MEDIUM: resolvers: hash the records before inserting them into the tree

We're using an XXH32() on the record to insert it into or look it up from
the tree. This way we don't change the rest of the code, the comparisons
are still made on all fields and the next node is visited on mismatch. This
also allows to continue to use roundrobin between identical nodes.

Just doing this is sufficient to see the CPU usage go down from ~60-70% to
4% at ~2k DNS requests per second for farm with 300 servers. A larger
config with 12 backends of 2000 servers each shows ~8-9% CPU for 6-10000
DNS requests per second.

It would probably be possible to go further with multiple levels of indexing
but it's not worth it, and it's important to remember that tree nodes take
space (the struct answer_list went back from 576 to 600 bytes).

diff --git a/src/resolvers.c b/src/resolvers.c
index b71a81f7f47b6e00d4a85c49ed3ccecd428100cd..34005625a084d3d7999623ee54ba56575b62bbf6 100644 (file)
--- a/src/resolvers.c
+++ b/src/resolvers.c
@@ -47,6 +47,7 @@
 #include <haproxy/time.h>
 #include <haproxy/tools.h>
 #include <haproxy/vars.h>
+#include <haproxy/xxhash.h>
 
 
 struct list sec_resolvers  = LIST_HEAD_INIT(sec_resolvers);
@@ -882,6 +883,7 @@ static int resolv_validate_dns_response(unsigned char *resp, unsigned char *bufe
        struct resolv_answer_item *answer_record, *tmp_record;
        struct resolv_response *r_res;
        struct eb32_node *eb32;
+       uint32_t key = 0;
        int i, found = 0;
        int cause = RSLV_RESP_ERROR;
 
@@ -1116,6 +1118,7 @@ static int resolv_validate_dns_response(unsigned char *resp, unsigned char *bufe
 
                                answer_record->data.in4.sin_family = AF_INET;
                                memcpy(&answer_record->data.in4.sin_addr, reader, answer_record->data_len);
+                               key = XXH32(reader, answer_record->data_len, answer_record->type);
                                break;
 
                        case DNS_RTYPE_CNAME:
@@ -1139,6 +1142,7 @@ static int resolv_validate_dns_response(unsigned char *resp, unsigned char *bufe
 
                                memcpy(answer_record->data.target, tmpname, len);
                                answer_record->data.target[len] = 0;
+                               key = XXH32(tmpname, len, answer_record->type);
                                previous_dname = answer_record->data.target;
                                break;
 
@@ -1167,6 +1171,7 @@ static int resolv_validate_dns_response(unsigned char *resp, unsigned char *bufe
                                answer_record->data_len = len;
                                memcpy(answer_record->data.target, tmpname, len);
                                answer_record->data.target[len] = 0;
+                               key = XXH32(tmpname, len, answer_record->type);
                                if (answer_record->ar_item != NULL) {
                                        pool_free(resolv_answer_item_pool, answer_record->ar_item);
                                        answer_record->ar_item = NULL;
@@ -1180,6 +1185,7 @@ static int resolv_validate_dns_response(unsigned char *resp, unsigned char *bufe
 
                                answer_record->data.in6.sin6_family = AF_INET6;
                                memcpy(&answer_record->data.in6.sin6_addr, reader, answer_record->data_len);
+                               key = XXH32(reader, answer_record->data_len, answer_record->type);
                                break;
 
                } /* switch (record type) */
@@ -1197,7 +1203,7 @@ static int resolv_validate_dns_response(unsigned char *resp, unsigned char *bufe
                /* Lookup to see if we already had this entry */
                found = 0;
 
-               for (eb32 = eb32_first(&r_res->answer_tree); eb32 != NULL;  eb32 = eb32_next(eb32)) {
+               for (eb32 = eb32_lookup(&r_res->answer_tree, key); eb32 != NULL;  eb32 = eb32_next(eb32)) {
                        tmp_record = eb32_entry(eb32, typeof(*tmp_record), link);
                        if (tmp_record->type != answer_record->type)
                                continue;
@@ -1242,7 +1248,7 @@ static int resolv_validate_dns_response(unsigned char *resp, unsigned char *bufe
                else {
                        answer_record->last_seen = now_ms;
                        answer_record->ar_item = NULL;
-                       answer_record->link.key = 0; // will be set later
+                       answer_record->link.key = key;
                        eb32_insert(&r_res->answer_tree, &answer_record->link);
                        answer_record = NULL;
                }