On 1/14/25 12:18, Marcos Paulo de Souza wrote: > On Sat, 2025-01-11 at 15:42 -0300, Filipe Xavier wrote: >> This new test makes sure that ftrace can trace a >> function that was introduced by a livepatch. >> >> Signed-off-by: Filipe Xavier <felipeaggger@xxxxxxxxx> > > Thanks for the new test Filipe! > > I have some nits below, but these don't need to be addressed for the > test to be merged. Either way, > > Reviewed-by: Marcos Paulo de Souza <mpdesouza@xxxxxxxx> > Tested-by: Marcos Paulo de Souza <mpdesouza@xxxxxxxx> > > > >> --- >> Changes in v2: >> - functions.sh: added reset tracing on push and pop_config. >> - test-ftrace.sh: enabled tracing_on before test init. >> - nitpick: added double quotations on filenames and fixed some >> wording. >> - Link to v1: >> https://lore.kernel.org/r/20250102-ftrace-selftest-livepatch-v1-1-84880baefc1b@xxxxxxxxx >> --- >> tools/testing/selftests/livepatch/functions.sh | 14 ++++++++++ >> tools/testing/selftests/livepatch/test-ftrace.sh | 33 >> ++++++++++++++++++++++++ >> 2 files changed, 47 insertions(+) >> >> diff --git a/tools/testing/selftests/livepatch/functions.sh >> b/tools/testing/selftests/livepatch/functions.sh >> index >> e5d06fb402335d85959bafe099087effc6ddce12..e6c13514002dae5f8d7461f90b8 >> 241ab43024ea4 100644 >> --- a/tools/testing/selftests/livepatch/functions.sh >> +++ b/tools/testing/selftests/livepatch/functions.sh >> @@ -62,6 +62,9 @@ function push_config() { >> awk -F'[: ]' '{print "file " $1 " line " $2 >> " " $4}') >> FTRACE_ENABLED=$(sysctl --values kernel.ftrace_enabled) >> KPROBE_ENABLED=$(cat "$SYSFS_KPROBES_DIR/enabled") >> + TRACING_ON=$(cat "$SYSFS_DEBUG_DIR/tracing/tracing_on") >> + CURRENT_TRACER=$(cat >> "$SYSFS_DEBUG_DIR/tracing/current_tracer") >> + FTRACE_FILTER=$(cat >> "$SYSFS_DEBUG_DIR/tracing/set_ftrace_filter") >> } >> >> function pop_config() { >> @@ -74,6 +77,17 @@ function pop_config() { >> if [[ -n "$KPROBE_ENABLED" ]]; then >> echo "$KPROBE_ENABLED" > >> "$SYSFS_KPROBES_DIR/enabled" >> fi >> + if [[ -n "$TRACING_ON" ]]; then >> + echo "$TRACING_ON" > >> "$SYSFS_DEBUG_DIR/tracing/tracing_on" >> + fi >> + if [[ -n "$CURRENT_TRACER" ]]; then >> + echo "$CURRENT_TRACER" > >> "$SYSFS_DEBUG_DIR/tracing/current_tracer" >> + fi >> + if [[ "$FTRACE_FILTER" == *"#"* ]]; then >> + echo > "$SYSFS_DEBUG_DIR/tracing/set_ftrace_filter" >> + elif [[ -n "$FTRACE_FILTER" ]]; then >> + echo "$FTRACE_FILTER" > >> "$SYSFS_DEBUG_DIR/tracing/set_ftrace_filter" >> + fi >> } > > I believe that this could be a separate patch, since this is new > functionality that's being added to functions.sh, and not exactly > related to the new test. > >> >> function set_dynamic_debug() { >> diff --git a/tools/testing/selftests/livepatch/test-ftrace.sh >> b/tools/testing/selftests/livepatch/test-ftrace.sh >> index >> fe14f248913acbec46fb6c0fec38a2fc84209d39..66af5d726c52e48e5177804e182 >> b4ff31784d5ac 100755 >> --- a/tools/testing/selftests/livepatch/test-ftrace.sh >> +++ b/tools/testing/selftests/livepatch/test-ftrace.sh >> @@ -61,4 +61,37 @@ livepatch: '$MOD_LIVEPATCH': unpatching complete >> % rmmod $MOD_LIVEPATCH" >> >> >> +# - verify livepatch can load >> +# - check if traces have a patched function >> +# - unload livepatch and reset trace >> + >> +start_test "trace livepatched function and check that the live patch >> remains in effect" >> + >> +TRACE_FILE="$SYSFS_DEBUG_DIR/tracing/trace" >> +FUNCTION_NAME="livepatch_cmdline_proc_show" >> + >> +load_lp $MOD_LIVEPATCH >> + >> +echo 1 > "$SYSFS_DEBUG_DIR/tracing/tracing_on" >> +echo $FUNCTION_NAME > "$SYSFS_DEBUG_DIR/tracing/set_ftrace_filter" >> +echo "function" > "$SYSFS_DEBUG_DIR/tracing/current_tracer" >> +echo "" > "$TRACE_FILE" >> + >> +if [[ "$(cat /proc/cmdline)" != "$MOD_LIVEPATCH: this has been live >> patched" ]] ; then >> + echo -e "FAIL\n\n" >> + die "livepatch kselftest(s) failed" >> +fi >> + >> +grep -q $FUNCTION_NAME "$TRACE_FILE" >> +FOUND=$? >> + >> +disable_lp $MOD_LIVEPATCH >> +unload_lp $MOD_LIVEPATCH >> + >> +if [ "$FOUND" -eq 1 ]; then >> + echo -e "FAIL\n\n" >> + die "livepatch kselftest(s) failed" >> +fi >> + >> + >> exit 0 > > The test works, and that's very cool. But when running locally, I find > the if we miss check_result call it doesn't add a newline after the > "ok": > > ... > # timeout set to 0 > # selftests: livepatch: test-ftrace.sh > # TEST: livepatch interaction with ftrace_enabled sysctl ... ok > # TEST: trace livepatched function and check that the live patch > remains in effect ... ok 5 selftests: livepatch: test-ftrace.sh > # timeout set to 0 > # selftests: livepatch: test-sysfs.sh > ... > > If the check_result below is added the output if sane again: > > ... > # selftests: livepatch: test-ftrace.sh > # TEST: livepatch interaction with ftrace_enabled sysctl ... ok > # TEST: trace livepatched function and check that the live patch > remains in effect ... ok > ok 5 selftests: livepatch: test-ftrace.sh > ... > > I checked and this would be the only one test without using > check_result, so maybe we should add this either way? I'm not sure what > you guys think about it. > > > diff --git a/tools/testing/selftests/livepatch/test-ftrace.sh > b/tools/testing/selftests/livepatch/test-ftrace.sh > index 66af5d726c52..135c0fb17a98 100755 > --- a/tools/testing/selftests/livepatch/test-ftrace.sh > +++ b/tools/testing/selftests/livepatch/test-ftrace.sh > @@ -93,5 +93,18 @@ if [ "$FOUND" -eq 1 ]; then > die "livepatch kselftest(s) failed" > fi > > +check_result "% insmod test_modules/$MOD_LIVEPATCH.ko > +livepatch: enabling patch '$MOD_LIVEPATCH' > +livepatch: '$MOD_LIVEPATCH': initializing patching transition > +livepatch: '$MOD_LIVEPATCH': starting patching transition > +livepatch: '$MOD_LIVEPATCH': completing patching transition > +livepatch: '$MOD_LIVEPATCH': patching complete > +% echo 0 > $SYSFS_KLP_DIR/$MOD_LIVEPATCH/enabled > +livepatch: '$MOD_LIVEPATCH': initializing unpatching transition > +livepatch: '$MOD_LIVEPATCH': starting unpatching transition > +livepatch: '$MOD_LIVEPATCH': completing unpatching transition > +livepatch: '$MOD_LIVEPATCH': unpatching complete > +% rmmod $MOD_LIVEPATCH" > + > Ah good catch, I noticed the newline, too, but didn't notice that the test wasn't using check_result(). For consistency, let's make that change before merging. Thanks, -- Joe
