Re: [PATCH 10/10] selftest/bpf: Add kprobe_multi test for bpf_cookie values

[Jiri Olsa <olsajiri@gmail.com>]

On Fri, Mar 04, 2022 at 03:11:26PM -0800, Andrii Nakryiko wrote:
> On Tue, Feb 22, 2022 at 9:08 AM Jiri Olsa <jolsa@kernel.org> wrote:
> >
> > Adding bpf_cookie test for programs attached by kprobe_multi links.
> >
> > Signed-off-by: Jiri Olsa <jolsa@kernel.org>
> > ---
> >  .../selftests/bpf/prog_tests/bpf_cookie.c     | 72 +++++++++++++++++++
> >  .../bpf/progs/kprobe_multi_bpf_cookie.c       | 62 ++++++++++++++++
> >  2 files changed, 134 insertions(+)
> >  create mode 100644 tools/testing/selftests/bpf/progs/kprobe_multi_bpf_cookie.c
> >
> > diff --git a/tools/testing/selftests/bpf/prog_tests/bpf_cookie.c b/tools/testing/selftests/bpf/prog_tests/bpf_cookie.c
> > index cd10df6cd0fc..edfb9f8736c6 100644
> > --- a/tools/testing/selftests/bpf/prog_tests/bpf_cookie.c
> > +++ b/tools/testing/selftests/bpf/prog_tests/bpf_cookie.c
> > @@ -7,6 +7,7 @@
> >  #include <unistd.h>
> >  #include <test_progs.h>
> >  #include "test_bpf_cookie.skel.h"
> > +#include "kprobe_multi_bpf_cookie.skel.h"
> >
> >  /* uprobe attach point */
> >  static void trigger_func(void)
> > @@ -63,6 +64,75 @@ static void kprobe_subtest(struct test_bpf_cookie *skel)
> >         bpf_link__destroy(retlink2);
> >  }
> >
> > +static void kprobe_multi_subtest(void)
> > +{
> > +       DECLARE_LIBBPF_OPTS(bpf_link_create_opts, opts);
> > +       int err, prog_fd, link1_fd = -1, link2_fd = -1;
> > +       LIBBPF_OPTS(bpf_test_run_opts, topts);
> 
> consistency ftw, LIBBPF_OPTS

ok

> 
> 
> > +       struct kprobe_multi_bpf_cookie *skel = NULL;
> > +       __u64 addrs[8], cookies[8];
> > +
> 
> [..]
> 
> > diff --git a/tools/testing/selftests/bpf/progs/kprobe_multi_bpf_cookie.c b/tools/testing/selftests/bpf/progs/kprobe_multi_bpf_cookie.c
> > new file mode 100644
> > index 000000000000..d6f8454ba093
> > --- /dev/null
> > +++ b/tools/testing/selftests/bpf/progs/kprobe_multi_bpf_cookie.c
> > @@ -0,0 +1,62 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +#include <linux/bpf.h>
> > +#include <bpf/bpf_helpers.h>
> > +#include <bpf/bpf_tracing.h>
> > +
> > +char _license[] SEC("license") = "GPL";
> > +
> > +extern const void bpf_fentry_test1 __ksym;
> > +extern const void bpf_fentry_test2 __ksym;
> > +extern const void bpf_fentry_test3 __ksym;
> > +extern const void bpf_fentry_test4 __ksym;
> > +extern const void bpf_fentry_test5 __ksym;
> > +extern const void bpf_fentry_test6 __ksym;
> > +extern const void bpf_fentry_test7 __ksym;
> > +extern const void bpf_fentry_test8 __ksym;
> > +
> > +/* No tests, just to trigger bpf_fentry_test* through tracing test_run */
> > +SEC("fentry/bpf_modify_return_test")
> > +int BPF_PROG(test1)
> > +{
> > +       return 0;
> > +}
> > +
> > +__u64 test2_result = 0;
> > +
> > +SEC("kprobe.multi/bpf_fentry_tes??")
> > +int test2(struct pt_regs *ctx)
> > +{
> > +       __u64 cookie = bpf_get_attach_cookie(ctx);
> > +       __u64 addr = bpf_get_func_ip(ctx);
> > +
> > +       test2_result += (const void *) addr == &bpf_fentry_test1 && cookie == 1;
> > +       test2_result += (const void *) addr == &bpf_fentry_test2 && cookie == 2;
> > +       test2_result += (const void *) addr == &bpf_fentry_test3 && cookie == 3;
> > +       test2_result += (const void *) addr == &bpf_fentry_test4 && cookie == 4;
> > +       test2_result += (const void *) addr == &bpf_fentry_test5 && cookie == 5;
> > +       test2_result += (const void *) addr == &bpf_fentry_test6 && cookie == 6;
> > +       test2_result += (const void *) addr == &bpf_fentry_test7 && cookie == 7;
> > +       test2_result += (const void *) addr == &bpf_fentry_test8 && cookie == 8;
> 
> this is not parallel mode friendly
> 
> let's filter by pid, but also it's best to do count locally and just
> assign it (so that multiple calls of the program still produce the
> same value, instead of constantly increasing global variable with each
> run)

ah I did not think of the paralel run, right, will change

thanks,
jirka