On Sat, May 14, 2022 at 11:31 PM Yuntao Wang ytcoode@gmail.com wrote:
Currently the trampoline_count test doesn't include any fmod_ret bpf programs, fix it to make the test cover all possible trampoline program types.
Since fmod_ret bpf programs can't be attached to __set_task_comm function, as it's neither whitelisted for error injection nor a security hook, change it to bpf_modify_return_test.
This patch also does some other cleanups such as removing duplicate code, dropping inconsistent comments, etc.
Signed-off-by: Yuntao Wang ytcoode@gmail.com
include/linux/bpf.h | 2 +- .../bpf/prog_tests/trampoline_count.c | 121 ++++++------------ .../bpf/progs/test_trampoline_count.c | 16 ++- 3 files changed, 47 insertions(+), 92 deletions(-)
[...]
/* with E2BIG error */
ASSERT_EQ(err, -E2BIG, "proper error check");
ASSERT_EQ(libbpf_get_error(link), -E2BIG, "E2BIG"); ASSERT_EQ(link, NULL, "ptr_is_null");
/* and finaly execute the probe */if (CHECK_FAIL(prctl(PR_GET_NAME, comm, 0L, 0L, 0L)))goto cleanup_extra;CHECK_FAIL(test_task_rename());
we stopped testing that kernel function actually can be called properly, why don't you do bpf_prog_test_run() here to trigger bpf_modify_return_test in kernel?
CHECK_FAIL(prctl(PR_SET_NAME, comm, 0L, 0L, 0L));-cleanup_extra:
bpf_object__close(obj);cleanup:
if (i >= MAX_TRAMP_PROGS)i = MAX_TRAMP_PROGS - 1; for (; i >= 0; i--) {bpf_link__destroy(inst[i].link_fentry);bpf_link__destroy(inst[i].link_fexit);
bpf_link__destroy(inst[i].link); bpf_object__close(inst[i].obj); }}
[...]