haproxy/include/import/ebimtree.h

325 lines
10 KiB
C
Raw Normal View History

/*
* Elastic Binary Trees - macros for Indirect Multi-Byte data nodes.
* Version 6.0.6
* (C) 2002-2011 - Willy Tarreau <w@1wt.eu>
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation, version 2.1
* exclusively.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#ifndef _EBIMTREE_H
#define _EBIMTREE_H
#include <string.h>
#include "ebtree.h"
#include "ebpttree.h"
/* These functions and macros rely on Pointer nodes and use the <key> entry as
* a pointer to an indirect key. Most operations are performed using ebpt_*.
*/
/* The following functions are not inlined by default. They are declared
* in ebimtree.c, which simply relies on their inline version.
*/
struct ebpt_node *ebim_lookup(struct eb_root *root, const void *x, unsigned int len);
struct ebpt_node *ebim_insert(struct eb_root *root, struct ebpt_node *new, unsigned int len);
/* Find the first occurrence of a key of a least <len> bytes matching <x> in the
* tree <root>. The caller is responsible for ensuring that <len> will not exceed
* the common parts between the tree's keys and <x>. In case of multiple matches,
* the leftmost node is returned. This means that this function can be used to
* lookup string keys by prefix if all keys in the tree are zero-terminated. If
* no match is found, NULL is returned. Returns first node if <len> is zero.
*/
static forceinline struct ebpt_node *
__ebim_lookup(struct eb_root *root, const void *x, unsigned int len)
{
struct ebpt_node *node;
eb_troot_t *troot;
int pos, side;
int node_bit;
troot = root->b[EB_LEFT];
if (unlikely(troot == NULL))
goto ret_null;
if (unlikely(len == 0))
goto walk_down;
pos = 0;
while (1) {
if (eb_gettag(troot) == EB_LEAF) {
node = container_of(eb_untag(troot, EB_LEAF),
struct ebpt_node, node.branches);
BUG/MEDIUM: ebtree: use a byte-per-byte memcmp() to compare memory blocks As reported in issue #689, there is a subtle bug in the ebtree code used to compared memory blocks. It stems from the platform-dependent memcmp() implementation. Original implementations used to perform a byte-per-byte comparison and to stop at the first non-matching byte, as in this old example: https://www.retro11.de/ouxr/211bsd/usr/src/lib/libc/compat-sys5/memcmp.c.html The ebtree code has been relying on this to detect the first non-matching byte when comparing keys. This is made so that a zero-terminated string can fail to match against a longer string. Over time, especially with large busses and SIMD instruction sets, multi-byte comparisons have appeared, making the processor fetch bytes past the first different byte, which could possibly be a trailing zero. This means that it's possible to read past the allocated area for a string if it was allocated by strdup(). This is not correct and definitely confuses address sanitizers. In real life the problem doesn't have visible consequences. Indeed, multi-byte comparisons are implemented so that aligned words are loaded (e.g. 512 bits at once to process a cache line at a time). So there is no way such a multi-byte access will cross a page boundary and end up reading from an unallocated zone. This is why it was never noticed before. This patch addresses this by implementing a one-byte-at-a-time memcmp() variant for ebtree, called eb_memcmp(). It's optimized for both small and long strings and guarantees to stop after the first non-matching byte. It only needs 5 instructions in the loop and was measured to be 3.2 times faster than the glibc's AVX2-optimized memcmp() on short strings (1 to 257 bytes), since that latter one comes with a significant setup cost. The break-even seems to be at 512 bytes where both version perform equally, which is way longer than what's used in general here. This fix should be backported to stable versions and reintegrated into the ebtree code.
2020-06-16 09:10:53 +00:00
if (eb_memcmp(node->key + pos, x, len) != 0)
goto ret_null;
else
goto ret_node;
}
node = container_of(eb_untag(troot, EB_NODE),
struct ebpt_node, node.branches);
node_bit = node->node.bit;
if (node_bit < 0) {
/* We have a dup tree now. Either it's for the same
* value, and we walk down left, or it's a different
* one and we don't have our key.
*/
BUG/MEDIUM: ebtree: use a byte-per-byte memcmp() to compare memory blocks As reported in issue #689, there is a subtle bug in the ebtree code used to compared memory blocks. It stems from the platform-dependent memcmp() implementation. Original implementations used to perform a byte-per-byte comparison and to stop at the first non-matching byte, as in this old example: https://www.retro11.de/ouxr/211bsd/usr/src/lib/libc/compat-sys5/memcmp.c.html The ebtree code has been relying on this to detect the first non-matching byte when comparing keys. This is made so that a zero-terminated string can fail to match against a longer string. Over time, especially with large busses and SIMD instruction sets, multi-byte comparisons have appeared, making the processor fetch bytes past the first different byte, which could possibly be a trailing zero. This means that it's possible to read past the allocated area for a string if it was allocated by strdup(). This is not correct and definitely confuses address sanitizers. In real life the problem doesn't have visible consequences. Indeed, multi-byte comparisons are implemented so that aligned words are loaded (e.g. 512 bits at once to process a cache line at a time). So there is no way such a multi-byte access will cross a page boundary and end up reading from an unallocated zone. This is why it was never noticed before. This patch addresses this by implementing a one-byte-at-a-time memcmp() variant for ebtree, called eb_memcmp(). It's optimized for both small and long strings and guarantees to stop after the first non-matching byte. It only needs 5 instructions in the loop and was measured to be 3.2 times faster than the glibc's AVX2-optimized memcmp() on short strings (1 to 257 bytes), since that latter one comes with a significant setup cost. The break-even seems to be at 512 bytes where both version perform equally, which is way longer than what's used in general here. This fix should be backported to stable versions and reintegrated into the ebtree code.
2020-06-16 09:10:53 +00:00
if (eb_memcmp(node->key + pos, x, len) != 0)
goto ret_null;
else
goto walk_left;
}
/* OK, normal data node, let's walk down. We check if all full
* bytes are equal, and we start from the last one we did not
* completely check. We stop as soon as we reach the last byte,
* because we must decide to go left/right or abort.
*/
node_bit = ~node_bit + (pos << 3) + 8; // = (pos<<3) + (7 - node_bit)
if (node_bit < 0) {
/* This surprising construction gives better performance
* because gcc does not try to reorder the loop. Tested to
* be fine with 2.95 to 4.2.
*/
while (1) {
if (*(unsigned char*)(node->key + pos++) ^ *(unsigned char*)(x++))
goto ret_null; /* more than one full byte is different */
if (--len == 0)
goto walk_left; /* return first node if all bytes matched */
node_bit += 8;
if (node_bit >= 0)
break;
}
}
/* here we know that only the last byte differs, so node_bit < 8.
* We have 2 possibilities :
* - more than the last bit differs => return NULL
* - walk down on side = (x[pos] >> node_bit) & 1
*/
side = *(unsigned char *)x >> node_bit;
if (((*(unsigned char*)(node->key + pos) >> node_bit) ^ side) > 1)
goto ret_null;
side &= 1;
troot = node->node.branches.b[side];
}
walk_left:
troot = node->node.branches.b[EB_LEFT];
walk_down:
while (eb_gettag(troot) != EB_LEAF)
troot = (eb_untag(troot, EB_NODE))->b[EB_LEFT];
node = container_of(eb_untag(troot, EB_LEAF),
struct ebpt_node, node.branches);
ret_node:
return node;
ret_null:
return NULL;
}
/* Insert ebpt_node <new> into subtree starting at node root <root>.
* Only new->key needs be set with the key. The ebpt_node is returned.
* If root->b[EB_RGHT]==1, the tree may only contain unique keys. The
* len is specified in bytes.
*/
static forceinline struct ebpt_node *
__ebim_insert(struct eb_root *root, struct ebpt_node *new, unsigned int len)
{
struct ebpt_node *old;
unsigned int side;
eb_troot_t *troot;
eb_troot_t *root_right;
int diff;
int bit;
int old_node_bit;
side = EB_LEFT;
troot = root->b[EB_LEFT];
root_right = root->b[EB_RGHT];
if (unlikely(troot == NULL)) {
/* Tree is empty, insert the leaf part below the left branch */
root->b[EB_LEFT] = eb_dotag(&new->node.branches, EB_LEAF);
new->node.leaf_p = eb_dotag(root, EB_LEFT);
new->node.node_p = NULL; /* node part unused */
return new;
}
len <<= 3;
/* The tree descent is fairly easy :
* - first, check if we have reached a leaf node
* - second, check if we have gone too far
* - third, reiterate
* Everywhere, we use <new> for the node node we are inserting, <root>
* for the node we attach it to, and <old> for the node we are
* displacing below <new>. <troot> will always point to the future node
* (tagged with its type). <side> carries the side the node <new> is
* attached to below its parent, which is also where previous node
* was attached.
*/
bit = 0;
while (1) {
if (unlikely(eb_gettag(troot) == EB_LEAF)) {
eb_troot_t *new_left, *new_rght;
eb_troot_t *new_leaf, *old_leaf;
old = container_of(eb_untag(troot, EB_LEAF),
struct ebpt_node, node.branches);
new_left = eb_dotag(&new->node.branches, EB_LEFT);
new_rght = eb_dotag(&new->node.branches, EB_RGHT);
new_leaf = eb_dotag(&new->node.branches, EB_LEAF);
old_leaf = eb_dotag(&old->node.branches, EB_LEAF);
new->node.node_p = old->node.leaf_p;
/* Right here, we have 3 possibilities :
* - the tree does not contain the key, and we have
* new->key < old->key. We insert new above old, on
* the left ;
*
* - the tree does not contain the key, and we have
* new->key > old->key. We insert new above old, on
* the right ;
*
* - the tree does contain the key, which implies it
* is alone. We add the new key next to it as a
* first duplicate.
*
* The last two cases can easily be partially merged.
*/
bit = equal_bits(new->key, old->key, bit, len);
/* Note: we can compare more bits than the current node's because as
* long as they are identical, we know we descend along the correct
* side. However we don't want to start to compare past the end.
*/
diff = 0;
if (((unsigned)bit >> 3) < len)
diff = cmp_bits(new->key, old->key, bit);
if (diff < 0) {
new->node.leaf_p = new_left;
old->node.leaf_p = new_rght;
new->node.branches.b[EB_LEFT] = new_leaf;
new->node.branches.b[EB_RGHT] = old_leaf;
} else {
/* we may refuse to duplicate this key if the tree is
* tagged as containing only unique keys.
*/
if (diff == 0 && eb_gettag(root_right))
return old;
/* new->key >= old->key, new goes the right */
old->node.leaf_p = new_left;
new->node.leaf_p = new_rght;
new->node.branches.b[EB_LEFT] = old_leaf;
new->node.branches.b[EB_RGHT] = new_leaf;
if (diff == 0) {
new->node.bit = -1;
root->b[side] = eb_dotag(&new->node.branches, EB_NODE);
return new;
}
}
break;
}
/* OK we're walking down this link */
old = container_of(eb_untag(troot, EB_NODE),
struct ebpt_node, node.branches);
old_node_bit = old->node.bit;
/* Stop going down when we don't have common bits anymore. We
* also stop in front of a duplicates tree because it means we
* have to insert above. Note: we can compare more bits than
* the current node's because as long as they are identical, we
* know we descend along the correct side.
*/
if (old_node_bit < 0) {
/* we're above a duplicate tree, we must compare till the end */
bit = equal_bits(new->key, old->key, bit, len);
goto dup_tree;
}
else if (bit < old_node_bit) {
bit = equal_bits(new->key, old->key, bit, old_node_bit);
}
if (bit < old_node_bit) { /* we don't have all bits in common */
/* The tree did not contain the key, so we insert <new> before the node
* <old>, and set ->bit to designate the lowest bit position in <new>
* which applies to ->branches.b[].
*/
eb_troot_t *new_left, *new_rght;
eb_troot_t *new_leaf, *old_node;
dup_tree:
new_left = eb_dotag(&new->node.branches, EB_LEFT);
new_rght = eb_dotag(&new->node.branches, EB_RGHT);
new_leaf = eb_dotag(&new->node.branches, EB_LEAF);
old_node = eb_dotag(&old->node.branches, EB_NODE);
new->node.node_p = old->node.node_p;
/* Note: we can compare more bits than the current node's because as
* long as they are identical, we know we descend along the correct
* side. However we don't want to start to compare past the end.
*/
diff = 0;
if (((unsigned)bit >> 3) < len)
diff = cmp_bits(new->key, old->key, bit);
if (diff < 0) {
new->node.leaf_p = new_left;
old->node.node_p = new_rght;
new->node.branches.b[EB_LEFT] = new_leaf;
new->node.branches.b[EB_RGHT] = old_node;
}
else if (diff > 0) {
old->node.node_p = new_left;
new->node.leaf_p = new_rght;
new->node.branches.b[EB_LEFT] = old_node;
new->node.branches.b[EB_RGHT] = new_leaf;
}
else {
struct eb_node *ret;
ret = eb_insert_dup(&old->node, &new->node);
return container_of(ret, struct ebpt_node, node);
}
break;
}
/* walk down */
root = &old->node.branches;
side = (((unsigned char *)new->key)[old_node_bit >> 3] >> (~old_node_bit & 7)) & 1;
troot = root->b[side];
}
/* Ok, now we are inserting <new> between <root> and <old>. <old>'s
* parent is already set to <new>, and the <root>'s branch is still in
* <side>. Update the root's leaf till we have it. Note that we can also
* find the side by checking the side of new->node.node_p.
*/
/* We need the common higher bits between new->key and old->key.
* This number of bits is already in <bit>.
*/
new->node.bit = bit;
root->b[side] = eb_dotag(&new->node.branches, EB_NODE);
return new;
}
#endif /* _EBIMTREE_H */