selftests/bpf: Fix possible hang in sockopt_inherit
authorStanislav Fomichev <sdf@google.com>
Wed, 15 Jul 2020 22:41:07 +0000 (15:41 -0700)
committerDaniel Borkmann <daniel@iogearbox.net>
Thu, 16 Jul 2020 18:57:09 +0000 (20:57 +0200)
commite81e7a533742c30615f8b15390df6525cef96778
treeecf1d3c7989e38766aab4b931daefdd8062d833c
parentde40a8abf07bfa125b7496b20ce6f92c6370eecd
selftests/bpf: Fix possible hang in sockopt_inherit

Andrii reported that sockopt_inherit occasionally hangs up on 5.5 kernel [0].
This can happen if server_thread runs faster than the main thread.
In that case, pthread_cond_wait will wait forever because
pthread_cond_signal was executed before the main thread was blocking.
Let's move pthread_mutex_lock up a bit to make sure server_thread
runs strictly after the main thread goes to sleep.

(Not sure why this is 5.5 specific, maybe scheduling is less
deterministic? But I was able to confirm that it does indeed
happen in a VM.)

[0] https://lore.kernel.org/bpf/CAEf4BzY0-bVNHmCkMFPgObs=isUAyg-dFzGDY7QWYkmm7rmTSg@mail.gmail.com/

Reported-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/20200715224107.3591967-1-sdf@google.com
tools/testing/selftests/bpf/prog_tests/sockopt_inherit.c