Re: [PATCH bpf v2 2/2] bpf: Reject untrusted pointers in refcount_acquire

["Kumar Kartikeya Dwivedi" <memxor@gmail.com>]

On Mon Jul 27, 2026 at 1:50 AM CEST, Ning Ding wrote:
> After bpf_rcu_read_unlock(), a refcounted map kptr is marked
> PTR_UNTRUSTED because it is no longer protected by RCU. The
> refcounted-kptr argument check ignores PTR_UNTRUSTED and still allows
> bpf_refcount_acquire() on that pointer.
>
> However, if another thread removes the object and drops its last reference, the
> stale address can later be reused for another refcounted object. A
> CAP_BPF-only reproducer observed bpf_refcount_acquire() returning
> non-NULL through such a stale pointer.
>
> Reject PTR_UNTRUSTED refcounted-kptr arguments and add a regression
> test.
>
> Fixes: 7c50b1cb76ac ("bpf: Add bpf_refcount_acquire kfunc")
> Reported-by: sashiko-bot@kernel.org
> Link: https://lore.kernel.org/r/20260726021304.97ED91F000E9@smtp.kernel.org
> Assisted-by: Codex:gpt-5
> Signed-off-by: Ning Ding <dingning04@gmail.com>
> ---

Same comment as the new set; split kernel commits and selftest commits into
separate ones. As for the fix, I think it would make more sense if
type_is_ptr_alloc_obj() was fixed to PTR_UNTRUSTED by definition, instead of
having to add extra checks on top.

pw-bot: cr

>  kernel/bpf/verifier.c                         |  5 ++++
>  .../bpf/progs/refcounted_kptr_fail.c          | 27 +++++++++++++++++++
>  2 files changed, 32 insertions(+)
>
> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
> index 1e7343b625de..63b1d997fe80 100644
> --- a/kernel/bpf/verifier.c
> +++ b/kernel/bpf/verifier.c
> @@ -12414,6 +12414,11 @@ static int check_kfunc_args(struct bpf_verifier_env *env, struct bpf_kfunc_call_
>  			meta->subprogno = reg->subprogno;
>  			break;
>  		case KF_ARG_PTR_TO_REFCOUNTED_KPTR:
> +			if (reg->type & PTR_UNTRUSTED) {
> +				verbose(env, "%s is an untrusted refcounted kptr\n",
> +					reg_arg_name(env, argno));
> +				return -EACCES;
> +			}
>  			if (!type_is_ptr_alloc_obj(reg->type)) {
>  				verbose(env, "%s is neither owning or non-owning ref\n",
>  					reg_arg_name(env, argno));
> diff --git a/tools/testing/selftests/bpf/progs/refcounted_kptr_fail.c b/tools/testing/selftests/bpf/progs/refcounted_kptr_fail.c
> index acd3e81a3916..80b92d7ec9ca 100644
> --- a/tools/testing/selftests/bpf/progs/refcounted_kptr_fail.c
> +++ b/tools/testing/selftests/bpf/progs/refcounted_kptr_fail.c
> @@ -127,6 +127,33 @@ long refcount_acquire_rcu_map_kptr_unchecked_drop(void *ctx)
>  	return 0;
>  }
>
> +SEC("?tc")
> +__failure __msg("is an untrusted refcounted kptr")
> +long refcount_acquire_after_rcu_unlock(void *ctx)
> +{
> +	struct map_value_refcount_only *mapval;
> +	struct node_refcount_only *n, *m;
> +	int idx = 0;
> +
> +	mapval = bpf_map_lookup_elem(&stashed_refcount_only, &idx);
> +	if (!mapval)
> +		return 1;
> +
> +	bpf_rcu_read_lock();
> +	n = mapval->node;
> +	if (!n) {
> +		bpf_rcu_read_unlock();
> +		return 2;
> +	}
> +	bpf_rcu_read_unlock();
> +
> +	m = bpf_refcount_acquire(n);
> +	if (m)
> +		bpf_obj_drop(m);
> +
> +	return 0;
> +}
> +
>  SEC("?tc")
>  __failure __msg("Unreleased reference id=3 alloc_insn={{[0-9]+}}")
>  long rbtree_refcounted_node_ref_escapes_owning_input(void *ctx)