Received-SPF: pass (google.com: domain of linux-kernel+bounces-2593-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
From: Siddh Raman Pant <code@siddh.me>
To: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>,
	"David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	Jakub Kicinski <kuba@kernel.org>,
	Paolo Abeni <pabeni@redhat.com>,
	Suman Ghosh <sumang@marvell.com>
Cc: netdev@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: [PATCH net-next v6 0/2] nfc: Fix UAF during datagram sending caused by missing refcounting
Date: Sun, 17 Dec 2023 18:41:02 +0530
Message-ID: <cover.1702816635.git.code@siddh.me>
Precedence: bulk
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit

Changes in v6:
- Revert label introduction from v4, and thus also v5 entirely.

Changes in v5:
- Move reason = LLCP_DM_REJ under the fail_put_sock label.
- Checkpatch now warns about == NULL check for new_sk, so fix that,
  and also at other similar places in the same function.

Changes in v4:
- Fix put ordering and comments.
- Separate freeing in recv() into end labels.
- Remove obvious comment and add reasoning.
- Picked up r-bs by Suman.

Changes in v3:
- Fix missing freeing statements.

Changes in v2:
- Add net-next in patch subject.
- Removed unnecessary extra lock and hold nfc_dev ref when holding llcp_sock.
- Remove last formatting patch.
- Picked up r-b from Krzysztof for LLCP_BOUND patch.

---

For connectionless transmission, llcp_sock_sendmsg() codepath will
eventually call nfc_alloc_send_skb() which takes in an nfc_dev as
an argument for calculating the total size for skb allocation.

virtual_ncidev_close() codepath eventually releases socket by calling
nfc_llcp_socket_release() (which sets the sk->sk_state to LLCP_CLOSED)
and afterwards the nfc_dev will be eventually freed.

When an ndev gets freed, llcp_sock_sendmsg() will result in an
use-after-free as it

(1) doesn't have any checks in place for avoiding the datagram sending.

(2) calls nfc_llcp_send_ui_frame(), which also has a do-while loop
    which can race with freeing. This loop contains the call to
    nfc_alloc_send_skb() where we dereference the nfc_dev pointer.

nfc_dev is being freed because we do not hold a reference to it when
we hold a reference to llcp_local. Thus, virtual_ncidev_close()
eventually calls nfc_release() due to refcount going to 0.

Since state has to be LLCP_BOUND for datagram sending, we can bail out
early in llcp_sock_sendmsg().

Please review and let me know if any errors are there, and hopefully
this gets accepted.

Thanks,
Siddh

Siddh Raman Pant (2):
  nfc: llcp_core: Hold a ref to llcp_local->dev when holding a ref to
    llcp_local
  nfc: Do not send datagram if socket state isn't LLCP_BOUND

 net/nfc/llcp_core.c | 40 +++++++++++++++++++++++++++++++++++++---
 net/nfc/llcp_sock.c |  5 +++++
 2 files changed, 42 insertions(+), 3 deletions(-)

-- 
2.42.0