Minimize the performance impact of large numbers of BGP peering
routes by updating LPM's IPv4 code to use a hash like the IPv6 code.
Co-authored-by: Julien Charbon <jchar...@verisign.com>
Signed-off-by: Julien Charbon <jchar...@verisign.com>
Signed-off-by: Jeff Warrington <jwarring...@verisign.com>
Co-authored-by: Nicolas Witkowski <nwitkow...@verisign.com>
Signed-off-by: Nicolas Witkowski <nwitkow...@verisign.com>
Co-authored-by: Rohit Gupta <rogu...@verisign.com>
Signed-off-by: Rohit Gupta <rogu...@verisign.com>
---
lib/lpm/rte_lpm.c | 306 ++++++++++++++++++++++------------------------
lib/lpm/rte_lpm.h | 6 +
2 files changed, 149 insertions(+), 163 deletions(-)
diff --git a/lib/lpm/rte_lpm.c b/lib/lpm/rte_lpm.c
index cdcd1b7f9e..c951be8946 100644
--- a/lib/lpm/rte_lpm.c
+++ b/lib/lpm/rte_lpm.c
@@ -16,9 +16,13 @@
#include <rte_string_fns.h>
#include <rte_errno.h>
#include <rte_tailq.h>
+#include <rte_hash.h>
+#include <rte_jhash.h>
#include "rte_lpm.h"
+#define RULE_HASH_TABLE_EXTRA_SPACE 64
+
TAILQ_HEAD(rte_lpm_list, rte_tailq_entry);
static struct rte_tailq_elem rte_lpm_tailq = {
@@ -53,10 +57,11 @@ struct __rte_lpm {
/* LPM metadata. */
char name[RTE_LPM_NAMESIZE]; /**< Name of the lpm. */
uint32_t max_rules; /**< Max. balanced rules per lpm. */
+ uint32_t used_rules; /**< Used rules so far. */
uint32_t number_tbl8s; /**< Number of tbl8s. */
/**< Rule info table. */
struct rte_lpm_rule_info rule_info[RTE_LPM_MAX_DEPTH];
- struct rte_lpm_rule *rules_tbl; /**< LPM rules. */
+ struct rte_hash *rules_tbl; /**< LPM rules. */
/* RCU config. */
struct rte_rcu_qsbr *v; /* RCU QSBR variable. */
@@ -76,6 +81,47 @@ struct __rte_lpm {
#define VERIFY_DEPTH(depth)
#endif
+/*
+ * LPM rule hash function
+ *
+ * It's used as a hash function for the rte_hash
+ * containing rules
+ */
+static inline uint32_t
+rule_hash(const void *data, __rte_unused uint32_t data_len,
+ uint32_t init_val)
+{
+ return rte_jhash(data, sizeof(struct rte_lpm_rule_key), init_val);
+}
+
+static inline void
+rule_key_init(struct rte_lpm_rule_key *key, uint32_t ip, uint8_t depth)
+{
+ key->ip = ip;
+ key->depth = depth;
+}
+
+/* Find a rule */
+static inline int
+rule_find_with_key(struct rte_lpm *lpm, const struct rte_lpm_rule_key
*rule_key, uint32_t *next_hop)
+{
+ struct __rte_lpm *i_lpm;
+ uint64_t hash_val;
+ int ret;
+
+ i_lpm = container_of(lpm, struct __rte_lpm, lpm);
+
+ /* lookup for a rule */
+ ret = rte_hash_lookup_data(i_lpm->rules_tbl, (const void *) rule_key,
+ (void **) &hash_val);
+ if (ret >= 0) {
+ *next_hop = (uint32_t) hash_val;
+ return 1;
+ }
+
+ return 0;
+}
+
/*
* Converts a given depth value to its corresponding mask value.
*
@@ -150,8 +196,9 @@ rte_lpm_create(const char *name, int socket_id,
struct __rte_lpm *i_lpm;
struct rte_lpm *lpm = NULL;
struct rte_tailq_entry *te;
- uint32_t mem_size, rules_size, tbl8s_size;
+ uint32_t mem_size, tbl8s_size;
struct rte_lpm_list *lpm_list;
+ struct rte_hash *rules_tbl = NULL;
lpm_list = RTE_TAILQ_CAST(rte_lpm_tailq.head, rte_lpm_list);
@@ -164,7 +211,25 @@ rte_lpm_create(const char *name, int socket_id,
return NULL;
}
- snprintf(mem_name, sizeof(mem_name), "LPM_%s", name);
+ snprintf(mem_name, sizeof(mem_name), "LRHv4_%s", name);
+ struct rte_hash_parameters rule_hash_tbl_params = {
+ .entries = config->max_rules * 1.2 +
RULE_HASH_TABLE_EXTRA_SPACE,
+ .key_len = sizeof(struct rte_lpm_rule_key),
+ .hash_func = rule_hash,
+ .hash_func_init_val = 0,
+ .name = mem_name,
+ .reserved = 0,
+ .socket_id = socket_id,
+ .extra_flag = 0
+ };
+
+ rules_tbl = rte_hash_create(&rule_hash_tbl_params);
+ if (rules_tbl == NULL) {
+ RTE_LOG(ERR, LPM, "LPM rules hash table allocation failed: %s
(%d)",
+ rte_strerror(rte_errno), rte_errno);
+
+ return NULL;
+ }
rte_mcfg_tailq_write_lock();
@@ -182,7 +247,6 @@ rte_lpm_create(const char *name, int socket_id,
/* Determine the amount of memory to allocate. */
mem_size = sizeof(*i_lpm);
- rules_size = sizeof(struct rte_lpm_rule) * config->max_rules;
tbl8s_size = sizeof(struct rte_lpm_tbl_entry) *
RTE_LPM_TBL8_GROUP_NUM_ENTRIES * config->number_tbl8s;
@@ -204,18 +268,6 @@ rte_lpm_create(const char *name, int socket_id,
goto exit;
}
- i_lpm->rules_tbl = rte_zmalloc_socket(NULL,
- (size_t)rules_size, RTE_CACHE_LINE_SIZE, socket_id);
-
- if (i_lpm->rules_tbl == NULL) {
- RTE_LOG(ERR, LPM, "LPM rules_tbl memory allocation failed\n");
- rte_free(i_lpm);
- i_lpm = NULL;
- rte_free(te);
- rte_errno = ENOMEM;
- goto exit;
- }
-
i_lpm->lpm.tbl8 = rte_zmalloc_socket(NULL,
(size_t)tbl8s_size, RTE_CACHE_LINE_SIZE, socket_id);
@@ -230,6 +282,7 @@ rte_lpm_create(const char *name, int socket_id,
}
/* Save user arguments. */
+ i_lpm->rules_tbl = rules_tbl;
i_lpm->max_rules = config->max_rules;
i_lpm->number_tbl8s = config->number_tbl8s;
strlcpy(i_lpm->name, name, sizeof(i_lpm->name));
@@ -277,7 +330,7 @@ rte_lpm_free(struct rte_lpm *lpm)
if (i_lpm->dq != NULL)
rte_rcu_qsbr_dq_delete(i_lpm->dq);
rte_free(i_lpm->lpm.tbl8);
- rte_free(i_lpm->rules_tbl);
+ rte_hash_free(i_lpm->rules_tbl);
rte_free(i_lpm);
rte_free(te);
}
@@ -359,106 +412,64 @@ rte_lpm_rcu_qsbr_add(struct rte_lpm *lpm, struct
rte_lpm_rcu_config *cfg)
* NOTE: Valid range for depth parameter is 1 .. 32 inclusive.
*/
static int32_t
-rule_add(struct __rte_lpm *i_lpm, uint32_t ip_masked, uint8_t depth,
- uint32_t next_hop)
+rule_add(struct rte_lpm *lpm, uint32_t ip_masked, uint8_t depth, uint32_t
next_hop)
{
- uint32_t rule_gindex, rule_index, last_rule;
- int i;
+ int ret, rule_exist;
+ struct rte_lpm_rule_key rule_key;
+ uint32_t unused;
VERIFY_DEPTH(depth);
- /* Scan through rule group to see if rule already exists. */
- if (i_lpm->rule_info[depth - 1].used_rules > 0) {
+ /* init a rule key */
+ rule_key_init(&rule_key, ip_masked, depth);
- /* rule_gindex stands for rule group index. */
- rule_gindex = i_lpm->rule_info[depth - 1].first_rule;
- /* Initialise rule_index to point to start of rule group. */
- rule_index = rule_gindex;
- /* Last rule = Last used rule in this rule group. */
- last_rule = rule_gindex + i_lpm->rule_info[depth -
1].used_rules;
+ /* Scan through rule list to see if rule already exists. */
+ rule_exist = rule_find_with_key(lpm, &rule_key, &unused);
- for (; rule_index < last_rule; rule_index++) {
+ struct __rte_lpm *i_lpm = container_of(lpm, struct __rte_lpm, lpm);
- /* If rule already exists update next hop and return. */
- if (i_lpm->rules_tbl[rule_index].ip == ip_masked) {
+ if (!rule_exist && i_lpm->used_rules == i_lpm->max_rules)
+ return -ENOSPC;
- if (i_lpm->rules_tbl[rule_index].next_hop
- == next_hop)
- return -EEXIST;
- i_lpm->rules_tbl[rule_index].next_hop =
next_hop;
+ ret = rte_hash_add_key_data(i_lpm->rules_tbl, &rule_key,
+ (void *)(uintptr_t) next_hop);
- return rule_index;
- }
- }
-
- if (rule_index == i_lpm->max_rules)
- return -ENOSPC;
- } else {
- /* Calculate the position in which the rule will be stored. */
- rule_index = 0;
-
- for (i = depth - 1; i > 0; i--) {
- if (i_lpm->rule_info[i - 1].used_rules > 0) {
- rule_index = i_lpm->rule_info[i - 1].first_rule
- + i_lpm->rule_info[i -
1].used_rules;
- break;
- }
- }
- if (rule_index == i_lpm->max_rules)
- return -ENOSPC;
-
- i_lpm->rule_info[depth - 1].first_rule = rule_index;
- }
-
- /* Make room for the new rule in the array. */
- for (i = RTE_LPM_MAX_DEPTH; i > depth; i--) {
- if (i_lpm->rule_info[i - 1].first_rule
- + i_lpm->rule_info[i - 1].used_rules ==
i_lpm->max_rules)
- return -ENOSPC;
-
- if (i_lpm->rule_info[i - 1].used_rules > 0) {
- i_lpm->rules_tbl[i_lpm->rule_info[i - 1].first_rule
- + i_lpm->rule_info[i - 1].used_rules]
- = i_lpm->rules_tbl[i_lpm->rule_info[i -
1].first_rule];
- i_lpm->rule_info[i - 1].first_rule++;
- }
- }
-
- /* Add the new rule. */
- i_lpm->rules_tbl[rule_index].ip = ip_masked;
- i_lpm->rules_tbl[rule_index].next_hop = next_hop;
+ if (ret < 0)
+ return ret;
/* Increment the used rules counter for this rule group. */
- i_lpm->rule_info[depth - 1].used_rules++;
+ if (!rule_exist) {
+ i_lpm->rule_info[depth - 1].used_rules++;
+ i_lpm->used_rules++;
+ return 1;
+ }
- return rule_index;
+ return 0;
}
/*
* Delete a rule from the rule table.
* NOTE: Valid range for depth parameter is 1 .. 32 inclusive.
*/
-static void
-rule_delete(struct __rte_lpm *i_lpm, int32_t rule_index, uint8_t depth)
+static int
+rule_delete(struct __rte_lpm *i_lpm, int32_t ip, uint8_t depth)
{
- int i;
+ int ret;
+ struct rte_lpm_rule_key rule_key;
VERIFY_DEPTH(depth);
- i_lpm->rules_tbl[rule_index] =
- i_lpm->rules_tbl[i_lpm->rule_info[depth - 1].first_rule
- + i_lpm->rule_info[depth - 1].used_rules - 1];
+ rule_key_init(&rule_key, ip, depth);
- for (i = depth; i < RTE_LPM_MAX_DEPTH; i++) {
- if (i_lpm->rule_info[i].used_rules > 0) {
- i_lpm->rules_tbl[i_lpm->rule_info[i].first_rule - 1] =
-
i_lpm->rules_tbl[i_lpm->rule_info[i].first_rule
- +
i_lpm->rule_info[i].used_rules - 1];
- i_lpm->rule_info[i].first_rule--;
- }
+ /* delete the rule */
+ ret = rte_hash_del_key(i_lpm->rules_tbl, (void *) &rule_key);
+
+ if (ret >= 0) {
+ i_lpm->rule_info[depth - 1].used_rules--;
+ i_lpm->used_rules--;
}
- i_lpm->rule_info[depth - 1].used_rules--;
+ return ret;
}
/*
@@ -466,24 +477,14 @@ rule_delete(struct __rte_lpm *i_lpm, int32_t rule_index,
uint8_t depth)
* NOTE: Valid range for depth parameter is 1 .. 32 inclusive.
*/
static int32_t
-rule_find(struct __rte_lpm *i_lpm, uint32_t ip_masked, uint8_t depth)
+rule_find(struct rte_lpm *lpm, uint32_t ip_masked, uint8_t depth, uint32_t
*next_hop)
{
- uint32_t rule_gindex, last_rule, rule_index;
-
- VERIFY_DEPTH(depth);
-
- rule_gindex = i_lpm->rule_info[depth - 1].first_rule;
- last_rule = rule_gindex + i_lpm->rule_info[depth - 1].used_rules;
+ struct rte_lpm_rule_key rule_key;
- /* Scan used rules at given depth to find rule. */
- for (rule_index = rule_gindex; rule_index < last_rule; rule_index++) {
- /* If rule is found return the rule index. */
- if (i_lpm->rules_tbl[rule_index].ip == ip_masked)
- return rule_index;
- }
+ /* init a rule key */
+ rule_key_init(&rule_key, ip_masked, depth);
- /* If rule is not found return -EINVAL. */
- return -EINVAL;
+ return rule_find_with_key(lpm, &rule_key, next_hop);
}
/*
@@ -799,7 +800,7 @@ int
rte_lpm_add(struct rte_lpm *lpm, uint32_t ip, uint8_t depth,
uint32_t next_hop)
{
- int32_t rule_index, status = 0;
+ int32_t status = 0;
struct __rte_lpm *i_lpm;
uint32_t ip_masked;
@@ -811,17 +812,11 @@ rte_lpm_add(struct rte_lpm *lpm, uint32_t ip, uint8_t
depth,
ip_masked = ip & depth_to_mask(depth);
/* Add the rule to the rule table. */
- rule_index = rule_add(i_lpm, ip_masked, depth, next_hop);
-
- /* Skip table entries update if The rule is the same as
- * the rule in the rules table.
- */
- if (rule_index == -EEXIST)
- return 0;
+ status = rule_add(lpm, ip_masked, depth, next_hop);
/* If the is no space available for new rule return error. */
- if (rule_index < 0) {
- return rule_index;
+ if (status < 0) {
+ return status;
}
if (depth <= MAX_DEPTH_TBL24) {
@@ -834,7 +829,7 @@ rte_lpm_add(struct rte_lpm *lpm, uint32_t ip, uint8_t depth,
* rule that was added to rule table.
*/
if (status < 0) {
- rule_delete(i_lpm, rule_index, depth);
+ rule_delete(i_lpm, ip_masked, depth);
return status;
}
@@ -850,9 +845,7 @@ int
rte_lpm_is_rule_present(struct rte_lpm *lpm, uint32_t ip, uint8_t depth,
uint32_t *next_hop)
{
- struct __rte_lpm *i_lpm;
uint32_t ip_masked;
- int32_t rule_index;
/* Check user arguments. */
if ((lpm == NULL) ||
@@ -861,44 +854,42 @@ uint32_t *next_hop)
return -EINVAL;
/* Look for the rule using rule_find. */
- i_lpm = container_of(lpm, struct __rte_lpm, lpm);
ip_masked = ip & depth_to_mask(depth);
- rule_index = rule_find(i_lpm, ip_masked, depth);
- if (rule_index >= 0) {
- *next_hop = i_lpm->rules_tbl[rule_index].next_hop;
+ if (rule_find(lpm, ip_masked, depth, next_hop))
return 1;
- }
/* If rule is not found return 0. */
return 0;
}
static int32_t
-find_previous_rule(struct __rte_lpm *i_lpm, uint32_t ip, uint8_t depth,
- uint8_t *sub_rule_depth)
+find_previous_rule(struct rte_lpm *lpm, uint32_t ip, uint8_t depth,
+ uint32_t *subrule_prev_hop, uint8_t *subrule_prev_depth)
{
- int32_t rule_index;
uint32_t ip_masked;
+ uint32_t prev_hop;
uint8_t prev_depth;
+ int found;
for (prev_depth = (uint8_t)(depth - 1); prev_depth > 0; prev_depth--) {
ip_masked = ip & depth_to_mask(prev_depth);
- rule_index = rule_find(i_lpm, ip_masked, prev_depth);
+ found = rule_find(lpm, ip_masked, prev_depth, &prev_hop);
- if (rule_index >= 0) {
- *sub_rule_depth = prev_depth;
- return rule_index;
+ if (found) {
+ *subrule_prev_hop = prev_hop;
+ *subrule_prev_depth = prev_depth;
+ return 1;
}
}
- return -1;
+ return 0;
}
static int32_t
delete_depth_small(struct __rte_lpm *i_lpm, uint32_t ip_masked,
- uint8_t depth, int32_t sub_rule_index, uint8_t sub_rule_depth)
+ uint8_t depth, int found, uint32_t sub_rule_hop, uint8_t sub_rule_depth)
{
#define group_idx next_hop
uint32_t tbl24_range, tbl24_index, tbl8_group_index, tbl8_index, i, j;
@@ -912,7 +903,7 @@ delete_depth_small(struct __rte_lpm *i_lpm, uint32_t
ip_masked,
* Firstly check the sub_rule_index. A -1 indicates no replacement rule
* and a positive number indicates a sub_rule_index.
*/
- if (sub_rule_index < 0) {
+ if (!found) {
/*
* If no replacement rule exists then invalidate entries
* associated with this rule.
@@ -949,7 +940,7 @@ delete_depth_small(struct __rte_lpm *i_lpm, uint32_t
ip_masked,
*/
struct rte_lpm_tbl_entry new_tbl24_entry = {
- .next_hop = i_lpm->rules_tbl[sub_rule_index].next_hop,
+ .next_hop = sub_rule_hop,
.valid = VALID,
.valid_group = 0,
.depth = sub_rule_depth,
@@ -959,8 +950,7 @@ delete_depth_small(struct __rte_lpm *i_lpm, uint32_t
ip_masked,
.valid = VALID,
.valid_group = VALID,
.depth = sub_rule_depth,
- .next_hop = i_lpm->rules_tbl
- [sub_rule_index].next_hop,
+ .next_hop = sub_rule_hop,
};
for (i = tbl24_index; i < (tbl24_index + tbl24_range); i++) {
@@ -1052,7 +1042,7 @@ tbl8_recycle_check(struct rte_lpm_tbl_entry *tbl8,
static int32_t
delete_depth_big(struct __rte_lpm *i_lpm, uint32_t ip_masked,
- uint8_t depth, int32_t sub_rule_index, uint8_t sub_rule_depth)
+ uint8_t depth, int found, uint32_t sub_rule_hop, uint8_t sub_rule_depth)
{
#define group_idx next_hop
uint32_t tbl24_index, tbl8_group_index, tbl8_group_start, tbl8_index,
@@ -1071,7 +1061,7 @@ delete_depth_big(struct __rte_lpm *i_lpm, uint32_t
ip_masked,
tbl8_index = tbl8_group_start + (ip_masked & 0xFF);
tbl8_range = depth_to_range(depth);
- if (sub_rule_index < 0) {
+ if (!found) {
/*
* Loop through the range of entries on tbl8 for which the
* rule_to_delete must be removed or modified.
@@ -1086,7 +1076,7 @@ delete_depth_big(struct __rte_lpm *i_lpm, uint32_t
ip_masked,
.valid = VALID,
.depth = sub_rule_depth,
.valid_group =
i_lpm->lpm.tbl8[tbl8_group_start].valid_group,
- .next_hop = i_lpm->rules_tbl[sub_rule_index].next_hop,
+ .next_hop = sub_rule_hop,
};
/*
@@ -1142,9 +1132,10 @@ delete_depth_big(struct __rte_lpm *i_lpm, uint32_t
ip_masked,
int
rte_lpm_delete(struct rte_lpm *lpm, uint32_t ip, uint8_t depth)
{
- int32_t rule_to_delete_index, sub_rule_index;
+ int32_t found;
struct __rte_lpm *i_lpm;
uint32_t ip_masked;
+ uint32_t sub_rule_hop;
uint8_t sub_rule_depth;
/*
* Check input arguments. Note: IP must be a positive integer of 32
@@ -1157,21 +1148,9 @@ rte_lpm_delete(struct rte_lpm *lpm, uint32_t ip, uint8_t
depth)
i_lpm = container_of(lpm, struct __rte_lpm, lpm);
ip_masked = ip & depth_to_mask(depth);
- /*
- * Find the index of the input rule, that needs to be deleted, in the
- * rule table.
- */
- rule_to_delete_index = rule_find(i_lpm, ip_masked, depth);
-
- /*
- * Check if rule_to_delete_index was found. If no rule was found the
- * function rule_find returns -EINVAL.
- */
- if (rule_to_delete_index < 0)
- return -EINVAL;
-
/* Delete the rule from the rule table. */
- rule_delete(i_lpm, rule_to_delete_index, depth);
+ if (rule_delete(i_lpm, ip_masked, depth) < 0)
+ return -EINVAL;
/*
* Find rule to replace the rule_to_delete. If there is no rule to
@@ -1179,18 +1158,18 @@ rte_lpm_delete(struct rte_lpm *lpm, uint32_t ip,
uint8_t depth)
* entries associated with this rule.
*/
sub_rule_depth = 0;
- sub_rule_index = find_previous_rule(i_lpm, ip, depth, &sub_rule_depth);
+ found = find_previous_rule(lpm, ip, depth, &sub_rule_hop,
&sub_rule_depth);
/*
* If the input depth value is less than 25 use function
* delete_depth_small otherwise use delete_depth_big.
*/
if (depth <= MAX_DEPTH_TBL24) {
- return delete_depth_small(i_lpm, ip_masked, depth,
- sub_rule_index, sub_rule_depth);
- } else { /* If depth > MAX_DEPTH_TBL24 */
- return delete_depth_big(i_lpm, ip_masked, depth, sub_rule_index,
- sub_rule_depth);
+ return delete_depth_small(i_lpm, ip_masked, depth, found,
+ sub_rule_hop, sub_rule_depth);
+ } else {
+ return delete_depth_big(i_lpm, ip_masked, depth, found,
+ sub_rule_hop, sub_rule_depth);
}
}
@@ -1205,6 +1184,7 @@ rte_lpm_delete_all(struct rte_lpm *lpm)
i_lpm = container_of(lpm, struct __rte_lpm, lpm);
/* Zero rule information. */
memset(i_lpm->rule_info, 0, sizeof(i_lpm->rule_info));
+ i_lpm->used_rules = 0;
/* Zero tbl24. */
memset(i_lpm->lpm.tbl24, 0, sizeof(i_lpm->lpm.tbl24));
@@ -1214,5 +1194,5 @@ rte_lpm_delete_all(struct rte_lpm *lpm)
* RTE_LPM_TBL8_GROUP_NUM_ENTRIES * i_lpm->number_tbl8s);
/* Delete all rules form the rules table. */
- memset(i_lpm->rules_tbl, 0, sizeof(i_lpm->rules_tbl[0]) *
i_lpm->max_rules);
+ rte_hash_reset(i_lpm->rules_tbl);
}
diff --git a/lib/lpm/rte_lpm.h b/lib/lpm/rte_lpm.h
index 75e27ff164..605445f3a7 100644
--- a/lib/lpm/rte_lpm.h
+++ b/lib/lpm/rte_lpm.h
@@ -108,6 +108,12 @@ struct rte_lpm_tbl_entry {
#endif
+/** Rules tbl entry key. */
+struct rte_lpm_rule_key {
+ uint32_t ip; /**< Rule IP address. */
+ uint32_t depth; /**< Rule depth. */
+};
+
/** LPM configuration structure. */
struct rte_lpm_config {
uint32_t max_rules; /**< Max number of rules. */
--
2.37.0 (Apple Git-136)
