[PATCH net-next v2] selftests: net: exit with test failure status

[PATCH net-next v2] selftests: net: exit with test failure status

[Hangbin Liu]

From: Hangbin Liu <liuhangbin@kylinos.cn>

The changed four tests can print [FAIL] subtests while still exiting with
status 0. Their log_test(), aka log_test_expected() in lib.sh, helpers
record a failure in ret, but the scripts never propagate ret to the exit
code, so a kselftest runner treats the run as a pass.

Initialize ret and exit with it at the end of each script so failed
subtests are reflected in the exit status.

Signed-off-by: Hangbin Liu <liuhangbin@kylinos.cn>
---
The patch doesn't fix a regression, nothing is failing (or passing when it
shouldn't) because of it. So target the patch to net-next.
---
Changes in v2:
- Also update fib_nexthop_multiprefix.sh (Ido Schimmel)
- Link to v1: https://lore.kernel.org/r/20261014-self_ret-v1-1-6b5a95ba6106@kylinos.cn

To: "David S. Miller" <davem@davemloft.net>
To: Eric Dumazet <edumazet@kernel.org>
To: Jakub Kicinski <kuba@kernel.org>
To: Paolo Abeni <pabeni@redhat.com>
To: Simon Horman <horms@kernel.org>
To: Shuah Khan <shuah@kernel.org>
To: David Ahern <dsahern@kernel.org>
To: Ido Schimmel <idosch@nvidia.com>
Cc: Hangbin Liu <hangbin.liu@linux.dev>
Cc: netdev@vger.kernel.org
Cc: linux-kselftest@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
---
 tools/testing/selftests/net/fdb_flush.sh               | 3 +++
 tools/testing/selftests/net/fib-onlink-tests.sh        | 3 +++
 tools/testing/selftests/net/fib_nexthop_multiprefix.sh | 1 +
 tools/testing/selftests/net/l2tp.sh                    | 3 +++
 4 files changed, 10 insertions(+)

diff --git a/tools/testing/selftests/net/fdb_flush.sh b/tools/testing/selftests/net/fdb_flush.sh
index 4965e52d5ef5..0827fdd60f8c 100755
--- a/tools/testing/selftests/net/fdb_flush.sh
+++ b/tools/testing/selftests/net/fdb_flush.sh
@@ -6,6 +6,7 @@
 # some combinations of arguments.
 
 source lib.sh
+ret=0
 
 FLUSH_BY_STATE_TESTS="
 	vxlan_test_flush_by_permanent
@@ -778,3 +779,5 @@ for t in $TESTS
 do
 	setup; $t; cleanup;
 done
+
+exit $ret
diff --git a/tools/testing/selftests/net/fib-onlink-tests.sh b/tools/testing/selftests/net/fib-onlink-tests.sh
index a26075abf7d2..325da693c918 100755
--- a/tools/testing/selftests/net/fib-onlink-tests.sh
+++ b/tools/testing/selftests/net/fib-onlink-tests.sh
@@ -6,6 +6,7 @@
 source lib.sh
 PAUSE_ON_FAIL=${PAUSE_ON_FAIL:=no}
 VERBOSE=0
+ret=0
 
 # Network interfaces
 # - odd in current namespace; even in peer ns
@@ -473,3 +474,5 @@ if [ "$TESTS" != "none" ]; then
 	printf "\nTests passed: %3d\n" ${nsuccess}
 	printf "Tests failed: %3d\n"   ${nfail}
 fi
+
+exit $ret
diff --git a/tools/testing/selftests/net/fib_nexthop_multiprefix.sh b/tools/testing/selftests/net/fib_nexthop_multiprefix.sh
index bd0b8a053ea1..2671cff270e5 100755
--- a/tools/testing/selftests/net/fib_nexthop_multiprefix.sh
+++ b/tools/testing/selftests/net/fib_nexthop_multiprefix.sh
@@ -269,3 +269,4 @@ if [ $ret -eq 0 ]; then
 fi
 
 cleanup
+exit $ret
diff --git a/tools/testing/selftests/net/l2tp.sh b/tools/testing/selftests/net/l2tp.sh
index 41e8b19d1bd3..de4b5da2da1b 100755
--- a/tools/testing/selftests/net/l2tp.sh
+++ b/tools/testing/selftests/net/l2tp.sh
@@ -16,6 +16,7 @@
 source lib.sh
 VERBOSE=0
 PAUSE_ON_FAIL=no
+ret=0
 
 which ping6 > /dev/null 2>&1 && ping6=$(which ping6) || ping6=$(which ping)
 
@@ -357,3 +358,5 @@ cleanup
 
 printf "\nTests passed: %3d\n" ${nsuccess}
 printf "Tests failed: %3d\n"   ${nfail}
+
+exit $ret

---
base-commit: 98c2ad4f5d805dbc1ea436a334ee59311806a48c
change-id: 20261014-self_ret-83b030f32ba3

Best regards,
--  
Hangbin Liu <liuhangbin@kylinos.cn>

Re: [PATCH net-next v2] selftests: net: exit with test failure status

[Ido Schimmel]

On Mon, Oct 19, 2026 at 08:41:03AM +0800, Hangbin Liu wrote:
> From: Hangbin Liu <liuhangbin@kylinos.cn>
> 
> The changed four tests can print [FAIL] subtests while still exiting with
> status 0. Their log_test(), aka log_test_expected() in lib.sh, helpers
> record a failure in ret, but the scripts never propagate ret to the exit
> code, so a kselftest runner treats the run as a pass.
> 
> Initialize ret and exit with it at the end of each script so failed
> subtests are reflected in the exit status.
> 
> Signed-off-by: Hangbin Liu <liuhangbin@kylinos.cn>

Reviewed-by: Ido Schimmel <idosch@nvidia.com>