| #include <stdio.h> |
| #include <stdlib.h> |
| #include "kerncompat.h" |
| #include "radix-tree.h" |
| #include "ctree.h" |
| #include "disk-io.h" |
| #include "print-tree.h" |
| |
| static int split_node(struct ctree_root *root, struct ctree_path *path, |
| int level); |
| static int split_leaf(struct ctree_root *root, struct ctree_path *path, |
| int data_size); |
| static int push_node_left(struct ctree_root *root, struct tree_buffer *dst, |
| struct tree_buffer *src); |
| static int balance_node_right(struct ctree_root *root, |
| struct tree_buffer *dst_buf, |
| struct tree_buffer *src_buf); |
| static int del_ptr(struct ctree_root *root, struct ctree_path *path, int level, |
| int slot); |
| |
| inline void init_path(struct ctree_path *p) |
| { |
| memset(p, 0, sizeof(*p)); |
| } |
| |
| void release_path(struct ctree_root *root, struct ctree_path *p) |
| { |
| int i; |
| for (i = 0; i < MAX_LEVEL; i++) { |
| if (!p->nodes[i]) |
| break; |
| tree_block_release(root, p->nodes[i]); |
| } |
| memset(p, 0, sizeof(*p)); |
| } |
| |
| int btrfs_cow_block(struct ctree_root *root, |
| struct tree_buffer *buf, |
| struct tree_buffer *parent, |
| int parent_slot, |
| struct tree_buffer **cow_ret) |
| { |
| struct tree_buffer *cow; |
| |
| if (!list_empty(&buf->dirty)) { |
| *cow_ret = buf; |
| return 0; |
| } |
| cow = alloc_free_block(root); |
| memcpy(&cow->node, &buf->node, sizeof(buf->node)); |
| btrfs_set_header_blocknr(&cow->node.header, cow->blocknr); |
| *cow_ret = cow; |
| btrfs_inc_ref(root, buf); |
| if (buf == root->node) { |
| root->node = cow; |
| cow->count++; |
| if (buf != root->commit_root) |
| free_extent(root, buf->blocknr, 1); |
| tree_block_release(root, buf); |
| } else { |
| parent->node.blockptrs[parent_slot] = cow->blocknr; |
| BUG_ON(list_empty(&parent->dirty)); |
| free_extent(root, buf->blocknr, 1); |
| } |
| tree_block_release(root, buf); |
| return 0; |
| } |
| |
| /* |
| * The leaf data grows from end-to-front in the node. |
| * this returns the address of the start of the last item, |
| * which is the stop of the leaf data stack |
| */ |
| static inline unsigned int leaf_data_end(struct leaf *leaf) |
| { |
| u32 nr = btrfs_header_nritems(&leaf->header); |
| if (nr == 0) |
| return sizeof(leaf->data); |
| return leaf->items[nr-1].offset; |
| } |
| |
| /* |
| * The space between the end of the leaf items and |
| * the start of the leaf data. IOW, how much room |
| * the leaf has left for both items and data |
| */ |
| int leaf_free_space(struct leaf *leaf) |
| { |
| int data_end = leaf_data_end(leaf); |
| int nritems = btrfs_header_nritems(&leaf->header); |
| char *items_end = (char *)(leaf->items + nritems + 1); |
| return (char *)(leaf->data + data_end) - (char *)items_end; |
| } |
| |
| /* |
| * compare two keys in a memcmp fashion |
| */ |
| int comp_keys(struct key *k1, struct key *k2) |
| { |
| if (k1->objectid > k2->objectid) |
| return 1; |
| if (k1->objectid < k2->objectid) |
| return -1; |
| if (k1->flags > k2->flags) |
| return 1; |
| if (k1->flags < k2->flags) |
| return -1; |
| if (k1->offset > k2->offset) |
| return 1; |
| if (k1->offset < k2->offset) |
| return -1; |
| return 0; |
| } |
| |
| int check_node(struct ctree_path *path, int level) |
| { |
| int i; |
| struct node *parent = NULL; |
| struct node *node = &path->nodes[level]->node; |
| int parent_slot; |
| u32 nritems = btrfs_header_nritems(&node->header); |
| |
| if (path->nodes[level + 1]) |
| parent = &path->nodes[level + 1]->node; |
| parent_slot = path->slots[level + 1]; |
| BUG_ON(nritems == 0); |
| if (parent) { |
| struct key *parent_key; |
| parent_key = &parent->keys[parent_slot]; |
| BUG_ON(memcmp(parent_key, node->keys, sizeof(struct key))); |
| BUG_ON(parent->blockptrs[parent_slot] != |
| btrfs_header_blocknr(&node->header)); |
| } |
| BUG_ON(nritems > NODEPTRS_PER_BLOCK); |
| for (i = 0; nritems > 1 && i < nritems - 2; i++) { |
| BUG_ON(comp_keys(&node->keys[i], &node->keys[i+1]) >= 0); |
| } |
| return 0; |
| } |
| |
| int check_leaf(struct ctree_path *path, int level) |
| { |
| int i; |
| struct leaf *leaf = &path->nodes[level]->leaf; |
| struct node *parent = NULL; |
| int parent_slot; |
| u32 nritems = btrfs_header_nritems(&leaf->header); |
| |
| if (path->nodes[level + 1]) |
| parent = &path->nodes[level + 1]->node; |
| parent_slot = path->slots[level + 1]; |
| BUG_ON(leaf_free_space(leaf) < 0); |
| |
| if (nritems == 0) |
| return 0; |
| |
| if (parent) { |
| struct key *parent_key; |
| parent_key = &parent->keys[parent_slot]; |
| BUG_ON(memcmp(parent_key, &leaf->items[0].key, |
| sizeof(struct key))); |
| BUG_ON(parent->blockptrs[parent_slot] != |
| btrfs_header_blocknr(&leaf->header)); |
| } |
| for (i = 0; nritems > 1 && i < nritems - 2; i++) { |
| BUG_ON(comp_keys(&leaf->items[i].key, |
| &leaf->items[i+1].key) >= 0); |
| BUG_ON(leaf->items[i].offset != leaf->items[i + 1].offset + |
| leaf->items[i + 1].size); |
| if (i == 0) { |
| BUG_ON(leaf->items[i].offset + leaf->items[i].size != |
| LEAF_DATA_SIZE); |
| } |
| } |
| return 0; |
| } |
| |
| int check_block(struct ctree_path *path, int level) |
| { |
| if (level == 0) |
| return check_leaf(path, level); |
| return check_node(path, level); |
| } |
| |
| /* |
| * search for key in the array p. items p are item_size apart |
| * and there are 'max' items in p |
| * the slot in the array is returned via slot, and it points to |
| * the place where you would insert key if it is not found in |
| * the array. |
| * |
| * slot may point to max if the key is bigger than all of the keys |
| */ |
| int generic_bin_search(char *p, int item_size, struct key *key, |
| int max, int *slot) |
| { |
| int low = 0; |
| int high = max; |
| int mid; |
| int ret; |
| struct key *tmp; |
| |
| while(low < high) { |
| mid = (low + high) / 2; |
| tmp = (struct key *)(p + mid * item_size); |
| ret = comp_keys(tmp, key); |
| |
| if (ret < 0) |
| low = mid + 1; |
| else if (ret > 0) |
| high = mid; |
| else { |
| *slot = mid; |
| return 0; |
| } |
| } |
| *slot = low; |
| return 1; |
| } |
| |
| /* |
| * simple bin_search frontend that does the right thing for |
| * leaves vs nodes |
| */ |
| int bin_search(struct node *c, struct key *key, int *slot) |
| { |
| if (btrfs_is_leaf(c)) { |
| struct leaf *l = (struct leaf *)c; |
| return generic_bin_search((void *)l->items, sizeof(struct item), |
| key, btrfs_header_nritems(&c->header), |
| slot); |
| } else { |
| return generic_bin_search((void *)c->keys, sizeof(struct key), |
| key, btrfs_header_nritems(&c->header), |
| slot); |
| } |
| return -1; |
| } |
| |
| struct tree_buffer *read_node_slot(struct ctree_root *root, |
| struct tree_buffer *parent_buf, |
| int slot) |
| { |
| struct node *node = &parent_buf->node; |
| if (slot < 0) |
| return NULL; |
| if (slot >= btrfs_header_nritems(&node->header)) |
| return NULL; |
| return read_tree_block(root, node->blockptrs[slot]); |
| } |
| |
| static int balance_level(struct ctree_root *root, struct ctree_path *path, |
| int level) |
| { |
| struct tree_buffer *right_buf; |
| struct tree_buffer *mid_buf; |
| struct tree_buffer *left_buf; |
| struct tree_buffer *parent_buf = NULL; |
| struct node *right = NULL; |
| struct node *mid; |
| struct node *left = NULL; |
| struct node *parent = NULL; |
| int ret = 0; |
| int wret; |
| int pslot; |
| int orig_slot = path->slots[level]; |
| u64 orig_ptr; |
| |
| if (level == 0) |
| return 0; |
| |
| mid_buf = path->nodes[level]; |
| mid = &mid_buf->node; |
| orig_ptr = mid->blockptrs[orig_slot]; |
| |
| if (level < MAX_LEVEL - 1) |
| parent_buf = path->nodes[level + 1]; |
| pslot = path->slots[level + 1]; |
| |
| if (!parent_buf) { |
| struct tree_buffer *child; |
| u64 blocknr = mid_buf->blocknr; |
| |
| if (btrfs_header_nritems(&mid->header) != 1) |
| return 0; |
| |
| /* promote the child to a root */ |
| child = read_node_slot(root, mid_buf, 0); |
| BUG_ON(!child); |
| root->node = child; |
| path->nodes[level] = NULL; |
| /* once for the path */ |
| tree_block_release(root, mid_buf); |
| /* once for the root ptr */ |
| tree_block_release(root, mid_buf); |
| clean_tree_block(root, mid_buf); |
| return free_extent(root, blocknr, 1); |
| } |
| parent = &parent_buf->node; |
| |
| if (btrfs_header_nritems(&mid->header) > NODEPTRS_PER_BLOCK / 4) |
| return 0; |
| |
| left_buf = read_node_slot(root, parent_buf, pslot - 1); |
| right_buf = read_node_slot(root, parent_buf, pslot + 1); |
| |
| /* first, try to make some room in the middle buffer */ |
| if (left_buf) { |
| btrfs_cow_block(root, left_buf, parent_buf, |
| pslot - 1, &left_buf); |
| left = &left_buf->node; |
| orig_slot += btrfs_header_nritems(&left->header); |
| wret = push_node_left(root, left_buf, mid_buf); |
| if (wret < 0) |
| ret = wret; |
| } |
| |
| /* |
| * then try to empty the right most buffer into the middle |
| */ |
| if (right_buf) { |
| btrfs_cow_block(root, right_buf, parent_buf, |
| pslot + 1, &right_buf); |
| right = &right_buf->node; |
| wret = push_node_left(root, mid_buf, right_buf); |
| if (wret < 0) |
| ret = wret; |
| if (btrfs_header_nritems(&right->header) == 0) { |
| u64 blocknr = right_buf->blocknr; |
| tree_block_release(root, right_buf); |
| clean_tree_block(root, right_buf); |
| right_buf = NULL; |
| right = NULL; |
| wret = del_ptr(root, path, level + 1, pslot + 1); |
| if (wret) |
| ret = wret; |
| wret = free_extent(root, blocknr, 1); |
| if (wret) |
| ret = wret; |
| } else { |
| memcpy(parent->keys + pslot + 1, right->keys, |
| sizeof(struct key)); |
| BUG_ON(list_empty(&parent_buf->dirty)); |
| } |
| } |
| if (btrfs_header_nritems(&mid->header) == 1) { |
| /* |
| * we're not allowed to leave a node with one item in the |
| * tree during a delete. A deletion from lower in the tree |
| * could try to delete the only pointer in this node. |
| * So, pull some keys from the left. |
| * There has to be a left pointer at this point because |
| * otherwise we would have pulled some pointers from the |
| * right |
| */ |
| BUG_ON(!left_buf); |
| wret = balance_node_right(root, mid_buf, left_buf); |
| if (wret < 0) |
| ret = wret; |
| BUG_ON(wret == 1); |
| } |
| if (btrfs_header_nritems(&mid->header) == 0) { |
| /* we've managed to empty the middle node, drop it */ |
| u64 blocknr = mid_buf->blocknr; |
| tree_block_release(root, mid_buf); |
| clean_tree_block(root, mid_buf); |
| mid_buf = NULL; |
| mid = NULL; |
| wret = del_ptr(root, path, level + 1, pslot); |
| if (wret) |
| ret = wret; |
| wret = free_extent(root, blocknr, 1); |
| if (wret) |
| ret = wret; |
| } else { |
| /* update the parent key to reflect our changes */ |
| memcpy(parent->keys + pslot, mid->keys, sizeof(struct key)); |
| BUG_ON(list_empty(&parent_buf->dirty)); |
| } |
| |
| /* update the path */ |
| if (left_buf) { |
| if (btrfs_header_nritems(&left->header) > orig_slot) { |
| left_buf->count++; // released below |
| path->nodes[level] = left_buf; |
| path->slots[level + 1] -= 1; |
| path->slots[level] = orig_slot; |
| if (mid_buf) |
| tree_block_release(root, mid_buf); |
| } else { |
| orig_slot -= btrfs_header_nritems(&left->header); |
| path->slots[level] = orig_slot; |
| } |
| } |
| /* double check we haven't messed things up */ |
| check_block(path, level); |
| if (orig_ptr != path->nodes[level]->node.blockptrs[path->slots[level]]) |
| BUG(); |
| |
| if (right_buf) |
| tree_block_release(root, right_buf); |
| if (left_buf) |
| tree_block_release(root, left_buf); |
| return ret; |
| } |
| |
| /* |
| * 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 |
| * level of the path (level 0) |
| * |
| * If the key isn't found, the path points to the slot where it should |
| * be inserted, and 1 is returned. If there are other errors during the |
| * search a negative error number is returned. |
| * |
| * if ins_len > 0, nodes and leaves will be split as we walk down the |
| * tree. if ins_len < 0, nodes will be merged as we walk down the tree (if |
| * possible) |
| */ |
| int search_slot(struct ctree_root *root, struct key *key, |
| struct ctree_path *p, int ins_len, int cow) |
| { |
| struct tree_buffer *b; |
| struct tree_buffer *cow_buf; |
| struct node *c; |
| int slot; |
| int ret; |
| int level; |
| |
| again: |
| b = root->node; |
| b->count++; |
| while (b) { |
| level = btrfs_header_level(&b->node.header); |
| if (cow) { |
| int wret; |
| wret = btrfs_cow_block(root, b, p->nodes[level + 1], |
| p->slots[level + 1], &cow_buf); |
| b = cow_buf; |
| } |
| BUG_ON(!cow && ins_len); |
| c = &b->node; |
| p->nodes[level] = b; |
| ret = check_block(p, level); |
| if (ret) |
| return -1; |
| ret = bin_search(c, key, &slot); |
| if (!btrfs_is_leaf(c)) { |
| if (ret && slot > 0) |
| slot -= 1; |
| p->slots[level] = slot; |
| if (ins_len > 0 && btrfs_header_nritems(&c->header) == |
| NODEPTRS_PER_BLOCK) { |
| int sret = split_node(root, p, level); |
| BUG_ON(sret > 0); |
| if (sret) |
| return sret; |
| b = p->nodes[level]; |
| c = &b->node; |
| slot = p->slots[level]; |
| } else if (ins_len < 0) { |
| int sret = balance_level(root, p, level); |
| if (sret) |
| return sret; |
| b = p->nodes[level]; |
| if (!b) |
| goto again; |
| c = &b->node; |
| slot = p->slots[level]; |
| BUG_ON(btrfs_header_nritems(&c->header) == 1); |
| } |
| b = read_tree_block(root, c->blockptrs[slot]); |
| } else { |
| struct leaf *l = (struct leaf *)c; |
| p->slots[level] = slot; |
| if (ins_len > 0 && leaf_free_space(l) < |
| sizeof(struct item) + ins_len) { |
| int sret = split_leaf(root, p, ins_len); |
| BUG_ON(sret > 0); |
| if (sret) |
| return sret; |
| } |
| BUG_ON(root->node->count == 1); |
| return ret; |
| } |
| } |
| BUG_ON(root->node->count == 1); |
| return 1; |
| } |
| |
| /* |
| * adjust the pointers going up the tree, starting at level |
| * making sure the right key of each node is points to 'key'. |
| * This is used after shifting pointers to the left, so it stops |
| * fixing up pointers when a given leaf/node is not in slot 0 of the |
| * higher levels |
| * |
| * If this fails to write a tree block, it returns -1, but continues |
| * fixing up the blocks in ram so the tree is consistent. |
| */ |
| static int fixup_low_keys(struct ctree_root *root, |
| struct ctree_path *path, struct key *key, |
| int level) |
| { |
| int i; |
| int ret = 0; |
| for (i = level; i < MAX_LEVEL; i++) { |
| struct node *t; |
| int tslot = path->slots[i]; |
| if (!path->nodes[i]) |
| break; |
| t = &path->nodes[i]->node; |
| memcpy(t->keys + tslot, key, sizeof(*key)); |
| BUG_ON(list_empty(&path->nodes[i]->dirty)); |
| if (tslot != 0) |
| break; |
| } |
| return ret; |
| } |
| |
| /* |
| * try to push data from one node into the next node left in the |
| * tree. |
| * |
| * returns 0 if some ptrs were pushed left, < 0 if there was some horrible |
| * error, and > 0 if there was no room in the left hand block. |
| */ |
| static int push_node_left(struct ctree_root *root, struct tree_buffer *dst_buf, |
| struct tree_buffer *src_buf) |
| { |
| struct node *src = &src_buf->node; |
| struct node *dst = &dst_buf->node; |
| int push_items = 0; |
| int src_nritems; |
| int dst_nritems; |
| int ret = 0; |
| |
| src_nritems = btrfs_header_nritems(&src->header); |
| dst_nritems = btrfs_header_nritems(&dst->header); |
| push_items = NODEPTRS_PER_BLOCK - dst_nritems; |
| if (push_items <= 0) { |
| return 1; |
| } |
| |
| if (src_nritems < push_items) |
| push_items = src_nritems; |
| |
| memcpy(dst->keys + dst_nritems, src->keys, |
| push_items * sizeof(struct key)); |
| memcpy(dst->blockptrs + dst_nritems, src->blockptrs, |
| push_items * sizeof(u64)); |
| if (push_items < src_nritems) { |
| memmove(src->keys, src->keys + push_items, |
| (src_nritems - push_items) * sizeof(struct key)); |
| memmove(src->blockptrs, src->blockptrs + push_items, |
| (src_nritems - push_items) * sizeof(u64)); |
| } |
| btrfs_set_header_nritems(&src->header, src_nritems - push_items); |
| btrfs_set_header_nritems(&dst->header, dst_nritems + push_items); |
| BUG_ON(list_empty(&src_buf->dirty)); |
| BUG_ON(list_empty(&dst_buf->dirty)); |
| return ret; |
| } |
| |
| /* |
| * try to push data from one node into the next node right in the |
| * tree. |
| * |
| * returns 0 if some ptrs were pushed, < 0 if there was some horrible |
| * error, and > 0 if there was no room in the right hand block. |
| * |
| * this will only push up to 1/2 the contents of the left node over |
| */ |
| static int balance_node_right(struct ctree_root *root, |
| struct tree_buffer *dst_buf, |
| struct tree_buffer *src_buf) |
| { |
| struct node *src = &src_buf->node; |
| struct node *dst = &dst_buf->node; |
| int push_items = 0; |
| int max_push; |
| int src_nritems; |
| int dst_nritems; |
| int ret = 0; |
| |
| src_nritems = btrfs_header_nritems(&src->header); |
| dst_nritems = btrfs_header_nritems(&dst->header); |
| push_items = NODEPTRS_PER_BLOCK - dst_nritems; |
| if (push_items <= 0) { |
| return 1; |
| } |
| |
| max_push = src_nritems / 2 + 1; |
| /* don't try to empty the node */ |
| if (max_push > src_nritems) |
| return 1; |
| if (max_push < push_items) |
| push_items = max_push; |
| |
| memmove(dst->keys + push_items, dst->keys, |
| dst_nritems * sizeof(struct key)); |
| memmove(dst->blockptrs + push_items, dst->blockptrs, |
| dst_nritems * sizeof(u64)); |
| memcpy(dst->keys, src->keys + src_nritems - push_items, |
| push_items * sizeof(struct key)); |
| memcpy(dst->blockptrs, src->blockptrs + src_nritems - push_items, |
| push_items * sizeof(u64)); |
| |
| btrfs_set_header_nritems(&src->header, src_nritems - push_items); |
| btrfs_set_header_nritems(&dst->header, dst_nritems + push_items); |
| |
| BUG_ON(list_empty(&src_buf->dirty)); |
| BUG_ON(list_empty(&dst_buf->dirty)); |
| return ret; |
| } |
| |
| /* |
| * helper function to insert a new root level in the tree. |
| * A new node is allocated, and a single item is inserted to |
| * point to the existing root |
| * |
| * returns zero on success or < 0 on failure. |
| */ |
| static int insert_new_root(struct ctree_root *root, |
| struct ctree_path *path, int level) |
| { |
| struct tree_buffer *t; |
| struct node *lower; |
| struct node *c; |
| struct key *lower_key; |
| |
| BUG_ON(path->nodes[level]); |
| BUG_ON(path->nodes[level-1] != root->node); |
| |
| t = alloc_free_block(root); |
| c = &t->node; |
| memset(c, 0, sizeof(c)); |
| btrfs_set_header_nritems(&c->header, 1); |
| btrfs_set_header_level(&c->header, level); |
| btrfs_set_header_blocknr(&c->header, t->blocknr); |
| btrfs_set_header_parentid(&c->header, |
| btrfs_header_parentid(&root->node->node.header)); |
| lower = &path->nodes[level-1]->node; |
| if (btrfs_is_leaf(lower)) |
| lower_key = &((struct leaf *)lower)->items[0].key; |
| else |
| lower_key = lower->keys; |
| memcpy(c->keys, lower_key, sizeof(struct key)); |
| c->blockptrs[0] = path->nodes[level-1]->blocknr; |
| /* the super has an extra ref to root->node */ |
| tree_block_release(root, root->node); |
| root->node = t; |
| t->count++; |
| path->nodes[level] = t; |
| path->slots[level] = 0; |
| return 0; |
| } |
| |
| /* |
| * worker function to insert a single pointer in a node. |
| * the node should have enough room for the pointer already |
| * |
| * slot and level indicate where you want the key to go, and |
| * blocknr is the block the key points to. |
| * |
| * returns zero on success and < 0 on any error |
| */ |
| static int insert_ptr(struct ctree_root *root, |
| struct ctree_path *path, struct key *key, |
| u64 blocknr, int slot, int level) |
| { |
| struct node *lower; |
| int nritems; |
| |
| BUG_ON(!path->nodes[level]); |
| lower = &path->nodes[level]->node; |
| nritems = btrfs_header_nritems(&lower->header); |
| if (slot > nritems) |
| BUG(); |
| if (nritems == NODEPTRS_PER_BLOCK) |
| BUG(); |
| if (slot != nritems) { |
| memmove(lower->keys + slot + 1, lower->keys + slot, |
| (nritems - slot) * sizeof(struct key)); |
| memmove(lower->blockptrs + slot + 1, lower->blockptrs + slot, |
| (nritems - slot) * sizeof(u64)); |
| } |
| memcpy(lower->keys + slot, key, sizeof(struct key)); |
| lower->blockptrs[slot] = blocknr; |
| btrfs_set_header_nritems(&lower->header, nritems + 1); |
| if (lower->keys[1].objectid == 0) |
| BUG(); |
| BUG_ON(list_empty(&path->nodes[level]->dirty)); |
| return 0; |
| } |
| |
| /* |
| * split the node at the specified level in path in two. |
| * The path is corrected to point to the appropriate node after the split |
| * |
| * Before splitting this tries to make some room in the node by pushing |
| * left and right, if either one works, it returns right away. |
| * |
| * returns 0 on success and < 0 on failure |
| */ |
| static int split_node(struct ctree_root *root, struct ctree_path *path, |
| int level) |
| { |
| struct tree_buffer *t; |
| struct node *c; |
| struct tree_buffer *split_buffer; |
| struct node *split; |
| int mid; |
| int ret; |
| int wret; |
| u32 c_nritems; |
| |
| t = path->nodes[level]; |
| c = &t->node; |
| if (t == root->node) { |
| /* trying to split the root, lets make a new one */ |
| ret = insert_new_root(root, path, level + 1); |
| if (ret) |
| return ret; |
| } |
| c_nritems = btrfs_header_nritems(&c->header); |
| split_buffer = alloc_free_block(root); |
| split = &split_buffer->node; |
| btrfs_set_header_flags(&split->header, btrfs_header_flags(&c->header)); |
| btrfs_set_header_blocknr(&split->header, split_buffer->blocknr); |
| btrfs_set_header_parentid(&split->header, |
| btrfs_header_parentid(&root->node->node.header)); |
| mid = (c_nritems + 1) / 2; |
| memcpy(split->keys, c->keys + mid, |
| (c_nritems - mid) * sizeof(struct key)); |
| memcpy(split->blockptrs, c->blockptrs + mid, |
| (c_nritems - mid) * sizeof(u64)); |
| btrfs_set_header_nritems(&split->header, c_nritems - mid); |
| btrfs_set_header_nritems(&c->header, mid); |
| ret = 0; |
| |
| BUG_ON(list_empty(&t->dirty)); |
| wret = insert_ptr(root, path, split->keys, split_buffer->blocknr, |
| path->slots[level + 1] + 1, level + 1); |
| if (wret) |
| ret = wret; |
| |
| if (path->slots[level] >= mid) { |
| path->slots[level] -= mid; |
| tree_block_release(root, t); |
| path->nodes[level] = split_buffer; |
| path->slots[level + 1] += 1; |
| } else { |
| tree_block_release(root, split_buffer); |
| } |
| return ret; |
| } |
| |
| /* |
| * how many bytes are required to store the items in a leaf. start |
| * and nr indicate which items in the leaf to check. This totals up the |
| * space used both by the item structs and the item data |
| */ |
| static int leaf_space_used(struct leaf *l, int start, int nr) |
| { |
| int data_len; |
| int end = start + nr - 1; |
| |
| if (!nr) |
| return 0; |
| data_len = l->items[start].offset + l->items[start].size; |
| data_len = data_len - l->items[end].offset; |
| data_len += sizeof(struct item) * nr; |
| return data_len; |
| } |
| |
| /* |
| * push some data in the path leaf to the right, trying to free up at |
| * least data_size bytes. returns zero if the push worked, nonzero otherwise |
| * |
| * returns 1 if the push failed because the other node didn't have enough |
| * room, 0 if everything worked out and < 0 if there were major errors. |
| */ |
| static int push_leaf_right(struct ctree_root *root, struct ctree_path *path, |
| int data_size) |
| { |
| struct tree_buffer *left_buf = path->nodes[0]; |
| struct leaf *left = &left_buf->leaf; |
| struct leaf *right; |
| struct tree_buffer *right_buf; |
| struct tree_buffer *upper; |
| int slot; |
| int i; |
| int free_space; |
| int push_space = 0; |
| int push_items = 0; |
| struct item *item; |
| u32 left_nritems; |
| u32 right_nritems; |
| |
| slot = path->slots[1]; |
| if (!path->nodes[1]) { |
| return 1; |
| } |
| upper = path->nodes[1]; |
| if (slot >= btrfs_header_nritems(&upper->node.header) - 1) { |
| return 1; |
| } |
| right_buf = read_tree_block(root, upper->node.blockptrs[slot + 1]); |
| right = &right_buf->leaf; |
| free_space = leaf_free_space(right); |
| if (free_space < data_size + sizeof(struct item)) { |
| tree_block_release(root, right_buf); |
| return 1; |
| } |
| /* cow and double check */ |
| btrfs_cow_block(root, right_buf, upper, slot + 1, &right_buf); |
| right = &right_buf->leaf; |
| free_space = leaf_free_space(right); |
| if (free_space < data_size + sizeof(struct item)) { |
| tree_block_release(root, right_buf); |
| return 1; |
| } |
| |
| left_nritems = btrfs_header_nritems(&left->header); |
| for (i = left_nritems - 1; i >= 0; i--) { |
| item = left->items + i; |
| if (path->slots[0] == i) |
| push_space += data_size + sizeof(*item); |
| if (item->size + sizeof(*item) + push_space > free_space) |
| break; |
| push_items++; |
| push_space += item->size + sizeof(*item); |
| } |
| if (push_items == 0) { |
| tree_block_release(root, right_buf); |
| return 1; |
| } |
| right_nritems = btrfs_header_nritems(&right->header); |
| /* push left to right */ |
| push_space = left->items[left_nritems - push_items].offset + |
| left->items[left_nritems - push_items].size; |
| push_space -= leaf_data_end(left); |
| /* make room in the right data area */ |
| memmove(right->data + leaf_data_end(right) - push_space, |
| right->data + leaf_data_end(right), |
| LEAF_DATA_SIZE - leaf_data_end(right)); |
| /* copy from the left data area */ |
| memcpy(right->data + LEAF_DATA_SIZE - push_space, |
| left->data + leaf_data_end(left), |
| push_space); |
| memmove(right->items + push_items, right->items, |
| right_nritems * sizeof(struct item)); |
| /* copy the items from left to right */ |
| memcpy(right->items, left->items + left_nritems - push_items, |
| push_items * sizeof(struct item)); |
| |
| /* update the item pointers */ |
| right_nritems += push_items; |
| btrfs_set_header_nritems(&right->header, right_nritems); |
| push_space = LEAF_DATA_SIZE; |
| for (i = 0; i < right_nritems; i++) { |
| right->items[i].offset = push_space - right->items[i].size; |
| push_space = right->items[i].offset; |
| } |
| left_nritems -= push_items; |
| btrfs_set_header_nritems(&left->header, left_nritems); |
| |
| BUG_ON(list_empty(&left_buf->dirty)); |
| BUG_ON(list_empty(&right_buf->dirty)); |
| memcpy(upper->node.keys + slot + 1, |
| &right->items[0].key, sizeof(struct key)); |
| BUG_ON(list_empty(&upper->dirty)); |
| |
| /* then fixup the leaf pointer in the path */ |
| if (path->slots[0] >= left_nritems) { |
| path->slots[0] -= left_nritems; |
| tree_block_release(root, path->nodes[0]); |
| path->nodes[0] = right_buf; |
| path->slots[1] += 1; |
| } else { |
| tree_block_release(root, right_buf); |
| } |
| return 0; |
| } |
| /* |
| * 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 |
| */ |
| static int push_leaf_left(struct ctree_root *root, struct ctree_path *path, |
| int data_size) |
| { |
| struct tree_buffer *right_buf = path->nodes[0]; |
| struct leaf *right = &right_buf->leaf; |
| struct tree_buffer *t; |
| struct leaf *left; |
| int slot; |
| int i; |
| int free_space; |
| int push_space = 0; |
| int push_items = 0; |
| struct item *item; |
| u32 old_left_nritems; |
| int ret = 0; |
| int wret; |
| |
| slot = path->slots[1]; |
| if (slot == 0) { |
| return 1; |
| } |
| if (!path->nodes[1]) { |
| return 1; |
| } |
| t = read_tree_block(root, path->nodes[1]->node.blockptrs[slot - 1]); |
| left = &t->leaf; |
| free_space = leaf_free_space(left); |
| if (free_space < data_size + sizeof(struct item)) { |
| tree_block_release(root, t); |
| return 1; |
| } |
| |
| /* cow and double check */ |
| btrfs_cow_block(root, t, path->nodes[1], slot - 1, &t); |
| left = &t->leaf; |
| free_space = leaf_free_space(left); |
| if (free_space < data_size + sizeof(struct item)) { |
| tree_block_release(root, t); |
| return 1; |
| } |
| |
| for (i = 0; i < btrfs_header_nritems(&right->header); i++) { |
| item = right->items + i; |
| if (path->slots[0] == i) |
| push_space += data_size + sizeof(*item); |
| if (item->size + sizeof(*item) + push_space > free_space) |
| break; |
| push_items++; |
| push_space += item->size + sizeof(*item); |
| } |
| if (push_items == 0) { |
| tree_block_release(root, t); |
| return 1; |
| } |
| /* push data from right to left */ |
| memcpy(left->items + btrfs_header_nritems(&left->header), |
| right->items, push_items * sizeof(struct item)); |
| push_space = LEAF_DATA_SIZE - right->items[push_items -1].offset; |
| memcpy(left->data + leaf_data_end(left) - push_space, |
| right->data + right->items[push_items - 1].offset, |
| push_space); |
| old_left_nritems = btrfs_header_nritems(&left->header); |
| BUG_ON(old_left_nritems < 0); |
| |
| for(i = old_left_nritems; i < old_left_nritems + push_items; i++) { |
| left->items[i].offset -= LEAF_DATA_SIZE - |
| left->items[old_left_nritems -1].offset; |
| } |
| btrfs_set_header_nritems(&left->header, old_left_nritems + push_items); |
| |
| /* fixup right node */ |
| push_space = right->items[push_items-1].offset - leaf_data_end(right); |
| memmove(right->data + LEAF_DATA_SIZE - push_space, right->data + |
| leaf_data_end(right), push_space); |
| memmove(right->items, right->items + push_items, |
| (btrfs_header_nritems(&right->header) - push_items) * |
| sizeof(struct item)); |
| btrfs_set_header_nritems(&right->header, |
| btrfs_header_nritems(&right->header) - |
| push_items); |
| push_space = LEAF_DATA_SIZE; |
| |
| for (i = 0; i < btrfs_header_nritems(&right->header); i++) { |
| right->items[i].offset = push_space - right->items[i].size; |
| push_space = right->items[i].offset; |
| } |
| |
| BUG_ON(list_empty(&t->dirty)); |
| BUG_ON(list_empty(&right_buf->dirty)); |
| |
| wret = fixup_low_keys(root, path, &right->items[0].key, 1); |
| if (wret) |
| ret = wret; |
| |
| /* then fixup the leaf pointer in the path */ |
| if (path->slots[0] < push_items) { |
| path->slots[0] += old_left_nritems; |
| tree_block_release(root, path->nodes[0]); |
| path->nodes[0] = t; |
| path->slots[1] -= 1; |
| } else { |
| tree_block_release(root, t); |
| path->slots[0] -= push_items; |
| } |
| BUG_ON(path->slots[0] < 0); |
| return ret; |
| } |
| |
| /* |
| * split the path's leaf in two, making sure there is at least data_size |
| * available for the resulting leaf level of the path. |
| * |
| * returns 0 if all went well and < 0 on failure. |
| */ |
| static int split_leaf(struct ctree_root *root, struct ctree_path *path, |
| int data_size) |
| { |
| struct tree_buffer *l_buf; |
| struct leaf *l; |
| u32 nritems; |
| int mid; |
| int slot; |
| struct leaf *right; |
| struct tree_buffer *right_buffer; |
| int space_needed = data_size + sizeof(struct item); |
| int data_copy_size; |
| int rt_data_off; |
| int i; |
| int ret; |
| int wret; |
| |
| l_buf = path->nodes[0]; |
| l = &l_buf->leaf; |
| |
| /* did the pushes work? */ |
| if (leaf_free_space(l) >= sizeof(struct item) + data_size) |
| return 0; |
| |
| if (!path->nodes[1]) { |
| ret = insert_new_root(root, path, 1); |
| if (ret) |
| return ret; |
| } |
| slot = path->slots[0]; |
| nritems = btrfs_header_nritems(&l->header); |
| mid = (nritems + 1)/ 2; |
| right_buffer = alloc_free_block(root); |
| BUG_ON(!right_buffer); |
| BUG_ON(mid == nritems); |
| right = &right_buffer->leaf; |
| memset(right, 0, sizeof(*right)); |
| if (mid <= slot) { |
| /* FIXME, just alloc a new leaf here */ |
| if (leaf_space_used(l, mid, nritems - mid) + space_needed > |
| LEAF_DATA_SIZE) |
| BUG(); |
| } else { |
| /* FIXME, just alloc a new leaf here */ |
| if (leaf_space_used(l, 0, mid + 1) + space_needed > |
| LEAF_DATA_SIZE) |
| BUG(); |
| } |
| btrfs_set_header_nritems(&right->header, nritems - mid); |
| btrfs_set_header_blocknr(&right->header, right_buffer->blocknr); |
| btrfs_set_header_level(&right->header, 0); |
| btrfs_set_header_parentid(&right->header, |
| btrfs_header_parentid(&root->node->node.header)); |
| data_copy_size = l->items[mid].offset + l->items[mid].size - |
| leaf_data_end(l); |
| memcpy(right->items, l->items + mid, |
| (nritems - mid) * sizeof(struct item)); |
| memcpy(right->data + LEAF_DATA_SIZE - data_copy_size, |
| l->data + leaf_data_end(l), data_copy_size); |
| rt_data_off = LEAF_DATA_SIZE - |
| (l->items[mid].offset + l->items[mid].size); |
| |
| for (i = 0; i < btrfs_header_nritems(&right->header); i++) |
| right->items[i].offset += rt_data_off; |
| |
| btrfs_set_header_nritems(&l->header, mid); |
| ret = 0; |
| wret = insert_ptr(root, path, &right->items[0].key, |
| right_buffer->blocknr, path->slots[1] + 1, 1); |
| if (wret) |
| ret = wret; |
| BUG_ON(list_empty(&right_buffer->dirty)); |
| BUG_ON(list_empty(&l_buf->dirty)); |
| BUG_ON(path->slots[0] != slot); |
| if (mid <= slot) { |
| tree_block_release(root, path->nodes[0]); |
| path->nodes[0] = right_buffer; |
| path->slots[0] -= mid; |
| path->slots[1] += 1; |
| } else |
| tree_block_release(root, right_buffer); |
| BUG_ON(path->slots[0] < 0); |
| return ret; |
| } |
| |
| /* |
| * Given a key and some data, insert an item into the tree. |
| * This does all the path init required, making room in the tree if needed. |
| */ |
| int insert_item(struct ctree_root *root, struct key *key, |
| void *data, int data_size) |
| { |
| int ret = 0; |
| int slot; |
| int slot_orig; |
| struct leaf *leaf; |
| struct tree_buffer *leaf_buf; |
| u32 nritems; |
| unsigned int data_end; |
| struct ctree_path path; |
| |
| /* create a root if there isn't one */ |
| if (!root->node) |
| BUG(); |
| init_path(&path); |
| ret = search_slot(root, key, &path, data_size, 1); |
| if (ret == 0) { |
| release_path(root, &path); |
| return -EEXIST; |
| } |
| if (ret < 0) |
| goto out; |
| |
| slot_orig = path.slots[0]; |
| leaf_buf = path.nodes[0]; |
| leaf = &leaf_buf->leaf; |
| |
| nritems = btrfs_header_nritems(&leaf->header); |
| data_end = leaf_data_end(leaf); |
| |
| if (leaf_free_space(leaf) < sizeof(struct item) + data_size) |
| BUG(); |
| |
| slot = path.slots[0]; |
| BUG_ON(slot < 0); |
| if (slot != nritems) { |
| int i; |
| unsigned int old_data = leaf->items[slot].offset + |
| leaf->items[slot].size; |
| |
| /* |
| * item0..itemN ... dataN.offset..dataN.size .. data0.size |
| */ |
| /* first correct the data pointers */ |
| for (i = slot; i < nritems; i++) |
| leaf->items[i].offset -= data_size; |
| |
| /* shift the items */ |
| memmove(leaf->items + slot + 1, leaf->items + slot, |
| (nritems - slot) * sizeof(struct item)); |
| |
| /* shift the data */ |
| memmove(leaf->data + data_end - data_size, leaf->data + |
| data_end, old_data - data_end); |
| data_end = old_data; |
| } |
| /* copy the new data in */ |
| memcpy(&leaf->items[slot].key, key, sizeof(struct key)); |
| leaf->items[slot].offset = data_end - data_size; |
| leaf->items[slot].size = data_size; |
| memcpy(leaf->data + data_end - data_size, data, data_size); |
| btrfs_set_header_nritems(&leaf->header, nritems + 1); |
| |
| ret = 0; |
| if (slot == 0) |
| ret = fixup_low_keys(root, &path, key, 1); |
| |
| BUG_ON(list_empty(&leaf_buf->dirty)); |
| if (leaf_free_space(leaf) < 0) |
| BUG(); |
| check_leaf(&path, 0); |
| out: |
| release_path(root, &path); |
| return ret; |
| } |
| |
| /* |
| * delete the pointer from a given node. |
| * |
| * If the delete empties a node, the node is removed from the tree, |
| * continuing all the way the root if required. The root is converted into |
| * a leaf if all the nodes are emptied. |
| */ |
| static int del_ptr(struct ctree_root *root, struct ctree_path *path, int level, |
| int slot) |
| { |
| struct node *node; |
| struct tree_buffer *parent = path->nodes[level]; |
| u32 nritems; |
| int ret = 0; |
| int wret; |
| |
| node = &parent->node; |
| nritems = btrfs_header_nritems(&node->header); |
| if (slot != nritems -1) { |
| memmove(node->keys + slot, node->keys + slot + 1, |
| sizeof(struct key) * (nritems - slot - 1)); |
| memmove(node->blockptrs + slot, |
| node->blockptrs + slot + 1, |
| sizeof(u64) * (nritems - slot - 1)); |
| } |
| nritems--; |
| btrfs_set_header_nritems(&node->header, nritems); |
| if (nritems == 0 && parent == root->node) { |
| BUG_ON(btrfs_header_level(&root->node->node.header) != 1); |
| /* just turn the root into a leaf and break */ |
| btrfs_set_header_level(&root->node->node.header, 0); |
| } else if (slot == 0) { |
| wret = fixup_low_keys(root, path, node->keys, level + 1); |
| if (wret) |
| ret = wret; |
| } |
| BUG_ON(list_empty(&parent->dirty)); |
| return ret; |
| } |
| |
| /* |
| * delete the item at the leaf level in path. If that empties |
| * the leaf, remove it from the tree |
| */ |
| int del_item(struct ctree_root *root, struct ctree_path *path) |
| { |
| int slot; |
| struct leaf *leaf; |
| struct tree_buffer *leaf_buf; |
| int doff; |
| int dsize; |
| int ret = 0; |
| int wret; |
| u32 nritems; |
| |
| leaf_buf = path->nodes[0]; |
| leaf = &leaf_buf->leaf; |
| slot = path->slots[0]; |
| doff = leaf->items[slot].offset; |
| dsize = leaf->items[slot].size; |
| nritems = btrfs_header_nritems(&leaf->header); |
| |
| if (slot != nritems - 1) { |
| int i; |
| int data_end = leaf_data_end(leaf); |
| memmove(leaf->data + data_end + dsize, |
| leaf->data + data_end, |
| doff - data_end); |
| for (i = slot + 1; i < nritems; i++) |
| leaf->items[i].offset += dsize; |
| memmove(leaf->items + slot, leaf->items + slot + 1, |
| sizeof(struct item) * |
| (nritems - slot - 1)); |
| } |
| btrfs_set_header_nritems(&leaf->header, nritems - 1); |
| nritems--; |
| /* delete the leaf if we've emptied it */ |
| if (nritems == 0) { |
| if (leaf_buf == root->node) { |
| btrfs_set_header_level(&leaf->header, 0); |
| BUG_ON(list_empty(&leaf_buf->dirty)); |
| } else { |
| clean_tree_block(root, leaf_buf); |
| wret = del_ptr(root, path, 1, path->slots[1]); |
| if (wret) |
| ret = wret; |
| wret = free_extent(root, leaf_buf->blocknr, 1); |
| if (wret) |
| ret = wret; |
| } |
| } else { |
| int used = leaf_space_used(leaf, 0, nritems); |
| if (slot == 0) { |
| wret = fixup_low_keys(root, path, |
| &leaf->items[0].key, 1); |
| if (wret) |
| ret = wret; |
| } |
| BUG_ON(list_empty(&leaf_buf->dirty)); |
| |
| /* delete the leaf if it is mostly empty */ |
| if (used < LEAF_DATA_SIZE / 3) { |
| /* push_leaf_left fixes the path. |
| * make sure the path still points to our leaf |
| * for possible call to del_ptr below |
| */ |
| slot = path->slots[1]; |
| leaf_buf->count++; |
| wret = push_leaf_left(root, path, 1); |
| if (wret < 0) |
| ret = wret; |
| if (path->nodes[0] == leaf_buf && |
| btrfs_header_nritems(&leaf->header)) { |
| wret = push_leaf_right(root, path, 1); |
| if (wret < 0) |
| ret = wret; |
| } |
| if (btrfs_header_nritems(&leaf->header) == 0) { |
| u64 blocknr = leaf_buf->blocknr; |
| clean_tree_block(root, leaf_buf); |
| wret = del_ptr(root, path, 1, slot); |
| if (wret) |
| ret = wret; |
| tree_block_release(root, leaf_buf); |
| wret = free_extent(root, blocknr, 1); |
| if (wret) |
| ret = wret; |
| } else { |
| tree_block_release(root, leaf_buf); |
| } |
| } |
| } |
| return ret; |
| } |
| |
| /* |
| * walk up the tree as far as required to find the next leaf. |
| * returns 0 if it found something or 1 if there are no greater leaves. |
| * returns < 0 on io errors. |
| */ |
| int next_leaf(struct ctree_root *root, struct ctree_path *path) |
| { |
| int slot; |
| int level = 1; |
| u64 blocknr; |
| struct tree_buffer *c; |
| struct tree_buffer *next = NULL; |
| |
| while(level < MAX_LEVEL) { |
| if (!path->nodes[level]) |
| return 1; |
| slot = path->slots[level] + 1; |
| c = path->nodes[level]; |
| if (slot >= btrfs_header_nritems(&c->node.header)) { |
| level++; |
| continue; |
| } |
| blocknr = c->node.blockptrs[slot]; |
| if (next) |
| tree_block_release(root, next); |
| next = read_tree_block(root, blocknr); |
| break; |
| } |
| path->slots[level] = slot; |
| while(1) { |
| level--; |
| c = path->nodes[level]; |
| tree_block_release(root, c); |
| path->nodes[level] = next; |
| path->slots[level] = 0; |
| if (!level) |
| break; |
| next = read_tree_block(root, next->node.blockptrs[0]); |
| } |
| return 0; |
| } |
| |
| |