[PATCH v2 07/58] objtool/klp: Give each run one working directory, one per test inside it
From: Song Liu <song@kernel.org>
Date: 2026-09-14 06:26:07
Subsystem:
objtool, the rest · Maintainers:
Josh Poimboeuf, Peter Zijlstra, Linus Torvalds
Each test made its own mktemp directory, so a run scattered them through
/tmp among everything else using the same names, and there was no way to
inspect what a failing test built -- the objects are gone the moment it
exits, which is right every time but the one that matters.
Give the run one directory and each test a subdirectory of it, named for
the test and nested the way the source is:
/tmp/klp-tests.3zC6oyfd/
generic/test-basic/{orig.o,patched.o,out.o,Module.symvers,...}
x86/test-kcfi/...
--keep then reports one path instead of forty, and removing it is one
command. A test run by hand still falls back to a temp directory of its
own, and still says where that went.
Detecting a leak comes free. Cleanup is now "remove each test's directory,
then rmdir the run's", and the rmdir fails if anything is left -- so a test
which dies without running its own cleanup is reported rather than quietly
leaving something behind.
preflight names the directory the tests build under, since mktemp honours
TMPDIR and a run in a container is otherwise silent about where its work
went. The runner says what to do only when it is relevant: on a failure
without --keep, that the option exists; with it, how many directories are
waiting. Usage text is not read while something is broken.
Assisted-by: Claude:claude-opus-5
Signed-off-by: Song Liu <song@kernel.org>
---
tools/objtool/tests/lib.sh | 23 ++++++++++++++++++-----
tools/objtool/tests/run-tests.sh | 26 ++++++++++++++++++++++++--
2 files changed, 42 insertions(+), 7 deletions(-)
diff --git a/tools/objtool/tests/lib.sh b/tools/objtool/tests/lib.sh
index 40fe74082158..7b8b5b2f0c92 100644
--- a/tools/objtool/tests/lib.sh
+++ b/tools/objtool/tests/lib.sh@@ -120,6 +120,7 @@ klp_preflight() # objtool $OBJTOOL (klp: yes) # compiler $cc_version # arch $KLP_TEST_ARCH$([ "$arch" = "$host" ] || echo " (host $host, cross)") +# tmpdir ${TMPDIR:-/tmp} (each test builds in a fresh directory here) EOF }
@@ -179,15 +180,27 @@ xpass() exit 1 } -cleanup() { [ -n "$workdir" ] && rm -rf "$workdir"; } +cleanup() +{ + [ -n "$workdir" ] || return 0 + + if [ -n "${KLP_TEST_KEEP:-}" ]; then + [ -n "${KLP_TEST_WORKDIR:-}" ] || echo "# kept $workdir" + return 0 + fi + + rm -rf "$workdir" +} # setup [exported symbol...] setup() { - # The environment was checked once when this file was sourced, so there - # is nothing to verify here: objtool exists at the resolved path, has - # klp support, and $CC works. - workdir="$(mktemp -d)" || fail "mktemp failed" + if [ -n "${KLP_TEST_WORKDIR:-}" ]; then + workdir="$KLP_TEST_WORKDIR" + mkdir -p "$workdir" || fail "cannot create $workdir" + else + workdir="$(mktemp -d)" || fail "mktemp failed" + fi trap cleanup EXIT export_syms "$@"
diff --git a/tools/objtool/tests/run-tests.sh b/tools/objtool/tests/run-tests.sh
index 6b937fc7f5bd..e6f1ac1b5d0d 100755
--- a/tools/objtool/tests/run-tests.sh
+++ b/tools/objtool/tests/run-tests.sh@@ -23,13 +23,17 @@ export LC_ALL=C usage() { cat <<EOF -usage: $(basename "$0") [test...] +usage: $(basename "$0") [-k|--keep] [test...] Run the objtool klp tests for this architecture: everything in generic/, plus everything in the directory named for it. With no arguments, runs all of them. A test may be named with or without its "test-" prefix and ".sh" suffix, and is looked for in both directories. +Options: + -k, --keep do not delete each test's working directory; print its path, + so the objects a failing test built can be looked at + Environment: OBJTOOL objtool binary to test (default ../objtool) CC compiler used to build fixtures (default gcc)
@@ -45,6 +49,7 @@ cd "$(dirname "$0")" || exit 1 while [ $# -gt 0 ]; do case "$1" in -h|--help) usage ;; + -k|--keep) export KLP_TEST_KEEP=1; shift ;; --) shift; break ;; -*) echo "unknown option: $1" >&2; usage 1 ;; *) break ;;
@@ -98,12 +103,18 @@ else done fi +# One directory for the whole run, one per test inside it, mirroring the +# source layout. A run then leaves a single thing behind instead of 39 +# scattered among everything else using mktemp. +rundir="$(mktemp -d "${TMPDIR:-/tmp}/klp-tests.XXXXXXXX")" || + { echo "Bail out! cannot create a working directory" >&2; exit 1; } + echo "1..${#tests[@]}" pass=0 fail=0 static_skip=0 probe_skip=0 xfail=0 xpass=0 for t in "${tests[@]}"; do - out="$(./"$t" 2>&1)" + out="$(KLP_TEST_WORKDIR="$rundir/${t%.sh}" ./"$t" 2>&1)" rc=$? # A test prints one result line, but it is not necessarily the only
@@ -156,4 +167,15 @@ done echo "# pass:$pass fail:$fail static-skip:$static_skip" \ "probe-skip:$probe_skip xfail:$xfail xpass:$xpass" +# A failure is the one time the objects matter, and by default they are +# already gone. Say so then rather than in the usage text nobody reads while +# something is broken. +if [ -n "${KLP_TEST_KEEP:-}" ]; then + echo "# working directories kept in $rundir -- inspect, then rm -rf it" +elif ! rmdir "$rundir"/*/ "$rundir" 2>/dev/null; then + echo "# $rundir was not empty; a test did not clean up after itself" +elif [ "$fail" != 0 ] || [ "$xpass" != 0 ]; then + echo "# re-run with --keep to hold on to what a failing test built" +fi + [ "$fail" = 0 ] && [ "$xpass" = 0 ]
--
2.53.0-Meta