Update seed IPs, based on bitcoin.sipa.be crawler data

Rebased-From: be22b3da1df15757ef15e4c4963507bcdff6a872
Github-Pull: #5532

2 files changed, 108 insertions, 25 deletions

diff --git a/contrib/seeds/README.md b/contrib/seeds/README.md
index f9a0c277e2..bc88201f0f 100644
--- a/contrib/seeds/README.md
+++ b/contrib/seeds/README.md
@@ -1,11 +1,8 @@
 ### Seeds ###
 
-Utility to generate the pnSeed[] array that is compiled into the client
-(see [src/net.cpp](/src/net.cpp)).
+Utility to generate the seeds.txt list that is compiled into the client
+(see [src/chainparamsseeds.h](/src/chainparamsseeds.h) and [share/seeds](/share/seeds)).
 
-The 600 seeds compiled into the 0.8 release were created from sipa's DNS seed data, like this:
+The 512 seeds compiled into the 0.10 release were created from sipa's DNS seed data, like this:
 
-	curl -s http://bitcoin.sipa.be/seeds.txt | head -1000 | makeseeds.py
-
-The input to makeseeds.py is assumed to be approximately sorted from most-reliable to least-reliable,
-with IP:port first on each line (lines that don't match IPv4:port are ignored).
+	curl -s http://bitcoin.sipa.be/seeds.txt | makeseeds.py
diff --git a/contrib/seeds/makeseeds.py b/contrib/seeds/makeseeds.py
index 1d01fd7d20..b831395f2c 100755
--- a/contrib/seeds/makeseeds.py
+++ b/contrib/seeds/makeseeds.py
@@ -1,32 +1,118 @@
 #!/usr/bin/env python
 #
-# Generate pnSeed[] from Pieter's DNS seeder
+# Generate seeds.txt from Pieter's DNS seeder
 #
 
-NSEEDS=600
+NSEEDS=512
+
+MAX_SEEDS_PER_ASN=2
+
+MIN_BLOCKS = 337600
+
+# These are hosts that have been observed to be behaving strangely (e.g.
+# aggressively connecting to every node).
+SUSPICIOUS_HOSTS = set([
+    "130.211.129.106", "178.63.107.226",
+    "83.81.130.26", "88.198.17.7", "148.251.238.178", "176.9.46.6",
+    "54.173.72.127", "54.174.10.182", "54.183.64.54", "54.194.231.211",
+    "54.66.214.167", "54.66.220.137", "54.67.33.14", "54.77.251.214",
+    "54.94.195.96", "54.94.200.247"
+])
 
 import re
 import sys
-from subprocess import check_output
+import dns.resolver
+
+PATTERN_IPV4 = re.compile(r"^((\d{1,3})\.(\d{1,3})\.(\d{1,3})\.(\d{1,3})):8333$")
+PATTERN_AGENT = re.compile(r"^(\/Satoshi:0.8.6\/|\/Satoshi:0.9.(2|3)\/|\/Satoshi:0.10.\d{1,2}\/)$")
+
+def parseline(line):
+    sline = line.split()
+    if len(sline) < 11:
+       return None
+    # Match only IPv4
+    m = PATTERN_IPV4.match(sline[0])
+    if m is None:
+        return None
+    # Do IPv4 sanity check
+    ip = 0
+    for i in range(0,4):
+        if int(m.group(i+2)) < 0 or int(m.group(i+2)) > 255:
+            return None
+        ip = ip + (int(m.group(i+2)) << (8*(3-i)))
+    if ip == 0:
+        return None
+    # Skip bad results.
+    if sline[1] == 0:
+        return None
+    # Extract uptime %.
+    uptime30 = float(sline[7][:-1])
+    # Extract Unix timestamp of last success.
+    lastsuccess = int(sline[2])
+    # Extract protocol version.
+    version = int(sline[10])
+    # Extract user agent.
+    agent = sline[11][1:-1]
+    # Extract service flags.
+    service = int(sline[9], 16)
+    # Extract blocks.
+    blocks = int(sline[8])
+    # Construct result.
+    return {
+        'ip': m.group(1),
+        'ipnum': ip,
+        'uptime': uptime30,
+        'lastsuccess': lastsuccess,
+        'version': version,
+        'agent': agent,
+        'service': service,
+        'blocks': blocks,
+    }
+
+# Based on Greg Maxwell's seed_filter.py
+def filterbyasn(ips, max_per_asn, max_total):
+    result = []
+    asn_count = {}
+    for ip in ips:
+        if len(result) == max_total:
+            break
+        try:
+            asn = int([x.to_text() for x in dns.resolver.query('.'.join(reversed(ip['ip'].split('.'))) + '.origin.asn.cymru.com', 'TXT').response.answer][0].split('\"')[1].split(' ')[0])
+            if asn not in asn_count:
+                asn_count[asn] = 0
+            if asn_count[asn] == max_per_asn:
+                continue
+            asn_count[asn] += 1
+            result.append(ip)
+        except:
+            sys.stderr.write('ERR: Could not resolve ASN for "' + ip['ip'] + '"\n')
+    return result
 
 def main():
     lines = sys.stdin.readlines()
+    ips = [parseline(line) for line in lines]
+
+    # Skip entries with valid IPv4 address.
+    ips = [ip for ip in ips if ip is not None]
+    # Skip entries from suspicious hosts.
+    ips = [ip for ip in ips if ip['ip'] not in SUSPICIOUS_HOSTS]
+    # Enforce minimal number of blocks.
+    ips = [ip for ip in ips if ip['blocks'] >= MIN_BLOCKS]
+    # Require service bit 1.
+    ips = [ip for ip in ips if (ip['service'] & 1) == 1]
+    # Require at least 50% 30-day uptime.
+    ips = [ip for ip in ips if ip['uptime'] > 50]
+    # Require a known and recent user agent.
+    ips = [ip for ip in ips if PATTERN_AGENT.match(ip['agent'])]
+    # Sort by availability (and use last success as tie breaker)
+    ips.sort(key=lambda x: (x['uptime'], x['lastsuccess'], x['ip']), reverse=True)
+    # Look up ASNs and limit results, both per ASN and globally.
+    ips = filterbyasn(ips, MAX_SEEDS_PER_ASN, NSEEDS)
+    # Sort the results by IP address (for deterministic output).
+    ips.sort(key=lambda x: (x['ipnum']))
 
-    ips = []
-    pattern = re.compile(r"^(\d{1,3})\.(\d{1,3})\.(\d{1,3})\.(\d{1,3}):8333")
-    for line in lines:
-        m = pattern.match(line)
-        if m is None:
-            continue
-        ip = 0
-        for i in range(0,4):
-            ip = ip + (int(m.group(i+1)) << (8*(i)))
-        if ip == 0:
-            continue
-        ips.append(ip)
-
-    for row in range(0, min(NSEEDS,len(ips)), 8):
-        print "    " + ", ".join([ "0x%08x"%i for i in ips[row:row+8] ]) + ","
+    for ip in ips:
+        print ip['ip']
 
 if __name__ == '__main__':
     main()