#include "disk-io.h"
#include "transaction.h"
#include "print-tree.h"
+#include "locking.h"
static int split_node(struct btrfs_trans_handle *trans, struct btrfs_root
*root, struct btrfs_path *path, int level);
kmem_cache_free(btrfs_path_cachep, p);
}
-void btrfs_release_path(struct btrfs_root *root, struct btrfs_path *p)
+void noinline btrfs_release_path(struct btrfs_root *root, struct btrfs_path *p)
{
int i;
+
for (i = 0; i < BTRFS_MAX_LEVEL; i++) {
+ p->slots[i] = 0;
if (!p->nodes[i])
- break;
+ continue;
+ if (p->locks[i]) {
+ btrfs_tree_unlock(p->nodes[i]);
+ p->locks[i] = 0;
+ }
free_extent_buffer(p->nodes[i]);
+ p->nodes[i] = NULL;
}
- memset(p, 0, sizeof(*p));
+}
+
+struct extent_buffer *btrfs_root_node(struct btrfs_root *root)
+{
+ struct extent_buffer *eb;
+ spin_lock(&root->node_lock);
+ eb = root->node;
+ extent_buffer_get(eb);
+ spin_unlock(&root->node_lock);
+ return eb;
+}
+
+struct extent_buffer *btrfs_lock_root_node(struct btrfs_root *root)
+{
+ struct extent_buffer *eb;
+
+ while(1) {
+ eb = btrfs_root_node(root);
+ btrfs_tree_lock(eb);
+
+ spin_lock(&root->node_lock);
+ if (eb == root->node) {
+ spin_unlock(&root->node_lock);
+ break;
+ }
+ spin_unlock(&root->node_lock);
+
+ btrfs_tree_unlock(eb);
+ free_extent_buffer(eb);
+ }
+ return eb;
}
static void add_root_to_dirty_list(struct btrfs_root *root)
} else {
first_key.objectid = 0;
}
- cow = __btrfs_alloc_free_block(trans, new_root, buf->len,
+ cow = btrfs_alloc_free_block(trans, new_root, buf->len,
new_root_objectid,
trans->transid, first_key.objectid,
level, buf->start, 0);
btrfs_clear_header_flag(cow, BTRFS_HEADER_FLAG_WRITTEN);
WARN_ON(btrfs_header_generation(buf) > trans->transid);
- ret = btrfs_inc_ref(trans, new_root, buf);
+ ret = btrfs_inc_ref(trans, new_root, buf, 0);
kfree(new_root);
if (ret)
return 0;
}
-int __btrfs_cow_block(struct btrfs_trans_handle *trans,
+int noinline __btrfs_cow_block(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct extent_buffer *buf,
struct extent_buffer *parent, int parent_slot,
struct extent_buffer **cow_ret,
- u64 search_start, u64 empty_size)
+ u64 search_start, u64 empty_size,
+ u64 prealloc_dest)
{
u64 root_gen;
struct extent_buffer *cow;
int ret = 0;
int different_trans = 0;
int level;
+ int unlock_orig = 0;
struct btrfs_key first_key;
+ if (*cow_ret == buf)
+ unlock_orig = 1;
+
+ WARN_ON(!btrfs_tree_locked(buf));
+
if (root->ref_cows) {
root_gen = trans->transid;
} else {
} else {
first_key.objectid = 0;
}
- cow = __btrfs_alloc_free_block(trans, root, buf->len,
- root->root_key.objectid,
- root_gen, first_key.objectid, level,
- search_start, empty_size);
+ if (prealloc_dest) {
+ struct btrfs_key ins;
+
+ ins.objectid = prealloc_dest;
+ ins.offset = buf->len;
+ ins.type = BTRFS_EXTENT_ITEM_KEY;
+
+ ret = btrfs_alloc_reserved_extent(trans, root,
+ root->root_key.objectid,
+ root_gen, level,
+ first_key.objectid,
+ &ins);
+ BUG_ON(ret);
+ cow = btrfs_init_new_buffer(trans, root, prealloc_dest,
+ buf->len);
+ } else {
+ cow = btrfs_alloc_free_block(trans, root, buf->len,
+ root->root_key.objectid,
+ root_gen, first_key.objectid,
+ level, search_start, empty_size);
+ }
if (IS_ERR(cow))
return PTR_ERR(cow);
WARN_ON(btrfs_header_generation(buf) > trans->transid);
if (btrfs_header_generation(buf) != trans->transid) {
different_trans = 1;
- ret = btrfs_inc_ref(trans, root, buf);
+ ret = btrfs_inc_ref(trans, root, buf, 1);
if (ret)
return ret;
} else {
}
if (buf == root->node) {
+ WARN_ON(parent && parent != buf);
root_gen = btrfs_header_generation(buf);
+
+ spin_lock(&root->node_lock);
root->node = cow;
extent_buffer_get(cow);
+ spin_unlock(&root->node_lock);
+
if (buf != root->commit_root) {
btrfs_free_extent(trans, root, buf->start,
buf->len, root->root_key.objectid,
btrfs_header_owner(parent), root_gen,
0, 0, 1);
}
+ if (unlock_orig)
+ btrfs_tree_unlock(buf);
free_extent_buffer(buf);
btrfs_mark_buffer_dirty(cow);
*cow_ret = cow;
return 0;
}
-int btrfs_cow_block(struct btrfs_trans_handle *trans,
+int noinline btrfs_cow_block(struct btrfs_trans_handle *trans,
struct btrfs_root *root, struct extent_buffer *buf,
struct extent_buffer *parent, int parent_slot,
- struct extent_buffer **cow_ret)
+ struct extent_buffer **cow_ret, u64 prealloc_dest)
{
u64 search_start;
u64 header_trans;
!btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) {
*cow_ret = buf;
spin_unlock(&root->fs_info->hash_lock);
+ WARN_ON(prealloc_dest);
return 0;
}
spin_unlock(&root->fs_info->hash_lock);
search_start = buf->start & ~((u64)(1024 * 1024 * 1024) - 1);
ret = __btrfs_cow_block(trans, root, buf, parent,
- parent_slot, cow_ret, search_start, 0);
+ parent_slot, cow_ret, search_start, 0,
+ prealloc_dest);
return ret;
}
struct btrfs_key *progress)
{
struct extent_buffer *cur;
- struct extent_buffer *tmp;
u64 blocknr;
u64 gen;
u64 search_start = *last_ret;
other = btrfs_node_blockptr(parent, i - 1);
close = close_blocks(blocknr, other, blocksize);
}
- if (close && i < end_slot - 2) {
+ if (!close && i < end_slot - 2) {
other = btrfs_node_blockptr(parent, i + 1);
close = close_blocks(blocknr, other, blocksize);
}
cur = btrfs_find_tree_block(root, blocknr, blocksize);
if (cur)
- uptodate = btrfs_buffer_uptodate(cur);
+ uptodate = btrfs_buffer_uptodate(cur, gen);
else
uptodate = 0;
if (!cur || !uptodate) {
if (search_start == 0)
search_start = last_block;
+ btrfs_tree_lock(cur);
err = __btrfs_cow_block(trans, root, cur, parent, i,
- &tmp, search_start,
+ &cur, search_start,
min(16 * blocksize,
- (end_slot - i) * blocksize));
+ (end_slot - i) * blocksize), 0);
if (err) {
+ btrfs_tree_unlock(cur);
free_extent_buffer(cur);
break;
}
- search_start = tmp->start;
- last_block = tmp->start;
+ search_start = cur->start;
+ last_block = cur->start;
*last_ret = search_start;
- if (parent_level == 1)
- btrfs_clear_buffer_defrag(tmp);
- free_extent_buffer(tmp);
+ btrfs_tree_unlock(cur);
+ free_extent_buffer(cur);
}
if (parent->map_token) {
unmap_extent_buffer(parent, parent->map_token,
*
* slot may point to max if the key is bigger than all of the keys
*/
-static int generic_bin_search(struct extent_buffer *eb, unsigned long p,
- int item_size, struct btrfs_key *key,
- int max, int *slot)
+static noinline int generic_bin_search(struct extent_buffer *eb,
+ unsigned long p,
+ int item_size, struct btrfs_key *key,
+ int max, int *slot)
{
int low = 0;
int high = max;
return -1;
}
-static struct extent_buffer *read_node_slot(struct btrfs_root *root,
+static noinline struct extent_buffer *read_node_slot(struct btrfs_root *root,
struct extent_buffer *parent, int slot)
{
int level = btrfs_header_level(parent);
btrfs_node_ptr_generation(parent, slot));
}
-static int balance_level(struct btrfs_trans_handle *trans,
+static noinline int balance_level(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_path *path, int level)
{
return 0;
mid = path->nodes[level];
+ WARN_ON(!path->locks[level]);
WARN_ON(btrfs_header_generation(mid) != trans->transid);
orig_ptr = btrfs_node_blockptr(mid, orig_slot);
/* promote the child to a root */
child = read_node_slot(root, mid, 0);
+ btrfs_tree_lock(child);
BUG_ON(!child);
- ret = btrfs_cow_block(trans, root, child, mid, 0, &child);
+ ret = btrfs_cow_block(trans, root, child, mid, 0, &child, 0);
BUG_ON(ret);
+ spin_lock(&root->node_lock);
root->node = child;
+ spin_unlock(&root->node_lock);
+
add_root_to_dirty_list(root);
+ btrfs_tree_unlock(child);
+ path->locks[level] = 0;
path->nodes[level] = NULL;
clean_tree_block(trans, root, mid);
+ btrfs_tree_unlock(mid);
/* once for the path */
free_extent_buffer(mid);
ret = btrfs_free_extent(trans, root, mid->start, mid->len,
left = read_node_slot(root, parent, pslot - 1);
if (left) {
+ btrfs_tree_lock(left);
wret = btrfs_cow_block(trans, root, left,
- parent, pslot - 1, &left);
+ parent, pslot - 1, &left, 0);
if (wret) {
ret = wret;
goto enospc;
}
right = read_node_slot(root, parent, pslot + 1);
if (right) {
+ btrfs_tree_lock(right);
wret = btrfs_cow_block(trans, root, right,
- parent, pslot + 1, &right);
+ parent, pslot + 1, &right, 0);
if (wret) {
ret = wret;
goto enospc;
u32 blocksize = right->len;
clean_tree_block(trans, root, right);
+ btrfs_tree_unlock(right);
free_extent_buffer(right);
right = NULL;
wret = del_ptr(trans, root, path, level + 1, pslot +
u64 root_gen = btrfs_header_generation(parent);
u64 bytenr = mid->start;
u32 blocksize = mid->len;
+
clean_tree_block(trans, root, mid);
+ btrfs_tree_unlock(mid);
free_extent_buffer(mid);
mid = NULL;
wret = del_ptr(trans, root, path, level + 1, pslot);
if (left) {
if (btrfs_header_nritems(left) > orig_slot) {
extent_buffer_get(left);
+ /* left was locked after cow */
path->nodes[level] = left;
path->slots[level + 1] -= 1;
path->slots[level] = orig_slot;
- if (mid)
+ if (mid) {
+ btrfs_tree_unlock(mid);
free_extent_buffer(mid);
+ }
} else {
orig_slot -= btrfs_header_nritems(left);
path->slots[level] = orig_slot;
btrfs_node_blockptr(path->nodes[level], path->slots[level]))
BUG();
enospc:
- if (right)
+ if (right) {
+ btrfs_tree_unlock(right);
free_extent_buffer(right);
- if (left)
+ }
+ if (left) {
+ if (path->nodes[level] != left)
+ btrfs_tree_unlock(left);
free_extent_buffer(left);
+ }
return ret;
}
/* first, try to make some room in the middle buffer */
if (left) {
u32 left_nr;
+
+ btrfs_tree_lock(left);
left_nr = btrfs_header_nritems(left);
if (left_nr >= BTRFS_NODEPTRS_PER_BLOCK(root) - 1) {
wret = 1;
} else {
ret = btrfs_cow_block(trans, root, left, parent,
- pslot - 1, &left);
+ pslot - 1, &left, 0);
if (ret)
wret = 1;
else {
path->nodes[level] = left;
path->slots[level + 1] -= 1;
path->slots[level] = orig_slot;
+ btrfs_tree_unlock(mid);
free_extent_buffer(mid);
} else {
orig_slot -=
btrfs_header_nritems(left);
path->slots[level] = orig_slot;
+ btrfs_tree_unlock(left);
free_extent_buffer(left);
}
return 0;
}
+ btrfs_tree_unlock(left);
free_extent_buffer(left);
}
- right= read_node_slot(root, parent, pslot + 1);
+ right = read_node_slot(root, parent, pslot + 1);
/*
* then try to empty the right most buffer into the middle
*/
if (right) {
u32 right_nr;
+ btrfs_tree_lock(right);
right_nr = btrfs_header_nritems(right);
if (right_nr >= BTRFS_NODEPTRS_PER_BLOCK(root) - 1) {
wret = 1;
} else {
ret = btrfs_cow_block(trans, root, right,
parent, pslot + 1,
- &right);
+ &right, 0);
if (ret)
wret = 1;
else {
path->slots[level + 1] += 1;
path->slots[level] = orig_slot -
btrfs_header_nritems(mid);
+ btrfs_tree_unlock(mid);
free_extent_buffer(mid);
} else {
+ btrfs_tree_unlock(right);
free_extent_buffer(right);
}
return 0;
}
+ btrfs_tree_unlock(right);
free_extent_buffer(right);
}
return 1;
/*
* readahead one full node of leaves
*/
-static void reada_for_search(struct btrfs_root *root, struct btrfs_path *path,
- int level, int slot, u64 objectid)
+static noinline void reada_for_search(struct btrfs_root *root,
+ struct btrfs_path *path,
+ int level, int slot, u64 objectid)
{
struct extent_buffer *node;
struct btrfs_disk_key disk_key;
return;
node = path->nodes[level];
+
search = btrfs_node_blockptr(node, slot);
blocksize = btrfs_level_size(root, level - 1);
eb = btrfs_find_tree_block(root, search, blocksize);
highest_read = search;
}
}
+
+static noinline void unlock_up(struct btrfs_path *path, int level,
+ int lowest_unlock)
+{
+ int i;
+ int skip_level = level;
+ int no_skips = 0;
+ struct extent_buffer *t;
+
+ for (i = level; i < BTRFS_MAX_LEVEL; i++) {
+ if (!path->nodes[i])
+ break;
+ if (!path->locks[i])
+ break;
+ if (!no_skips && path->slots[i] == 0) {
+ skip_level = i + 1;
+ continue;
+ }
+ if (!no_skips && path->keep_locks) {
+ u32 nritems;
+ t = path->nodes[i];
+ nritems = btrfs_header_nritems(t);
+ if (nritems < 1 || path->slots[i] >= nritems - 1) {
+ skip_level = i + 1;
+ continue;
+ }
+ }
+ if (skip_level < i && i >= lowest_unlock)
+ no_skips = 1;
+
+ t = path->nodes[i];
+ if (i >= lowest_unlock && i > skip_level && path->locks[i]) {
+ btrfs_tree_unlock(t);
+ path->locks[i] = 0;
+ }
+ }
+}
+
/*
* look for key in the tree. path is filled in with nodes along the way
* if key is found, we return zero and you can find the item in the leaf
ins_len, int cow)
{
struct extent_buffer *b;
+ struct extent_buffer *tmp;
int slot;
int ret;
int level;
int should_reada = p->reada;
+ int lowest_unlock = 1;
+ int blocksize;
u8 lowest_level = 0;
+ u64 blocknr;
+ u64 gen;
+ struct btrfs_key prealloc_block;
lowest_level = p->lowest_level;
WARN_ON(lowest_level && ins_len);
WARN_ON(p->nodes[0] != NULL);
- WARN_ON(!mutex_is_locked(&root->fs_info->fs_mutex));
+ WARN_ON(cow && root == root->fs_info->extent_root &&
+ !mutex_is_locked(&root->fs_info->alloc_mutex));
+ if (ins_len < 0)
+ lowest_unlock = 2;
+
+ prealloc_block.objectid = 0;
+
again:
- b = root->node;
- extent_buffer_get(b);
+ if (p->skip_locking)
+ b = btrfs_root_node(root);
+ else
+ b = btrfs_lock_root_node(root);
+
while (b) {
level = btrfs_header_level(b);
+
+ /*
+ * setup the path here so we can release it under lock
+ * contention with the cow code
+ */
+ p->nodes[level] = b;
+ if (!p->skip_locking)
+ p->locks[level] = 1;
+
if (cow) {
int wret;
+
+ /* is a cow on this block not required */
+ spin_lock(&root->fs_info->hash_lock);
+ if (btrfs_header_generation(b) == trans->transid &&
+ !btrfs_header_flag(b, BTRFS_HEADER_FLAG_WRITTEN)) {
+ spin_unlock(&root->fs_info->hash_lock);
+ goto cow_done;
+ }
+ spin_unlock(&root->fs_info->hash_lock);
+
+ /* ok, we have to cow, is our old prealloc the right
+ * size?
+ */
+ if (prealloc_block.objectid &&
+ prealloc_block.offset != b->len) {
+ btrfs_free_reserved_extent(root,
+ prealloc_block.objectid,
+ prealloc_block.offset);
+ prealloc_block.objectid = 0;
+ }
+
+ /*
+ * for higher level blocks, try not to allocate blocks
+ * with the block and the parent locks held.
+ */
+ if (level > 1 && !prealloc_block.objectid &&
+ btrfs_path_lock_waiting(p, level)) {
+ u32 size = b->len;
+ u64 hint = b->start;
+
+ btrfs_release_path(root, p);
+ ret = btrfs_reserve_extent(trans, root,
+ size, size, 0,
+ hint, (u64)-1,
+ &prealloc_block, 0);
+ BUG_ON(ret);
+ goto again;
+ }
+
wret = btrfs_cow_block(trans, root, b,
p->nodes[level + 1],
p->slots[level + 1],
- &b);
+ &b, prealloc_block.objectid);
+ prealloc_block.objectid = 0;
if (wret) {
free_extent_buffer(b);
- return wret;
+ ret = wret;
+ goto done;
}
}
+cow_done:
BUG_ON(!cow && ins_len);
if (level != btrfs_header_level(b))
WARN_ON(1);
level = btrfs_header_level(b);
+
p->nodes[level] = b;
+ if (!p->skip_locking)
+ p->locks[level] = 1;
+
ret = check_block(root, p, level);
- if (ret)
- return -1;
+ if (ret) {
+ ret = -1;
+ goto done;
+ }
+
ret = bin_search(b, key, level, &slot);
if (level != 0) {
if (ret && slot > 0)
BTRFS_NODEPTRS_PER_BLOCK(root) - 3) {
int sret = split_node(trans, root, p, level);
BUG_ON(sret > 0);
- if (sret)
- return sret;
+ if (sret) {
+ ret = sret;
+ goto done;
+ }
b = p->nodes[level];
slot = p->slots[level];
} else if (ins_len < 0) {
int sret = balance_level(trans, root, p,
level);
- if (sret)
- return sret;
+ if (sret) {
+ ret = sret;
+ goto done;
+ }
b = p->nodes[level];
if (!b) {
btrfs_release_path(NULL, p);
slot = p->slots[level];
BUG_ON(btrfs_header_nritems(b) == 1);
}
+ unlock_up(p, level, lowest_unlock);
+
/* this is only true while dropping a snapshot */
- if (level == lowest_level)
+ if (level == lowest_level) {
break;
+ }
- if (should_reada)
- reada_for_search(root, p, level, slot,
- key->objectid);
+ blocknr = btrfs_node_blockptr(b, slot);
+ gen = btrfs_node_ptr_generation(b, slot);
+ blocksize = btrfs_level_size(root, level - 1);
- b = read_node_slot(root, b, slot);
+ tmp = btrfs_find_tree_block(root, blocknr, blocksize);
+ if (tmp && btrfs_buffer_uptodate(tmp, gen)) {
+ b = tmp;
+ } else {
+ /*
+ * reduce lock contention at high levels
+ * of the btree by dropping locks before
+ * we read.
+ */
+ if (level > 1) {
+ btrfs_release_path(NULL, p);
+ if (tmp)
+ free_extent_buffer(tmp);
+ if (should_reada)
+ reada_for_search(root, p,
+ level, slot,
+ key->objectid);
+
+ tmp = read_tree_block(root, blocknr,
+ blocksize, gen);
+ if (tmp)
+ free_extent_buffer(tmp);
+ goto again;
+ } else {
+ if (tmp)
+ free_extent_buffer(tmp);
+ if (should_reada)
+ reada_for_search(root, p,
+ level, slot,
+ key->objectid);
+ b = read_node_slot(root, b, slot);
+ }
+ }
+ if (!p->skip_locking)
+ btrfs_tree_lock(b);
} else {
p->slots[level] = slot;
if (ins_len > 0 && btrfs_leaf_free_space(root, b) <
int sret = split_leaf(trans, root, key,
p, ins_len, ret == 0);
BUG_ON(sret > 0);
- if (sret)
- return sret;
+ if (sret) {
+ ret = sret;
+ goto done;
+ }
}
- return ret;
+ unlock_up(p, level, lowest_unlock);
+ goto done;
}
}
- return 1;
+ ret = 1;
+done:
+ if (prealloc_block.objectid) {
+ btrfs_free_reserved_extent(root,
+ prealloc_block.objectid,
+ prealloc_block.offset);
+ }
+
+ return ret;
}
/*
u64 lower_gen;
struct extent_buffer *lower;
struct extent_buffer *c;
+ struct extent_buffer *old;
struct btrfs_disk_key lower_key;
BUG_ON(path->nodes[level]);
else
btrfs_node_key(lower, &lower_key, 0);
- c = __btrfs_alloc_free_block(trans, root, root->nodesize,
+ c = btrfs_alloc_free_block(trans, root, root->nodesize,
root->root_key.objectid,
- root_gen, lower_key.objectid, level,
- root->node->start, 0);
+ root_gen, le64_to_cpu(lower_key.objectid),
+ level, root->node->start, 0);
if (IS_ERR(c))
return PTR_ERR(c);
+
memset_extent_buffer(c, 0, 0, root->nodesize);
btrfs_set_header_nritems(c, 1);
btrfs_set_header_level(c, level);
btrfs_mark_buffer_dirty(c);
- /* the super has an extra ref to root->node */
- free_extent_buffer(root->node);
+ spin_lock(&root->node_lock);
+ old = root->node;
root->node = c;
+ spin_unlock(&root->node_lock);
+
+ /* the super has an extra ref to root->node */
+ free_extent_buffer(old);
+
add_root_to_dirty_list(root);
extent_buffer_get(c);
path->nodes[level] = c;
+ path->locks[level] = 1;
path->slots[level] = 0;
if (root->ref_cows && lower_gen != trans->transid) {
struct btrfs_path *back_path = btrfs_alloc_path();
int ret;
+ mutex_lock(&root->fs_info->alloc_mutex);
ret = btrfs_insert_extent_backref(trans,
root->fs_info->extent_root,
path, lower->start,
root->root_key.objectid,
trans->transid, 0, 0);
BUG_ON(ret);
+ mutex_unlock(&root->fs_info->alloc_mutex);
btrfs_free_path(back_path);
}
return 0;
*
* returns 0 on success and < 0 on failure
*/
-static int split_node(struct btrfs_trans_handle *trans, struct btrfs_root
- *root, struct btrfs_path *path, int level)
+static noinline int split_node(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path, int level)
{
u64 root_gen;
struct extent_buffer *c;
root_gen = 0;
btrfs_node_key(c, &disk_key, 0);
- split = __btrfs_alloc_free_block(trans, root, root->nodesize,
+ split = btrfs_alloc_free_block(trans, root, root->nodesize,
root->root_key.objectid,
root_gen,
btrfs_disk_key_objectid(&disk_key),
if (path->slots[level] >= mid) {
path->slots[level] -= mid;
+ btrfs_tree_unlock(c);
free_extent_buffer(c);
path->nodes[level] = split;
path->slots[level + 1] += 1;
} else {
+ btrfs_tree_unlock(split);
free_extent_buffer(split);
}
return ret;
* the start of the leaf data. IOW, how much room
* the leaf has left for both items and data
*/
-int btrfs_leaf_free_space(struct btrfs_root *root, struct extent_buffer *leaf)
+int noinline btrfs_leaf_free_space(struct btrfs_root *root,
+ struct extent_buffer *leaf)
{
int nritems = btrfs_header_nritems(leaf);
int ret;
if (slot >= btrfs_header_nritems(upper) - 1)
return 1;
+ WARN_ON(!btrfs_tree_locked(path->nodes[1]));
+
right = read_node_slot(root, upper, slot + 1);
+ btrfs_tree_lock(right);
free_space = btrfs_leaf_free_space(root, right);
- if (free_space < data_size + sizeof(struct btrfs_item)) {
- free_extent_buffer(right);
- return 1;
- }
+ if (free_space < data_size + sizeof(struct btrfs_item))
+ goto out_unlock;
/* cow and double check */
ret = btrfs_cow_block(trans, root, right, upper,
- slot + 1, &right);
- if (ret) {
- free_extent_buffer(right);
- return 1;
- }
+ slot + 1, &right, 0);
+ if (ret)
+ goto out_unlock;
+
free_space = btrfs_leaf_free_space(root, right);
- if (free_space < data_size + sizeof(struct btrfs_item)) {
- free_extent_buffer(right);
- return 1;
- }
+ if (free_space < data_size + sizeof(struct btrfs_item))
+ goto out_unlock;
left_nritems = btrfs_header_nritems(left);
- if (left_nritems == 0) {
- free_extent_buffer(right);
- return 1;
- }
+ if (left_nritems == 0)
+ goto out_unlock;
if (empty)
nr = 0;
left->map_token = NULL;
}
- if (push_items == 0) {
- free_extent_buffer(right);
- return 1;
- }
+ if (push_items == 0)
+ goto out_unlock;
if (!empty && push_items == left_nritems)
WARN_ON(1);
/* then fixup the leaf pointer in the path */
if (path->slots[0] >= left_nritems) {
path->slots[0] -= left_nritems;
+ if (btrfs_header_nritems(path->nodes[0]) == 0)
+ clean_tree_block(trans, root, path->nodes[0]);
+ btrfs_tree_unlock(path->nodes[0]);
free_extent_buffer(path->nodes[0]);
path->nodes[0] = right;
path->slots[1] += 1;
} else {
+ btrfs_tree_unlock(right);
free_extent_buffer(right);
}
return 0;
+
+out_unlock:
+ btrfs_tree_unlock(right);
+ free_extent_buffer(right);
+ return 1;
}
+
/*
* push some data in the path leaf to the left, trying to free up at
* least data_size bytes. returns zero if the push worked, nonzero otherwise
return 1;
}
+ WARN_ON(!btrfs_tree_locked(path->nodes[1]));
+
left = read_node_slot(root, path->nodes[1], slot - 1);
+ btrfs_tree_lock(left);
free_space = btrfs_leaf_free_space(root, left);
if (free_space < data_size + sizeof(struct btrfs_item)) {
- free_extent_buffer(left);
- return 1;
+ ret = 1;
+ goto out;
}
/* cow and double check */
ret = btrfs_cow_block(trans, root, left,
- path->nodes[1], slot - 1, &left);
+ path->nodes[1], slot - 1, &left, 0);
if (ret) {
/* we hit -ENOSPC, but it isn't fatal here */
- free_extent_buffer(left);
- return 1;
+ ret = 1;
+ goto out;
}
free_space = btrfs_leaf_free_space(root, left);
if (free_space < data_size + sizeof(struct btrfs_item)) {
- free_extent_buffer(left);
- return 1;
+ ret = 1;
+ goto out;
}
if (empty)
}
if (push_items == 0) {
- free_extent_buffer(left);
- return 1;
+ ret = 1;
+ goto out;
}
if (!empty && push_items == btrfs_header_nritems(right))
WARN_ON(1);
/* then fixup the leaf pointer in the path */
if (path->slots[0] < push_items) {
path->slots[0] += old_left_nritems;
+ if (btrfs_header_nritems(path->nodes[0]) == 0)
+ clean_tree_block(trans, root, path->nodes[0]);
+ btrfs_tree_unlock(path->nodes[0]);
free_extent_buffer(path->nodes[0]);
path->nodes[0] = left;
path->slots[1] -= 1;
} else {
+ btrfs_tree_unlock(left);
free_extent_buffer(left);
path->slots[0] -= push_items;
}
BUG_ON(path->slots[0] < 0);
return ret;
+out:
+ btrfs_tree_unlock(left);
+ free_extent_buffer(left);
+ return ret;
}
/*
*
* returns 0 if all went well and < 0 on failure.
*/
-static int split_leaf(struct btrfs_trans_handle *trans, struct btrfs_root
- *root, struct btrfs_key *ins_key,
- struct btrfs_path *path, int data_size, int extend)
+static noinline int split_leaf(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_key *ins_key,
+ struct btrfs_path *path, int data_size,
+ int extend)
{
u64 root_gen;
struct extent_buffer *l;
btrfs_item_key(l, &disk_key, 0);
- right = __btrfs_alloc_free_block(trans, root, root->leafsize,
+ right = btrfs_alloc_free_block(trans, root, root->leafsize,
root->root_key.objectid,
- root_gen, disk_key.objectid, 0,
- l->start, 0);
+ root_gen,
+ le64_to_cpu(disk_key.objectid),
+ 0, l->start, 0);
if (IS_ERR(right)) {
BUG_ON(1);
return PTR_ERR(right);
path->slots[1] + 1, 1);
if (wret)
ret = wret;
+
+ btrfs_tree_unlock(path->nodes[0]);
free_extent_buffer(path->nodes[0]);
path->nodes[0] = right;
path->slots[0] = 0;
path->slots[1], 1);
if (wret)
ret = wret;
+ btrfs_tree_unlock(path->nodes[0]);
free_extent_buffer(path->nodes[0]);
path->nodes[0] = right;
path->slots[0] = 0;
BUG_ON(path->slots[0] != slot);
if (mid <= slot) {
+ btrfs_tree_unlock(path->nodes[0]);
free_extent_buffer(path->nodes[0]);
path->nodes[0] = right;
path->slots[0] -= mid;
path->slots[1] += 1;
- } else
+ } else {
+ btrfs_tree_unlock(right);
free_extent_buffer(right);
+ }
BUG_ON(path->slots[0] < 0);
total_data += data_size[i];
}
- /* create a root if there isn't one */
- if (!root->node)
- BUG();
-
- total_size = total_data + (nr - 1) * sizeof(struct btrfs_item);
+ total_size = total_data + (nr * sizeof(struct btrfs_item));
ret = btrfs_search_slot(trans, root, cpu_key, path, total_size, 1);
- if (ret == 0) {
+ if (ret == 0)
return -EEXIST;
- }
if (ret < 0)
goto out;
nritems = btrfs_header_nritems(leaf);
data_end = leaf_data_end(root, leaf);
- if (btrfs_leaf_free_space(root, leaf) <
- sizeof(struct btrfs_item) + total_size) {
+ if (btrfs_leaf_free_space(root, leaf) < total_size) {
btrfs_print_leaf(root, leaf);
printk("not enough freespace need %u have %d\n",
total_size, btrfs_leaf_free_space(root, leaf));
BUG_ON(slot < 0);
if (slot != nritems) {
- int i;
unsigned int old_data = btrfs_item_end_nr(leaf, slot);
if (old_data < data_end) {
btrfs_print_leaf(root, leaf);
BUG();
}
-
out:
return ret;
}
nritems = btrfs_header_nritems(leaf);
if (slot + nr != nritems) {
- int i;
int data_end = leaf_data_end(root, leaf);
memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
btrfs_set_header_level(leaf, 0);
} else {
u64 root_gen = btrfs_header_generation(path->nodes[1]);
- clean_tree_block(trans, root, leaf);
wret = del_ptr(trans, root, path, 1, path->slots[1]);
if (wret)
ret = wret;
root_gen = btrfs_header_generation(
path->nodes[1]);
- clean_tree_block(trans, root, leaf);
-
wret = del_ptr(trans, root, path, 1, slot);
if (wret)
ret = wret;
if (wret)
ret = wret;
} else {
- btrfs_mark_buffer_dirty(leaf);
+ /* if we're still in the path, make sure
+ * we're dirty. Otherwise, one of the
+ * push_leaf functions must have already
+ * dirtied this buffer
+ */
+ if (path->nodes[0] == leaf)
+ btrfs_mark_buffer_dirty(leaf);
free_extent_buffer(leaf);
}
} else {
}
/*
- * walk up the tree as far as required to find the previous leaf.
+ * search the tree again to find a leaf with lesser keys
* returns 0 if it found something or 1 if there are no lesser leaves.
* returns < 0 on io errors.
*/
int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path)
{
+ struct btrfs_key key;
+ struct btrfs_disk_key found_key;
+ int ret;
+
+ btrfs_item_key_to_cpu(path->nodes[0], &key, 0);
+
+ if (key.offset > 0)
+ key.offset--;
+ else if (key.type > 0)
+ key.type--;
+ else if (key.objectid > 0)
+ key.objectid--;
+ else
+ return 1;
+
+ btrfs_release_path(root, path);
+ ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+ if (ret < 0)
+ return ret;
+ btrfs_item_key(path->nodes[0], &found_key, 0);
+ ret = comp_keys(&found_key, &key);
+ if (ret < 0)
+ return 0;
+ return 1;
+}
+
+/*
+ * A helper function to walk down the tree starting at min_key, and looking
+ * for nodes or leaves that are either in cache or have a minimum
+ * transaction id. This is used by the btree defrag code, but could
+ * also be used to search for blocks that have changed since a given
+ * transaction id.
+ *
+ * This does not cow, but it does stuff the starting key it finds back
+ * into min_key, so you can call btrfs_search_slot with cow=1 on the
+ * key and get a writable path.
+ *
+ * This does lock as it descends, and path->keep_locks should be set
+ * to 1 by the caller.
+ *
+ * This honors path->lowest_level to prevent descent past a given level
+ * of the tree.
+ *
+ * returns zero if something useful was found, < 0 on error and 1 if there
+ * was nothing in the tree that matched the search criteria.
+ */
+int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key,
+ struct btrfs_key *max_key,
+ struct btrfs_path *path, int cache_only,
+ u64 min_trans)
+{
+ struct extent_buffer *cur;
+ struct btrfs_key found_key;
+ int slot;
+ int sret;
+ u32 nritems;
+ int level;
+ int ret = 1;
+
+again:
+ cur = btrfs_lock_root_node(root);
+ level = btrfs_header_level(cur);
+ WARN_ON(path->nodes[level]);
+ path->nodes[level] = cur;
+ path->locks[level] = 1;
+
+ if (btrfs_header_generation(cur) < min_trans) {
+ ret = 1;
+ goto out;
+ }
+ while(1) {
+ nritems = btrfs_header_nritems(cur);
+ level = btrfs_header_level(cur);
+ sret = bin_search(cur, min_key, level, &slot);
+
+ /* at level = 0, we're done, setup the path and exit */
+ if (level == 0) {
+ if (slot >= nritems)
+ goto find_next_key;
+ ret = 0;
+ path->slots[level] = slot;
+ btrfs_item_key_to_cpu(cur, &found_key, slot);
+ goto out;
+ }
+ if (sret && slot > 0)
+ slot--;
+ /*
+ * check this node pointer against the cache_only and
+ * min_trans parameters. If it isn't in cache or is too
+ * old, skip to the next one.
+ */
+ while(slot < nritems) {
+ u64 blockptr;
+ u64 gen;
+ struct extent_buffer *tmp;
+ struct btrfs_disk_key disk_key;
+
+ blockptr = btrfs_node_blockptr(cur, slot);
+ gen = btrfs_node_ptr_generation(cur, slot);
+ if (gen < min_trans) {
+ slot++;
+ continue;
+ }
+ if (!cache_only)
+ break;
+
+ if (max_key) {
+ btrfs_node_key(cur, &disk_key, slot);
+ if (comp_keys(&disk_key, max_key) >= 0) {
+ ret = 1;
+ goto out;
+ }
+ }
+
+ tmp = btrfs_find_tree_block(root, blockptr,
+ btrfs_level_size(root, level - 1));
+
+ if (tmp && btrfs_buffer_uptodate(tmp, gen)) {
+ free_extent_buffer(tmp);
+ break;
+ }
+ if (tmp)
+ free_extent_buffer(tmp);
+ slot++;
+ }
+find_next_key:
+ /*
+ * we didn't find a candidate key in this node, walk forward
+ * and find another one
+ */
+ if (slot >= nritems) {
+ path->slots[level] = slot;
+ sret = btrfs_find_next_key(root, path, min_key, level,
+ cache_only, min_trans);
+ if (sret == 0) {
+ btrfs_release_path(root, path);
+ goto again;
+ } else {
+ goto out;
+ }
+ }
+ /* save our key for returning back */
+ btrfs_node_key_to_cpu(cur, &found_key, slot);
+ path->slots[level] = slot;
+ if (level == path->lowest_level) {
+ ret = 0;
+ unlock_up(path, level, 1);
+ goto out;
+ }
+ cur = read_node_slot(root, cur, slot);
+
+ btrfs_tree_lock(cur);
+ path->locks[level - 1] = 1;
+ path->nodes[level - 1] = cur;
+ unlock_up(path, level, 1);
+ }
+out:
+ if (ret == 0)
+ memcpy(min_key, &found_key, sizeof(found_key));
+ return ret;
+}
+
+/*
+ * this is similar to btrfs_next_leaf, but does not try to preserve
+ * and fixup the path. It looks for and returns the next key in the
+ * tree based on the current path and the cache_only and min_trans
+ * parameters.
+ *
+ * 0 is returned if another key is found, < 0 if there are any errors
+ * and 1 is returned if there are no higher keys in the tree
+ *
+ * path->keep_locks should be set to 1 on the search made before
+ * calling this function.
+ */
+int btrfs_find_next_key(struct btrfs_root *root, struct btrfs_path *path,
+ struct btrfs_key *key, int lowest_level,
+ int cache_only, u64 min_trans)
+{
+ int level = lowest_level;
int slot;
- int level = 1;
struct extent_buffer *c;
- struct extent_buffer *next = NULL;
while(level < BTRFS_MAX_LEVEL) {
if (!path->nodes[level])
return 1;
- slot = path->slots[level];
+ slot = path->slots[level] + 1;
c = path->nodes[level];
- if (slot == 0) {
+next:
+ if (slot >= btrfs_header_nritems(c)) {
level++;
- if (level == BTRFS_MAX_LEVEL)
+ if (level == BTRFS_MAX_LEVEL) {
return 1;
+ }
continue;
}
- slot--;
-
- if (next)
- free_extent_buffer(next);
+ if (level == 0)
+ btrfs_item_key_to_cpu(c, key, slot);
+ else {
+ u64 blockptr = btrfs_node_blockptr(c, slot);
+ u64 gen = btrfs_node_ptr_generation(c, slot);
- next = read_node_slot(root, c, slot);
- break;
- }
- path->slots[level] = slot;
- while(1) {
- level--;
- c = path->nodes[level];
- free_extent_buffer(c);
- slot = btrfs_header_nritems(next);
- if (slot != 0)
- slot--;
- path->nodes[level] = next;
- path->slots[level] = slot;
- if (!level)
- break;
- next = read_node_slot(root, next, slot);
+ if (cache_only) {
+ struct extent_buffer *cur;
+ cur = btrfs_find_tree_block(root, blockptr,
+ btrfs_level_size(root, level - 1));
+ if (!cur || !btrfs_buffer_uptodate(cur, gen)) {
+ slot++;
+ if (cur)
+ free_extent_buffer(cur);
+ goto next;
+ }
+ free_extent_buffer(cur);
+ }
+ if (gen < min_trans) {
+ slot++;
+ goto next;
+ }
+ btrfs_node_key_to_cpu(c, key, slot);
+ }
+ return 0;
}
- return 0;
+ return 1;
}
/*
- * walk up the tree as far as required to find the next leaf.
+ * search the tree again to find a leaf with greater keys
* returns 0 if it found something or 1 if there are no greater leaves.
* returns < 0 on io errors.
*/
int level = 1;
struct extent_buffer *c;
struct extent_buffer *next = NULL;
+ struct btrfs_key key;
+ u32 nritems;
+ int ret;
+
+ nritems = btrfs_header_nritems(path->nodes[0]);
+ if (nritems == 0) {
+ return 1;
+ }
+
+ btrfs_item_key_to_cpu(path->nodes[0], &key, nritems - 1);
+
+ btrfs_release_path(root, path);
+ path->keep_locks = 1;
+ ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+ path->keep_locks = 0;
+
+ if (ret < 0)
+ return ret;
+
+ nritems = btrfs_header_nritems(path->nodes[0]);
+ /*
+ * by releasing the path above we dropped all our locks. A balance
+ * could have added more items next to the key that used to be
+ * at the very end of the block. So, check again here and
+ * advance the path if there are now more items available.
+ */
+ if (nritems > 0 && path->slots[0] < nritems - 1) {
+ path->slots[0]++;
+ goto done;
+ }
while(level < BTRFS_MAX_LEVEL) {
if (!path->nodes[level])
c = path->nodes[level];
if (slot >= btrfs_header_nritems(c)) {
level++;
- if (level == BTRFS_MAX_LEVEL)
+ if (level == BTRFS_MAX_LEVEL) {
return 1;
+ }
continue;
}
- if (next)
+ if (next) {
+ btrfs_tree_unlock(next);
free_extent_buffer(next);
+ }
- if (path->reada)
+ if (level == 1 && (path->locks[1] || path->skip_locking) &&
+ path->reada)
reada_for_search(root, path, level, slot, 0);
next = read_node_slot(root, c, slot);
+ if (!path->skip_locking) {
+ WARN_ON(!btrfs_tree_locked(c));
+ btrfs_tree_lock(next);
+ }
break;
}
path->slots[level] = slot;
while(1) {
level--;
c = path->nodes[level];
+ if (path->locks[level])
+ btrfs_tree_unlock(c);
free_extent_buffer(c);
path->nodes[level] = next;
path->slots[level] = 0;
+ if (!path->skip_locking)
+ path->locks[level] = 1;
if (!level)
break;
- if (path->reada)
- reada_for_search(root, path, level, 0, 0);
+ if (level == 1 && path->locks[1] && path->reada)
+ reada_for_search(root, path, level, slot, 0);
next = read_node_slot(root, next, 0);
+ if (!path->skip_locking) {
+ WARN_ON(!btrfs_tree_locked(path->nodes[level]));
+ btrfs_tree_lock(next);
+ }
}
+done:
+ unlock_up(path, 0, 1);
return 0;
}
+/*
+ * this uses btrfs_prev_leaf to walk backwards in the tree, and keeps
+ * searching until it gets past min_objectid or finds an item of 'type'
+ *
+ * returns 0 if something is found, 1 if nothing was found and < 0 on error
+ */
int btrfs_previous_item(struct btrfs_root *root,
struct btrfs_path *path, u64 min_objectid,
int type)
{
struct btrfs_key found_key;
struct extent_buffer *leaf;
+ u32 nritems;
int ret;
while(1) {
path->slots[0]--;
}
leaf = path->nodes[0];
+ nritems = btrfs_header_nritems(leaf);
+ if (nritems == 0)
+ return 1;
+ if (path->slots[0] == nritems)
+ path->slots[0]--;
+
btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
if (found_key.type == type)
return 0;
+ if (found_key.objectid < min_objectid)
+ break;
+ if (found_key.objectid == min_objectid &&
+ found_key.type < type)
+ break;
}
return 1;
}
-
projects / linux-2.6-block.git / blobdiff