- head->node.bytenr,
- head->node.num_bytes);
+ head->bytenr,
+ head->num_bytes);
}
}
btrfs_delayed_ref_unlock(head);
- btrfs_put_delayed_ref(&head->node);
+ btrfs_put_delayed_ref_head(head);
return 0;
}
@@ -2672,10 +2672,10 @@ static noinline int __btrfs_run_delayed_refs(struct btrfs_trans_handle *trans,
switch (ref->action) {
case BTRFS_ADD_DELAYED_REF:
case BTRFS_ADD_DELAYED_EXTENT:
- locked_ref->node.ref_mod -= ref->ref_mod;
+ locked_ref->ref_mod -= ref->ref_mod;
break;
case BTRFS_DROP_DELAYED_REF:
- locked_ref->node.ref_mod += ref->ref_mod;
+ locked_ref->ref_mod += ref->ref_mod;
break;
default:
WARN_ON(1);
@@ -3005,32 +3005,14 @@ again:
}
count = (unsigned long)-1;
- while (node) {
- head = rb_entry(node, struct btrfs_delayed_ref_head,
- href_node);
- if (btrfs_delayed_ref_is_head(&head->node)) {
- struct btrfs_delayed_ref_node *ref;
-
- ref = &head->node;
- atomic_inc(&ref->refs);
-
- spin_unlock(&delayed_refs->lock);
- /*
- * Mutex was contended, block until it's
- * released and try again
- */
- mutex_lock(&head->mutex);
- mutex_unlock(&head->mutex);
-
- btrfs_put_delayed_ref(ref);
- cond_resched();
- goto again;
- } else {
- WARN_ON(1);
- }
- node = rb_next(node);
- }
+ head = rb_entry(node, struct btrfs_delayed_ref_head,
+ href_node);
+ atomic_inc(&head->refs);
spin_unlock(&delayed_refs->lock);
+
+ /* Mutex was contended, block until it's released and retry. */
+ mutex_lock(&head->mutex);
+ mutex_unlock(&head->mutex);
cond_resched();
goto again;
}
@@ -3095,7 +3077,7 @@ static noinline int check_delayed_ref(struct btrfs_root *root,
}
if (!mutex_trylock(&head->mutex)) {
- atomic_inc(&head->node.refs);
+ atomic_inc(&head->refs);
spin_unlock(&delayed_refs->lock);
btrfs_release_path(path);
@@ -3106,7 +3088,7 @@ static noinline int check_delayed_ref(struct btrfs_root *root,
*/
mutex_lock(&head->mutex);
mutex_unlock(&head->mutex);
- btrfs_put_delayed_ref(&head->node);
+ btrfs_put_delayed_ref_head(head);
btrfs_put_transaction(cur_trans);
return -EAGAIN;
}
@@ -7156,9 +7138,8 @@ static noinline int check_ref_cleanup(struct btrfs_trans_handle *trans,
* at this point we have a head with no other entries. Go
* ahead and process it.
*/
- head->node.in_tree = 0;
rb_erase(&head->href_node, &delayed_refs->href_root);
-
+ RB_CLEAR_NODE(&head->href_node);
atomic_dec(&delayed_refs->num_entries);
/*
@@ -7177,7 +7158,7 @@ static noinline int check_ref_cleanup(struct btrfs_trans_handle *trans,
ret = 1;
mutex_unlock(&head->mutex);
- btrfs_put_delayed_ref(&head->node);
+ btrfs_put_delayed_ref_head(head);
return ret;
out:
spin_unlock(&head->lock);
diff --git a/include/trace/events/btrfs.h b/include/trace/events/btrfs.h
index 8db632e0af60..3e144480cabd 100644
--- a/include/trace/events/btrfs.h
+++ b/include/trace/events/btrfs.h
@@ -649,11 +649,10 @@ DEFINE_EVENT(btrfs_delayed_data_ref, run_delayed_data_ref,
DECLARE_EVENT_CLASS(btrfs_delayed_ref_head,
TP_PROTO(const struct btrfs_fs_info *fs_info,
- const struct btrfs_delayed_ref_node *ref,
const struct btrfs_delayed_ref_head *head_ref,
int action),
- TP_ARGS(fs_info, ref, head_ref, action),
+ TP_ARGS(fs_info, head_ref, action),
TP_STRUCT__entry_btrfs(
__field( u64, bytenr )
@@ -663,8 +662,8 @@ DECLARE_EVENT_CLASS(btrfs_delayed_ref_head,
),
TP_fast_assign_btrfs(fs_info,
- __entry->bytenr = ref->bytenr;
- __entry->num_bytes = ref->num_bytes;
+ __entry->bytenr = head_ref->bytenr;
+ __entry->num_bytes = head_ref->num_bytes;
__entry->action = action;
__entry->is_data = head_ref->is_data;
),
@@ -679,21 +678,19 @@ DECLARE_EVENT_CLASS(btrfs_delayed_ref_head,
DEFINE_EVENT(btrfs_delayed_ref_head, add_delayed_ref_head,
TP_PROTO(const struct btrfs_fs_info *fs_info,
- const struct btrfs_delayed_ref_node *ref,
const struct btrfs_delayed_ref_head *head_ref,
int action),
- TP_ARGS(fs_info, ref, head_ref, action)
+ TP_ARGS(fs_info, head_ref, action)
);
DEFINE_EVENT(btrfs_delayed_ref_head, run_delayed_ref_head,
TP_PROTO(const struct btrfs_fs_info *fs_info,
- const struct btrfs_delayed_ref_node *ref,
const struct btrfs_delayed_ref_head *head_ref,
int action),
- TP_ARGS(fs_info, ref, head_ref, action)
+ TP_ARGS(fs_info, head_ref, action)
);
#define show_chunk_type(type) \