Re: [PATCH bpf-next v2 11/16] bpf, verifier: Remove side effects from may_access_direct_pkt_data
From: Eduard Zingerman <eddyz87@gmail.com>
Date: 2026-01-05 18:23:07
Also in:
bpf
On Mon, 2026-01-05 at 13:14 +0100, Jakub Sitnicki wrote:
The may_access_direct_pkt_data() helper sets env->seen_direct_write as a side effect, which creates awkward calling patterns: - check_special_kfunc() has a comment warning readers about the side effect - specialize_kfunc() must save and restore the flag around the call Make the helper a pure function by moving the seen_direct_write flag setting to call sites that need it. Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com> ---
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
quoted hunk ↗ jump to hunk
kernel/bpf/verifier.c | 33 ++++++++++++--------------------- 1 file changed, 12 insertions(+), 21 deletions(-)diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c index 9394b0de2ef0..52d76a848f65 100644 --- a/kernel/bpf/verifier.c +++ b/kernel/bpf/verifier.c@@ -6151,13 +6151,9 @@ static bool may_access_direct_pkt_data(struct bpf_verifier_env *env, if (meta) return meta->pkt_access; - env->seen_direct_write = true;
Note to reviewers: the call to may_access_direct_pkt_data() in check_func_arg() always has a non-NULL 'meta', so it is correct to skip setting 'env->seen_direct_write' there, behavior does not change.
return true; case BPF_PROG_TYPE_CGROUP_SOCKOPT: - if (t == BPF_WRITE) - env->seen_direct_write = true; - return true; default:
[...]