| Message ID | 20180606161244.2649812-1-yhs@fb.com |
|---|---|
| State | Accepted, archived |
| Delegated to: | BPF Maintainers |
| Headers | show |
| Series | [bpf] tools/bpf: fix selftest get_cgroup_id_user | expand |
On 06/06/2018 06:12 PM, Yonghong Song wrote: > Commit f269099a7e7a ("tools/bpf: add a selftest for > bpf_get_current_cgroup_id() helper") added a test > for bpf_get_current_cgroup_id() helper. The bpf program > is attached to tracepoint syscalls/sys_enter_nanosleep > and will record the cgroup id if the tracepoint is hit. > The test program creates a cgroup and attachs itself to > this cgroup and expects that the test program process > cgroup id is the same as the cgroup_id retrieved > by the bpf program. > > In a light system where no other processes called > nanosleep syscall, the test case can pass. > In a busy system where many different processes can hit > syscalls/sys_enter_nanosleep tracepoint, the cgroup id > recorded by bpf program may not match the test program > process cgroup_id. > > This patch fixed an issue by communicating the test program > pid to bpf program. The bpf program only records > cgroup id if the current task pid is the same as > passed-in pid. This ensures that the recorded cgroup_id > is for the cgroup within which the test program resides. > > Fixes: f269099a7e7a ("tools/bpf: add a selftest for bpf_get_current_cgroup_id() helper") > Signed-off-by: Yonghong Song <yhs@fb.com> Applied to bpf, thanks Yonghong!
diff --git a/tools/testing/selftests/bpf/get_cgroup_id_kern.c b/tools/testing/selftests/bpf/get_cgroup_id_kern.c index 2cf8cb2..014dba1 100644 --- a/tools/testing/selftests/bpf/get_cgroup_id_kern.c +++ b/tools/testing/selftests/bpf/get_cgroup_id_kern.c @@ -11,12 +11,24 @@ struct bpf_map_def SEC("maps") cg_ids = { .max_entries = 1, }; +struct bpf_map_def SEC("maps") pidmap = { + .type = BPF_MAP_TYPE_ARRAY, + .key_size = sizeof(__u32), + .value_size = sizeof(__u32), + .max_entries = 1, +}; + SEC("tracepoint/syscalls/sys_enter_nanosleep") int trace(void *ctx) { - __u32 key = 0; + __u32 pid = bpf_get_current_pid_tgid(); + __u32 key = 0, *expected_pid; __u64 *val; + expected_pid = bpf_map_lookup_elem(&pidmap, &key); + if (!expected_pid || *expected_pid != pid) + return 0; + val = bpf_map_lookup_elem(&cg_ids, &key); if (val) *val = bpf_get_current_cgroup_id(); diff --git a/tools/testing/selftests/bpf/get_cgroup_id_user.c b/tools/testing/selftests/bpf/get_cgroup_id_user.c index ea19a42..e8da7b3 100644 --- a/tools/testing/selftests/bpf/get_cgroup_id_user.c +++ b/tools/testing/selftests/bpf/get_cgroup_id_user.c @@ -50,13 +50,13 @@ int main(int argc, char **argv) const char *probe_name = "syscalls/sys_enter_nanosleep"; const char *file = "get_cgroup_id_kern.o"; int err, bytes, efd, prog_fd, pmu_fd; + int cgroup_fd, cgidmap_fd, pidmap_fd; struct perf_event_attr attr = {}; - int cgroup_fd, cgidmap_fd; struct bpf_object *obj; __u64 kcgid = 0, ucgid; + __u32 key = 0, pid; int exit_code = 1; char buf[256]; - __u32 key = 0; err = setup_cgroup_environment(); if (CHECK(err, "setup_cgroup_environment", "err %d errno %d\n", err, @@ -81,6 +81,14 @@ int main(int argc, char **argv) cgidmap_fd, errno)) goto close_prog; + pidmap_fd = bpf_find_map(__func__, obj, "pidmap"); + if (CHECK(pidmap_fd < 0, "bpf_find_map", "err %d errno %d\n", + pidmap_fd, errno)) + goto close_prog; + + pid = getpid(); + bpf_map_update_elem(pidmap_fd, &key, &pid, 0); + snprintf(buf, sizeof(buf), "/sys/kernel/debug/tracing/events/%s/id", probe_name); efd = open(buf, O_RDONLY, 0);
Commit f269099a7e7a ("tools/bpf: add a selftest for bpf_get_current_cgroup_id() helper") added a test for bpf_get_current_cgroup_id() helper. The bpf program is attached to tracepoint syscalls/sys_enter_nanosleep and will record the cgroup id if the tracepoint is hit. The test program creates a cgroup and attachs itself to this cgroup and expects that the test program process cgroup id is the same as the cgroup_id retrieved by the bpf program. In a light system where no other processes called nanosleep syscall, the test case can pass. In a busy system where many different processes can hit syscalls/sys_enter_nanosleep tracepoint, the cgroup id recorded by bpf program may not match the test program process cgroup_id. This patch fixed an issue by communicating the test program pid to bpf program. The bpf program only records cgroup id if the current task pid is the same as passed-in pid. This ensures that the recorded cgroup_id is for the cgroup within which the test program resides. Fixes: f269099a7e7a ("tools/bpf: add a selftest for bpf_get_current_cgroup_id() helper") Signed-off-by: Yonghong Song <yhs@fb.com> --- tools/testing/selftests/bpf/get_cgroup_id_kern.c | 14 +++++++++++++- tools/testing/selftests/bpf/get_cgroup_id_user.c | 12 ++++++++++-- 2 files changed, 23 insertions(+), 3 deletions(-)