host_firewall: drop inbound traffic from internet-wide scanners (OpenFilters internet-scanners), v4 and v6, daily refresh

host_firewall/defaults/main.yml

@@ -54,7 +54,7 @@ firewall:
     # Lets the kernel log the nft log statements of container network
     # namespaces (LXC guests) into the host kernel log.
     all_netns: false
-    # Shared by the Spamhaus (fw-spamhaus-in: , fw-spamhaus-out: ) and FireHOL (fw-firehol-in: , fw-firehol-out: ) chains, one limit each.
+    # Shared by the Spamhaus (fw-spamhaus-in: , fw-spamhaus-out: ), FireHOL (fw-firehol-in: , fw-firehol-out: ) and scanner (fw-scanners-in: ) chains, one limit each.
     spamhaus_rate: "6/minute burst 10 packets"
   # Spamhaus DROP, DROPv6 and ASN-DROP: traffic from the listed networks drops
   # (input, forward), traffic to them is rejected (output, forward). The sets
@@ -65,6 +65,16 @@ firewall:
   # private or reserved space and ones holding an address of the host are left
   # out. Refreshed daily by nft-host-firehol.
   firehol: true
+  # Internet-wide scanners (Censys, Shodan, ONYPHE, Shadowserver, ...), IPv4
+  # and IPv6: the cidr files of github.com/OpenFilters/internet-scanners
+  # (Unlicense). Inbound traffic from the listed networks drops (input,
+  # forward) once the established rule has passed; outbound traffic is never
+  # filtered. Log prefix fw-scanners-in: with the Spamhaus log limit, which
+  # CrowdSec whitelists. Entries in private or reserved space, overlapping the
+  # admin, scraper and cluster sets, the CrowdSec whitelist or the public_ips
+  # of any host, and ones holding an address of the host are left out
+  # (/etc/nft-host-scanners.allowlist). Refreshed daily by nft-host-scanners.
+  scanners: true
   # Reject outbound TCP 25; the mail group turns it off.
   block_smtp_out: true
   # EU-27 plus CH, NO, GB, IS, LI (ipdeny country codes). Empty leaves the geo

host_firewall/files/nft-host-scanners

@@ -0,0 +1,182 @@
+#!/usr/bin/python3
+# Managed by Ansible (role host_firewall), do not edit.
+"""Rebuilds the internet scanner sets scanners_v4 and scanners_v6 of table inet host.
+
+Source: the cidr directory of OpenFilters/internet-scanners (Unlicense), read
+from the tarball of the default branch. A failed download or sanity check
+exits non-zero before anything changes, so the previous sets stay in place.
+"""
+import io
+import ipaddress
+import json
+import os
+import re
+import shutil
+import subprocess
+import sys
+import tarfile
+import tempfile
+import time
+import urllib.request
+
+PERSISTED = "/etc/nftables.d/host-scanners.nft"
+# One CIDR or address per line: the public addresses of the hosts, the
+# CrowdSec whitelist and the admin, scraper and cluster sets of the firewall.
+ALLOWLIST = "/etc/nft-host-scanners.allowlist"
+URL = "https://codeload.github.com/OpenFilters/internet-scanners/tar.gz/refs/heads/main"
+USER_AGENT = "ruekov-host-firewall (nft-host-scanners)"
+MEMBER = re.compile(r"^[^/]+/cidr/[a-z0-9]+_v([46])\.txt$")
+MAX_BYTES = 50 * 1024 * 1024
+
+# Lower bounds that catch an empty or truncated answer (the list holds about
+# 21000 IPv4 and 3000 IPv6 entries).
+MIN_ENTRIES = {4: 5000, 6: 500}
+# A prefix this wide is a parse error, not a scanner network.
+MIN_PREFIXLEN = {4: 12, 6: 24}
+# Private, CGNAT (NetBird), loopback, link-local, ULA and multicast/reserved
+# space: an entry overlapping one of these never lands in a set, so the own
+# infrastructure cannot block itself.
+EXCLUDED = [ipaddress.ip_network(n) for n in (
+    "0.0.0.0/8", "10.0.0.0/8", "100.64.0.0/10", "127.0.0.0/8",
+    "169.254.0.0/16", "172.16.0.0/12", "192.168.0.0/16", "224.0.0.0/3",
+    "::/128", "::1/128", "::ffff:0:0/96", "fc00::/7", "fe80::/10", "ff00::/8")]
+
+
+def fail(message):
+    print(f"nft-host-scanners: {message}", file=sys.stderr)
+    sys.exit(1)
+
+
+def fetch():
+    error = None
+    for attempt in range(3):
+        try:
+            request = urllib.request.Request(URL, headers={"User-Agent": USER_AGENT})
+            with urllib.request.urlopen(request, timeout=120) as response:
+                data = response.read(MAX_BYTES + 1)
+            if len(data) > MAX_BYTES:
+                fail(f"answer from {URL} exceeds {MAX_BYTES} bytes")
+            return data
+        except OSError as e:
+            error = e
+            time.sleep(5 * (attempt + 1))
+    fail(f"download of {URL} failed: {error}")
+
+
+def parse(data):
+    """Returns {4: [networks], 6: [networks]} from the cidr/*_v4.txt and *_v6.txt files."""
+    networks = {4: [], 6: []}
+    try:
+        archive = tarfile.open(fileobj=io.BytesIO(data), mode="r:gz")
+    except (tarfile.TarError, OSError) as e:
+        fail(f"unreadable archive: {e}")
+    with archive:
+        for member in archive:
+            match = MEMBER.match(member.name)
+            if not match or not member.isfile():
+                continue
+            version = int(match.group(1))
+            for line in archive.extractfile(member).read().decode("utf-8").splitlines():
+                line = line.strip()
+                if not line or line.startswith("#"):
+                    continue
+                try:
+                    # Host bits set (34.19.127.176/27) are masked, not rejected.
+                    network = ipaddress.ip_network(line, strict=False)
+                except ValueError:
+                    fail(f"unparsable line in {member.name}: {line!r}")
+                if network.version != version:
+                    fail(f"IPv{network.version} entry in {member.name}: {line!r}")
+                networks[version].append(network)
+    return networks
+
+
+def read_allowlist():
+    entries = []
+    with open(ALLOWLIST, encoding="utf-8") as f:
+        for line in f:
+            line = line.strip()
+            if not line or line.startswith("#"):
+                continue
+            try:
+                entries.append(ipaddress.ip_network(line, strict=False))
+            except ValueError:
+                fail(f"unparsable line in {ALLOWLIST}: {line!r}")
+    return entries
+
+
+def own_addresses():
+    result = subprocess.run(["ip", "-j", "address", "show"], capture_output=True, text=True)
+    if result.returncode != 0:
+        fail(f"ip address show failed: {result.stderr.strip()}")
+    return [ipaddress.ip_address(a["local"]) for link in json.loads(result.stdout)
+            for a in link.get("addr_info", []) if a.get("family") in ("inet", "inet6")]
+
+
+def main():
+    listed = parse(fetch())
+    for version, minimum in MIN_ENTRIES.items():
+        if len(listed[version]) < minimum:
+            fail(f"only {len(listed[version])} IPv{version} entries")
+
+    protected = EXCLUDED + read_allowlist()
+    own = own_addresses()
+    merged = {}
+    for version in (4, 6):
+        kept, protected_skipped, own_skipped = [], 0, 0
+        for network in listed[version]:
+            if network.prefixlen < MIN_PREFIXLEN[version]:
+                fail(f"suspiciously wide prefix {network}")
+            if any(p.version == version and network.overlaps(p) for p in protected):
+                protected_skipped += 1
+            elif any(a.version == version and a in network for a in own):
+                own_skipped += 1
+            else:
+                kept.append(network)
+        merged[version] = list(ipaddress.collapse_addresses(kept))
+        print(f"scanners_v{version}: {len(listed[version])} entries, {protected_skipped} in private, reserved or "
+              f"allowlisted space and {own_skipped} holding an own address skipped -> {len(merged[version])} merged")
+
+    persisted = "# Generated by nft-host-scanners, do not edit.\ntable inet host {\n"
+    for version in (4, 6):
+        elements = ",\n\t\t".join(str(n) for n in merged[version]) + ","
+        persisted += (f"\tset scanners_v{version} {{ type ipv{version}_addr; flags interval; auto-merge; elements = {{\n"
+                      f"\t\t{elements}\n\t}} }}\n")
+    persisted += "}\n"
+
+    # Checked under another table name: a dry run against the live table
+    # name changes the live sets for its duration and unlisted addresses match.
+    check_rules = persisted.replace("table inet host {", "table inet host_scanners_check {")
+    checked = subprocess.run(["nft", "-c", "-f", "-"], input=check_rules, capture_output=True, text=True)
+    if checked.returncode != 0:
+        fail(f"nft rejects the sets: {checked.stderr.strip()}")
+
+    with tempfile.TemporaryDirectory() as tmp:
+        previous = None
+        if os.path.exists(PERSISTED):
+            previous = os.path.join(tmp, "previous.nft")
+            shutil.copyfile(PERSISTED, previous)
+        with open(PERSISTED + ".new", "w", encoding="utf-8") as f:
+            f.write(persisted)
+        os.chmod(PERSISTED + ".new", 0o644)
+        os.replace(PERSISTED + ".new", PERSISTED)
+
+        # A reload replaces the whole table, sets included, in one transaction.
+        active = subprocess.run(["systemctl", "-q", "is-active", "nft-host.service"]).returncode == 0
+        if active:
+            reloaded = subprocess.run(["systemctl", "reload", "nft-host.service"], capture_output=True, text=True)
+            if reloaded.returncode != 0:
+                if previous:
+                    shutil.copyfile(previous, PERSISTED)
+                else:
+                    os.remove(PERSISTED)
+                fail(f"reloading the rule set failed, kept the previous sets: {reloaded.stderr.strip()}")
+        print("reloaded the rule set" if active else "persisted only")
+
+
+try:
+    main()
+except SystemExit:
+    raise
+except Exception as e:  # keep the previous sets on any surprise
+    fail(f"{type(e).__name__}: {e}")

host_firewall/files/nft-host-scanners.service

@@ -0,0 +1,8 @@
+[Unit]
+Description=Refresh the internet scanner sets of the host firewall
+Wants=network-online.target
+After=network-online.target
+
+[Service]
+Type=oneshot
+ExecStart=/usr/local/sbin/nft-host-scanners

host_firewall/files/nft-host-scanners.timer

@@ -0,0 +1,10 @@
+[Unit]
+Description=Daily refresh of the host firewall internet scanner sets
+
+[Timer]
+OnCalendar=daily
+RandomizedDelaySec=6h
+Persistent=true
+
+[Install]
+WantedBy=timers.target

host_firewall/tasks/main.yml

@@ -43,7 +43,7 @@
     src: "{{ item }}"
     dest: "/etc/systemd/system/{{ item }}"
     mode: "0644"
-  loop: [nft-host.service, nft-host-geo.service, nft-host-geo.timer, nft-host-spamhaus.service, nft-host-spamhaus.timer, nft-host-firehol.service, nft-host-firehol.timer]
+  loop: [nft-host.service, nft-host-geo.service, nft-host-geo.timer, nft-host-spamhaus.service, nft-host-spamhaus.timer, nft-host-firehol.service, nft-host-firehol.timer, nft-host-scanners.service, nft-host-scanners.timer]
   register: firewall_units
 
 - name: Install the rollback script
@@ -211,6 +211,66 @@
         path: /etc/nftables.d/host-firehol.nft
         state: absent
 
+- name: Internet scanner sets
+  when: firewall.scanners
+  block:
+    - name: Install the internet scanner set updater
+      ansible.builtin.copy:
+        src: nft-host-scanners
+        dest: /usr/local/sbin/nft-host-scanners
+        mode: "0755"
+
+    # The updater leaves out entries overlapping these addresses.
+    - name: Install the internet scanner allowlist
+      ansible.builtin.template:
+        src: nft-host-scanners.allowlist.j2
+        dest: /etc/nft-host-scanners.allowlist
+        mode: "0644"
+
+    # The rule set includes this file, so a missing one would fail the load at boot.
+    - name: Build the internet scanner sets on first run
+      ansible.builtin.command: /usr/local/sbin/nft-host-scanners
+      args:
+        creates: /etc/nftables.d/host-scanners.nft
+
+    - name: Read the internet scanner sets file
+      ansible.builtin.stat:
+        path: /etc/nftables.d/host-scanners.nft
+      register: firewall_scanners
+
+    - name: Require non-empty internet scanner sets
+      ansible.builtin.assert:
+        that: firewall_scanners.stat.exists and firewall_scanners.stat.size > 0
+        fail_msg: /etc/nftables.d/host-scanners.nft is missing or empty, refusing to apply the firewall
+      # Check mode neither installs the updater nor builds the sets.
+      when: not ansible_check_mode
+
+    - name: Enable the internet scanner set timer
+      ansible.builtin.systemd:
+        name: nft-host-scanners.timer
+        enabled: true
+        state: started
+        daemon_reload: "{{ firewall_units is changed }}"
+      # Check mode does not install the unit.
+      ignore_errors: "{{ ansible_check_mode }}"
+
+- name: Stop the internet scanner set refresh
+  when: not firewall.scanners
+  block:
+    - name: Disable the internet scanner set timer
+      ansible.builtin.systemd:
+        name: nft-host-scanners.timer
+        enabled: false
+        state: stopped
+        daemon_reload: "{{ firewall_units is changed }}"
+      # Check mode does not install the unit.
+      ignore_errors: "{{ ansible_check_mode }}"
+
+    - name: Remove the internet scanner sets file
+      ansible.builtin.file:
+        path: /etc/nftables.d/host-scanners.nft
+        state: absent
+
 - name: Render the rule set
   ansible.builtin.template:
     src: host.nft.j2

host_firewall/templates/host.nft.j2

@@ -38,6 +38,9 @@ include "/etc/nftables.d/host-spamhaus.nft"
 {% if firewall.firehol %}
 include "/etc/nftables.d/host-firehol.nft"
 {% endif %}
+{% if firewall.scanners %}
+include "/etc/nftables.d/host-scanners.nft"
+{% endif %}
 
 table inet host {
 	{{ addr_set('admin_v4', 'ipv4_addr', firewall.admin.v4) }}
@@ -69,6 +72,13 @@ table inet host {
 
 {% endif %}
 		ct state established,related accept
+{% if firewall.scanners %}
+
+		# After the established rule: only inbound connections of scanners drop.
+		ip saddr @scanners_v4 jump scanners_in
+		ip6 saddr @scanners_v6 jump scanners_in
+
+{% endif %}
 		ct state invalid jump drop_log
 		iif lo accept
 
@@ -143,6 +153,16 @@ table inet host {
 		reject
 	}
 {% endif %}
+{% if firewall.scanners %}
+
+	# Own limit and prefix, so scanners use up neither the fw-in-drop budget
+	# nor the one of the blocklists. CrowdSec whitelists the prefix: scans of
+	# these networks are known and need no decision.
+	chain scanners_in {
+		limit rate {{ firewall.log.spamhaus_rate }} log prefix "fw-scanners-in: "
+		drop
+	}
+{% endif %}
 {% if firewall.block_smtp_out or blocklists %}
 
 	# Outbound mail goes through the mail host only. Loopback stays open for a
@@ -220,7 +240,7 @@ table inet host {
 		ip6 saddr @geo_v6 accept
 {% endif %}
 	}
-{% if forward_guard or blocklists %}
+{% if forward_guard or blocklists or firewall.scanners %}
 
 	chain forward {
 		type filter hook forward priority filter; policy accept;
@@ -236,6 +256,11 @@ table inet host {
 		ip saddr @firehol_v4 jump firehol_in
 		ip daddr @firehol_v4 jump firehol_out
 {% endif %}
+{% if firewall.scanners %}
+
+		ct state new ip saddr @scanners_v4 jump scanners_in
+		ct state new ip6 saddr @scanners_v6 jump scanners_in
+{% endif %}
 {% for guard in forward_guard %}
 
 		# {{ guard.key }}

host_firewall/templates/nft-host-scanners.allowlist.j2

@@ -0,0 +1,11 @@
+# {{ ansible_managed }}
+{% set maps = [
+  (crowdsec | default({}, true)).whitelist | default({}, true),
+  firewall.admin.v4 | default({}, true), firewall.admin.v6 | default({}, true),
+  firewall.scrapers.v4 | default({}, true), firewall.scrapers.v6 | default({}, true),
+  firewall.cluster.v4 | default({}, true), firewall.cluster.v6 | default({}, true)] %}
+{% set entries = (maps | map('dict2items') | sum(start=[]) | rejectattr('value', 'none') | map(attribute='value') | map('string') | list
+  + groups['all'] | map('extract', hostvars, 'public_ips') | select('defined') | flatten) | unique %}
+{% for entry in entries %}
+{{ entry }}
+{% endfor %}