From: bert hubert <bert.hubert@netherlabs.nl>
Date: Wed, 9 Sep 2015 09:17:45 +0000 (+0200)
Subject: Port the 0x20 hashing to 3.7.3
X-Git-Tag: rec-3.7.4~12^2
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=ef49a7cc45f3b9bbacfcfc79d329eacbe771e105;p=pdns

Port the 0x20 hashing to 3.7.3

This is a combination of the commits:
- 6200227b6b3cfd8796d6cc75e8cf29bcec586f0c
- af891b3d6d3f6ee6e848411cc9e477c96e955f98

And a port of the testcase
---

diff --git a/pdns/dns.cc b/pdns/dns.cc
index 1678f7072..41da1ba36 100644
--- a/pdns/dns.cc
+++ b/pdns/dns.cc
@@ -142,7 +142,7 @@ uint32_t hashQuestion(const char* packet, uint16_t len, uint32_t init)
   while((labellen=*pos++) && pos < end) { 
     if(pos + labellen + 1 > end) // include length field  in hash
       return 0;
-    ret=burtle(pos, labellen+1, ret);
+    ret=burtleCI(pos, labellen+1, ret);
     pos += labellen;
   }
   return ret;
diff --git a/pdns/misc.cc b/pdns/misc.cc
index 01529cf05..5dc51c39b 100644
--- a/pdns/misc.cc
+++ b/pdns/misc.cc
@@ -911,6 +911,47 @@ uint32_t burtle(const unsigned char* k, uint32_t length, uint32_t initval)
   return c;
 }
 
+uint32_t burtleCI(const unsigned char* k, uint32_t length, uint32_t initval)
+{
+  uint32_t a,b,c,len;
+
+   /* Set up the internal state */
+  len = length;
+  a = b = 0x9e3779b9;  /* the golden ratio; an arbitrary value */
+  c = initval;         /* the previous hash value */
+
+  /*---------------------------------------- handle most of the key */
+  while (len >= 12) {
+    a += (dns_tolower(k[0]) +((uint32_t)dns_tolower(k[1])<<8) +((uint32_t)dns_tolower(k[2])<<16) +((uint32_t)dns_tolower(k[3])<<24));
+    b += (dns_tolower(k[4]) +((uint32_t)dns_tolower(k[5])<<8) +((uint32_t)dns_tolower(k[6])<<16) +((uint32_t)dns_tolower(k[7])<<24));
+    c += (dns_tolower(k[8]) +((uint32_t)dns_tolower(k[9])<<8) +((uint32_t)dns_tolower(k[10])<<16)+((uint32_t)dns_tolower(k[11])<<24));
+    burtlemix(a,b,c);
+    k += 12; len -= 12;
+  }
+
+  /*------------------------------------- handle the last 11 bytes */
+  c += length;
+  switch(len) {             /* all the case statements fall through */
+  case 11: c+=((uint32_t)dns_tolower(k[10])<<24);
+  case 10: c+=((uint32_t)dns_tolower(k[9])<<16);
+  case 9 : c+=((uint32_t)dns_tolower(k[8])<<8);
+    /* the first byte of c is reserved for the length */
+  case 8 : b+=((uint32_t)dns_tolower(k[7])<<24);
+  case 7 : b+=((uint32_t)dns_tolower(k[6])<<16);
+  case 6 : b+=((uint32_t)dns_tolower(k[5])<<8);
+  case 5 : b+=dns_tolower(k[4]);
+  case 4 : a+=((uint32_t)dns_tolower(k[3])<<24);
+  case 3 : a+=((uint32_t)dns_tolower(k[2])<<16);
+  case 2 : a+=((uint32_t)dns_tolower(k[1])<<8);
+  case 1 : a+=dns_tolower(k[0]);
+    /* case 0: nothing left to add */
+  }
+  burtlemix(a,b,c);
+  /*-------------------------------------------- report the result */
+  return c;
+}
+
+
 void setSocketTimestamps(int fd)
 {
 #ifdef SO_TIMESTAMP
diff --git a/pdns/misc.hh b/pdns/misc.hh
index 515e35fe5..dcf7ba561 100644
--- a/pdns/misc.hh
+++ b/pdns/misc.hh
@@ -544,5 +544,6 @@ unsigned int getFilenumLimit(bool hardOrSoft=0);
 void setFilenumLimit(unsigned int lim);
 bool readFileIfThere(const char* fname, std::string* line);
 uint32_t burtle(const unsigned char* k, uint32_t lengh, uint32_t init);
+uint32_t burtleCI(const unsigned char* k, uint32_t lengh, uint32_t init);
 void setSocketTimestamps(int fd);
 #endif
diff --git a/pdns/test-misc_hh.cc b/pdns/test-misc_hh.cc
index 8732264f3..22739a836 100644
--- a/pdns/test-misc_hh.cc
+++ b/pdns/test-misc_hh.cc
@@ -5,8 +5,11 @@
 #include <boost/assign/list_of.hpp>
 #include <boost/foreach.hpp>
 #include <boost/tuple/tuple.hpp>
+#include <math.h>
 #include "misc.hh"
 #include "dns.hh"
+#include "dnswriter.hh"
+#include "dnsrecords.hh"
 #include <utility>
 
 using std::string;
@@ -119,6 +122,40 @@ BOOST_AUTO_TEST_CASE(test_AtomicCounter) {
     BOOST_CHECK_EQUAL(ac, 2);
 }
 
+BOOST_AUTO_TEST_CASE(test_QuestionHash) {
+  vector<unsigned char> packet;
+  reportBasicTypes();
+  DNSPacketWriter dpw1(packet, "www.ds9a.nl.", QType::AAAA);
+  
+  uint32_t hash1=hashQuestion((char*)&packet[0], packet.size(), 0);
+  DNSPacketWriter dpw2(packet, "wWw.Ds9A.nL.", QType::AAAA);
+  uint32_t hash2=hashQuestion((char*)&packet[0], packet.size(), 0);
+  BOOST_CHECK_EQUAL(hash1, hash2);
+ 
+  vector<uint32_t> counts(1500);
+ 
+  for(unsigned int n=0; n < 100000; ++n) {
+    packet.clear();
+    DNSPacketWriter dpw1(packet, lexical_cast<string>(n)+"."+lexical_cast<string>(n*2)+".", QType::AAAA);
+    counts[hashQuestion((char*)&packet[0], packet.size(), 0) % counts.size()]++;
+  }
+  
+  double sum = 0.0;
+  BOOST_FOREACH(double d, counts){
+    sum += d;
+  }
+
+  double m =  sum / counts.size();
+
+  double accum = 0.0;
+  BOOST_FOREACH(double d, counts) {
+      accum += (d - m) * (d - m);
+  };
+      
+  double stdev = sqrt(accum / (counts.size()-1));
+  BOOST_CHECK(stdev < 10);
+}
+
 BOOST_AUTO_TEST_CASE(test_endianness) {
   uint32_t i = 1;
 #if BYTE_ORDER == BIG_ENDIAN