1 /*
2 * Elastic Binary Trees - macros for Indirect Multi-Byte data nodes.
3 * Version 6.0.6
4 * (C) 2002-2011 - Willy Tarreau <w@1wt.eu>
5 *
6 * This library is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Lesser General Public
8 * License as published by the Free Software Foundation, version 2.1
9 * exclusively.
10 *
11 * This library is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Lesser General Public License for more details.
15 *
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with this library; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19 */
20
21 #ifndef _EBIMTREE_H
22 #define _EBIMTREE_H
23
24 #include <string.h>
25 #include "ebtree.h"
26 #include "ebpttree.h"
27
28 /* These functions and macros rely on Pointer nodes and use the <key> entry as
29 * a pointer to an indirect key. Most operations are performed using ebpt_*.
30 */
31
32 /* The following functions are not inlined by default. They are declared
33 * in ebimtree.c, which simply relies on their inline version.
34 */
35 REGPRM3 struct ebpt_node *ebim_lookup(struct eb_root *root, const void *x, unsigned int len);
36 REGPRM3 struct ebpt_node *ebim_insert(struct eb_root *root, struct ebpt_node *new, unsigned int len);
37
38 /* Find the first occurrence of a key of a least <len> bytes matching <x> in the
39 * tree <root>. The caller is responsible for ensuring that <len> will not exceed
40 * the common parts between the tree's keys and <x>. In case of multiple matches,
41 * the leftmost node is returned. This means that this function can be used to
42 * lookup string keys by prefix if all keys in the tree are zero-terminated. If
43 * no match is found, NULL is returned. Returns first node if <len> is zero.
44 */
45 static forceinline struct ebpt_node *
__ebim_lookup(struct eb_root * root,const void * x,unsigned int len)46 __ebim_lookup(struct eb_root *root, const void *x, unsigned int len)
47 {
48 struct ebpt_node *node;
49 eb_troot_t *troot;
50 int pos, side;
51 int node_bit;
52
53 troot = root->b[EB_LEFT];
54 if (unlikely(troot == NULL))
55 goto ret_null;
56
57 if (unlikely(len == 0))
58 goto walk_down;
59
60 pos = 0;
61 while (1) {
62 if (eb_gettag(troot) == EB_LEAF) {
63 node = container_of(eb_untag(troot, EB_LEAF),
64 struct ebpt_node, node.branches);
65 if (memcmp(node->key + pos, x, len) != 0)
66 goto ret_null;
67 else
68 goto ret_node;
69 }
70 node = container_of(eb_untag(troot, EB_NODE),
71 struct ebpt_node, node.branches);
72
73 node_bit = node->node.bit;
74 if (node_bit < 0) {
75 /* We have a dup tree now. Either it's for the same
76 * value, and we walk down left, or it's a different
77 * one and we don't have our key.
78 */
79 if (memcmp(node->key + pos, x, len) != 0)
80 goto ret_null;
81 else
82 goto walk_left;
83 }
84
85 /* OK, normal data node, let's walk down. We check if all full
86 * bytes are equal, and we start from the last one we did not
87 * completely check. We stop as soon as we reach the last byte,
88 * because we must decide to go left/right or abort.
89 */
90 node_bit = ~node_bit + (pos << 3) + 8; // = (pos<<3) + (7 - node_bit)
91 if (node_bit < 0) {
92 /* This surprising construction gives better performance
93 * because gcc does not try to reorder the loop. Tested to
94 * be fine with 2.95 to 4.2.
95 */
96 while (1) {
97 if (*(unsigned char*)(node->key + pos++) ^ *(unsigned char*)(x++))
98 goto ret_null; /* more than one full byte is different */
99 if (--len == 0)
100 goto walk_left; /* return first node if all bytes matched */
101 node_bit += 8;
102 if (node_bit >= 0)
103 break;
104 }
105 }
106
107 /* here we know that only the last byte differs, so node_bit < 8.
108 * We have 2 possibilities :
109 * - more than the last bit differs => return NULL
110 * - walk down on side = (x[pos] >> node_bit) & 1
111 */
112 side = *(unsigned char *)x >> node_bit;
113 if (((*(unsigned char*)(node->key + pos) >> node_bit) ^ side) > 1)
114 goto ret_null;
115 side &= 1;
116 troot = node->node.branches.b[side];
117 }
118 walk_left:
119 troot = node->node.branches.b[EB_LEFT];
120 walk_down:
121 while (eb_gettag(troot) != EB_LEAF)
122 troot = (eb_untag(troot, EB_NODE))->b[EB_LEFT];
123 node = container_of(eb_untag(troot, EB_LEAF),
124 struct ebpt_node, node.branches);
125 ret_node:
126 return node;
127 ret_null:
128 return NULL;
129 }
130
131 /* Insert ebpt_node <new> into subtree starting at node root <root>.
132 * Only new->key needs be set with the key. The ebpt_node is returned.
133 * If root->b[EB_RGHT]==1, the tree may only contain unique keys. The
134 * len is specified in bytes.
135 */
136 static forceinline struct ebpt_node *
__ebim_insert(struct eb_root * root,struct ebpt_node * new,unsigned int len)137 __ebim_insert(struct eb_root *root, struct ebpt_node *new, unsigned int len)
138 {
139 struct ebpt_node *old;
140 unsigned int side;
141 eb_troot_t *troot;
142 eb_troot_t *root_right;
143 int diff;
144 int bit;
145 int old_node_bit;
146
147 side = EB_LEFT;
148 troot = root->b[EB_LEFT];
149 root_right = root->b[EB_RGHT];
150 if (unlikely(troot == NULL)) {
151 /* Tree is empty, insert the leaf part below the left branch */
152 root->b[EB_LEFT] = eb_dotag(&new->node.branches, EB_LEAF);
153 new->node.leaf_p = eb_dotag(root, EB_LEFT);
154 new->node.node_p = NULL; /* node part unused */
155 return new;
156 }
157
158 len <<= 3;
159
160 /* The tree descent is fairly easy :
161 * - first, check if we have reached a leaf node
162 * - second, check if we have gone too far
163 * - third, reiterate
164 * Everywhere, we use <new> for the node node we are inserting, <root>
165 * for the node we attach it to, and <old> for the node we are
166 * displacing below <new>. <troot> will always point to the future node
167 * (tagged with its type). <side> carries the side the node <new> is
168 * attached to below its parent, which is also where previous node
169 * was attached.
170 */
171
172 bit = 0;
173 while (1) {
174 if (unlikely(eb_gettag(troot) == EB_LEAF)) {
175 eb_troot_t *new_left, *new_rght;
176 eb_troot_t *new_leaf, *old_leaf;
177
178 old = container_of(eb_untag(troot, EB_LEAF),
179 struct ebpt_node, node.branches);
180
181 new_left = eb_dotag(&new->node.branches, EB_LEFT);
182 new_rght = eb_dotag(&new->node.branches, EB_RGHT);
183 new_leaf = eb_dotag(&new->node.branches, EB_LEAF);
184 old_leaf = eb_dotag(&old->node.branches, EB_LEAF);
185
186 new->node.node_p = old->node.leaf_p;
187
188 /* Right here, we have 3 possibilities :
189 * - the tree does not contain the key, and we have
190 * new->key < old->key. We insert new above old, on
191 * the left ;
192 *
193 * - the tree does not contain the key, and we have
194 * new->key > old->key. We insert new above old, on
195 * the right ;
196 *
197 * - the tree does contain the key, which implies it
198 * is alone. We add the new key next to it as a
199 * first duplicate.
200 *
201 * The last two cases can easily be partially merged.
202 */
203 bit = equal_bits(new->key, old->key, bit, len);
204
205 /* Note: we can compare more bits than the current node's because as
206 * long as they are identical, we know we descend along the correct
207 * side. However we don't want to start to compare past the end.
208 */
209 diff = 0;
210 if (((unsigned)bit >> 3) < len)
211 diff = cmp_bits(new->key, old->key, bit);
212
213 if (diff < 0) {
214 new->node.leaf_p = new_left;
215 old->node.leaf_p = new_rght;
216 new->node.branches.b[EB_LEFT] = new_leaf;
217 new->node.branches.b[EB_RGHT] = old_leaf;
218 } else {
219 /* we may refuse to duplicate this key if the tree is
220 * tagged as containing only unique keys.
221 */
222 if (diff == 0 && eb_gettag(root_right))
223 return old;
224
225 /* new->key >= old->key, new goes the right */
226 old->node.leaf_p = new_left;
227 new->node.leaf_p = new_rght;
228 new->node.branches.b[EB_LEFT] = old_leaf;
229 new->node.branches.b[EB_RGHT] = new_leaf;
230
231 if (diff == 0) {
232 new->node.bit = -1;
233 root->b[side] = eb_dotag(&new->node.branches, EB_NODE);
234 return new;
235 }
236 }
237 break;
238 }
239
240 /* OK we're walking down this link */
241 old = container_of(eb_untag(troot, EB_NODE),
242 struct ebpt_node, node.branches);
243 old_node_bit = old->node.bit;
244
245 /* Stop going down when we don't have common bits anymore. We
246 * also stop in front of a duplicates tree because it means we
247 * have to insert above. Note: we can compare more bits than
248 * the current node's because as long as they are identical, we
249 * know we descend along the correct side.
250 */
251 if (old_node_bit < 0) {
252 /* we're above a duplicate tree, we must compare till the end */
253 bit = equal_bits(new->key, old->key, bit, len);
254 goto dup_tree;
255 }
256 else if (bit < old_node_bit) {
257 bit = equal_bits(new->key, old->key, bit, old_node_bit);
258 }
259
260 if (bit < old_node_bit) { /* we don't have all bits in common */
261 /* The tree did not contain the key, so we insert <new> before the node
262 * <old>, and set ->bit to designate the lowest bit position in <new>
263 * which applies to ->branches.b[].
264 */
265 eb_troot_t *new_left, *new_rght;
266 eb_troot_t *new_leaf, *old_node;
267
268 dup_tree:
269 new_left = eb_dotag(&new->node.branches, EB_LEFT);
270 new_rght = eb_dotag(&new->node.branches, EB_RGHT);
271 new_leaf = eb_dotag(&new->node.branches, EB_LEAF);
272 old_node = eb_dotag(&old->node.branches, EB_NODE);
273
274 new->node.node_p = old->node.node_p;
275
276 /* Note: we can compare more bits than the current node's because as
277 * long as they are identical, we know we descend along the correct
278 * side. However we don't want to start to compare past the end.
279 */
280 diff = 0;
281 if (((unsigned)bit >> 3) < len)
282 diff = cmp_bits(new->key, old->key, bit);
283
284 if (diff < 0) {
285 new->node.leaf_p = new_left;
286 old->node.node_p = new_rght;
287 new->node.branches.b[EB_LEFT] = new_leaf;
288 new->node.branches.b[EB_RGHT] = old_node;
289 }
290 else if (diff > 0) {
291 old->node.node_p = new_left;
292 new->node.leaf_p = new_rght;
293 new->node.branches.b[EB_LEFT] = old_node;
294 new->node.branches.b[EB_RGHT] = new_leaf;
295 }
296 else {
297 struct eb_node *ret;
298 ret = eb_insert_dup(&old->node, &new->node);
299 return container_of(ret, struct ebpt_node, node);
300 }
301 break;
302 }
303
304 /* walk down */
305 root = &old->node.branches;
306 side = (((unsigned char *)new->key)[old_node_bit >> 3] >> (~old_node_bit & 7)) & 1;
307 troot = root->b[side];
308 }
309
310 /* Ok, now we are inserting <new> between <root> and <old>. <old>'s
311 * parent is already set to <new>, and the <root>'s branch is still in
312 * <side>. Update the root's leaf till we have it. Note that we can also
313 * find the side by checking the side of new->node.node_p.
314 */
315
316 /* We need the common higher bits between new->key and old->key.
317 * This number of bits is already in <bit>.
318 */
319 new->node.bit = bit;
320 root->b[side] = eb_dotag(&new->node.branches, EB_NODE);
321 return new;
322 }
323
324 #endif /* _EBIMTREE_H */
325