The test did not catch non-zero exit status of the spawned coprocess. To
make it happen, Drop the line killing it (it will exit anyway) and pass
its PID to 'wait'.
While being at it, put the sleep into the correct spot (otherwise the
check for chain 'foo' existence fails as it runs too early) and make
said chain existence check effective.
Fixes: 4e3c11a6f5a94 ("nft: Fix for ruleset flush while restoring")
Signed-off-by: Phil Sutter <phil@xxxxxx>
---
.../shell/testcases/nft-only/0007-mid-restore-flush_0 | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/iptables/tests/shell/testcases/nft-only/0007-mid-restore-flush_0 b/iptables/tests/shell/testcases/nft-only/0007-mid-restore-flush_0
index 43880ffbc5851..981f007f205b9 100755
--- a/iptables/tests/shell/testcases/nft-only/0007-mid-restore-flush_0
+++ b/iptables/tests/shell/testcases/nft-only/0007-mid-restore-flush_0
@@ -13,11 +13,11 @@ COMMIT
:foo [0:0]
EOF
-$XT_MULTI iptables-save | grep -q ':foo'
+sleep 1
+$XT_MULTI iptables-save | grep -q ':foo' || exit 1
nft flush ruleset
echo "COMMIT" >&"${COPROC[1]}"
-sleep 1
-
-[[ -n $COPROC_PID ]] && kill $COPROC_PID
-wait
+# close the pipe to make iptables-restore exit if it didn't error out yet
+eval "exec ${COPROC[1]}>&-"
+wait $COPROC_PID
--
2.41.0
