Re: [RFC PATCH bpf-next 2/2] selftests/livepatch: update test-ftrace.sh for deprecated ftrace_enabled

[Miroslav Benes <mbenes@suse.cz>]

On Thu, 30 Jul 2026, Steven Rostedt wrote:

> On Thu, 30 Jul 2026 12:35:44 -0400
> Andrey Grodzovsky <andrey.grodzovsky@crowdstrike.com> wrote:
> 
> > kernel.ftrace_enabled=0 is now always refused, so the old scenario
> > (disable, fail to load a livepatch, re-enable, load) no longer applies.
> > Replace it with a single check that the write is refused with the new
> > error, and drop the now-stale disable/reload flow.
> > 
> > Assisted-by: Claude:claude-sonnet-5
> > Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@crowdstrike.com>
> > ---
> >  .../selftests/livepatch/test-ftrace.sh        | 49 ++-----------------
> >  1 file changed, 3 insertions(+), 46 deletions(-)
> > 
> > diff --git a/tools/testing/selftests/livepatch/test-ftrace.sh b/tools/testing/selftests/livepatch/test-ftrace.sh
> > index d2c3dea63104..c88f4d13a134 100755
> > --- a/tools/testing/selftests/livepatch/test-ftrace.sh
> > +++ b/tools/testing/selftests/livepatch/test-ftrace.sh
> > @@ -9,56 +9,13 @@ MOD_LIVEPATCH=test_klp_livepatch
> >  setup_config
> >  
> >  
> > -# - turn ftrace_enabled OFF and verify livepatches can't load
> > -# - turn ftrace_enabled ON and verify livepatch can load
> > -# - verify that ftrace_enabled can't be turned OFF while a livepatch is loaded
> > +# - verify that kernel.ftrace_enabled=0 is refused (deprecated knob)
> >  
> > -start_test "livepatch interaction with ftrace_enabled sysctl"
> > +start_test "ftrace_enabled sysctl write is refused (deprecated)"
> >  
> > -set_ftrace_enabled 0
> > -load_failing_mod $MOD_LIVEPATCH
> > -
> > -set_ftrace_enabled 1
> > -load_lp $MOD_LIVEPATCH
> > -if [[ "$(cat /proc/cmdline)" != "$MOD_LIVEPATCH: this has been live patched" ]] ; then
> > -	echo -e "FAIL\n\n"
> > -	die "livepatch kselftest(s) failed"
> > -fi
> > -
> 
> This test should work for both old and new kernels. I would have it still
> do everything here but if setting ftrace_enabled returns -EOPNOTSUPP then
> do the simple path.

Yes, please. It would be ideal.

Miroslav