Re: [External] Re: [RFC PATCH bpf-next 2/2] selftests/bpf: Add tests for duplicate kprobe symbol handling

[Jiri Olsa <olsajiri@gmail.com>]

On Tue, Feb 24, 2026 at 11:54:24AM +0100, Jiri Olsa wrote:
> On Mon, Feb 23, 2026 at 01:56:15PM -0500, Andrey Grodzovsky wrote:
> > On Mon, Feb 23, 2026 at 4:19 AM Jiri Olsa <olsajiri@gmail.com> wrote:
> > >
> > > On Wed, Feb 18, 2026 at 05:56:17PM -0500, Andrey Grodzovsky wrote:
> > >
> > > SNIP
> > >
> > > > diff --git a/tools/testing/selftests/bpf/test_kmods/bpf_testmod_dup_sym.c b/tools/testing/selftests/bpf/test_kmods/bpf_testmod_dup_sym.c
> > > > new file mode 100644
> > > > index 000000000000..98b3e085ae90
> > > > --- /dev/null
> > > > +++ b/tools/testing/selftests/bpf/test_kmods/bpf_testmod_dup_sym.c
> > > > @@ -0,0 +1,45 @@
> > > > +// SPDX-License-Identifier: GPL-2.0
> > > > +/* Copyright (c) 2025 CrowdStrike */
> > > > +/* Test module for duplicate kprobe symbol handling */
> > > > +#include <linux/init.h>
> > > > +#include <linux/module.h>
> > > > +#include <linux/kernel.h>
> > > > +
> > > > +/* Duplicate symbol to test kprobe attachment with duplicate symbols.
> > > > + * This creates a duplicate of the syscall wrapper used in attach_probe tests.
> > > > + * The libbpf fix should handle this by preferring the vmlinux symbol.
> > > > + * This function should NEVER be called - kprobes should attach to vmlinux version.
> > > > + */
> > > > +#ifdef __x86_64__
> > > > +noinline int __x64_sys_nanosleep(void)
> > > > +#elif defined(__s390x__)
> > > > +noinline int __s390x_sys_nanosleep(void)
> > > > +#elif defined(__aarch64__)
> > > > +noinline int __arm64_sys_nanosleep(void)
> > > > +#elif defined(__riscv)
> > > > +noinline int __riscv_sys_nanosleep(void)
> > > > +#else
> > > > +noinline int sys_nanosleep(void)
> > > > +#endif
> > > > +{
> > > > +     WARN_ONCE(1, "bpf_testmod_dup_sym: dummy nanosleep symbol called - this should never execute!\n");
> > > > +     return -EINVAL;
> > > > +}
> > >
> > > hi,
> > > this fails to build for me:
> > >
> > >   CC [M]  bpf_testmod_dup_sym.o
> > > bpf_testmod_dup_sym.c:14:14: error: no previous prototype for ‘__x64_sys_nanosleep’ [-Werror=missing-prototypes]
> > >    14 | noinline int __x64_sys_nanosleep(void)
> > >       |              ^~~~~~~~~~~~~~~~~~~
> > >
> > > jirka
> > 
> > Hey Jiri, ran i on my local mirror of CI repo  with latest rebased to
> > origin/bpf-next_base with command
> > tools/testing/selftests/bpf/vmtest.sh -- ./test_progs -t attach_probe
> > and (arch is x86) and eveyrhing pass.
> > Then i also ran in guthub
> > https://github.com/kernel-patches/bpf/pull/11138 and the only single
> > failure i got
> > is during runtime is some unrealted test failure.
> > https://github.com/kernel-patches/bpf/actions/runs/22318486040/job/64571209293?pr=11138
> > This seems to be because of me defining sys_nanosleep duplicate
> > without static qualifier and so if your
> > local build has -Werror=missing-prototypes,  this warning cause a
> > failure? Do you think we should
> > add static here ?
> 
> I think adding declaration will fix it, but if CI is fine,
> it's something in my setup.. please ignore, I'll check

ok it's just warning, but does not break the build, I can see it
in CI run as well..  but I guess we still want to fix it.. let's
add the missing externs?

jirka