BPF List
 help / color / mirror / Atom feed
From: "Alexei Starovoitov" <alexei.starovoitov@gmail.com>
To: "Justin Suess" <utilityemal77@gmail.com>,
	"Daniel Borkmann" <daniel@iogearbox.net>,
	"Andrii Nakryiko" <andrii@kernel.org>,
	"Eduard Zingerman" <eddyz87@gmail.com>,
	"Kumar Kartikeya Dwivedi" <memxor@gmail.com>
Cc: "Song Liu" <song@kernel.org>,
	"Puranjay Mohan" <puranjay@kernel.org>, <bpf@vger.kernel.org>,
	<stable@vger.kernel.org>
Subject: Re: [PATCH bpf-next v2 1/6] bpf: Carry MEM_RCU for trusted fields with MEM_RCU parents
Date: Wed, 21 Oct 2026 00:37:58 +0000	[thread overview]
Message-ID: <DMA3HLS1N8FF.9YVQJB6RM2A0@gmail.com> (raw)
In-Reply-To: <20261020145059.2146170-2-utilityemal77@gmail.com>

On Tue, Oct 20, 2026 at 07:50 AM Justin Suess <utilityemal77@gmail.com> wrote:
> An RCU-protected parent also guarantees less about the field's value
> than a trusted one. __safe_trusted promises a non-NULL field for as
> long as the parent object is valid, but RCU only keeps the parent's
> memory valid: struct file is allocated SLAB_TYPESAFE_BY_RCU, so a file
> read under RCU can be freed and reused within the same CS.

It guarantees less about the field's lifetime too. A MEM_RCU parent
only means that its memory is not freed yet, so the field is safe
only if the object it points to is RCU freed itself.
inode is. mm_struct is not, __mmdrop() does kmem_cache_free() without
a grace period. With this patch the first MEM_RCU vma hands out
vma->vm_mm as rcu_ptr_mm_struct.

That's what the __safe_rcu_or_null list is for. Let's put f_inode
there and not apply the trusted lists to a parent that is not trusted.
exe_file->f_inode is the same rcu_ptr_or_null_inode, so the tests in
patch 2 don't change, and the ternaries go away.

Something like this (compile tested only):

diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index 01434e1..219d027 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -6476,6 +6476,11 @@ BTF_TYPE_SAFE_RCU_OR_NULL(struct request_sock) {
 	struct sock *sk;
 };

+/* struct file is SLAB_TYPESAFE_BY_RCU. A reused file has f_inode == NULL. */
+BTF_TYPE_SAFE_RCU_OR_NULL(struct file) {
+	struct inode *f_inode;
+};
+
 /* full trusted: these fields are trusted even outside of RCU CS and never NULL */
 BTF_TYPE_SAFE_TRUSTED(struct bpf_iter_meta) {
 	struct seq_file *seq;
@@ -6552,6 +6557,7 @@ static bool type_is_rcu_or_null(struct bpf_verifier_env *env,
 	BTF_TYPE_EMIT(BTF_TYPE_SAFE_RCU_OR_NULL(struct mm_struct));
 	BTF_TYPE_EMIT(BTF_TYPE_SAFE_RCU_OR_NULL(struct sk_buff));
 	BTF_TYPE_EMIT(BTF_TYPE_SAFE_RCU_OR_NULL(struct request_sock));
+	BTF_TYPE_EMIT(BTF_TYPE_SAFE_RCU_OR_NULL(struct file));

 	return btf_nested_type_is_trusted(&env->log, reg, field_name, btf_id, "__safe_rcu_or_null");
 }
@@ -6757,10 +6763,14 @@ static int check_ptr_to_btf_access(struct bpf_verifier_env *env,
 		 *
 		 * A regular RCU-protected pointer with __rcu tag can also be deemed
 		 * trusted if we are in an RCU CS. Such pointer can be NULL.
+		 *
+		 * If the parent is not trusted, the field can't regain trusted status.
 		 */
-		if (type_is_trusted(env, reg, field_name, btf_id)) {
+		bool trusted = is_trusted_reg(env, reg);
+
+		if (trusted && type_is_trusted(env, reg, field_name, btf_id)) {
 			flag |= PTR_TRUSTED;
-		} else if (type_is_trusted_or_null(env, reg, field_name, btf_id)) {
+		} else if (trusted && type_is_trusted_or_null(env, reg, field_name, btf_id)) {
 			flag |= PTR_TRUSTED | PTR_MAYBE_NULL;
 		} else if (in_rcu_cs(env) && !type_may_be_null(reg->type)) {
 			if (type_is_rcu(env, reg, field_name, btf_id)) {

[...]

> Cc: stable@vger.kernel.org

That shouldn't go to stable.

pw-bot: cr

  reply	other threads:[~2026-10-21  0:38 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-10-20 14:50 [PATCH bpf-next v2 0/6] bpf: bound trusted fields to their parent's lifetime Justin Suess
2026-10-20 14:50 ` [PATCH bpf-next v2 1/6] bpf: Carry MEM_RCU for trusted fields with MEM_RCU parents Justin Suess
2026-10-21  0:37   ` Alexei Starovoitov [this message]
2026-10-20 14:50 ` [PATCH bpf-next v2 2/6] selftests/bpf: Test MEM_RCU carry-over for trusted fields Justin Suess
2026-10-20 14:50 ` [PATCH bpf-next v2 3/6] bpf: Bind trusted fields to their parent's reference Justin Suess
2026-10-21  1:16   ` Kumar Kartikeya Dwivedi
2026-10-20 14:50 ` [PATCH bpf-next v2 4/6] selftests/bpf: Test trusted-field invalidation on parent release Justin Suess
2026-10-20 14:50 ` [PATCH bpf-next v2 5/6] bpf: Bind iterator elements to their iterator Justin Suess
2026-10-20 15:25   ` bot+bpf-ci
2026-10-20 15:29   ` sashiko-bot
2026-10-21  0:57   ` Kumar Kartikeya Dwivedi
2026-10-20 14:50 ` [PATCH bpf-next v2 6/6] selftests/bpf: Test iterator element invalidation on destroy Justin Suess

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=DMA3HLS1N8FF.9YVQJB6RM2A0@gmail.com \
    --to=alexei.starovoitov@gmail.com \
    --cc=andrii@kernel.org \
    --cc=bpf@vger.kernel.org \
    --cc=daniel@iogearbox.net \
    --cc=eddyz87@gmail.com \
    --cc=memxor@gmail.com \
    --cc=puranjay@kernel.org \
    --cc=song@kernel.org \
    --cc=stable@vger.kernel.org \
    --cc=utilityemal77@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox