On Tue, 9 Jul 2024 at 20:54, Kumar Kartikeya Dwivedi <memxor@xxxxxxxxx> wrote:
>
> Add a selftest that tries to trigger a situation where two timer
> callbacks are attempting to cancel each other's timer. By running them
> continuously, we hit a condition where both run in parallel and cancel
> each other. Without the fix in the previous patch, this would cause a
> lockup as hrtimer_cancel on either side will wait for forward progress
> from the callback.
>
> Ensure that this situation leads to a EDEADLK error.
>
> Signed-off-by: Kumar Kartikeya Dwivedi <memxor@xxxxxxxxx>
> ---
> .../selftests/bpf/prog_tests/timer_lockup.c | 65 ++++++++++++++
> .../selftests/bpf/progs/timer_lockup.c | 85 +++++++++++++++++++
> 2 files changed, 150 insertions(+)
> create mode 100644 tools/testing/selftests/bpf/prog_tests/timer_lockup.c
> create mode 100644 tools/testing/selftests/bpf/progs/timer_lockup.c
>
> diff --git a/tools/testing/selftests/bpf/prog_tests/timer_lockup.c b/tools/testing/selftests/bpf/prog_tests/timer_lockup.c
> new file mode 100644
> index 000000000000..73e376fc5bbd
> --- /dev/null
> +++ b/tools/testing/selftests/bpf/prog_tests/timer_lockup.c
> @@ -0,0 +1,65 @@
> +// SPDX-License-Identifier: GPL-2.0
> +#define _GNU_SOURCE
> +#include <sched.h>
> +#include <test_progs.h>
> +#include <pthread.h>
> +#include <network_helpers.h>
> +#include "timer_lockup.skel.h"
> +
> +long cpu;
> +int *timer1_err;
> +int *timer2_err;
> +
> +static void *timer_lockup_thread(void *arg)
> +{
> + LIBBPF_OPTS(bpf_test_run_opts, opts,
> + .data_in = &pkt_v4,
> + .data_size_in = sizeof(pkt_v4),
> + .repeat = 10000,
> + );
> + int prog_fd = *(int *)arg;
> + cpu_set_t cpuset;
> +
> + CPU_ZERO(&cpuset);
> + CPU_SET(__sync_fetch_and_add(&cpu, 1), &cpuset);
> + ASSERT_OK(pthread_setaffinity_np(pthread_self(), sizeof(cpuset), &cpuset), "cpu affinity");
> +
> + while (!*timer1_err && !*timer2_err)
> + bpf_prog_test_run_opts(prog_fd, &opts);
> +
> + return NULL;
> +}
> +
> +void test_timer_lockup(void)
> +{
> + struct timer_lockup *skel;
> + pthread_t thrds[2];
> + void *ret;
> +
> + skel = timer_lockup__open_and_load();
> + if (!ASSERT_OK_PTR(skel, "timer_lockup__open_and_load"))
> + return;
> +
> + int timer1_prog = bpf_program__fd(skel->progs.timer1_prog);
> + int timer2_prog = bpf_program__fd(skel->progs.timer2_prog);
> +
> + timer1_err = &skel->bss->timer1_err;
> + timer2_err = &skel->bss->timer2_err;
> +
> + if (!ASSERT_OK(pthread_create(&thrds[0], NULL, timer_lockup_thread, &timer1_prog), "pthread_create thread1"))
> + return;
> + if (!ASSERT_OK(pthread_create(&thrds[1], NULL, timer_lockup_thread, &timer2_prog), "pthread_create thread2")) {
> + pthread_exit(&thrds[0]);
> + return;
A goto out: timer_lockup___destroy(skel) is missing here and above
this. Will wait for a day or so before respinning.
> [...]
