Re: [PATCH bpf-next v3 12/17] bpf: Report Execution Context Safety errors

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Incorrect diagnostic constraint explanation for global function calls under locks
--

--- Patch [12]: [PATCH bpf-next v3 12/17] bpf: Report Execution Context Safety errors ---
commit 4c2714bc706956dcdd83f54e5db8f96839c37f55
Author: Kumar Kartikeya Dwivedi <memxor@gmail.com>

bpf: Report Execution Context Safety errors

This commit augments selected sleepability and critical-section failures
with Execution Context Safety reports. It adds source context, path history,
and suggestions tied to the active context while keeping existing verifier
messages.

> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
> index 07e2c1ed8fbe6..e253fbc733a48 100644
> --- a/kernel/bpf/verifier.c
> +++ b/kernel/bpf/verifier.c
> @@ -10109,17 +10111,32 @@ static int check_func_call(struct bpf_verifier_env *env, struct bpf_insn *insn,
>  	if (err == -EFAULT)
>  		return err;
>  	if (bpf_subprog_is_global(env, subprog)) {
> +		const char *context;
>  		const char *sub_name = bpf_verifier_subprog_name(env, subprog);
> +		const char *operation;
>  
>  		if (env->cur_state->active_locks) {
>  			verbose(env, "global function calls are not allowed while holding a lock,\n"
>  				     "use static function instead\n");
> +			operation =
> +				bpf_diag_scratch_printf(env, 1, "global function %s()", sub_name);
> +			bpf_diag_ctx(env, BPF_DIAG_CTX_FORBIDDEN, *insn_idx, operation,
> +				     BPF_DIAG_CONTEXT_LOCK, "lock region",
> +				     "Release the lock before calling the global function, or use "
> +				     "a static function instead.");

[Severity: Medium]
Will this result in an incorrect diagnostic reason being printed?

Since BPF_DIAG_CTX_FORBIDDEN triggers diag_ctx_forbidden(), which in turn
uses diag_context_constraint() for the constraint explanation:

kernel/bpf/diagnostics.c:diag_context_constraint() {
	...
	case BPF_DIAG_CONTEXT_LOCK:
		return "code holding a BPF spin lock cannot call operations that may sleep";
	...
}

Doesn't this incorrectly tell the developer that their global function call
was rejected because it "may sleep"? Global function calls are generally
disallowed under locks due to verifier state limitations rather than just
sleepability.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260713153910.2556007-1-memxor@gmail.com?part=12