From 3de611b507628abac081630d8180a9a369c83668 Mon Sep 17 00:00:00 2001 From: Petr Machata Date: Thu, 27 Feb 2020 08:50:07 +0100 Subject: [PATCH] selftests: mlxsw: Add a RED selftest This tests that below the queue minimum length, there is no dropping / marking, and above max, everything is dropped / marked. The test is structured as a core file with topology and test code, and three wrappers: one for RED used as a root Qdisc, and two for testing (W)RED under PRIO and ETS. Signed-off-by: Petr Machata Signed-off-by: Ido Schimmel Signed-off-by: Jiri Pirko Signed-off-by: David S. Miller --- .../drivers/net/mlxsw/sch_red_core.sh | 499 ++++++++++++++++++ .../drivers/net/mlxsw/sch_red_ets.sh | 83 +++ .../drivers/net/mlxsw/sch_red_prio.sh | 5 + .../drivers/net/mlxsw/sch_red_root.sh | 60 +++ tools/testing/selftests/net/forwarding/lib.sh | 10 + 5 files changed, 657 insertions(+) create mode 100644 tools/testing/selftests/drivers/net/mlxsw/sch_red_core.sh create mode 100755 tools/testing/selftests/drivers/net/mlxsw/sch_red_ets.sh create mode 100755 tools/testing/selftests/drivers/net/mlxsw/sch_red_prio.sh create mode 100755 tools/testing/selftests/drivers/net/mlxsw/sch_red_root.sh diff --git a/tools/testing/selftests/drivers/net/mlxsw/sch_red_core.sh b/tools/testing/selftests/drivers/net/mlxsw/sch_red_core.sh new file mode 100644 index 000000000000..ebf7752f6d93 --- /dev/null +++ b/tools/testing/selftests/drivers/net/mlxsw/sch_red_core.sh @@ -0,0 +1,499 @@ +# SPDX-License-Identifier: GPL-2.0 + +# This test sends a >1Gbps stream of traffic from H1, to the switch, which +# forwards it to a 1Gbps port. This 1Gbps stream is then looped back to the +# switch and forwarded to the port under test $swp3, which is also 1Gbps. +# +# This way, $swp3 should be 100% filled with traffic without any of it spilling +# to the backlog. Any extra packets sent should almost 1:1 go to backlog. That +# is what H2 is used for--it sends the extra traffic to create backlog. +# +# A RED Qdisc is installed on $swp3. The configuration is such that the minimum +# and maximum size are 1 byte apart, so there is a very clear border under which +# no marking or dropping takes place, and above which everything is marked or +# dropped. +# +# The test uses the buffer build-up behavior to test the installed RED. +# +# In order to test WRED, $swp3 actually contains RED under PRIO, with two +# different configurations. Traffic is prioritized using 802.1p and relies on +# the implicit mlxsw configuration, where packet priority is taken 1:1 from the +# 802.1p marking. +# +# +--------------------------+ +--------------------------+ +# | H1 | | H2 | +# | + $h1.10 | | + $h2.10 | +# | | 192.0.2.1/28 | | | 192.0.2.2/28 | +# | | | | | | +# | | $h1.11 + | | | $h2.11 + | +# | | 192.0.2.17/28 | | | | 192.0.2.18/28 | | +# | | | | | | | | +# | \______ ______/ | | \______ ______/ | +# | \ / | | \ / | +# | + $h1 | | + $h2 | +# +-------------|------------+ +-------------|------------+ +# | >1Gbps | +# +-------------|------------------------------------------------|------------+ +# | SW + $swp1 + $swp2 | +# | _______/ \___________ ___________/ \_______ | +# | / \ / \ | +# | +-|-----------------+ | +-|-----------------+ | | +# | | + $swp1.10 | | | + $swp2.10 | | | +# | | | | .-------------+ $swp5.10 | | | +# | | BR1_10 | | | | | | | +# | | | | | | BR2_10 | | | +# | | + $swp2.10 | | | | | | | +# | +-|-----------------+ | | | + $swp3.10 | | | +# | | | | +-|-----------------+ | | +# | | +-----------------|-+ | | +-----------------|-+ | +# | | | $swp1.11 + | | | | $swp2.11 + | | +# | | | | | .-----------------+ $swp5.11 | | +# | | | BR1_11 | | | | | | | +# | | | | | | | | BR2_11 | | +# | | | $swp2.11 + | | | | | | | +# | | +-----------------|-+ | | | | $swp3.11 + | | +# | | | | | | +-----------------|-+ | +# | \_______ ___________/ | | \___________ _______/ | +# | \ / \ / \ / | +# | + $swp4 + $swp5 + $swp3 | +# +-------------|----------------------|-------------------------|------------+ +# | | | 1Gbps +# \________1Gbps_________/ | +# +----------------------------|------------+ +# | H3 + $h3 | +# | _____________________/ \_______ | +# | / \ | +# | | | | +# | + $h3.10 $h3.11 + | +# | 192.0.2.3/28 192.0.2.19/28 | +# +-----------------------------------------+ + +NUM_NETIFS=8 +CHECK_TC="yes" +lib_dir=$(dirname $0)/../../../net/forwarding +source $lib_dir/lib.sh +source $lib_dir/devlink_lib.sh +source qos_lib.sh + +ipaddr() +{ + local host=$1; shift + local vlan=$1; shift + + echo 192.0.2.$((16 * (vlan - 10) + host)) +} + +host_create() +{ + local dev=$1; shift + local host=$1; shift + + simple_if_init $dev + mtu_set $dev 10000 + + vlan_create $dev 10 v$dev $(ipaddr $host 10)/28 + ip link set dev $dev.10 type vlan egress 0:0 + + vlan_create $dev 11 v$dev $(ipaddr $host 11)/28 + ip link set dev $dev.11 type vlan egress 0:1 +} + +host_destroy() +{ + local dev=$1; shift + + vlan_destroy $dev 11 + vlan_destroy $dev 10 + mtu_restore $dev + simple_if_fini $dev +} + +h1_create() +{ + host_create $h1 1 +} + +h1_destroy() +{ + host_destroy $h1 +} + +h2_create() +{ + host_create $h2 2 + + # Some of the tests in this suite use multicast traffic. As this traffic + # enters BR2_10 resp. BR2_11, it is flooded to all other ports. Thus + # e.g. traffic ingressing through $swp2 is flooded to $swp3 (the + # intended destination) and $swp5 (which is intended as ingress for + # another stream of traffic). + # + # This is generally not a problem, but if the $swp5 throughput is lower + # than $swp2 throughput, there will be a build-up at $swp5. That may + # cause packets to fail to queue up at $swp3 due to shared buffer + # quotas, and the test to spuriously fail. + # + # Prevent this by setting the speed of $h2 to 1Gbps. + + ethtool -s $h2 speed 1000 autoneg off +} + +h2_destroy() +{ + ethtool -s $h2 autoneg on + host_destroy $h2 +} + +h3_create() +{ + host_create $h3 3 + ethtool -s $h3 speed 1000 autoneg off +} + +h3_destroy() +{ + ethtool -s $h3 autoneg on + host_destroy $h3 +} + +switch_create() +{ + local intf + local vlan + + ip link add dev br1_10 type bridge + ip link add dev br1_11 type bridge + + ip link add dev br2_10 type bridge + ip link add dev br2_11 type bridge + + for intf in $swp1 $swp2 $swp3 $swp4 $swp5; do + ip link set dev $intf up + mtu_set $intf 10000 + done + + for intf in $swp1 $swp4; do + for vlan in 10 11; do + vlan_create $intf $vlan + ip link set dev $intf.$vlan master br1_$vlan + ip link set dev $intf.$vlan up + done + done + + for intf in $swp2 $swp3 $swp5; do + for vlan in 10 11; do + vlan_create $intf $vlan + ip link set dev $intf.$vlan master br2_$vlan + ip link set dev $intf.$vlan up + done + done + + ip link set dev $swp4.10 type vlan egress 0:0 + ip link set dev $swp4.11 type vlan egress 0:1 + for intf in $swp1 $swp2 $swp5; do + for vlan in 10 11; do + ip link set dev $intf.$vlan type vlan ingress 0:0 1:1 + done + done + + for intf in $swp2 $swp3 $swp4 $swp5; do + ethtool -s $intf speed 1000 autoneg off + done + + ip link set dev br1_10 up + ip link set dev br1_11 up + ip link set dev br2_10 up + ip link set dev br2_11 up + + local size=$(devlink_pool_size_thtype 0 | cut -d' ' -f 1) + devlink_port_pool_th_set $swp3 8 $size +} + +switch_destroy() +{ + local intf + local vlan + + devlink_port_pool_th_restore $swp3 8 + + tc qdisc del dev $swp3 root 2>/dev/null + + ip link set dev br2_11 down + ip link set dev br2_10 down + ip link set dev br1_11 down + ip link set dev br1_10 down + + for intf in $swp5 $swp4 $swp3 $swp2; do + ethtool -s $intf autoneg on + done + + for intf in $swp5 $swp3 $swp2 $swp4 $swp1; do + for vlan in 11 10; do + ip link set dev $intf.$vlan down + ip link set dev $intf.$vlan nomaster + vlan_destroy $intf $vlan + done + + mtu_restore $intf + ip link set dev $intf down + done + + ip link del dev br2_11 + ip link del dev br2_10 + ip link del dev br1_11 + ip link del dev br1_10 +} + +setup_prepare() +{ + h1=${NETIFS[p1]} + swp1=${NETIFS[p2]} + + swp2=${NETIFS[p3]} + h2=${NETIFS[p4]} + + swp3=${NETIFS[p5]} + h3=${NETIFS[p6]} + + swp4=${NETIFS[p7]} + swp5=${NETIFS[p8]} + + h3_mac=$(mac_get $h3) + + vrf_prepare + + h1_create + h2_create + h3_create + switch_create +} + +cleanup() +{ + pre_cleanup + + switch_destroy + h3_destroy + h2_destroy + h1_destroy + + vrf_cleanup +} + +ping_ipv4() +{ + ping_test $h1.10 $(ipaddr 3 10) " from host 1, vlan 10" + ping_test $h1.11 $(ipaddr 3 11) " from host 1, vlan 11" + ping_test $h2.10 $(ipaddr 3 10) " from host 2, vlan 10" + ping_test $h2.11 $(ipaddr 3 11) " from host 2, vlan 11" +} + +get_tc() +{ + local vlan=$1; shift + + echo $((vlan - 10)) +} + +get_qdisc_handle() +{ + local vlan=$1; shift + + local tc=$(get_tc $vlan) + local band=$((8 - tc)) + + # Handle is 107: for TC1, 108: for TC0. + echo "10$band:" +} + +get_qdisc_backlog() +{ + local vlan=$1; shift + + qdisc_stats_get $swp3 $(get_qdisc_handle $vlan) .backlog +} + +get_mc_transmit_queue() +{ + local vlan=$1; shift + + local tc=$(($(get_tc $vlan) + 8)) + ethtool_stats_get $swp3 tc_transmit_queue_tc_$tc +} + +get_nmarked() +{ + local vlan=$1; shift + + ethtool_stats_get $swp3 ecn_marked +} + +get_qdisc_npackets() +{ + local vlan=$1; shift + + busywait_for_counter 1100 +1 \ + qdisc_stats_get $swp3 $(get_qdisc_handle $vlan) .packets +} + +# This sends traffic in an attempt to build a backlog of $size. Returns 0 on +# success. After 10 failed attempts it bails out and returns 1. It dumps the +# backlog size to stdout. +build_backlog() +{ + local vlan=$1; shift + local size=$1; shift + local proto=$1; shift + + local tc=$((vlan - 10)) + local band=$((8 - tc)) + local cur=-1 + local i=0 + + while :; do + local cur=$(busywait 1100 until_counter_is $((cur + 1)) \ + get_qdisc_backlog $vlan) + local diff=$((size - cur)) + local pkts=$(((diff + 7999) / 8000)) + + if ((cur >= size)); then + echo $cur + return 0 + elif ((i++ > 10)); then + echo $cur + return 1 + fi + + $MZ $h2.$vlan -p 8000 -a own -b $h3_mac \ + -A $(ipaddr 2 $vlan) -B $(ipaddr 3 $vlan) \ + -t $proto -q -c $pkts "$@" + done +} + +check_marking() +{ + local vlan=$1; shift + local cond=$1; shift + + local npackets_0=$(get_qdisc_npackets $vlan) + local nmarked_0=$(get_nmarked $vlan) + sleep 5 + local npackets_1=$(get_qdisc_npackets $vlan) + local nmarked_1=$(get_nmarked $vlan) + + local nmarked_d=$((nmarked_1 - nmarked_0)) + local npackets_d=$((npackets_1 - npackets_0)) + local pct=$((100 * nmarked_d / npackets_d)) + + echo $pct + ((pct $cond)) +} + +do_ecn_test() +{ + local vlan=$1; shift + local limit=$1; shift + local backlog + local pct + + # Main stream. + start_tcp_traffic $h1.$vlan $(ipaddr 1 $vlan) $(ipaddr 3 $vlan) \ + $h3_mac tos=0x01 + + # Build the below-the-limit backlog using UDP. We could use TCP just + # fine, but this way we get a proof that UDP is accepted when queue + # length is below the limit. The main stream is using TCP, and if the + # limit is misconfigured, we would see this traffic being ECN marked. + RET=0 + backlog=$(build_backlog $vlan $((2 * limit / 3)) udp) + check_err $? "Could not build the requested backlog" + pct=$(check_marking $vlan "== 0") + check_err $? "backlog $backlog / $limit Got $pct% marked packets, expected == 0." + log_test "TC $((vlan - 10)): ECN backlog < limit" + + # Now push TCP, because non-TCP traffic would be early-dropped after the + # backlog crosses the limit, and we want to make sure that the backlog + # is above the limit. + RET=0 + backlog=$(build_backlog $vlan $((3 * limit / 2)) tcp tos=0x01) + check_err $? "Could not build the requested backlog" + pct=$(check_marking $vlan ">= 95") + check_err $? "backlog $backlog / $limit Got $pct% marked packets, expected >= 95." + log_test "TC $((vlan - 10)): ECN backlog > limit" + + # Up there we saw that UDP gets accepted when backlog is below the + # limit. Now that it is above, it should all get dropped, and backlog + # building should fail. + RET=0 + build_backlog $vlan $((2 * limit)) udp >/dev/null + check_fail $? "UDP traffic went into backlog instead of being early-dropped" + log_test "TC $((vlan - 10)): ECN backlog > limit: UDP early-dropped" + + stop_traffic + sleep 1 +} + +do_red_test() +{ + local vlan=$1; shift + local limit=$1; shift + local backlog + local pct + + # Use ECN-capable TCP to verify there's no marking even though the queue + # is above limit. + start_tcp_traffic $h1.$vlan $(ipaddr 1 $vlan) $(ipaddr 3 $vlan) \ + $h3_mac tos=0x01 + + # Pushing below the queue limit should work. + RET=0 + backlog=$(build_backlog $vlan $((2 * limit / 3)) tcp tos=0x01) + check_err $? "Could not build the requested backlog" + pct=$(check_marking $vlan "== 0") + check_err $? "backlog $backlog / $limit Got $pct% marked packets, expected == 0." + log_test "TC $((vlan - 10)): RED backlog < limit" + + # Pushing above should not. + RET=0 + backlog=$(build_backlog $vlan $((3 * limit / 2)) tcp tos=0x01) + check_fail $? "Traffic went into backlog instead of being early-dropped" + pct=$(check_marking $vlan "== 0") + check_err $? "backlog $backlog / $limit Got $pct% marked packets, expected == 0." + local diff=$((limit - backlog)) + pct=$((100 * diff / limit)) + ((0 <= pct && pct <= 5)) + check_err $? "backlog $backlog / $limit expected <= 5% distance" + log_test "TC $((vlan - 10)): RED backlog > limit" + + stop_traffic + sleep 1 +} + +do_mc_backlog_test() +{ + local vlan=$1; shift + local limit=$1; shift + local backlog + local pct + + RET=0 + + start_tcp_traffic $h1.$vlan $(ipaddr 1 $vlan) $(ipaddr 3 $vlan) bc + start_tcp_traffic $h2.$vlan $(ipaddr 2 $vlan) $(ipaddr 3 $vlan) bc + + qbl=$(busywait 5000 until_counter_is 500000 \ + get_qdisc_backlog $vlan) + check_err $? "Could not build MC backlog" + + # Verify that we actually see the backlog on BUM TC. Do a busywait as + # well, performance blips might cause false fail. + local ebl + ebl=$(busywait 5000 until_counter_is 500000 \ + get_mc_transmit_queue $vlan) + check_err $? "MC backlog reported by qdisc not visible in ethtool" + + stop_traffic + stop_traffic + + log_test "TC $((vlan - 10)): Qdisc reports MC backlog" +} diff --git a/tools/testing/selftests/drivers/net/mlxsw/sch_red_ets.sh b/tools/testing/selftests/drivers/net/mlxsw/sch_red_ets.sh new file mode 100755 index 000000000000..af83efe9ccf1 --- /dev/null +++ b/tools/testing/selftests/drivers/net/mlxsw/sch_red_ets.sh @@ -0,0 +1,83 @@ +#!/bin/bash +# SPDX-License-Identifier: GPL-2.0 + +ALL_TESTS=" + ping_ipv4 + ecn_test + red_test + mc_backlog_test +" +: ${QDISC:=ets} +source sch_red_core.sh + +# do_ecn_test first build 2/3 of the requested backlog and expects no marking, +# and then builds 3/2 of it and does expect marking. The values of $BACKLOG1 and +# $BACKLOG2 are far enough not to overlap, so that we can assume that if we do +# see (do not see) marking, it is actually due to the configuration of that one +# TC, and not due to configuration of the other TC leaking over. +BACKLOG1=200000 +BACKLOG2=500000 + +install_qdisc() +{ + local -a args=("$@") + + tc qdisc add dev $swp3 root handle 10: $QDISC \ + bands 8 priomap 7 6 5 4 3 2 1 0 + tc qdisc add dev $swp3 parent 10:8 handle 108: red \ + limit 1000000 min $BACKLOG1 max $((BACKLOG1 + 1)) \ + probability 1.0 avpkt 8000 burst 38 "${args[@]}" + tc qdisc add dev $swp3 parent 10:7 handle 107: red \ + limit 1000000 min $BACKLOG2 max $((BACKLOG2 + 1)) \ + probability 1.0 avpkt 8000 burst 63 "${args[@]}" + sleep 1 +} + +uninstall_qdisc() +{ + tc qdisc del dev $swp3 parent 10:7 + tc qdisc del dev $swp3 parent 10:8 + tc qdisc del dev $swp3 root +} + +ecn_test() +{ + install_qdisc ecn + + do_ecn_test 10 $BACKLOG1 + do_ecn_test 11 $BACKLOG2 + + uninstall_qdisc +} + +red_test() +{ + install_qdisc + + do_red_test 10 $BACKLOG1 + do_red_test 11 $BACKLOG2 + + uninstall_qdisc +} + +mc_backlog_test() +{ + install_qdisc + + # Note that the backlog numbers here do not correspond to RED + # configuration, but are arbitrary. + do_mc_backlog_test 10 $BACKLOG1 + do_mc_backlog_test 11 $BACKLOG2 + + uninstall_qdisc +} + +trap cleanup EXIT + +setup_prepare +setup_wait + +bail_on_lldpad +tests_run + +exit $EXIT_STATUS diff --git a/tools/testing/selftests/drivers/net/mlxsw/sch_red_prio.sh b/tools/testing/selftests/drivers/net/mlxsw/sch_red_prio.sh new file mode 100755 index 000000000000..76820a0e9a1b --- /dev/null +++ b/tools/testing/selftests/drivers/net/mlxsw/sch_red_prio.sh @@ -0,0 +1,5 @@ +#!/bin/bash +# SPDX-License-Identifier: GPL-2.0 + +QDISC=prio +source sch_red_ets.sh diff --git a/tools/testing/selftests/drivers/net/mlxsw/sch_red_root.sh b/tools/testing/selftests/drivers/net/mlxsw/sch_red_root.sh new file mode 100755 index 000000000000..b2217493a88e --- /dev/null +++ b/tools/testing/selftests/drivers/net/mlxsw/sch_red_root.sh @@ -0,0 +1,60 @@ +#!/bin/bash +# SPDX-License-Identifier: GPL-2.0 + +ALL_TESTS=" + ping_ipv4 + ecn_test + red_test + mc_backlog_test +" +source sch_red_core.sh + +BACKLOG=300000 + +install_qdisc() +{ + local -a args=("$@") + + tc qdisc add dev $swp3 root handle 108: red \ + limit 1000000 min $BACKLOG max $((BACKLOG + 1)) \ + probability 1.0 avpkt 8000 burst 38 "${args[@]}" + sleep 1 +} + +uninstall_qdisc() +{ + tc qdisc del dev $swp3 root +} + +ecn_test() +{ + install_qdisc ecn + do_ecn_test 10 $BACKLOG + uninstall_qdisc +} + +red_test() +{ + install_qdisc + do_red_test 10 $BACKLOG + uninstall_qdisc +} + +mc_backlog_test() +{ + install_qdisc + # Note that the backlog value here does not correspond to RED + # configuration, but is arbitrary. + do_mc_backlog_test 10 $BACKLOG + uninstall_qdisc +} + +trap cleanup EXIT + +setup_prepare +setup_wait + +bail_on_lldpad +tests_run + +exit $EXIT_STATUS diff --git a/tools/testing/selftests/net/forwarding/lib.sh b/tools/testing/selftests/net/forwarding/lib.sh index f80f384978ce..aff3178edf6d 100644 --- a/tools/testing/selftests/net/forwarding/lib.sh +++ b/tools/testing/selftests/net/forwarding/lib.sh @@ -607,6 +607,16 @@ ethtool_stats_get() ethtool -S $dev | grep "^ *$stat:" | head -n 1 | cut -d: -f2 } +qdisc_stats_get() +{ + local dev=$1; shift + local handle=$1; shift + local selector=$1; shift + + tc -j -s qdisc show dev "$dev" \ + | jq '.[] | select(.handle == "'"$handle"'") | '"$selector" +} + humanize() { local speed=$1; shift -- 2.34.1