On Wed, Nov 23, 2022 at 11:07 AM Yonghong Song <yhs@xxxxxxxx> wrote:
>
>
>
> On 11/23/22 9:17 AM, sdf@xxxxxxxxxx wrote:
> > On 11/23, Yonghong Song wrote:
> >
> >
> >> On 11/23/22 4:23 AM, Jiri Olsa wrote:
> >> > On Mon, Nov 21, 2022 at 10:03:40AM -0800, Stanislav Fomichev wrote:
> >> > > LWT_XMIT to test L3 case, TC to test L2 case.
> >> > >
> >> > > v2:
> >> > > - s/veth_ifindex/ipip_ifindex/ in two places (Martin)
> >> > > - add comment about which condition triggers the rejection (Martin)
> >> > >
> >> > > Signed-off-by: Stanislav Fomichev <sdf@xxxxxxxxxx>
> >> >
> >> > hi,
> >> > I'm getting selftest fails and it looks like it's because of this test:
> >> >
> >> > [root@qemu bpf]# ./test_progs -n 62,98
> >> > #62 empty_skb:OK
> >> > execute_one_variant:PASS:skel_open 0 nsec
> >> > execute_one_variant:PASS:my_pid_map_update 0 nsec
> >> > libbpf: failed to determine tracepoint 'raw_syscalls/sys_enter'
> >> perf event ID: No such file or directory
> >> > libbpf: prog 'handle_legacy': failed to create tracepoint
> >> 'raw_syscalls/sys_enter' perf event: No such file or directory
> >> > libbpf: prog 'handle_legacy': failed to auto-attach: -2
> >> > execute_one_variant:FAIL:skel_attach unexpected error: -2 (errno 2)
> >> > test_legacy_printk:FAIL:legacy_case unexpected error: -2 (errno 2)
> >> > execute_one_variant:PASS:skel_open 0 nsec
> >> > libbpf: failed to determine tracepoint 'raw_syscalls/sys_enter'
> >> perf event ID: No such file or directory
> >> > libbpf: prog 'handle_modern': failed to create tracepoint
> >> 'raw_syscalls/sys_enter' perf event: No such file or directory
> >> > libbpf: prog 'handle_modern': failed to auto-attach: -2
> >> > execute_one_variant:FAIL:skel_attach unexpected error: -2 (errno 2)
> >> > #98 legacy_printk:FAIL
> >> >
> >> > All error logs:
> >> > execute_one_variant:PASS:skel_open 0 nsec
> >> > execute_one_variant:PASS:my_pid_map_update 0 nsec
> >> > libbpf: failed to determine tracepoint 'raw_syscalls/sys_enter'
> >> perf event ID: No such file or directory
> >> > libbpf: prog 'handle_legacy': failed to create tracepoint
> >> 'raw_syscalls/sys_enter' perf event: No such file or directory
> >> > libbpf: prog 'handle_legacy': failed to auto-attach: -2
> >> > execute_one_variant:FAIL:skel_attach unexpected error: -2 (errno 2)
> >> > test_legacy_printk:FAIL:legacy_case unexpected error: -2 (errno 2)
> >> > execute_one_variant:PASS:skel_open 0 nsec
> >> > libbpf: failed to determine tracepoint 'raw_syscalls/sys_enter'
> >> perf event ID: No such file or directory
> >> > libbpf: prog 'handle_modern': failed to create tracepoint
> >> 'raw_syscalls/sys_enter' perf event: No such file or directory
> >> > libbpf: prog 'handle_modern': failed to auto-attach: -2
> >> > execute_one_variant:FAIL:skel_attach unexpected error: -2 (errno 2)
> >> > #98 legacy_printk:FAIL
> >> > Summary: 1/0 PASSED, 0 SKIPPED, 1 FAILED
> >> >
> >> > when I run separately it passes:
> >> >
> >> > [root@qemu bpf]# ./test_progs -n 98
> >> > #98 legacy_printk:OK
> >> > Summary: 1/0 PASSED, 0 SKIPPED, 0 FAILED
> >> >
> >> >
> >> > it seems that the open_netns/close_netns does not work properly,
> >> > and screw up access to tracefs for following tests
> >> >
> >> > if I comment out all the umounts in setns_by_fd, it does not fail
> >
> >> Agreed with the above observations.
> >> With the current bpf-next, I can easily hit the above perf event ID
> >> issue.
> >
> >> But if I backout the following two patches:
> >> 68f8e3d4b916531ea3bb8b83e35138cf78f2fce5 selftests/bpf: Make sure
> >> zero-len
> >> skbs aren't redirectable
> >> 114039b342014680911c35bd6b72624180fd669a bpf: Move skb->len == 0
> >> checks into
> >> __bpf_redirect
> >
> >
> >> and run a few times with './test_progs -j' and I didn't hit any issues.
> >
> > My guess would be that we need to remount debugfs in setns_by_fd?
> >
> > diff --git a/tools/testing/selftests/bpf/network_helpers.c
> > b/tools/testing/selftests/bpf/network_helpers.c
> > index bec15558fd93..1f37adff7632 100644
> > --- a/tools/testing/selftests/bpf/network_helpers.c
> > +++ b/tools/testing/selftests/bpf/network_helpers.c
> > @@ -426,6 +426,10 @@ static int setns_by_fd(int nsfd)
> > if (!ASSERT_OK(err, "mount /sys/fs/bpf"))
> > return err;
> >
> > + err = mount("debugfs", "/sys/kernel/debug", "debugfs", 0, NULL);
> > + if (!ASSERT_OK(err, "mount /sys/kernel/debug"))
> > + return err;
> > +
> > return 0;
> > }
>
> Ya, this does fix the problem. Could you craft a patch for this?
Sure, give me a second..
