Tested-by: Stefan Schantl <stefan.schantl(a)ipfire.org>
> NFQUEUE does not let the packet continue where it was processed, but
> inserts it back into iptables at the start. That is why we need an
> extra IPSBYPASS chain which has the following tasks:
> 
> * Make the BYPASS bit permanent for the entire connection
> * Clear the REPEAT bit
> 
> The latter is more of cosmetic nature so that we can identify packets
> that have come from suricata again and those which have bypassed the
> IPS
> straight away.
> 
> The IPS_* chain will now only be sent traffic to, when none of the
> two
> relevant bits has been set. Otherwise the packet has already been
> processed by suricata in the first pass or suricata has decided to
> bypass the connection.
> 
> This massively reduces load on the IPS which allows many common
> connections (TLS connections with downloads) to bypass the IPS
> bringing
> us back to line speed.
> 
> Signed-off-by: Michael Tremer <michael.tremer(a)ipfire.org>
> ---
>  src/initscripts/system/firewall | 23 ++++++++++++++++++++---
>  src/initscripts/system/suricata | 27 +++------------------------
>  2 files changed, 23 insertions(+), 27 deletions(-)
> 
> diff --git a/src/initscripts/system/firewall
> b/src/initscripts/system/firewall
> index ce428393d..530e8f1d6 100644
> --- a/src/initscripts/system/firewall
> +++ b/src/initscripts/system/firewall
> @@ -17,6 +17,11 @@ NAT_MASK="0x0f000000"
>  IPSEC_MARK="0x00800000"
>  IPSEC_MASK="${IPSEC_MARK}"
>  
> +IPS_REPEAT_MARK="0x80000000"
> +IPS_REPEAT_MASK="0x80000000"
> +IPS_BYPASS_MARK="0x40000000"
> +IPS_BYPASS_MASK="0x40000000"
> +
>  function iptables() {
>         /sbin/iptables --wait "$@"
>  }
> @@ -41,6 +46,17 @@ iptables_init() {
>         modprobe nf_log_ipv4
>         sysctl -q -w net.netfilter.nf_log.2=nf_log_ipv4
>  
> +       # IPS Bypass Chain which stores the BYPASS bit in connection
> tracking
> +       iptables -N IPSBYPASS
> +       iptables -A IPSBYPASS -j MARK --set-xmark "0/$((
> IPS_REPEAT_MASK ))"
> +       iptables -A IPSBYPASS -j CONNMARK --save-mark
> +
> +       # Jump into bypass chain when the BYPASS bit is set
> +       for chain in INPUT FORWARD OUTPUT; do
> +               iptables -A "${chain}" -m mark \
> +                       --mark "$(( IPS_REPEAT_MARK | IPS_BYPASS_MARK
> ))/$(( IPS_REPEAT_MASK | IPS_BYPASS_MASK ))" -j IPSBYPASS
> +       done
> +
>         # Empty LOG_DROP and LOG_REJECT chains
>         iptables -N LOG_DROP
>         iptables -A LOG_DROP   -m limit --limit 10/second -j LOG
> @@ -147,9 +163,10 @@ iptables_init() {
>         iptables -N IPS_INPUT
>         iptables -N IPS_FORWARD
>         iptables -N IPS_OUTPUT
> -       iptables -A INPUT -j IPS_INPUT
> -       iptables -A FORWARD -j IPS_FORWARD
> -       iptables -A OUTPUT -j IPS_OUTPUT
> +
> +       for chain in INPUT FORWARD OUTPUT; do
> +               iptables -A "${chain}" -m mark --mark "0x0/$((
> IPS_REPEAT_MASK | IPS_BYPASS_MASK ))" -j "IPS_${chain}"
> +       done
>  
>         # OpenVPN transfer network translation
>         iptables -t nat -N OVPNNAT
> diff --git a/src/initscripts/system/suricata
> b/src/initscripts/system/suricata
> index 72d01b91d..13fcc7f34 100644
> --- a/src/initscripts/system/suricata
> +++ b/src/initscripts/system/suricata
> @@ -34,12 +34,6 @@ network_zones=( red green blue orange ovpn )
>  # Array to store the network zones weather the IPS is enabled for.
>  enabled_ips_zones=()
>  
> -# Mark and Mask options.
> -REPEAT_MARK="0x80000000"
> -REPEAT_MASK="0x80000000"
> -BYPASS_MARK="0x40000000"
> -BYPASS_MASK="0x40000000"
> -
>  # PID file of suricata.
>  PID_FILE="/var/run/suricata.pid"
>  
> @@ -134,34 +128,19 @@ function generate_fw_rules {
>         # Flush the firewall chains.
>         flush_fw_chain
>  
> -       # Skip anything that has the bypass bit set
> -       local chain
> -       for chain in "${IPS_INPUT_CHAIN}" "${IPS_FORWARD_CHAIN}"
> "${IPS_OUTPUT_CHAIN}"; do
> -               iptables -w -A "${chain}" -m mark --mark
> "${BYPASS_MARK}/${BYPASS_MASK}" -j RETURN
> -       done
> -
>         # Check if the array of enabled_ips_zones contains any
> elements.
>         if [[ ${enabled_ips_zones[@]} ]]; then
>                 # Loop through the array and create firewall rules.
>                 for enabled_ips_zone in "${enabled_ips_zones[@]}"; do
>                         # Create rules queue input and output related
> traffic and pass it to the IPS.
> -                       iptables -w -A "$IPS_INPUT_CHAIN" -i
> "$enabled_ips_zone" -m mark ! --mark "${REPEAT_MARK}/${REPEAT_MASK}"
> -j NFQUEUE $NFQ_OPTIONS
> -                       iptables -w -A "$IPS_OUTPUT_CHAIN" -o
> "$enabled_ips_zone" -m mark ! --mark "${REPEAT_MARK}/${REPEAT_MASK}"
> -j NFQUEUE $NFQ_OPTIONS
> +                       iptables -w -A "$IPS_INPUT_CHAIN" -i
> "$enabled_ips_zone" -j NFQUEUE $NFQ_OPTIONS
> +                       iptables -w -A "$IPS_OUTPUT_CHAIN" -o
> "$enabled_ips_zone" -j NFQUEUE $NFQ_OPTIONS
>  
>                         # Create rules which are required to handle
> forwarded traffic.
>                         for enabled_ips_zone_forward in
> "${enabled_ips_zones[@]}"; do
> -                               iptables -w -A "$IPS_FORWARD_CHAIN" -
> i "$enabled_ips_zone" -o "$enabled_ips_zone_forward" -m mark ! --mark
> "${REPEAT_MARK}/${REPEAT_MASK}" -j NFQUEUE $NFQ_OPTIONS
> +                               iptables -w -A "$IPS_FORWARD_CHAIN" -
> i "$enabled_ips_zone" -o "$enabled_ips_zone_forward" -j NFQUEUE
> $NFQ_OPTIONS
>                         done
>                 done
> -
> -               # Add common rules at the end of the chain
> -               for chain in "${IPS_INPUT_CHAIN}"
> "${IPS_FORWARD_CHAIN}" "${IPS_OUTPUT_CHAIN}"; do
> -                       # Clear repeat bit
> -                       iptables -w -A "${chain}" -j MARK --set-xmark
> "0x0/${REPEAT_MASK}"
> -
> -                       # Store bypass bit in CONNMARK
> -                       iptables -w -A "${chain}" -m mark --mark
> "${BYPASS_MARK}/${BYPASS_MASK}" -j CONNMARK --save-mark
> -               done
>         fi
>  }
>