[PATCH net-next 0/4] Improve tsn_lib selftests for future distributed tasks

[PATCH net-next 0/4] Improve tsn_lib selftests for future distributed tasks

[Vladimir Oltean]

Some of the boards I am working with are limited in the number of ports
that they offer, and as more TSN related selftests are added, it is
important to be able to distribute the work among multiple boards.
A large part of implementing that is ensuring network-wide
synchronization, but also permitting more streams of data to flow
through the network. There is the more important aspect of also
coordinating the timing characteristics of those streams, and that is
also something that is tackled, although not in this modest patch set.
The goal here is not to introduce new selftests yet, but just to lay a
better foundation for them. These patches are a part of the cleanup work
I've done while working on selftests for frame preemption. They are
regression-tested with psfp.sh.

Vladimir Oltean (4):
  selftests: net: tsn_lib: don't overwrite isochron receiver extra args
    with UDS
  selftests: net: tsn_lib: allow running ptp4l on multiple interfaces
  selftests: net: tsn_lib: allow multiple isochron receivers
  selftests: net: tsn_lib: run phc2sys in automatic mode

 .../selftests/drivers/net/ocelot/psfp.sh      |  2 +-
 .../selftests/net/forwarding/tsn_lib.sh       | 52 ++++++++++++-------
 2 files changed, 34 insertions(+), 20 deletions(-)

-- 
2.34.1

[PATCH net-next 1/4] selftests: net: tsn_lib: don't overwrite isochron receiver extra args with UDS

[Vladimir Oltean]

The extra_args argument ($3) of isochron_recv_start is overwritten with
uds ($2), if that argument exists.

This is currently not a problem, because the only TSN selftest
(ocelot/psfp.sh) omits remote sync so it does not specify to the
receiver a UNIX domain socket for ptp4l. So $uds is currently an empty
string.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
 tools/testing/selftests/net/forwarding/tsn_lib.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/testing/selftests/net/forwarding/tsn_lib.sh b/tools/testing/selftests/net/forwarding/tsn_lib.sh
index 60a1423e8116..1c8e36c56f32 100644
--- a/tools/testing/selftests/net/forwarding/tsn_lib.sh
+++ b/tools/testing/selftests/net/forwarding/tsn_lib.sh
@@ -139,7 +139,7 @@ isochron_recv_start()
 	local extra_args=$3
 
 	if ! [ -z "${uds}" ]; then
-		extra_args="--unix-domain-socket ${uds}"
+		extra_args="${extra_args} --unix-domain-socket ${uds}"
 	fi
 
 	isochron rcv \
-- 
2.34.1

[PATCH net-next 2/4] selftests: net: tsn_lib: allow running ptp4l on multiple interfaces

[Vladimir Oltean]

Switch ports will want to act as Boundary Clocks, which are configured
using ptp4l by specifying the "-i" argument multiple times.

Since we track a log file and a pid file for each ptp4l instance, and we
want to be compatible with the existing single-port callers of
ptp4l_start and ptp4l_stop, pass the interface list as a single string
of space-separated values. Based on this, we create a label for each
ptp4l instance, where the spaces are replaced with underscores
(ptp4l_start "eth0 eth1" generates "ptp4l_pid_eth0_eth1").

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
 .../selftests/net/forwarding/tsn_lib.sh       | 27 +++++++++++++------
 1 file changed, 19 insertions(+), 8 deletions(-)

diff --git a/tools/testing/selftests/net/forwarding/tsn_lib.sh b/tools/testing/selftests/net/forwarding/tsn_lib.sh
index 1c8e36c56f32..ace9c4f06805 100644
--- a/tools/testing/selftests/net/forwarding/tsn_lib.sh
+++ b/tools/testing/selftests/net/forwarding/tsn_lib.sh
@@ -53,15 +53,27 @@ phc2sys_stop()
 	rm "${phc2sys_log}" 2> /dev/null
 }
 
+# Replace space separators from interface list with underscores
+if_names_to_label()
+{
+	local if_name_list="$1"
+
+	echo "${if_name_list/ /_}"
+}
+
 ptp4l_start()
 {
-	local if_name=$1
+	local if_names="$1"
 	local slave_only=$2
 	local uds_address=$3
-	local log="ptp4l_log_${if_name}"
-	local pid="ptp4l_pid_${if_name}"
+	local log="ptp4l_log_$(if_names_to_label ${if_names})"
+	local pid="ptp4l_pid_$(if_names_to_label ${if_names})"
 	local extra_args=""
 
+	for if_name in ${if_names}; do
+		extra_args="${extra_args} -i ${if_name}"
+	done
+
 	if [ "${slave_only}" = true ]; then
 		extra_args="${extra_args} -s"
 	fi
@@ -71,7 +83,6 @@ ptp4l_start()
 	declare -g "${log}=$(mktemp)"
 
 	chrt -f 10 ptp4l -m -2 -P \
-		-i ${if_name} \
 		--step_threshold 0.00002 \
 		--first_step_threshold 0.00002 \
 		--tx_timestamp_timeout 100 \
@@ -80,16 +91,16 @@ ptp4l_start()
 		> "${!log}" 2>&1 &
 	declare -g "${pid}=$!"
 
-	echo "ptp4l for interface ${if_name} logs to ${!log} and has pid ${!pid}"
+	echo "ptp4l for interfaces ${if_names} logs to ${!log} and has pid ${!pid}"
 
 	sleep 1
 }
 
 ptp4l_stop()
 {
-	local if_name=$1
-	local log="ptp4l_log_${if_name}"
-	local pid="ptp4l_pid_${if_name}"
+	local if_names="$1"
+	local log="ptp4l_log_$(if_names_to_label ${if_names})"
+	local pid="ptp4l_pid_$(if_names_to_label ${if_names})"
 
 	{ kill ${!pid} && wait ${!pid}; } 2> /dev/null
 	rm "${!log}" 2> /dev/null
-- 
2.34.1

[PATCH net-next 3/4] selftests: net: tsn_lib: allow multiple isochron receivers

[Vladimir Oltean]

Move the PID variable for the isochron receiver into a separate
namespace per stats port, to allow multiple receivers (and/or
orchestration daemons) to be instantiated by the same script.

Preserve the existing behavior by making isochron_do() use the default
stats TCP port of 5000.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
 .../testing/selftests/net/forwarding/tsn_lib.sh  | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)

diff --git a/tools/testing/selftests/net/forwarding/tsn_lib.sh b/tools/testing/selftests/net/forwarding/tsn_lib.sh
index ace9c4f06805..20c2b411ba36 100644
--- a/tools/testing/selftests/net/forwarding/tsn_lib.sh
+++ b/tools/testing/selftests/net/forwarding/tsn_lib.sh
@@ -147,7 +147,9 @@ isochron_recv_start()
 {
 	local if_name=$1
 	local uds=$2
-	local extra_args=$3
+	local stats_port=$3
+	local extra_args=$4
+	local pid="isochron_pid_${stats_port}"
 
 	if ! [ -z "${uds}" ]; then
 		extra_args="${extra_args} --unix-domain-socket ${uds}"
@@ -158,16 +160,20 @@ isochron_recv_start()
 		--sched-priority 98 \
 		--sched-fifo \
 		--utc-tai-offset ${UTC_TAI_OFFSET} \
+		--stats-port ${stats_port} \
 		--quiet \
 		${extra_args} & \
-	isochron_pid=$!
+	declare -g "${pid}=$!"
 
 	sleep 1
 }
 
 isochron_recv_stop()
 {
-	{ kill ${isochron_pid} && wait ${isochron_pid}; } 2> /dev/null
+	local stats_port=$1
+	local pid="isochron_pid_${stats_port}"
+
+	{ kill ${!pid} && wait ${!pid}; } 2> /dev/null
 }
 
 isochron_do()
@@ -219,7 +225,7 @@ isochron_do()
 
 	cpufreq_max ${ISOCHRON_CPU}
 
-	isochron_recv_start "${h2}" "${receiver_uds}" "${receiver_extra_args}"
+	isochron_recv_start "${h2}" "${receiver_uds}" 5000 "${receiver_extra_args}"
 
 	isochron send \
 		--interface ${sender_if_name} \
@@ -240,7 +246,7 @@ isochron_do()
 		${extra_args} \
 		--quiet
 
-	isochron_recv_stop
+	isochron_recv_stop 5000
 
 	cpufreq_restore ${ISOCHRON_CPU}
 }
-- 
2.34.1

[PATCH net-next 4/4] selftests: net: tsn_lib: run phc2sys in automatic mode

[Vladimir Oltean]

We can make the phc2sys helper not only synchronize a PHC to
CLOCK_REALTIME, which is what it currently does, but also CLOCK_REALTIME
to a PHC, which is going to be needed in distributed TSN tests.

Instead of making the complexity of the arguments passed to
phc2sys_start() explode, we can let it figure out the sync direction
automatically, based on ptp4l's port states.

Towards that goal, pass just the path to the desired ptp4l instance's
UNIX domain socket, and remove the $if_name argument (from which it
derives the PHC). Also adapt the one caller from the ocelot psfp.sh
test. In the case of psfp.sh, phc2sys_start is able to properly figure
out that CLOCK_REALTIME is the source clock and swp1's PHC is the
destination, because of the way in which ptp4l_start for the
UDS_ADDRESS_SWP1 was called: with slave_only=false, so it will always
win the BMCA and always become the sync master between itself and $h1.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
 tools/testing/selftests/drivers/net/ocelot/psfp.sh | 2 +-
 tools/testing/selftests/net/forwarding/tsn_lib.sh  | 7 ++-----
 2 files changed, 3 insertions(+), 6 deletions(-)

diff --git a/tools/testing/selftests/drivers/net/ocelot/psfp.sh b/tools/testing/selftests/drivers/net/ocelot/psfp.sh
index 5a5cee92c665..bed748dde4b0 100755
--- a/tools/testing/selftests/drivers/net/ocelot/psfp.sh
+++ b/tools/testing/selftests/drivers/net/ocelot/psfp.sh
@@ -181,7 +181,7 @@ setup_prepare()
 
 	# Set up swp1 as a master PHC for h1, synchronized to the local
 	# CLOCK_REALTIME.
-	phc2sys_start ${swp1} ${UDS_ADDRESS_SWP1}
+	phc2sys_start ${UDS_ADDRESS_SWP1}
 
 	# Assumption true for LS1028A: h1 and h2 use the same PHC. So by
 	# synchronizing h1 to swp1 via PTP, h2 is also implicitly synchronized
diff --git a/tools/testing/selftests/net/forwarding/tsn_lib.sh b/tools/testing/selftests/net/forwarding/tsn_lib.sh
index 20c2b411ba36..b91bcd8008a9 100644
--- a/tools/testing/selftests/net/forwarding/tsn_lib.sh
+++ b/tools/testing/selftests/net/forwarding/tsn_lib.sh
@@ -22,8 +22,7 @@ fi
 
 phc2sys_start()
 {
-	local if_name=$1
-	local uds_address=$2
+	local uds_address=$1
 	local extra_args=""
 
 	if ! [ -z "${uds_address}" ]; then
@@ -33,9 +32,7 @@ phc2sys_start()
 	phc2sys_log="$(mktemp)"
 
 	chrt -f 10 phc2sys -m \
-		-c ${if_name} \
-		-s CLOCK_REALTIME \
-		-O ${UTC_TAI_OFFSET} \
+		-a -rr \
 		--step_threshold 0.00002 \
 		--first_step_threshold 0.00002 \
 		${extra_args} \
-- 
2.34.1

[PATCH net-next 4/4] selftests: net: tsn_lib: run phc2sys in automatic mode

[Vladimir Oltean]

We can make the phc2sys helper not only synchronize a PHC to
CLOCK_REALTIME, which is what it currently does, but also CLOCK_REALTIME
to a PHC, which is going to be needed in distributed TSN tests.

Instead of making the complexity of the arguments passed to
phc2sys_start() explode, we can let it figure out the sync direction
automatically, based on ptp4l's port states.

Towards that goal, pass just the path to the desired ptp4l instance's
UNIX domain socket, and remove the $if_name argument (from which it
derives the PHC). Also adapt the one caller from the ocelot psfp.sh
test. In the case of psfp.sh, phc2sys_start is able to properly figure
out that CLOCK_REALTIME is the source clock and swp1's PHC is the
destination, because of the way in which ptp4l_start for the
UDS_ADDRESS_SWP1 was called: with slave_only=false, so it will always
win the BMCA and always become the sync master between itself and $h1.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
 tools/testing/selftests/drivers/net/ocelot/psfp.sh | 2 +-
 tools/testing/selftests/net/forwarding/tsn_lib.sh  | 7 ++-----
 2 files changed, 3 insertions(+), 6 deletions(-)

diff --git a/tools/testing/selftests/drivers/net/ocelot/psfp.sh b/tools/testing/selftests/drivers/net/ocelot/psfp.sh
index 5a5cee92c665..bed748dde4b0 100755
--- a/tools/testing/selftests/drivers/net/ocelot/psfp.sh
+++ b/tools/testing/selftests/drivers/net/ocelot/psfp.sh
@@ -181,7 +181,7 @@ setup_prepare()
 
 	# Set up swp1 as a master PHC for h1, synchronized to the local
 	# CLOCK_REALTIME.
-	phc2sys_start ${swp1} ${UDS_ADDRESS_SWP1}
+	phc2sys_start ${UDS_ADDRESS_SWP1}
 
 	# Assumption true for LS1028A: h1 and h2 use the same PHC. So by
 	# synchronizing h1 to swp1 via PTP, h2 is also implicitly synchronized
diff --git a/tools/testing/selftests/net/forwarding/tsn_lib.sh b/tools/testing/selftests/net/forwarding/tsn_lib.sh
index 20c2b411ba36..b91bcd8008a9 100644
--- a/tools/testing/selftests/net/forwarding/tsn_lib.sh
+++ b/tools/testing/selftests/net/forwarding/tsn_lib.sh
@@ -22,8 +22,7 @@ fi
 
 phc2sys_start()
 {
-	local if_name=$1
-	local uds_address=$2
+	local uds_address=$1
 	local extra_args=""
 
 	if ! [ -z "${uds_address}" ]; then
@@ -33,9 +32,7 @@ phc2sys_start()
 	phc2sys_log="$(mktemp)"
 
 	chrt -f 10 phc2sys -m \
-		-c ${if_name} \
-		-s CLOCK_REALTIME \
-		-O ${UTC_TAI_OFFSET} \
+		-a -rr \
 		--step_threshold 0.00002 \
 		--first_step_threshold 0.00002 \
 		${extra_args} \
-- 
2.34.1

Re: [PATCH net-next 0/4] Improve tsn_lib selftests for future distributed tasks

[Kurt Kanzenbach]

[-- Attachment #1: Type: text/plain, Size: 917 bytes --]

On Sat Sep 24 2022, Vladimir Oltean wrote:
> Some of the boards I am working with are limited in the number of ports
> that they offer, and as more TSN related selftests are added, it is
> important to be able to distribute the work among multiple boards.
> A large part of implementing that is ensuring network-wide
> synchronization, but also permitting more streams of data to flow
> through the network. There is the more important aspect of also
> coordinating the timing characteristics of those streams, and that is
> also something that is tackled, although not in this modest patch set.
> The goal here is not to introduce new selftests yet, but just to lay a
> better foundation for them. These patches are a part of the cleanup work
> I've done while working on selftests for frame preemption. They are
> regression-tested with psfp.sh.

For this series,

Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de>

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 861 bytes --]

Re: [PATCH net-next 0/4] Improve tsn_lib selftests for future distributed tasks

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net-next.git (master)
by Jakub Kicinski <kuba@kernel.org>:

On Sat, 24 Sep 2022 00:00:11 +0300 you wrote:
> Some of the boards I am working with are limited in the number of ports
> that they offer, and as more TSN related selftests are added, it is
> important to be able to distribute the work among multiple boards.
> A large part of implementing that is ensuring network-wide
> synchronization, but also permitting more streams of data to flow
> through the network. There is the more important aspect of also
> coordinating the timing characteristics of those streams, and that is
> also something that is tackled, although not in this modest patch set.
> The goal here is not to introduce new selftests yet, but just to lay a
> better foundation for them. These patches are a part of the cleanup work
> I've done while working on selftests for frame preemption. They are
> regression-tested with psfp.sh.
> 
> [...]

Here is the summary with links:
  - [net-next,1/4] selftests: net: tsn_lib: don't overwrite isochron receiver extra args with UDS
    https://git.kernel.org/netdev/net-next/c/7d45b5fd27b4
  - [net-next,2/4] selftests: net: tsn_lib: allow running ptp4l on multiple interfaces
    https://git.kernel.org/netdev/net-next/c/7ff9396ee82c
  - [net-next,3/4] selftests: net: tsn_lib: allow multiple isochron receivers
    https://git.kernel.org/netdev/net-next/c/a7ce95ac837d
  - [net-next,4/4] selftests: net: tsn_lib: run phc2sys in automatic mode
    https://git.kernel.org/netdev/net-next/c/162d52dfee44

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