/* rrl.c - Response Rate Limiting for NSD.
* By W.C.A. Wijngaards
* Copyright 2012, NLnet Labs.
* BSD, see LICENSE.
*/
#include "config.h"
#include <errno.h>
#include "rrl.h"
#include "util.h"
#include "lookup3.h"
#include "options.h"

#ifdef RATELIMIT

#ifdef HAVE_MMAP
#include <sys/mman.h>
#if defined(MAP_ANON) && !defined(MAP_ANONYMOUS)
#define MAP_ANONYMOUS   MAP_ANON
#endif
#endif /* HAVE_MMAP */


/**
* The rate limiting data structure bucket, this represents one rate of
* packets from a single source.
* Smoothed average rates.
*/
struct rrl_bucket {
       /* the source netmask */
       uint64_t source;
       /* rate, in queries per second, which due to rate=r(t)+r(t-1)/2 is
        * equal to double the queries per second */
       uint32_t rate;
       /* the full hash */
       uint32_t hash;
       /* counter for queries arrived in this second */
       uint32_t counter;
       /* timestamp, which time is the time of the counter, the rate is from
        * one timestep before that. */
       int32_t stamp;
       /* flags for the source mask and type */
       uint16_t flags;
};

/* the (global) array of RRL buckets */
static struct rrl_bucket* rrl_array = NULL;
static size_t rrl_array_size = RRL_BUCKETS;
static uint32_t rrl_ratelimit = RRL_LIMIT; /* 2x qps */
static uint8_t rrl_slip_ratio = RRL_SLIP;
static uint8_t rrl_ipv4_prefixlen = RRL_IPV4_PREFIX_LENGTH;
static uint8_t rrl_ipv6_prefixlen = RRL_IPV6_PREFIX_LENGTH;
static uint64_t rrl_ipv6_mask; /* max prefixlen 64 */
static uint32_t rrl_whitelist_ratelimit = RRL_WLIST_LIMIT; /* 2x qps */

/* the array of mmaps for the children (saved between reloads) */
static void** rrl_maps = NULL;
static size_t rrl_maps_num = 0;

void rrl_mmap_init(int numch, size_t numbuck, size_t lm, size_t wlm, size_t sm,
       size_t plf, size_t pls)
{
#ifdef HAVE_MMAP
       size_t i;
#endif
       if(numbuck != 0)
               rrl_array_size = numbuck;
       rrl_ratelimit = lm*2;
       rrl_slip_ratio = sm;
       rrl_ipv4_prefixlen = plf;
       rrl_ipv6_prefixlen = pls;
       if (pls <= 32) {
               rrl_ipv6_mask = ((uint64_t) htonl(0xffffffff << (32-pls))) << 32;
       } else {
               rrl_ipv6_mask =  ((uint64_t) htonl(0xffffffff << (64-pls))) |
                       (((uint64_t)0xffffffff)<<32);
       }
       rrl_whitelist_ratelimit = wlm*2;
#ifdef HAVE_MMAP
       /* allocate the ratelimit hashtable in a memory map so it is
        * preserved across reforks (every child its own table) */
       rrl_maps_num = (size_t)numch;
       rrl_maps = (void**)xmallocarray(rrl_maps_num, sizeof(void*));
       for(i=0; i<rrl_maps_num; i++) {
               rrl_maps[i] = mmap(NULL,
                       sizeof(struct rrl_bucket)*rrl_array_size,
                       PROT_READ|PROT_WRITE, MAP_SHARED|MAP_ANONYMOUS, -1, 0);
               if(rrl_maps[i] == MAP_FAILED) {
                       log_msg(LOG_ERR, "rrl: mmap failed: %s",
                               strerror(errno));
                       exit(1);
               }
               memset(rrl_maps[i], 0,
                       sizeof(struct rrl_bucket)*rrl_array_size);
       }
#else
       (void)numch;
       rrl_maps_num = 0;
       rrl_maps = NULL;
#endif
}

void rrl_mmap_deinit(void)
{
#ifdef HAVE_MMAP
       size_t i;
       for(i=0; i<rrl_maps_num; i++) {
               munmap(rrl_maps[i], sizeof(struct rrl_bucket)*rrl_array_size);
               rrl_maps[i] = NULL;
       }
       free(rrl_maps);
       rrl_maps = NULL;
#endif
}

void rrl_mmap_deinit_keep_mmap(void)
{
#ifdef HAVE_MMAP
       free(rrl_maps);
       rrl_maps = NULL;
#endif
}

void rrl_set_limit(size_t lm, size_t wlm, size_t sm)
{
       rrl_ratelimit = lm*2;
       rrl_whitelist_ratelimit = wlm*2;
       rrl_slip_ratio = sm;
}

void rrl_init(size_t ch)
{
       if(!rrl_maps || ch >= rrl_maps_num)
           rrl_array = xalloc_array_zero(sizeof(struct rrl_bucket),
               rrl_array_size);
#ifdef HAVE_MMAP
       else rrl_array = (struct rrl_bucket*)rrl_maps[ch];
#endif
}

void rrl_deinit(size_t ch)
{
       if(!rrl_maps || ch >= rrl_maps_num)
               free(rrl_array);
       rrl_array = NULL;
}

/** return the source netblock of the query, this is the genuine source
* for genuine queries and the target for reflected packets */
static uint64_t rrl_get_source(query_type* query, uint16_t* c2)
{
       /* note there is an IPv6 subnet, that maps
        * to the same buckets as IPv4 space, but there is a flag in c2
        * that makes the hash different */
#ifdef INET6
       if( ((struct sockaddr_in*)&query->client_addr)->sin_family == AF_INET) {
               *c2 = 0;
               return ((struct sockaddr_in*)&query->client_addr)->
                       sin_addr.s_addr & htonl(0xffffffff << (32-rrl_ipv4_prefixlen));
       } else {
               uint64_t s;
               *c2 = rrl_ip6;
               memmove(&s, &((struct sockaddr_in6*)&query->client_addr)->sin6_addr,
                       sizeof(s));
               return s & rrl_ipv6_mask;
       }
#else
       *c2 = 0;
       return query->client_addr.sin_addr.s_addr & htonl(0xffffffff << (32-rrl_ipv4_prefixlen));
#endif
}

/** debug source to string */
static const char* rrlsource2str(uint64_t s, uint16_t c2)
{
       static char buf[64];
       struct in_addr a4;
#ifdef INET6
       if(c2) {
               /* IPv6 */
               struct in6_addr a6;
               memset(&a6, 0, sizeof(a6));
               memmove(&a6, &s, sizeof(s));
               if(!inet_ntop(AF_INET6, &a6, buf, sizeof(buf)))
                       strlcpy(buf, "[ip6 ntop failed]", sizeof(buf));
               else {
                       static char prefix[5];
                       snprintf(prefix, sizeof(prefix), "/%d", rrl_ipv6_prefixlen);
                       strlcat(buf, &prefix[0], sizeof(buf));
               }
               return buf;
       }
#else
       (void)c2;
#endif
       /* ipv4 */
       a4.s_addr = (uint32_t)s;
       if(!inet_ntop(AF_INET, &a4, buf, sizeof(buf)))
               strlcpy(buf, "[ip4 ntop failed]", sizeof(buf));
       else {
               static char prefix[5];
               snprintf(prefix, sizeof(prefix), "/%d", rrl_ipv4_prefixlen);
               strlcat(buf, &prefix[0], sizeof(buf));
       }
       return buf;
}

enum rrl_type rrlstr2type(const char* s)
{
       if(strcmp(s, "nxdomain")==0) return rrl_type_nxdomain;
       else if(strcmp(s, "error")==0) return rrl_type_error;
       else if(strcmp(s, "referral")==0) return rrl_type_referral;
       else if(strcmp(s, "any")==0) return rrl_type_any;
       else if(strcmp(s, "wildcard")==0) return rrl_type_wildcard;
       else if(strcmp(s, "nodata")==0) return rrl_type_nodata;
       else if(strcmp(s, "dnskey")==0) return rrl_type_dnskey;
       else if(strcmp(s, "positive")==0) return rrl_type_positive;
       else if(strcmp(s, "rrsig")==0) return rrl_type_rrsig;
       else if(strcmp(s, "all")==0) return rrl_type_all;
       return 0; /* unknown */
}

const char* rrltype2str(enum rrl_type c)
{
       switch(c & 0x0fff) {
               case rrl_type_nxdomain: return "nxdomain";
               case rrl_type_error: return "error";
               case rrl_type_referral: return "referral";
               case rrl_type_any: return "any";
               case rrl_type_wildcard: return "wildcard";
               case rrl_type_nodata: return "nodata";
               case rrl_type_dnskey: return "dnskey";
               case rrl_type_positive: return "positive";
               case rrl_type_rrsig: return "rrsig";
               case rrl_type_all: return "all";
       }
       return "unknown";
}

/** classify the query in a number of different types, each has separate
* ratelimiting, so that positive queries are not impeded by others */
static uint16_t rrl_classify(query_type* query, const uint8_t** d,
       size_t* d_len)
{
       if(RCODE(query->packet) == RCODE_NXDOMAIN) {
               if(query->zone && query->zone->apex) {
                       *d = dname_name(domain_dname(query->zone->apex));
                       *d_len = domain_dname(query->zone->apex)->name_size;
               }
               return rrl_type_nxdomain;
       }
       if(RCODE(query->packet) != RCODE_OK) {
               if(query->zone && query->zone->apex) {
                       *d = dname_name(domain_dname(query->zone->apex));
                       *d_len = domain_dname(query->zone->apex)->name_size;
               }
               return rrl_type_error;
       }
       if(query->delegation_domain) {
               *d = dname_name(domain_dname(query->delegation_domain));
               *d_len = domain_dname(query->delegation_domain)->name_size;
               return rrl_type_referral;
       }
       if(query->qtype == TYPE_ANY) {
               if(query->qname) {
                       *d = dname_name(query->qname);
                       *d_len = query->qname->name_size;
               }
               return rrl_type_any;
       }
       if(query->qtype == TYPE_RRSIG) {
               if(query->qname) {
                       *d = dname_name(query->qname);
                       *d_len = query->qname->name_size;
               }
               return rrl_type_rrsig;
       }
       if(query->wildcard_domain) {
               *d = dname_name(domain_dname(query->wildcard_domain));
               *d_len = domain_dname(query->wildcard_domain)->name_size;
               return rrl_type_wildcard;
       }
       if(ANCOUNT(query->packet) == 0) {
               if(query->zone && query->zone->apex) {
                       *d = dname_name(domain_dname(query->zone->apex));
                       *d_len = domain_dname(query->zone->apex)->name_size;
               }
               return rrl_type_nodata;
       }
       if(query->qtype == TYPE_DNSKEY) {
               if(query->qname) {
                       *d = dname_name(query->qname);
                       *d_len = query->qname->name_size;
               }
               return rrl_type_dnskey;
       }
       /* positive */
       if(query->qname) {
               *d = dname_name(query->qname);
               *d_len = query->qname->name_size;
       }
       return rrl_type_positive;
}

/** Examine the query and return hash and source of netblock. */
static void examine_query(query_type* query, uint32_t* hash, uint64_t* source,
       uint16_t* flags, uint32_t* lm)
{
       /* compile a binary string representing the query */
       uint16_t c, c2;
       /* size with 16 bytes to spare */
       uint8_t buf[MAXDOMAINLEN + sizeof(*source) + sizeof(c) + 16];
       const uint8_t* dname = NULL; size_t dname_len = 0;
       uint32_t r = 0x267fcd16;

       *source = rrl_get_source(query, &c2);
       c = rrl_classify(query, &dname, &dname_len);
       if(query->zone && query->zone->opts &&
               (query->zone->opts->pattern->rrl_whitelist & c))
               *lm = rrl_whitelist_ratelimit;
       if(*lm == 0) return;
       c |= c2;
       *flags = c;
       memmove(buf, source, sizeof(*source));
       memmove(buf+sizeof(*source), &c, sizeof(c));

       DEBUG(DEBUG_QUERY, 1, (LOG_INFO, "rrl_examine type %s name %s", rrltype2str(c), dname?wiredname2str(dname):"NULL"));

       /* and hash it */
       if(dname && dname_len <= MAXDOMAINLEN) {
               memmove(buf+sizeof(*source)+sizeof(c), dname, dname_len);
               *hash = hashlittle(buf, sizeof(*source)+sizeof(c)+dname_len, r);
       } else
               *hash = hashlittle(buf, sizeof(*source)+sizeof(c), r);
}

/* age the bucket because elapsed time steps have gone by */
static void rrl_attenuate_bucket(struct rrl_bucket* b, int32_t elapsed)
{
       if(elapsed > 16) {
               b->rate = 0;
       } else {
               /* divide rate /2 for every elapsed time step, because
                * the counters in the inbetween steps were 0 */
               /* r(t) = 0 + 0/2 + 0/4 + .. + oldrate/2^dt */
               b->rate >>= elapsed;
               /* we know that elapsed >= 2 */
               b->rate += (b->counter>>(elapsed-1));
       }
}

/** log a message about ratelimits */
static void
rrl_msg(query_type* query, const char* str)
{
       uint16_t c, c2, wl = 0;
       const uint8_t* d = NULL;
       size_t d_len;
       uint64_t s;
       char address[128];
       if(verbosity < 1) return;
       addr2str(&query->client_addr, address, sizeof(address));
       s = rrl_get_source(query, &c2);
       c = rrl_classify(query, &d, &d_len) | c2;
       if(query->zone && query->zone->opts &&
               (query->zone->opts->pattern->rrl_whitelist & c))
               wl = 1;
       log_msg(LOG_INFO, "ratelimit %s %s type %s%s target %s query %s %s",
               str, d?wiredname2str(d):"", rrltype2str(c),
               wl?"(whitelisted)":"", rrlsource2str(s, c2),
               address, rrtype_to_string(query->qtype));
}

/** true if the query used to be blocked by the ratelimit */
static int
used_to_block(uint32_t rate, uint32_t counter, uint32_t lm)
{
       return rate >= lm || counter+rate/2 >= lm;
}

/** update the rate in a ratelimit bucket, return actual rate */
uint32_t rrl_update(query_type* query, uint32_t hash, uint64_t source,
       uint16_t flags, int32_t now, uint32_t lm)
{
       struct rrl_bucket* b = &rrl_array[hash % rrl_array_size];

       DEBUG(DEBUG_QUERY, 1, (LOG_INFO, "source %llx hash %x oldrate %d oldcount %d stamp %d",
               (long long unsigned)source, hash, b->rate, b->counter, b->stamp));

       /* check if different source */
       if(b->source != source || b->flags != flags || b->hash != hash) {
               /* initialise */
               /* potentially the wrong limit here, used lower nonwhitelim */
               if(verbosity >= 1 &&
                       used_to_block(b->rate, b->counter, rrl_ratelimit)) {
                       char address[128];
                       addr2str(&query->client_addr, address, sizeof(address));
                       log_msg(LOG_INFO, "ratelimit unblock ~ type %s target %s query %s %s (%s collision)",
                               rrltype2str(b->flags),
                               rrlsource2str(b->source, b->flags),
                               address, rrtype_to_string(query->qtype),
                               (b->hash!=hash?"bucket":"hash"));
               }
               b->hash = hash;
               b->source = source;
               b->flags = flags;
               b->counter = 1;
               b->rate = 0;
               b->stamp = now;
               return 1;
       }
       /* this is the same source */

       /* check if old, zero or smooth it */
       /* circular arith for time */
       if(now - b->stamp == 1) {
               /* very busy bucket and time just stepped one step */
               int oldblock = used_to_block(b->rate, b->counter, lm);
               b->rate = b->rate/2 + b->counter;
               if(oldblock && b->rate < lm)
                       rrl_msg(query, "unblock");
               b->counter = 1;
               b->stamp = now;
       } else if(now - b->stamp > 0) {
               /* older bucket */
               int olderblock = used_to_block(b->rate, b->counter, lm);
               rrl_attenuate_bucket(b, now - b->stamp);
               if(olderblock && b->rate < lm)
                       rrl_msg(query, "unblock");
               b->counter = 1;
               b->stamp = now;
       } else if(now != b->stamp) {
               /* robust, timestamp from the future */
               if(used_to_block(b->rate, b->counter, lm))
                       rrl_msg(query, "unblock");
               b->rate = 0;
               b->counter = 1;
               b->stamp = now;
       } else {
               /* bucket is from the current timestep, update counter */
               b->counter ++;

               /* log what is blocked for operational debugging */
               if(b->counter + b->rate/2 == lm && b->rate < lm)
                       rrl_msg(query, "block");
       }

       /* return max from current rate and projected next-value for rate */
       /* so that if the rate increases suddenly very high, it is
        * stopped halfway into the time step */
       if(b->counter > b->rate/2)
               return b->counter + b->rate/2;
       return b->rate;
}

int rrl_process_query(query_type* query)
{
       uint64_t source;
       uint32_t hash;
       /* we can use circular arithmetic here, so int32 works after 2038 */
       int32_t now = (int32_t)time(NULL);
       uint32_t lm = rrl_ratelimit;
       uint16_t flags;
       if(rrl_ratelimit == 0 && rrl_whitelist_ratelimit == 0)
               return 0;

       /* examine query */
       examine_query(query, &hash, &source, &flags, &lm);

       if(lm == 0)
               return 0; /* no limit for this */

       /* update rate */
       return (rrl_update(query, hash, source, flags, now, lm) >= lm);
}

query_state_type rrl_slip(query_type* query)
{
       /* discard number the packets, randomly */
#ifdef HAVE_ARC4RANDOM_UNIFORM
       if((rrl_slip_ratio > 0) && ((rrl_slip_ratio == 1) || ((arc4random_uniform(rrl_slip_ratio)) == 0))) {
#elif HAVE_ARC4RANDOM
       if((rrl_slip_ratio > 0) && ((rrl_slip_ratio == 1) || ((arc4random() % rrl_slip_ratio) == 0))) {
#else
       if((rrl_slip_ratio > 0) && ((rrl_slip_ratio == 1) || ((random() % rrl_slip_ratio) == 0))) {
#endif
               /* set TC on the rest */
               TC_SET(query->packet);
               ANCOUNT_SET(query->packet, 0);
               NSCOUNT_SET(query->packet, 0);
               ARCOUNT_SET(query->packet, 0);
               if(query->qname)
                       /* header, type, class, qname */
                       buffer_set_position(query->packet,
                               QHEADERSZ+4+query->qname->name_size);
               else    buffer_set_position(query->packet, QHEADERSZ);
               return QUERY_PROCESSED;
       }
       return QUERY_DISCARDED;
}

#endif /* RATELIMIT */

You are viewing proxied material from ftp.icm.edu.pl. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.