[PATCH net-next v3] selftests/net: synchronize udpgro tests' tx and rx connection

[PATCH net-next v3] selftests/net: synchronize udpgro tests' tx and rx connection

[Lucas Karpinski]

The sockets used by udpgso_bench_tx aren't always ready when
udpgso_bench_tx transmits packets. This issue is more prevalent in -rt
kernels, but can occur in both. Replace the hacky sleep calls with a
function that checks whether the ports in the namespace are ready for
use.

Suggested-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Lucas Karpinski <lkarpins@redhat.com>
---
https://lore.kernel.org/all/6ceki76bcv7qz6de5rxc26ot6aezdmeoz2g4ubtve7qwozmyyw@zibbg64wsdjp/
v3:
- moved redundant code to net_helper.sh
- replaced awk with grep
- changed commit message to specify udpgro
- added target tree

v2:
- applied synchronization method suggested by Paolo
- changed commit message to code

tools/testing/selftests/net/net_helper.sh     | 22 +++++++++++++++++++
 tools/testing/selftests/net/udpgro.sh         | 13 +++++------
 tools/testing/selftests/net/udpgro_bench.sh   |  5 +++--
 tools/testing/selftests/net/udpgro_frglist.sh |  5 +++--
 4 files changed, 34 insertions(+), 11 deletions(-)
 create mode 100755 tools/testing/selftests/net/net_helper.sh

diff --git a/tools/testing/selftests/net/net_helper.sh b/tools/testing/selftests/net/net_helper.sh
new file mode 100755
index 000000000000..4fe0befa13fb
--- /dev/null
+++ b/tools/testing/selftests/net/net_helper.sh
@@ -0,0 +1,22 @@
+#!/bin/bash
+# SPDX-License-Identifier: GPL-2.0
+#
+# Helper functions
+
+wait_local_port_listen()
+{
+	local listener_ns="${1}"
+	local port="${2}"
+	local protocol="${3}"
+	local port_hex
+	local i
+
+	port_hex="$(printf "%04X" "${port}")"
+	for i in $(seq 10); do
+		if ip netns exec "${listener_ns}" cat /proc/net/"${protocol}"* | \
+		   grep -q "${port_hex}"; then
+			break
+		fi
+		sleep 0.1
+	done
+}
diff --git a/tools/testing/selftests/net/udpgro.sh b/tools/testing/selftests/net/udpgro.sh
index 0c743752669a..af5dc57c8ce9 100755
--- a/tools/testing/selftests/net/udpgro.sh
+++ b/tools/testing/selftests/net/udpgro.sh
@@ -3,6 +3,8 @@
 #
 # Run a series of udpgro functional tests.
 
+source net_helper.sh
+
 readonly PEER_NS="ns-peer-$(mktemp -u XXXXXX)"
 
 BPF_FILE="../bpf/xdp_dummy.bpf.o"
@@ -51,8 +53,7 @@ run_one() {
 		echo "ok" || \
 		echo "failed" &
 
-	# Hack: let bg programs complete the startup
-	sleep 0.2
+	wait_local_port_listen ${PEER_NS} 8000 udp
 	./udpgso_bench_tx ${tx_args}
 	ret=$?
 	wait $(jobs -p)
@@ -97,7 +98,7 @@ run_one_nat() {
 		echo "ok" || \
 		echo "failed"&
 
-	sleep 0.1
+	wait_local_port_listen "${PEER_NS}" 8000 udp
 	./udpgso_bench_tx ${tx_args}
 	ret=$?
 	kill -INT $pid
@@ -118,11 +119,9 @@ run_one_2sock() {
 		echo "ok" || \
 		echo "failed" &
 
-	# Hack: let bg programs complete the startup
-	sleep 0.2
+	wait_local_port_listen "${PEER_NS}" 12345 udp
 	./udpgso_bench_tx ${tx_args} -p 12345
-	sleep 0.1
-	# first UDP GSO socket should be closed at this point
+	wait_local_port_listen "${PEER_NS}" 8000 udp
 	./udpgso_bench_tx ${tx_args}
 	ret=$?
 	wait $(jobs -p)
diff --git a/tools/testing/selftests/net/udpgro_bench.sh b/tools/testing/selftests/net/udpgro_bench.sh
index 894972877e8b..cb664679b434 100755
--- a/tools/testing/selftests/net/udpgro_bench.sh
+++ b/tools/testing/selftests/net/udpgro_bench.sh
@@ -3,6 +3,8 @@
 #
 # Run a series of udpgro benchmarks
 
+source net_helper.sh
+
 readonly PEER_NS="ns-peer-$(mktemp -u XXXXXX)"
 
 BPF_FILE="../bpf/xdp_dummy.bpf.o"
@@ -40,8 +42,7 @@ run_one() {
 	ip netns exec "${PEER_NS}" ./udpgso_bench_rx ${rx_args} -r &
 	ip netns exec "${PEER_NS}" ./udpgso_bench_rx -t ${rx_args} -r &
 
-	# Hack: let bg programs complete the startup
-	sleep 0.2
+	wait_local_port_listen "${PEER_NS}" 8000 udp
 	./udpgso_bench_tx ${tx_args}
 }
 
diff --git a/tools/testing/selftests/net/udpgro_frglist.sh b/tools/testing/selftests/net/udpgro_frglist.sh
index 0a6359bed0b9..dd47fa96f6b3 100755
--- a/tools/testing/selftests/net/udpgro_frglist.sh
+++ b/tools/testing/selftests/net/udpgro_frglist.sh
@@ -3,6 +3,8 @@
 #
 # Run a series of udpgro benchmarks
 
+source net_helper.sh
+
 readonly PEER_NS="ns-peer-$(mktemp -u XXXXXX)"
 
 BPF_FILE="../bpf/xdp_dummy.bpf.o"
@@ -45,8 +47,7 @@ run_one() {
         echo ${rx_args}
 	ip netns exec "${PEER_NS}" ./udpgso_bench_rx ${rx_args} -r &
 
-	# Hack: let bg programs complete the startup
-	sleep 0.2
+	wait_local_port_listen "${PEER_NS}" 8000 udp
 	./udpgso_bench_tx ${tx_args}
 }
 
-- 
2.41.0

Re: [PATCH net-next v3] selftests/net: synchronize udpgro tests' tx and rx connection

[Willem de Bruijn]

Lucas Karpinski wrote:
> The sockets used by udpgso_bench_tx aren't always ready when
> udpgso_bench_tx transmits packets. This issue is more prevalent in -rt
> kernels, but can occur in both. Replace the hacky sleep calls with a
> function that checks whether the ports in the namespace are ready for
> use.
> 
> Suggested-by: Paolo Abeni <pabeni@redhat.com>
> Signed-off-by: Lucas Karpinski <lkarpins@redhat.com>

Reviewed-by: Willem de Bruijn <willemb@google.com>

Re: [PATCH net-next v3] selftests/net: synchronize udpgro tests' tx and rx connection

[patchwork-bot+netdevbpf]

Hello:

This patch was applied to netdev/net-next.git (main)
by David S. Miller <davem@davemloft.net>:

On Tue, 14 Nov 2023 10:11:31 -0500 you wrote:
> The sockets used by udpgso_bench_tx aren't always ready when
> udpgso_bench_tx transmits packets. This issue is more prevalent in -rt
> kernels, but can occur in both. Replace the hacky sleep calls with a
> function that checks whether the ports in the namespace are ready for
> use.
> 
> Suggested-by: Paolo Abeni <pabeni@redhat.com>
> Signed-off-by: Lucas Karpinski <lkarpins@redhat.com>
> 
> [...]

Here is the summary with links:
  - [net-next,v3] selftests/net: synchronize udpgro tests' tx and rx connection
    https://git.kernel.org/netdev/net-next/c/3bdd9fd29cb0

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html