Re: [PATCH v5 6/9] selftests: livepatch: Adapt atomic replace tests to provides/obsoletes

On Sun, Aug 9, 2026 at 5:33 PM <[email protected]> wrote:
>
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [Medium] The 'replace' feature detection in test_modules/Makefile fails for out-of-tree (O=) kernel builds, breaking tests on older kernels.
> --
>
> --- Patch [6]: [PATCH v5 6/9] selftests: livepatch: Adapt atomic replace tests to provides/obsoletes ---
> commit e4d15e9f2792340ea543b405f9d074e729665c9b
> Author: Yafang Shao <[email protected]>
>
> selftests: livepatch: Adapt atomic replace tests to provides/obsoletes
>
> The legacy "replace" field in struct klp_patch was replaced by the
> provides/obsoletes mechanism. This commit adapts selftests to build and
> run on both old and new kernels by detecting the presence of the
> "replace" field.
>
> > diff --git a/tools/testing/selftests/livepatch/test_modules/Makefile b/tools/testing/selftests/livepatch/test_modules/Makefile
> > index a13d398585dcb..29c55df360467 100644
> > --- a/tools/testing/selftests/livepatch/test_modules/Makefile
> > +++ b/tools/testing/selftests/livepatch/test_modules/Makefile
> > @@ -16,6 +16,16 @@ obj-m += test_klp_atomic_replace.o \
> >       test_klp_state3.o \
> >       test_klp_syscall.o
> >
> > +# The legacy "replace" field was replaced by provides/obsoletes. When the
> > +# kernel (KDIR) still supports "replace", pass KLP_HAS_REPLACE to the
> > +# modules so that they can #ifdef the legacy "replace" code. Otherwise
> > +# the modules are built with the provides/obsoletes based code (to be
> > +# added later).
> > +KLP_HAS_REPLACE := $(shell grep -q 'bool replace' $(KDIR)/include/linux/livepatch.h 2>/dev/null && echo y)
>
> [Severity: Medium]
> Will this grep fail for out-of-tree (O=...) kernel builds?
>
> If $(KDIR) points to the build directory rather than the source directory,
> $(KDIR)/include/linux/livepatch.h might not exist.
>
> When tests are run, detect_replace_attr() will still correctly identify the
> replace sysfs attribute on the running kernel and set HAS_REPLACE_ATTR=1.
>
> However, since the module would be compiled without KLP_HAS_REPLACE, it lacks
> the replace module parameter. When the test scripts attempt to load the module
> with replace=1, insmod will fail, causing the test suite to abort.

I believe this case is invalid. We always run `make run_tests` to
execute the full livepatch selftests, which ensures all kernel modules
are rebuilt against the running kernel.

-- 
Regards
Yafang