Adding upstream version 1.0.9.upstream/1.0.9upstream

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

3 files changed, 294 insertions, 0 deletions

diff --git a/tests/shell/helpers/nft-valgrind-wrapper.sh b/tests/shell/helpers/nft-valgrind-wrapper.sh
new file mode 100755
index 0000000..98bbdf4
--- /dev/null
+++ b/tests/shell/helpers/nft-valgrind-wrapper.sh
@@ -0,0 +1,31 @@
+#!/bin/bash -e
+
+SUFFIX="$(date "+%H%M%S.%6N").$$"
+
+rc=0
+libtool \
+	--mode=execute \
+	valgrind \
+		--log-file="$NFT_TEST_TESTTMPDIR/valgrind.$SUFFIX.%p.log" \
+		--trace-children=yes \
+		--leak-check=full \
+		--show-leak-kinds=all \
+		--num-callers=100 \
+		--error-exitcode=122 \
+		--vgdb-prefix="$_NFT_TEST_VALGRIND_VGDB_PREFIX-$SUFFIX" \
+		$NFT_TEST_VALGRIND_OPTS \
+		"$NFT_REAL" \
+		"$@" \
+	|| rc=$?
+
+if [ "$rc" -eq 122 ] ; then
+	shopt -s nullglob
+	FILES=( "$NFT_TEST_TESTTMPDIR/valgrind.$SUFFIX."*".log" )
+	shopt -u nullglob
+	(
+		printf '%s\n' "args: $*"
+		printf '%s\n' "${FILES[*]}"
+	) >> "$NFT_TEST_TESTTMPDIR/rc-failed-valgrind"
+fi
+
+exit $rc
diff --git a/tests/shell/helpers/random-source.sh b/tests/shell/helpers/random-source.sh
new file mode 100755
index 0000000..91a8248
--- /dev/null
+++ b/tests/shell/helpers/random-source.sh
@@ -0,0 +1,40 @@
+#!/bin/bash
+
+# Commands like `sort` and `shuf` have a "--random-source" argument, for
+# generating a stable, reproducible output. However, they require an input
+# that provides sufficiently many bytes (depending on the input).
+#
+# This script generates a stream that can be used like
+#
+#     shuf --random-source=<($0 "$seed")
+
+seed=""
+for a; do
+	seed="$seed${#a}:$a\n"
+done
+
+if command -v openssl &>/dev/null ; then
+	# We have openssl. Use it.
+	# https://www.gnu.org/software/coreutils/manual/html_node/Random-sources.html#Random-sources
+	#
+	# Note that we don't care that different installations/architectures generate the
+	# same output.
+	openssl enc -aes-256-ctr -pass "pass:$seed" -nosalt </dev/zero 2>/dev/null
+else
+	# Hack something. It's much slower.
+	idx=0
+	while : ; do
+		idx="$((idx++))"
+		seed="$(sha256sum <<<"$idx.$seed")"
+		echo ">>>$seed" >> a
+		seed="${seed%% *}"
+		LANG=C awk -v s="$seed" 'BEGIN{
+			for (i=1; i <= length(s); i+=2) {
+				xchar = substr(s, i, 2);
+				decnum = strtonum("0x"xchar);
+				printf("%c", decnum);
+			}
+		}' || break
+	done
+fi
+exit 0
diff --git a/tests/shell/helpers/test-wrapper.sh b/tests/shell/helpers/test-wrapper.sh
new file mode 100755
index 0000000..13b918f
--- /dev/null
+++ b/tests/shell/helpers/test-wrapper.sh
@@ -0,0 +1,223 @@
+#!/bin/bash -e
+
+# This wrapper wraps the invocation of the test. It is called by run-tests.sh,
+# and already in the unshared namespace.
+#
+# For some printf debugging, you can also patch this file.
+
+array_contains() {
+	local needle="$1"
+	local a
+	shift
+	for a; do
+		[ "$a" = "$needle" ] && return 0
+	done
+	return 1
+}
+
+TEST="$1"
+TESTBASE="$(basename "$TEST")"
+TESTDIR="$(dirname "$TEST")"
+
+START_TIME="$(cut -d ' ' -f1 /proc/uptime)"
+
+export TMPDIR="$NFT_TEST_TESTTMPDIR"
+
+CLEANUP_UMOUNT_VAR_RUN=n
+
+cleanup() {
+	if [ "$CLEANUP_UMOUNT_VAR_RUN" = y ] ; then
+		umount "/var/run" &>/dev/null || :
+	fi
+}
+
+trap cleanup EXIT
+
+printf '%s\n' "$TEST" > "$NFT_TEST_TESTTMPDIR/name"
+
+read tainted_before < /proc/sys/kernel/tainted
+
+if [ "$NFT_TEST_HAS_UNSHARED_MOUNT" = y ] ; then
+	# We have a private mount namespace. We will mount /var/run/ as a tmpfs.
+	#
+	# The main purpose is so that we can create /var/run/netns, which is
+	# required for `ip netns add` to work.  When running as rootless, this
+	# is necessary to get such tests to pass. When running rootful, it's
+	# still useful to not touch the "real" /var/run/netns of the system.
+	#
+	# Note that this also hides everything that might reside in /var/run.
+	# That is desirable, as tests should not depend on content there (or if
+	# they do, we need to explicitly handle it as appropriate).
+	if mount -t tmpfs --make-private "/var/run" ; then
+		CLEANUP_UMOUNT_VAR_RUN=y
+	fi
+	mkdir -p /var/run/netns
+fi
+
+TEST_TAGS_PARSED=0
+ensure_TEST_TAGS() {
+	if [ "$TEST_TAGS_PARSED" = 0 ] ; then
+		TEST_TAGS_PARSED=1
+		TEST_TAGS=( $(sed -n '1,10 { s/^.*\<\(NFT_TEST_REQUIRES\|NFT_TEST_SKIP\)\>\s*(\s*\(NFT_TEST_SKIP_[a-zA-Z0-9_]\+\|NFT_TEST_HAVE_[a-zA-Z0-9_]\+\)\s*).*$/\1(\2)/p }' "$1" 2>/dev/null || : ) )
+	fi
+}
+
+rc_test=0
+
+if [ "$rc_test" -eq 0 ] ; then
+	for KEY in $(compgen -v | grep '^NFT_TEST_HAVE_') ; do
+		if [ "${!KEY}" != n ]; then
+			continue
+		fi
+		ensure_TEST_TAGS "$TEST"
+		if array_contains "NFT_TEST_REQUIRES($KEY)" "${TEST_TAGS[@]}" ; then
+			echo "Test skipped due to $KEY=n (test has \"NFT_TEST_REQUIRES($KEY)\" tag)" >> "$NFT_TEST_TESTTMPDIR/testout.log"
+			rc_test=77
+			break
+		fi
+	done
+fi
+
+if [ "$rc_test" -eq 0 ] ; then
+	for KEY in $(compgen -v | grep '^NFT_TEST_SKIP_') ; do
+		if [ "${!KEY}" != y ]; then
+			continue
+		fi
+		ensure_TEST_TAGS "$TEST"
+		if array_contains "NFT_TEST_SKIP($KEY)" "${TEST_TAGS[@]}" ; then
+			echo "Test skipped due to $KEY=y (test has \"NFT_TEST_SKIP($KEY)\" tag)" >> "$NFT_TEST_TESTTMPDIR/testout.log"
+			rc_test=77
+			break
+		fi
+	done
+fi
+
+if [ "$rc_test" -eq 0 ] ; then
+	"$TEST" &>> "$NFT_TEST_TESTTMPDIR/testout.log" || rc_test=$?
+fi
+
+$NFT list ruleset > "$NFT_TEST_TESTTMPDIR/ruleset-after"
+
+read tainted_after < /proc/sys/kernel/tainted
+
+DUMPPATH="$TESTDIR/dumps"
+DUMPFILE="$DUMPPATH/$TESTBASE.nft"
+NODUMPFILE="$DUMPPATH/$TESTBASE.nodump"
+
+dump_written=
+
+# The caller can request a re-geneating of the dumps, by setting
+# DUMPGEN=y.
+#
+# This only will happen if the command completed with success.
+#
+# It also will only happen for tests, that have a "$DUMPPATH" directory. There
+# might be tests, that don't want to have dumps created. The existence of the
+# directory controls that. Tests that have a "$NODUMPFILE" file, don't get a dump generated.
+if [ "$rc_test" -eq 0 -a "$DUMPGEN" = y -a -d "$DUMPPATH" -a ! -f "$NODUMPFILE" ] ; then
+	dump_written=y
+	if [ ! -f "$DUMPFILE" ] ; then
+		# No dumpfile exists yet. We generate both a .nft and a .nodump
+		# file. The user can pick which one to commit to git.
+		: > "$NODUMPFILE"
+	fi
+	cat "$NFT_TEST_TESTTMPDIR/ruleset-after" > "$DUMPFILE"
+fi
+
+rc_dump=0
+if [ "$rc_test" -ne 77 -a -f "$DUMPFILE" ] ; then
+	if [ "$dump_written" != y ] ; then
+		if ! $DIFF -u "$DUMPFILE" "$NFT_TEST_TESTTMPDIR/ruleset-after" &> "$NFT_TEST_TESTTMPDIR/ruleset-diff" ; then
+			rc_dump=1
+		else
+			rm -f "$NFT_TEST_TESTTMPDIR/ruleset-diff"
+		fi
+	fi
+fi
+if [ "$rc_dump" -ne 0 ] ; then
+	echo "$DUMPFILE" > "$NFT_TEST_TESTTMPDIR/rc-failed-dump"
+fi
+
+rc_chkdump=0
+# check that a flush after the test succeeds. We anyway need a clean ruleset
+# for the `nft --check` next.
+$NFT flush ruleset &> "$NFT_TEST_TESTTMPDIR/rc-failed-chkdump" || rc_chkdump=1
+if [ -f "$DUMPFILE" ] ; then
+	# We have a dumpfile. Call `nft --check` to possibly cover new code
+	# paths.
+	if [ "$rc_test" -eq 77 ] ; then
+		# The test was skipped. Possibly we don't have the required
+		# features to process this file. Ignore any output and exit
+		# code, but still call the program (for valgrind or sanitizer
+		# issue we hope to find).
+		$NFT --check -f "$DUMPFILE" &>/dev/null || :
+	else
+		$NFT --check -f "$DUMPFILE" &>> "$NFT_TEST_TESTTMPDIR/rc-failed-chkdump" || rc_chkdump=1
+	fi
+fi
+if [ -s "$NFT_TEST_TESTTMPDIR/rc-failed-chkdump" ] ; then
+	# Non-empty output? That is wrong.
+	rc_chkdump=1
+elif [ "$rc_chkdump" -eq 0 ] ; then
+	rm -rf "$NFT_TEST_TESTTMPDIR/rc-failed-chkdump"
+fi
+if [ "$rc_chkdump" -ne 0 ] ; then
+	# Ensure we don't have empty output files. Always write something, so
+	# that `grep ^ -R` lists the file.
+	echo -e "<<<<<\n\nCalling \`nft --check\` (or \`nft flush ruleset\`) failed for \"$DUMPFILE\"" >> "$NFT_TEST_TESTTMPDIR/rc-failed-chkdump"
+fi
+
+rc_valgrind=0
+[ -f "$NFT_TEST_TESTTMPDIR/rc-failed-valgrind" ] && rc_valgrind=1
+
+rc_tainted=0
+if [ "$tainted_before" != "$tainted_after" ] ; then
+	echo "$tainted_after" > "$NFT_TEST_TESTTMPDIR/rc-failed-tainted"
+	rc_tainted=1
+fi
+
+if [ "$rc_valgrind" -ne 0 ] ; then
+	rc_exit=122
+elif [ "$rc_tainted" -ne 0 ] ; then
+	rc_exit=123
+elif [ "$rc_test" -ge 118 -a "$rc_test" -le 124 ] ; then
+	# Special exit codes are reserved. Coerce them.
+	rc_exit=125
+elif [ "$rc_test" -ne 0 ] ; then
+	rc_exit="$rc_test"
+elif [ "$rc_dump" -ne 0 ] ; then
+	rc_exit=124
+elif [ "$rc_chkdump" -ne 0 ] ; then
+	rc_exit=121
+else
+	rc_exit=0
+fi
+
+
+# We always write the real exit code of the test ($rc_test) to one of the files
+# rc-{ok,skipped,failed}, depending on which it is.
+#
+# Note that there might be other rc-failed-{dump,tainted,valgrind} files with
+# additional errors. Note that if such files exist, the overall state will
+# always be failed too (and an "rc-failed" file exists).
+#
+# On failure, we also write the combined "$rc_exit" code from "test-wrapper.sh"
+# to "rc-failed-exit" file.
+#
+# This means, failed tests will have a "rc-failed" file, and additional
+# "rc-failed-*" files exist for further information.
+if [ "$rc_exit" -eq 0 ] ; then
+	RC_FILENAME="rc-ok"
+elif [ "$rc_exit" -eq 77 ] ; then
+	RC_FILENAME="rc-skipped"
+else
+	RC_FILENAME="rc-failed"
+	echo "$rc_exit" > "$NFT_TEST_TESTTMPDIR/rc-failed-exit"
+fi
+echo "$rc_test" > "$NFT_TEST_TESTTMPDIR/$RC_FILENAME"
+
+END_TIME="$(cut -d ' ' -f1 /proc/uptime)"
+WALL_TIME="$(awk -v start="$START_TIME" -v end="$END_TIME" "BEGIN { print(end - start) }")"
+printf "%s\n" "$WALL_TIME" "$START_TIME" "$END_TIME" > "$NFT_TEST_TESTTMPDIR/times"
+
+exit "$rc_exit"