On 29/07/2024 18:47, Alexis Lothoré wrote: > Hello Alan, thanks for the review > > On 7/29/24 19:29, Alan Maguire wrote: >> On 29/07/2024 09:20, Alexis Lothoré (eBPF Foundation) wrote: >>> test_dev_cgroup is defined as a standalone test program, and so is not >>> executed in CI. >>> >>> Convert it to test_progs framework so it is tested automatically in CI, and >>> remove the old test. In order to be able to run it in test_progs, /dev/null >>> must remain usable, so change the new test to test operations on devices >>> 1:3 as valid, and operations on devices 1:5 (/dev/zero) as invalid. >>> >>> Signed-off-by: Alexis Lothoré (eBPF Foundation) <alexis.lothore@xxxxxxxxxxx> >> >> A few small suggestions but looks great! >> >> Reviewed-by: Alan Maguire <alan.maguire@xxxxxxxxxx> > > [...] > >>> + unlink(path); >>> + ret = mknod(path, mode, makedev(dev_major, dev_minor)); >>> + ASSERT_EQ(ret, expected_ret, "mknod"); >> no need to unlink unless "if (!ret)" > > Indeed, you are right. > > [...] > >>> + skel = dev_cgroup__open_and_load(); >>> + if (!ASSERT_OK_PTR(skel, "load program")) >>> + goto cleanup_cgroup; >>> + >>> + if (!ASSERT_OK(bpf_prog_attach(bpf_program__fd(skel->progs.bpf_prog1), >>> + cgroup_fd, BPF_CGROUP_DEVICE, 0), >>> + "attach_program")) >> >> I'd suggest using bpf_program__attach_cgroup() here as you can assign >> the link in the skeleton; see prog_tests/cgroup_v1v2.c. > > Ah yes, thanks for the hint ! > > >>> + goto cleanup_progs; >>> + >>> + if (test__start_subtest("deny-mknod")) >>> + test_mknod("/dev/test_dev_cgroup_zero", S_IFCHR, 1, 5, -EPERM); >>> + >> >> nit: group with other deny subtests. > > ACK > >>> + if (test__start_subtest("allow-mknod")) >>> + test_mknod("/dev/test_dev_cgroup_null", S_IFCHR, 1, 3, 0); >>> + >>> + if (test__start_subtest("allow-read")) >>> + test_read("/dev/urandom", buf, TEST_BUFFER_SIZE, TEST_BUFFER_SIZE); >>> + >> >> Nit: should we have a separate garbage buffer for the successful >> /dev/urandom read? We're not validating buffer contents anywhere but we >> will overwrite our test string I think and it'll end up non-null terminated. > > True, but since the tests aren't performing any string operation on it, is it > really a big deal ? I can even switch the string to a byte array, if it can > prevent any mistake. > There's no need, don't worry. As long as the size limits ensure we don't overrun the buffer, we're good. > If that's ok for you, I can bring all the suggestions discussed here in a new > revision and keep your review tag. > Sounds great, thanks! Alan