From: Luiz Augusto von Dentz <luiz.dentz@gmail.com>
To: linux-bluetooth@vger.kernel.org
Subject: [PATCH v1] Bluetooth: L2CAP: Hold incoming data until encryption key size is known
Date: Thu, 22 Oct 2026 11:46:21 -0400 [thread overview]
Message-ID: <20261022154621.3958803-2-luiz.dentz@gmail.com> (raw)
In-Reply-To: <20261022154621.3958803-1-luiz.dentz@gmail.com>
From: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Once encryption is enabled on an ACL link, the encryption key size is
read from the controller, using hdev->min_enc_key_size in the meantime.
Data received before the reply is processed with that key size, so a
peer starting SMP over BR/EDR right after encryption, which it may do,
gets its Pairing Request rejected with "Encryption Key Size" whenever
the pending security level is FIPS, as that requires a 16 octet key:
> HCI Event: Encryption Change (0x08) plen 4
> ACL Data RX: Handle 1 flags 0x02 dlen 11
BR/EDR SMP: Pairing Request (0x01) len 6
@ MGMT Event: Authentication Failed (0x0011) plen 8
Status: Authentication Failed (0x05)
< HCI Command: Read Encryption Key Size (0x05|0x0008) plen 2
> HCI Event: Command Complete (0x0e) plen 7
Key size: 16
< ACL Data TX: Handle 1 flags 0x00 dlen 6
BR/EDR SMP: Pairing Failed (0x05) len 1
Reason: Encryption key size (0x06)
Mark the connection with HCI_CONN_ENC_KEY_SIZE_PEND while the key size
is read and hold incoming L2CAP data back in pending_rx meanwhile,
processing it once the reply is handled.
Found with mgmt-tester "Pairing Acceptor - SMP over BR/EDR 2" under User
Mode Linux time travel, where the emulated peer always replies before
the key size is read.
Fixes: e3f6a257a73d ("Bluetooth: Use actual encryption key size for SMP over BR/EDR")
Assisted-by: OpenCode:claude-opus-5.5
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
---
include/net/bluetooth/hci_core.h | 1 +
net/bluetooth/hci_event.c | 12 +++++++++++-
net/bluetooth/l2cap_core.c | 16 ++++++++++++++--
3 files changed, 26 insertions(+), 3 deletions(-)
diff --git a/include/net/bluetooth/hci_core.h b/include/net/bluetooth/hci_core.h
index 8f32c860d492..f6fa6a327e11 100644
--- a/include/net/bluetooth/hci_core.h
+++ b/include/net/bluetooth/hci_core.h
@@ -1017,6 +1017,7 @@ enum {
HCI_CONN_CREATE_PA_SYNC,
HCI_CONN_PA_SYNC,
HCI_CONN_PA_SYNC_FAILED,
+ HCI_CONN_ENC_KEY_SIZE_PEND,
};
static inline bool hci_conn_ssp_enabled(struct hci_conn *conn)
diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
index 56ad10ab96e3..565668d8d7cc 100644
--- a/net/bluetooth/hci_event.c
+++ b/net/bluetooth/hci_event.c
@@ -724,6 +724,8 @@ static u8 hci_cc_read_enc_key_size(struct hci_dev *hdev, void *data,
goto done;
}
+ clear_bit(HCI_CONN_ENC_KEY_SIZE_PEND, &conn->flags);
+
/* While unexpected, the read_enc_key_size command may fail. The most
* secure approach is to then assume the key size is 0 to force a
* disconnection.
@@ -3190,6 +3192,7 @@ static int hci_read_enc_key_size(struct hci_dev *hdev, struct hci_conn *conn)
{
struct hci_cp_read_enc_key_size cp;
u8 *key_enc_size = hci_conn_key_enc_size(conn);
+ int err;
if (!read_key_size_capable(hdev)) {
conn->enc_key_size = HCI_LINK_KEY_SIZE;
@@ -3211,7 +3214,14 @@ static int hci_read_enc_key_size(struct hci_dev *hdev, struct hci_conn *conn)
else
conn->enc_key_size = hdev->min_enc_key_size;
- return hci_send_cmd(hdev, HCI_OP_READ_ENC_KEY_SIZE, sizeof(cp), &cp);
+ /* L2CAP holds incoming data back until the key size is known */
+ set_bit(HCI_CONN_ENC_KEY_SIZE_PEND, &conn->flags);
+
+ err = hci_send_cmd(hdev, HCI_OP_READ_ENC_KEY_SIZE, sizeof(cp), &cp);
+ if (err)
+ clear_bit(HCI_CONN_ENC_KEY_SIZE_PEND, &conn->flags);
+
+ return err;
}
static void hci_conn_complete_evt(struct hci_dev *hdev, void *data,
diff --git a/net/bluetooth/l2cap_core.c b/net/bluetooth/l2cap_core.c
index 4dd56e0074fd..361ed3ebe19e 100644
--- a/net/bluetooth/l2cap_core.c
+++ b/net/bluetooth/l2cap_core.c
@@ -7389,7 +7389,11 @@ static void l2cap_recv_frame(struct l2cap_conn *conn, struct sk_buff *skb)
u16 cid, len;
__le16 psm;
- if (hcon->state != BT_CONNECTED) {
+ /* Also hold data back while the encryption key size is read, as
+ * the security of the link, e.g. for SMP over BR/EDR, depends on it.
+ */
+ if (hcon->state != BT_CONNECTED ||
+ test_bit(HCI_CONN_ENC_KEY_SIZE_PEND, &hcon->flags)) {
BT_DBG("queueing pending rx skb");
skb_queue_tail(&conn->pending_rx, skb);
return;
@@ -7452,7 +7456,11 @@ static void process_pending_rx(struct work_struct *work)
mutex_lock(&conn->lock);
- while ((skb = skb_dequeue(&conn->pending_rx)))
+ /* l2cap_recv_frame() would queue it all again while the encryption
+ * key size is read, l2cap_security_cfm() schedules this once known.
+ */
+ while (!test_bit(HCI_CONN_ENC_KEY_SIZE_PEND, &conn->hcon->flags) &&
+ (skb = skb_dequeue(&conn->pending_rx)))
l2cap_recv_frame(conn, skb);
mutex_unlock(&conn->lock);
@@ -8041,6 +8049,10 @@ static void l2cap_security_cfm(struct hci_conn *hcon, u8 status, u8 encrypt)
}
mutex_unlock(&conn->lock);
+
+ /* Process data held back while the encryption key size was read */
+ if (!skb_queue_empty(&conn->pending_rx))
+ queue_work(hcon->hdev->workqueue, &conn->pending_rx_work);
}
/* Append fragment into frame respecting the maximum len of rx_skb */
--
2.56.0
next prev parent reply other threads:[~2026-10-22 15:46 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-10-22 15:46 [PATCH v1] Bluetooth: btintel_pcie: Allow COMPILE_TEST without ACPI Luiz Augusto von Dentz
2026-10-22 15:46 ` Luiz Augusto von Dentz [this message]
2026-10-22 18:20 ` [v1] Bluetooth: L2CAP: Hold incoming data until encryption key size is known bluez.test.bot
2026-10-23 21:00 ` [PATCH v1] " patchwork-bot+bluetooth
2026-10-22 18:08 ` [v1] Bluetooth: btintel_pcie: Allow COMPILE_TEST without ACPI bluez.test.bot
2026-10-23 21:00 ` [PATCH v1] " patchwork-bot+bluetooth
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=20261022154621.3958803-2-luiz.dentz@gmail.com \
--to=luiz.dentz@gmail.com \
--cc=linux-bluetooth@vger.kernel.org \
/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