diff --git a/heartbeat/IPaddr2 b/heartbeat/IPaddr2 index 2aaebe8b7..5fd8dba11 100755 --- a/heartbeat/IPaddr2 +++ b/heartbeat/IPaddr2 @@ -1,1229 +1,1238 @@ #!/bin/sh # # $Id: IPaddr2.in,v 1.24 2006/08/09 13:01:54 lars Exp $ # # OCF Resource Agent compliant IPaddr2 script. # # Based on work by Tuomo Soini, ported to the OCF RA API by Lars # Marowsky-Brée. Implements Cluster Alias IP functionality too. # # Cluster Alias IP cleanup, fixes and testing by Michael Schwartzkopff # # # Copyright (c) 2003 Tuomo Soini # Copyright (c) 2004-2006 SUSE LINUX AG, Lars Marowsky-Brée # All Rights Reserved. # # This program is free software; you can redistribute it and/or modify # it under the terms of version 2 of the GNU General Public License as # published by the Free Software Foundation. # # This program is distributed in the hope that it would be useful, but # WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. # # Further, this software is distributed without any warranty that it is # free of the rightful claim of any third person regarding infringement # or the like. Any license provided herein, whether implied or # otherwise, applies only to this software file. Patent licenses, if # any, provided herein do not apply to combinations of this program with # other software, or any other product whatsoever. # # You should have received a copy of the GNU General Public License # along with this program; if not, write the Free Software Foundation, # Inc., 59 Temple Place - Suite 330, Boston MA 02111-1307, USA. # # # TODO: # - There ought to be an ocf_run_cmd function which does all logging, # timeout handling etc for us # - Make this the standard IP address agent on Linux; the other # platforms simply should ignore the additional parameters OR can use # the legacy heartbeat resource script... # - Check LVS <-> clusterip incompatibilities. # # OCF parameters are as below # OCF_RESKEY_ip # OCF_RESKEY_broadcast # OCF_RESKEY_nic # OCF_RESKEY_cidr_netmask # OCF_RESKEY_iflabel # OCF_RESKEY_mac # OCF_RESKEY_clusterip_hash # OCF_RESKEY_arp_interval # OCF_RESKEY_arp_count # OCF_RESKEY_arp_bg # OCF_RESKEY_preferred_lft # # OCF_RESKEY_CRM_meta_clone # OCF_RESKEY_CRM_meta_clone_max ####################################################################### # Initialization: : ${OCF_FUNCTIONS_DIR=${OCF_ROOT}/lib/heartbeat} . ${OCF_FUNCTIONS_DIR}/ocf-shellfuncs . ${OCF_FUNCTIONS_DIR}/findif.sh # Defaults OCF_RESKEY_lvs_support_default=false OCF_RESKEY_lvs_ipv6_addrlabel_default=false OCF_RESKEY_lvs_ipv6_addrlabel_value_default=99 OCF_RESKEY_clusterip_hash_default="sourceip-sourceport" OCF_RESKEY_unique_clone_address_default=false OCF_RESKEY_arp_interval_default=200 OCF_RESKEY_arp_count_default=5 OCF_RESKEY_arp_count_refresh_default=0 OCF_RESKEY_arp_bg_default=true OCF_RESKEY_run_arping_default=false OCF_RESKEY_preferred_lft_default="forever" : ${OCF_RESKEY_lvs_support=${OCF_RESKEY_lvs_support_default}} : ${OCF_RESKEY_lvs_ipv6_addrlabel=${OCF_RESKEY_lvs_ipv6_addrlabel_default}} : ${OCF_RESKEY_lvs_ipv6_addrlabel_value=${OCF_RESKEY_lvs_ipv6_addrlabel_value_default}} : ${OCF_RESKEY_clusterip_hash=${OCF_RESKEY_clusterip_hash_default}} : ${OCF_RESKEY_unique_clone_address=${OCF_RESKEY_unique_clone_address_default}} : ${OCF_RESKEY_arp_interval=${OCF_RESKEY_arp_interval_default}} : ${OCF_RESKEY_arp_count=${OCF_RESKEY_arp_count_default}} : ${OCF_RESKEY_arp_count_refresh=${OCF_RESKEY_arp_count_refresh_default}} : ${OCF_RESKEY_arp_bg=${OCF_RESKEY_arp_bg_default}} : ${OCF_RESKEY_run_arping=${OCF_RESKEY_run_arping_default}} : ${OCF_RESKEY_preferred_lft=${OCF_RESKEY_preferred_lft_default}} ####################################################################### SENDARP=$HA_BIN/send_arp SENDUA=$HA_BIN/send_ua FINDIF=findif VLDIR=$HA_RSCTMP SENDARPPIDDIR=$HA_RSCTMP CIP_lockfile=$HA_RSCTMP/IPaddr2-CIP-${OCF_RESKEY_ip} ####################################################################### meta_data() { cat < 1.0 This Linux-specific resource manages IP alias IP addresses. It can add an IP alias, or remove one. In addition, it can implement Cluster Alias IP functionality if invoked as a clone resource. If used as a clone, you should explicitly set clone-node-max >= 2, and/or clone-max < number of nodes. In case of node failure, clone instances need to be re-allocated on surviving nodes. This would not be possible if there is already an instance on those nodes, and clone-node-max=1 (which is the default). Manages virtual IPv4 and IPv6 addresses (Linux specific version) The IPv4 (dotted quad notation) or IPv6 address (colon hexadecimal notation) example IPv4 "192.168.1.1". example IPv6 "2001:db8:DC28:0:0:FC57:D4C8:1FFF". IPv4 or IPv6 address The base network interface on which the IP address will be brought online. If left empty, the script will try and determine this from the routing table. Do NOT specify an alias interface in the form eth0:1 or anything here; rather, specify the base interface only. If you want a label, see the iflabel parameter. Prerequisite: There must be at least one static IP address, which is not managed by the cluster, assigned to the network interface. If you can not assign any static IP address on the interface, modify this kernel parameter: sysctl -w net.ipv4.conf.all.promote_secondaries=1 # (or per device) Network interface The netmask for the interface in CIDR format (e.g., 24 and not 255.255.255.0) If unspecified, the script will also try to determine this from the routing table. CIDR netmask Broadcast address associated with the IP. It is possible to use the special symbols '+' and '-' instead of the broadcast address. In this case, the broadcast address is derived by setting/resetting the host bits of the interface prefix. Broadcast address You can specify an additional label for your IP address here. This label is appended to your interface name. The kernel allows alphanumeric labels up to a maximum length of 15 characters including the interface name and colon (e.g. eth0:foobar1234) A label can be specified in nic parameter but it is deprecated. If a label is specified in nic name, this parameter has no effect. Interface label Enable support for LVS Direct Routing configurations. In case a IP address is stopped, only move it to the loopback device to allow the local node to continue to service requests, but no longer advertise it on the network. Notes for IPv6: It is not necessary to enable this option on IPv6. Instead, enable 'lvs_ipv6_addrlabel' option for LVS-DR usage on IPv6. Enable support for LVS DR Enable adding IPv6 address label so IPv6 traffic originating from the address's interface does not use this address as the source. This is necessary for LVS-DR health checks to realservers to work. Without it, the most recently added IPv6 address (probably the address added by IPaddr2) will be used as the source address for IPv6 traffic from that interface and since that address exists on loopback on the realservers, the realserver response to pings/connections will never leave its loopback. See RFC3484 for the detail of the source address selection. See also 'lvs_ipv6_addrlabel_value' parameter. Enable adding IPv6 address label. Specify IPv6 address label value used when 'lvs_ipv6_addrlabel' is enabled. The value should be an unused label in the policy table which is shown by 'ip addrlabel list' command. You would rarely need to change this parameter. IPv6 address label value. Set the interface MAC address explicitly. Currently only used in case of the Cluster IP Alias. Leave empty to chose automatically. Cluster IP MAC address Specify the hashing algorithm used for the Cluster IP functionality. Cluster IP hashing function If true, add the clone ID to the supplied value of IP to create a unique address to manage Create a unique address for cloned instances Specify the interval between unsolicited ARP packets in milliseconds. This parameter is deprecated and used for the backward compatibility only. It is effective only for the send_arp binary which is built with libnet, and send_ua for IPv6. It has no effect for other arp_sender. ARP packet interval in ms (deprecated) Number of unsolicited ARP packets to send at resource initialization. ARP packet count sent during initialization Number of unsolicited ARP packets to send during resource monitoring. Doing so helps mitigate issues of stuck ARP caches resulting from split-brain situations. ARP packet count sent during monitoring Whether or not to send the ARP packets in the background. ARP from background The program to send ARP packets with on start. Available options are: - send_arp: default - ipoibarping: default for infiniband interfaces if ipoibarping is available - iputils_arping: use arping in iputils package - libnet_arping: use another variant of arping based on libnet ARP sender Extra options to pass to the arp_sender program. Available options are vary depending on which arp_sender is used. A typical use case is specifying '-A' for iputils_arping to use ARP REPLY instead of ARP REQUEST as Gratuitous ARPs. Options for ARP sender Flush the routing table on stop. This is for applications which use the cluster IP address and which run on the same physical host that the IP address lives on. The Linux kernel may force that application to take a shortcut to the local loopback interface, instead of the interface the address is really bound to. Under those circumstances, an application may, somewhat unexpectedly, continue to use connections for some time even after the IP address is deconfigured. Set this parameter in order to immediately disable said shortcut when the IP address goes away. Flush kernel routing table on stop Whether or not to run arping for IPv4 collision detection check. Run arping for IPv4 collision detection check For IPv6, set the preferred lifetime of the IP address. This can be used to ensure that the created IP address will not be used as a source address for routing. Expects a value as specified in section 5.5.4 of RFC 4862. IPv6 preferred lifetime END exit $OCF_SUCCESS } ip_init() { local rc if [ X`uname -s` != "XLinux" ]; then ocf_exit_reason "IPaddr2 only supported Linux." exit $OCF_ERR_INSTALLED fi if [ X"$OCF_RESKEY_ip" = "X" ] && [ "$__OCF_ACTION" != "stop" ]; then ocf_exit_reason "IP address (the ip parameter) is mandatory" exit $OCF_ERR_CONFIGURED fi if case $__OCF_ACTION in start|stop) ocf_is_root;; *) true;; esac then : YAY! else ocf_exit_reason "You must be root for $__OCF_ACTION operation." exit $OCF_ERR_PERM fi BASEIP="$OCF_RESKEY_ip" BRDCAST="$OCF_RESKEY_broadcast" NIC="$OCF_RESKEY_nic" # Note: We had a version out there for a while which used # netmask instead of cidr_netmask. Don't remove this aliasing code! if [ ! -z "$OCF_RESKEY_netmask" -a -z "$OCF_RESKEY_cidr_netmask" ] then OCF_RESKEY_cidr_netmask=$OCF_RESKEY_netmask export OCF_RESKEY_cidr_netmask fi NETMASK="$OCF_RESKEY_cidr_netmask" IFLABEL="$OCF_RESKEY_iflabel" IF_MAC="$OCF_RESKEY_mac" IP_INC_GLOBAL=${OCF_RESKEY_CRM_meta_clone_max:-1} IP_INC_NO=`expr ${OCF_RESKEY_CRM_meta_clone:-0} + 1` if ocf_is_true ${OCF_RESKEY_lvs_support} && [ $IP_INC_GLOBAL -gt 1 ]; then ocf_exit_reason "LVS and load sharing do not go together well" exit $OCF_ERR_CONFIGURED fi if ocf_is_decimal "$IP_INC_GLOBAL" && [ $IP_INC_GLOBAL -gt 0 ]; then : else ocf_exit_reason "Invalid meta-attribute clone_max [$IP_INC_GLOBAL], should be positive integer" exit $OCF_ERR_CONFIGURED fi echo $OCF_RESKEY_ip | grep -qs ":" if [ $? -ne 0 ];then FAMILY=inet if ocf_is_true $OCF_RESKEY_lvs_ipv6_addrlabel ;then ocf_exit_reason "IPv4 does not support lvs_ipv6_addrlabel" exit $OCF_ERR_CONFIGURED fi else FAMILY=inet6 if ocf_is_true $OCF_RESKEY_lvs_support ;then ocf_exit_reason "The IPv6 does not support lvs_support" exit $OCF_ERR_CONFIGURED fi if ocf_is_true $OCF_RESKEY_lvs_ipv6_addrlabel ;then if ocf_is_decimal "$OCF_RESKEY_lvs_ipv6_addrlabel_value" && [ $OCF_RESKEY_lvs_ipv6_addrlabel_value -ge 0 ]; then : else ocf_exit_reason "Invalid lvs_ipv6_addrlabel_value [$OCF_RESKEY_lvs_ipv6_addrlabel_value], should be positive integer" exit $OCF_ERR_CONFIGURED fi fi fi # support nic:iflabel format in nic parameter case $NIC in *:*) IFLABEL=`echo $NIC | sed 's/[^:]*://'` NIC=`echo $NIC | sed 's/:.*//'` # only the base name should be passed to findif OCF_RESKEY_nic=$NIC ;; esac # $FINDIF takes its parameters from the environment # NICINFO=`$FINDIF` rc=$? if [ $rc -eq 0 ] then NICINFO=`echo "$NICINFO" | sed -e 's/netmask\ //;s/broadcast\ //'` NIC=`echo "$NICINFO" | cut -d" " -f1` NETMASK=`echo "$NICINFO" | cut -d" " -f2` BRDCAST=`echo "$NICINFO" | cut -d" " -f3` else # findif couldn't find the interface if ocf_is_probe; then ocf_log info "[$FINDIF] failed" exit $OCF_NOT_RUNNING elif [ "$__OCF_ACTION" = stop ]; then ocf_log warn "[$FINDIF] failed" exit $OCF_SUCCESS else ocf_exit_reason "[$FINDIF] failed" exit $rc fi fi SENDARPPIDFILE="$SENDARPPIDDIR/send_arp-$OCF_RESKEY_ip" if [ -n "$IFLABEL" ]; then IFLABEL=${NIC}:${IFLABEL} if [ ${#IFLABEL} -gt 15 ]; then ocf_exit_reason "Interface label [$IFLABEL] exceeds maximum character limit of 15" exit $OCF_ERR_CONFIGURED fi fi if [ "$IP_INC_GLOBAL" -gt 1 ] && ! ocf_is_true "$OCF_RESKEY_unique_clone_address"; then IP_CIP="yes" IP_CIP_HASH="${OCF_RESKEY_clusterip_hash}" if [ -z "$IF_MAC" ]; then # Choose a MAC # 1. Concatenate some input together # 2. This doesn't need to be a cryptographically # secure hash. # 3. Drop everything after the first 6 octets (12 chars) # 4. Delimit the octets with ':' # 5. Make sure the first octet is odd, # so the result is a multicast MAC IF_MAC=`echo $OCF_RESKEY_ip $NETMASK $BRDCAST | \ md5sum | \ sed -e 's#\(............\).*#\1#' \ -e 's#..#&:#g; s#:$##' \ -e 's#^\(.\)[02468aAcCeE]#\11#'` fi IP_CIP_FILE="/proc/net/ipt_CLUSTERIP/$OCF_RESKEY_ip" fi } # # Find out which interfaces serve the given IP address and netmask. # The arguments are an IP address and a netmask. # Its output are interface names devided by spaces (e.g., "eth0 eth1"). # find_interface() { local ipaddr="$1" local netmask="$2" # # List interfaces but exclude FreeS/WAN ipsecN virtual interfaces # local iface="`$IP2UTIL -o -f $FAMILY addr show \ | grep "\ $ipaddr/$netmask" \ | cut -d ' ' -f2 \ | grep -v '^ipsec[0-9][0-9]*$'`" echo "$iface" return 0 } # # Delete an interface # delete_interface () { ipaddr="$1" iface="$2" netmask="$3" CMD="$IP2UTIL -f $FAMILY addr delete $ipaddr/$netmask dev $iface" ocf_run $CMD || return $OCF_ERR_GENERIC if ocf_is_true $OCF_RESKEY_flush_routes; then ocf_run $IP2UTIL route flush cache fi if [ "$FAMILY" = "inet6" ] && ocf_is_true $OCF_RESKEY_lvs_ipv6_addrlabel ;then delete_ipv6_addrlabel $ipaddr fi return $OCF_SUCCESS } # # Add an interface # add_interface () { local cmd msg ipaddr netmask broadcast iface label ipaddr="$1" netmask="$2" broadcast="$3" iface="$4" label="$5" if [ "$FAMILY" = "inet" ] && ocf_is_true $OCF_RESKEY_run_arping && check_binary arping; then arping -q -c 2 -w 3 -D -I $iface $ipaddr if [ $? = 1 ]; then ocf_log err "IPv4 address collision $ipaddr [DAD]" return $OCF_ERR_CONFIGURED fi fi if [ "$FAMILY" = "inet6" ] && ocf_is_true $OCF_RESKEY_lvs_ipv6_addrlabel ;then add_ipv6_addrlabel $ipaddr fi cmd="$IP2UTIL -f $FAMILY addr add $ipaddr/$netmask dev $iface" msg="Adding $FAMILY address $ipaddr/$netmask to device $iface" if [ "$broadcast" != "none" ]; then cmd="$IP2UTIL -f $FAMILY addr add $ipaddr/$netmask brd $broadcast dev $iface" msg="Adding $FAMILY address $ipaddr/$netmask with broadcast address $broadcast to device $iface" fi if [ ! -z "$label" ]; then cmd="$cmd label $label" msg="${msg} (with label $label)" fi if [ "$FAMILY" = "inet6" ] ;then cmd="$cmd preferred_lft $OCF_RESKEY_preferred_lft" msg="${msg} (with preferred_lft $OCF_RESKEY_preferred_lft)" fi ocf_log info "$msg" ocf_run $cmd || return $OCF_ERR_GENERIC msg="Bringing device $iface up" cmd="$IP2UTIL link set $iface up" ocf_log info "$msg" ocf_run $cmd || return $OCF_ERR_GENERIC return $OCF_SUCCESS } # # Delete a route # delete_route () { prefix="$1" iface="$2" CMD="$IP2UTIL route delete $prefix dev $iface" ocf_log info "$CMD" $CMD return $? } # On Linux systems the (hidden) loopback interface may # conflict with the requested IP address. If so, this # unoriginal code will remove the offending loopback address # and save it in VLDIR so it can be added back in later # when the IPaddr is released. # # TODO: This is very ugly and should be controlled by an additional # instance parameter. Or even: multi-state, with the IP only being # "active" on the master!? # remove_conflicting_loopback() { ipaddr="$1" netmask="$2" broadcast="$3" ifname="$4" ocf_log info "Removing conflicting loopback $ifname." if echo "$ipaddr $netmask $broadcast $ifname" > "$VLDIR/$ipaddr" then : Saved loopback information in $VLDIR/$ipaddr else ocf_log err "Could not save conflicting loopback $ifname." \ "it will not be restored." fi delete_interface "$ipaddr" "$ifname" "$netmask" # Forcibly remove the route (if it exists) to the loopback. delete_route "$ipaddr" "$ifname" } # # On Linux systems the (hidden) loopback interface may # need to be restored if it has been taken down previously # by remove_conflicting_loopback() # restore_loopback() { ipaddr="$1" if [ -s "$VLDIR/$ipaddr" ]; then ifinfo=`cat "$VLDIR/$ipaddr"` ocf_log info "Restoring loopback IP Address " \ "$ifinfo." add_interface $ifinfo rm -f "$VLDIR/$ipaddr" fi } add_ipv6_addrlabel() { local cmd ipaddr value ipaddr="$1" value="$OCF_RESKEY_lvs_ipv6_addrlabel_value" cmd="$IP2UTIL addrlabel add prefix $ipaddr label $value" ocf_log info "Adding IPv6 address label prefix $ipaddr label $value" ocf_run $cmd || ocf_log warn "$cmd failed." } delete_ipv6_addrlabel() { local cmd ipaddr value ipaddr="$1" value="$OCF_RESKEY_lvs_ipv6_addrlabel_value" cmd="$IP2UTIL addrlabel del prefix $ipaddr label $value" ocf_run $cmd # an error can be ignored } is_infiniband() { $IP2UTIL link show $NIC | grep link/infiniband >/dev/null } log_arp_sender() { local cmdline local output local rc cmdline="$@" output=$($cmdline 2>&1) rc=$? if [ $rc -ne 0 ] && \ [ "$ARP_SENDER" != "libnet_arping" ] ; then # libnet_arping always return an error as no answers ocf_log err "Could not send gratuitous arps: rc=$rc" fi ocf_log $LOGLEVEL "$output" } # wrapper function to manage PID file to run arping in background run_with_pidfile() { local cmdline local pid local rc cmdline="$@" $cmdline & pid=$! echo "$pid" > $SENDARPPIDFILE wait $pid rc=$? rm -f $SENDARPPIDFILE return $rc } build_arp_sender_cmd() { case "$ARP_SENDER" in send_arp) if [ "x$IP_CIP" = "xyes" ] ; then if [ x = "x$IF_MAC" ] ; then MY_MAC=auto else + # send_arp.linux should return without doing anything in this case MY_MAC=`echo ${IF_MAC} | sed -e 's/://g'` fi else MY_MAC=auto fi ARGS="$OCF_RESKEY_send_arp_opts -i $OCF_RESKEY_arp_interval -r $ARP_COUNT -p $SENDARPPIDFILE $NIC $OCF_RESKEY_ip $MY_MAC not_used not_used" ARP_SENDER_CMD="$SENDARP $ARGS" ;; iputils_arping) ARGS="$OCF_RESKEY_send_arp_opts -U -c $ARP_COUNT -I $NIC $OCF_RESKEY_ip" ARP_SENDER_CMD="run_with_pidfile arping $ARGS" ;; libnet_arping) ARGS="$OCF_RESKEY_send_arp_opts -U -c $ARP_COUNT -i $NIC -S $OCF_RESKEY_ip $OCF_RESKEY_ip" ARP_SENDER_CMD="run_with_pidfile arping $ARGS" ;; ipoibarping) ARGS="-q -c $ARP_COUNT -U -I $NIC $OCF_RESKEY_ip" ARP_SENDER_CMD="ipoibarping $ARGS" ;; *) # should not occur ocf_exit_reason "unrecognized arp_sender value: $ARP_SENDER" exit $OCF_ERR_GENERIC ;; esac } # # Send Unsolicited ARPs to update neighbor's ARP cache # run_arp_sender() { if [ "x$1" = "xrefresh" ] ; then ARP_COUNT=$OCF_RESKEY_arp_count_refresh LOGLEVEL=debug else ARP_COUNT=$OCF_RESKEY_arp_count LOGLEVEL=info fi if [ $ARP_COUNT -eq 0 ] ; then return fi + # do not need to send Gratuitous ARPs in the Cluster IP configuration + # except send_arp.libnet binary to retain the old behavior + if [ "x$IP_CIP" = "xyes" ] && \ + [ "x$ARP_SENDER" != "xsend_arp" ] ; then + ocf_log info "Gratuitous ARPs are not sent in the Cluster IP configuration" + return + fi + # prepare arguments for each arp sender program # $ARP_SENDER_CMD should be set build_arp_sender_cmd ocf_log $LOGLEVEL "$ARP_SENDER_CMD" if ocf_is_true $OCF_RESKEY_arp_bg; then log_arp_sender $ARP_SENDER_CMD & else log_arp_sender $ARP_SENDER_CMD fi } # # Run send_ua to note send ICMPv6 Unsolicited Neighbor Advertisements. # run_send_ua() { local i # Duplicate Address Detection [DAD] # Kernel will flag the IP as 'tentative' until it ensured that # there is no duplicates. # If there is, it will flag it as 'dadfailed' for i in $(seq 1 10); do ipstatus=$($IP2UTIL -o -f $FAMILY addr show dev $NIC to $OCF_RESKEY_ip/$NETMASK) case "$ipstatus" in *dadfailed*) ocf_log err "IPv6 address collision $OCF_RESKEY_ip [DAD]" $IP2UTIL -f $FAMILY addr del dev $NIC $OCF_RESKEY_ip/$NETMASK if [ $? -ne 0 ]; then ocf_log err "Could not delete IPv6 address" fi return $OCF_ERR_GENERIC ;; *tentative*) if [ $i -eq 10 ]; then ocf_log warn "IPv6 address : DAD is still in tentative" fi ;; *) break ;; esac sleep 1 done # Now the address should be usable ARGS="-i $OCF_RESKEY_arp_interval -c $OCF_RESKEY_arp_count $OCF_RESKEY_ip $NETMASK $NIC" ocf_log info "$SENDUA $ARGS" $SENDUA $ARGS || ocf_log err "Could not send ICMPv6 Unsolicited Neighbor Advertisements." } # Do we already serve this IP address on the given $NIC? # # returns: # ok = served (for CIP: + hash bucket) # partial = served and no hash bucket (CIP only) # partial2 = served and no CIP iptables rule # no = nothing # ip_served() { if [ -z "$NIC" ]; then # no nic found or specified echo "no" return 0 fi cur_nic="`find_interface $OCF_RESKEY_ip $NETMASK`" if [ -z "$cur_nic" ]; then echo "no" return 0 fi if [ -z "$IP_CIP" ]; then for i in $cur_nic; do # only mark as served when on the same interfaces as $NIC [ "$i" = "$NIC" ] || continue echo "ok" return 0 done # There used to be logic here to pretend "not served", # if ${OCF_RESKEY_lvs_support} was enabled, and the IP was # found active on "lo*" only. With lvs_support on, you should # have NIC != lo, so thats already filtered # by the continue above. echo "no" return 0 fi # Special handling for the CIP: if [ ! -e $IP_CIP_FILE ]; then echo "partial2" return 0 fi if egrep -q "(^|,)${IP_INC_NO}(,|$)" $IP_CIP_FILE ; then echo "ok" return 0 else echo "partial" return 0 fi exit $OCF_ERR_GENERIC } ####################################################################### ip_usage() { cat <$IP_CIP_FILE fi if [ "$ip_status" = "no" ]; then if ocf_is_true ${OCF_RESKEY_lvs_support}; then for i in `find_interface $OCF_RESKEY_ip 32`; do case $i in lo*) remove_conflicting_loopback $OCF_RESKEY_ip 32 255.255.255.255 lo ;; esac done fi add_interface $OCF_RESKEY_ip $NETMASK ${BRDCAST:-none} $NIC $IFLABEL rc=$? if [ $rc -ne $OCF_SUCCESS ]; then ocf_exit_reason "Failed to add $OCF_RESKEY_ip" exit $rc fi fi case $NIC in lo*) : no need to run send_arp on loopback ;; *) if [ $FAMILY = "inet" ];then run_arp_sender else if [ -x $SENDUA ]; then run_send_ua if [ $? -ne 0 ]; then ocf_exit_reason "run_send_ua failed." exit $OCF_ERR_GENERIC fi fi fi ;; esac exit $OCF_SUCCESS } ip_stop() { local ip_del_if="yes" if [ -n "$IP_CIP" ]; then # Cluster IPs need special processing when the last bucket # is removed from the node... take a lock to make sure only one # process executes that code ocf_take_lock $CIP_lockfile ocf_release_lock_on_exit $CIP_lockfile fi if [ -f "$SENDARPPIDFILE" ] ; then kill `cat "$SENDARPPIDFILE"` if [ $? -ne 0 ]; then ocf_log warn "Could not kill previously running send_arp for $OCF_RESKEY_ip" else ocf_log info "killed previously running send_arp for $OCF_RESKEY_ip" fi rm -f "$SENDARPPIDFILE" fi local ip_status=`ip_served` ocf_log info "IP status = $ip_status, IP_CIP=$IP_CIP" if [ $ip_status = "no" ]; then : Requested interface not in use exit $OCF_SUCCESS fi if [ -n "$IP_CIP" ] && [ $ip_status != "partial2" ]; then if [ $ip_status = "partial" ]; then exit $OCF_SUCCESS fi echo "-$IP_INC_NO" >$IP_CIP_FILE if [ "x$(cat $IP_CIP_FILE)" = "x" ]; then ocf_log info $OCF_RESKEY_ip, $IP_CIP_HASH i=1 while [ $i -le $IP_INC_GLOBAL ]; do ocf_log info $i $IPTABLES -D INPUT -d $OCF_RESKEY_ip -i $NIC -j CLUSTERIP \ --new \ --clustermac $IF_MAC \ --total-nodes $IP_INC_GLOBAL \ --local-node $i \ --hashmode $IP_CIP_HASH i=`expr $i + 1` done else ip_del_if="no" fi fi if [ "$ip_del_if" = "yes" ]; then delete_interface $OCF_RESKEY_ip $NIC $NETMASK if [ $? -ne 0 ]; then ocf_exit_reason "Unable to remove IP [${OCF_RESKEY_ip} from interface [ $NIC ]" exit $OCF_ERR_GENERIC fi if ocf_is_true ${OCF_RESKEY_lvs_support}; then restore_loopback "$OCF_RESKEY_ip" fi fi exit $OCF_SUCCESS } ip_monitor() { # TODO: Implement more elaborate monitoring like checking for # interface health maybe via a daemon like FailSafe etc... local ip_status=`ip_served` case $ip_status in ok) run_arp_sender refresh return $OCF_SUCCESS ;; partial|no|partial2) exit $OCF_NOT_RUNNING ;; *) # Errors on this interface? return $OCF_ERR_GENERIC ;; esac } # make sure that we have something to send ARPs with set_send_arp_program() { ARP_SENDER=send_arp if [ -n "$OCF_RESKEY_arp_sender" ]; then case "$OCF_RESKEY_arp_sender" in send_arp) check_binary $SENDARP ;; iputils_arping) check_binary arping ;; libnet_arping) check_binary arping ;; ipoibarping) check_binary ipoibarping ;; *) ocf_exit_reason "unrecognized arp_sender value: $OCF_RESKEY_arp_sender" exit $OCF_ERR_CONFIGURED ;; esac ARP_SENDER="$OCF_RESKEY_arp_sender" else if is_infiniband; then ARP_SENDER=ipoibarping if ! have_binary ipoibarping; then [ "$__OCF_ACTION" = start ] && ocf_log warn "using send_arp for infiniband because ipoibarping is not available (set arp_sender to \"send_arp\" to suppress this message)" check_binary $SENDARP ARP_SENDER=send_arp fi fi fi } ip_validate() { check_binary $IP2UTIL IP_CIP= ip_init set_send_arp_program if [ -n "$IP_CIP" ]; then check_binary $IPTABLES check_binary $MODPROBE fi # $BASEIP, $NETMASK, $NIC , $IP_INC_GLOBAL, and $BRDCAST have been checked within ip_init, # do not bother here. if ocf_is_true "$OCF_RESKEY_unique_clone_address" && ! ocf_is_true "$OCF_RESKEY_CRM_meta_globally_unique"; then ocf_exit_reason "unique_clone_address makes sense only with meta globally_unique set" exit $OCF_ERR_CONFIGURED fi if ocf_is_decimal "$OCF_RESKEY_arp_interval" && [ $OCF_RESKEY_arp_interval -gt 0 ]; then : else ocf_exit_reason "Invalid OCF_RESKEY_arp_interval [$OCF_RESKEY_arp_interval]" exit $OCF_ERR_CONFIGURED fi if ocf_is_decimal "$OCF_RESKEY_arp_count" && [ $OCF_RESKEY_arp_count -gt 0 ]; then : else ocf_exit_reason "Invalid OCF_RESKEY_arp_count [$OCF_RESKEY_arp_count]" exit $OCF_ERR_CONFIGURED fi if [ -z "$OCF_RESKEY_preferred_lft" ]; then ocf_exit_reason "Empty value is invalid for OCF_RESKEY_preferred_lft" exit $OCF_ERR_CONFIGURED fi if [ -n "$IP_CIP" ]; then local valid=1 case $IP_CIP_HASH in sourceip|sourceip-sourceport|sourceip-sourceport-destport) ;; *) ocf_exit_reason "Invalid OCF_RESKEY_clusterip_hash [$IP_CIP_HASH]" exit $OCF_ERR_CONFIGURED ;; esac if ocf_is_true ${OCF_RESKEY_lvs_support}; then ocf_exit_reason "LVS and load sharing not advised to try" exit $OCF_ERR_CONFIGURED fi case $IF_MAC in [0-9a-zA-Z][13579bBdDfF][!0-9a-zA-Z][0-9a-zA-Z][0-9a-zA-Z][!0-9a-zA-Z][0-9a-zA-Z][0-9a-zA-Z][!0-9a-zA-Z][0-9a-zA-Z][0-9a-zA-Z][!0-9a-zA-Z][0-9a-zA-Z][0-9a-zA-Z][!0-9a-zA-Z][0-9a-zA-Z][0-9a-zA-Z]) ;; *) valid=0 ;; esac if [ $valid -eq 0 ]; then ocf_exit_reason "Invalid IF_MAC [$IF_MAC]" exit $OCF_ERR_CONFIGURED fi fi } if ocf_is_true "$OCF_RESKEY_unique_clone_address"; then prefix=`echo $OCF_RESKEY_ip | awk -F. '{print $1"."$2"."$3}'` suffix=`echo $OCF_RESKEY_ip | awk -F. '{print $4}'` suffix=`expr ${OCF_RESKEY_CRM_meta_clone:-0} + $suffix` OCF_RESKEY_ip="$prefix.$suffix" fi case $__OCF_ACTION in meta-data) meta_data ;; usage|help) ip_usage exit $OCF_SUCCESS ;; esac ip_validate case $__OCF_ACTION in start) ip_start ;; stop) ip_stop ;; status) ip_status=`ip_served` if [ $ip_status = "ok" ]; then echo "running" exit $OCF_SUCCESS else echo "stopped" exit $OCF_NOT_RUNNING fi ;; monitor) ip_monitor ;; validate-all) ;; *) ip_usage exit $OCF_ERR_UNIMPLEMENTED ;; esac # vi:sw=4:ts=8: diff --git a/tools/send_arp.linux.c b/tools/send_arp.linux.c index ba3f13e4e..2aa9b5d44 100644 --- a/tools/send_arp.linux.c +++ b/tools/send_arp.linux.c @@ -1,1330 +1,1336 @@ /* * arping.c * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version * 2 of the License, or (at your option) any later version. * * Authors: Alexey Kuznetsov, * YOSHIFUJI Hideaki */ /* Andrew Beekhof, Lars Ellenberg: * Based on arping from iputils, * adapted to the command line conventions established by the libnet based * send_arp tool as used by the IPaddr and IPaddr2 resource agents. * The libnet based send_arp, and its command line argument convention, * was first added to the heartbeat project by Matt Soffen. * * Latest "resync" with iputils as of: * git://git.linux-ipv6.org/gitroot/iputils.git * 511f8356e22615479c3cc16bca64d72d204f6df3 * Fri Jul 24 10:48:47 2015 * To get various bugfixes and support for infiniband and other link layer * addresses which do not fit into plain "sockaddr_ll", and broadcast addresses * that may be different from memset(,0xff,). */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef CAPABILITIES #include #include #endif #include #include #include #include #include #include #include #include #ifdef USE_SYSFS #include struct sysfs_devattr_values; #endif #ifndef WITHOUT_IFADDRS #include #endif #ifdef USE_IDN #include #include #endif static char SNAPSHOT[] = "s20121221"; static void usage(void) __attribute__((noreturn)); #ifndef DEFAULT_DEVICE #define DEFAULT_DEVICE "eth0" #endif #ifdef DEFAULT_DEVICE # define DEFAULT_DEVICE_STR DEFAULT_DEVICE #else # define DEFAULT_DEVICE NULL #endif struct device { const char *name; int ifindex; #ifndef WITHOUT_IFADDRS struct ifaddrs *ifa; #endif #ifdef USE_SYSFS struct sysfs_devattr_values *sysfs; #endif }; int quit_on_reply=0; struct device device = { .name = DEFAULT_DEVICE, }; char *source; struct in_addr src, dst; char *target; int dad, unsolicited, advert; int quiet; int count=-1; int timeout; int unicasting; int s; int broadcast_only; struct sockaddr_storage me; struct sockaddr_storage he; struct timeval start, last; int sent, brd_sent; int received, brd_recv, req_recv; #ifndef CAPABILITIES static uid_t euid; #endif #define MS_TDIFF(tv1,tv2) ( ((tv1).tv_sec-(tv2).tv_sec)*1000 + \ ((tv1).tv_usec-(tv2).tv_usec)/1000 ) #define OFFSET_OF(name,ele) ((size_t)&((name *)0)->ele) static socklen_t sll_len(size_t halen) { socklen_t len = OFFSET_OF(struct sockaddr_ll, sll_addr) + halen; if (len < sizeof(struct sockaddr_ll)) len = sizeof(struct sockaddr_ll); return len; } #define SLL_LEN(hln) sll_len(hln) #if 1 /* hb_mode: always print hb_mode usage in this binary */ static char print_usage[]={ "send_arp: sends out custom ARP packet.\n" " usage: send_arp [-i repeatinterval-ms] [-r repeatcount] [-p pidfile] \\\n" " device src_ip_addr src_hw_addr broadcast_ip_addr netmask\n" "\n" " where:\n" " repeatinterval-ms: ignored\n" "\n" " repeatcount: how many ARP packets to send.\n" "\n" " pidfile: pid file to use\n" "\n" " device: network interface to use\n" "\n" " src_ip_addr: source ip address\n" "\n" " src_hw_addr: only \"auto\" is supported.\n" +" If other specified, it will exit without sending any ARP packets.\n" "\n" " broadcast_ip_addr: ignored\n" "\n" " netmask: ignored\n" "\n" " Notes: Other options of iputils-arping may be accepted but it's not\n" " intended to be supported in this binary.\n" "\n" }; void usage(void) { fprintf(stderr, "%s\n", print_usage); exit(2); } #else /* hb_mode */ void usage(void) { fprintf(stderr, "Usage: arping [-fqbDUAV] [-c count] [-w timeout] [-I device] [-s source] destination\n" " -f : quit on first reply\n" " -q : be quiet\n" " -b : keep broadcasting, don't go unicast\n" " -D : duplicate address detection mode\n" " -U : Unsolicited ARP mode, update your neighbours\n" " -A : ARP answer mode, update your neighbours\n" " -V : print version and exit\n" " -c count : how many packets to send\n" " -w timeout : how long to wait for a reply\n" " -I device : which ethernet device to use" #ifdef DEFAULT_DEVICE_STR " (" DEFAULT_DEVICE_STR ")" #endif "\n" " -s source : source ip address\n" " destination : ask for what ip address\n" ); exit(2); } #endif /* hb_mode */ static void set_signal(int signo, void (*handler)(void)) { struct sigaction sa; memset(&sa, 0, sizeof(sa)); sa.sa_handler = (void (*)(int))handler; sa.sa_flags = SA_RESTART; sigaction(signo, &sa, NULL); } #ifdef CAPABILITIES static const cap_value_t caps[] = { CAP_NET_RAW, }; static cap_flag_value_t cap_raw = CAP_CLEAR; #endif static void limit_capabilities(void) { #ifdef CAPABILITIES cap_t cap_p; cap_p = cap_get_proc(); if (!cap_p) { perror("arping: cap_get_proc"); exit(-1); } cap_get_flag(cap_p, CAP_NET_RAW, CAP_PERMITTED, &cap_raw); if (cap_raw != CAP_CLEAR) { if (cap_clear(cap_p) < 0) { perror("arping: cap_clear"); exit(-1); } cap_set_flag(cap_p, CAP_PERMITTED, 1, caps, CAP_SET); if (cap_set_proc(cap_p) < 0) { perror("arping: cap_set_proc"); if (errno != EPERM) exit(-1); } } if (prctl(PR_SET_KEEPCAPS, 1) < 0) { perror("arping: prctl"); exit(-1); } if (setuid(getuid()) < 0) { perror("arping: setuid"); exit(-1); } if (prctl(PR_SET_KEEPCAPS, 0) < 0) { perror("arping: prctl"); exit(-1); } cap_free(cap_p); #else euid = geteuid(); #endif } static int modify_capability_raw(int on) { #ifdef CAPABILITIES cap_t cap_p; if (cap_raw != CAP_SET) return on ? -1 : 0; cap_p = cap_get_proc(); if (!cap_p) { perror("arping: cap_get_proc"); return -1; } cap_set_flag(cap_p, CAP_EFFECTIVE, 1, caps, on ? CAP_SET : CAP_CLEAR); if (cap_set_proc(cap_p) < 0) { perror("arping: cap_set_proc"); return -1; } cap_free(cap_p); #else if (setuid(on ? euid : getuid())) { perror("arping: setuid"); return -1; } #endif return 0; } static int enable_capability_raw(void) { return modify_capability_raw(1); } static int disable_capability_raw(void) { return modify_capability_raw(0); } static void drop_capabilities(void) { #ifdef CAPABILITIES cap_t cap_p = cap_init(); if (!cap_p) { perror("arping: cap_init"); exit(-1); } if (cap_set_proc(cap_p) < 0) { perror("arping: cap_set_proc"); exit(-1); } cap_free(cap_p); #else if (setuid(getuid()) < 0) { perror("arping: setuid"); exit(-1); } #endif } static int send_pack(int s, struct in_addr src, struct in_addr dst, struct sockaddr_ll *ME, struct sockaddr_ll *HE) { int err; struct timeval now; unsigned char buf[256]; struct arphdr *ah = (struct arphdr*)buf; unsigned char *p = (unsigned char *)(ah+1); ah->ar_hrd = htons(ME->sll_hatype); if (ah->ar_hrd == htons(ARPHRD_FDDI)) ah->ar_hrd = htons(ARPHRD_ETHER); ah->ar_pro = htons(ETH_P_IP); ah->ar_hln = ME->sll_halen; ah->ar_pln = 4; ah->ar_op = advert ? htons(ARPOP_REPLY) : htons(ARPOP_REQUEST); memcpy(p, &ME->sll_addr, ah->ar_hln); p+=ME->sll_halen; memcpy(p, &src, 4); p+=4; if (advert) memcpy(p, &ME->sll_addr, ah->ar_hln); else memcpy(p, &HE->sll_addr, ah->ar_hln); p+=ah->ar_hln; memcpy(p, &dst, 4); p+=4; gettimeofday(&now, NULL); err = sendto(s, buf, p-buf, 0, (struct sockaddr*)HE, SLL_LEN(ah->ar_hln)); if (err == p-buf) { last = now; sent++; if (!unicasting) brd_sent++; } return err; } static void finish(void) { if (!quiet) { printf("Sent %d probes (%d broadcast(s))\n", sent, brd_sent); printf("Received %d response(s)", received); if (brd_recv || req_recv) { printf(" ("); if (req_recv) printf("%d request(s)", req_recv); if (brd_recv) printf("%s%d broadcast(s)", req_recv ? ", " : "", brd_recv); printf(")"); } printf("\n"); fflush(stdout); } fflush(stdout); if (dad) exit(!!received); if (unsolicited) exit(0); exit(!received); } static void catcher(void) { struct timeval tv, tv_s, tv_o; gettimeofday(&tv, NULL); if (start.tv_sec==0) start = tv; timersub(&tv, &start, &tv_s); tv_o.tv_sec = timeout; tv_o.tv_usec = 500 * 1000; if (count-- == 0 || (timeout && timercmp(&tv_s, &tv_o, >))) finish(); timersub(&tv, &last, &tv_s); tv_o.tv_sec = 0; if (last.tv_sec==0 || timercmp(&tv_s, &tv_o, >)) { send_pack(s, src, dst, (struct sockaddr_ll *)&me, (struct sockaddr_ll *)&he); if (count == 0 && unsolicited) finish(); } alarm(1); } static void print_hex(unsigned char *p, int len) { int i; for (i=0; isll_pkttype != PACKET_HOST && FROM->sll_pkttype != PACKET_BROADCAST && FROM->sll_pkttype != PACKET_MULTICAST) return 0; /* Only these types are recognised */ if (ah->ar_op != htons(ARPOP_REQUEST) && ah->ar_op != htons(ARPOP_REPLY)) return 0; /* ARPHRD check and this darned FDDI hack here :-( */ if (ah->ar_hrd != htons(FROM->sll_hatype) && (FROM->sll_hatype != ARPHRD_FDDI || ah->ar_hrd != htons(ARPHRD_ETHER))) return 0; /* Protocol must be IP. */ if (ah->ar_pro != htons(ETH_P_IP)) return 0; if (ah->ar_pln != 4) return 0; if (ah->ar_hln != ((struct sockaddr_ll *)&me)->sll_halen) return 0; if (len < sizeof(*ah) + 2*(4 + ah->ar_hln)) return 0; memcpy(&src_ip, p+ah->ar_hln, 4); memcpy(&dst_ip, p+ah->ar_hln+4+ah->ar_hln, 4); if (!dad) { if (src_ip.s_addr != dst.s_addr) return 0; if (src.s_addr != dst_ip.s_addr) return 0; if (memcmp(p+ah->ar_hln+4, ((struct sockaddr_ll *)&me)->sll_addr, ah->ar_hln)) return 0; } else { /* DAD packet was: src_ip = 0 (or some src) src_hw = ME dst_ip = tested address dst_hw = We fail, if receive request/reply with: src_ip = tested_address src_hw != ME if src_ip in request was not zero, check also that it matches to dst_ip, otherwise dst_ip/dst_hw do not matter. */ if (src_ip.s_addr != dst.s_addr) return 0; if (memcmp(p, ((struct sockaddr_ll *)&me)->sll_addr, ((struct sockaddr_ll *)&me)->sll_halen) == 0) return 0; if (src.s_addr && src.s_addr != dst_ip.s_addr) return 0; } if (!quiet) { int s_printed = 0; printf("%s ", FROM->sll_pkttype==PACKET_HOST ? "Unicast" : "Broadcast"); printf("%s from ", ah->ar_op == htons(ARPOP_REPLY) ? "reply" : "request"); printf("%s [", inet_ntoa(src_ip)); print_hex(p, ah->ar_hln); printf("] "); if (dst_ip.s_addr != src.s_addr) { printf("for %s ", inet_ntoa(dst_ip)); s_printed = 1; } if (memcmp(p+ah->ar_hln+4, ((struct sockaddr_ll *)&me)->sll_addr, ah->ar_hln)) { if (!s_printed) printf("for "); printf("["); print_hex(p+ah->ar_hln+4, ah->ar_hln); printf("]"); } if (last.tv_sec) { long usecs = (tv.tv_sec-last.tv_sec) * 1000000 + tv.tv_usec-last.tv_usec; long msecs = (usecs+500)/1000; usecs -= msecs*1000 - 500; printf(" %ld.%03ldms\n", msecs, usecs); } else { printf(" UNSOLICITED?\n"); } fflush(stdout); } received++; if (FROM->sll_pkttype != PACKET_HOST) brd_recv++; if (ah->ar_op == htons(ARPOP_REQUEST)) req_recv++; if (quit_on_reply || (count == 0 && received == sent)) finish(); if(!broadcast_only) { memcpy(((struct sockaddr_ll *)&he)->sll_addr, p, ((struct sockaddr_ll *)&me)->sll_halen); unicasting=1; } return 1; } #ifdef USE_SYSFS union sysfs_devattr_value { unsigned long ulong; void *ptr; }; enum { SYSFS_DEVATTR_IFINDEX, SYSFS_DEVATTR_FLAGS, SYSFS_DEVATTR_ADDR_LEN, #if 0 SYSFS_DEVATTR_TYPE, SYSFS_DEVATTR_ADDRESS, #endif SYSFS_DEVATTR_BROADCAST, SYSFS_DEVATTR_NUM }; struct sysfs_devattr_values { char *ifname; union sysfs_devattr_value value[SYSFS_DEVATTR_NUM]; }; static int sysfs_devattr_ulong_dec(char *ptr, struct sysfs_devattr_values *v, unsigned idx); static int sysfs_devattr_ulong_hex(char *ptr, struct sysfs_devattr_values *v, unsigned idx); static int sysfs_devattr_macaddr(char *ptr, struct sysfs_devattr_values *v, unsigned idx); struct sysfs_devattrs { const char *name; int (*handler)(char *ptr, struct sysfs_devattr_values *v, unsigned int idx); int free; } sysfs_devattrs[SYSFS_DEVATTR_NUM] = { [SYSFS_DEVATTR_IFINDEX] = { .name = "ifindex", .handler = sysfs_devattr_ulong_dec, }, [SYSFS_DEVATTR_ADDR_LEN] = { .name = "addr_len", .handler = sysfs_devattr_ulong_dec, }, [SYSFS_DEVATTR_FLAGS] = { .name = "flags", .handler = sysfs_devattr_ulong_hex, }, #if 0 [SYSFS_DEVATTR_TYPE] = { .name = "type", .handler = sysfs_devattr_ulong_dec, }, [SYSFS_DEVATTR_ADDRESS] = { .name = "address", .handler = sysfs_devattr_macaddr, .free = 1, }, #endif [SYSFS_DEVATTR_BROADCAST] = { .name = "broadcast", .handler = sysfs_devattr_macaddr, .free = 1, }, }; #endif static void byebye(int nsig) { /* Avoid an "error exit" log message if we're killed */ nsig = 0; exit(nsig); } /* * find_device() * * This function checks 1) if the device (if given) is okay for ARP, * or 2) find fist appropriate device on the system. * * Return value: * >0 : Succeeded, and appropriate device not found. * device.ifindex remains 0. * 0 : Succeeded, and approptiate device found. * device.ifindex is set. * <0 : Failed. Support not found, or other * : system error. Try other method. * * If an appropriate device found, it is recorded inside the * "device" variable for later reference. * * We have several implementations for this. * by_ifaddrs(): requires getifaddr() in glibc, and rtnetlink in * kernel. default and recommended for recent systems. * by_sysfs(): requires libsysfs , and sysfs in kernel. * by_ioctl(): unable to list devices without ipv4 address; this * means, you need to supply the device name for * DAD purpose. */ /* Common check for ifa->ifa_flags */ static int check_ifflags(unsigned int ifflags, int fatal) { if (!(ifflags & IFF_UP)) { if (fatal) { if (!quiet) printf("Interface \"%s\" is down\n", device.name); exit(2); } return -1; } if (ifflags & (IFF_NOARP | IFF_LOOPBACK)) { if (fatal) { if (!quiet) printf("Interface \"%s\" is not ARPable\n", device.name); exit(dad ? 0 : 2); } return -1; } return 0; } static int find_device_by_ifaddrs(void) { #ifndef WITHOUT_IFADDRS int rc; struct ifaddrs *ifa0, *ifa; int count = 0; rc = getifaddrs(&ifa0); if (rc) { perror("getifaddrs"); return -1; } for (ifa = ifa0; ifa; ifa = ifa->ifa_next) { if (!ifa->ifa_addr) continue; if (ifa->ifa_addr->sa_family != AF_PACKET) continue; if (device.name && ifa->ifa_name && strcmp(ifa->ifa_name, device.name)) continue; if (check_ifflags(ifa->ifa_flags, device.name != NULL) < 0) continue; if (!((struct sockaddr_ll *)ifa->ifa_addr)->sll_halen) continue; if (!ifa->ifa_broadaddr) continue; device.ifa = ifa; if (count++) break; } if (count == 1 && device.ifa) { device.ifindex = if_nametoindex(device.ifa->ifa_name); if (!device.ifindex) { perror("arping: if_nametoindex"); freeifaddrs(ifa0); return -1; } device.name = device.ifa->ifa_name; return 0; } return 1; #else return -1; #endif } #ifdef USE_SYSFS static void sysfs_devattr_values_init(struct sysfs_devattr_values *v, int do_free) { int i; if (do_free) { free(v->ifname); for (i = 0; i < SYSFS_DEVATTR_NUM; i++) { if (sysfs_devattrs[i].free) free(v->value[i].ptr); } } memset(v, 0, sizeof(*v)); } static int sysfs_devattr_ulong(char *ptr, struct sysfs_devattr_values *v, unsigned int idx, unsigned int base) { unsigned long *p; char *ep; if (!ptr || !v) return -1; p = &v->value[idx].ulong; errno = 0; *p = strtoul(ptr, &ep, base); if ((*ptr && isspace(*ptr & 0xff)) || errno || (*ep != '\0' && *ep != '\n')) goto out; return 0; out: return -1; } static int sysfs_devattr_ulong_dec(char *ptr, struct sysfs_devattr_values *v, unsigned int idx) { int rc = sysfs_devattr_ulong(ptr, v, idx, 10); return rc; } static int sysfs_devattr_ulong_hex(char *ptr, struct sysfs_devattr_values *v, unsigned int idx) { int rc = sysfs_devattr_ulong(ptr, v, idx, 16); return rc; } static int sysfs_devattr_macaddr(char *ptr, struct sysfs_devattr_values *v, unsigned int idx) { unsigned char *m; int i; unsigned int addrlen; if (!ptr || !v) return -1; addrlen = v->value[SYSFS_DEVATTR_ADDR_LEN].ulong; m = malloc(addrlen); for (i = 0; i < addrlen; i++) { if (i && *(ptr + i * 3 - 1) != ':') goto out; if (sscanf(ptr + i * 3, "%02hhx", &m[i]) != 1) goto out; } v->value[idx].ptr = m; return 0; out: free(m); return -1; } #endif static int find_device_by_sysfs(void) { int rc = -1; #ifdef USE_SYSFS struct sysfs_class *cls_net; struct dlist *dev_list; struct sysfs_class_device *dev; struct sysfs_attribute *dev_attr; struct sysfs_devattr_values sysfs_devattr_values; int count = 0; if (!device.sysfs) { device.sysfs = malloc(sizeof(*device.sysfs)); sysfs_devattr_values_init(device.sysfs, 0); } cls_net = sysfs_open_class("net"); if (!cls_net) { perror("sysfs_open_class"); return -1; } dev_list = sysfs_get_class_devices(cls_net); if (!dev_list) { perror("sysfs_get_class_devices"); goto out; } sysfs_devattr_values_init(&sysfs_devattr_values, 0); dlist_for_each_data(dev_list, dev, struct sysfs_class_device) { int i; int rc = -1; if (device.name && strcmp(dev->name, device.name)) goto do_next; sysfs_devattr_values_init(&sysfs_devattr_values, 1); for (i = 0; i < SYSFS_DEVATTR_NUM; i++) { dev_attr = sysfs_get_classdev_attr(dev, sysfs_devattrs[i].name); if (!dev_attr) { perror("sysfs_get_classdev_attr"); rc = -1; break; } if (sysfs_read_attribute(dev_attr)) { perror("sysfs_read_attribute"); rc = -1; break; } rc = sysfs_devattrs[i].handler(dev_attr->value, &sysfs_devattr_values, i); if (rc < 0) break; } if (rc < 0) goto do_next; if (check_ifflags(sysfs_devattr_values.value[SYSFS_DEVATTR_FLAGS].ulong, device.name != NULL) < 0) goto do_next; if (!sysfs_devattr_values.value[SYSFS_DEVATTR_ADDR_LEN].ulong) goto do_next; if (device.sysfs->value[SYSFS_DEVATTR_IFINDEX].ulong) { if (device.sysfs->value[SYSFS_DEVATTR_FLAGS].ulong & IFF_RUNNING) goto do_next; } sysfs_devattr_values.ifname = strdup(dev->name); if (!sysfs_devattr_values.ifname) { perror("malloc"); goto out; } sysfs_devattr_values_init(device.sysfs, 1); memcpy(device.sysfs, &sysfs_devattr_values, sizeof(*device.sysfs)); sysfs_devattr_values_init(&sysfs_devattr_values, 0); if (count++) break; continue; do_next: sysfs_devattr_values_init(&sysfs_devattr_values, 1); } if (count == 1) { device.ifindex = device.sysfs->value[SYSFS_DEVATTR_IFINDEX].ulong; device.name = device.sysfs->ifname; } rc = !device.ifindex; out: sysfs_close_class(cls_net); #endif return rc; } static int check_device_by_ioctl(int s, struct ifreq *ifr) { if (ioctl(s, SIOCGIFFLAGS, ifr) < 0) { perror("ioctl(SIOCGIFINDEX"); return -1; } if (check_ifflags(ifr->ifr_flags, device.name != NULL) < 0) return 1; if (ioctl(s, SIOCGIFINDEX, ifr) < 0) { perror("ioctl(SIOCGIFINDEX"); return -1; } return 0; } static int find_device_by_ioctl(void) { int s; struct ifreq *ifr0, *ifr, *ifr_end; size_t ifrsize = sizeof(*ifr); struct ifconf ifc; static struct ifreq ifrbuf; int count = 0; s = socket(AF_INET, SOCK_DGRAM, 0); if (s < 0) { perror("socket"); return -1; } memset(&ifrbuf, 0, sizeof(ifrbuf)); if (device.name) { strncpy(ifrbuf.ifr_name, device.name, sizeof(ifrbuf.ifr_name) - 1); if (check_device_by_ioctl(s, &ifrbuf)) goto out; count++; } else { do { int rc; ifr0 = malloc(ifrsize); if (!ifr0) { perror("malloc"); goto out; } ifc.ifc_buf = (char *)ifr0; ifc.ifc_len = ifrsize; rc = ioctl(s, SIOCGIFCONF, &ifc); if (rc < 0) { perror("ioctl(SIOCFIFCONF"); goto out; } if (ifc.ifc_len + sizeof(*ifr0) + sizeof(struct sockaddr_storage) - sizeof(struct sockaddr) <= ifrsize) break; ifrsize *= 2; free(ifr0); ifr0 = NULL; } while(ifrsize < INT_MAX / 2); if (!ifr0) { fprintf(stderr, "arping: too many interfaces!?\n"); goto out; } ifr_end = (struct ifreq *)(((char *)ifr0) + ifc.ifc_len - sizeof(*ifr0)); for (ifr = ifr0; ifr <= ifr_end; ifr++) { if (check_device_by_ioctl(s, &ifrbuf)) continue; memcpy(&ifrbuf.ifr_name, ifr->ifr_name, sizeof(ifrbuf.ifr_name)); if (count++) break; } } close(s); if (count == 1) { device.ifindex = ifrbuf.ifr_ifindex; device.name = ifrbuf.ifr_name; } return !device.ifindex; out: close(s); return -1; } static int find_device(void) { int rc; rc = find_device_by_ifaddrs(); if (rc >= 0) goto out; rc = find_device_by_sysfs(); if (rc >= 0) goto out; rc = find_device_by_ioctl(); out: return rc; } /* * set_device_broadcast() * * This fills the device "broadcast address" * based on information found by find_device() funcion. */ static int set_device_broadcast_ifaddrs_one(struct device *device, unsigned char *ba, size_t balen, int fatal) { #ifndef WITHOUT_IFADDRS struct ifaddrs *ifa; struct sockaddr_ll *sll; if (!device) return -1; ifa = device->ifa; if (!ifa) return -1; sll = (struct sockaddr_ll *)ifa->ifa_broadaddr; if (sll->sll_halen != balen) { if (fatal) { if (!quiet) printf("Address length does not match...\n"); exit(2); } return -1; } memcpy(ba, sll->sll_addr, sll->sll_halen); return 0; #else return -1; #endif } static int set_device_broadcast_sysfs(struct device *device, unsigned char *ba, size_t balen) { #ifdef USE_SYSFS struct sysfs_devattr_values *v; if (!device) return -1; v = device->sysfs; if (!v) return -1; if (v->value[SYSFS_DEVATTR_ADDR_LEN].ulong != balen) return -1; memcpy(ba, v->value[SYSFS_DEVATTR_BROADCAST].ptr, balen); return 0; #else return -1; #endif } static int set_device_broadcast_fallback(struct device *device, unsigned char *ba, size_t balen) { if (!quiet) fprintf(stderr, "WARNING: using default broadcast address.\n"); memset(ba, -1, balen); return 0; } static void set_device_broadcast(struct device *dev, unsigned char *ba, size_t balen) { if (!set_device_broadcast_ifaddrs_one(dev, ba, balen, 0)) return; if (!set_device_broadcast_sysfs(dev, ba, balen)) return; set_device_broadcast_fallback(dev, ba, balen); } int main(int argc, char **argv) { int socket_errno; int ch; int hb_mode = 0; signal(SIGTERM, byebye); signal(SIGPIPE, byebye); limit_capabilities(); #ifdef USE_IDN setlocale(LC_ALL, ""); #endif enable_capability_raw(); s = socket(PF_PACKET, SOCK_DGRAM, 0); socket_errno = errno; disable_capability_raw(); while ((ch = getopt(argc, argv, "h?bfDUAqc:w:s:I:Vr:i:p:")) != EOF) { switch(ch) { case 'b': broadcast_only=1; break; case 'D': dad++; quit_on_reply=1; break; case 'U': unsolicited++; break; case 'A': advert++; unsolicited++; break; case 'q': quiet++; break; case 'r': /* send_arp.libnet compatibility option */ hb_mode = 1; /* fall-through */ case 'c': count = atoi(optarg); break; case 'w': timeout = atoi(optarg); break; case 'I': device.name = optarg; break; case 'f': quit_on_reply=1; break; case 's': source = optarg; break; case 'V': printf("send_arp utility, based on arping from iputils-%s\n", SNAPSHOT); exit(0); case 'p': case 'i': hb_mode = 1; /* send_arp.libnet compatibility options, ignore */ break; case 'h': case '?': default: usage(); } } if(hb_mode) { /* send_arp.libnet compatibility mode */ if (argc - optind != 5) { usage(); return 1; } /* * argv[optind+1] DEVICE dc0,eth0:0,hme0:0, * argv[optind+2] IP 192.168.195.186 * argv[optind+3] MAC ADDR 00a0cc34a878 * argv[optind+4] BROADCAST 192.168.195.186 * argv[optind+5] NETMASK ffffffffffff */ unsolicited = 1; device.name = argv[optind]; target = argv[optind+1]; + if (strcmp(argv[optind+2], "auto")) { + fprintf(stderr, "send_arp.linux: Gratuitous ARPs are not sent in the Cluster IP configuration\n"); + /* return success to suppress an error log by the RA */ + exit(0); + } } else { argc -= optind; argv += optind; if (argc != 1) usage(); target = *argv; } if (device.name && !*device.name) device.name = NULL; if (s < 0) { errno = socket_errno; perror("arping: socket"); exit(2); } if (find_device() < 0) exit(2); if (!device.ifindex) { if (device.name) { fprintf(stderr, "arping: Device %s not available.\n", device.name); exit(2); } fprintf(stderr, "arping: device (option -I) is required.\n"); usage(); } if (inet_aton(target, &dst) != 1) { struct hostent *hp; char *idn = target; #ifdef USE_IDN int rc; rc = idna_to_ascii_lz(target, &idn, 0); if (rc != IDNA_SUCCESS) { fprintf(stderr, "arping: IDN encoding failed: %s\n", idna_strerror(rc)); exit(2); } #endif hp = gethostbyname2(idn, AF_INET); if (!hp) { fprintf(stderr, "arping: unknown host %s\n", target); exit(2); } #ifdef USE_IDN free(idn); #endif memcpy(&dst, hp->h_addr, 4); } if (source && inet_aton(source, &src) != 1) { fprintf(stderr, "arping: invalid source %s\n", source); exit(2); } if (!dad && unsolicited && src.s_addr == 0) src = dst; if (!dad || src.s_addr) { struct sockaddr_in saddr; int probe_fd = socket(AF_INET, SOCK_DGRAM, 0); if (probe_fd < 0) { perror("socket"); exit(2); } if (device.name) { enable_capability_raw(); if (setsockopt(probe_fd, SOL_SOCKET, SO_BINDTODEVICE, device.name, strlen(device.name)+1) == -1) perror("WARNING: interface is ignored"); disable_capability_raw(); } memset(&saddr, 0, sizeof(saddr)); saddr.sin_family = AF_INET; if (src.s_addr) { saddr.sin_addr = src; if (bind(probe_fd, (struct sockaddr*)&saddr, sizeof(saddr)) == -1) { perror("bind"); exit(2); } } else if (!dad) { int on = 1; socklen_t alen = sizeof(saddr); saddr.sin_port = htons(1025); saddr.sin_addr = dst; if (setsockopt(probe_fd, SOL_SOCKET, SO_DONTROUTE, (char*)&on, sizeof(on)) == -1) perror("WARNING: setsockopt(SO_DONTROUTE)"); if (connect(probe_fd, (struct sockaddr*)&saddr, sizeof(saddr)) == -1) { perror("connect"); exit(2); } if (getsockname(probe_fd, (struct sockaddr*)&saddr, &alen) == -1) { perror("getsockname"); exit(2); } src = saddr.sin_addr; } close(probe_fd); }; ((struct sockaddr_ll *)&me)->sll_family = AF_PACKET; ((struct sockaddr_ll *)&me)->sll_ifindex = device.ifindex; ((struct sockaddr_ll *)&me)->sll_protocol = htons(ETH_P_ARP); if (bind(s, (struct sockaddr*)&me, sizeof(me)) == -1) { perror("bind"); exit(2); } if (1) { socklen_t alen = sizeof(me); if (getsockname(s, (struct sockaddr*)&me, &alen) == -1) { perror("getsockname"); exit(2); } } if (((struct sockaddr_ll *)&me)->sll_halen == 0) { if (!quiet) printf("Interface \"%s\" is not ARPable (no ll address)\n", device.name); exit(dad?0:2); } he = me; set_device_broadcast(&device, ((struct sockaddr_ll *)&he)->sll_addr, ((struct sockaddr_ll *)&he)->sll_halen); if (!quiet) { printf("ARPING %s ", inet_ntoa(dst)); printf("from %s %s\n", inet_ntoa(src), device.name ? : ""); } if (!src.s_addr && !dad) { fprintf(stderr, "arping: no source address in not-DAD mode\n"); exit(2); } drop_capabilities(); set_signal(SIGINT, finish); set_signal(SIGALRM, catcher); catcher(); while(1) { sigset_t sset, osset; unsigned char packet[4096]; struct sockaddr_storage from; socklen_t alen = sizeof(from); int cc; if ((cc = recvfrom(s, packet, sizeof(packet), 0, (struct sockaddr *)&from, &alen)) < 0) { perror("arping: recvfrom"); continue; } sigemptyset(&sset); sigaddset(&sset, SIGALRM); sigaddset(&sset, SIGINT); sigprocmask(SIG_BLOCK, &sset, &osset); recv_pack(packet, cc, (struct sockaddr_ll *)&from); sigprocmask(SIG_SETMASK, &osset, NULL); } }