Re: [PATCH bpf-next v13 5/6] selftests/bpf: add usage example for cpu time counter kfuncs

[sashiko-bot@kernel.org]

Sashiko AI review found 1 potential issue(s):
- [Medium] Incorrect and redundant file descriptor validation in BPF skeleton test.
--

commit 0a7806f6db918db501677b21fa0e15ae508d1ada
Author: Vadim Fedorenko <vadim.fedorenko@linux.dev>

selftests/bpf: add usage example for cpu time counter kfuncs

This commit adds a selftest example demonstrating how to measure the
latency of BPF kfuncs or helpers using the time stamp counter and how
to convert the measured values into nanoseconds.

> diff --git a/tools/testing/selftests/bpf/prog_tests/test_cpu_cycles.c b/tools/testing/selftests/bpf/prog_tests/test_cpu_cycles.c
> new file mode 100644
> index 0000000000000..067307f0c4c29
> --- /dev/null
> +++ b/tools/testing/selftests/bpf/prog_tests/test_cpu_cycles.c
[ ... ]
> +static void cpu_cycles(void)
> +{
> +	LIBBPF_OPTS(bpf_test_run_opts, opts);
> +	struct test_cpu_cycles *skel;
> +	int err, pfd;
> +
> +	skel = test_cpu_cycles__open_and_load();
> +	if (!ASSERT_OK_PTR(skel, "test_cpu_cycles open and load"))
> +		return;
> +
> +	pfd = bpf_program__fd(skel->progs.bpf_cpu_cycles);
> +	if (!ASSERT_GT(pfd, 0, "test_cpu_cycles fd"))
> +		goto fail;

Is it necessary to validate the file descriptor here? The BPF skeleton API
guarantees that after a successful open_and_load(), all programs and maps
are created with valid file descriptors, making this validation redundant.

Additionally, if a file descriptor check were needed, shouldn't it use
ASSERT_GE rather than ASSERT_GT? 0 is technically a valid file descriptor,
so if pfd evaluated to 0, this test would falsely fail.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260418131614.1501848-1-puranjay@kernel.org?part=5