Merge branch 'nexthop-API-sysctl'

Roopa Prabhu says:

====================
New sysctl to turn off nexthop API compat mode

Currently route nexthop API maintains user space compatibility
with old route API by default. Dumps and netlink notifications
support both new and old API format. In systems which have
moved to the new API, this compatibility mode cancels some
of the performance benefits provided by the new nexthop API.

This patch adds new sysctl nexthop_compat_mode which is on
by default but provides the ability to turn off compatibility
mode allowing systems to run entirely with the new routing
API if they wish to. Old route API behaviour and support is
not modified by this sysctl

v4:
	- Use davids note for Documenting the sysctl
	- test with latest iproute2 and adjust 'pref'

v3:
	- Document new sysctl
	- move sysctl to use proc_dointvec_minmax with 0 and 1 values
	- selftest: remove pref medium in ipv6 test

v2:
       - Incorporate David Aherns pointers on covering dumps and
         nexthop deletes. Also use one ipv4 sysctl to cover
         both ipv4 and ipv6 (I see it is done that way for many
         others)
       - Added a selftest to cover dump and notfications for nexthop
	 api compat mode
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 196 insertions, 2 deletions

diff --git a/tools/testing/selftests/net/fib_nexthops.sh b/tools/testing/selftests/net/fib_nexthops.sh
index b785241127df..dd0e5fec6367 100755
--- a/tools/testing/selftests/net/fib_nexthops.sh
+++ b/tools/testing/selftests/net/fib_nexthops.sh
@@ -19,8 +19,8 @@ ret=0
 ksft_skip=4
 
 # all tests in this script. Can be overridden with -t option
-IPV4_TESTS="ipv4_fcnal ipv4_grp_fcnal ipv4_withv6_fcnal ipv4_fcnal_runtime"
-IPV6_TESTS="ipv6_fcnal ipv6_grp_fcnal ipv6_fcnal_runtime"
+IPV4_TESTS="ipv4_fcnal ipv4_grp_fcnal ipv4_withv6_fcnal ipv4_fcnal_runtime ipv4_compat_mode"
+IPV6_TESTS="ipv6_fcnal ipv6_grp_fcnal ipv6_fcnal_runtime ipv6_compat_mode"
 
 ALL_TESTS="basic ${IPV4_TESTS} ${IPV6_TESTS}"
 TESTS="${ALL_TESTS}"
@@ -253,6 +253,33 @@ check_route6()
 	check_output "${out}" "${expected}"
 }
 
+start_ip_monitor()
+{
+	local mtype=$1
+
+	# start the monitor in the background
+	tmpfile=`mktemp /var/run/nexthoptestXXX`
+	mpid=`($IP monitor $mtype > $tmpfile & echo $!) 2>/dev/null`
+	sleep 0.2
+	echo "$mpid $tmpfile"
+}
+
+stop_ip_monitor()
+{
+	local mpid=$1
+	local tmpfile=$2
+	local el=$3
+
+	# check the monitor results
+	kill $mpid
+	lines=`wc -l $tmpfile | cut "-d " -f1`
+	test $lines -eq $el
+	rc=$?
+	rm -rf $tmpfile
+
+	return $rc
+}
+
 ################################################################################
 # basic operations (add, delete, replace) on nexthops and nexthop groups
 #
@@ -883,6 +910,173 @@ ipv4_fcnal_runtime()
 	log_test $? 0 "IPv4 route with MPLS encap, v6 gw - check"
 }
 
+sysctl_nexthop_compat_mode_check()
+{
+	local sysctlname="net.ipv4.nexthop_compat_mode"
+	local lprefix=$1
+
+	IPE="ip netns exec me"
+
+	$IPE sysctl -q $sysctlname 2>&1 >/dev/null
+	if [ $? -ne 0 ]; then
+		echo "SKIP: kernel lacks nexthop compat mode sysctl control"
+		return $ksft_skip
+	fi
+
+	out=$($IPE sysctl $sysctlname 2>/dev/null)
+	log_test $? 0 "$lprefix default nexthop compat mode check"
+	check_output "${out}" "$sysctlname = 1"
+}
+
+sysctl_nexthop_compat_mode_set()
+{
+	local sysctlname="net.ipv4.nexthop_compat_mode"
+	local mode=$1
+	local lprefix=$2
+
+	IPE="ip netns exec me"
+
+	out=$($IPE sysctl -w $sysctlname=$mode)
+	log_test $? 0 "$lprefix set compat mode - $mode"
+	check_output "${out}" "net.ipv4.nexthop_compat_mode = $mode"
+}
+
+ipv6_compat_mode()
+{
+	local rc
+
+	echo
+	echo "IPv6 nexthop api compat mode test"
+	echo "--------------------------------"
+
+	sysctl_nexthop_compat_mode_check "IPv6"
+	if [ $? -eq $ksft_skip ]; then
+		return $ksft_skip
+	fi
+
+	run_cmd "$IP nexthop add id 62 via 2001:db8:91::2 dev veth1"
+	run_cmd "$IP nexthop add id 63 via 2001:db8:91::3 dev veth1"
+	run_cmd "$IP nexthop add id 122 group 62/63"
+	ipmout=$(start_ip_monitor route)
+
+	run_cmd "$IP -6 ro add 2001:db8:101::1/128 nhid 122"
+	# route add notification should contain expanded nexthops
+	stop_ip_monitor $ipmout 3
+	log_test $? 0 "IPv6 compat mode on - route add notification"
+
+	# route dump should contain expanded nexthops
+	check_route6 "2001:db8:101::1" "2001:db8:101::1 nhid 122 metric 1024 pref medium nexthop via 2001:db8:91::2 dev veth1 weight 1 nexthop via 2001:db8:91::3 dev veth1 weight 1"
+	log_test $? 0 "IPv6 compat mode on - route dump"
+
+	# change in nexthop group should generate route notification
+	run_cmd "$IP nexthop add id 64 via 2001:db8:91::4 dev veth1"
+	ipmout=$(start_ip_monitor route)
+	run_cmd "$IP nexthop replace id 122 group 62/64"
+	stop_ip_monitor $ipmout 3
+
+	log_test $? 0 "IPv6 compat mode on - nexthop change"
+
+	# set compat mode off
+	sysctl_nexthop_compat_mode_set 0 "IPv6"
+
+	run_cmd "$IP -6 ro del 2001:db8:101::1/128 nhid 122"
+
+	run_cmd "$IP nexthop add id 62 via 2001:db8:91::2 dev veth1"
+	run_cmd "$IP nexthop add id 63 via 2001:db8:91::3 dev veth1"
+	run_cmd "$IP nexthop add id 122 group 62/63"
+	ipmout=$(start_ip_monitor route)
+
+	run_cmd "$IP -6 ro add 2001:db8:101::1/128 nhid 122"
+	# route add notification should not contain expanded nexthops
+	stop_ip_monitor $ipmout 1
+	log_test $? 0 "IPv6 compat mode off - route add notification"
+
+	# route dump should not contain expanded nexthops
+	check_route6 "2001:db8:101::1" "2001:db8:101::1 nhid 122 metric 1024 pref medium"
+	log_test $? 0 "IPv6 compat mode off - route dump"
+
+	# change in nexthop group should not generate route notification
+	run_cmd "$IP nexthop add id 64 via 2001:db8:91::4 dev veth1"
+	ipmout=$(start_ip_monitor route)
+	run_cmd "$IP nexthop replace id 122 group 62/64"
+	stop_ip_monitor $ipmout 0
+	log_test $? 0 "IPv6 compat mode off - nexthop change"
+
+	# nexthop delete should not generate route notification
+	ipmout=$(start_ip_monitor route)
+	run_cmd "$IP nexthop del id 122"
+	stop_ip_monitor $ipmout 0
+	log_test $? 0 "IPv6 compat mode off - nexthop delete"
+
+	# set compat mode back on
+	sysctl_nexthop_compat_mode_set 1 "IPv6"
+}
+
+ipv4_compat_mode()
+{
+	local rc
+
+	echo
+	echo "IPv4 nexthop api compat mode"
+	echo "----------------------------"
+
+	sysctl_nexthop_compat_mode_check "IPv4"
+	if [ $? -eq $ksft_skip ]; then
+		return $ksft_skip
+	fi
+
+	run_cmd "$IP nexthop add id 21 via 172.16.1.2 dev veth1"
+	run_cmd "$IP nexthop add id 22 via 172.16.1.2 dev veth1"
+	run_cmd "$IP nexthop add id 122 group 21/22"
+	ipmout=$(start_ip_monitor route)
+
+	run_cmd "$IP ro add 172.16.101.1/32 nhid 122"
+	stop_ip_monitor $ipmout 3
+
+	# route add notification should contain expanded nexthops
+	log_test $? 0 "IPv4 compat mode on - route add notification"
+
+	# route dump should contain expanded nexthops
+	check_route "172.16.101.1" "172.16.101.1 nhid 122 nexthop via 172.16.1.2 dev veth1 weight 1 nexthop via 172.16.1.2 dev veth1 weight 1"
+	log_test $? 0 "IPv4 compat mode on - route dump"
+
+	# change in nexthop group should generate route notification
+	run_cmd "$IP nexthop add id 23 via 172.16.1.3 dev veth1"
+	ipmout=$(start_ip_monitor route)
+	run_cmd "$IP nexthop replace id 122 group 21/23"
+	stop_ip_monitor $ipmout 3
+	log_test $? 0 "IPv4 compat mode on - nexthop change"
+
+	sysctl_nexthop_compat_mode_set 0 "IPv4"
+
+	# cleanup
+	run_cmd "$IP ro del 172.16.101.1/32 nhid 122"
+
+	ipmout=$(start_ip_monitor route)
+	run_cmd "$IP ro add 172.16.101.1/32 nhid 122"
+	stop_ip_monitor $ipmout 1
+	# route add notification should not contain expanded nexthops
+	log_test $? 0 "IPv4 compat mode off - route add notification"
+
+	# route dump should not contain expanded nexthops
+	check_route "172.16.101.1" "172.16.101.1 nhid 122"
+	log_test $? 0 "IPv4 compat mode off - route dump"
+
+	# change in nexthop group should not generate route notification
+	ipmout=$(start_ip_monitor route)
+	run_cmd "$IP nexthop replace id 122 group 21/22"
+	stop_ip_monitor $ipmout 0
+	log_test $? 0 "IPv4 compat mode off - nexthop change"
+
+	# nexthop delete should not generate route notification
+	ipmout=$(start_ip_monitor route)
+	run_cmd "$IP nexthop del id 122"
+	stop_ip_monitor $ipmout 0
+	log_test $? 0 "IPv4 compat mode off - nexthop delete"
+
+	sysctl_nexthop_compat_mode_set 1 "IPv4"
+}
+
 basic()
 {
 	echo