linux.git/tools/testing/selftests/bpf/progs/struct_ops_assoc_in_timer.c, branch master Linux kernel source tree selftests/bpf: Migrate struct_ops_assoc test to KF_IMPLICIT_ARGS 2026-01-21T00:22:38+00:00 Ihor Solodrai ihor.solodrai@linux.dev 2026-01-20T22:26:36+00:00 bd06b977e02d80fe0dec303cc219d007121a4526 A test kfunc named bpf_kfunc_multi_st_ops_test_1_impl() is a user of __prog suffix. Subsequent patch removes __prog support in favor of KF_IMPLICIT_ARGS, so migrate this kfunc to use implicit argument. Signed-off-by: Ihor Solodrai <ihor.solodrai@linux.dev> Link: https://lore.kernel.org/r/20260120222638.3976562-12-ihor.solodrai@linux.dev Signed-off-by: Alexei Starovoitov <ast@kernel.org> 
A test kfunc named bpf_kfunc_multi_st_ops_test_1_impl() is a user of
__prog suffix. Subsequent patch removes __prog support in favor of
KF_IMPLICIT_ARGS, so migrate this kfunc to use implicit argument.

Signed-off-by: Ihor Solodrai <ihor.solodrai@linux.dev>
Link: https://lore.kernel.org/r/20260120222638.3976562-12-ihor.solodrai@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
selftests/bpf: Test getting associated struct_ops in timer callback 2025-12-06T00:17:58+00:00 Amery Hung ameryhung@gmail.com 2025-12-03T23:37:48+00:00 0e841d19263ab6e1ca2b280109832f57624e48d1 Make sure 1) a timer callback can also reference the associated struct_ops, and then make sure 2) the timer callback cannot get a dangled pointer to the struct_ops when the map is freed. The test schedules a timer callback from a struct_ops program since struct_ops programs do not pin the map. It is possible for the timer callback to run after the map is freed. The timer callback calls a kfunc that runs .test_1() of the associated struct_ops, which should return MAP_MAGIC when the map is still alive or -1 when the map is gone. The first subtest added in this patch schedules the timer callback to run immediately, while the map is still alive. The second subtest added schedules the callback to run 500ms after syscall_prog runs and then frees the map right after syscall_prog runs. Both subtests then wait until the callback runs to check the return of the kfunc. Signed-off-by: Amery Hung <ameryhung@gmail.com> Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Link: https://lore.kernel.org/bpf/20251203233748.668365-7-ameryhung@gmail.com 
Make sure 1) a timer callback can also reference the associated
struct_ops, and then make sure 2) the timer callback cannot get a
dangled pointer to the struct_ops when the map is freed.

The test schedules a timer callback from a struct_ops program since
struct_ops programs do not pin the map. It is possible for the timer
callback to run after the map is freed. The timer callback calls a
kfunc that runs .test_1() of the associated struct_ops, which should
return MAP_MAGIC when the map is still alive or -1 when the map is
gone.

The first subtest added in this patch schedules the timer callback to
run immediately, while the map is still alive. The second subtest added
schedules the callback to run 500ms after syscall_prog runs and then
frees the map right after syscall_prog runs. Both subtests then wait
until the callback runs to check the return of the kfunc.

Signed-off-by: Amery Hung <ameryhung@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20251203233748.668365-7-ameryhung@gmail.com