Re: [PATCH bpf-next v5 2/4] selftests/bpf: Add tests for bpf_call_rcu()

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] bpf: selftest sleepable_callback fails to properly test callback restriction
--

commit 8dcd9155da335ebebbf25f7c26266814bb232e74
Author: Puranjay Mohan <puranjay@kernel.org>

selftests/bpf: Add tests for bpf_call_rcu()
    
This commit adds selftests to cover bpf_call_rcu() functionality, checking
grace periods, callback re-arming, teardown handling, and negative tests
including verifier rejection of a sleepable callback.

> diff --git a/tools/testing/selftests/bpf/progs/call_rcu_fail.c b/tools/testing/selftests/bpf/progs/call_rcu_fail.c
> new file mode 100644
> index 0000000000000..bf60a27731fb3
> --- /dev/null
> +++ b/tools/testing/selftests/bpf/progs/call_rcu_fail.c

[ ... ]

> +SEC("syscall")
> +__failure __msg("sleepable helper bpf_copy_from_user") __msg("in non-sleepable prog")
> +int sleepable_callback(void *ctx)
> +{

[Severity: Medium]
Does SEC("syscall") cause this test to pass for the wrong reason?

Since the main program is marked as non-sleepable, the verifier will reject
the sleepable helper regardless of the callback context. Should this be
changed to SEC("syscall.s") so the main program is sleepable, which would
properly force the verifier to reject the helper purely because it is inside
the non-sleepable callback context?

> +	struct elem *e;
> +
> +	e = bpf_map_lookup_elem(&arr, &key);
> +	if (!e)
> +		return 0;
> +	bpf_call_rcu(&e->rh, &arr, sleepable_reclaim);
> +	return 0;
> +}

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260921191407.1742386-1-puranjay@kernel.org?part=2